Transgene‐free CRISPR/Cas9‐mediated gene editing through protoplast‐to‐plant regeneration enhances active compounds in Salvia miltiorrhiza

Salvia miltiorrhiza (red sage, Chinese pinyin; danshen) is used in Eastern medicine to treat cardiovascular diseases. S. miltiorrhiza contains water‐soluble and lipid‐soluble bioactive compounds, including phenolic acids and diterpenoid tanshinones, respectively (Shi et al., 2021); the latter gives its root surface a red colour (Skała and Wysokińska, 2005).

Several studies have sought to inactivate specific biosynthetic or transcription factor genes related to the bioactive compounds in S. miltiorrhiza by introducing a clustered regularly interspaced short palindromic repeat (CRISPR)‐/CRISPR‐associated nuclease 9 (Cas9)‐based genome editing cassette via Agrobacterium‐mediated hairy root transformation (Deng et al., 2020). Nevertheless, chimaeras in transformation and removing transgenes in plants with high‐genetic heterozygosity like S. miltiorrhiza present significant challenges (Su et al., 2023).

Here, we established a protoplast regeneration system for S. miltiorrhiza using either in vitro‐assembled sgRNA‐Cas9 ribonucleoprotein (RNP) complexes or plasmids carrying CRISPR/Cas9 system genes to target one or multiple sites for editing the genes through a single transfection event. As transcription factors regulating entire metabolic pathways are generally recognized as valuable tools for engineering elevated metabolite levels (Broun and Somerville, 2001), seven transcription factor genes—MYB28, MYB36, MYB39, MYB100, basic leucine zipper 1 (bZIP1), bZIP2 and MYB98—were selected as targets for mutagenesis. Of the 23 target sites selected (Table S1), three were previously reported for bZIP1, bZIP2 and MYB98 (Deng et al., 2020; Hao et al., 2020; Shi et al., 2021). We successfully established our S. miltiorrhiza protoplast regeneration system, which requires ~6 months from transfection of protoplasts to whole plants (Figure 1a; Data S1).

Figure 1.

Regeneration of transgene‐free gene‐edited S. miltiorrhiza protoplasts into whole plants. (a) Establishment of a protoplast‐to‐plant regeneration system. Protoplasts were isolated from in vitro‐grown plantlets through enzymatic leaf digestion; transgene‐free CRISPR/Cas9 gene editing reagents were then introduced (Day 2). Alginate‐Ca²⁺ gel with embedded protoplasts was incubated in a liquid medium in the dark (Week 3). Calli >5 mm were transferred to a fresh solid medium containing thidiazuron and incubated in light (Month 2). Multiple bud formation was observed in light (Month 3). Roots differentiated from calli with multiple buds in a phytohormone‐free medium in light (Month 4). Plantlets transferred to pots (Month 6). (b, c) Editing efficiencies (number of validated regenerated gene‐edited plants, in parentheses), as a percentage of total plants analysed, for individual target sites in bZIP1, bZIP2 and MYB98 transfected with plasmid (b) and in MYB36, MYB39, MYB28 and MYB100 transfected with RNP (c). Total, total plants analysed; T1–T5, target sites 1–5; edited, validated gene‐edited plants; null, knockout plant confirmed by Sanger sequencing. (d, e) Characterization of 12 regenerated plants with MYB28 edited through transgene‐free CRISPR‐/Cas9‐mediated mutagenesis. (d) Pre‐validation of edited target sites 3 and 4 by cleavage with their respective RNP complexes in vitro. PCR‐amplified MYB28 DNA fragments were incubated with RNP complex and then separated by electrophoresis. M, DNA size marker; WT, wildtype; +, −, WT genomic DNA incubated with and without RNP complex, respectively. White and black arrowheads indicate sizes of edited genomic DNA (cannot be cleaved by the RNP complex) and unedited genomic DNA after cleavage by the RNP complex, respectively. (e) RT‐PCR analysis of MYB28 expression in six independent myb28 knockout plants. Actin served as a reference transcript. (f) Representative vegetatively propagated plant myb28#4‐3 from stem cutting of myb28#4. (g) Genome editing of regenerated myb28 plants was stably inherited by their vegetatively propagated derivatives. (h) Phenotypes of 3‐month‐old wild‐type and knockout mutants cultivated in the growth chamber. (i) Active compound contents in root extracts (mean ± SD; n = 4). Different letters signify statistically significant differences (P < 0.05). (j, k) Representative images of flowers from wild‐type and null mutants (j) and white flowers of three clonally propagated plants from two independent regenerated knockout mutants of myb36 (#20, #62; k).

