Introduction

Soybean (Glycine max) is an important crop worldwide, and Brazil is the leading producer, with an annual production of about 138 and 153 million tonnes in 2020/21 and 2022/23, respectively (Shahbandeh 2023). It is the crop with the highest protein content (40–42%) and the food legume with the second highest oil concentration (18–22%) (Pagano and Miransari 2016). In addition, soybean can be used for biofuel production and is a major source of protein for human and animal consumption (Pagano and Miransari 2016). Plant parasitic nematodes (PPNs) are among the most damaging agricultural pests worldwide, particularly in the soybean crop (Arraes et al. 2022). PPNs have sophisticated parasitism strategies, as they can successfully infect many host plant species (Warmerdam et al. 2018; Basso et al. 2020a). PPNs are hypothetically grouped according to their feeding behavior into ectoparasitic, semi-endoparasitic, and endoparasitic nematodes, possibly a sedentary or non-sedentary parasitism. Importantly, PPNs can cause annual losses of about $78.0 billion worldwide and crop yield losses of about 10–15% (Lima et al. 2017; Dias-Arieira et al. 2021).

Currently, soybean production and yield are mainly challenged by nematodes of the genera Meloidogyne, Heterodera, Pratylenchus, and Rotylenchulus (Elhady et al. 2018). Root-knot nematodes (RKNs) are ubiquitous, sedentary endoparasites of the genus Meloidogyne that induce permanent feeding sites in roots to complete their life cycle (Sato et al. 2019; Moreira et al. 2023). The pre-parasitic second-stage juvenile (ppJ2) infects plant roots and migrates to the vascular cylinder, inducing cell dedifferentiation into multinucleated giant cells known as feeding cells (Escobar et al. 2015; Sato et al. 2019). Meloidogyne incognita is one of the most abundant and important species in soybean crops (Abad et al. 2008; Escobar et al. 2015; Basso et al. 2022a; Moreira et al. 2022). Varietal resistance to nematodes in soybean is often a horizontal resistance that does not persist for many years, as it is conferred by multiple small effect genes or quantitative trait loci (QTL), which are widely involved in multiple disease control (Lin et al. 2022). For this reason, there is a high demand for the development of tools to manage and support the integrated management of these nematodes. New biotechnological tools based on classical transgenesis (RNA interference and gene overexpression) or genome editing (endogenous soybean genes) are powerful strategies that can be used in the integrated nematode management (Basso et al. 2020b; Mendes et al. 2021a). As a requirement for this, the exploration and validation of new molecules (e.g., genes and promoter sequences) associated with resistance or susceptibility of soybean to M. incognita can contribute to the development of less susceptible soybean cultivars (Arraes et al. 2022).

Agrobacterium rhizogenes is a symbiotic gram-negative soil bacterium that causes hairy root disease in the host plants, characterized by adventitious root overgrowth at the site of infection (Young et al. 2001; Guillon et al. 2006). Hairy roots are induced by the gene transfer (T-DNA) of the bacterial root-inducing (Ri) plasmid through the type IV secretion system into the plant nuclear genome (Chilton et al. 1982; Gelvin 2009). The hairy root transformation (composite roots) mediated by A. rhizogenes has been mainly used for functional validation of target genes, single-guide RNA (sgRNA) and plant genome editing strategies using clustered regularly interspaced short palindromic repeat (CRISPR) with CRISPR-associated protein 9 (Cas9) (CRISPR/Cas9) technologies (Georgiev et al. 2012; Häkkinen and Oksman-Caldentey 2018; Li et al. 2019; Bai et al. 2020; Gutierrez-Valdes et al. 2020). To date, several types of soybean explants have been used to induce hairy roots, including hypocotyls (Kereszt et al. 2007; Yang et al. 2021), cotyledons (Chen et al. 2018; Lin et al. 2023), root tips (Weber and Bodanese-Zanettini 2011), and shoots (Song et al. 2021). Previous studies have successfully induced hairy roots from soybean cotyledons (Chen et al. 2018) and hypocotyls (Kereszt et al. 2007) under conventional in vitro conditions and from 5-day-old soybean seedlings maintained in a greenhouse (Yang et al. 2021). However, these hairy roots induced from the cotyledons are not suitable for root-knot nematode studies. On the other hand, hairy roots induced from hypocotyl are suitable for nematode bioassay, but it is a time and space consuming method as hairy roots need to be cultured to form a soybean plant. Therefore, there is a great need for the development of a simple, rapid, and easy system for hairy root production from soybean for the validation of genes and technologies in conjunction with nematode bioassays. To support this, a hairy root induction protocol from detached leaves under ex vitro conditions was successfully developed for the validation of target genes in Arachis hypogaea associated with nematode bioassays (Guimaraes et al. 2017b). More recently, the same ex vitro hairy root protocol was also established in detached soybean leaves and validated for gene expression and RNA interference strategies associated with nematode bioassays (Pereira et al. 2023).

Herein is presented an ex vitro hairy root system generated from petioles of detached soybean leaves for screening candidate genes related to plant resistance to RKNs using classical transgenesis and CRISPR technologies. This method for soybean hairy roots was established based on the procedures previously used for functional validation of candidate genes for plant resistance to RKNs in peanuts and soybean (Guimaraes et al. 2017b; Pereira et al. 2023). Subsequently, this ex vitro hairy root system was validated using both classical transgenesis and CRISPR-activation (CRISPRa) strategies. Transgenic soybean hairy roots were used to overexpress the GmPR10 (G. max pathogenesis-related protein 10) and GmGST (G. max glutathione S-transferase) genes in association with nematode bioassays. In addition, soybean hairy roots were used in CRISPR/dCas9 for transcriptional activation of the GmExpA (G. max expansin A) and GmExpLB (G. max expansin-like B) genes for nematodes bioassays. The ex vitro hairy root system proved to be a very efficient and simple method for screening target genes, using conventional overexpression and CRISPR/dCas9 approaches, and for validating CRISPR sgRNAs directly in the target crop plant.

