Search for Resistant Genotypes to Cuscuta campestris Infection in Two Legume Species, Vicia sativa and Vicia ervilia

The dodders (Cuscuta spp.) are parasitic plants that feed on the stems of their host plants. Cuscuta campestris is one of the most damaging parasitic plants for the worldwide agricultural production of broad-leaved crops. Its control is limited or non-existent, therefore resistance breeding is the best alternative both economically and environmentally. Common vetch (Vicia sativa) and bitter vetch (Vicia ervilia) are highly susceptible to C. campestris, but no resistant genotypes have been identified. Thus, the aim of this study was to identify in V. sativa and V. ervilia germplasm collections genotypes resistant to C. campestris infection for use in combating this parasitic plant. Three greenhouse screening were conducted to: (1) identify resistant responses in a collection of 154 accessions of bitter vetch and a collection of 135 accessions of common vetch genotypes against infection of C. campestris; (2) confirm the resistant response identified in common vetch accessions; and (3) characterize the effect of C. campestris infection on biomass of V. sativa resistant and susceptible accessions. Most common vetch and bitter vetch genotypes tested were susceptible to C. campestris. However, the V. sativa genotype Vs.1 exhibited high resistance. The resistant phenotype was characterized by a delay in the development of C. campestris posthaustorial growth and a darkening resembling a hypersensitive-like response at the penetration site. The resistant mechanism was effective in limiting the growth of C. campestris as the ratio of parasite/host shoot dry biomass was more significantly reduced than the rest of the accessions. To the best or our knowledge, this is the first identification of Cuscuta resistance in V. sativa genotypes.


Introduction
The common vetch (Vicia sativa L.) is an annual legume native to the Mediterranean Basin. It is widely cultivated in many areas worldwide due to its high nutritional value as a grain legume or forage crop [1][2][3] and its ability to grow over a wide range of climatic and soil conditions [4]. Worldwide cultivation of common vetch area reached nearly 540,762 ha in 2018 with 34% of this area being cultivated in the Mediterranean Basin [5]. Spain is the main producing country of common vetch for which the growing area was 103,100 and 143,200 hectares for grain and forage production, respectively, in 2018 [6]. In traditional rain-fed areas in the Mediterranean Basin, common vetch is cultivated either as a monocrop or intercropped with cereals for improved forage harvesting and yield [7,8]. Bitter vetch (Vicia ervilia (L.) Willd.) is one of the oldest cultivated grain legume crops, originating in the Mediterranean and Middle East area [9][10][11]. Spanish cultivation of bitter vetch reached 54,900 ha during 2018 [12]. It is an annual, predominantly self-pollinated species, tolerant to marginal soils, and drought and cold climate conditions [13]. The inclusion of V. sativa and V. ervilia in crop rotations contributes to the increase in sustainability in agricultural systems, reducing the need for fertilizers and pesticides by improving soil fertility and reducing the incidence of pests and weeds [11,14]. The multiple benefits of V. sativa and V. ervilia are threatened by their high susceptibility to infection by parasitic weeds [15][16][17][18]. Despite the many advantages of the cultivation of these Vicia species in low input cropping systems, its cultivation is in decline, mainly due to the lack of investment in breeding programs to register elite cultivars [19].
The dodders (Cuscuta spp.) are stem parasitic plants from the Convolvulaceae family with none or reduced photosynthetic activity and as a consequence, they are obliged to obtain nutrients by parasitizing the stem of other plants [20]. The genus Cuscuta contains over 170 species distributed throughout the world [21]. Among them, C. campestris is one of the most damaging species worldwide for the agricultural production of dicotyledonous crops [22]. After germination, Cuscuta seedlings coil around host stems and differentiate prehaustoria that penetrate the host and connect to its vascular system for nutrient diversion [23]. The control of Cuscuta is difficult because of their persistent seedbanks formed by long living seeds with hard coats as well as their capacity to infect a broad host range including other weeds and the intimacy of haustorial connections with the host, which makes the application of control methods selective enough to kill the parasite without affecting the crop difficult [24][25][26]. For most crop-Cuscuta species pairs, control is limited or non-existent and for those crops where Cuscuta control is possible, the control strategies are mainly based in either phloem-mobile herbicides applied to herbicide-resistant crops [27][28][29] or infection-resistant crops [22]. Legume resistance to Cuscuta has only been reported in chickpea and the resistance mechanism is characterized by the failure of the prehaustorium to penetrate the host stem [22]. Resistance in V. sativa and V. ervilia has not been reported against the infection of stem parasitic weeds, but it has been frequently reported against the infection of root parasitic plants [15][16][17][18][30][31][32][33]. Resistant phenotype in vetch to Phelipanche aegyptiaca has been described with the appearance of a dark substance between host and parasite cells, and is associated with an increase in peroxidase activity and the increase in the concentrations of phenolics and lignin [30,31]. Resistant V. sativa to O. crenata is associated with lignification of the endodermal cells [32] and mucilage production inside vetch vessels, leading to the obstruction of the parasite nutritive supply [15]. Resistant V. ervilia to Orobanche crenata is associated with reduced induction of O. crenata germination and failure to the O. crenata haustorium to penetrate the host root [17,18].
V. sativa and V. ervilia are highly susceptible to C. campestris, but to the best of our knowledge, no resistant genotypes have been identified thus far in these crop species. Thus, the objective of this research was to identify resistant genotypes against C. campestris in two Vicia germplasm collections: a collection of 135 genotypes of V. sativa and a collection of 154 accessions of V. ervilia.

