In Mexico, the production of chili peppers is affected by pests and diseases. In recent years, the whitefly has become the main transmitter of diseases induced by the genus Begomovirus and has caused production losses for at least 95% (^{García et al., 2010}). In order to control this pest, different insecticides have been used; however, their use leads to contamination and to resistance in even more aggressive resistance of insect populations (^{George et al., 2015}). On the other hand, in plantations of habanero peppers (Capsicum chinense) with denomination of origin of the Yucatan peninsula, there is an interest in implementing technologies based on plant resistance to reduce losses in production caused by virosis from this interaction. A strategy for the management of virosis and to improve the production in this crop is the use of grafted plants, which uses the virtues conferred by the rootstock and those of the graft of commercial interest. Previous investigations in watermelon and tomato plantations showed the efficiency of the use of rootstocks, in which tolerance was conferred to a virus complex such as: Cucumber mosaic virus (CMV), Watermelon mosaic virus II (WMV-II), Zucchini yellow mosaic virus (ZYMV) and Tomato yellow leaf curl virus (TYLCV) (^{Wang et al., 2002}). In this regard, the option of using tolerant wild rootstocks is presented as a solution to the problem of virosis in the field. The aim of this investigation was to estimate, under field conditions, the tolerance of the graft of Capsicum chinense-Capsicum annuum var. glabriusculum to begomovirus.

Area of study and genetic material. The study was carried out using a cooperating farmer in the state of Yucatan, Mexico, in a field located between 20° 58´ 40” N and 89° 26´ 30” W at an altitude of 8 m above sea level. Two varieties were used as grafts: habanero criollo and habanero jaguar (C. chinense), and as rootstock, amashito and muela (C. annuum var. glabriusculum) gathered in backyard orchards and the adjacent wilderness.

Obtaining grafts. In order to favor the germination of rootstock seeds, they were submerged in gibberellic acid at 500 ppm for 24 h. Later, they were planted in polystyrene trays with 200 cavities in Sunshine #3^® + Agrolita^® in a 2:1 ratio as a substrate. Eight days later, the commercial varieties of habanero peppers, used as grafts, were planted. The plants were grafted using the terminal plectrum technique when the stems of the rootstocks and the grafts were 2 mm in length. Grafted plants were kept for 21 days in a humid adaptation chamber at a humidity of 80% and 27 °C. After this time, they remained for 14 days in a tunnel-style greenhouse with 450 lux for their adaptation and growth.

Establishment of the experiment. The plantation was established under a system of drip irrigation and with a regional production, a distance of 1.2 m between furrows, and 0.40 m between plants. The soil texture, according to analyses, was clay-loam. The fertilization program consisted of N-N0_3, 9.29, N-NH_4, 1.33, P-PO_4, 1.42, K, 4.85, Ca, 4.32, Mg, 1.68, S-SO4, 2.56, Na⁺, 0.19, Cl^-, 0.13 and HCO3^-, 0.01 meq L^-1, respectively; with modifications according to crop requirements. A complete random block design was implemented; six treatments were established with four repetitions: amashito + habanero criollo, amashito + habanero jaguar, muela + habanero criollo, muela + habanero jaguar, and as controls habanero criollo and habanero jaguar. Each repetition consisted of 30 plants. The production cycle was spring-summer (February 12 to August 25, 2018) and without the use of pesticides.

Fluctuation of whitefly population. In order to associate the presence of begomovirus with the vector, starting on the first day of transplant, and until the end of the planting cycle, 10 plants were monitored for each repetition every 10 days, using a monitoring visor, which consists of a cylindrical 40 L bucket, open on both ends and with a hole on one side. One of the ends is covered with transparent polyethylene and the other end is placed over the plant; through the hole, the plant is shaken with a wooden rod and the flies fly onto the polyethylene, where they are counted.

Incidence and severity of virosis. In each treatment, a quantification was performed (every 10 days) on plants with deformities, golden mosaics, wrinkles and leaf curling. Plants with these characteristics were divided between the total number of plants times 100. The viral severity was also estimated, using a pictorial severity scale of five types, created by the work group (Figure 1). The mean points of severity of each type of the proposed scale were obtained using the program 2-Log Ver. 1.0 (^{Mora et al., 2000}). Using the data obtained, disease progress curves were created, and the intensity of the disease was estimated using the following epidemiological parameters: Area Under the Disease Progress Curve (AUDPC), apparent infection rate (with the Weibull model, with its parameter 1/b) and final severity (Y_final) (^{Morales, 2011}).