Materials and methods

Classical transgenesis binary vectors

The pPZP-GmPR10-eGFP-HygR and pPZP-GmGST-eGFP-HygR binary vectors were synthesized and assembled by Epoch Life Science (Sugar Land, TX, USA). These constructs contained GmPR10 (477 bp coding sequence; Glyma.17G030400.1) and GmGST (669 bp coding sequence; Glyma.18G190300) coding sequences under the Cauliflower mosaic virus (CaMV) 35S promoter, respectively (Fig. 1a). The hptII selectable marker gene for hygromycin B resistance was used under the control of the Arabidopsis thaliana Ubi3 promoter. The enhanced green fluorescent protein (eGFP) gene was used as a molecular marker protein under the control of the CaMV 35S promoter. As a negative control, pPZP-eGFP-HygR (hereafter referred to as pPZP empty vector) was used, which contains the same backbone as the first two vectors. Both binary vectors were transferred into A. rhizogenes strain K599 and A. tumefaciens strain GV3101 and used for the transformation of soybean hairy roots and Nicotiana tabacum var. Petit Havana, respectively.

Fig. 1
figure 1

Overexpression of the GmPR10 and GmGST genes in transgenic soybean hairy roots and tobacco lines. a Overview of the T-DNA cloned into the pPZP binary vector for constitutive overexpression of the GmPR10 and GmGST genes. b, d Relative expression of the GmPR10 (b) and GmGST (d) genes in transgenic soybean hairy roots overexpressing the GmPR10 and GmGST genes compared to transgenic soybean hairy roots transformed with pPZP empty vector (n = 6). c, e Number of galls per gram of root in transgenic soybean hairy roots overexpressing the GmPR10 (c) and GmGST (e) genes compared to transgenic soybean hairy roots transformed with the empty pPZP vector at 30 dpi (n = 8). f Bright field and eGFP fluorescence images generated from transgenic soybean hairy roots overexpressing the GmPR10 and GmGST genes compared to transgenic soybean hairy roots transformed with the empty pPZP vector. Scale bars = 2 mm. g, i Relative expression of the GmPR10 (g) and GmGST (i) genes in transgenic tobacco lines overexpressing the GmPR10 and GmGST genes compared to non-transgenic (NT) control plants (n = 6). h, j Number of galls per gram of root in transgenic tobacco lines overexpressing the GmPR10 (h) and GmGST (j) genes compared to NT control plants at 60 dpi (n = 10). k Bright field and eGFP fluorescence images generated from transgenic tobacco lines overexpressing the GmPR10 and GmGST genes compared to NT tobacco lines. Scale bars = 2 mm. Data are displayed as mean ± SD. Asterisks represent the significance level according to the unpaired t-test (P value < 0.05). Different letters indicate significant differences by Tukey’s post-hoc test with multiple comparisons, and one-way ANOVA (P value < 0.05)

Full size image

CRISPR/dCas9 binary vectors

The CRISPR/dCas9 binary vectors were designed for the transcriptional activation of the Glycine max expansin A (GmExpA, Glyma.06G021700; GmExpA2, Glyma.04G021600) and Glycine max expansin-like B (GmExpLB; Glyma.01G204800) genes (Fig. 2a). The sgRNAs were designed using the CRISPR-P 2.0 software (Lei et al. 2014). The sgRNA sequences were blasted against the soybean genome using Phytozome 13 (https://phytozome-next.jgi.doe.gov/) to check for potential off-target effects. Three sgRNAs for each target gene were cloned separately into the CRISPR-Act2.0-pYPQ141A2.0 vector using Golden Gate assembly (Suppl. Table S1) (Lowder et al. 2018). Subsequently, each construct was recombined with the pYPQ173 (CaMV 35S::pco-dCas9-NLS-VP64-T2A-MS2-NLS-VP64) and pMDC32-eGFP (CaMV 35S::eGFP) destination vectors using Gateway assembly (Lowder et al. 2018). The CRISPR-Act2.0 system consists of the dCas9 (a deactivated Cas9 protein from Streptococcus pyogenes) fused to the transcriptional activator VP16/64 and a modified guide RNA2.0 scaffold with two MS2 aptamers to recruit the MS2 bacteriophage coat protein (MCP) fused to a VP16/64 activator (Fig. 2b). The plant codon optimized dCas9 gene was fused to NLS-VP64-T2A-MS2-NLS-VP64 and cloned under the control of the CaMV 35S promoter. The sgRNAs were cloned under the control of the AtU6-1 promoter, fused to scaffold 2 × MS2 aptamers, and transcription was stopped by the AtU6 terminator. The final CRISPR/dCas9 binary vectors containing the sgRNA were transferred into A. rhizogenes strain K599 and used to transform soybean hairy roots.

