Boosting genome editing in perennial plants by CRISPR-Combo mediated morphogenic gene activation

Abstract

A major bottleneck in genome editing of many perennial plants is their recalcitrance to transformation and regeneration. To boost genome editing in such perennial crops, transcriptional reprogramming of morphogenic genes is introduced by CRISPR-Combo, a versatile system for simultaneous genome editing and transcriptional activation in plant cells. In potato, we screen 17 morphogenic genes and identify 4 genes (WOX11/12, ARF5, ABI3-1, and ABI3-2) that promote regeneration of genome-edited hairy roots, and 3 of the 4 genes are also found to boost shoot regeneration by Agrobacterium-mediated stable transformation. Similarly, screening of 10 morphogenic genes in citrus leads to the identification of 5 genes (BBM3, FUS3, IPT1, SERK1, and STM) that enhance plant regeneration upon activation. In wild strawberry, we demonstrate that simultaneous activation of Baby Boom genes (BBM1 and BBM2) or of GRF3 and GIF1 reduces the generation time of genome-edited plants by over one month. Moreover, in poplar, we show that simultaneous activation of WUS and WOX11 synergistically promotes plant regeneration without exogenous plant hormones, which leads to a protocol of generating genome-edited poplar shoots in less than one month. Collectively, this study provides efficient strategies for boosting genome editing in four perennial crops.

Perennial plants are often difficult to improve genetically because they are hard to transform and regenerate. Here, the authors used CRISPR-Combo-mediated activation of morphogenic genes to accelerate genome editing and plant recovery in potato, citrus, strawberry, and poplar.

Introduction

Over the past decade, CRISPR-Cas genome-editing tools have been developed and demonstrated for plants other than model plants like Arabidopsis and rice^1–3. However, the deployment of such genome-editing tools to non-model plants is often constrained by limitations in plant transformation and regeneration^4,5. Plant regeneration recalcitrance could be overcome by ectopic expression of morphogenic genes that promote somatic embryogenesis or organogenesis⁶. Well-characterized morphogenic genes include (i) members of the AP2/ERF TF family such as BABY BOOM (BBM) whose overexpression can induce organogenesis^7–9 and embryo formation^9–11, (ii) members of the homeobox TF family such as WUSCHEL (WUS) and WOX genes, which maintain stem cell populations in the meristems^12–22, (iii) GRF/GIF TF family members that stimulate meristem growth^23,24, (iv) AGL15 that promotes somatic embryogenesis via a cell-identity circuitry^25–27, RKD/PLT5 genes that induces embryogenesis or embryo-like structures^28,29, and (v) peptide factors such as Regeneration factor 1 (REF1)³⁰ which initiate wound-induced tissue repair and regeneration mediated by WIND1^31–33.

These morphogenic genes often regulate interconnected transcriptional regulatory networks and interactions with plant hormones to induce cell regeneration and cell fate transitions. Consequently, it is not surprising that the simultaneous expression of multiple morphogenic genes may result in more robust plant regeneration. For example, co-expression of BABY BOOM (BBM) and Wuschel2 (WUS2) promotes plant regeneration in many monocots, including maize, sorghum, sugarcane, teff, switchgrass, foxtail millet, pearl millet, and rye^34–36. Similarly, co-expression of a GRF-GIF gene pair promoted shoot regeneration in monocots, including wheat^37,38, maize³⁹ and sorghum⁴⁰, and the dicots citrus³⁷ and tomato⁴¹. In nearly all these studies, these morphogenic gene cassettes were stacked with the CRISPR-Cas9 cassette to support the regeneration of genome-edited plants.

We previously developed a CRISPR-Combo system that confers simultaneous genome editing and gene activation^42–44, representing a great upgrade from the CRISPRa systems^45,46. CRISPR-Combo is a powerful system for promoting the regeneration of genome-edited plants via morphogenic gene activation, as we demonstrated in rice and poplar⁴². Here, we sought to extend CRISPR-Combo to support genome editing experiments in perennial plants because of several features of this system. First, we aimed to demonstrate that CRISPR activation (CRISPRa), a major built-in component of CRISPR-Combo, is a highly efficient approach for systematically screening a list of morphogenic genes for their potential to promote plant regeneration in individual species. Second, CRISPR-Combo is highly amenable to multiplexed targeting via programmed guide RNAs (gRNAs), enabling us to easily build multiplexed gene activation systems and test their efficacy in achieving robust plant regeneration. Third, CRISPR-Combo relies on a single Cas9 protein for genome editing and gene activation. Due to this feature, plants regenerated through morphogenic gene activation may enrich genome-editing events⁴². Fourth, using CRISPRa rather than conventional overexpression cassettes helps streamline vector construction, as only gRNAs that target morphogenic genes need to be expressed. Fifth, gene activation by CRISPRa acts on the endogenous chromatin environment, which may help alleviate potential detrimental or pleiotropic phenotypes associated with the conventional morphogenic gene overexpression cassettes^7,21,34,47. Finally, the conventional approach of ectopic expression of morphogenic genes has not been widely used to promote regeneration in perennial plants. We thus sought to boost genome editing in perennial plants by simultaneously activating morphogenic genes to benchmark this powerful approach.

Results

Screening morphogenic genes in potato using a hairy root system

To assess the effectiveness of the CRISPR-Combo system to screen multiple morphogenic genes, we used an efficient Rhizobium rhizogenes-mediated hairy root system. Although hairy root induction and shoot organogenesis are fundamentally distinct developmental processes, we reasoned that the hairy root system could serve as a robust functional filter to identify genes with broad reprogramming potential. We prioritized candidate morphogenic genes based on their conserved roles in regulating plant developmental plasticity, focusing on five key regulatory layers: (i) embryogenic regulating factors (e.g., BABY BOOM/BBM, LEC1, and ABI3/FUS3 families); (ii) meristematic stem cell factors (e.g., WUSCHEL/WUS and WOX homeobox genes); (iii) hormone signaling and biosynthesis mediators (e.g., ARF5/MP and IPT); (iv) somatic embryogenesis receptor kinases (e.g., SERK1); and (v) wound-induced reprogramming factors (e.g., WIND1). By selecting representative paralogs from these interconnected networks and applying our screening framework, we aimed to efficiently identify endogenous morphogenic genes whose CRISPR-Combo-mediated activation would help overcome recalcitrance in species-specific regeneration, characteristic of many crop species. Hence, we first tested the hairy root system in potato⁴⁸, which is a vegetatively propagated herbaceous perennial species, but grown as an annual crop. Based on a comprehensive analysis (see “Methods”), we selected 17 morphogenic genes for activation and 5 target genes for editing (Supplementary Table 1). These two functionalities were programmed with two different gRNA systems: gRNA1.0 with 20-nt spacers for genome editing and gRNA2.0 with 15-nt spacers for gene activation (Supplementary Fig. 1a). Each of the CRISPR-Combo T-DNAs contains 4 gRNAs, with 3 gRNAs dedicated for gene activation and the remaining gRNA for genome editing (Supplementary Fig. 1b, and Supplementary Table 2). Scrambled gRNAs were used to construct the CRISPR-Combo control vector. These T-DNA vectors were divided into two batches based on their ability to induce hairy roots upon transformation with R. rhizogenes (Supplementary Fig. 2a, b). From this screen, four morphogenic genes (StWOX11/12, StARF5, StABI3-1, and StABI3-2) were identified to promote hairy root formation compared to the Combo control (Supplementary Fig. 2c, d).