We determined the mutagenesis efficiency for each target site in regenerated plants by analysing the size of PCR products containing the sites (with smaller products indicative of deletion) or by endonuclease cleavage of PCR products (Figure 1b–d; Table S2). We validated any mutation using Sanger sequencing (Data S2). However, some of the regenerated gene‐edited plants using plasmids were not transgene‐free according to PCR analysis (Figure S1), indicating transgene‐free gene editing should be performed by RNP.

We next checked the expression of each targeted gene in the respective putative knockout plants (Figures 1e and S2; Table S3) and assessed whether the mutagenized site was inherited via vegetative propagation of stem cuttings (Figures 1f,g and S3). We focused on the knockout of MYB28 by multiplex editing with RNP. Half the plants tested carried a homozygous MYB28 mutation as determined by in vitro RNP cleavage (Liang et al., 2018) and Sanger sequencing (Figure 1c,d and Data S2). RT‐PCR indicated that MYB28 was barely expressed in the mutagenized plants (Figure 1e). Five individuals propagated from stem cuttings showed an in vitro RNP cleavage pattern identical to that of the original regenerated plant (Figure 1f,g), indicating stable transmission of gene editing through vegetative propagation. Target gene expression levels were comparable in the original regenerated plant and vegetatively propagated derivatives (Figure S4).

We examined the phenotypes of the knockout plants (Figure 1h,i). MYB98 is a positive regulator of the biosynthetic pathways of phenolic acids and tanshinones (Hao et al., 2020). Knockout plant myb98#17, which was homozygous for a mutation in MYB98, exhibited a paler red root colour and had lower levels of most water‐ and lipid‐soluble compounds than the wild type (Figure 1h,i). bZIP1 is a negative regulator of tanshinone biosynthesis and a positive regulator of phenolic acid biosynthesis (Deng et al., 2020), whereas bZIP2 is a negative regulator of phenolic acid biosynthesis (Shi et al., 2021). Plants harbouring knockout mutations of either gene had deeper red roots, with higher levels of lipid‐soluble compounds (Figure 1h,i). Overexpressing MYB36 in hairy roots promotes tanshinone accumulation while decreasing phenolic acid levels (Li et al., 2022). However, MYB36 is primarily expressed in flowers and is homologous to some Arabidopsis MYB genes (Li and Lu, 2014). Our regenerated MYB36 knockout plants and vegetatively propagated derivatives all displayed white flowers (Figure 1j,k), indicating that MYB36 may function as a positive regulator of anthocyanin biosynthesis in flowers.

More broadly, we demonstrated high‐efficiency gene editing of both gene copies of various transcription factor genes regulating the biosynthesis of bioactive compounds in S. miltiorrhiza through a single transfection event using transgene‐free CRISPR/Cas9 reagents in protoplasts, and regeneration of knockout plants through a newly established protoplast‐to‐plant regeneration system. This paves the way for enhancing the contents of active compounds in heterozygous transformation‐recalcitrant medicinal plants.

Conflict of interest

The authors declare no conflict of interest.

Author contributions

HJH, SZL, CCC and CSL conceived the research. CTH, CSL, PYH and FHW performed the protoplast regeneration and gene editing. SC, CYL, YCL and YLY performed bioactive ingredient analysis. CCC and CSL analysed the data. CCC, CSL, PYH, HJH and SZL prepared the manuscript.

Supporting information

Data S1 Materials and methods.

Data S2 Results of sanger sequencing.

Table S1 Sequences of the 23 target sites used in this study.

Table S2 Primer sequences used to amplify genomic fragments spanning the target sites by nested PCR.

Table S3 Primers used for RT‐PCR analyses.

Figure S1 Presence of Cas9 DNA fragment in some regenerated plants transfected by plasmids.

Figure S2 RT‐PCR analyses of the expression of the six genes in their corresponding regenerated gene‐edited plants.

Figure S3 Vegetatively propagated plants exhibit the same genotype as their respective original regenerated gene‐edited plants.

Figure S4 RT‐PCR analyses of vegetatively propagated plants derived from regenerated gene‐edited plants for bZIP2 and MYB98.

Data availability statement

The data that supports the findings of this study are available in the supplementary material of this article.