Fig. 2
figure 2

Transcriptional activation of endogenous genes in transgenic soybean hairy roots using CRISPR/dCas9 technologies. a Overview of the T-DNA cloned into the binary CRISPR-Act2.0 vector. b Overview of the dCas9-Act2.0 system for transcriptional activation of target genes. In brief, the CRISPR-Act 2.0 system consists of a dCas9 protein fused to a VP64 activator domain, a single-guide RNA (sgRNA) in tandem with a sgRNA scaffold 2.0 containing two MS2 aptamers, which acts to recruit the MS2 bacteriophage coat protein (MCP) fused to a VP54 activator domain. c Bright field and eGFP fluorescence images generated of transgenic soybean hairy roots transformed with different CRISPR/dCas9 binary vectors. Scale bars = 2 mm. d, g Transcriptional activation levels of GmExpA (sgRNA16, 26, and 55) and GmExpA2 (sgRNA16) (d), GmExpLB (g), and dCas9 (e/h) genes in transgenic soybean hairy roots transformed with CRISPR/dCas9 binary vectors (n = 4 to 5). f, i Susceptibility levels to M. incognita of the transgenic soybean hairy roots transformed with the CRISPR/dCas9-GmExpA-sgRNA16 (upregulation of the GmExpA and GmExpA2 genes) (f) and CRISPR/dCas9-GmExpLB-sgRNA45 (upregulation of the GmExpLB gene) (i) binary vectors at 30 dpi (n = 10). Data are displayed as mean ± SD. Asterisks represent the significance level according to the unpaired t-test (P < 0.05). Different letters indicate significant differences by Tukey’s post-hoc test with multiple comparisons, and one-way ANOVA (P value < 0.05)

Full size image

Agrobacterium rhizogenes and A. tumefaciens cultures

A single colony harboring each binary vector was plated on Petri plates containing a solid LB medium supplemented with appropriate antibiotics and grown at 28 °C for 48 h. An aliquot of 1 mL LB medium containing 15% glycerol (v/v) was added to the culture and homogenized. A bacterial paste was then prepared by plating 500 µL of the mixture onto Petri plates containing fresh solid LB medium supplemented with appropriate antibiotics and grown at 28 °C for 16 h. The A. tumefaciens clones were grown in LB medium with appropriate antibiotics overnight at 28 °C.

Hairy root induction from soybean cotyledon and hypocotyl

Soybean cv. Williams 82 seeds were sterilized with 1–2% sodium hypochlorite for 8 min and washed 6 times with sterile water. Sterile seeds were germinated in 150 × 25 mm Petri plates containing two sterile filter papers soaked in 0.5X liquid MS medium (Murashige and Skoog 1962). Subsequently, unfolded green cotyledons were isolated from plants incubated for 7–10 days at 26 °C under a 12/12 h photoperiod. Multiple 1–3 mm superficial incisions were made on the abaxial region of the cotyledons co-cultured with A. rhizogenes as previously described (Liu et al. 2019). For the cotyledon, a culture of the A. rhizogenes clone containing the empty pPZP vector was used at OD600 0.5 resuspended in 0.5X liquid MS medium. Meanwhile, the central part of the hypocotyl, close to the cotyledonary node, was isolated and stabbed with a sterile needle (23¾ gauge) carrying a drop of the bacterial paste (A. rhizogenes clone containing the empty pPZP vector). Inoculated cotyledons and hypocotyls were transferred to a 150 × 25 mm Petri plate and moist cotton was placed on the wound, but removed 5 days after inoculation. Petri plates were maintained at 26 ± 2 °C under a 12/12 h photoperiod for 10–15 days (Kereszt et al. 2007). Transgenic hairy roots were assessed 10–15 days after co-cultivation by eGFP fluorescence at 488 nm under a fluorescence stereomicroscope M205 (Leica Microsystems, Wetzlar, Germany).

Hairy root induction from petioles of detached soybean leaves

Soybean cv. Williams 82 plants were grown in a greenhouse for 18–20 days. Initially, three leaf types were tested for their potential to induce hairy roots: (i) central leaflet of the trefoil, (ii) unifoliate leaf, and (iii) complete trefoil. The expanded leaves were harvested at V1 and V2 vegetative growth stages by cutting at the petiole-stem junction. The petioles were then inoculated with a small amount of bacterial paste (A. rhizogenes clone containing the empty pPZP vector). Inoculated leaves were transferred to 150 × 15 mm Petri plates containing moist cotton covered with Germitest paper. The inoculated area of the petiole was covered with moist cotton and placed under a glass slide, which was watered every 2 days. The Petri plates were maintained for 12–15 days in a growth chamber at 25 ± 2 °C, 12/12 h photoperiod, and 120 µmol m−2 s−1 light intensity. Transgenic hairy roots were screened for eGFP fluorescence as described previously.

Gene overexpression and CRISPR/dCas9-based transcriptional activation in hairy roots produced from petioles of detached soybean leaves

Hairy root production from soybean detached leaves was repeated as described above using the other binary vectors for classical transgenesis and transcriptional activation. At 12–15 days after inoculation, the eGFP-negative hairy roots were systematically removed. Transgenic hairy roots were collected for RNA and gene expression analysis. eGFP-positive hairy roots were covered with vermiculite and acclimatized for 5 days in the growth room under the same conditions as above. Hairy roots were then inoculated with ppJ2 from M. incognita race 1 (Guimaraes et al. 2017b).

Stable genetic transformation of Nicotiana tabacum

Stable transgenic lines of the N. tabacum var. Petit Havana were generated using the binary vectors pPZP-GmPR10-eGFP-HygR and pPZP-GmGST-eGFP-HygR. Plant transformation was performed using the leaf disc method (Horsch et al. 1985) under in vitro selection with 50 mg/L hygromycin B (Harrison et al. 2006). Transgenic plants were characterized by PCR (transgene insertion) and real-time PCR (gene expression) using target-specific primers (Suppl. Table S2). T1 generation transgenic lines were subjected to bioassays with M. incognita race 3.

