Tolerance Screening for Phytophthora capsici and Fusarium solani f. sp. cucurbitae in Cucurbita spp. and Gene Expression Responses in Mutant Families

: Cucurbita species can be affected by soil-borne pathogens, such as Phytopthora capsici and Fusarium solani f. sp. cucurbitae ( Fsc ). Diverse commercial and conserved lines of Cucurbita spp. were tested. C. pepo subsp. pepo genotypes showed the highest susceptibility to both pathogens. The tolerance to P. capsici and Fsc was then screened in a zucchini mutant population. Two M3 mutant lines (Cp107 and Cp116) with a high occurrence of tolerant individuals to Phytophthora capsici were obtained from a screening of 160 M2 mutant lines. The M3 lines presented higher tolerance than the background MUCU-16. Furthermore, in the inoculated samples, both mutants overexpressed CpDEF and expressed more CpPAL and CpChiIV than the susceptible control. It has been previously shown that this expression pattern could be associated with tolerance in the P. capsici - Cucurbita spp. pathosystem. The M3 lines obtained could be applied in breeding programs, as they are likely to be compatible with the highly susceptible C. pepo subsp. pepo genotype.


Introduction
The Cucurbitaceae family includes a group of cultivable species within the Cucurbita genus, such as C. moschata, C. maxima, and C. pepo. There are several morphotypes for each species, with the zucchini morphotype (C. pepo subsp. pepo) being one of the most economically relevant. Cucurbita production can be affected by soil-borne pathogens, which can lead to plant death. Two important pathogens are the oomycete Phytophthora capsici [1] and the fungus Fusarium solani f. sp. cucurbitae (Fsc) [2]. Both Fsc and P. capsici are able to affect a wide range of cucurbit species and have analogous symptomatology, initially causing damage localized to the crown that leads to constriction and plant wilting due to the inability to transport water and nutrients [3][4][5]. These pathogens can remain in soil for long periods of time [1,5], and their management is currently through chemical treatment, crop rotation, or bio/solarization [5][6][7].
Screening of resistance to soil-borne pathogenic fungi and oomycetes (SBPFO) in cucurbits, mainly through existing germplasms and wild species, has been successfully used to inform breeding via rootstock candidates or introgression programs [3]. Grafting has proven to be very valuable in cucurbits, but plant resistance through resistance genes will always be the preferred objective, because it is an easy-handling solution, despite its complications [8]. Most studies on SBPFO in cucurbits have been focused on the Cucumis genus and the different formae speciales of Fusarium oxysporum, although there have been some focused on Cucurbita spp. with P. capsici [9][10][11][12][13][14][15][16][17] and Fsc [2,18]. Cucurbita spp. accessions have been identified with high frequency of individuals with resistance to Fsc [18]. Also, some C. lundelliana and C. okeechobeensis wild species [15], C. pepo [9], and C. moschata [11] genotypes showed reduced crown rot damage caused by P. capsici.
Mutant germplasms were obtained from a second-generation mutant (M2) population of MUCU-16 C. pepo subsp. pepo [23]. The mutations were randomly distributed [23] in the 160 mutant families tested. Currently, the M2 and M3 families are conserved in the IFAPA-La Mojonera germplasm collection.
Phytophthora capsici (MI0211, GenBank MG012233) [6] and Fsc race 1 (Fs511, GenBank AM940070) isolates [35], obtained from infected plants in greenhouses in Almería province (Spain) were morphologically and molecularly described in previous studies [6,35]. Both pathogens were preserved in the IFAPA-La Mojonera laboratories and subsequently inoculated on C. pepo, and re-isolated and identified again from dead plants prior to inoculation.

