Role of different trichome style in the resistance of Lycopersicon hirsutum genotypes to Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae)

The tomato leaf miner (TLM), Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae), is one of the major pests that attacks commercial tomato. TLM is becoming resistant to many of the pesticides used in the tomato fields. A potential alternative method of control is host plant resistance, which may be mediated by glandular trichomes. We studied glandular and non-glandular trichomes potentially related to tomato resistance of nine tomato genotypes (Lycopersicon hirsutum Humb & Bonpl), extensively cultivated in western Iran. The antixenosis and/or antibiosis effect of the genotypes Mobil, Falat 3, Cal J N3, Dehghan, Super Strain B, Río Grande, King Ston, Early Urbana and Peto Mech were evaluated under greenhouse conditions (25 ± 1°C, 65 ± 5% R.H., photoperiod of L16: D8 h). TLM elicited the lowest egg-laying response on Peto Mech, Río Grande and King Ston; and the highest one on Dehghan. King Ston and Mobil allowed the lowest and highest larvae density on the leaves, respectively. The highest preference was observed on Mobil and the lowest one on Río Grande and King Ston. Moreover, trichome type and density of the assayed genotypes appeared to be related to TLM population density: the most infested genotype (Mobil and Cal J N3) displayed the lowest IV and VI trichome style ranks of leaf blade, vein and domatia (glandular trichomes). Significant negative relations were found between egg and larvae density with leaf I style trichome; and also larvae and adult density with IV and V style trichome. Finally, this preliminary study screened two genotypes (Río Grande and King Ston) quite promising for developing resistance programs in western Iran against TLM.


Introduction
Several phytophagous organisms are associated with Tomato (Lycopersicon hirsutum Humb & Bonpl), but a few arthropods can actually threaten its production during its growth period.Among them, the tomato leaf miner (TLM), Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae), is considered a key pest of tomato crops both in the field and under protected conditions (Khanjani 2013).Both yield and fruit quality can be significantly reduced by direct feeding and the secondary pathogens which may enter through the wounds made by the larvae (Colomo and Berta 1995;Silva et al. 1998).Females lay eggs on aerial parts of their host

Research Article
plants.Neonate larvae penetrate into tomato fruits, leaves or stems on which they feed and develop, thus creating conspicuous mines and galleries.Tomato plants can be attacked at any developmental stage, from seedlings to mature plants.The larva is generally easily found because it prefers apical buds, flowers or new fruits, on which the black feces are visible (Khanjani 2013).
Integrated Pest Management (IPM) strategies are being developed for the control of T. absoluta.Studies are being done on the use of synthetic sex pheromones (Salas 2004), chemical control (Siqueira et al. 2000), biological control (Villas Boas and Franca 1996;Torres et al. 2002), and bagging fruits (Leite et al. 2014).Concerning chemical control, several treatments are required per growing season and TLM resistance to some insecticides has been reported in several countries, for example to abamectin, cartap, permethrin and deltamethrin (Siqueira et al. 2000, Lietti et al. 2005).
Due to the hazards and adverse consequences caused by the use of chemical pesticides, researchers are currently working on more eco-friendly control methods such as plant resistance with less environmental impact.Host-plant resistance within IPM strategies can be a suitable method for pest control.Plant resistance can be expressed by antixenosis, antibiosis, tolerance, or combinations of these mechanisms (Smith 2005).Host-plant resistance to TLM has been reported for some host plants like tomato (Solanum licopersicon L.) (Oliveira et al. 2009), L. hirsutum Humb & Bonpl (Gilardón et al. 2001;Leite et al. 2001), L. pennellii Cor.(Resende et al. 2000), and L. peruvianum L. (Silva et al. 1998;Suinaga et al. 1999Suinaga et al. , 2004)).
Glandular trichomes are known to secrete a variety of secondary metabolites that are able to physically entangle insects and also, are usually associated with a toxin that kill the invaders.In Lycopersicon species, especially L. hirsutum Humb & Bonpl and L. pennellii (Cor.)D.Arcy, resistance to many arthropods depends on the type, density and exudates composition of type IV and VI glandular trichomes (Williams et al. 1980;Kennedy and Dimock 1983;Weston et al. 1989;Eigenbrode and Trumble 1993;Simmons et al. 2003).
The objective of this study was to discover/ascertain tomato (Lycopersicon hirsutum) resistance to TLM under controlled conditions by assessing the relationship between tomato genotype and glandular type I, IV, VI, VII and non-glandular type V trichomes of Lycopersicon species in order to increase the level of host resistance in cultivated tomato, by introgression with resistant accessions.
A free-choice bioassay was carried out to assess the cultivars preference of TLM adults.To evaluate/assess antixenosis, 30 T. absoluta adults (female: male ratio 2:1) were released weekly in the greenhouse and the effects of these genotypes on TLM adult performance were studied under controlled conditions (25±1°C, R.H. of 65±5% and a photoperiod of 16L: 8D h).The effects of genotypes on TLM population density and growth (number of eggs, larvae, pupae and adults) as antibiosis assay were studied in the same condition.After one month since moth releasing, leaves and branches were used for counting the stages.
Density and types of trichomes of tomato leaves were assessed.We classified the trichomes according to Luckwill (1943) method.The pubescence was evaluated on three distal-shoot leaves per plant.Per each leaf, trichomes were measured on four vein sections (each section was 5 mm long and was selected randomly) and on four circular areas (each area had 10 mm of diameter and was selected randomly) between veins on the blade.Different trichomes density ranks of either vein or blade were assessed under 25x magnifications.The measurements were replicated on twelve plants per each genotype and each plant represented a replicate.