Meloidogyne incognita inoculation

Meloidogyne incognita races 1 and 3 were maintained in tomato plants cv. Santa Clara. Eggs were collected and hatched as described by Mendes et al. (2021a) and Moreira et al. (2022). Soybean hairy roots were inoculated with 1000 ppJ2 of M. incognita race 1 while transgenic tobacco plants were inoculated with 1000 ppJ2 of M. incognita race 3. Soybean hairy roots and transgenic tobacco roots were collected for gall counting at 30 and 60 days post-inoculation (dpi), respectively.

Gene expression

The RNA was extracted using the PureLink™ Plant RNA Reagent (Thermo Fisher Scientific, Waltham, MA, USA) and quantified using a NanoDrop 2000 (Thermo Fisher Scientific). Subsequently, 1 μg of RNA treated with Ambion DNase I (Invitrogen, Waltham, MA, USA) was used for cDNA synthesis using oligo-(dT)20 primers and the M-MLV Reverse Transcriptase Kit (Invitrogen). The cDNA samples were diluted 1:10 and real-time PCR assays were performed on an Applied Biosystems 7500 Fast Real-Time PCR System (Applied Biosystems, Waltham, MA, USA) using 0.2 µM primers (Suppl. Tables S2 and S3), qPCR SybrMaster mix (Cellco Biotec, São Carlos, São Paulo, Brazil), and 2 μL cDNA in a final volume of 10 µL. Relative expression levels were calculated using the 2−ΔCt method (Schmittgen and Livak 2008). The GmCYP2, GmELF1α, and GmAct1 were used as endogenous reference genes for soybean (Miranda et al. 2013) (Suppl. Table S3) while NtActin1, NtActin2, and NtL25 genes were used for tobacco plants (Suppl. Table S2). Four to six biological replicates were used for each treatment, while all samples were carried out in technical triplicates. Target-specific amplification was confirmed by a single and distinct peak in the melting curve analysis. Real-time PCR Miner was used to determine the ∆Ct values and the PCR efficiency (Zhao and Fernald 2005).

Microscopy analysis

Galls were collected and fixed in 2% glutaraldehyde prepared in 50 mM PIPES buffer pH 6.9. The material was dehydrated in serial dilutions of cold ethanol (15, 30, 50, 75, 85, and 100%) and embedded in Technovit 7100 plastic resin (EMS, Hatfield, Hertfordshire, UK). Polymerized blocks were sectioned at 5 µm using a microtome and floated in drops of water. Each slide was stained with 0.05% (w/v) toluidine blue and sealed with Depex. Sections were photographed using a digital Axiocam (Zeiss, Oberkochen, Germany) with standard bright-field or dark-field optics.

Statistical analysis

Galling number data in soybean hairy roots were evaluated by the unpaired t-test (n = 8 to 10), while in transgenic tobacco lines they were evaluated by the Tukey’s test (n = 10). Real-time PCR data from hairy roots overexpressing GmPR10 and GmGST genes were evaluated by the unpaired t-test (n = 6), while in transgenic tobacco lines and CRISPR/dCas9 soybean hairy roots were evaluated by the Tukey’s test (n = 4 to 6). Real-time PCR data were analyzed using the SATqPCR software (Rancurel et al. 2019). Differences at P value < 0.05 were considered statistically significant.

Results

Hairy root induction from soybean cotyledon, hypocotyl, and petioles of detached leaves

Hairy roots were induced from soybean cotyledon and hypocotyl under in vitro conditions, whereas hairy roots from petioles of detached leaves were induced under ex vivo conditions (Fig. 3a–l). In this first test, only the A. rhizogenes culture containing the empty pPZP vector was used. Cotyledons showed a hairy root induction rate and transformation efficiency of 100 and 91%, respectively (Table 1). Hypocotyls showed a hairy root induction rate and transformation efficiency of 100 and 21%, respectively (Table 1). Similarly, petioles of detached leaves showed a hairy root induction rate and transformation efficiency of 96.7 and 86.2%, respectively (Table 1). Induction of hairy roots from cotyledons did not allow the development of soybean seedlings and this system was not compatible with nematode bioassays (Fig. 3a–d). In contrast, the hairy root induction from soybean hypocotyls allowed seedlings to form, but the hairy roots had to be cultured in a suitable substrate to allow branches and leaves to develop (Fig. 3e–h). Unfortunately, in this system, the roots can only be used for the nematode bioassay after seedling formation, which is very time- and space-consuming. Differently, hairy root induction from petioles of detached leaves allowed the nematode bioassays to be carried out without seedling formation (Fig. 3i–l). To improve hairy root induction from petioles of detached leaves, leaves of soybean plants at different vegetative stages were tested. Soybean plants at V1 or V2 were preferentially used, as younger leaves are healthier and therefore survive and are more likely to give hairy roots than those at V3 stage. The central leaf of the trefoil (Fig. 4a and b) and the unifoliate leaf (Fig. 4a and c) showed well-developed hairy roots already 13 days after inoculation with A. rhizogenes. Particularly, the central leaf of the trefoil exhibited longer and less branched hairy roots (Fig. 4b). Fortunately, the unifoliate leaves showed shorter and more numerous adventitious hairy roots, which are very suitable for carrying out bioassays with nematodes in Petri plates (Fig. 4c). In contrast, the whole trefoil was not able to induce hairy roots (Fig. 4d). Therefore, this ex vitro hairy root system from petioles of detached leaves (unifoliate leaves) proved to be more practical and space efficient as it is compatible with the bioassays with nematodes in Petri plates. Therefore, considerable emphasis was placed on the ex vitro hairy root system for its validation through the screening of target genes, using classical transgenesis and CRISPR strategies, associated with nematode bioassays.

