To characterize broomrape attachment in tomato, we used a transparent, soil-less co-cultivation system, similar to previously established soil-less rhizotron approaches with slight modifications. This system consistently supported tomato root growth and maintained adequate moisture and nutrient levels throughout the experimental period (Fig. 1a and b). Following preconditioning in the dark and germination stimulation with GR24 (a synthetic strigolactone analog), branched broomrape seeds exhibited synchronized germination within five days of co-culture. Germinated branched broomrape seeds with an emerging radicle were observed along the primary and lateral roots of tomato, and haustoria attachment was confirmed using stereomicroscopy (Fig. 1c). By eight days post-inoculation, early-stage tubercles (localized swellings ~0.5 mm) were consistently observed along tomato primary and lateral roots, indicating host penetration and parasite development. By Day 14, tubercles began transitioning to the “early spider” stage, characterized by radial expansion and lateral branching of parasite tissue (Fig. 1d). This approach proved reproducible and allowed for non-destructive monitoring of broomrape development across multiple stages. Quantification of broomrape attachment across four commercially available tomato cultivars (Heinz hybrids H9553, H9775, and HM Clause HM58841 and HM4885) demonstrated parasite attachment and statistically significant differences observed between cultivars (Fig. 1e). The transparent pouch assembly further enabled repeated imaging and facilitated downstream applications including sampling for transcriptomic profiling and spatially targeted imaging and analysis of host–parasite interfaces.

To enable root-specific genetic and functional analysis in tomato during parasite infection, we evaluated the potential of hairy roots as a host system for branched broomrape. R. rhizogenes-induced tomato hairy roots were generated and maintained under sterile in vitro conditions. Media-based in vitro infections have precedent, including growing excised host roots on nutrient medium with parasite seeds presented on a low-nutrient surface, as well as callus-based infections on media, in which differentiated Orobanche calli were generated to infect the host rather than proceeding from parasite seeds (Losner-Goshen et al. 1996, Zhou et al. 2004). We retain the split-media logic but differ in three practical respects: (i) Host tissue: CRISPR-edited transgenic hairy roots enabling direct gene function analysis in tomato-broomrape interactions; (ii) Source of infection: infections initiated from preconditioned branched broomrape seeds to maintain the normal developmental sequence from seed to radicle to haustorium trajectory rather than callus-derived infections which is shown to lower infection rates; (iii) Microbial minimization: Earlier systems applied broad-spectrum antibiotic cocktails to the media to suppress contamination from broomrape seeds, here instead 0.06% Plant Preservative Mixture (PPM) was applied to the parasite (low-nutrient) compartment during preconditioning/placement. Higher PPM concentrations impaired broomrape germination and viability, confirming a narrow window for effective use. This localized PPM minimizes microbial growth yet preserves germination/attachment, and it keeps the host compartment antibiotic-free, reducing the risk of confounding effects on the infection process (Fig. 2a). Hairy roots grow and extend from the rich medium into the low-nutrient zone, where they encounter preconditioned broomrape seeds, which support parasite attachment to the host. This setup improved assay reproducibility and extended the duration of co-culture, allowing broomrape seedlings to attach successfully to tomato hairy roots within 10 days, with early-stage tubercles visible under stereomicroscopy (Fig. 2b and c). While this system effectively captured early-stage infection events, it did not support later-stage parasite development beyond the early spider stage, likely due to nutrient limitations in the split-media setup or a lack of photosynthetic tissue. Nonetheless, these results demonstrate that tomato hairy roots, cultured in a spatially partitioned media, provide a reliable system for dissecting host genetics in early-stage broomrape parasitism.

To verify that observed haustoria attachments resulted in functional parasitism, we performed confocal laser scanning microscopy on transverse sections of broomrape infected wild-type hairy roots. Samples were stained with Direct Yellow 96 to visualize cell walls and Basic Fuchsin to detect lignin deposition. In infected wild-type roots, broomrape intrusive cells were observed penetrating the host cortex and reaching the central vascular cylinder (Fig. 3a). Xylem continuity between host and parasite was also observed, indicating the formation of functional xylem–xylem bridges (Fig. 3b). To preserve the spatial organization of host–parasite interactions that is often disrupted during sectioning, cleared whole-mount roots were examined using differential interference contrast (DIC) microscopy. Rapid Potassium Hydroxide (KOH) clearing followed by toluidine blue staining enabled clear visualization of xylem–xylem connections along the longitudinal axis of the interface, marked by characteristic secondary wall thickenings (Fig. 3c and d). These results validate that the hairy root in vitro system supports functional parasitism and is suitable for investigating host genetic responses to broomrape.

One plant defense response to broomrape infection is cell wall remodeling, a well-established mechanism that restricts parasite penetration and haustorial development. Lignification, suberization, and reinforcement of the cell wall at the host–parasite interface have been reported in resistant species, acting as physical and biochemical barriers to infection (Mutuku et al. 2019, Jhu et al. 2022, Kawa et al. 2024). Recent work shows that tomato roots have an exodermis layer that forms a polar lignin cap (PLC), an apoplastic diffusion barrier functionally analogous to an endodermal Casparian strip, yet genetically independent. Two transcription factors, SCHIZORIZA (SlSCZ) and SlEXO1, were shown to regulate PLC polarity and restrict lignification to the exodermis (Manzano et al. 2025). We hypothesized that loss of SlSCZ would further induce lignin deposition outside the exodermal PLC upon broomrape challenge. To test this, we used hairy root mutagenesis to analyze CRISPR-edited roots targeting SlSCZ (Manzano et al. 2025).