Statistical analysis
All data were submitted to analyses of variance (ANOVA) (PROC GLM SAS Institute 2003).If significance differences were detected, multiple comparisons were made using the Tukey's test (p < 0.05).Graphs were made by Sigma-Plot (v11.0;Systat Software 2008).A dendrogram based on biological parameters of T. absoluta on different tomato genotypes was constructed after cluster analysis by Ward's method using the statistical software SPSS 13.0 (SPSS, 2004).Pearson's correlation coefficient was calculated to assess the relationship between moth population density (number of eggs, larvae, pupae + adults) and leaf trichome density, using SPSS 13.0 (2004).
The lower densities of trichome style I, IV and VI on vein were detected on Mobil; and of trichome V and VII on vein were observed on Cal J N3.The higher densities were found on leaves of (Peto Mech, Super Strain B, King Ston, Dehghan and Early Urbana respectively) (Tables 1-3).The lowest mean total number of style I, IV, V, VI and VII on domatia observed on Mobil, Cal J N3 and Falat 3 while Peto Mech, Dehghan, Super Strain B genotypes were showed the highest domatia rank.Different leaf blade trichome style of Mobil and Cal J N3 genotypes as well as trichme rank of them on domatia and vein showed the lowest rank.We observed the highest blade trichom ranks on Early Urbana, Dehghan, King Ston Genotypes (Tables 1-3).
Correlation analysis of moth population density (number of eggs, larvae, pupae + adults) and trichome features assessed on leaves of un-infested plants of the same tomato genotypes were calculated.The eggs and vein I style trichome densities were negatively related (Pearson's coefficient = -0.25;P < 0.01) (table 4).Correlation analysis indicated that egg density was positively related with IV and VI kind trichome in whole parts of leaf (0.45, 0.30, 0.56 and 0.51, 0.52 and 0.59; P < 0.01 on vein, domatia and blade respectively); V on domatia (0.24; P < 0.01) and VII on leaf blade (0.3; P < 0.01) (Table 4).No correlation was found between egg density and I style on domatia (-0.09;P > 0.05) and blade (0.08; P > 0.05); V style on vein (0.17; P > 0.05) and blade (0.14; P > 0.05) and VII on vein (0.12; P > 0.05) and domatia (0.03; P > 0.05) (Table 4).Light negative but significant relationships were found between TLM larvae abundance and Style I trichome density on vein (-0.22; P < 0.01), domatia (-0.22;P < 0.01); Style IV (-0.19;P < 0.01) and Style V (-0.23;P < 0.01) on leaf blade vs. other kinds of trichome on different parts of leaves that showed no significant correlation with larvae density (table 4).Significant negative correlations were found between TLM adults and trichomes style IV (-0.24;P < 0.01) and V (-0.27;P < 0.01) on leaf blade.No significant correlation was found between moth adult density and the other kind of trichomes on different place of leave (Table 4).