Fig. 3
figure 3

Hairy root induction from soybean cotyledon, hypocotyl, and petioles of detached leaves. a, b Hairy root induction from soybean cotyledon under in vitro conditions. e, f Hairy root induction from soybean hypocotyl under in vitro conditions. i, j Hairy root induction from petioles of detached soybean leaves under ex vitro conditions. Bright field and eGFP fluorescence images of transgenic soybean hairy roots induced from cotyledon (c, d), hypocotyl (g, h), petioles of detached leaves (k, l). Scale bars = 5 mm

Full size image
Table 1 Hairy root induction rate and transformation efficiency from soybean cotyledon, hypocotyl, and petioles of detached leaves using A. rhizogenes transformed with the empty pPZP vector
Full size table
Fig. 4
figure 4

Hairy root induction from petioles of detached soybean leaves, comparing the different leaf types. a Trefoil of a soybean plant in V1 developmental stage. b Hairy roots induced from the central leaves of the trefoil. c Hairy roots induced from unifoliate leaves. d Failure of hairy root induction using whole trefoils. Bar = 3 cm

Full size image

Genetic transformation of hairy roots induced from petioles of detached leaves

Up to 36 petioles of soybean leaves were inoculated with A. rhizogenes carrying the binary vectors pPZP-empty, pPZP-GmPR10-eGFP-HygR, and pPZP-GmGST-eGFP-HygR and used in bioassays with M. incognita. Transgenic hairy roots were successfully screened for eGFP fluorescence (Fig. 1f). In the first round, hairy root induction rate and transformation efficiency were 88.9–100% and 83.3–96.9%, respectively (Table 2). For gene activation, up to 12 leaf petioles were inoculated with A. rhizogenes carrying binary CRISPR/dCas9 vectors. Using different CRISPR/dCas9 constructs containing sgRNA, hairy root induction rates reached transformation efficiencies of 83.3–100% and 33.3–63.6%, respectively (Table 2). The binary vectors containing the sgRNA that allowed higher GmExpA and GmExpLB gene activation were then selected for a new round of hairy root transformation and nematode bioassays. From up to 36 leaf petioles for each CRISPR/dCas9 binary vector, the hairy root induction rate and transformation efficiency were 86.1–88.9% and 71.9–80.6%, respectively (Table 2). In addition, using the empty pPZP vector as a negative control in both CRISPR/dCas9 experiments, hairy root induction rate and transformation efficiency of 73.3–87.5% and 77.3–100% were obtained, respectively (Table 2). Therefore, these collective data showed that the ex vitro hairy root system from petioles of unifoliate detached leaves was effective for soybean hairy root induction and transformation, and provided support for conducting the gene expression and transcriptional activation assays.

Table 2 Hairy root induction rate and transformation efficiency from soybean petioles of detached leaves
Full size table

Overexpression of the GmPR10 and GmGST genes by classical transgenesis

The GmPR10 and GmGST genes were significantly overexpressed in soybean hairy roots transformed with appropriate binary vectors compared to hairy roots transformed with the empty pPZP vector (Fig. 1b and d). In addition, transgenic soybean hairy roots were successfully screened for eGFP fluorescence (Fig. 1f). The relative expression of the GmPR10 and GmGST genes was 6 and 68-fold higher in transgenic soybean hairy roots compared with the controls, respectively. The expression profile of the GmPR10 and GmGST genes overexpressed in transgenic tobacco lines was also measured compared to non-transgenic (NT) plants (Fig. 1g and i). The eGFP fluorescence was successfully confirmed in transgenic tobacco plants (Fig. 1k). Therefore, these gene expression data showed that the ex vitro hairy root system allowed the overexpression of target genes and supported the performance of the nematode bioassays.

Transcriptional activation of the GmExpA and GmExpLB genes by CRISPR/dCas9

The transcriptional activation levels of the GmExpA gene in transgenic soybean hairy roots transformed with the CRISPR/dCas9 constructs carrying the sgRNA16 (which is also capable of activating the transcription of its paralog gene GmExpA2, Glyma.04G021600), sgRNA26, and sgRNA55 were 26.8-, 9.4-, and 45.1-fold higher, respectively, compared to soybean hairy roots transformed with the empty pPZP vector alone (Fig. 2d). Likewise, the transcriptional activation levels of the GmExpLB gene in soybean hairy roots transformed with the CRISPR/dCas9 constructs carrying the sgRNA13, sgRNA21, and sgRNA45 were 77.4-, 998-, and 3,818-fold higher, respectively, compared to soybean hairy roots transformed with the empty pPZP vector (Fig. 2g). Not surprisingly, dCas9 gene expression was proportionally higher in soybean hairy roots transformed with CRISPR/dCas9 constructs compared to soybean hairy roots transformed with the empty pPZP vector alone (Fig. 2e and h). As a result, soybean hairy roots transformed with the CRISPR/dCas9 constructs containing the sgRNA16 (GmExpA and GmExpA2) and sgRNA45 (GmExpLB) were selected to perform bioassays with nematodes. Therefore, these transcriptional activation data showed that the ex vitro hairy root system allowed a significant upregulation of the target genes using CRISPR/dCas9 strategies and the performance of  nematode bioassays. In addition, these data show that this hairy root system can also be used to validate CRISPR sgRNAs for CRISPR-based genome editing and CRISPR interference (CRISPRi) approaches.

