A total of 27 melon cultivars ( Supplementary Table 1 ) were evaluated for resistance to P. aegyptiaca, together with details of the source of the seed and their origin. The melon cultivar Naxigan was used as a susceptible control for both rhizotron and pot assays (Peng et al., 2018). The origin and description of six P. aegyptiaca populations used in this study are shown in Supplementary Table 2 . Total genomic DNA was extracted from the seeds of six broomrape populations, and the ribosomal protein S2 (rps2) and ribosomal DNA internal transcribed spacer (ITS) region were amplified by PCR using the primer pairs rps2F/rps2R, ITS1/ITS4 (Anderson and Cairney, 2004; Park et al., 2007). Based on phylogenetic analysis of plastid ITS and rps2 (Schneeweiss et al., 2004; Park et al., 2007), we confirmed that our experimental species was P. aegyptiaca ( Supplementary Figure 1 ). O. cumana seeds were collected from mature plants parasitizing sunflowers at 188 Regiment (87° 33′15″ E, 46° 6′ 9″ N, 500 m elevation), Xinjiang, China in 2019. O. cumana seeds were used as a control to detect the effects of root exudates of different melon cultivars on the germination of P. aegyptiaca.

Rhizotron experiments were carried out under laboratory conditions as described by Labrousse et al. (2001) with minor modifications to observe the attachment and development of P. aegyptiaca throughout the study. Seeds per melon cultivar were sterilized with sodium hypochlorite solution (5% v / v) for 20 min, rinsed thrice in sterile deionized water, and germinated in petri dishes (15 cm diameter) on wet filter paper and kept in the dark at 30°C for 2–3 d (until the radicle reached 1 cm in length). Germinated seeds were grown in a 50-hole (5 × 5 × 5 cm) plastic cave dish containing vermiculite and were watered regularly with Hoagland nutrient solution. The seedlings were maintained in a growth chamber at 25°C with a day/night regime of 16/8 h and a 10 000 Lx light source. Briefly, 2-week-old seedlings were transferred to a rhizotron comprising a 15-cm petri dish filled with a sponge overlaid with glass fiber filter paper, and the plant roots were evenly spread out over the surface of the filter paper. To prevent exposure of host roots to direct light, the dishes containing plants were covered with aluminum foil. The rhizotrons were then kept in growth chambers at a temperature cycle of 28°C/25°C for a 16-h-light/8-h-dark cycle. After 1 week, the growing melon plants were inoculated with sterilized P. aegyptiaca seeds (Plot 2; approximately 400 seeds each) and were carefully placed next to the roots of each melon plant. The rhizotrons containing inoculated plants were placed back into the growth chambers and incubated under the same conditions until investigation.

Developmental stages of broomrape were recorded and classified at 35 d after inoculation (DAI) according to their stage of development using a scale from 0 to 9 (Westwood et al., 2012) with slight modifications: S0, P. aegyptiaca seeds were not germinated; S1, seed germination; S2, terminal haustorium formation (before the radical had firmly adhered to or penetrated the host roots); S3, broomrape had firmly adhered to or penetrated the host roots but had not formed a vascular connection; S4, broomrape had formed a vascular connection to the host roots and a tubercle had formed (before the development of secondary roots); S5, “spider” stage (tubercles with secondary roots) ≤ 1 cm; S6, sprout already visible remaining underground > 2 cm; S7, the start of the development of a spike; S8, flowering; S9, setting of seeds (Westwood et al., 2012).

Pot experiments were carried out to evaluate the resistance of 27 melon cultivars in a greenhouse in 2017 and 2018. The seeds of P. aegyptiaca were collected from Plot 2. Germinated melon seeds were sown in 1.5-L pots containing a mixture of sand-vermiculite-compost (1:1:1, v:v:v) and P. aegyptiaca seeds (15 mg of P. aegyptiaca seeds per 1 kg of the substrate for the 2017 experiment and 30 mg of P. aegyptiaca seeds per 1 kg of the substrate for the 2018 experiment). Noninfected plants were grown and evaluated in parallel. The pots were arranged in a completely randomized design under greenhouse conditions. The plants were grown in a growth chamber at 25°C–30°C, with a day/night regime of 16/8h and with a 10 000 Lx light source (800–900 μmol m⁻² s⁻¹). Pots were watered with about 150 ml of tap water every 5-7 days. Sixty days later, the plants were divided from the pots, and the roots were washed carefully in water. Attachments on host roots were counted and classified according to their stage of development using a scale from zero to nine as mentioned above.

The parasitism rate (percentage of melon plants infected by P. aegyptiaca) and the parasitic degree (the number of P. aegyptiaca per plant) were calculated. According to the resistance standards ( Supplementary Table 3 ), the resistance levels of different melon varieties to P. aegyptiaca were identified (Peng et al., 2018).

In addition, morphological traits of melon (plant height (PH), stem diameter (SD), fresh (FGD) and dry weights of total aboveground parts (DGD), fresh (RFW), and dry weight of roots (RDW)) were measured. The fresh and dry weights of broomrape were also measured.

Field screening experiments were carried out to assess the behavior of melon cultivars selected for their partial resistance to P. aegyptiaca. A sketch of the field layout is provided in Supplementary Figure 2 .