Search for Resistant Genotypes in Vicia ervilia
In order to identify V. ervilia genotypes with the capacity to inhibit C. campestris infection through resistance against haustorium formation and penetration, a collection of 154 V. ervilia accessions was studied in a greenhouse. Each V. ervilia accession was planted in triplicated pots and grown in a greenhouse for sixteen days before two day-old C. campestris seedings were placed around each V. ervilia plant. Table 1 shows that C. campestris seedlings were able to coil around the stem of all plants tested from each of the 154 V. ervilia accessions, revealing the absence of allelopathic mechanisms that could cause C. campestris repellency against the V. ervilia plants. Once the C. campestris stem is coiled around the host stem, prehaustoria forms and penetrates the stem to form vascular connections. From the nutrients extracted from the host through the vascular connections, C. campestris develops filamentous stems at the Cuscuta-host connection sites, called in this work posthaustorial growth ( Figure 1). Table 1 shows the data recorded at seven days after inoculation in the 24 V. ervilia most resistant accessions in which at least 50% of the plants resisted the penetration of the haustorium and inhibited the posthaustorial growth of nine day-old C. campestris seedlings, with only one accession, the accession Ve.136, able to completely resist in all their plants the formation and penetration of the haustorium. The remaining 130 V. ervilia accessions allowed in more than 50% of their plants the development of nine day-old C. campestris posthaustorial growth. These differing responses among V. ervilia accessions observed against nine day-old C. campestris were not maintained in the successive days displaying all the V. ervilia accessions susceptible responses 14 days later.

Search for Resistant Genotypes in Vicia sativa
In contrast to the results observed in the screening of V. ervilia accessions, the screening of 135 accessions of V. sativa revealed that resistance is very scarce, but exists in this legume species (Table 2). Each V. sativa accession was planted and inoculated as described for V. ervilia (see Materials and Methods section). As occurred in V. ervilia plants, C. campestris seedlings were able to coil quickly around the stem of all V. sativa accessions, indicating the absence of a mechanism of repellency in V. sativa plants against C. campestris. Table 2 also shows data on the posthaustorial growth of nine day-old C. campestris recorded in the 23 most resistant V. sativa accessions in which at least 50% of the plants had resisted the penetration of the haustorium and inhibited the posthaustorial growth. Accession Vs.1 and Vs.4 were able to resist in all their plants tested the infection of nine dayold C. campestris. The remaining 112 V. sativa accessions allowed in more than 50% of their plants the development of posthaustorial growth in nine day-old C. campestris seedlings. Fourteen days later, C. campestris seedlings were able to penetrate all the accessions including those accessions that resisted the infection, however, contrary to what happened

