Recessive Resistance Derived from Tomato cv. Tyking-Limits Drastically the Spread of Tomato Yellow Leaf Curl Virus

The tomato yellow leaf curl disease (TYLCD) causes severe damage to tomato (Solanum lycopersicum L.) crops throughout tropical and subtropical regions of the world. TYLCD is associated with a complex of single-stranded circular DNA plant viruses of the genus Begomovirus (family Geminiviridae) transmitted by the whitefy Bemisia tabaci Gennadius (Hemiptera: Aleyrodidae). The tomato inbred line TX 468-RG is a source of monogenic recessive resistance to begomoviruses derived from the hybrid cv. Tyking F1. A detailed analysis of this germplasm source against tomato yellow leaf curl virus-Israel (TYLCV-IL), a widespread TYLCD-associated virus, showed a significant restriction to systemic virus accumulation even under continuous virus supply. The resistance was effective in limiting the onset of TYLCV-IL in tomato, as significantly lower primary spread of the virus occurred in resistant plants. Also, even if a limited number of resistant plants could result infected, they were less efficient virus sources for secondary spread owing to the impaired TYLCV-IL accumulation. Therefore, the incorporation of this resistance into breeding programs might help TYLCD management by drastically limiting TYLCV-IL spread.


Introduction
The tomato yellow leaf curl disease (TYLCD) is one of the major yield-limiting factors of tomato (Solanum lycopersicum L.) crops in tropical and subtropical regions [1,2]. This devastating disease is caused by a complex of single-stranded circular DNA plant viruses of the genus Begomovirus (family Geminiviridae) that are transmitted in a persistent circulative manner [3] by whitefly (Hemiptera: Aleyrodidae) members of the Bemisia tabaci Gennadius cryptic species. Among them, the monopartite tomato yellow leaf curl virus (TYLCV) is the most widespread and economically important worldwide [4]. Tomato plants affected by TYLCD exhibit characteristic symptoms of stunting, yellowing, upward curling of leaves, and suffer premature dropping of flowers and reduction of marketable fruits that can result in 100% yield loss when infections occur during early growth stages [1,5].
The control of TYLCD is often based on intensive chemical treatments to limit vector population and/or by using physical barriers, both with limited success. In addition, chemical control result in deleterious environmental effects and can determine selection of insecticide-resistant B. tabaci populations [6,7]. Therefore, in this scenario, the use of resistant tomato cultivars is the most environmentally sustainable and economically viable approach to reduce TYLCD damage.
Susceptibility to TYLCV requires virus replication in the plant cell nucleus via a double-stranded DNA intermediate, movement to adjacent cells through plasmodesmata, long-distance movement through the phloem, and further acquisition by vectors for transmission from plant to plant to reinitiate the infection cycle. A block at any of these steps, either by active defense responses or by incompatible interactions of viral and host factors, may lead to virus resistance [8]. A number of virus resistance factors have been derived from wild tomato relative species that can help restricting TYLCD-associated virus infection and limit disease damage [9]. The partially dominant Ty-1 resistance gene derived from the S. chilense accession LA1969 [10] is so far the most widely used commercially. However, the performance of Ty-1 or other plant host resistance genes varies depending upon the TYLCD-causing virus [11,12]. Also, breakdown of Ty-1 resistance can occur under high inoculum pressure [13] and it often shows lower effectiveness in heterozygous plants used commercially [14]. Furthermore, emergence of recombinant TYLCD-viruses with novel pathogenic characteristics [15,16] might pose a threat to available resistance genes. Recently, we reported an alternative tomato source of begomovirus resistance named TX 468-RG, which was derived via selfing from the commercial F 1 hybrid "Tyking" (released by Royal Sluis, The Netherlands). This inbred line displayed high levels of resistance to bipartite and monopartite begomoviruses associated to TYLCD based on single recessive gene control [17,18]. A previous work allowed us to clarify the genetic control of the TX 468-RG resistance and to demonstrate its effectiveness to a range of TYLCD-associated monopartite begomoviruses [17]. However, the resistance mechanism remains unsolved. In the present study, a detailed set of analyses was conducted to understand the restriction to accumulation of an isolate of the Israel strain of TYLCV (TYLCV-IL) in TX 468-RG and to evaluate to what extent this resistance can help to limit virus spread under field conditions. We conclude that although no effect in inoculated leaves was observed, systemic infection of TYLCV-IL was impaired in TX 468-RG plants, resulting in reduced virus accumulation. As a consequence, TX 468-RG resistance was effective to limit primary and secondary spread of the virus. Therefore, this resistance is highly recommended for breeding purposes to control damage caused by TYLCV-IL in tomato.