The P. aegyptiaca field trials were conducted at two sites; one was at the experimental station of the Shihezi University (May–Aug) in 2018, and the other was at Majiaping farm in Shihezi (May–Aug) in 2019 and 2021 ( Supplementary Table 4 ). All field trials were replicated three times using a completely randomized block design. For control experiments, melon cultivars were simultaneously planted in P. aegyptiaca-free plots within the same locations. At the experimental station, each plot representing an individual cultivar measured 3 m × 3 m (9 m²) and contained one row of 10 melon plants with a plant distance of 0.3 m × 0.3 m ( Supplementary Figure 2 ). At Majiaping, each plot measured 6.5 m × 3 m (19.5 m²) with two rows of 20 melon plants with the same plant distance as in the experimental station ( Supplementary Figure 2 ). Plots were separated by one open row of 0.6 m to avoid neighbor effects and to allow easy access. From sowing onwards, each trial was regularly hand weeded (at least every 2–3 weeks) to remove all weeds other than P. aegyptiaca. The number of aboveground P. aegyptiaca plants in each plot was counted at 45, 60, and 80 d. At the maturity stage of melon, the fruit yields of noninfested and naturally infested fields by P. aegyptiaca were recorded.

Three melon cultivars (KR1326, Huangpi 9818, K1076) were used to evaluate the resistance against broomrape from six populations ( Supplementary Table 2 ) in pot experiments. The KR1326 and Huangpi 9818 cultivars were selected as potential resistant candidates showing postattachment resistance, and K1076 was selected as a susceptible cultivar based on the results of rhizotron, pot, and field experiments. A 1.5 L plastic pot was filled with 1.5 L of a mixture of compost, vermiculite, and sand (2:1:1 v:v:v) together with P. aegyptiaca seeds (50 mg seeds). Growth conditions and measurements of traits were the same as in the pot experiment.

The collection method was as reported in Yoneyama et al. (2008) with several changes. Furthermore, 2-week-old seedlings (n = 12) of each cultivar (KR1326, K1076) were transferred to a 12-hole hydroponic box (16 × 20 × 7 cm, width × length × height (W × L × H)) containing 750 ml of 1/2 Hoagland nutrient solution as the culture medium in a growth chamber with a day/night regime of 16/8 h and a 10 000 Lx light source at 25°C. The Hoagland medium was exchanged every 3 d. After 10 d, melon seedlings were transferred to 1/2 Hoagland medium without P and grown for 3 d. Then, 3 d later, the roots of seedlings were rinsed with tap water, and tap water was used as the culture medium. Root exudates were collected every 3 d, four times in total, and each cultivar was repeated three times. Root exudates were immediately extracted with an equal volume of ethyl acetate three times. The ethyl acetate extracts were combined, dried over anhydrous MgSO₄, and concentrated in vacuo. These crude extracts were stored in sealed glass vials at 4°C until use.

At the end of the final extraction, 2 g of the hydroponically cultivated melon roots were collected and used for quantitative detection of strigolactones (strigol and 5-deoxystrigol) in melon roots. After liquid nitrogen freezing and grinding, 15 ml of precooled acetone was added to the roots. The mixture was vibrated for 30 min by ultrasound, vortexed and mixed, and extracted overnight at 4°C in a shaker. After centrifugation (18514 g for 5 mins at 4°C), the acetone supernatant was taken and evaporated by nitrogen flow. When dry, 2 ml 20% acetone was added, and the solution was passed through an HLB (Hydrophile-Lipophile Balance) column (the passed fraction contained SLs) and repeated three times. Then, 2 ml of 100% acetone was added. Finally, the acetone was blown dry using nitrogen, re-dissolved with 200 μl of acetonitrile, and detected by HPLC-MS/MS (Yoneyama et al., 2008; Motonami et al., 2013). Characterization of strigolactones in the roots of melon was conducted by comparing retention times of germination stimulants on HPLC with those of synthetic standards (strigol and 5-deoxystrigol, strigolab) and by using HPLC-MS/MS.

Germination assays of P. aegyptiaca (Plot 2) and O. cumana seeds were conducted as reported previously (Bao et al., 2017). The crude extract of root exudates was dissolved with 1 ml isopropanol and then diluted with sterile water 4, 10, 1 × 10², 1 × 10³, 1 × 10⁴, 1 × 10⁵, and 1 × 10⁶ times. Then, 1 × 10^-7 M GR24 (positive control) and sterile water (negative control) were used ascontrols. To exclude the effect of isopropanol on seed germination, isopropanol (4, 10, 1 × 10², 1 × 10³, 1 × 10⁴, 1 × 10⁵, and 1 × 10⁶ times, diluted with sterile water) was also assessed. Then, 500 μl of the test solution was used for P. aegyptiaca germination, with three replications for each test solution. The germination rates of P. aegyptiaca and O. cumana seeds were examined with a stereo microscope at 7 d after treatment.

The characterization of the resistant/susceptible phenotypes of melon was conducted by rhizotron infection experiments. The melon-P. aegyptiaca system was observed and photographed with a stereo microscope (Zeiss Discovery.V20) equipped with a digital camera (AxiaoCam ICc 5).