Inoculation
P. capsici and Fsc inocula followed the same inoculation procedures. Two inoculation methods were used: root dipping and substrate infestation ( Table 1). The first screening with different Cucurbita spp. genotypes and the initial M2 mutant germplasm were inocu-lated by root dipping (Table 1). In this method, the radicles were immersed for 30 s in a suspension containing the pathogen propagules. This suspension was obtained by grinding several colonies fully covering the petri dish surface with each isolate, which had been previously grown for 10 days in PDA (1 petri dish per 400 mL of standard nutrient solution). The inocula concentrations were calculated by means of the dilution plate technique on PDA, obtaining ca. 10 6 UFC mL −1 for Fsc and ca. 10 4 UFC mL −1 for P. capsici. 1 (32) The same procedure was used to obtain the inoculum for the subsequent mutant screenings and gene expression quantification analysis (test 3-5, Table 1). However, instead of radicle immersion, the inoculum was added by irrigation to the substrate at 2 cm depth near the plant with 50 mL of the inoculum suspension (substrate infestation method) once the plants reached the development stage of 2-3 true leaves ( Table 1). The presence of Fsc macroconidia and Phytophthora sporangia on the dead plants' roots and crown was checked under a microscope, and they were reisolated for the following inoculations after each test. Control plants were watered with standard nutrient solution.

Test Conditions
All experiments were performed in a growth chamber with a 14 h photoperiod (>12,000 lux), 23-33 • C, and 40-75% relative humidity. The temperature and humidity were measured using a HOBO data logger. Before planting, the seeds were disinfected by immersing them in a sodium hypochlorite (35 g L −1 active chloride) solution 1:1 for 20 min, and subsequently they were rinsed with autoclaved (30 min, 121 • C) sterile water and incubated at 28 • C in the dark until the roots reached 3-5 cm in length. Only germinated seeds were used.
In order to inoculate all plants at the same age, different sowing dates were applied depending on previous knowledge about the early development of each genotype. This step is important, because different plant development stages could affect their resistance responses [36]. The germination rate after one week for commercial and non-mutated germplasms was higher than 90%, while the M2 mutant germplasms had a germination rate between 25-95%, depending on the mutant family. Once the seeds had germinated, the seedlings reached 2-3 true leaves after 10-15 days. The substrate used for all of the experiments was vermiculite at field capacity, with a standard nutrient solution (1.5 dS m −1 ). For all tests, all samples were randomly distributed in separate trays in the chamber beside non-inoculated controls. Non-inoculated controls were included in all of the tests. For test 1, 8 control replicates per genotype were used. At least one plant per mutant family served as the control for tests 2, 3, and 4. Test 5 was completed with the same number (32) of control repetitions.
For the different Cucurbita spp. genotype tolerance screenings (test 1), two plants per 1 L pot (eight plants per pathogen, two trials) were used, and these plants were inoculated by the root-dipping method (Table 1). Susceptible and tolerant reference controls for tests 4 and 5 were chosen based on the responses of these plants (test 1).
The mutant screening and the selection of tolerant families consisted of three steps ( Table 1): initial screening with seedlings (test 2), second screening with developed plants (test 3), and final selection (test 4). Initial mutant screening (test 2) was performed with 160 mutant families (three replicates) in 60 mL 96-well trays ( Table 1). The mutant families that showed less symptomatology or that had tolerant individuals by day 14 postinoculation were selected for the next test. Subsequent screenings were undertaken in 1 L pots (two plants per pot) with a decreasing number of selected families but increasing numbers of replicates per family (Table 1). For the next test, 50 selected families for P. capsici and 26 for Fsc were inoculated when the plants reached 2-3 true leaves (test 3). For the final mutant selection test (test 4), six families were selected for each pathogen, and genotypes M64 and AMA acted as the tolerant and susceptible controls, respectively. The mutant individuals that survived the P. capsici substrate infestation were reinoculated, transplanted into a 25 L pot, and developed in a greenhouse. This allowed the plants to reach a large enough size to produce, after self-pollination, a mature zucchini fruit with grown M3 seeds. The two M3 families with fertile seeds obtained in the study were used in the expression analysis test. For the gene expression analysis (test 5), the M3 mutant lines Cp107 and Cp116, and genotypes M63 (tolerant control) and MUCU-16 (susceptible control) were used (Table 1).