Tomato Plants, Virus Isolate, and Whitefly Population
A tomato F 8 inbred line (named as "TX 468-RG") was derived via repeated selfing and selection steps from the commercial F 1 hybrid "Tyking" (released in the 1990s by Royal Sluis, The Netherlands) and used in all assays. This germplasm displayed high levels of resistance to bipartite begomoviruses [18] as well as to a range of TYLCD-associated monopartite begomoviruses [17]. The open-pollinated tomato cv. Moneymaker (MM) (IHSM-UMA-CSIC seed bank) was used as susceptible control in the experiments.
The infectious clone of the isolate [ES:Alm:Pep:99] of TYLCV-IL (TYLCV-IL [ES:Alm:Pep:99], from now on, TYLCV-IL) (GenBank accession number AJ489258), has been described elsewhere [19]. Healthy B. tabaci adult individuals were obtained from a colony of the Mediterranean (MED) species (formerly known as Q biotype) originated from field individuals collected in Malaga, Spain. Whiteflies were reared on melon (Cucumis melo L. cv. ANC42, IHSM-UMA-CSIC seed bank) plants within wooden cages covered with insect-proof nets, in an insect-proof glasshouse with temperature control (22-27 • C day and 17-20 • C night) and supplemental light when needed.

Virus Inoculation
Agrobacterium tumefaciens-mediated stem puncture inoculation (agroinjection) with TYLCV-IL was conducted on plants at the three-leaf growth stage as described previously [20]. Also, for TYLCV-IL local accumulation studies, A. tumefaciens-mediated leaf tissue infiltration (agroinfiltration) studies were conducted following Tomás et al. [21]. Either for agroinjection or for agroinfiltration, bacterial suspensions at OD 600 = 1.0 were used. For B. tabaci-mediated inoculation, viruliferous whiteflies were obtained by providing insect adults with a 48 h acquisition access period (AAP) on systemically infected young leaves of MM plants agroinjected with TYLCV-IL three weeks earlier.
For graft inoculation, healthy scions of TX 468-RG and MM were grafted onto MM and TX 468-RG plants previously infected with TYLCV-IL by agroinjection. Scions consisted of a stem piece containing a leaf with its associated lateral shoot meristem obtained from healthy test plants. Grafts of healthy MM or TX 468-RG scions were made on both MM and TX 468-RG infected rootstocks by splice-grafting. The grafted plants were then kept in a shaded and humid environment within a growth chamber for seven days and then moved to the insect-proof glasshouse (see below) until evaluation . Evaluations of TYLCD  symptom and presence/accumulation of TYLCV-IL in young developing leaves of the scions was done  at 28 days post-grafting by visual observation and by tissue blot and dot blot hybridization analyses of  individual plants, respectively. Plant inoculations were performed in a growth chamber (25 • C day and 20 • C night, 70% relative humidity, with a 16 h photoperiod at 250 µmol· s −1 · m −2 photosynthetically active radiation), and the inoculated plants were then kept until analyzed in an insect-proof glasshouse with temperature control (approximately 16 h day length, at 22 to 27 • C during the day and 17 to 20 • C at night) and supplemental light when needed.