To examine the stage of parasite development on the host roots, small sections of tissue were dissected from the host root system at the S4 and S5 stages. The fresh root samples were fixed in FAA (5% formaldehyde + 10% acetic acid + 50% ethanol, in water) for 48 h. Paraffin sections were prepared as described previously (Bai et al., 2020). Alcian green-safranin double staining of sections was performed as described previously (Yoshida and Shirasu, 2009).

The sections were examined under a light microscope (Axio Imager M2, Zeiss, Germany) and photographed using a XiaoCam MRC digital camera. For visualization of total lignin autofluorescence, the sections were observed by epi-fluorescence under excitation at 450–490 nm. To obtain comparable images, fluorescence intensity was kept constant between samples.

For callose detection, root segments with attached haustoria were excised, and we hand-sectioned plants at 200–300 μm thickness using razorblades and kept the sections in 4°C water before staining. The sections were stained for 15–30 min in a solution of 0.1% aniline blue fluorochrome in water (Bordallo et al., 2002). Aniline blue fluorochrome was used for callose detection under UV fluorescence (340–380 nm).

ANOVA was performed using the SPSS 19.0 software. In the rhizotron evaluation and pot experiments, the number of P. aegyptiaca seedlings in each growth stage was examined by Tukey’s HSD test with melon cultivar as a factor. The data of rhizotron evaluation were presented as relative means ± SE in the form of graphs using the statistical package R (version 4.1.2, ggplot2). In the pot experiments, PH, SD, GDW, GFW, and RDW of melon cultivars were analyzed by Tukey’s HSD test. Student t test P values > 0.10, < 0.1, < 0.05 and < 0.01, < 0.001 were marked by NS, +, *, **, and ***, respectively. Photoshop (version 2019) was used for image processing. The details concerning statistics used in data analyses performed in this study are described in the respective sections of results and methods.

Large phenotypic differences were observed among 27 melon cultivars in the various steps of development, from P. aegyptiaca seed germination to establishment on melon roots ( Figure 1 ). Of the seed inocula, failure to germinate was highest in KR1222 (47.05%) and lowest in Naxigan (24.15%).

Development of the parasite at stage S1 (seeds germinated but fail to attach to host roots) was common among all cultivars. Arrested development at this stage was highest in Jingpinxiaoxiangfei (46.54%) followed by Naxigan (46.35%) and lowest in Jingtianmi No.17 (24.55%). Many of the attachments did not develop past the S1 stage due to a lack of haustorium formation.

Arrested development of the parasite at stage S3 was observed, where seeds had firmly adhered to or penetrated the host roots but had not formed a vascular connection. Successful attachment but without the formation of a tubercle was highest in Tianmicui (36.08%) and lowest in Xinmi No.28 (13.84%).

The arrested development of P. aegyptiaca at stage S4, necrosis of the tubercles, was observed in five cultivars (Jingpin 2010 0.027%, Qinghuami 0.052%, KR1326 1.37%, Huangpi 9818 0.54%, Tianmicui 1.32%) and was highest in KR1326 (1.37%) ( Figure 1 ; Supplementary Figure 3F ). In this cultivar, 1.55% of the inoculated parasites underwent arrested development at stage S4, of which the necrotic tubercles of broomrapes accounted for 88.39%.

The P. aegyptiaca seedlings at stage S5, the “spider” stage (tubercles with secondary roots), were observed in 20 cultivars and varied widely among cultivars, representing 0–2.97% of the original inocula ( Figure 1 ). In seven cultivars (Xuemi, Huangpi 9818, Baimei, Qingcuimi, KR1327, Xinmi No.28, and KR1326), the inoculated parasites could not grow up to the S5 stage. Some necrotic broomrape at the S5 stage was observed in five cultivars (K1076, Qinghuami, Jingpinxiaoxiangfei, Mibao No.1, and Jingpin 2010).

The number and biomass of P. aegyptiaca and the plant height, stem diameter, and dry weight parameters of the host were investigated to evaluate the resistance of melon to P. aegyptiaca in pot experiments ( Figure 2 and Supplementary Tables 5 , 6 ).

As shown in Figure 2 , the results of the 2018 pot experiment were similar to those in 2017. Among these cultivars, the parasitism rate and degree of nine cultivars (K1217, KR1328, Mibao No.1, Naxigan, Qingcuimi, Tianmicui, Baimei, Fengwei No.8, and Xinmi 28) in 2017 pot experiments were higher than those in 2018, and the parasitism rate and degree of the other cultivars were lower than those in 2018. The mean values of parasitism rate and parasitism degree in the two years varied from 20.54 ± 10.69 to 100.00 ± 0.00 and from 0.27 ± 0.15 to 10.17 ± 2.42, respectively. According to the mean values of parasitism rate and parasitism degree of 27 melon cultivars, these could be divided into four groups: 1) a highly sensitive group comprising KR1222 with the highest parasitism degree (10.17 ± 2.42); 2) a sensitive group of 16 cultivars including K1076, K1217, Huang No.25, K1237, KR1328, Huangfei, Tianmicui, and others; 3) a resistant group composed of K1386, K1526, Jingpin 2010, Xinxuelihong, K986, Xuemi, Jingtianmi No.17, and K1238; 4) a highly resistant group including KR1326, and Huangpi 9818. Huangpi 9818 showed a lower parasitism rate and degree over the two consecutive years, significantly lower than other cultivars, and had no significant difference with KR1326.