Our initial screening results, albeit promising, were from vectors that contained different genome editing targets. To further validate the results, we generated 4 T-DNA CRISPR-Combo vectors targeting the same gene (StVINV, encoding a vacuolar Invertase) with the same gRNA, while the same 3 gRNAs per gene were used to activate each of the 4 morphogenic genes: StWOX11/12, StARF5, StABI3-1, and StABI3-2. A CRISPR-Cas9 vector targeting StVINV alone was added as an additional control. Our data confirmed that activation of these 4 morphogenic genes significantly increased hairy root production per explant as compared to the Combo control (Fig. 1a, b). Expression analysis of the transgenic hairy roots by qRT-PCR showed that the targeted morphogenic genes were indeed activated (Fig. 1c–f). Genome editing efficiency was also assessed by deep sequencing of PCR amplicons. The sequence data showed that the average editing efficiencies of over 60% for the CRISPR-Combo constructs were similar to the CRISPR-Cas9 control (Fig. 1g). Further sequence analysis showed that up to 3 of the 4 copies of StVINV in tetraploid potato were edited in some transgenic hairy roots (Fig. 1h), suggesting efficient editing of the tetraploid genome of potato with this transient assay. Together, these data indicate that the R. rhizogenes-mediated hairy root system can be used for a CRISPR-Combo-based screen of morphogenic genes in potato.

Fig. 1. Hairy root-based screening and stable transformation-mediated functional validation of promising morphogenic genes using the CRISPR-Combo system.

a, b Rhizobium rhizogenes mediated transformation of potato with the CRISPR-Combo system showed that the activation of morphogenic genes (StWOX11/12, StARF5, StABI3-1, and StABI3-2) promoted the induction of hairy roots compared to the combo control. a Representative hairy roots transformed by different constructs 30 days after transformation (DAT). Scale bars, 1 cm. b Quantitative analysis of hairy root production at two time points (30 days and 40 days) after transformation. Error bars represent ± standard error of the mean (n = 4 independent plates, and each plate contains 10–16 explants). c–f Confirmation of target gene activation by qRT-PCR. Error bar, mean ± s.d. (n = 4 independent replicates). g Indel frequency at the target gene, StVINV, in potato hairy roots by different constructs. Error bars signify mean ± s.d. (n = 5 independent replicates). Each dot represents an individual potato hairy line. The maxima, center, and minima of the box refer to the upper quartile, median, and lower quartile. h Representative genotypes of CRISPR-Combo-mediated T₀ hairy root mutants at the StVINV site in potato. The red dashes and letters indicate a nucleotide deletion and insertion, respectively. The bold letter indicates PAM (protospacer adjacent motif), and the protospacer sequences are underlined. i Agrobacterium-mediated stable transformation, three of the four promising constructs identified in the hairy root screen (StWOX11/12, StARF5, and StABI3-2) significantly enhanced shoot induction compared to the Combo control (representative image after 45 DAT). j, Average regeneration efficiency of promising constructs quantified at three points (25-, 35-, and 45 DAT). Error bars signify mean ± s.d. (n = 3 independent plates, and each plate contains 12-18 explants). Scale bars, 1 cm. p values in b–g, j were obtained using the two-tailed Student’s t-test comparison of treatments against the Combo control. ^*p < 0.05, ^**p < 0.01. Source data are provided as a Source Data file.

Evaluation of candidate morphogenic genes in potato stable transformation

Plant regeneration in potato relies heavily on efficient shoot organogenesis, which remains a key bottleneck in stable transformation systems. To determine whether morphogenic genes identified from the hairy root screen could enhance shoot regeneration, we performed Agrobacterium tumefaciens-mediated transformation of potato explants. Stable integration of the T-DNA in the potato lines was verified by PCR amplification of the nptII selection marker (Supplementary Fig. 3). Consistent with the hairy root results, activation of selected morphogenic genes promoted shoot regeneration efficiency in potato explants (Fig. 1i, and Supplementary Fig. 4a). Notably, StWOX11/12, StARF5/MP and StABI3-2 showed increase in regeneration efficiency compared to the Combo control in multiple time points (25, 35, and 45 days after transformation; DAT) (Fig. 1j). Overall, our CRISPR-Combo constructs improved the regeneration efficiency (45-70%), compared to the Combo-control (~30–35%). In contrast, StABI3-1 exhibited a comparatively lower regeneration efficiency, despite showing positive activity in the hairy root assay. This difference could be explained by potential functional divergence among ABI3 paralogs, with StABI3-1 possibly preferentially regulating embryogenic or seed-associated developmental programs rather than shoot organogenesis under tissue culture conditions. Importantly, amplicon sequencing confirmed gene editing in regenerated potato plants by these three promising constructs (Supplementary Fig. 4b,c). Together, these results provide supporting evidence that CRISPR-Combo-mediated activation of endogenous morphogenic genes can improve regeneration in potato stable transformation, while also highlighting gene-specific and context-dependent differences in morphogenic potential in different starting tissues.

Screening morphogenic genes in citrus using a hairy root system

Encouraged by the potato data, we next sought to use CRISPR-Combo to screen morphogenic genes in citrus, a perennial fruit tree in which plant transformation and regeneration remain challenging. Ten morphogenic genes were selected for CRISPR activation (“Methods”, Supplementary Table 3). Twenty CRISPR-Combo vectors were constructed, with each expressing 4 gRNAs: 2 gRNAs-2.0 with 15-nt spacers for morphogenic gene activation and 2 gRNAs-1.0 with 20-nt spacers for genome editing of CsNPR3⁴⁹ (Supplementary Table 4). These T-DNA vectors, together with CRISPR-Cas9 and CRISPR-Combo (with scrambled gRNAs for activation) controls, were evaluated for R. rhizogenes-mediated hairy root production in citrus⁵⁰. Our initial screen discovered potential lethality for the vectors that edit CsLOB1⁵¹. We subsequently focused on the group of 10 CRISPR-Combo vectors in which CsNPR3 is the target for editing. Five morphogenic genes (CsBBM3, CsFUS, CsIPT, CsSERK1L, and CsSTM) conferred more production of hairy roots upon activation (Fig. 2a), which was statistically significant compared to controls (Fig. 2b). By contrast, activation of two other BBM genes (CsBBM1 and CsBBM2) and three WUS genes (CsWUS1, CsWUS2, and CsWUS3) did not enhance hairy root formation and resembled control levels (Fig. 2a, b). These differences among closely related family members may reflect divergence in their regulatory roles, expression patterns, or downstream interaction, although the underlying mechanisms remain to be fully elucidated in this system. Further, repeated experiments focusing on the five promising CRISPR-Combo vectors confirmed these observations (Supplementary Fig. 5). Genotyping of the resulting hairy roots showed that CsNPR3 was edited (Supplementary Fig. 6), supporting that the enhanced hairy root production by these constructs is likely due to in planta expression of CRISPR-Combo.