Primary and Secondary TYLCV-IL Spread Experiments
Primary and secondary spread experiments were conducted at "La Mayora" Experimental Station (Malaga, southern coastal Spain) essentially as described by Rodríguez-López et al. [22]. Briefly, primary spread of TYLCV-IL, i.e., virus spread to healthy plants from external source of viruliferous vectors [23], was simulated in medium-scale experiments conducted within insect-proof net, walk-in structures

Virus Detection and Determination of TYLCV-IL Accumulation Levels in Infected Tomato Plants
Virus presence was analyzed in young leaves of inoculated plants by tissue-blot hybridization of freshly cross-sectioned leaf petioles or by Southern blot hybridization of total nucleic acid (TNA), according to Tomás et al. [21] using a probe specific to TYLCV-IL. To monitor virus accumulation in young leaf tissues of TX 468-RG and MM plants inoculated with TYLC-IL, dot blot hybridization analyses were conducted. For this, TNA extracts were prepared from the second leaf from the apex, as described by Celix et al. [25], except that TNA were dissolved in 50 µL of sterile H 2 O in the final step. TNA extracts were quantified, the concentrations standardized at 200 ng/µL, and used for dot blot hybridization. For estimation of TYLCV-IL accumulation levels, different groups of inoculated plants were sampled in a destructive manner at 7, 14, 21, and 28 days post inoculation (dpi) (about 14 plants per evaluation date). Two microliter per sample of each dilution of a dilution series 1:1, 1:2, 1:4, 1:8, 1:16, 1:32, 1:64, 1:128, 1:256, and 1:512 of TNA were applied to positively charged nylon membranes (Roche Diagnostics GmbH, Mannheim, Germany). Two replicated membranes were prepared for DIG-labeled DNA-probe hybridization, one for hybridization with a probe to TYLCV [26] and the other with a probe to a gene fragment coding for 18S ribosomal RNA (18S rRNA) [27], as loading control. Healthy MM plants were used as negative controls. Assessment of TYCLV-IL DNA content was done by densitometric measure of hybridization signals obtained in autoradiographs. Plant ribosomal RNA signals were used as an internal standard to equilibrate TNA loading among samples. Densitometry measurements were expressed in pixels measured using Quantity One Software v 4.6.7 (VersaDoc MP 4000 Imaging System; BioRad), giving an arbitrary value of 1000 to that of one of the MM infected plants per date analyzed, and referring the other values to this one. In every case, the densitometry values used for estimation of TYLCV-IL accumulation levels were those from a dilution that fell within the linear range of the relationship between dilution and densitometry measure. Distribution of values was represented by Box-and-Whisker plots [24].

Data Analysis
Statistical effects of genotypes and/or treatments in the different experiments were analyzed with the IBM SPSS Statistics v. 22 software by applying Generalized Linear Models (GzLM), in which all possible pair-wise comparisons were performed using the sequential Bonferroni method for error correction. For the case of comparisons of the TYLCV-IL accumulation levels estimated, the GzLM used Logarithm as the link function and Normal as the underlying distribution. For TYLCV incidence in the primary and secondary spread experiments, data set was expressed as the number of infected and non-infected plants at each time point and were analyzed by GzLM using Logit as the link function and Binomial as the underlying distribution. Also, a general disease incidence pressure was estimated for combinations of genotypes and treatments in those experiments by calculating the Area Under the Disease Progression Curve (AUDPC) using the formula AUDPC = Σ((Yi + Y(i+1))(T(i+1) -Ti))/2, where Yi = proportion of infected plants at date i, and Ti = time (in days) at date i. For comparing means of AUDPC values, the GzLM used Identity as the link function and Normal as the underlying distribution. Finally, for comparison of the number of whiteflies visiting the two tomato genotypes in the preference experiment, whiteflies per plant data were analyzed by GzLM using Logarithm as the link function and Negative Binomial as the underlying distribution.