Wild-type hairy roots, when challenged with broomrape, did not trigger ectopic lignin deposition at the attachment site, and a xylem–xylem bridge was established (Fig. 4a). In uninfected slscz mutant alleles, ectopic lignin deposition is observed and restricted to a few cells in the inner cortex (Fig. 4b and c), whereas broomrape infection triggered ectopic lignin deposition across multiple cortical layers surrounding the invading haustorium, supporting that broad lignification is infection-induced (Fig. 4d and e). Lignin signal was clearly higher at sites of parasite attachment in the slscz mutant alleles compared to infected wild-type roots (Fig. 4f). Tubercles formed on slscz roots developed late compared to the wild-type roots and failed to grow beyond the initial swelling (Fig. 4g). This data suggests that SlSCZ disruption primes cortical cells for a lignin-based defense response upon broomrape invasion and that lignin accumulation at the interface may prevent or delay parasite access to host vasculature. These results provide a framework for dissecting post-attachment resistance in tomato and for testing additional genes involved in exodermal barrier formation in parasite susceptibility.

Transparent germination pouches commercially supplied sterile (height: 30 cm, width: 25 cm—PhytoAB—SKU CYG-98LB) served as cultivation chambers, allowing direct visualization of parasitic developmental stages. Only the initial pouch assembly was carried out in the laminar flow hood to minimize contamination, and the system was not maintained under sterile conditions afterwards. A small front flap was cut with a razor to permit precise seed placement and repeated observation. GF/A glass-fiber sheets (Cytiva, 1820-915) and all nutrient solutions were autoclaved. Each pouch lined with a single autoclaved GF/A sheet was sealed with lab tape after plants were transferred, leaving only a minimal basal port for nutrient solution replenishment. This configuration maintained uniform moisture and supported robust tomato root growth without contamination for up to 3 weeks (adapted from Clarke et al. 2020 with modifications).

Branched broomrape seeds were collected from a broomrape-infested field in Yolo County, California, USA. Seeds were surface sterilized using chlorine gas. To generate chlorine gas, approximately 20 mL of household bleach (Clorox—8.25% sodium hypochlorite) was placed in a small glass beaker inside a desiccator positioned under a chemical fume hood. Then, 600 μL of concentrated hydrochloric acid (12N HCl) was carefully added to the bleach. The reaction released chlorine gas (Cl₂), and the desiccator was immediately sealed to contain the gas. Seeds were placed in open Eppendorf tubes held in a tube rack within the desiccator and exposed to chlorine gas for 3 h. After sterilization, the desiccator was opened under a fume hood to safely dissipate the gas. Seeds were then rinsed thoroughly using a 70 μm cell strainer (Sigma—CLS431751) with sterile distilled water to remove any residual chlorine before use. Sterilized seeds were preconditioned by spreading them on moistened filter paper (Whatman No. 1) in sterile Petri dishes (Sigma—PD20047S0). About 1 mL of 0.06% PPM (Plant Cell Technology) was added to each plate. Plates were sealed with Parafilm and incubated at 22°C in darkness for 7 days. Following preconditioning, seeds were rinsed with sterile water to remove excess PPM, and 1 mL of 10 μM GR24 (PhytoTech Labs, G3324) was applied to the dish to stimulate germination. Seeds were further incubated at 25°C in darkness for 24 h prior to co-cultivation with host tissue.

Tomato seeds (S. lycopersicum cv. M82; Heinz lines H9775, H9553, and HM Clause lines HM58841 and HM4885) were first germinated on half-strength Murashige and Skoog (MS) (Plant Cell Technology—Basal Salts) medium. At 7 days post-germination, seedlings were transferred into germination pouches and supplied with ¼ strength Hoagland solution containing 0.05 mM KH₂PO₄. Seedlings were grown in the pouches at 25°C (16/8 light cycle), with roots kept in darkness inside a filing folder hung inside a box while shoots were exposed to light. Once plants developed two true leaves, preconditioned seeds suspended in water were then applied with a paintbrush near the roots of tomato seedlings in the mid-section of germination pouches to promote host–parasite interactions. The pouches were sealed with tape and incubated vertically at 25°C. The bags were inspected every 2–3 days to monitor moisture levels. Beginning one week after transfer, 30 mL of ½ strength Hoagland solution (0.05 mM KH₂PO₄) was added as needed through a small opening at the base of each pouch to maintain moisture while minimizing parasite seed displacement.

Haustoria/tubercle formation was assessed under a stereomicroscope at 9 days post-inoculation. Late-stage tubercles were confirmed and quantified at 14 days post-inoculation.