Yield. In each plantation, the number of fruits per plant and potential yield (t ha^-1) were registered. Thirty fruits were selected at random per treatment and their length and equatorial diameter were measured, following the Mexican Official Norms NOM-189-SCFI-2017 for habanero peppers from the Yucatan peninsula.

Statistical analyses. The data obtained underwent an analysis of variance and the separation of means was carried out using the SNK method (P≤0.05) with the SAS ver. 9.11 statistical package.

Molecular identification of begomovirus. The extraction of DNA was carried out following the DNAzol ^® Reagent protocol, according to the supplier (Life technologies), with modifications. Out of each treatment and for each repetition, a young foliar tissue sample was taken from new sprouts (10 g) from 24 plants with similar virosis characteristics 120 days after transplanting (dat). The detection of begomovirus was carried out (in the 24 samples taken) by Polymerase Chain Reaction (PCR) under the following conditions: one cycle at 94 °C for 30 s, 35 cycles at 94 °C for 30 s, 55 °C for 30 s, and finally, one cycle at 72 °C for 30 s. In an Eppendorf tube, a reaction mixture was prepared with primers AV494 (5´-GCCYATRTAYAGRAAGCCMAG-3´) and AC1094 (5´-GGRTTDGARGCATGHGTACA TG-3´), which amplified a segment of the 600 pb protein of the capsid of the genus Begomovirus (^{Villanueva et al., 2013}). In order to observe the final products of PCR, electrophoresis was carried out in 1% agarose gel in Applied Biosystems equipment, stained with ethidium bromide, and in a UVP (Científica Sienna) photodocumenter. Out of the positive samples, six were chosen at random for the sequencing in both directions to Macrogen, Korea. The analysis in silico of the nucleotide sequences obtained was carried out by comparing the sequence of the NCBI, (National Center for Biotechnology Information) database, using the program BLASTN (Basic Local Alignment Search Tool).

Figure 1 Pictorial severity in the Capsicum chinense-begomovirus pathosystem. 

Fluctuation of the whitefly population. The first population record in the plantation was 40 dat and at 90 dat the largest average of insects per plant was quantified in habanero jaguar with 32 insects. At the end of the experiment, habanero jaguar and habanero criollo (ungrafted treatments), presented the highest average of the putative vector with 22 and 17 individuals per plant, respectively; whereas the treatments with grafting (muela + habanero jaguar, muela + habanero criollo, amashito + habanero jaguar y amashito + habanero criollo) displayed lower populations that fluctuated between 5.5 and 14.5 insects per plant (Figure 2). A study with tomato grafts included a report of lower populations of Bactericera cockerelli, Liriomyza spp. and aphids (^{Cortes, 2010}). This effect has been explained as an antixenosis, meaning that the insect feeds off least preferred foods, and even reproduces in them, but when there are plants with higher susceptibilities, such as the ungrafted controls, these are preferred by the insects.

Incidence and severity of virosis. At the end of the plantation, the lowest incidences of the disease were observed in muela + habanero criollo with 65%, and muela + habanero jaguar with 73% (Figure 3), in the remaining treatments, incidences were between 95 and 100%.

Figure 2.  Population fluctuation of whitefly in grafted plants and control habanero criollo and jaguar plants, planted under field conditions in Yucatan. Cycle: Spring-Summer. 

Severity was recorded starting 40 dat, and symptoms consisted of deformities, golden mosaics, wrinkling and leaf curling. According to the pictorial severity scale used, the treatments that presented the highest percentages of final virosis severity were in habanero jaguar and habanero criollo with 62 and 58 % (type 2). In these treatments, the severity increased by 44 and 48%, respectively, in relation to muela + habanero criollo, which allowed the 24% of final severity (type 1) (Figure 4).

Figure 3.  Incidence of virosis in grafted Capsicum chinense-Capsicum annuum cv. glabriusculum plants, under field conditions in Yucatan. Cycle: Spring-Summer. 