Discussion
The main purpose of this study was to evaluate the resistance of some native (Falat 3 and Dehghan) (Iranian) and non-native tomato gnotypes to TLM and to develop an experimental protocol in order to reduce the influence of the environmental factors (e.g., soil nutrients, light exposition, biocenotic interactions) on the responses of the tested genotypes.Tomato genotypes resistance to TLM reared in similar conditions are scarce on the contrary of some greenhouse observations (Oliveira et al. 2009).Even though antixenosis assay does not give a clear estimation of the TLM resistance, it was performed preliminarily on a quite large number of genotype following a free-choice experimental design aimed at selecting genotype for further assays.The scarce TLM adult infestation on King Ston, Río Grande and Falat 3 genotypes might be related to the scarce attractiveness of chemical compounds produced by the selected genotypes such as their morphological features (glandular and non-glandular trichome, wax content) which might adverse TLM success.These genotypes can express short distance attractiveness and the role of biogenic volatiles must be clarified in field where moths adopt mainly airborne dispersal.A significant positive relationship between egg density and glandular trichome types IV, VI on whole parts of leaf and VII on blade; and non-glandular trichome type V on domatia were obtained.This relationship suggests that the glandular trichome types IV, VI and VII and non-glandular trichome type V in L. hirsutum play a major role in the TLM preference.
Several studies reported that the principal compounds participating in the substrate selection for T. absoluta female oviposition was produced in the glandular trichomes of some wild and commercial tomato varieties.The principal compounds identified were 2-tridecanone (2-TD) and 2-undecanone (2-UD) (in L. hirsutum principally, trichome type VI) (Giustolin and Vendramim 1996;Maluf et al. 1997;Magalhães et al. 2001), sesquiterpenes (Freitas et al. 1998) (mainly in L. peruvianum) and acylsugars (in L. peruvianum and S. lycopersicon -commercial tomatoes) (Resende et al. 2000).Suinaga et al. (1999) demonstrated that heptadecane was the principal compound associated with reduced number and viability of T. absoluta eggs in L. peruvianum.TLM population density of larvae varied largely in the current no choice assay among the genotypes and these variations should be strictly related to the plant features, reflecting differences either in nutrients required by the pest or in secondary compounds, taking in count that all cultivars were cultivated under the same environmental conditions.Previous studies demonstrated that quantity and quality of the food sources influence fitness (survival rates, developmental duration, egg laying and other parameters) of moth (Gonçalves et al. 1998;Leite et al. 1999Leite et al. , 2001;;Ecole et al. 2000Ecole et al. , 2001;;Suinaga et al. 2004;Oliveira et al. 2009;).
We found higher density of eggs on Dehghan and Falat 3 but larva density in these genotypes was lower than egg density which lower number of larvae may reflect higher mortality of T. absoluta during leaf mesophyll feeding (Suinaga et al. 1999(Suinaga et al. , 2004a;;Ecole et al. 2000Ecole et al. , 2001)).King Ston and Early Urbana showed the lowest larvae density while these genotypes had higher oviposition rate than larvae density that it may be due to lethal effects of glandular trichom.Simmons et al. (2003Simmons et al. ( , 2004) ) reported the entrapment of M. persicae and H. armigera larvae by sticky exudates of type IV glandular trichome on L. pennellii accessions.
Based on the results from all assays performed on the interactions between tomato genotypes and T. absoluta, three distinct groups of genotype can be separated.The resistant group (cluster A) includes Río Grande, King Ston and Falat 3 on which the mean larva, adult population and egg density of the moth were largely lower than that of the other genotypes vs in these genotypes percentage of population reduction were largely higher than that.The susceptible group (cluster C) consisted of Cal J N3, Mobil and Peto Mech; the mean larva, adult population density and laid eggs of TLM on two initial genotypes were largely the highest while % of population reduction on theses genotypes were low.Finally, Early Urbana, Super Strain B, and Dehghan were included in an intermediate group (cluster B).
It was found that susceptible genotypes/the susceptible group (Cal J N3 and Mobil) showed little density of glandular trichomes IV and VI on different parts of the leaf, while the resistant one (Río Grande, King Ston and Falat 3) showed more density of these trichome types than sensitive one.Our result was the same/similar as previous studies where they observed that type IV glandular trichomes were proved to be a basis of resistance in L. pennellii to TSSM (Weston et al. 1989), M. persicae (Simmons et al. 2003), H. armigera (Simmons et al. 2004) and resistance of L. hirsutum to Spodoptera exigua (Hübner) (Eigenbrode and Trumble 1993).
There was a significant negative relationship between egg and larva density and type I non-glandular trichomes on leaf vein, which probably caused a negative effect on individual moth population fitness like egg deposition and larvae feeding.Moth population (i.e.larvae and adults) appeared to be largely negative related to the leaf physical structures like IV (glandular) and V (non-glandular) trichome style in the assayed genotypes.Leite et al. (2000) found that some tomato varieties resistant to TLM had denser pubescent leaves than susceptible varieties (mainly glandular trichome versus non glandular trichome).Giustolin and Vendramim (1994) observed that the resistance of L. hirsutum to T. absoluta is usually attributed to the allomones 2-TD and 2-UD present in the leaf glandular trichomes of this tomato species, which are absent in L. esculentum.Lycopersicon esculentum, in general, does not have the trichome kind VI that produces these compounds.
Finally, the current investigation set up an experimental protocol and should be considered preliminary on the study of a pool of tomato genotypes about resistance to TLM.Further studies on these genotypes are required in order to detail the factors (biochemical and morphological ones) which actually act for resistance and susceptibility.

Figure 1 .
Figure 1.Tuta absoluta density (mean + SE), A) adult, B) egg, C) larvae, and D) % of population reduction; (n=10) after one month of population development on the nine tomato genotypes.Different letters above the bars indicate significant differences among genotypes (P< 0.05, Tukey's test).

Figure 2 .
Figure 2. Cultivars Ward's dendrogram generated from tomato leaf miner populations developed on the assayed cultivars.

Table 1 .
Mean (± SE) trichome rank density (number of trichomes per cm 2 ) on leaves vein of nine tomato genotypes, separated by trichome type.

Table 2 .
Mean (± SE) trichome rank density (number of trichomes per cm 2 ) on leaves domatia of nine tomato genotypes, separated by trichome type.

Table 3 .
Mean (± SE) trichome rank density (number of trichomes per cm 2 ) on leaves blade of nine tomato genotypes, separated by trichome type.
Means within columns followed by different letters are significantly different (Tukey's test: P< 0.05).

Table 4 .
Correlation coefficient (Pearson's coefficient) between tomato leaf miner egg, larva and adult abundances/ densities and leaf trichome rank densities of nine tomato genotypes.