As shown in Supplementary Table 5 , P. aegyptiaca infection in various melon hosts was variable with respect to the number and biomass of P. aegyptiaca attachments. Susceptible melon had more broomrape at S6 and S7 stages and greater broomrape biomass. The arrested development of broomrape at S4, the tubercle stage, was observed in Xuemi and KR1326 and was lowest in KR1326 (0.50 ± 0.27). There were nine cultivars (K1526, Jingpin 2010, KR1328, Huangpi 9818, Mibao No.1, Qinghuami, Baimei, Jingtianmi No.17, and K1217) in which arrested development of P. aegyptiaca was observed at S5, the “spider” stage (tubercles with secondary roots). The arrested development of the parasite at S6 (sprout stage) was observed in 11 cultivars (K1237, K1076, KR1222, Jingpinxiaoxiangfei, Tianmicui, Xinmi 28, Fengwei No.8, K986, KR1327, and Qingcuimi, K1238). In five cultivars (Huang No.25, Huangfei, Naxigan, Xinxuelihong, and K1386), broomrape developed to the S7 stage (spike stage). The lowest number of broomrape attachments (0.50 ± 0.27) and broomrape biomass were observed in KR1326. The heaviest attachment dry weight of broomrape was observed in Huang No. 25 and was significantly higher than the other cultivars, followed by K1237, Huangfei, K1076, Naxigan, and KR1222. The attachment dry weight of broomrape on K1238, Huangpi 9818, Mibao No.1, Qinghuami, Baimei, Xuemi, Jingtianmi No.17, K1217, and KR1326 was close to zero.

The broomrape had no significant adverse effects (P > 0.10) on the total aboveground parts dry weight (GDW), root dry weight (RDW), plant height (PH), or stem diameter (SD) in Qingcuimi, Jingpin 2010, Tianmicui, KR1327, KR1326, Xuemi, Huangpi 9818, and Baimei ( Supplementary Table 6 ). However, there were significant differences (P < 0.001, 0.01, or 0.05) in GDW, RDW, and PH in KR1222, K986, K1076, Huangfei, K1237, Naxigan, Jingpinxiaoxiangfei, and Huang No.25. The broomrape also had no significant adverse effects (P > 0.10) on the GDW in Xinxuelihong, Fengwei No.8, Xinmi 28, K1526, Qinghuami, Mibao No.1, K1238, K1217, KR1328, Jingtianmi No.17, Qingcuimi. The parasite reduced the RDW of Xinxuelihong, Fengwei No.8, Xinmi 28, K1526, Qinghuami, and Mibao No.1 by 45.71% (P < 0.01), 54.17% (P < 0.01), 44.83% (P < 0.05), 43.75% (P < 0.05), 28.57% (P < 0.05), and 36.36% (P < 0.05), respectively. P. aegyptiaca, also significantly reduced PH of eight cultivars, and the observed reductions were 17.65% (P < 0.01), 9.66% (P < 0.05), 41.50% (P < 0.01), 26.03% (P < 0.05), 21.42% (P < 0.001), 44.54% (P < 0.01), 16.55% (P < 0.05), 28.72% (P < 0.05) for Xinxuelihong, Fengwei No.8, Xinmi 28, K1526, K1238, K1217, KR1328, and Jingtianmi No.17, respectively. In addition, SD was reduced by 27.04% (P < 0.01) and 10.68% (P < 0.05) in Naxigan and Huang No.25, respectively.

Taken together, our results suggested that for the cultivars that were screened, K1217, Jingpin 2010, KR1326, Xuemi, Huangpi 9818, Baimei, and Jingtianmi No.17 were promising cultivars against P. aegyptiaca with low parasitism rate and parasitism degree and the growth is less affected by broomrape. In particular, Huangpi 9818 and KR1326 could provide valuable resistance against P. aegyptiaca.

Based on the numbers of aboveground P. aegyptiaca observed in the field, KR1326 was the most resistant cultivar in both the experimental station (2018) and Majiaping (2019, 2021) ( Figure 3A ). In 2018, the numbers of aboveground P. aegyptiaca of KR1326, Huangpi 9818, Jingtianmi No.17, K1217, and Xuemi (1.25 ± 0.26, 1.06 ± 0.38, 1.18 ± 0.18, 2.45 ± 0.44, 1.59 ± 0.47) were significantly lower than those of K1076 and Baimei.

In 2019, the aboveground numbers of P. aegyptiaca of KR1326 (1.44 ± 0.26) were notably lower than those of K1076, Xuemi, and Baimei (12.00 ± 1.18, 8.75 ± 1.39, and 5.80 ± 0.56) but not significantly different from Huangpi 9818, K1217, and Jingtianmi No.17 (2.33 ± 0.40, 3.61 ± 0.64, and 17 3.26 ± 0.56).