Fig. 2. CRISPR-Combo-based screening of morphogenic genes via the citrus hairy root system.

a, b Rhizobium rhizogenes mediated transformation of citrus with the CRISPR-Combo system identified five out of ten morphogenic genes (CsBBM3, CsFUS3, CsIPT, CsSERK1L, and CsSTM) whose activation promoted the generation of more hairy roots compared to the controls. a Representative images of hairy root formation by different constructs, 45 days after transformation. Scale bars, 1 cm. b Quantitative analysis of hairy root production at three time points (30-, 45- and 60-days) after transformation. Error bars represent ± standard error of the mean (n = 15, 16, and 17 independent biological replicates for the CsBBM1/CsNPR3, CsBBM3/CsNPR3, CsFUS3/CsNPR3, and CsIPT/CsNPR3 treatments; the Cas9/CsNPR3, CsBBM2/CsNPR3, CsSERK1L/CsNPR3, CsSTM/CsNPR3, CsWUS1/CsNPR3, CsWUS2/CsNPR3 and CsWUS3/CsNPR3 treatments; and the Combo control, respectively). p values were calculated by the two-tailed Student’s t-test comparison of treatments against the Combo control. ^*p < 0.05, ^**p < 0.01. Source data are provided as a Source Data file.

We next tested these five CRISPR-Combo vectors in an ex vivo system that induces hairy roots from citron petioles. Activation of these morphogenic genes all resulted in rapid induction of hairy roots upon R. rhizogenes inoculation (Fig. 3a). The numbers of hairy roots per plant leaf were significantly increased with about threefold more roots production than the controls for CRISPR-Combo vectors that activate CsFUS3, CsSERK1L, and CsSTM (Fig. 3b). Interestingly, more genome-edited hairy roots were detected in the five CRISPR-Combo constructs than the controls (Fig. 3c). Gene expression analysis by qRT-PCR confirmed effective transcriptional activation of the target morphogenic genes, with up to sixfold enhancement (Fig. 3d). We found that variability in morphogenic gene expression levels across independent hairy root lines, which could be attributed to the inherent heterogeneity of independent transformation events, including variations in T-DNA copy number and position effects within the native chromatin environment.

Fig. 3. Ex vivo hairy root transformation of citron with promising CRISPR-Combo constructs.

a, b Testing promising morphogenic genes (CsBBM3, CsFUS3, CsIPT, CsSERK1L, and CsSTM) and their efficacy in another cultivar (i.e., citron variety) showed significant enhancement in hairy root induction using Rhizobium rhizogenes-mediated transformation. a Representative images of hairy root formation by different constructs, 30 days after transformation. Scale bars, 1 cm. b Quantitative analysis of hairy root (HR) per plant. Error bars represent ± standard error of the mean (n = 19, 20, and 22 independent biological replicates for the Combo (control), the CsBBM3/CsNPR3, CsFUS3/CsNPR3, CsIPT/CsNPR3, CsSERK1L/CsNPR3, and CsSTM/CsNPR3 treatments, and the Cas9/CsNPR3 treatments, respectively). c Indel frequency at the target gene, CsNPR3, in citrus hairy roots by different constructs. d Activation of morphogenic genes verified by qRT-PCR in citrus hairy root transformation. Error bar, mean ± s.d. (n = 5 independent biological replicates). p values in b, d were calculated by the two-tailed Student’s t-test comparison of treatments relative to Combo control ^*p < 0.01, ^** p < 0.001. Source data are provided as a Source Data file.

Promoting citrus transformation via morphogenic gene activation

Plant regeneration based on organogenesis requires both shoot and root induction and shoot formation is often the most critical step^52,53. We next evaluated these morphogenic gene activation CRISPR-Combo constructs in a citrus stable transformation system, using Agrobacterium-mediated transformation of petiole explants. Strikingly, more shoots were induced by the CRISPR-Combo constructs (Fig. 4a). Based on the number of regenerated shoots 45 days after transformation (DAT), the control constructs generated regeneration efficiency under 60% (Fig. 4b). In contrast, all five morphogenic gene CRISPR-Combo constructs produced ~80% or greater regeneration efficiency, which is significantly higher than controls (Fig. 4b). The presence of the T-DNA in the regenerated citrus lines was verified by PCR amplification of the nptII selection marker, confirming stable integration across independent transgenic events (Supplementary Fig. 7). In the regenerated transgenic plants, genome editing by CsNPR3-gRNA1 and CsNPR3-gRNA2 was confirmed by amplicon sequencing (Fig. 4c–e), indicating that CRISPR-Combo was effective in regenerating more genome-edited plants by activation of morphogenic genes. Overall, hairy root-based screening followed by validation in stable transformation demonstrates that CRISPR-Combo-mediated activation of specific morphogenic genes effectively enhances and accelerates shoot regeneration. However, morphogenic genes that did not improve hairy root induction, such as CsBBM1/2 and CsWUS1/3, were not advanced to stable transformation experiments. Future studies evaluating these candidates across diverse tissue types will be important to fully assess the accuracy and tissue-specific limitations of the hairy root screening system.

Fig. 4. CRISPR-Combo enhances citrus regeneration and recovery of heritable targeted mutations.

a Representative images of Agrobacterium-mediated transformation showing promising CRISPR-Combo constructs promote shoot regeneration compared to the controls. Scale bars, 1 cm. b Average regeneration efficiency of different constructs by Agrobacterium-mediated transformation quantified at three points (25-, 35- and 45-days after transformation). All explants were selected under kanamycin plant selection and recovered under the same conditions. Error bars represent ± standard error of the mean (n = 3 independent plates, and each plate contains 25-30 explants). c, d Indel frequency at the two target sites in CsNPR3 among transgenic citrus plants. Error bars signify mean ± s.d. (n = 7, 8, and 10 independent biological replicates for the non-transgenic treatments, Combo control, and Cas9/CsNPR3, CsBBM3/CsNPR3, CsFUS3/CsNPR3, CsIPT/CsNPR3, and CsSERK1L/CsNPR3, and CsSTM/CsNPR3 treatments, respectively). Each dot represents an individual citrus transgenic line. The maxima, center, and minima of the box refer to the upper quartile, median, and lower quartile. e Representative genome editing genotypes of top CRISPR-Combo lines. The red dashes and letters indicate a nucleotide deletion and insertion, respectively. The bold letter indicates PAM (protospacer adjacent motif) and the protospacer sequences are underlined. p values at b–d were calculated by the two-tailed Student’s t-test comparison of treatments against the Combo control. ^*p < 0.05, ^**p < 0.01. Source data are provided as a Source Data file.