Search for Resistant Genotypes in Vicia sativa
In contrast to the results observed in the screening of V. ervilia accessions, the screening of 135 accessions of V. sativa revealed that resistance is very scarce, but exists in this legume species (Table 2). Each V. sativa accession was planted and inoculated as described for V. ervilia (see Materials and Methods section). As occurred in V. ervilia plants, C. campestris seedlings were able to coil quickly around the stem of all V. sativa accessions, indicating the absence of a mechanism of repellency in V. sativa plants against C. campestris. Table  2 also shows data on the posthaustorial growth of nine day-old C. campestris recorded in the 23 most resistant V. sativa accessions in which at least 50% of the plants had resisted the penetration of the haustorium and inhibited the posthaustorial growth. Accession Vs.1 and Vs.4 were able to resist in all their plants tested the infection of nine day-old C. campestris. The remaining 112 V. sativa accessions allowed in more than 50% of their plants the development of posthaustorial growth in nine day-old C. campestris seedlings. Fourteen days later, C. campestris seedlings were able to penetrate all the accessions including those accessions that resisted the infection, however, contrary to what happened in the V. ervilia collection, a darkening at the penetration site resembling a hypersensitive-like response ( Table 3 shows the results of a second greenhouse screening performed on accessions Vs.1, Vs.4, Vs.6, Vs.9, Vs.68, and Vs.84 to confirm the resistant response identified by the first greenhouse screening. In addition, we included the accessions Vs.11, Vs.51, and Vs.80 that also showed hypersensitive-like response but it was observed that plant segregation being the resistant phenotype was only visible in some of their plants, possibly due to lack of homogeneity of these V. sativa accessions stored in the germplasm collections of the Germplasm banks. Two susceptible accessions, Vs.8 and Vs.121, without induction of the hypersensitive-like response were also included. In this screening, we confirmed the delay of C. campestris early development and the induction of hypersensitive-like response in later stages of C. campestris in all plants of accessions Vs.1, Vs.4, and Vs.6 while these resistant responses were not observed in the susceptible accession Vs.8. The susceptible accession Vs.121 was very sensitive to Cuscuta infection and died before the evaluation date. Accessions Vs.9, Vs.11, Vs.51, Vs.68, Vs.80, and Vs.84 showed segregation being the resistant response only present in some of their plants, indicating that these accessions were not genetically homogeneous for this resistant character. Resistant plants were selected and multiplied in the absence of pollinators to initiate a breeding program for resistance against Cuscuta infection. Table 3. Compatibility of Cuscuta campestris seedlings with a selected collection of resistant and susceptible Vicia sativa accessions. Data expressed as the ( †) percentage of V. sativa plants with Cuscuta seedlings coiled around their stems; ( † †) percentage of V. sativa plants with Cuscuta adhesive disks and posthaustorial growth emerging from the penetration site; ( † † †) percentage of Vicia sativa plants showing hypersensitive-like response at the penetration site. Analysis of variance was applied to replicate data. Differences among genotypes with the susceptible genotype Vs.8 were assessed by Dunnett's test. *, **, and *** indicate significant differences at p < 0.05, 0.01, and 0.001, respectively; ns indicates no significant difference when comparing each common vetch accession with the Vs.8.

Effects of Resistant Response in Vicia sativa and Cuscuta campestris Trophic Relations
A third experiment was carried out in V. sativa to characterize the effect that the resistance response has in the trophic relations during infection between V. sativa accessions and Cuscuta (Table 4). The weight of dry biomass of host root, host aboveground tissues, and Cuscuta tissues were estimated in resistant accessions Vs.1, Vs.4, Vs.68, and Vs. 84 and compared with the dry weight of corresponding compartments in susceptible accessions Vs.8 and Vs.121. In addition, root and aboveground biomass of each accession was determined in uninfected control plants. The extreme susceptibility of accession Vs.121 to Cuscuta infection caused the death of the infected Vs.121 plants before the harvesting date. Cuscuta infection greatly reduced the total biomass of susceptible plants. The total dry weight of V. sativa accessions Vs.8 and Vs.121 plants infected with Cuscuta was respectively reduced by 84.3% and 79.3% in comparison with their corresponding uninfected control plants. The biomass reduction of infected plants was visible both in aerial biomass (86.4% and 82.0%) and root biomass (81.3% and 75.3%), respectively, for the Vs.8 and Vs.121 accessions (Figure 3). The ratio of parasite/host shoot dry biomass was 2.89 and 2.22, respectively, for Vs.8 and Vs.121 accessions. The biomass gain of Cuscuta did not account for the difference in biomass between infected and uninfected V. sativa plants. Combining V. sativa and Cuscuta total biomass revealed that the combined biomass of the infected system in accessions Vs.8 and Vs.121 was respectively 56.9% and 55.5% lower than that of the Vs.8 and Vs.121 uninfected plants. Besides the reduction in host biomass, Cuscuta also modified host allometric relationships among above and belowground organs. In uninfected V. sativa susceptible accessions Vs.8 and Vs.121, the percentage of dry weight allocated in aboveground organs with respect to total V. sativa dry weight of the entire plant was 61.0% and 62.1%, respectively. Cuscuta infection reduced this percentage to 20.7% in accession Vs.8 and to 24.3% in accession Vs.121. When adding the parasite sink as an aerial organ of the infected system, the ratio of combined host and parasite aboveground dry biomass to total combined biomass increased up to 79.1% for Vs.8 and 77.2% for Vs.121 infected plants.