In 2021 ( Figures 3A, B ), the numbers of aboveground P. aegyptiaca of all cultivars were significantly higher than in 2019. The lowest was KR1326 (9.74 ± 1.58) followed by Huangpi 9818 (14.32 ± 1.61) and K1217 (16.43 ± 3.12), and the highest was K1076 (32.24 ± 5.46). The resistance to P. aegyptiaca among cultivars varied; for example, Xuemi, KR1326, and K1076 showed different resistance in 2021 ( Figure 3C ). KR1326 exhibited the highest resistance to P. aegyptiaca and was more resistant than Jingtianmi No.17 and Xuemi ( Figures 3A, C ). This cultivar had very few emerging broomrape shoots through the soil ( Figure 3C ).

In summary, the field experiments confirmed the laboratory results, i.e., that some melon cultivars were resistant to P. aegyptiaca. In particular, KR1326 showed the most resistance at all field sites.

In the P. aegyptiaca infested fields, the average single melon weight and yield per cultivar ranged from 0.07 to 0.86 kg (average: 0.42 kg) and 0.70 to 8.81 t ha^-1 (4.3 t ha^-1) in 2018, and from 0.69 to 1.91 kg (average: 1.55 kg) and 7.06 to 19.58 t ha^-1 (15.40 t ha^-1) in 2021 ( Figures 3D–G ). The single melon weight and yield per cultivar in 2018 were significantly lower than in 2021. This difference was due to planting density and soil fertility.

Broomrape affected the yields of all melon cultivars compared with their controls, and the effect was more severe on the susceptible cultivars. Single melon weight and yield were significantly decreased for K1076, Xuemi, and Baimei in 2018 (at the experimental station) and 2021 (at Majiaping). Huangpi 9818 and KR1326 yield decreased by 12.03% and 14.98% in 2018 and by 15.88% and 10.88% in 2021, respectively. However, in some cultivars, the effects of broomrape were more significant and had a close relationship with the field trial results. For example, broomrape infestation reduced the K1217 yield by 25.87% at the experimental station in 2018, but not at Majiaping in 2021. Jingtianmi No.17 yield decreased by 13.94% (2018) with nonsignificant differences compared with the control but decreased significantly by 28.41% in 2021. These results suggest that broomrape altered the yields of susceptible cultivars more drastically compared to resistant cultivars. In addition, Huangpi 9818 and KR1326 yield loss was minor in K1076 in all field sites.

The results of rhizotron and pot experiments showed that KR1326 and Huangpi 9818 had better resistance to broomrape from Plot 2 that had been collected from processing tomatoes growing in 3rd company, Junhu, Xinjiang. To determine whether the resistance of KR1326 and Huangpi 9818 was broad-spectrum or specific to the Plot 2 P. aegyptica population, KR1326, Huangpi 9818, and K1076 (the most susceptible cultivars) were infected with P. aegyptiaca from six different populations (including Plot 2).

Figure 4 shows the attachment number and the biomass of each P. aegyptiaca group on the different melon cultivars. KR1326 and Huangpi 9818 showed the same pattern of resistance to Plot 1, Plot 2, Plot 4, and Plot 6 broomrape populations, although they were collected from different host species and different regions ( Supplementary Table 2 ). K1076 was very susceptible to all populations of P. aegyptiaca, especially to the Plot 5 population. The amount of P. aegyptiaca biomass (AFW and ADW) of Plot 5 on K1076 was significantly higher than for other broomrape populations. Similar to K1076, the number of Plot 5 P. aegyptiaca at S5 and S6 stages and the amount of P. aegyptiaca biomass on KR1326 were significantly higher than in plants treated with P. aegyptiaca from other populations. In addition, the pattern of resistance of Huangpi 9818 to Plot 3 P. aegyptiaca differed from those of Plot 1, Plot 2, Plot 4, Plot 5, and Plot 6 P. aegyptiaca. A large number of attachments (S4 and S5 stage) of broomrape were observed on Huangpi 9818 roots, significantly higher than in other broomrape populations. A small number of broomrape plants from Plot 1 developed to S6 and S7 stages, while the development of others stagnated at the S4 and S5 stages. These results suggested that the differences in the number of different developmental stages and biomass of parasites from the same P. aegyptiaca population on different hosts were related to the strength of host-P. aegyptiaca interactions. In addition, six P. aegyptiaca populations showed consistency in the resistance response of KR1326 and Huangpi 9818 to P. aegyptiaca.

The plant height, stem diameter, and dry biomass of the infected melon cultivars were compared with those of the noninfected plants to evaluate the impact of P. aegyptiaca infection on the growth of the host. P. aegyptiaca affected the growth of KR1326, Huangpi 9818, and K1076, and more severely affected the susceptible cultivar K1076 ( Figure 5E ). The parasitism by broomrape had a greater effect on the total aboveground parts dry weight, root dry weight, and plant height of the hosts, while the effect on the stem diameter was less, although the difference was significant ( Figures 5A–D ). Among these three cultivars, the plant height, total aboveground parts dry weight, and root dry weight of the susceptible cultivar K1076 were most influenced by broomrape, being reduced by 33.17–72.97%, 30.51–54.63%, and 62.42–83.35%, respectively. In the two most resistant cultivars, KR1326 and Huangpi 9818, total aboveground parts dry weight, root dry weight, and plant height were slightly suppressed, but still significantly ( Figures 5A–D ). These results suggest that P. aegyptiaca altered the biomass of susceptible cultivars more compared to resistant cultivars.