TYLCV-IL Accumulation is Highly Restricted in TX 468-RG
Systemic TYLCV-IL infection was observed in all inoculated MM plants, whereas lower incidence levels were observed in TX 468-RG plants, especially at initial evaluation dates (see number of plants infected vs. number of plants inoculated at the bottom of boxes, Figure 1A). Also, no symptoms were observed in any TX 468-RG-infected plant, whereas typical TYLCD symptoms were observed in all infected MM plant from 21 dpi. Interestingly, viral DNA accumulation in infected plants revealed a strong restriction in TX 468-RG. Although viral DNA could be detected in young tissues of a number of these plants, confirming that they are not immune (sensu [28]) to TYLCV-IL, significantly lower accumulation levels were observed ( Figure 1A) when compared to plants of the susceptible MM control (see e.g., dot blot hybridization for two representative plants from the two genotypes at 28 dpi in Figure 1B). Thus, at 28 dpi, about eight times lower viral DNA accumulation was detected in young tissues of systemically infected TX 468-RG than in the equivalent ones of MM plants.
By contrast, no major differences in accumulation levels of viral forms derived from the input DNA was observed at local agroinfiltrated tissues of TX 468-RG and MM plants (Figure 2), suggesting that constraints at initial steps of viral infection does not seem to be the basis of the resistance mechanism. Therefore, differences in virus accumulation might occur during the systemic infection process.
To investigate whether virus resistance in TX 468-RG occurred during the process of systemic virus translocation, grafting of TX 468-RG and MM healthy scions onto TYLC-IL-infected rootstock were analyzed. Although tissue blot hybridization is not a quantitative assay, the results shown in Figure 3 strongly suggest that resistance in TX 468-RG impaired TYLC-IL accumulation during systemic translocation and was operative even under a continuous virus supply. Thus, only traces of virus were observed in all TX 468-RG scions either grafted onto MM or TX 468-RG rootstocks infected with TYLCV-IL, supporting that the level of virus accumulation in this genotype was not related to the amount of inoculum supplied. Also, no TYLCD symptoms were ever observed in any of the TX 468-RG grafted scions. In contrast, severe symptoms and higher virus accumulation was observed in MM scions grafted either onto MM or onto TX 468-RG rootstocks infected with TYLCV-IL, even if in the latter case low virus titers occurred in the rootstock (see Figure 1). Therefore, the begomovirus-resistance present in TX 468-RG was able to impair the TYLCV-IL systemic translocation within the plant even under high viral loads.

TX 468-RG Results in a Reduced Primary Spread of TYLCV-IL
The results of the study conducted to estimate the effect of plant genotype on primary virus spread showed that significantly lower incidence levels were observed in TX 468-RG in relation to MM, either in choice or in no-choice conditions in all the three independent experiments conducted ( Figure 4). Over 80% incidence was achieved at 28 dpi in MM, whereas about half of this value was observed for TX 468-RG. The resistant line also displayed a significant delay in the onset of the infection with a significantly reduced number of infected plants at 10 dpi. All this resulted in significantly lower AUDPC values for TX 468-RG when compared to MM (5.2 ± 0.3 vs. 18.3 ± 0.6 for no-choice and 2.7 ± 1.4 vs. 18.4 ± 0.9 for free-choice) and, therefore, lower disease pressure in TX 468-RG under our experimental conditions. Thus, the use of TX 468-RG-derived resistance would result in a significantly reduced primary spread of TYLCV-IL and lower virus pressure under field conditions.

TX 468-RG Limits Secondary Spread of TYLCV-IL from Infected Source Plants to either TX 468-RG or MM
Because the previous results indicated that TYLCV-IL could infect TX 468-RG plants during virus spread, secondary spread of TYLCV-IL to both TX 468-RG and MM healthy test plants was assessed. As summarized in Figure 5, the results of the three independent experiments conducted clearly showed that significantly lower secondary spread occurred when TX 468-RG was present, as source and/or test plant, with no infections observed when this genotype was used as source and test plant. Therefore, theTYLCV-IL resistance of TX 468-RG is strongly effective in reducing the secondary spread of this virus. As a result, a significantly lower infection pressure occurred in test plants (AUDPC 16.0 ± 2.9 for secondary spread from MM to MM vs. 5.5 ± 1.3, 0.8 ± 0.8, and 0, for secondary spread from TX 468-RG to MM, MM to TX 468-RG, and TX 468-RG to TX 468-RG, respectively). Therefore, lower TYLCD infection pressure under field conditions is expected when using this resistance.