In muela + habanero criollo, a value was presented below 746.6 % per day of the AUDPC with regard to habanero criollo and habanero jaguar, which displayed the highest intensity of the disease, with 2115.1 and 2341.2 % per day. The lowest apparent infection rate was registered in grafted treatments, particularly in muela + habanero criollo, and therefore, the health of the plants was better than in the rest of the treatments. At the end of the plantation, the grafted treatments had a lower trend in the Y_final; likewise, in muela + habanero criollo, it helped obtain the lowest average for the Y_final (Table1). Investigations carried out with watermelon grafts reduced the AUDPC and improved crop yield (^{Winmer et al., 2015}). Another study showed how the use of plants grafted onto bell pepper plants improved the agronomic behavior and the management of pests and diseases (^{Sánchez et al., 2015}). The use of grafts provides vigorous plants, able to allow crop development in the presence of phytoparasites (^{Acevedo and Sánchez, 2017}). In addition, it has been reported that grafted plants can activate defense mechanisms; in tobacco plants, the grafts promoted the accumulation of phenolic compounds and salicylic acid, which stimulated the peroxidase activity, and the replication of the Potato Y virus (PVY^NTN) was reduced, resulting in the manifestation of only mild symptoms (^{Spoustová et al., 2015}). On the other hand, the resistance mechanisms in grafted tomato plants against Pseudomonas solanacearum were related to the rootstock, particularly in the lower section of the stem, where tyloses were formed (^{Grimault et al., 1994}). Grafted plants are known to share epigenetic traits, which makes them maintain their original genomes, even when they do not share DNA, although the epigenetic information is established within the plants (^{Fuentes et al., 2014}), which may explain the tolerant response to begomovirus in this study with grafted plants.

Figure 4. Severity of virosis in Capsicum chinense-Capsicum annuum cv. glabriusculum grafted plants, under field conditions in Yucatan. Cycle: Spring-Summer. 

Yield. The planting cycle was 130 days after grafting and six harvests were carried out. The analysis of variance displayed significant differences (P≤0.01). The grafted treatments had a higher yield, particularly amashito + habanero criollo, amashito + habanero jaguar and muela + habanero criollo with an estimated potential yield of 18.3, 18.4 and 18.4 t ha^-1 (Table 2). The grafted plants are tolerant to various environmental stresses, such as drought, salinity, pests and diseases, and therefore improve water and nutrient absorption, and thus have a vigorous growth, a prolongued growth period, and an increased yield (^{Acevedo and Sánchez, 2017}). In muela + habanero jaguar, the lowest estimated yield potential was found, with 9.8 t ha^-1. Although the rootstocks are often carried out to provide greater yield, they may occasionally reduce it, due to compatibility with the plant being grafted. In solanaceae and cucurbits, the graft on vigorous patterns increased the production, although it depended on the rootstock-variety combination and the conditions of plantations (^{Di Gioia et al., 2010}).

Fruits. Significant differences (P ≤ 0.01) were observed in fruit sizes (Table 2). The NOM-189-SCFI-2017 establishes acceptable parameters between 3.8 and 5.5 cm in length and between 2.5 and 3.0 cm in width for habanero peppers. The results obtained in this study show that all treatments fulfill the parameters established by the NOM, with a range of 4.6 to 4.8 cm in length and 2.8 to 3.0 cm in diameter (Table 2).

Sequencing, analysis and identification of begomovirus. The sequencing of the six samples that gave positive for begomovirus displayed a homology of 83 to 91% of nucleotidic investigation with two isolations, DZI03W20 and GU355941.1 of the GenBank by BLASTn which corresponded to the Pepper golden mosaic virus (PepGMV) and the Tomato yellow leaf curl virus (TYLCV). Out of the total, 84% were bipartite and corresponded to the treatments amashito + habanero criollo, amashito + habanero jaguar, muela + habanero criollo, muela + habanero jaguar and habanero jaguar, and the remaining 16%, monopartite, detected in the treatment of habanero criollo. The sequences helped securely identify the presence of begomovirus, although it is recommended to use specific oligonucleotides for each one of the preliminarily identified viruses. Only one of the products of PCR sent for sequencing turned out to have a percentage of homology of 83%, a percentage which may be related to the presence of undefined bases (N) in some of the positions of the resulting sequences or to another species of begomovirus. However, this percentage is actually higher if we consider the editing based on the corresponding chromatogram, a process which was not carried out in the data reported. In the case of begomovirus, it has been established that the sequence of one species with a homology ≥91% with the whole genome or part of the DNA-A component, means that we are dealing with the species in question, whereas if there is a homology <91%, then we are dealing with a new species (^{Brown et al., 2015}).

Based on the results, the muela + habanero criollo grafts conferred a higher tolerance to virosis and had a lower white fly population. The amashito + habanero jaguar, amashito + habanero criollo and muela + creole habanero grafts displayed a better agronomic behavior and crop productivity.