Fast-track production of genome-edited strawberry plants by CRISPR-Combo-mediated combinational morphogenic gene activation

Identification of promising morphogenic genes provides opportunities for simultaneous activation by CRISPR-Combo. Previously, robust plant regeneration was observed with co-expression of two morphogenic genes, such as BBM with WUS2^34–36 or GRF-GIF^37–41, in many plant species. We decided to test multiplexed morphogenic gene activation in strawberry, a perennial crop in which genome editing holds promising applications^54–56. In addition to multiplexed CRISPR-Cas9 genome editing constructs (Fig. 5a), we generated two multiplexed CRISPR-Combo constructs that simultaneously activate endogenous BBM1 and BBM2 genes (two gRNAs-2.0 per gene), while editing gene(s) (HB40 or NAC42A/NAC42B) with another two gRNAs (Fig. 5b). Similarly, five CRISPR-Combo constructs were made for simultaneous activating GRF3 and GIF1 (with two gRNAs-2.0 per gene), while editing a gene of interest (bHLH36, ATC, or AGL6) with another two gRNAs (Fig. 5c). It is of note that some constructs use two expression cassettes for the same gRNA to ensure robust editing outcomes (Fig. 5a, c). These constructs were introduced into a diploid wild strawberry (Fragaria vesca) via Agrobacterium-mediated transformation (Supplementary Table 5). Transgenic plants were confirmed by PCR-based genotyping. We found that co-activation of BBM1/BBM2 or GRF3/GIF1 enabled the production of well-developed transgenic plants in 4 months, shortening the transformation and regeneration timeline by 1 month (Fig. 5d, e). Because distinct target genes and gRNAs were used across the constructs, we could not directly compare genome editing efficiency. However, we found that genome editing efficiency of five gRNAs by the CRISPR-Cas9 constructs (JH19-1 and JH19-2) ranged from 3.45% to 45.83% (Fig. 5f). In contrast, genome editing efficiency of four gRNAs in the Combo-BBMs constructs ranged from 29.41% to 100% (Fig. 5f). Similarly, four of the five Combo-GRF3-GIF1 constructs generated genome editing efficiency ranging from 54.55% to 100%, while one gRNA (ATC-sgRNA1) failed to produce gene edits (Fig. 5f). We carried out independent transformation of these constructs and observed similar results (Supplementary Fig. 8). Collectively, these findings suggest that strawberry genome editing can be improved by accelerated plant regeneration via simultaneous activation of morphogenic gene pairs.

Fig. 5. Accelerated production of genome-edited strawberry plants by CRISPR-Combo-based multiplexed activation of morphogenic genes.

a In the JH19 control vector, the expression of Cas9 is driven by the Arabidopsis ubiquitin 10 promoter. A sgRNA1.0 scaffold is constitutively induced by either a Fragaria vesca U6 or Arabidopsis U6 promoter. Up to four sgRNAs in one vector are incorporated into the JH19. b, c Schematic diagram of the CRISPR-Combo vectors. Cas9-SunTag is driven by the Arabidopsis ubiquitin 10 promoter, and the activator complex is driven by the maize ubiquitin 1 promoter. The activator complex includes a 2xTAL activation domain (TAD) with a single-chain variable fragment of GCN4 antibody fused to a super-folder GFP. Two types of single guide (sgRNA) scaffolds sgRNA1.0 and sgRNA2.0 are driven by the Arabidopsis U3 promoter to edit and activate target genes, respectively. FveBBM1 and FveBBM2 are used as morphogenic genes in b, while FveGRF3 and FveGIF1 are used as morphogenic genes in c. d Regeneration time of transformed strawberry calli with different vectors. The Y-axis indicates the number of days from cotyledon infiltration by Agrobacterium to calli with shoot regeneration. Error bars represent ± standard error of the mean (n = 4, 4, and 5 biological replicates for JH19, Combo-BBMs, and Combo-GRF3-GIF1, respectively). p values were calculated by the two-tailed Student’s t-test comparison of combo against the JH19 control. ^*p = 0.0010 for comparison between JH19 and Combo-BBMs. ^*p = 0.0085 for comparison between JH19 and Combo-GRF3-GIF1. ^*p < 0.01. e Photographs of shoot regeneration from 4-month-old calli transformed with the JH19 control construct (above) and the Combo-BBMs-1 construct. f Editing efficiency in the T₀ transgenic strawberry plants with different vectors. Source data are provided as a Source Data file.

Improving poplar transformation and genome editing by CRISPR-Combo-mediated combinational gene activation

To further test whether combinatorial activation of morphogenic genes is a robust method for promoting plant regeneration, we extended our investigation to poplar. Previously, we showed that activation of WUS or WOX11 promotes plant regeneration in a hybrid poplar⁴². Based on this previous success, we wanted to know whether there are any additive or synergistic effects when both morphogenic genes are activated simultaneously. Furthermore, we wanted to know whether simultaneous activation of WUS and WOX11 can promote plant regeneration in the absence of exogenous plant hormones. To this end, we conducted hormone-free poplar transformation by comparing a CRISPR-Combo construct with WUS & WOX11 activation to single-gene-activation CRISPR-Combo constructs and a CRISPR-Cas9 genome-editing control construct (Supplementary Table 6). Compared to single-gene activation, simultaneous activation of WUS and WOX11 led to the generation of more shoots and roots, based on the analysis of all calli and regenerated T₀ shoots (Fig. 6a–d), in the absence of exogenous plant hormones. Examination of the regenerated lines from the combinational activation CRISPR-Combo construct showed that WOX11 and WUS were highly activated in the majority of 10 randomly selected T₀ lines (Fig. 6e, f), with the transgene confirmed by PCR (Supplementary Fig. 9). Remarkably, these lines also exhibited the highest biallelic editing efficiency at the target gene, 4CL1 (Fig. 6g).

Fig. 6. Simultaneous activation of WOX11 and WUS enhances poplar tissue culture regeneration and gene editing efficiency under hormone-free conditions using the CRISPR-Combo system.

a, b Shoot regeneration from poplar calli transformed with four different CRISPR-Combo constructs targeting gene editing of 4CL1 and activation of WOX11 and/or WUS. Scale bars, 1 cm. Regeneration was assessed after six weeks on hormone-free DKW medium. Data are presented as mean values ± SEM (n = 3 biological replicates). One-way ANOVA followed by Tukey’s multiple-comparison test was used. Bars with different letters (a, b) differ significantly (p < 0.05). Exact P values were summarized in Supplementary Table 8. All statistical tests were two-sided. c, d Rooting performance of all T₀ regenerated shoots transferred to hormone-free ½-strength LS medium for one week. Scale bars, 1 cm. One-way ANOVA followed by Tukey’s multiple-comparison test was used. Bars with different letters (a, b) differ significantly (p < 0.05). Data are presented as mean values ± SEM (n = 3 biological replicates). One-way ANOVA followed by Tukey’s multiple-comparison test was used. Bars with different letters (a–c) differ significantly (p < 0.05). Exact P values were summarized in Supplementary Table 9. All statistical tests were two-sided. e, f Relative expression levels of WOX11 and WUS in 10 randomly selected CRISPR-Combo (Cas9-GE(4CL1)+Act (WOX11 + WUS)) T₀ lines. Gene expression was quantified via qRT-PCR. Data are presented as mean values ±SEM (n = 3 technical replicates). Student’s t-tests were performed to compare expression levels between activation lines and control lines (CTRL) and exact P values were indicated above the bars. All statistical tests were two-sided. g Genotyping and zygosity analysis of 4CL1 gene in 10 randomly selected CRISPR-Combo co-activation lines, showing proportions of biallelic edit (indel frequency>70%), monoallelic edit (indel frequency >30% to 70%) and chimeric edit (indel frequency 1%-30%). h Schematics illustrating an improved poplar transformation and genome editing pipeline by simultaneous activation without the need for exogenous hormones. Created in BioRender. Qi, Y. (2026) https://BioRender.com/wxxzm79. Source data are provided as a Source Data file.