Tolerance Evaluation
The tolerance evaluation was based on three parameters: the disease severity index (DSI), percentage of plants with symptomatology, and relativized area under the disease progress curve (RAUDPC).
The DSI is based on a visual non-parametric scale ranging from 0 to 4: 0 = no damage, 1 = leaf epinasty/soaked crown, 2 = constricted/girdling crown, 3 = wilt/chlorosis, and 4 = death; this is analogous to other authors [6,9,18]. The percentage of plants with symptomatology plotted against time produced an area under the disease progress curve (AUDPC) that represented the disease evolution. The AUDPC was relativized (RAUDPC) using the sample with the highest AUDPC value in each trial [2,6,14]. In test 1, plants that had an RAUDPC between 0-10% were considered tolerant, 11-30% low susceptibility, 31-70% intermediate susceptibility, and 71-100% high susceptibility. The RAUDPC was submitted to an analysis of variance (ANOVA); when significant differences (p < 0.05) were found, Tukey's HSD was carried out. For two-level comparison, Student's t-test was used. The data were tested for a normal distribution and homoscedasticity by the Shapiro-Wilk and Levene tests. Statistix 9.0 software (Tallahassee, FL, USA) was used for all statistical analyses. Tests 1 and 5 were duplicated (Table 1), and the data were pooled. Additionally, the damage level caused by both pathogens was analyzed once the plant died ( Figure A1).

qPCR Sample Collection and Conditions
The test conditions are shown in Table 1 (test 5). Root and crown samples were collected on 0, 3, 10, and 14 days post inoculation (dpi), as the disease evolution was observed. Each sample was extracted from eight different plants, which decreased the number of plants in the test by eight after each sample collection day. Sample collection, RNA extraction, and cDNA obtention were performed as described in previous work [34]. The primers (Sigma-Aldrich, St. Louis, MO, USA), except for the UFP and EFP-1α reference genes [37], were obtained from previous work [34]. The qPCR cycling conditions were: 50 • C 2 ; 95 • C 10 ; 40× (95 • C 15", 55 • C 15 , and 60 • C 1"); and melting (95 • C 15", Horticulturae 2022, 8,191 5 of 14 60 • C 1 , and 95 • C 1"). Gene expression was relativized with UFP (ubiquitin fusion protein) and EFP-1α (elongation factor-1α) reference genes [37] following data processing with efficiency correction, multiple reference gene normalization, and the use of error propagation rules [38,39]. The normalized relative quantities (NRQ) for each sample were obtained and transformed with LOG 10 (NRQ inoculated /NRQ non-inoculated ). Sample data were pooled (3, 10, and 14 dpi and crown and root tissues) by genotype and compared against the susceptible control by applying Student's t-test (p < 0.005). Homoscedasticity and normal distribution were checked with Statistix 9.0 software.

