1. Plants face many environmental stresses that can impact their survival, development, and fitness. Insects are the most diverse, abundant and threatening herbivores in nature. As a consequence, plants produce direct chemical and physical defenses to reduce herbivory. They also release volatiles to recruit natural enemies that indirectly protect them from herbivory. The recruitment of parasitic wasps can benefit plant fitness because they ultimately kill their insect hosts.
2. Recently, studies showed that parasitoids can indirectly mediate plant defenses by modulating herbivore oral secretions. In addition to the direct benefits of parasitoids in terms of reducing herbivore survival, we tested if the reduction in induced defenses by parasitized caterpillars compared to non-parasitized caterpillars may reduce the costs associated with defense expression.
3. We provide evidence that tomato plants treated with saliva from parasitized caterpillars have significantly higher fitness parameters including increased flower numbers (16.3%) and heavier fruit weight (15%), compared to plants treated with saliva from non-parasitized caterpillars. Since plants were grown without actual herbivores, the higher values for these fitness parameters were due to lower costs of induced defenses and not due to reduced herbivory by parasitized caterpillars. Furthermore, the resulting seed germination time was shorter and the germination rate was higher when the maternal plants were previously exposed to parasitized herbivore treatment compared to control (non-treated) plants.
4. Overall, application of saliva did not result in transgenerational priming of offspring defense responses. However, offspring of parents exposed to caterpillar saliva had lower constitutive levels and higher induced levels of trypsin inhibitor than offspring from unexposed parents.
5. This study shows that the saliva of parasitized caterpillars can modulate plant defenses and further demonstrates that the lower induction of plant defenses is associated with elevated plant fitness in the absence of herbivore feeding, suggesting that induced plant defenses are costly. 

Caterpillar glucose oxidase enzyme activities in labial glands

   GOX activities were assessed as described by Eichenseer et al. (1999). The changes in absorbance values were recorded at 460 nm in a plate reader. Protein in each sample was quantified by Bradford assays using BSA (bovine serum albumin, Fraction V, Omnipur) as the protein standard (Bradford 1976).

Plant defense response

Twenty-four hours after saliva application, plant tissues were collected from the treated leaflet. RNA extraction, cDNA synthesis and qRT-PCR analysis were processed as described (Tan et al., 2018).

Forty-eight hours after saliva application, plant tissues were collected from the third terminal leaflet for peroxidase (POD), polyphenol oxidase (PPO) and trypsin inhibitor (TI) assays. PPO and POD assays were performed as described by Felton et al. (1989). For trypsin inhibitor (TI) activity assays, samples were prepared and the absorbance values were recorded at 247 nm in a plate reader. Protein concentration in each sample was quantified by using Bradford assays.

Plant fitness

A serrated wounding tool was used to wound the third terminal leaflet of tomato plants and immediately 15 µl of insect saliva from parasitized or non-parasitized caterpillars was applied with a pipette. Two days later, the same process was repeated on two leaflets of the 4^th leaf. Four days after the first application, insect saliva was applied to two leaflets of the 5^th leaf. In total, insect saliva was applied three times to five leaflets to simulate caterpillar feeding. Plants were watered as needed and pots were rotated randomly every week. Testing plants were excluded from other biotic and/or abiotic factors which might have varied with natural conditions.

The first flowering date, flower number, fruit weight, seed weight and seed number were recorded to represent plant fitness. The first flowering date is the number of days between the first saliva application and appearance of the first flower. 

Second generation performance

Thirty seeds from each maternal plant were sown in potting soil in the greenhouse and a total of 10 maternal plants of each treatment were used. Pots were placed on trays and water was added to the tray to maintain soil moisture. Seeds were observed daily for germination time. Germination rates of seeds were calculated at 11 days after sowing as follows: (germinated seed #/ total seed # (30 seeds)) x 100.

In the offspring defense response assay, there were three maternal plant treatments (C, P, and NP) and two herbivore treatments (N: no herbivore treatment, and H: herbivore treatment). Seeds were planted as described above. Plants with 3 fully expanded leaves were used for the experiment. Labial glands were collected and homogenized from non-parasitized caterpillars. For the herbivore treatment (H), 15 µl (1 µg/µl protein) of insect saliva was applied after mechanical wounding of the third terminal leaflet (counting from the bottom) of tomato plants, while the other half plants were without herbivore (N) treatment. 

Twenty-four and 48 hours after saliva application, plant tissues were collected from the third terminal leaflets for polyphenol oxidase (PPO), peroxidase (POD) and trypsin inhibitor (TI) assays as described above. Each plant was only used once.

Statistical analyses

Data were transformed as needed to obtain a normal distribution and to address homogeneity of residual variances; SAS 9.4 (SAS Institute Inc) was used for data analyses. The GOX activities in labial glands were analyzed by Student’s t test. Plant fitness was analyzed by One-way ANOVA (Proc GLM), and plant defense enzyme activities and seed germination rate were analyzed by Two-way ANOVA (Proc GLM), followed by means comparisons using the Tukey’s Least Significant Difference (LSD) test (significance level, P<0.05). Seed germination days were analyzed by Kruskal-Wallis tests, followed by pairwise multiple comparison (DSCF) tests (significance level, P<0.05).