Discussion
From the screening of 135 accessions of V. sativa and 154 accessions of V. ervilia have identified in this work a resistant phenotype against C. campestris infection in V tiva accessions. The resistant phenotype was observed after Cuscuta germination, co and prehaustorium development on the host, with a darkening of the host-parasite  Table 4. Varietal differences in the severity of Cuscuta infection were determined with (i) Cuscuta dry matter, and the (ii) ratio of parasite/host shoot dry biomass. Varietal differences in the effect of Cuscuta infection in allometric relationships were determined by calculating the (i) host aboveground biomass index (ratio of aboveground host dry matter/total host dry matter), and (ii) the combined aboveground biomass index (ratio of aboveground host and Cuscuta dry matter/total combined host-parasite dry matter). Varietal differences in the Cuscuta-induced changes in productivity were studied analyzing four parameters: (i) reduction in total host biomass (ratio of total host biomass of infected plants/total host biomass of uninfected plants); (ii) reduction in aboveground host biomass (ratio of aboveground host dry matter of infected plants/aboveground host dry matter of uninfected plants); (iii) reduction in host root biomass (ratio of host root dry matter of infected plants/host root dry matter of uninfected plants); (iv) reduction in combined biomass (ratio of total combined host-parasite dry matter in infected plants/total host biomass of uninfected plants). Analysis of variance was applied to replicate data. Differences among genotypes with the susceptible genotype Vs.8 were assessed by Dunnett's test. *, **, and *** indicate significant differences at p < 0.05, 0.01, and 0.001, respectively; ns indicates no significant difference when comparing each common vetch accession with the Vs.8. Least significant difference (LSD) value (p< 0.05) is provided for comparison among accessions. The Cuscuta biomass gain and the effect of Cuscuta infection on the reduction of V. sativa aboveground and root dry weight, biomass loss of infected system, and changes in allometric relationships were not significantly different between susceptible accessions Vs.8 and Vs.121 and the accessions responding with hypersensitive-like reactions Vs.4, Vs.68, and Vs.84. In contrast, the resistant response observed in accession Vs.1 was effective in limiting the growth of C. campestris, being the ratio of parasite/host shoot dry biomass, to only 0.59, significantly lower than the ratio of parasite/host shoot dry biomass in the rest of the V. sativa accessions. The reduction of parasite growth by the resistant mechanism in Vs.1 had positive effects on the host biomass, and the reduction of Vs.1 biomass by infection was significantly lower than the biomass reduction observed in the rest of the V. sativa accessions. Limited biomass loss due to infection in Vs.1 was observed both in the host aboveground and root biomass and in the combined biomass of the infected system in comparison with the susceptible accessions (Table 4).