Based on these data, we proposed an improved genome-editing pipeline for poplar (Fig. 6h). In the standard protocol, regular CRISPR-Cas constructs are introduced into poplar callus tissues cultured on shoot-induction medium containing exogenous plant hormones, which usually takes 10–12 weeks for shoot regeneration (Fig. 6h). In the improved protocol, CRISPR-Combo WUS & WOX11 activation constructs combined with gRNAs for genome editing, can reduce shoot regeneration time to 4 weeks while also enhancing genome editing efficiency of the target gene (Fig. 6h). Impressively, these improvements were achieved by using shoot-induction medium without exogenous plant hormones.

To assess whether the CRISPR activation of WUS and WOX11 causes any pleiotropic effects, transgenic plants were closely monitored in the greenhouse. No visible defects were observed in overall plant architecture or in above- and below-ground tissue organization compared with controls (Fig. 7a, b). Interestingly, CRISPR-Combo WUS & WOX11 activation lines are significantly taller than the controls (Fig. 7a, c). This phenotype is consistent with the enhanced biomass observed in shoot and root tissues of these lines (Fig. 7d, e). Hence, simultaneous and continuous activation of WUS and WOX11 seems to augment the growth benefit in poplar that was already observed in the single gene activation lines⁴², suggesting additive or synergistic effects in promoting plant growth.

Fig. 7. Simultaneous activation of WOX11 and WUS greatly enhances biomass production in poplar.

Representative images of CRISPR-Combo (Cas9-GE(4CL1)+Act (WOX11 + WUS) T₀ plants after 3 months of growth in the greenhouse conditions. Scale bars, 10 cm (a) and 5 cm (b). c–e Quantification of plant height, shoot biomass, and root biomass in T₀ lines grown in the greenhouse. Data are presented as mean values ± SEM (n = 3 biological replicates). One-way ANOVA followed by Tukey’s multiple-comparison test was used. Bars with different letters (a, b) differ significantly (p < 0.05). Exact P values were summarized in Supplementary Tables 10–12. All statistical tests were two-sided. Source data are provided as a Source Data file.

Discussion

Harnessing the power of morphogenic genes to reprogram plant cell regeneration and development can help overcome the plant transformation bottleneck in achieving genome editing in many plants^{3,5,6,34,36,37,57}. In this study, we demonstrated CRISPR-Combo as a powerful tool to efficiently screen morphogenic genes that could enhance plant regeneration and genome editing. We identified different morphogenic genes capable of promoting organogenesis upon activation in four perennial plant species. The facile and effective activation of these morphogenic genes through CRISPR-Combo directly improves genome editing pipelines that rely on Agrobacterium-mediated plant transformation and regeneration. We have demonstrated this in potato, citrus, strawberry, and poplar. Moreover, we show that combinational activation of morphogenic genes holds great promise for further improving plant regeneration and genome editing, particularly in perennial crops.

Conventionally, gene stacking has been used to express exogenous morphogenic genes either with constitutive promoters or inducible promoters^34–41. This approach, while widely used, has its limitations, such as spatially and temporally uncontrolled expression, pleiotropic and developmental abnormalities, large vector sizes (especially when combinatorial morphogenic gene expression is used), and limited scalability for large-scale gene screening. By contrast, our CRISPR-Combo approach seeks to activate endogenous morphogenic genes in their native chromatin environments. Based on this fundamental difference, CRISPR-Combo may be advantageous in multiple aspects, including (1) rapid screening of many putative morphogenic genes via a gRNA library, (2) high flexibility for multiplexed gene activation, (3) simultaneous and efficient genome editing via hormone-free plant regeneration, and (4) streamlined and compact vector construction processes as gene activation and genome editing are both programmed by gRNAs.

Although boosting plant regeneration with ectopic expression of morphogenic genes is an established strategy across plant species^{3,5,6,34,36,37,57}, our approach differs in that CRISPR-Combo activates endogenous morphogenic genes to enhance regeneration of genome-edited plants. The species-specific effects of morphogenic genes observed in our results reflect differences in endogenous regulatory networks, chromatin context, or regenerative competence. This suggests that the phenotypic outcome of a given morphogenic gene activation depends not only on its intrinsic CRISPRa activity but also on the gene-specific native regulation confined by the spatiotemporal dynamics in plant developmental programming and the epigenetic chromatin status. All these features of CRISPR-Combo or CRISPRa differ from the conventional morphogenic gene expression methods, either with constitutive promoters or inducible promoters. Due to the nature of CRISPR activation, we might have missed achieving positive results on certain morphogenic genes due to gRNA design and/or intrinsic regulation of the native genes. In this regard, a conventional overexpression approach may be explored to investigate whether such genes, when overexpressed or highly induced, could promote plant regeneration. Recently, an inducible CRISPRa system has been adopted to screen morphogenic genes to augment plant transformation in multiple plant species⁵⁸. However, our CRISPR-Combo system simultaneously promotes morphogenic gene activation while achieving or enhancing genome editing efficiency. All other systems still must stack the morphogenic expression/activation cassettes with the genome editing cassettes, making them cumbersome and less versatile.

We anticipate several exciting future directions for improving genome editing using this powerful CRISPR-Combo-based morphogenic gene activation strategy. First, the number of morphogenic genes that we screened and tested here was limited. It is easy to conceive that an extensive sgRNA library can be cloned into an all-in-one CRISPRa vector⁵⁹ for the discovery of additional promising genes that promote plant regeneration. Candidate genes to be included in the library can go beyond conventional morphogenic genes to include, for example, microRNA genes^60–63. These genes can be co-activated via multiplexed sgRNA expression to ensure robust organogenesis or somatic embryogenesis toward whole-plant regeneration. Second, more plant species and recalcitrant cultivars could benefit from such research. In this study, the plants we worked with are transformable, though the processes are lengthy and less efficient. We, however, envision using the CRISPR-Combo system to efficiently identify combinations of morphogenic genes that enhance the regeneration of highly recalcitrant plants, taking inspiration from the four Yamanaka factors that can reprogram somatic cells into induced pluripotent stem cells (iPSCs) in mammals^64,65. This is particularly important for commercially valuable genotypes, where recalcitrance to transformation remains a major barrier to the deployment of genome editing for crop improvement and trait engineering. Third, our CRISPR-Combo systems may be applied to improve in planta transformation and genome editing^66–68. It will be interesting to test whether the same morphogenic genes identified with the callus system can promote organogenesis via in planta transformation. Fourth, we want to note that CRISPR-Combo’s unexplored ability to knock down genes (by targeting genes with 15-nt spacers in the gRNA1.0 scaffold) may also be used to promote plant regeneration. For example, some genes, such as WOX13⁶⁹ and miR396⁷⁰, are negative regulators of plant shoot regeneration. While completely knocking out such genes may be detrimental to plants, CRISPR-Combo allows simultaneous knockdown, potentially in combination with morphogenic gene activation, to enhance robust plant regeneration and genome editing. Finally, it is important yet challenging to achieve transgene-free genome-edited perennial plants, especially in commercial varieties. We envision that the morphogenic gene-activating CRISPR-Combo system can be delivered into protoplasts as circular plasmids or as a ribonucleoprotein (RNP) complex⁷¹, which may boost the regeneration of edited protoplasts. Alternatively, Agrobacterium-mediated transient expression of CRISPR-Combo vectors may confer transgene-free genome editing when an herbicide marker-based co-editing strategy is applied^72–74, as CRISPR-Combo is capable of precise base editing^42–44.