Bioassays with nematodes in transgenic soybean hairy roots and tobacco plants

Transgenic soybean hairy roots overexpressing the GmPR10 gene showed a 41% reduction in the number of galls compared to transgenic hairy roots transformed with the empty pPZP vector (Fig. 1c; Suppl. Fig. S1c). In contrast, transgenic soybean hairy roots overexpressing the GmGST gene showed a 39% reduction in the number of galls compared to transgenic hairy roots transformed with the pPZP empty vector (Fig. 1e; Suppl. Fig. S1d). In comparison, stably transformed tobacco lines overexpressing the GmPR10 gene showed a 51.6–57.8% reduction in the number of galls compared to NT control plants (Fig. 1h). Meanwhile, transgenic tobacco lines overexpressing the GmGST gene showed a 52.2 to 56% reduction in the number of galls compared to NT control plants (Fig. 1j). These initial superexpression data showed that the ex vitro hairy root system could be an alternative to the stable transformation system for gene validation when nematode bioassays are required, as it is a more straightforward approach. On the other hand, transgenic soybean hairy roots transformed with the CRISPR/dCas9 vector carrying the sgRNA16 (GmExpA and GmExpA2) showed a 58.7% reduction in the number of galls compared to transgenic hairy roots transformed with the empty pPZP vector (Fig. 2f; Suppl. Fig. S1e). Similarly, transgenic soybean hairy roots transformed with the CRISPR/dCas9 vector carrying the sgRNA45 (GmExpLB) showed a 67.4% reduction in the number of galls compared to transgenic hairy roots transformed with the control vector (Fig. 2i; Suppl. Fig. S1f). Similar to classical transgenesis, the CRISPR/dCas9 data also revealed that the ex vitro hairy root system can be an alternative to the stable transformation system when bioassays with nematodes are required. These data showed that the ex vitro hairy root system is compatible for screening target genes with nematode bioassays.

Morphology and histology of galls produced in soybean hairy roots

Transgenic soybean hairy roots transformed with the empty pPZP vector showed multiple large and well-developed galls, containing nematodes that completed their life cycle after 30 dpi (Fig. 5a), characterized by the formation of egg masses. In contrast, transgenic soybean hairy roots overexpressing the GmPR10, GmGST, GmExpA and GmExpA2, or GmExpLB genes showed smaller galls with delayed nematode development (Fig. 5b–e) and consequently fewer egg masses. These data on gall morphology are consistent with the results showing that transgenic soybean hairy roots overexpressing the GmPR10 (Suppl. Fig. S1c) and GmGST (Suppl. Fig. S1d) genes, and with transcriptional activation of the GmExpA and GmExpA2 (Suppl. Fig. S1e) and GmExpLB (Suppl. Fig. S1f) genes are less susceptible to the nematodes. In addition, gall histology data showed that galls formed in transgenic soybean hairy roots transformed with the empty pPZP vector had well developed feeding sites with giant cells containing a dense cytoplasm (Fig. 5a). Galls formed in transgenic soybean hairy roots overexpressing the GmPR10 gene showed giant cells with apparently reduced size (Fig. 5b). Galls overexpressing the GmGST gene also exhibited apparently smaller giant cells with a less dense cytoplasm that was punctate in color and largely vacuolated, suggesting that they contain fewer nutrients needed for nematode development (Fig. 5c). GmExpA/GmExpA2 and especially GmExpLB overexpressing galls showed giant cells with more intensely stained cytoplasm and apparently thicker cell walls compared to the control pPZP empty vector (Fig. 5a, d, and e). Also, the egg masses present in gall sections of GmExpLB (30 dpi) showed significantly fewer eggs compared to those normally seen in control roots. Furthermore, the ex vitro hairy root system from petioles of soybean unifoliate detached leaves allowed the consistent validation of genes, using classical transgenesis and CRISPR strategies (i.e., sgRNA validation, gene activation), associated with nematode bioassays in a simple and faster manner (Fig. 6a–g).

Fig. 5
figure 5

Histological analyses of M. incognita-induced galls in transgenic soybean hairy roots at 7, 21, and 30 dpi. Gall sections were stained with toluidine blue and imaged by bright field microscopy. Squares with dotted lines indicate feeding sites. a Galls formed in transgenic soybean hairy roots transformed with the empty pPZP vector, showing well-developed feeding sites and the presence of well-developed nematodes. b Galls formed in transgenic hairy roots overexpressing the GmPR10 gene showed a clear delay in gall development with an apparent reduced feeding site area and a consequent delay in nematode development. c Galls formed in transgenic hairy roots overexpressing the GmGST gene showed giant cells with punctate staining and less dense cytoplasm compared to transgenic hairy roots transformed with the empty pPZP vector. d Galls formed in transgenic hairy roots with upregulated expression levels of GmExpA and GmExpA2 genes (sgRNA16) showed apparently smaller feeding sites. e Galls formed in transgenic hairy roots with an upregulated expression of the GmExpLB gene (sgRNA45) showed giant cells with apparently thicker cell walls and more stained cytoplasm compared to transgenic hairy roots transformed with the empty pPZP vector. *, giant cell; n, nematode; EM, egg mass. Scale bars = 50 µm