The resistant cultivar KR1326 was selected for detailed germination-inducing activity and microscopic characterization. The germination induction effect of KR1326 root exudates in comparison with K1076 on seeds of P. aegyptiaca and O. cumana is shown in Figure 6 . In all cases, null germination was observed when both P. aegyptiaca and O. cumana seeds were treated with distilled water (negative control). After treatment with GR24 at 1 × 10^-7 M (positive control), the two parasites’ seed germination rates were 94.29% ± 1.35% and 42.79% ± 1.51%, respectively.

The results showed that different concentrations of isopropanol did not affect the germination of P. aegyptiaca or O. cumana seeds ( Figures 6A, B ). At the same time, the data showed that the germination percentage of P. aegyptiaca was higher than that of O. cumana treated with K1076 and KR1326 root exudates. Treatments with 4, 10, 1 × 10⁴, 1 × 10⁵, and 1 × 10⁶ times the concentration of root exudates of K1076 and KR1326 induced negligible to little (0–4.8%) germination, while 1 × 10² times the concentration of root exudates of K1076 and KR1326 induced high germination of P. aegyptiaca seeds (86.42 ± 2.10% and 87.70% ± 1.37%, P = 2.47) and O. cumana seeds (24.39% ± 4.46% and 21.34.01% ± 3.67%, P = 0.58) ( Figures 6A, B, D ). In addition, 1 × 10³ times the concentration of root exudates from K1076 (87.53% ± 0.70%) also highly induced seed germination of P. aegyptiaca, and the percentage was higher than in KR1326 (75.93% ± 4.05%) (P = 0.05). However, germination of O. cumana seeds with 1 × 10³ times the concentration of root exudates from K1076 (18.05% ± 2.58%) was significantly higher than that induced by KR1326 (1.33% ± 0.17%) (P = 0.003).

HPLC-MS/MS analyses of 5-DS and strigol of root extracts of K1076 and KR1326 are shown in Figure 6C . The 5-DS was detected in roots of both K1076 and KR1326. The concentration in KR1326 root extracts was 0.27 ng/g, slightly higher than for K1076 (0.22 ng/g) (P = 0.008). Strigol was not detected in the roots of either melon cultivar.

Melon cultivars displayed striking differences in their ability to support the growth and development of P. aegyptiaca. To understand how melon cultivars resisted P. aegyptiaca parasitism, we performed microscopic observations of P. aegyptiaca haustoria on K1076 and KR1326 roots at 7, 9, 16, and 23 DAI. Here, 7- and 9-d-old parasites on KR1326 and K1076 were phenotypically indistinguishable ( Figure 7A ). No clear discoloration was found on the roots of either cultivar at the parasite attachment site at this early stage ( Figure 7A ). This indicates that the hypersensitive response (HR) was not involved in the resistance response. At 16 DAI, broomrape had successfully penetrated the K1076 and KR1326 root cortex and established connections with the host vascular system. There was the formation of an enlarged tubercle, an indication of the successful vascular connection between the host and parasite. Compared with K1076, broomrape on KR1326 only showed minor distinguishable phenotypic differences of slight discoloration or browning on KR1326 ( Figure 7A , black arrow). However, resistant interactions had significantly more discolored tubercles (S4B) (P < 0.001) and lesser healthy tubercles (S4H) (P < 0.001) by comparison to susceptible interactions ( Figure 7B ). However, at 23 DAI, dramatically different phenotypic responses were observed among the susceptible and resistant interactions (P < 0.01 for S4H; P < 0.001 for S4DB and S5H). The parasite’s secondary roots thrived in susceptible interactions, and the haustoria significantly expanded ( Figure 7A ). In stark contrast, the appearance of the parasite on KR1326 showed severe browning and was significantly smaller ( Figure 7A , red arrow) than the parasite on K1076. This indicates that the resistance of KR1326 against P. aegyptiaca may be related to the inhibition of the growth and development of the parasite.

To clarify the changes in cell structure associated with the resistance mechanism by which the host root inhibits parasite development, histological analyses were performed on sections of infected roots containing an attached parasite tubercle ( Figure 8 ). At 16 DAI, in K1076 and KR1326 roots the P. aegyptiaca endophyte (en) and xylem cells invaded the central cylinder and established vascular connections ( Figures 8A, C ). The cells of broomrape occupied the host central cylinder tissues. Compared to the resistant cultivar KR1326, a large number of densely blue-stained cells (red arrow) are present in the apical part of the parasite in K1076 roots, indicating that the haustorial tissues had begun to differentiate ( Figure 8A ). A transverse section through a haustorium on K1076 at 23 DAI showed well-developed xylem (px), young stem meristem (red arrow), and root meristem (white arrow) ( Figure 8E ). Ultraviolet (UV) autofluorescence further confirmed that lignin autofluorescence was more robust in the xylem tissues of parasites (red triangle) in K1076 roots ( Figure 8F ) than in KR1326 ( Figure 8H ).