Methods

Plant material and growth conditions

Citrus (Citrus sinensis х Poncirus trifoliata var. Carrizo citrange, Citrus medica, var. Citron), potato (Solanum tuberosum L. var. Atlantic), wild strawberry (F. vesca) strain Hawaii 4 (H4), and poplar (Populus tremula × P. alba INRA 717-1B4) were used for hairy root transformation or stable transformation in the study. Transgenic citrus was grown in growth chambers with a 16/8 h (light/dark) photoperiod at 28 °C. Transgenic potatoes, strawberries and poplars were grown in chambers with a 16/8 h (light/dark) at 22 °C.

Selection of potential morphogenic genes in citrus and potato

The citrus (Cs) and potato (St) genes (listed in Supplementary Tables 1 and 3) included in this study were selected based on their conserved and well-documented roles in promoting plant growth through the regulation of organogenesis, somatic embryogenesis, meristem maintenance, hormone signaling, and cell proliferation across diverse plant species. Members of the BBM family, together with LEC1, FUS3, ABI3, ABI4, and AGL15, are core regulators of embryogenic competence and cell proliferation, and their overexpression has been shown to enhance somatic embryogenesis and regeneration efficiency in multiple plant species^8,27,75,76. Genes from the WUSCHEL, WOX, and KNOX families play central roles in maintaining stem cell populations and meristem activity, thereby supporting sustained plant growth and organ formation^21,42,77,78. ARF5/MP and members of the SERK1 family function as key components of hormone-mediated signaling pathways, particularly auxin-dependent and receptor kinase signaling, which are essential for coordinating cell division, pattern formation, and growth^79,80. In addition, IPT (isopentenyl transferase) genes were included as key regulators of cytokinin biosynthesis, a rate-limiting step in cytokinin production that promotes cell division, meristem activity, shoot regeneration, and overall plant growth⁸¹. Furthermore, WIND1- and UPB1-related genes regulate wound-induced dedifferentiation, redox homeostasis, and regeneration-associated growth responses, thereby facilitating tissue regeneration and developmental plasticity^82,83. Collectively, these citrus and potato genes represent complementary regulatory layers that integrate embryogenesis, meristem function, hormone-mediated growth control, and regeneration processes, ultimately promoting plant growth and developmental competence.

Construction of CRISPR-Combo vectors

Oligonucleotides corresponding to individual gRNAs were rationally designed (Supplementary Data 1), synthesized, and annealed to generate double-stranded inserts. gRNAs intended for citrus or potato gene activation were cloned into the BsmBI (Esp3I) sites of pYPQ131B2.0 (Addgene #99885) and pYPQ132B2.0 (Addgene #99888) using T4 DNA ligase, following a previously established protocol⁸⁴. sgRNAs for citrus or potato gene editing were cloned into pYPQ133B (Addgene #69283) and pYPQ134B (Addgene #179216) in a manner identical to each other. For strawberry, the four gene activation gRNAs were cloned into pYPQ131B2.0, pYPQ132B2.0, pYPQ133B2.0 (Addgene #99892), and pYPQ134B2.0 (Addgene #167158), and the two genome editing gRNAs were cloned into pYPQ135B and pYPQ136B. To construct the CRISPR-Combo vector for combinational PtWOX11/PtWUS activation (pLR4610), five gRNAs were used with four gene activation gRNAs cloned into pYPQ131B2.0, pYPQ132B2.0, pYPQ133B2.0 and pYPQ134B2.0, and the genome editing gRNA cloned into pYPQ135B. For multiplexed sgRNA expression constructs, sgRNA entry vectors were assembled into the acceptor vector pYPQ144 (Addgene #69296, for four gRNAs), pYPQ145 (Addgene #69297, for five gRNAs), and pYPQ146 (Addgene #69298, for six gRNAs) by Golden Gate assembly using BsaI-HF and T4 DNA ligase. Multisite Gateway LR reaction (LR Clonase II; Invitrogen) was used to generate the final T-DNA constructs, by recombining the attL1-attL5 entry vector pYPQ-Cas9-Act3.0 (Addgene #178954), the assembled attR5-attL2 pYPQ144/145/146 entry vector, and the attR1-attR2 destination vector pCGS710 (for potato and citrus), pYPQ202 (Addgene #86198, for poplar), and pJH23⁸⁵ (for strawberry). The detailed procedures for the other control vectors were previously described for strawberry (JH19 vectors)⁵⁶ and for poplar (pLR2070, pLR4071, and pLR4073)⁴². Vector identities and configurations are summarized in Supplementary Tables 2 and 4. All these newly developed CRISPR-Combo constructs were verified by sequencing and then introduced into R. rhizogenes (strain ATCC 43056) and A. tumefaciens (strain GV3101) for the induction of hairy roots and stable transformation, respectively.

Rhizobium rhizogenes-mediated citrus hairy root transformation

R. rhizogenes, harboring CRISPR-Combo vectors, was used for citrus hairy root transformation. A single positive colony of R. rhizogenes strain (ATCC 43056) was grown from stock on Luria-Bertani (LB) solid media supplemented with kanamycin (50 mg/L) at 28 °C in the dark at 200 rpm for 12–16 h. The overnight culture was collected for an optical density (OD) of 1.0, centrifuged at 4304 × g for 20 min at 20 °C, and washed with ½ MS liquid media (1/2 MSLIQ, Supplementary Table 7) supplemented with 200 µM Acetosyringone. The cells were centrifuged again, resuspended in ½ MS liquid to a final OD₆₀₀ of 0.6. The culture was supplemented with 40 µM Acetosyringone and gently shaken for 1 h at 28 °C in the dark at 160 rpm.