Full size image
Fig. 6
figure 6

Overview of the ex vitro hairy root system from petioles of detached soybean leaves (unifoliate leaf). a Soybean seeds are sterilized and germinated. b Agrobacterium rhizogenes are transformed with the desired binary vectors. c Soybean unifoliate leaves are removed at the petiole-stem junction. d A bacterial paste is prepared from the A. rhizogenes clones. e The bacterial paste is used for mechanical inoculation with a needle in the central portion of the leaf petiole. f A. rhizogenes-inoculated leaves are placed on a Petri plates and covered with a thin cotton layer and a Germitest paper. In addition, a glass slide is placed under the abaxial region of the leaf and the petiole is covered with wet cotton. g After hairy root induction, unifoliate leaves with hairy roots are acclimatized in vermiculite for 5 days and then inoculated with the nematode ppJ2

Full size image

Discussion

The selection and combination of genetic elements to compose a biotechnological strategy are important requirements to obtain more promising performances from classical transgenesis or CRISPR-based technologies (Basso et al. 2019, 2020b). However, the stable genetic transformation and the generation advancement of soybean is an expensive, time-consuming, and genotype-dependent process (Paes-de-Melo et al. 2020; Fragoso et al. 2022). It is therefore recommended to first validate the strategies and molecules in a simpler and faster system before moving on to stable transformation of soybean (Mendes et al. 2021a, b). The hairy roots induction and transformation mediated by A. rhizogenes have been widely used in different crops to validate novel molecules with biotechnological potential and to elucidate certain chemical and physiological processes (Estrada-Navarrete et al. 2007; Chandra 2012; Guimaraes et al. 2017b; Niazian et al. 2022).

To date, a variety of soybean explants have been used for hairy root induction and genetic transformation under in vitro conditions, such as hypocotyl (Kereszt et al. 2007), cotyledon (Chen et al. 2018; Mendes et al. 2022), root tip (Weber and Bodanese-Zanettini 2011), and shoot (Song et al. 2021). However, the hairy root induction and transformation efficiencies from hypocotyl, cotyledon, root tip, and shoot ranged from 25 to 80%, 30 to 60%, 14 to 52%, and 90 to 100%, respectively, most of which are very low efficiencies and require in vitro culture. Clearly, a system that produces a large number of hairy roots with a high transformation frequency in a short time is desirable (Cao et al. 2009). There is a clear genotype dependence for the hairy root induction in soybean (Mazarei et al. 1998; Cho et al. 2000; Cao et al. 2009). In addition, transformation efficiencies of different soybean cultivars under ex vitro or in vitro conditions ranged from 13.3 to 83.3% and 14 to 52%, respectively (Weber and Bodanese-Zanettini 2011). Furthermore, these previous data emphasized that the efficiency of soybean hairy root induction and transformation depends on the type of explant and the vegetative stage of the soybean plant.

More recently, Pereira et al. (2023) used an ex vitro method to demonstrate the successful induction and transgenesis of soybean hairy roots in Williams 82 and Embrapa BRS 537 cultivars using A. rhizogenes strain K599. In this case, the transformation efficiency was 73–80% for the BRS 537 cultivar and 53–93% for the Williams 82 cultivar, using binary vectors to overexpress target genes or perform RNAi strategies. The Williams 82 soybean cultivar, in particular, has been widely used in genetic engineering because it is very responsive in tissue culture and genetic transformation, and its genome has been sequenced and is freely available (Schmutz et al. 2010; Haun et al. 2011). In addition, this soybean cultivar has shown promising results for hairy root induction under in vitro conditions (Chen et al. 2018; Fan et al. 2020; Cheng et al. 2021). In particular, the Williams 82 soybean cultivar is susceptible to RKN and cyst nematodes, making it a good model cultivar for conducting proof-of-concept bioassays with nematodes (Kruger et al. 2008; Yang et al. 2021). This system can also be used to induce hairy roots in other soybean cultivars. However, the A. rhizogenes strain could be different in some cases, which requires further testing (Niazian et al. 2022).

In this study, the soybean cultivar Williams 82 was used to induce hairy roots from soybean cotyledons, hypocotyls, and petioles of detached leaves. Hairy root induction from soybean hypocotyl and cotyledon under in vitro conditions resulted in efficiencies of 21 and 91%, respectively. Despite the induction efficiencies and their use for other purposes, these two hairy root methods are not widely compatible for performing nematode bioassays (Lozovaya et al. 2007; Liu et al. 2012). This is because, for bioassays with RKNs, hairy roots produced from soybean cotyledons must be only partially covered with a substrate due to their root disposition, which negatively affects the development of this biological system (Kereszt et al. 2007). Similarly, hairy roots produced from soybean hypocotyls need to be grown until soybean plantlets are formed, which requires more time and space (Kereszt et al. 2007). For this reason, the detached leaf petiole system was the focus of this study because it showed consistent hairy root induction with an 86.2% efficiency under ex vitro conditions and high compatibility with bioassays using RKNs. In this method, hairy roots are produced from petioles of detached soybean leaves up to 15 days after inoculation with A. rhizogenes. Transgenic roots are then separated from non-transgenic roots and acclimatized in vermiculite for 5 days. ppJ2 nematodes are then inoculated and bioassays can be maintained for 30 days (no more than 60 days) until the evaluation. In addition, the entire procedure is performed in Petri plates, which require less space and cost, and provide the high humidity necessary for successful hairy root induction and transformation.