Results
3.1. Cucurbita Genotype Tolerance Screening to P. capsici and F. solani f. sp. cucurbitae The RAUDPC caused by P. capsici and Fsc for different Cucurbita spp. genotypes after 37 dpi is shown in Table 2. Both pathogens, inoculated by root-dipping, showed damage to all genotypes tested. However, the C. moschata species showed less damage for both Fsc and P. capsici (Table 2). According to the phenotype classification, M64 and M63 were considered tolerant to P. capsici, while TZ3 was tolerant to Fsc (Table 2). In contrast, many C. pepo plants showed high susceptibility to both pathogens, although the two marrow morphotypes analyzed (PIC and ZEB) presented intermediate susceptibility to Fsc (Table 2). Just one hybrid of C. maxima × C. moschata showed high susceptibility to both pathogens, HER, while SHI showed intermediate susceptibility ( Table 2). Significant differences in the RAUDPC were observed between the TZ3 tolerant and AZM, CAM, HER, MUC, MIL and NAT susceptible genotypes after Fsc inoculation (Table 2). For P. capsici, significant differences also appeared between M63, M64, TZ3, TER, and ZEB with the highly susceptible MIL, MUC, NAT, SCA, and VIC genotypes ( Table 2). The first mutant screening included 160 families (test 2, Figure 1). Both P. capsici and Fsc produced symptomatology (Figure 2) in the entire inoculated population after 27 dpi. In some cases, plants were unable to develop true leaves due to the disease. Families with less damage or with individuals without damage on day 14 post-inoculation were selected for the next test (Figure 1). For P. capsici, 50 families were selected, while for Fsc, 26 families were selected. Fewer families were selected for Fsc because this fungus produced more damage by day 14 than P. capsici. The selected families were inoculated in a new test (test 3). In this test, the six families with the lowest RAUDPC by day 27 post-inoculation were selected for the subsequent screening (test 4, Figure 1).  For P. capsici, all the previously selected families from test 3 had individuals that survived to the inoculation of test 4 (Table 3); these plants were re-inoculated at 34 dpi and transferred to a greenhouse (Figure 1). In total, 15 plants from families Cp45 (1), Cp107 (2), Cp116 (3), Cp136 (3), Cp139 (2), and Cp144 (4) remained at 97 dpi. Mature fruits were obtained from families Cp107, Cp116, Cp139, and Cp144, but only Cp107 and Cp116 fruits had M3 fertile seeds. On the other hand, none of the six families selected for Fsc showed tolerance to the pathogen ( Table 3). The tolerant (M64) and susceptible (AMA) controls responded (Table 3) as in the previous test for both pathogens (Table 2).  (1); Fusarium solani f. sp. cucurbitae (left) and Phytophthora capsici (right) inoculated trays (2); individuals showing tolerance (3). Screening with selected plants (4)(5)(6): plant symptomatology 21 dpi-surviving individuals were moved to a greenhouse after 34 dpi and reinoculated (5); surviving plant growth until 97 dpi for fruit production (6).   The undergrown plant lesions evaluated after plant death indicated that both pathogens produced damage to the crown and root tissues. All dead plants analyzed for P. capsici and for Fsc presented more accentuated symptomatology in the crown, with at least soaking damage, while the secondary roots were the least affected tissue ( Figure A1).

M3 Mutant Tolerance Analysis
The M3 Cp107 and Cp116 mutant families (derived from self-pollinated M2 plants that showed tolerance) presented less damage caused by P. capsici than the non-mutated MUCU-16 (M0) (Figure 3). The mutants showed significant differences in the RAUDPC against the susceptible control 14 dpi (Table 4). No damage was observed until day 10 after inoculation for Cp107, Cp116, and M63 while, for MUCU-16, the symptoms started on day 3 (Table 4).