Citrus in vitro and ex vivo hairy root transformation was described previously^50,86,87. Briefly, 4-week-old green epicotyls were sliced at an angle 5 mm above the cotyledon and incubated vertically in the R. rhizogenes culture for 30 min at room temperature. The epicotyls were removed and carefully blotted dry on sterile filter paper before being placed on ½ MS co-cultivation media (HRCCM, Supplementary Table 7) and incubated at 22 °C in the dark for 3 days. The epicotyls were washed for 20 min at room temperature with ½ MS liquid supplemented with Cefotaxime (200 mg/L) to kill R. rhizogenes and blotted dry on sterile filter paper. The epicotyls were transferred to ½ MS hairy root selection media and placed under a 16/8 h (light/dark) photoperiod at 28 °C for 3–4 weeks; by then, some callus/root formation had started. The epicotyls were washed and transferred to a hairy root induction medium for continued root induction and elongation for 4–6 more weeks. For ex vivo hairy root transformation of citron (Citrus medica L.), fully expanded leaves with attached petioles were excised from healthy 1-year-old plants grown under greenhouse conditions. The cut ends of the petioles were dipped into a fresh culture of R. rhizogenes (OD₆₀₀ = 0.8) prepared as above, then vacuum-infiltrated for ~30 min at 30 psi and subsequently incubated on a rotary shaker at 50 rpm for 24 h to facilitate bacterial infection. Following incubation, all explants were transferred to the inert vermiculate matrix in a humidity dome tray with frequent misting to maintain high humidity. Hairy root induction was monitored, and quantitative analysis was performed 30 days after transformation.

Rhizobium rhizogenes-mediated potato hairy root transformation

R. rhizogenes-mediated transformation of the commercial potato cultivar Atlantic was performed as described by Bedre et al.⁸⁸. In brief, stem segments from 4-week-old in vitro-grown plantlets were co-cultured with a suspension of bacteria (OD₆₀₀ = 0.6) containing the CRISPR-Combo vectors for 20 min and then transferred into co-cultivation medium comprising MS salts supplemented with zeatin riboside (2.2 mg/L) (Gold Biotechnology, Inc., St. Louis, MO), indol-3-acetic acid (0.01 mg/L; Millipore Sigma), and gibberellic acid (0.2 mg/L; Caisson Labs). After 2 days of co-cultivation, the stem segments were transferred to a callus induction medium consisting of MS salts supplemented with carbenicillin (200 mg/L) (Caisson Labs), cefotaxime (250 mg/L) (Millipore Sigma), and kanamycin (50 mg/L) (Caisson Labs), zeatin (2.2 mg/L), IAA (0.01 mg/L) and GA3 (0.2 mg/L). The plates were incubated under fluorescent light (300 mmol s⁻¹ m⁻²) at 22 °C for 14 days. After 2 weeks, the stem segments were sub-cultured in MS salts supplemented with kanamycin (50 mg/L) without hormones. Hairy root induction was recorded every 10 days.

Agrobacterium-mediated stable transformation in potato

Agrobacterium-mediated transformation of the commercial potato cultivar Atlantic was performed as previously described⁸⁹. In brief, stem segments from 4-week-old in vitro-grown seedlings were co-cultured with a suspension of Agrobacteria containing each of the CRISPR-Combo constructs for 20 min and then transferred into co-cultivation medium comprising MS medium supplemented with zeatin riboside (2.2 mg/L) (Gold Biotechnology, Inc., St. Louis, MO), indol-3-acetic acid (0.01 mg/L; Millipore Sigma), and gibberellic acid (0.2 mg/L; Caisson Labs). After 2 days of co-cultivation, the stem segments were transferred to a selective regeneration medium consisting of MS medium supplemented with carbenicillin (200 mg/L) (Caisson Labs), cefotaxime (250 mg/L) (Millipore Sigma), and kanamycin (50 mg/L) (Caisson Labs). The plates were incubated under fluorescent light (300 mmol s⁻¹ m⁻²) at 22 °C. Stem segments were sub-cultured every 3 weeks until they regenerated. Plantlets of ~20 mm tall were transferred to the rooting medium for molecular analyzes.

Agrobacterium-mediated citrus for stable transformation

A. tumefaciens (strain GV3101) harboring CRISPR-Combo vectors was used for citrus (Carrizo (Citrus × poncirus)) stable transformation. The transformation of citrus was performed as described by Dominguez et al.⁹⁰. Briefly, 4-week-old in vitro-grown citrus bright-green epicotyls were cut transversely into 1 cm lengths and co-cultured with A. tumefaciens suspension (OD₆₀₀ = 0.6) for 20 min at room temperature. The cell suspension was removed, and the explants were blotted dry on sterile filter paper to remove excess bacteria. Explants were placed horizontally on DBA3 co-cultivation media (CCM; Supplementary Table 7) and incubated at 22 °C in the dark for 3 days. The explants were subsequently transferred to DBA3 selection media (SM; Supplementary Table 7) and maintained for 3 weeks under a 16/8 h (light/dark) photoperiod at 28 °C. Shoots began to develop after 2 weeks. Regenerated shoots with green leaves approximately 4 mm in height were carefully excised from the explants and laid flat on Shoot Elongation Media (SEM, Supplementary Table 7). The putative transgenic shoots were subjected to PCR analysis to detect the nptII gene. This process was repeated every 3 weeks for up to 4 months. The PCR-positive shoots were placed in rooting media (MSRM, Supplementary Table 7).

Agrobacterium-mediated strawberry for stable transformation

Agrobacterium-mediated transformation of woodland strawberry F. vesca was performed as previously described with minor modifications⁹¹ (Supplementary Table 7). Briefly, cotyledons or juvenile leaves of Hawaii 4 seedlings were sliced and cultured on Murashige and Skoog (MS) medium, followed by vacuum infiltration with A. tumefaciens strain GV3101 harboring the CRISPR-Combo or the control JH19 constructs using a syringe barrel and plunger. Infected explants were incubated on 5++ medium in the dark for 3 days. Subsequently, explants were rinsed with sterile water, transferred to CT medium, and maintained in the dark for an additional 7 days to induce callus formation. Transformed calli were then selected on CT medium containing 4 mg/L hygromycin with passages every 2–3 weeks until shoots emerged. All cultures were maintained in a growth chamber with a 16/8 h (light/dark) at 22 °C. Regenerated plantlets with 2–3 compound leaves were transferred to the rooting medium. After approximately 1–2 months, rooted plants were transplanted into soil and subjected to genotyping.

Agrobacterium-mediated poplar for stable transformation

P. tremula × P. alba clone 717-1B4 was used for stable transformation following our optimized protocol⁹² (Supplementary Table 7). The T-DNA vector was introduced into A. tumefaciens strain GV3101. Briefly, petiole explants cultured on callus induction medium (CIM) were used for Agrobacterium inoculation. Regenerated shoots were selected on hormone-free Driver and Kuniyuki Walnut medium (DKW-HF) supplemented with 20 μg/mL hygromycin, 250 mg/L cefotaxime, and 200 mg/L timentin. Hormone-free medium consisted of the respective basal salts and vitamins supplemented with a carbon source, without the addition of any exogenous plant growth regulators (e.g., auxins or cytokinins) or complex organic additives. All the calli were included in the assessment of shoot regeneration rate. Shoots reaching a height of 1–2 cm were subsequently transferred to hormone-free ½-strength LS medium (½ LS-HF) containing 20 μg/mL hygromycin and 200 mg/L timentin. All the regenerated shoots were included in the assessment of rooting rate. Rooted plants were propagated in ½ LS-HF and used for genotyping.