Given these advantages, this ex vitro hairy root system was used to screen candidate genes for improving plant resistance to nematodes using classical transgenesis and CRISPR/dCas9 approaches. For the gene overexpression strategy, the GmPR10 and GmGST genes were selected because of the positive correlation between the soybean proteomic and transcriptomic data and plant resistance to RKNs (Andrade et al. 2010; Xu et al. 2014; Morris et al. 2021; Arraes et al. 2022). Hairy root induction and transformation were successfully performed, and bioassay data confirmed that this ex vitro hairy root system was efficient for gene overexpression and allowed consistent screening of target genes. In addition to the reduced susceptibility data of the transgenic soybean hairy roots overexpressing these two genes to M. incognita, a significant correlation was obtained compared to stably transformed tobacco plants. These data support the potential of the hairy roots, produced from petioles of detached leaves under ex vitro conditions, for screening candidate genes through classical transgenesis to improve plant resistance to RKNs.

This ex vitro hairy root system was also evaluated for CRISPR/dCas9-mediated transcriptional activation systems. In particular, it was used to screen different sgRNAs for transcriptional activation of soybean target genes. The well-established CRISPR-Act2.0 system was used for this study (Lowder et al. 2018). Three sgRNAs were strategically designed to target each of the soybean GmExpA/GmExpA2 and GmExpLB genes. Particularly, the GmExpA (as well as GmExpA2) gene previously showed to be a molecule that is strongly downregulated during soybean parasitism by M. incognita in both soybean cultivars BRS133 (susceptible to nematodes) and PI595099 (resistant to nematodes) (Arraes et al. 2022). Consequently, the GmExpA gene was selected for transcriptional activation in transgenic hairy roots. Similarly, the GmExpLB gene corresponds to the ortholog of the Arachis duranensis AdEXLB8 gene, which has also been validated in A. duranensis as a resistance-associated gene against RKNs (Guimaraes et al. 2017a, b). However, this GmExpLB gene has not yet been evaluated for its potential to confer reduced susceptibility by overexpression in transgenic plants and was, therefore, also chosen for transcriptional activation in transgenic hairy roots. Successfully, transgenic soybean hairy roots showed high transcriptional activation of the GmExpA and GmExpLB genes, validating the functionality of the CRISPR/dCas9 strategy in the hairy root system. In the same sense, this ex vitro hairy root system proved to be effective for screening the best sgRNAs for subsequent use in transcriptional activation of stable transgenic plants. To date, sgRNAs have mainly been validated in soybean using protoplast transfection (Sun et al. 2015; Kim and Choi 2020) and in vitro hairy root systems (Li et al. 2019; Bai et al. 2020), both of which are considered costly and time-consuming methods.

In addition, transgenic hairy roots with upregulation of the GmExpA/GmExpA2 and GmExpLB genes showed reduced susceptibility to M. incognita, demonstrating that this system is also compatible with nematode bioassays. This characteristic greatly favors screening for target genetic elements in soybean hairy roots compared to the costly and laborious screening in stable transgenic plants (Homrich et al. 2012; Heenatigala et al. 2018). Similar results of reduced plant susceptibility to nematodes by overexpressing the expansin A gene in transgenic plants have already been demonstrated (Basso et al. 2022b). In particular, GmEXPA1 (Glyma.02G109100) gene expression was shown to be responsive to M. incognita infection, while its constitutive overexpression increased root lignification and reduced plant susceptibility to nematodes (Basso et al. 2022b). Therefore, the GmExpA/GmExpA2 proteins at basal or overaccumulated concentrations may act by conferring greater rigidity or relaxation of the root cell wall as a natural mechanism to adapt plant tissues to adverse (including plant parasitism by nematodes) or natural conditions (Cosgrove 2015). In the same vein, it is hypothesized that the GmExpLB gene may act similarly to its orthologous gene AdEXLB8 by interfering with phytohormone dynamics, leading to a defense primed state (Brasileiro et al. 2021).

Concerning gall histological studies, this ex vitro hairy root system from petioles of detached leaves efficiently demonstrated that reliable analyses of the plant-nematode interactions can be performed using this system. Overexpression of the GmPR10, GmGST, GmExpA/GmExpA2, and GmExpLB genes in soybean hairy roots visibly delayed the development of galls, nematodes, and egg masses. In addition, the presence of giant cells with less cytoplasmic content in transgenic hairy roots indicated a reduced ability of the nematode to parasitize the transgenic plant (Almeida-Engler et al. 2004; Banora et al. 2011; Kyndt et al. 2014). Therefore, all these data confirmed that this ex vitro hairy root system is an efficient method for screening candidate genes associated with bioassays involving RKNs using classical transgenesis and CRISPR/dCas9. Furthermore, the overexpression or upregulation (via CRISPR dCas9) of the GmPR10, GmGST, GmExpA/GmExpA2, and GmExpLB genes were highlighted as powerful strategies to reduce soybean susceptibility to RKNs.

Conclusions

This study presents an ex vitro system that is highly efficient for hairy root induction by A. rhizogenes-mediated genetic transformation. This system used hairy roots generated from petioles of detached soybean leaves and was validated using both classical transgenesis and CRISPR/dCas9 strategies. It proved to be a practical, time-saving and cost-effective method for screening target genes, sgRNAs, and CRISPR/dCas9 strategies directly in soybean. Furthermore, the GmPR10, GmGST, GmExpA/GmExpA2, and GmExpLB genes were also highlighted as potential targets for the development of biotechnological tools to reduce soybean susceptibility to RKNs.