M3 Gene Expression
The differential expression between inoculated and non-inoculated samples for all pooled samples (dpi: 3, 10, and 14) showed significant differences in MUCU-16 for the CpDEF, CpChiIV, and CpPAL genes. On the other hand, the CpACO and CpLPOX genes remained similar to the control (Figure 4). Mainly due to down-regulation of CpPAL in the susceptible after inoculation, the genotypes M63 (p = 0.021), Cp107 (p = 0.042) and Cp116 (p = 0.043) were expressed CpPAL more than MUCU-16 ( Figure 4). No significant differences were found between the C. pepo genotypes (MUCU-16, Cp107, and Cp116) in CpDEF expression in the inoculated samples, while the tolerant control showed higher up-regulation (p = 0.0117, Figure 4) with respect to the susceptible control. However, only the Cp107 (p = 0.046) and Cp116 (p = 0.035) mutant families expressed the CpChiIV gene more in the inoculated samples versus the control (Figure 4). CpDEF expression in the inoculated samples, while the tolerant control showed higher up-regulation (p = 0.0117, Figure 4) with respect to the susceptible control. However, only the Cp107 (p = 0.046) and Cp116 (p = 0.035) mutant families expressed the CpChiIV gene more in the inoculated samples versus the control (Figure 4).  Enzenbacher and Hausbeck [14] tested different isolates of P. capsici on C. pepo (zucchini and straightneck) and C. moschata (butternut) genotypes. Their results pointed out that the C. pepo straightneck morphotypes showed higher susceptibility and the C. moschata genotype showed higher tolerance [14]. Similarly, differences in the disease severity index were observed between the morphotypes of C. pepo and between the subspecies C. pepo subsp. pepo and C. pepo subsp. ovifera [4]. In this study, C. pepo subsp. pepo marrow morphotypes showed more tolerance to P. capsici, and the C. pepo subsp. ovifera were highly susceptible [4]. Our results indicated that the plant responses to these pathogens were also dependent on the species and morphotype. In our study, C. moschata was the most tolerant, while C. pepo subsp. ovifera was highly susceptible to P. capsici. The C. pepo zucchini morphotypes were the most susceptible, while the marrows were the least susceptible among the C. pepo subsp. pepo morphotypes. Furthermore, Perez-Hernandez et al. [2] tested some pathogenic isolates of Fsc that were shown to be highly aggressive to zucchini plants, while other cucurbit hosts showed differential susceptibility to the path-  Enzenbacher and Hausbeck [14] tested different isolates of P. capsici on C. pepo (zucchini and straightneck) and C. moschata (butternut) genotypes. Their results pointed out that the C. pepo straightneck morphotypes showed higher susceptibility and the C. moschata genotype showed higher tolerance [14]. Similarly, differences in the disease severity index were observed between the morphotypes of C. pepo and between the subspecies C. pepo subsp. pepo and C. pepo subsp. ovifera [4]. In this study, C. pepo subsp. pepo marrow morphotypes showed more tolerance to P. capsici, and the C. pepo subsp. ovifera were highly susceptible [4]. Our results indicated that the plant responses to these pathogens were also dependent on the species and morphotype. In our study, C. moschata was the most tolerant, while C. pepo subsp. ovifera was highly susceptible to P. capsici. The C. pepo zucchini morphotypes were the most susceptible, while the marrows were the least susceptible among the C. pepo subsp. pepo morphotypes. Furthermore, Perez-Hernandez et al. [2] tested some pathogenic isolates of Fsc that were shown to be highly aggressive to zucchini plants, while other cucurbit hosts showed differential susceptibility to the pathogen. These results support ours, since, in this study, zucchini genotypes were found to be highly susceptible to Fsc, while some genotypes of C. moschata showed tolerance to the fungus.
It has been observed in different screenings for resistance to P. capsici that some genotypes of C. lundelliana and C. okeechobeensis wild species [15], C. pepo [9], and C. moschata [11] show reduced crown damage. Padley et al. [9] obtained, from a resistance screening in C. pepo, 16 accessions that were less damaged by P. capsici, and selected two of them because they were more resistant [9]. All these accessions showed high genetic proximity, but one line was also close to the susceptible C. pepo subsp. pepo and C. pepo subsp. texana genotypes [16]. The backcrossing and its phenotypic segregation analysis indicated that the resistance to P. capsici was controlled by three dominant genes (R4, R5, and R6) [17]. Furthermore, resistance to P. capsici has been introduced into C. moschata from wild C. lundelliana and C. okeechobeensis, and its segregation also suggests that this trait is controlled by three dominant genes (Crr-1, Crr-2, Crr-3) [10].
In addition, the genotypes that present more tolerance could be used for grafting other compatible cucurbits. Cucurbita spp. germplasms that present high frequency of individuals with tolerance to Fsc were identified as rootstock candidates [18]. Additionally, grafted Citrullus lanatus on Ferro, GV 100 (C. maxima × C. moschata), and Just (Cucurbita spp.) rootstocks showed tolerance to Fsc, although this response depended on the plant development at the inoculation moment [41]. For watermelon, the Lagenaria siceraria (bottle gourd) rootstocks FR-Strong, Emphasis, Macis, and WMXP-3938 were resistant to crown and root rot caused by P. capsici [42]. Our results suggest that there are rootstock candidates for tolerance to P. capsici (M63 and M64) and Fsc (TZ3), but compatibility studies with zucchini are needed.