Tomato (S. lycopersicum cv. M82) hairy roots were generated via R. rhizogenes (ATCC: strain 15834)–mediated transformation following a previously published protocol (Ron et al. 2014). Briefly, electrocompetent R. rhizogenes cells were electroporated with the desired binary vector and plated on nutrient agar (BD 247940) supplemented with spectinomycin (100 mg L⁻¹). Plates were incubated for 2–3 days at 28–30°C. Colonies that grew under selection were then inoculated into 10 mL of nutrient broth (BD 90002-660) containing the same antibiotic and cultured overnight at 30°C with shaking at 200 rpm. This overnight culture was used to infect tomato cotyledons harvested from 8–10-day-old sterile-grown M82 seedlings, prior to the emergence of true leaves. Tomato cotyledons were excised using a sterile scalpel and submerged for 20 min in the R. rhizogenes suspension adjusted to an OD₆₀₀ in 1× MS liquid medium. Following infection, cotyledons were placed (adaxial side down) onto solid MS medium (1× MS with vitamins (Plant Cell Technology), 3% sucrose, 1% agar) without antibiotic selection and incubated in the dark at 25°C for 3 days. They were then transferred to a selection medium containing cefotaxime (200 mg L⁻¹) and kanamycin (100 mg L⁻¹), and incubated again at 25°C. Typically, three to five transformed roots developed per cotyledon. Kanamycin-resistant roots were excised and subcultured on fresh selection medium. Typically for one construct, 15 independent transgenic roots were isolated for downstream imaging and genotyping analyses.

For detailed methodology, refer to the Methods section in Manzano et al. 2025. Briefly, guide RNAs were designed using CRISPR-PLANT to meet GC content and mismatch criteria. Gene sequences (ITAG3.2) were retrieved from Phytozome, and gene maps were constructed in SnapGene. Primers for genotyping were designed with Primer-BLAST and checked for specificity in S. lycopersicum. Single guide RNA (sgRNA) oligos were cloned into pYPQ131-3, assembled via Gateway cloning into p278, and then into pMR286 containing Cas9 and KanR. Final vectors were transformed into R. rhizogenes for hairy root transformation in S. lycopersicum (cv. M82, LA3475). Mutations were verified by Sanger sequencing. Two null slscz mutant alleles in hairy roots were infected with preconditioned broomrape seeds using the split-media system and analyzed 21 days post-inoculation for parasite attachment and cellular-level infection site characterization.

A two-compartment split-media system was prepared using square Petri dishes 120 × 120 × 17 mm (Greiner BIO-One-688 161) divided into two zones. The middle portion was filled with a nutrient-rich medium (1× MS + 3% sucrose) using a sterile cookie cutter (2 inch—round) to support hairy root growth, while the rest of the plate was filled with a non-nutritive medium (1% agar; BD 214530) to promote broomrape seed germination. To reduce fungal and microbial contamination from broomrape seeds, 0.06% PPM was added to the low-nutrient medium. Hairy roots were positioned such that they extended from the nutrient-rich side into the low-nutrient compartment. Preconditioned branched broomrape seeds suspended in 0.1% agarose were applied with a pipette to the surface of the low-nutrient zone and positioned near the roots using a dissecting needle. Plates were sealed and incubated at 25°C in the dark.

For rapid assessment of broomrape attachments, harvested roots were either cleared overnight in 10% KOH or quickly cleared by incubating in 10% KOH at 95°C for 5 min. Cleared root segments were gently rinsed three times with distilled water and incubated with 0.1% Toluidine Blue solution for 1 min and rinsed again gently with distilled water. Root segments were mounted on a slide in 50% glycerol, covered with a cover glass, and sealed around the edges with nail polish. Observations were recorded on a Leica compound microscope under DIC. Xylem strands are visible with secondary wall thickenings in spirals or networks.

Infected tomato root segments containing attached branched broomrape were excised and embedded in 4% (w/v) low-melting-point agarose. Embedded tissue blocks were sectioned at 100 μm thickness using a vibratome. Sections were transferred to ClearSee solution (Ursache et al. 2018) and incubated at room temperature for 1–2 days to enhance tissue clearing. Cleared sections were then stained with 0.01% (w/v) Basic Fuchsin for 30 min to visualize lignified xylem, followed by 15 min in 0.01% (w/v) Direct Yellow 96 to stain cell walls. After staining, sections were rinsed in ClearSee 3 times and mounted in the same solution on glass slides. Imaging was performed using a LSM 700 laser scanning microscope (Carl Zeiss) with the following settings: excitation/emission settings optimized for DY-96 (488 nm) and Fuchsin (561 nm). Attachment sites were examined for haustoria penetration, host cortex invasion, and formation of xylem–xylem connections between host and parasite.

For each image, the red channel was used to identify areas with magenta signal (Basic Fuchsin) based on visual inspection. Regions were selected using the freehand selection tool, and the mean intensity and area were measured using the Measure function. The total signal was then normalized to the root area to obtain an area-normalized fluorescence intensity for comparison across samples. All sections were processed as described above. In each experiment, laser power was calibrated on control samples and then kept constant for all subsequent imaging.

The data underlying this article are available in the article and in its online supplementary material.