Breeding exceptionally fragrant soybeans for soy milk with strong aroma

Knockout of the soybean (Glycine max) betaine aldehyde dehydrogenase genes GmBADH1 and GmBADH2 using CRISPR/Cas12i3 enhances the aroma of soybeans. Soy milk made from the gmbadh1/2 double mutant seeds exhibits a much stronger aroma, which consumers prefer; this mutant has potential for enhancing quality in soy-based products. Soybean (Glycine max (L.) Merr.) is a major source of vegetable protein and oil in human diet and animal nutrition. Soybean seeds have been extensively used in various food products and snacks. Taste quality, particularly the aroma, affects cooking and eating, and ultimately influences consumer preference. Soy milk is particularly popular in China and has been gaining popularity in many other countries in the world. 2-Acetyl-1-pyrroline (2-AP) is a significant aromatic compound responsible for the delightful "popcorn-like" fragrance in crops. The production of this compound is affected by betaine aldehyde dehydrogenase (BADH), and genetic variations in this gene have been utilized in the breeding of several crops including rice, corn, sorghum and millet (Chen et al., 2008; Wang et al., 2021; Zhang et al., 2022, 2023). In soybeans, there are two homologs of rice OsBADHs, namely GmBADH1 and GmBADH2. It has been observed that natural aromatic soybeans owe their fragrant smell to variations in GmBADH2 (Juwattanasomran et al., 2012). Additionally, GmBADH2 could be knocked down in nonaromatic soybean varieties through RNA interference, resulting in the improvement of 2-AP levels (Arikit et al., 2011). GmBADH1 and GmBADH2 share 92% amino acid sequence similarity and their genes exhibit similar expression patterns as the homologs in rice, maize, and sorghum (Chen et al., 2008; Arikit et al., 2011; Wang et al., 2021; Zhang et al., 2022). However, it is currently unclear whether GmBADH1 affects the content of 2-AP in soybean, and whether knocking out two GmBADHs can further increase the content of 2-AP. We hypothesized that both GmBADH1 and GmBADH2 might be the candidate genes involved in the production of the 2-AP aroma. To test this hypothesis, we knocked out individual or both GmBADH1 and GmBADH2 in a nonaromatic elite soybean variety, Xudou20 (XD20), which is commonly grown in the Huang and Huai river basins of China. In our experiment, we utilized the GmEF1α promoter to drive the expression of Cas12i3 (Duan et al., 2024). We employed the GmU6 promoter to drive crRNA1 and the GmU3 promoter to drive crRNA2 (Figure 1A), which target the fourth exons of GmBADH1 and GmBADH2, respectively (Figure 1B, C). To assess the editing efficiency of Cas12i3 in soybean, we employed a hairy root test system, as previously described (Cao et al., 2022). We analyzed the GmBADH1 and GmBADH2 mutations in hairy roots by Sanger sequencing. In 20 transformed hairy roots examined, mutation frequencies were found to be 45.0% for GmBADH1 and 10.0% for GmBADH2 (Figure 1D). This result showed that the CRISPR/Cas12i3 gene editing system is active in soybean. Following the successful hairy roots experiment, the binary vector was introduced into XD20 through Agrobacterium tumefaciens-mediated stable transformation as previously described (Paz et al., 2006). As a result, we obtained 103 independent T0 transgenic lines, 24 of which were found to have mutations in either or both of the target genes (Figure 1E). CRISPR/Cas12i3-mediated editing of GmBADH1 and GmBADH2 in soybean enhances 2-AP content and aroma (A) Schematic of the CRISPR/Cas12i3 construct. (B) GmBADH1 crRNA design and mutant genotypes. (C) GmBADH2 crRNA design and mutant genotypes. (D, E) Editing efficiency in hair roots and T0 mutants. (F) Double mutant genotypes. (G) Photograph of XD20 at 16 DAE (days after emergence). (H) Photograph of gmbadh1-1 (left) and gmbadh1-2 (right) at 16 DAE. (I) Photograph of gmbadh2-1 (left) and gmbadh2-2 (right) at 16 DAE. (J) Photograph of gmbadh1/2 double mutant #3 (left) and #6 (right) at 16 DAE. (K) Photograph of XD20 at 74 DAE. (L) Photograph of gmbadh1/2 double mutant #3 (left) and #6 (right) at 74 DAE. (M) Photograph of XD20 at 98 DAE. (N) Photograph of gmbadh1/2 double mutant #3 (left) and #6 (right) at 98 DAE. (O–R) The plant height (O), flowering time (P), seeds weight (Q) and maturity time (R) of XD20 and the mutants. (S–U) The 2-AP content of dry leaves (S), dry seeds (T) and soybean milk (U) of XD20 and the mutants. (V) The number of volunteers choosing XD20 or the double mutant soy milk. Data (O–V) are shown as the means ± SE (n = 3). P-values were calculated using a t-test (V) or ANOVA (R–U). *P < 0.05. ****P < 0.0001. Red bases are target sequences and the underlined bases represent the protospacer adjacent motif (5′-TTN). Bars = 14 cm. We obtained T-DNA vector-free GmBADH mutant lines in the T1 generation. The gmbadh1-1 and gmbadh1-2 lines both had 11-bp deletions in GmBADH1, while the gmbadh2-1 line had an 11-bp deletion in GmBADH2 and the gmbadh2-2 line harbored a 40-bp deletion and an 1-bp insertion in GmBADH2. The gmbadh1/2 double mutant line #3 contained an 11-bp deletion in GmBADH1 and a 14-bp deletion in GmBADH2, while the gmbadh1/2 double mutant line #6 contained a 14-bp deletion in GmBADH1 and an 11-bp deletion in GmBADH2 (Figure 1B–D). All of these mutations resulted in gene knockout or alteration of gene products of GmBADHs. The soybean plants were cultivated in a climate chamber under conditions of 14 h/10 h (light/dark) lighting, with temperatures set at 25°C/20°C (light/dark). We observed no obvious differences in vegetative or reproductive growth in the single or double mutant lines compared with the wild type (Figure 1G–N). There were no obvious changes in plant height, flowering time or total seed weight, with the exception of a delay in maturation time of approximately 5 d in the double mutant lines compared with the wild type (Figure 1O–R). The 2-AP content in the soybean plants was quantified using HS-SPME-GC-MS. The analysis revealed that the gmbadh2 but not the gmbadh1 single mutant lines exhibited an increased 2-AP content. Interestingly, 2-AP content in the double mutants was more than four times as high as those in the gmbadh2 single mutants, both in the leaves and seeds (Figure 1S, T). These results suggested that the gmbadh1/2 double mutations caused a large enhancement in 2-AP content, supporting the idea that GmBADH1 also contributes to the control of the 2-AP content in soybean. The lack of a 2-AP increase in the gmbadh2 single mutant lines, but a very large increase in gmbadh1/2 double mutant lines, suggested that the two GmBADH genes had redundant roles. Our results showed that while GmBADH2 plays a predominant role in controlling 2-AP content in soybean, GmBADH1 clearly also contributes significantly to this control. Soybean is widely used to produce soy milk due to its pleasant taste and health benefits. As anticipated, soy milk made from the gmbadh1/2 double mutant seeds exhibited a much stronger aromatic smell that was consistent with its much higher 2-AP level (Figure 1U). This heightened aroma likely enhances the sensory experience of consuming soy milk, which is supported by the preferences of volunteers when given the choice between soy milk made from nonaromatic soybeans and that produced from the gmbadh1/2 soybean seeds (Figure 1V). According to a previous report, 2-AP content in the seeds of existing fragrant soybeans ranged between 0.05 and 0.5 mg/kg (Zhang et al., 2021). In contrast, the gmbadh1/2 seeds generated in this study had a marked 2-AP content of 1.59 mg/kg. This study demonstrates that by mutating both GmBADH1 and GmBADH2 in nonaromatic soybean varieties through gene editing, extraordinary aromatic soybeans can be generated for enhanced taste quality of soybean and soybean products. This work was supported by the National Natural Science Foundation of China (32188102 to J.-K.Z.), the Key R&D Program of Shandong Province, China (2021LZGC012-004 to H.X.), and Bellagen Biotechnology Co. Ltd., Jinan, China. The authors declare no conflict of interest. H.X., X.N., and J.-K.Z. conceived the project; M.S., X.C., S.L., X.W., and H.X. performed the experiments and collected the data; H.X., M.S., X.C., Q.N. J.Z., and J.-K.Z. wrote the manuscript. All authors read and approved its content. Additional Supporting Information may be found online in the supporting information tab for this article: http://onlinelibrary.wiley.com/doi/10.1111/jipb.13631/suppinfo Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.