Discussion
From the screening of 135 accessions of V. sativa and 154 accessions of V. ervilia, we have identified in this work a resistant phenotype against C. campestris infection in V. sativa accessions. The resistant phenotype was observed after Cuscuta germination, coiling and prehaustorium development on the host, with a darkening of the host-parasite contact site observed as a hypersensitive-like response. Hypersensitive-like response at the Viciaparasite contact has been previously observed against the root parasitic weed [30,31], but never against Cuscuta infection. The observed phenotype was exhibited by all plants in Vs.1, Vs.4, and Vs.6 accessions, while a degree in plant segregation in their hypersensitive-like response was observed in Vs.9, Vs.11, Vs.51, Vs.68, Vs.80, and Vs.84, being visible only in some of their plants possibly due to lack of homogeneity of these V. sativa accessions stored in the germplasm collections of the Germplasm banks. No resistant accessions were found in V. ervilia despite the large number of genotypes from diverse worldwide origins in the collection used for the screening. Resistance to C. campestris infection is very rare in cultivated species and to date, no varietal differences have been described in the responses to C. campestris infection in the majority of susceptible crops with few exceptions like in the greenhouse study of a collection of chickpea genotypes, which identified two resistant genotypes 'ICCV95333' and 'Hazera4' [22]. The resistant mechanism identified in chickpea by Goldwasser et al. [22] was different from that identified in our work in Vicia sativa genotypes and described as a seedling repellency-based mechanism after prehaustoria differentiation.
Both susceptible and resistant V. sativa plants infected with Cuscuta accumulated less dry matter than uninfected control plants, however, the loss of host biomass was significantly lower in resistant genotypes. The biomass accumulated by the parasite and the ratio of parasite/host aboveground dry biomass was much lower in resistant than in susceptible V. sativa plants. These observations agree with those reported in C. campestris-chickpea interactions [22], but they contrast to those reported in C. reflexa-host interactions [34,35]. The difference in host biomass between infected and uninfected plants did not equal that accumulated by the parasite, which agreed with Striga-host interactions [36,37], but disagreed with Orobanche-host interactions [38,39]. Besides the reduction in host biomass, Cuscuta also induced changes in the partitioning of dry weight between aboveground and belowground organs within the host, reducing the percentage of total V. sativa biomass allocated in aboveground organs with respect to total V. sativa dry weight. These findings are in contrast to those observed in Striga, which do not tend to cause large changes in the proportion of dry matter partitioned between photosynthetic and non-photosynthetic organs [40,41] and to those observed in Orobanche, which increases the ratio of photosynthetic to non-photosynthetic organs [38]. Resistance to Cuscuta parasitism in accession Vs.1 was absolute, Cuscuta development was strongly inhibited, and host biomass reduction was strongly inhibited.

Plant Material
Cuscuta seeds were collected in June 2018 from mature Cuscuta campestris plants parasitizing chickpea in a field at the IFAPA Center Alameda del Obispo of Córdoba, Spain. Cuscuta seeds were separated from dry capsules using a combination of winnowing with a fan and sifting with a 0.6 mm mesh-size sieve (Filtra, Barcelona, Spain). Cuscuta seeds were stored dry in the dark at room temperature until use for this work in the spring of 2020.
Screening for Cuscuta resistant genotypes was performed in germplasm collections of two legume species, bitter vetch (Vicia ervilia (L.) Willd.) and common vetch (Vicia sativa L.

Greenhouse Screening of Vicia ervilia and Vicia sativa Germplasm Collection
In a greenhouse at the IFAPA Research Center (Centro Alameda del Obispo, Córdoba Spain), 867 pots of 10.3 cm each side and 13.2 cm high containing 1 L of 1/1 sand and peat proportion were prepared for the screening of 154 accessions of bitter vetch and 135 accessions of common vetch. Three seeds of each accession of each vetch species were sown in triplicate pots in a complete randomized design. Vetch plants were grown for 16 days (10 • C-27 • C min and max temperature) before Cuscuta inoculation. They were irrigated with tap water every two days.
To promote Cuscuta germination, the hard coat of Cuscuta seeds was eliminated by scarification with sulfuric acid during 45 min, followed by throughout rinsing with sterile distilled water and air-dried. Scarified Cuscuta seeds were spread in wet filter paper inside Petri dishes and allowed to germinate in the dark during two days at 23 • C. Then, four germinated Cuscuta seedlings were manually placed using tweezers on the soil surface surrounding each of the three vetch plants per pot at 2 cm distance of the vetch stem. Seven days later, the Cuscuta seedlings were visually inspected and classified as either (i) unattached Cuscuta seedling, (ii) attached Cuscuta seedling without adhesion disks, (iii) attached Cuscuta seedling with adhesion disks without posthaustorial growth, and (iv) posthaustorial growth from the adhesion disks. At the end of the experiment, mature haustoria were inspected for hypersensitive-like response at the vetch-Cuscuta interface.