Mutant Selection
To our knowledge, none of the previous SBPFO resistance screening studies in cucurbits were performed with artificially induced mutants. All were based on germplasm bank collections of wild, conserved, commercial, or hybrid genotypes [2,9,14,15,18]. This study presents the first results for mutant C. pepo subsp. pepo lines with tolerance to P. capsici.
Few studies have been conducted in the Cucurbitaceae family on biotic stress with EMS-mutated populations, despite this being an excellent tool for breeding [3,27]. For example, one bottle gourd (Lagenaria siceraria) M4 EMS-induced inbred line (117-8) resistant to zucchini yellow mosaic virus (ZYMV) was obtained and showed stable resistance and no virus presence among inoculated plants [27]. Our selected families (Cp107 and Cp116) presented more tolerance to P. capsici than the highly susceptible C. pepo subsp. pepo from which they originated (Table 4). However, some characteristics, such as plant architecture, seemed to be affected with respect to the MUCU-16 background M0 (Figure 3), although specific phenotyping analyses still have to be performed.
Genes for resistance to SBPFO have been reported in cucurbits [3,43]. However, the mutant tolerance to P. capsici could be due to diverse causes, given the nature (random and numerous) of mutations caused by EMS. Plant development, which affects plant tolerance [36], plant resistance genes that respond to pathogen virulence factors [3,30,31], or plant susceptibility genes, which are targeted by the pathogen [33,[44][45][46][47], may have been affected by the mutations.
During soilborne disease-resistance screening with a large population, it is common to find individuals that show more tolerance, although the family as a whole may seem susceptible [18]. We also observed this situation, especially in the mutant population. In an M1 population, most of the randomly mutated genes would be in heterozygosis, and its segregation (M2) could show different phenotypes [23]. The individuals tolerant to P. capsici were self-fertilized and M3 seeds were obtained. This reduced the family phenotypic heterogeneity, although segregation studies should be performed to confirm this fact. It should be taken into account that, despite obtaining a mutant of interest, sometimes it will not be pos-sible to produce seeds due to the difficulty some mutants have in generating fertile seeds, as random mutations can alter their formation [23]. Finding a mutant material with tolerance to a disease in the species to be cultivated would reduce the steps in introgression programs or grafting. Nevertheless, the method for producing a mutant population for each crop must be previously known to avoid the loss of desirable characteristics of the background M0, while still allowing it to acquire new ones. Vicente-Dólera et al. [23] observed that, with the mutagenic agent EMS at a concentration of 0.4%, plants had more viability than at higher concentrations. Also, with this concentration, the plant population incorporated a mutation frequency of 1/135 kb [23]. For this reason, we chose this concentration so that the plants would not lose the useful characteristics acquired during domestication.

qPCR
The CpDEF gene was highly overexpressed in the inoculated plants, and CpPAL downregulation occurred only in the susceptible genotypes, which is in agreement with previous studies [34]. Moreover, The mutants Cp107 and Cp116 expressed more CpChiIV than MUCU-16 after inoculation. The mutant population may not have directly affected these genes, since this differential expression between the susceptible and tolerant phenotypes observed may be the consequence, rather than the cause, of tolerance.

Conclusions
The screening of an EMS-mutated population was proven to be a reliable method for obtaining new traits in complex phenomena, such as C. pepo subsp. pepo tolerance to P. capsici. The tolerance trait was found in the phenotype and gene expression analyses of the two M3 families obtained. These families presented significant differences in P. capsici tolerance, and in the expression of CpPAL and CpChiIV genes compared with the susceptible background. Figure A1. Underground symptomatology of P. capsici and Fsc. P. capsici (left) and Fsc (right) 14 dpi in seedlings (1). P. capsici (left) and Fsc (right) 21 dpi in plants inoculated when they reached 2-3 true leaves (2,3). Soaked and constricted crown and main root disaggregation caused by Fsc (4), main root browning caused by P. capsici (5) 14 dpi. Note that these cucurbits had a fibrous root system (generate more than one main root), but, in the young plants, a main root is clearly observable.