Validation of transgenic plants

To verify the presence and stable integration of the CRISPR-Combo T-DNA in regenerated lines across all four species, genomic DNA was extracted from leaf tissues. The presence of the transgene was confirmed via PCR amplification of the respective plant selection markers: the nptII gene for potato and citrus, and the hpt and Cas9 gene for strawberry and poplar, respectively, using specific primers (Supplementary Data 1).

Validation of CRISPR genome editing

To identify indels in targeted gene-edited lines, a 300–500 bp target region was amplified by PCR with gene-specific primers flanking the target sites (Supplementary Data 1) using Phusion Polymerase (New England Biolabs, Ipswich, MA, USA). PCR amplicons were purified using PCR clean-up kits (Zymo Research, Irvine, CA, USA) and sequenced using Sanger sequencing (Eton Bioscience, Inc., San Diego, CA, USA). The proportion of indels in StVINV-edited potato hairy root and stable transgenic lines was analyzed using the Inference of CRISPR Edits tool⁹³. Next-generation sequencing was used to analyze the targeted genomic regions that were directly amplified from genomic DNA extracted from leaf tissue using the CTAB method⁹⁴ and barcoded using the Hi-TOM primers. PCR products were verified by gel electrophoresis. Approximately 6–60 PCR products were pooled together and purified using a QIAquick PCR Purification Kit (QIAGEN), followed by quantification with a Nanodrop (Thermo Fisher). The purified PCR products were used for Illumina HiSeq2500 sequencing. For targeted mutagenesis, the NGS data were analyzed with CRISPRMatch⁹⁵ and CRISPResso2⁹⁶.

RNA extraction and qPCR analysis

For potato and citrus, the total RNA was extracted using the Direct-zol RNA MiniPrep Plus kit (Zymo Research) according to the manufacturer’s instructions. For strawberry plants, total RNA was isolated using a cetyltrimethylammonium bromide (CTAB)-based protocol as described by Gambino et al.⁹⁷. For cDNA synthesis, approximately 500–1000 ng of total RNA was reverse-transcribed using either SuperScript IV VILO Master Mix (Invitrogen, USA) or SuperScript III First-Strand Synthesis Kit with oligo (dT) primers (Thermo Fisher). Quantitative real-time polymerase chain reaction (RT–qPCR) was conducted on a Bio-Rad CFX96 real-time System (Bio-Rad, USA) with PowerUp SYBR Green Master Mix (Applied Biosystems, USA) or the AzuraQuant Green Fast qPCR Mix (Azura Genomics). The 10 μl reaction consists of 1 μl of diluted cDNA template, forward and reverse primer at 200 nM, 5 μl of 2× SYBR Green Master Mix, and nuclease-free water. The thermal cycling profile was: step 1, 50 °C for 2 min, step 2, 95 °C for 2 min, followed by 40 cycles of 95 °C for 15 s, 55 °C for 15 s, and 72 °C for 30 s. Relative gene expression levels were calculated using the 2⁻^ΔΔCt method⁹⁸. StRPL2 (RIBOSOMAL PROTEIN L2), CsGAPC2 (GLYCERALDEHYDE-3-PHOSPHATE DEHYDROGENASE C2), FvePP2a (FvH4_4g27700), and PtCDC2 (cell division control protein 2) were used as the endogenous control gene for potato, citrus, strawberry, and poplar, respectively. The fold changes of target genes were calculated by the 2^−ΔΔCt method⁹⁸. All primers used in this study are listed in Supplementary Data 1.

Statistics and reproducibility

For potato hairy root transformation, each construct was transformed with 3–4 independent plates, with each plate containing 10–12 explants. For potato stable transformation, each construct was transformed on 3 independent plates, with each plate containing 10–15 stem segments. For citrus hairy root transformation, each construct was transformed with 4–5 independent boxes, with each box containing 4 explants. For ex vivo citrus hairy root transformation, each construct was transformed with 20–22 explants. For citrus stable transformation, each construct was transformed on 3-4 plates, with each plate containing 25–30 explants. For strawberry transformation, 4–30 T₀ transgenic plants were regenerated for each construct. For poplar transformation, 12 T₀ lines were generated for analysis per construct. For stable transformation of potato, citrus, strawberry, and poplar, all transformation experiments were done twice independently, with similar results observed. All data analyzes, statistical tests, and graphs were generated as described in the figure legends using Microsoft Excel (2011) and GraphPad Prism. Unless otherwise stated, most statistical comparisons were performed using a two-tailed Student t-test or analysis of variance (ANOVA) followed by Tukey’s multiple-comparison test. Turkey’s multiple comparisons of the poplar data were presented in Supplementary Tables 8–12.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

Y.Q. proposed and coordinated the research. M.R. conducted Rhizobium rhizogenes-mediated hairy root transformation of potato and citrus, Agrobacterium-mediated transformation of potato, citrus, and molecular analysis of the resulting transgenic lines. G.L. generated the poplar CRISPR-Combo constructs, conducted poplar transformation, and analyzed the transgenic lines for genotypes and phenotypes. L.G. generated the strawberry CRISPR-Combo constructs, conducted strawberry transformation, and genotyped the transgenic lines. H. F. selected citrus morphogenic genes, made the corresponding CRISPR-Combo constructs, and did the initial testing of these constructs in citrus. A.E. selected potato morphogenic genes, made the corresponding CRISPR-Combo constructs, and did the initial testing of these constructs in potato. R.M. and Y.C. provided support in vector construction and genotyping. R.N. assisted in plant transformation and statistical analysis. J.C. provided resources for citrus research. G.D.C. provided resources for poplar research. Z.L., K.K.M. and Y.Q. provided resources and supervised the research. Y.Q. M.R., G.L., L.G., H.F. and A.E. wrote the manuscript with input from all authors. All authors reviewed and agreed on the final manuscript for submission.

Peer review

Peer review information

Nature Communications thanks Nagesh Sardesai and other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This research was mainly supported by the Foundation for Food & Agriculture Research Genotype-Independent Regeneration of Fertile Plants Program (award no. 21010111). It was also partly supported by the NSF Plant Genome Research Program (award no. IOS-2132693 and IOS-2224203), USDA-NIFA’s Emergency Citrus Disease Research & Extension (award no. 2020-70029- 33161, 2021-70029-36056, 2025-70029-44033), HATCH (TEX09621, TEX0-7790), Department of Energy (award no. DE-SC0023011), McIntire Stennis Forest Research Program (MD-PSLA-24014), Texas A&M AgriLife Institute for Advancing Health Through Agriculture, and Texas A&M AgriLife Research Insect-vectored Disease Program (114190- 96210).

Data availability

The modular CRISPR-Combo vectors were previously described⁴² and are available at Addgene [https://www.addgene.org/Yiping_Qi/]. The CRISPR-Combo T-DNA vectors (Supplementary Tables 2, 4, 5, and 6) are available upon request, pending an MTA. NGS data of poplar T₀ plants are accessible via NCBI with the accession number PRJNA1459490. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76367-w.