Confirmation of Resistant Phenotypes
A resistant phenotype was identified in some V. sativa accessions during the first screening. The resistant phenotype was characterized as a delay in the development of posthaustorial growth in nine day-old C. campestris seedlings ( Figure 1) and a subsequent darkening resembling a hypersensitive-like response at the penetration site of 23 day-old C. campestris (Figure 2). A second greenhouse screening was performed to confirm the resistant phenotypes in common vetch accessions, Vs.1, Vs.4, Vs.6, Vs.7, Vs.9, Vs.11, Vs.19, Vs.51, Vs.68, Vs.80, and Vs.84, identified by the first greenhouse screening. In addition, two highly susceptible common vetch accessions, Vs.8 and Vs.121, were included as susceptible controls. The experimental design, common vetch and Cuscuta cultivation, and resistance phenotyping were performed as described above.

Effects of Cuscuta Infection in Resistant and Susceptible Common Vetch Accessions
A third experiment was carried out to characterize the effect of Cuscuta infection on the biomass of common vetch resistant and susceptible accessions. In the greenhouse, 36 pots of 18 cm each side and 25.5 cm high containing 5.5 L of 1/1 sand and peat proportion were prepared to sow four resistant accessions (Vs.1, Vs.4, Vs. 68, and Vs.84) and two susceptible controls (Vs.8 and Vs.121). Cuscuta seeds were scarified, germinated, and manually inoculated on three pots per vetch accession as described before. Each accession was cultivated in three pots without Cuscuta as uninfected controls. At the end of the cultivation cycle, the consequences in common vetch productivity of Cuscuta parasitism was estimated by recording separately in each pot the Cuscuta and host biomass [39]. For each common vetch accession, Cuscuta dry weight, host aboveground, and host root tissue were collected separately and carried to the laboratory. Samples were dried at 70 • C for 48 hours and each biomass compartment weighed independently to determine five parameters: (i) host aboveground dry weight; (ii) host root dry weight; (iii) Cuscuta dry weight; (iv) total host biomass (aboveground and root host dry weight); and (v) combined biomass (total host and Cuscuta dry weight). Using these parameters, varietal differences in the severity of Cuscuta infection were determined with (i) Cuscuta dry weight, and the (ii) ratio of parasite/host shoot dry weight. Varietal differences in the effect of Cuscuta infection in allometric relationships were determined by calculating the (i) host aboveground biomass index (ratio of aboveground host dry matter/total host dry matter), and (ii) the combined aboveground biomass index (ratio of aboveground host and Cuscuta dry matter/combined dry matter). Varietal differences in the Cuscuta-induced changes in productivity were studied analyzing four parameters: (i) reduction in total host biomass (ratio of total host biomass of infected plants/total host biomass of uninfected plants); (ii) reduction in aboveground host biomass (ratio of aboveground host dry matter of infected plants/aboveground host dry matter of uninfected plants); (iii) reduction in host root biomass (ratio of host root dry matter of infected plants/host root dry matter of uninfected plants); and (iv) reduction in combined biomass (ratio of total combined host-parasite biomass in infected plants/total host biomass of uninfected plants).

Statistical Analysis
The experimental design was randomized complete blocks. Percentage data were transformed with arcsin ( √ (x/100) before analysis. Analysis of variance (one-way ANOVA) was applied to replicate data, with accession as the main factor using Statistix 9.1 software (Analytical software, Tallahassee, FL, USA). The significance of mean differences between each genotype against the control was evaluated by the two-sided Dunnett test. The significance of mean differences among genotypes was evaluated by the least significant difference (LSD) (p < 0.05).

Conclusions
The majority of V. sativa and V. ervilia accessions studied were susceptible to Cuscuta infection. We identified the V. sativa accession Vs.1 with high resistance to infection. The resistant phenotype is characterized by a hypersensitive-like response, resulting in inhibition of Cuscuta growth and reduction of biomass loss of infected Vs.1 plants. To the best of our knowledge, this is the first identification of Cuscuta resistance in V. sativa genotypes. Further histological and biochemical studies will continue this research to elucidate the exact mechanism involved.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.