In contrast, for the roots of KR1326, the haustorium differentiation of the parasite was not apparent. A more detailed observation of the resistant interactions showed that the host xylem cells in contact with the parasite intrusive cells presented a thickening of their walls stained intensely red with safranin ( Figure 8K ). We also observed the reddish coloration in the intercellular spaces between the host xylem and intrusive parasite cells (black triangles). This phenomenon may be caused by the accumulation of large amounts of secondary metabolites at this site. In addition, columns of discontinuous xylem (red box) and blue-stained parenchyma cells (white border) of the broomrape embedded between the endodermis and xylem of the host root were observed ( Figure 8K ). Observation of these sections under a fluorescence microscope showed an intense green fluorescence in both the xylem of the parasite inside the host (red triangle) and host (white triangle) ( Figure 8L ). The above results show that vascular continuity was established between parasite and host but that there was poor haustorium tissue differentiation on KR1326 compared with K1076. This raises the possibility that KR1326 either lacks key specific primary/secondary metabolites/signals necessary for haustorium differentiation or that the presence of a P. aegyptiaca-specific metabolite(s) prevents haustorium development.

Aniline blue fluorochrome staining was used to detect callose accumulation at the P. aegyptiaca-KR1326 resistant interaction sites ( Figures 9E–H ) and P. aegyptiaca-K1076 susceptible interaction sites ( Figures 9M–P ). The melon roots of non-infected K1076 (Figures 9I–L) and KR1326 ( Figures 9A–D ) were used as controls. Some callose deposition occurred in cells at the KR1326-parasite interface and host endodermis ( Figures 9G, H ). The resistant interaction showed stronger fluorescence than susceptible interaction ( Figures 9O, P ).

In many previous studies, the resistance level to Orobanche spp. in diverse cultivated crops was evaluated using different approaches and parameters based on the number of Orobanche tubercles/shoots per host plant and the impacts on yield and host development (Labrousse et al., 2001; Echevarría-Zomeño et al., 2006; Trabelsi et al., 2016; Bai et al., 2020).

In this study, a rhizotron experiment was performed to evaluate the resistance of 27 melon cultivars against P. aegyptiaca. Regarding the analyses of P. aegyptiaca resistance in the rhizotron experiment, the results showed that K986, Mibao No.1, Jingpin 2010, Xinxuelihong, K1217, Huangpi 9818, Baimei, Qingcuimi, KR1327, Xinmi No.28, and KR1326 with lower tubercles numbers were highly resistant against P. aegyptiaca compared to Naxigan. However, the resistance ranking by the parasitism degree and rate in the pot experiment showed that K1217, Mibao No.1, Qingcuimi, Baimei, KR1327, and Xinmi No.28 were susceptible to P. aegyptiaca. This inconsistency likely occurred because of the variability in broomrape sizes. For example, a melon cultivar may have multiple small-sized attachments (in tubercle stage) with postattachment resistance. This variation likely skewed the data and gave the impression that this cultivar was susceptible to broomrape. Previous studies showed that the mean number of attachments and the parasitism degree and rate did not accurately determine the resistance of host plants (Trabelsi et al., 2016; Mbuvi et al., 2017). For this reason, mean broomrape biomass seemed to be a good measure of resistance, as it considered both the number of attached parasites as well as their size. Rubiales et al. (1996) suggested that a screening based only on the number of emerged stems was misleading and that the health of the host plant must also be considered. In the pot experiment, K1217, Jingtianmi No.17, KR1326, Xuemi, Huangpi 9818, and Baimei showed better resistance. P. aegyptiaca had no significant adverse effects on GDW, RDW, PH, or SD in these cultivars, and the biomass of broomrape was significantly lower than in the susceptible control.

Laboratory analysis provided a unique opportunity to minimize environmental variation and ensure that the observed resistance rankings could be confidently associated with the cultivar (Mbuvi et al., 2017). However, these data did not always translate to the field due to other factors that may influence resistance through genetic χ environment interactions, which have been observed before for Striga and Phelipanche ramosa (Rodenburg and Bastiaans, 2011; Rodenburg et al., 2015; Stojanova et al., 2019). Therefore, field trials were carried out to verify the results of laboratory observations. The result was similar to the resistance rankings of melon cultivars screened in the laboratory. Field trial results showed that among seven tested cultivars (K1217, Jingtianmi No.17, KR1326, Xuemi, Huangpi 9818, Baimei, and K1076), Huangpi 9818 and KR1326 showed stable resistance levels to P. aegyptiaca at all field sites as expressed by a low number of emerged broomrape spikes. In addition, the yields of Huangpi 9818 and KR1326 were less affected by P. aegyptiaca, and the difference was not significant compared to the controls. By contrast, Xuemi and Baimei had 1.59–7.75 and 5.8–6.5 emerged broomrape spikes, significantly more than KR1326.