B. tabaci Exhibited no Preference for either MM or TX 468-RG
No significant difference in whitefly preference for either MM or TX 468-RG tomato genotypes was observed based on the data obtained from the no-choice preference experiment (MM, 9.5 ± 1.4 vs. TX 468-RG, 8.3 ± 1.3 whiteflies per plant; p = 0.553) ( Figure 6). Therefore, results from spread experiments were not influenced by differences in vector behavior depending on the tomato genotype.

Discussion
The use of tomato cultivars with genetic resistance has been the most effective strategy to minimize losses caused by viral diseases, including pathosystems involving Begomovirus species [9]. The current study advances our understanding about the benefits of using virus resistant genotypes to limit TYLCV spread in the field. A compatible virus interaction with a plant involves effective multiplication and movement of the virus from the site of infection and throughout the plant as well as effective transmission to other plants to guarantee maintenance in nature [29]. However, during the virus/host plant co-evolution process, distinct host plant mechanisms to restrict virus infection can be selected and sources of resistance can become predominant in natural plant populations under continuous disease pressure due to their selective advantages [28]. Here, we demonstrated that the previously reported monogenic recessive resistance against TYLCV-IL present in TX 468-RG [17] strongly impairs systemic virus infection, resulting in a significantly reduced primary and secondary spread of the virus. Therefore, this resistance is proposed as a good and effective alternative for breeding purposes to reduce TYLCD damage in commercial tomatoes.
selective advantages [28]. Here, we demonstrated that the previously reported monogenic recessive resistance against TYLCV-IL present in TX 468-RG [17] strongly impairs systemic virus infection, resulting in a significantly reduced primary and secondary spread of the virus. Therefore, this resistance is proposed as a good and effective alternative for breeding purposes to reduce TYLCD damage in commercial tomatoes.    Figure S1.  Previous observations for the usefulness of TX 468-RG resistance to control TYLCV-IL [17] were confirmed by the detailed studies conducted here. This is important because TYLCV-IL is one of the most widespread TYLCD-associated virus worldwide [4] causing severe economic losses to tomato production where present. TX 468-RG plants, however, are not immune and can be infected with this virus. Nevertheless, we show here that a severe restriction to systemic virus accumulation occurred in infected plants harboring this recessive resistance. Leaf agroinfiltration experiments demonstrated that local accumulation of TYLCV-IL was not impaired in TX 468-RG plants, suggesting no effect on the initial steps of infection and that the restriction occurred during the systemic infection process [30]. Based on our studies, however, we cannot rule out whether the latter restriction is due to impairment of viral movement and/or to the triggering of plant defenses, such as gene silencing [31]; further research will be needed to address this issue. Similar restriction to systemic infection was shown to be conferred to another monopartite begomovirus, Tomato leaf curl virus by the recessive resistance tgr-1 gene present in the line FLA-653 [32], derived from a cross between the resistant genotypes "Tyking" and S. chilense LA2779. In this case, however, Bian et al. [32] observed that in addition to a limitation of long-distance translocation a restriction also occurred at local level, which was associated with impaired cell-to-cell movement of the virus. Restriction to systemic virus accumulation has also been reported for the cultivar "Tyking" when confronted by other begomoviruses [33]. Interestingly, we showed that similarly to that reported for the breeding tomato line TY172 resistant to TYLCV [34], the plant defense mechanism operating in TX 468-RG against TYLCV-IL is not overcome by continuous supply of high loads of virus. Therefore, contrary to that observed for the Ty-1 gene widely used commercially [13], effectiveness of the recessive resistance studied here is expected even under high disease pressure, in accordance with preliminary field studies [17]. Previous observations for the usefulness of TX 468-RG resistance to control TYLCV-IL [17] were confirmed by the detailed studies conducted here. This is important because TYLCV-IL is one of the most widespread TYLCD-associated virus worldwide [4] causing severe economic losses to tomato production where present. TX 468-RG plants, however, are not immune and can be infected with this virus. Nevertheless, we show here that a severe restriction to systemic virus accumulation occurred in infected plants harboring this recessive resistance. Leaf agroinfiltration experiments demonstrated that local accumulation of TYLCV-IL was not impaired