Our study indicates that the resistance in KR1326 and Huangpi 9818 are relatively broad spectrum and therefore potentially of great interest to farmers in the short term as well as for longer-term studies of the genetic basis of resistance for breeding. And, we found that Plot 5 P. aegyptiaca showed more underground attachments and biomass on KR1326 and emerged shoots on K1076 than broomrape from other populations. This indicates that the parasitic ability of Plot 5 was stronger than in P. aegyptiaca from other broomrape populations. However, Plot 3, 4, and 5 broomrapes were collected from processing tomatoes in Bayinytgolin Mongolia Autonomous Prefecture. In other words, variation in parasitic ability occurred even within the same broomrape ecotype. This phenomenon was similar to that observed in Striga (Mbuvi et al., 2017). The authors found that a Striga ecotype from Alupe was significantly more virulent than Mbita but was less virulent than the ecotype from Kibos, and they believed that it was difficult to associate virulence with a specific eco-geographic region. Huang et al. (2012) suggested that the high genetic diversity of the Striga seed bank led to rapid evolution/build-up of virulent genotypes from a subset of the natural seed bank population. However, the results of this study clearly show that there are differences in parasitic ability among the same ecotype of broomrape populations. Therefore, spatial and temporal stability and durability of P. aegyptiaca resistance have to be further ascertained taking into account variability in environmental conditions and virulence between P. aegyptiaca ecotypes.

Previous studies have shown that low induction of germination plays a significant role in host resistance (Yoder and Scholes, 2010; Dor et al., 2011; Fernández-Aparicio et al., 2012; Trabelsi et al., 2016), but little is known about its involvement in melon response to infection by P. aegyptiaca. The results of our study indicated that the resistance in KR1326 was not related to low induction of P. aegyptiaca germination, since germination of P. aegyptiaca seeds was similarly stimulated by exudates of the susceptible cultivar K1076. In addition, 5-DS was detected in the roots of both cultivars, and the content was not significantly different; this result also further supports our conclusion that the resistance mechanism was not related to low induction of P. aegyptiaca germination. Our study did however detect differences between K1076 and KR1326 for germination induction of O. cumana, a broomrape species with a narrow host range that mainly parasitizes sunflowers and whose seeds germinate under low concentrations of sunflower-specific sesquiterpene lactones (Cala et al., 2017). The results showed that root exudates of K1076 induced higher levels of O. cumana germination in comparison with KR1326, but lower than positive control GR24. This result could explain the nonhost resistance of melon to O. cumana being due to low induction of germination of O. cumana.

Different numbers of attached parasites developed successfully (compatible) or underwent necrosis (incompatible) on K1076 and KR1326. The failure of parasites to develop after attachment was associated with the appearance of broomrape tubercles darkening at the site of attachment at 23 DAI. A similar resistant phenotype was observed in legumes infected by O. crenata (Pérez-de-Luque et al., 2005). In this case, the tubercle necrosis reaction was associated with the lignification of host endodermis and pericycle cells at the penetration site. Cissoko et al. (2011) also showed that Striga grew slowly and remained small after forming only a few connections to the host xylem in the Striga-rice (NERICA) resistance interaction. However, this type of resistance differs from that observed in cowpea-Striga (Li and Timko, 2009; Su et al., 2020), sunflower-O. cumana (Duriez et al., 2019), or tomato-Cuscuta (Jhu et al., 2022). Following parasite attachment, a proportion of the parasites died owing to the onset of a rapid hypersensitive reaction between parasite and host. In our study, no clear discoloration was observed on the KR1326 roots at the parasite attachment site at 7, 9, 16, and 23 d, indicating that HR was not involved in the resistance response.

In this study, transverse sections through incompatible parasite attachments showed that the host xylem cells in contact with the parasite intrusive cells presented a thickening of their walls that stained intensely red with safranin. In addition, the safranin staining substance was also observed in the intercellular spaces between host cells and intrusive parasite cells (black triangles). Pérez-de-Luque et al. (2005) suggest that this secretion is derived from the haustorial cells of broomrape. The accumulation of secretions at the infection site may lead to the activation of xylem occlusion that may cause further necrosis of established tubercles. The accumulation of defensive secondary metabolites is an effective plant defense and competition strategy for deterring Orobanchaceae parasites (Jhu and Sinha, 2022). This observation is reminiscent of a safranin-stainable substance accumulated in the contact zone between S. hermonthica and Lotus japonicus (Yoshida and Shirasu, 2009). Lignification of endodermal cell walls in contact with parasitic tissues was observed in resistant fava beans upon O. crenata attack (Pérez‐de‐Luque et al., 2007). This can be explained by delayed penetration and establishment as a result of the barriers activated by the resistant host; it would take more time for the parasite to establish a functional haustorium because it must overcome the resistance mechanisms.

Callose deposition was observed in the resistance interaction at the host–parasite interface and in the middle lamellae and cell walls of some xylem vessels. This suggested that callose may be implicated in cell wall reinforcement. Reinforcement of host cell walls by callose deposition, a β-1,3-glucan polymer, appears to be the main factor responsible for resistance in this case (Pérez-de-Luque et al., 2007). In both resistant fava bean and pea plants, callose accumulation to reinforce cell walls and hamper O. crenata penetration was also observed (Pérez-de-Luque et al., 2006; Pérez-de-Luque et al., 2007). Because the reinforcement of cell walls by callose requires cross-linking of hydroxyproline-rich glycoproteins (Brown et al., 1998), protein cross-linking may also take place. Therefore, in subsequent experiments, the presence of protein crosslinks in the cell wall should also be investigated.