in TX 468-RG plants, suggesting no effect on the initial steps of infection and that the restriction occurred during the systemic infection process [30]. Based on our studies, however, we cannot rule out whether the latter restriction is due to impairment of viral movement and/or to the triggering of plant defenses, such as gene silencing [31]; further research will be needed to address this issue. Similar restriction to systemic infection was shown to be conferred to another monopartite begomovirus, Tomato leaf curl virus by the recessive resistance tgr-1 gene present in the line FLA-653 [32], derived from a cross between the resistant genotypes "Tyking" and S. chilense LA2779. In this case, however, Bian et al. [32] observed that in addition to a limitation of long-distance translocation a restriction also occurred at local level, which was associated with impaired cell-to-cell movement of the virus. Restriction to systemic virus accumulation has also been reported for the cultivar "Tyking" when confronted by other begomoviruses [33]. Interestingly, we showed that similarly to that reported for the breeding tomato line TY172 resistant to TYLCV [34], the plant defense mechanism operating in TX 468-RG against TYLCV-IL is not overcome by continuous supply of high loads of virus. Therefore, contrary to that observed for the Ty-1 gene widely used commercially [13], effectiveness of the recessive resistance studied here is expected even under high disease pressure, in accordance with preliminary field studies [17].
Spread of TYLCD by whiteflies in the field generally occurs in two phases. In a first instance, onset of the disease in a crop occurs from external sources of inoculum, i.e., primary spread, which usually leads to a random distribution pattern of primary infections foci [23]. Then, secondary spread from these primary sources of infection can occur within the crop, whose intensity will strongly depend on the magnitude of the insect vector population present in the crop. Here, we demonstrate that the resistance present in TX 468-RG was effective to severely limit primary spread of TYLCV-IL, thus resulting into a first barrier to reduce TYLCD field epidemics. As no significant difference on the B. tabaci preference was observed between TX 468-RG and MM, spread restriction was mainly based on the resistance to TYLCV-IL present in the former genotype. We observed, however, that a small number of TX 468-RG plants could get infected, even though no virus damage occurs in these plants due to the resistance factor [17]. These infected TX-468-RG plants constitute potential sources for secondary virus spread within the crop or even for primary spread to nearby susceptible crops. In fact, there are reports supporting the threat of some tomato genotypes resistant but not immune to TYLCV as efficient sources for secondary virus spread [35,36]. However, we demonstrate here that the level of resistance expressed by TX 468-RG was effective to drastically restrict secondary spread of the virus. This finding suggests that TX 468-RG plants restricted virus accumulation to levels that resulted in impaired insect transmission [37], and is compatible with the low transmission rates observed by Lapidot et al. [35] from tomato plants highly resistant to TYLCV. This is an important aspect for TYLCV management, as virus transmission from infected resistant plants can determine the field success of the resistant genotype to control TYLCD epidemics [35,36]. In conclusion, the resistance present in TX 468-RG was effective to reduce primary and secondary spread of TYLCV-IL. Therefore, this resistance can help in effective management of TYLCD under field conditions. Even though some TYLCD-associated virus-resistant cultivars are available commercially, owed to the resistance breakdown occasionally observed under high disease pressure [13], control measures traditionally have emphasized reducing vector populations through chemical control [38]. Concerns exist, however, about intensive use of insecticides due to the environmental damage caused and to development of pesticide resistance in insect population [6,7]. Therefore, based on the results shown here, the TYLCV-IL resistance present in TX 468-RG offers a good alternative for TYLCD-resistance breeding programs to incorporate this character into new hybrids and cultivars. Moreover, since TX 468-RG has proven to also be resistant to bipartite begomoviruses [18] with a similar reduction in viral accumulation, the same epidemiological impact on their spread is expected. Therefore, this source of resistance can be an important component of broad management of begomoviruses in tomato.
wrote the manuscript. R.F.-M. performed statistical analyses of data. E.M. conceived and designed field experiments, supervised the studies and wrote the manuscript and R.O., supervised the studies and wrote the manuscript. All authors read and approved the final manuscript.