Spider mites suppress tomato defenses downstream of jasmonate and salicylate independently of hormonal crosstalk

Abstract

Plants respond to herbivory by mounting a defense. Some plant-eating spider mites (Tetranychus spp.) have adapted to plant defenses to maintain a high reproductive performance. From natural populations we selected three spider mite strains from two species, Tetranychus urticae and Tetranychus evansi, that can suppress plant defenses, using a fourth defense-inducing strain as a benchmark, to assess to which extent these strains suppress defenses differently. We characterized timing and magnitude of phytohormone accumulation and defense-gene expression, and determined if mites that cannot suppress defenses benefit from sharing a leaf with suppressors. The nonsuppressor strain induced a mixture of jasmonate- (JA) and salicylate (SA)-dependent defenses. Induced defense genes separated into three groups: 'early' (expression peak at 1 d postinfestation (dpi)); 'intermediate' (4 dpi); and 'late', whose expression increased until the leaf died. The T. evansi strains suppressed genes from all three groups, but the T. urticae strain only suppressed the late ones. Suppression occurred downstream of JA and SA accumulation, independently of the JA-SA antagonism, and was powerful enough to boost the reproductive performance of nonsuppressors up to 45%. Our results show that suppressing defenses not only brings benefits but, within herbivore communities, can also generate a considerable ecological cost when promoting the population growth of a competitor.

Keywords: Solanum lycopersicum (tomato); Tetranychus spp. (spider mite); defense suppression; herbivore communities; hormonal crosstalk; jasmonic acid (JA); salicylic acid (SA).

Fig 1

The reproductive performance of the jasmonic acid (JA)-defense-inducing and -susceptible Tetranychus urticae Santpoort-2 increases on tomato (Solanum lycopersicum) leaflets shared with spider mites from suppressor strains. The figure shows the average (+ SEM) number of eggs produced by adult female mites of strain T. urticae Santpoort-2 per 4 d on leaflets simultaneously coinfested with 15 adult females of T. urticae DeLier-1, T. evansi Viçosa-1, or T. evansi Algarrobo-1, or with T. urticae Santpoort-2 as a control. Numbers within the bars indicate the average egg production. Bars annotated with different letters were significantly different according to Fisher's least significant difference (LSD) test (P ≤ 0.05) after ANOVA.

Fig 2

The amounts of 12-oxo-phytodienoic acid (OPDA), jasmonic acid (JA), jasmonic acid-isoleucine (JA-Ile), and free salicylic acid (SA) accumulated in spider mite-infested tomato (Solanum lycopersicum) leaflets during the course of 7 d. The figure shows the average (+ SEM) amounts of OPDA (a), JA (b), JA-Ile (c), and free SA (d) at 1, 4, and 7 d postinfestation (dpi) of tomato leaflets with 15 adult females from strain Tetranychus urticae DeLier-1, T. evansi Viçosa-1, T. evansi Algarrobo-1, or T. urticae Santpoort-2. Uninfested leaflets were used as controls. OPDA was not detected at 1 dpi in control, T. urticae DeLier-1, and T. evansi Viçosa-1 samples. Bars annotated with different letters were significantly different according to Fisher's least significant difference (LSD) test (P ≤ 0.05) after ANOVA. Bars marked with ‘ns’ did not test differently in the ANOVA. Data were log-transformed before the statistical analysis.

Fig 3

Relative transcript abundances of various defense-related genes in spider mite-infested tomato (Solanum lycopersicum) leaflets during the course of 7 d. Based on the Tetranychus urticae Santpoort-2 samples, the genes separated into three groups: those whose transcript abundances were highest at 1 d postinfestation (dpi) were annotated as ‘early’; those with a peak at 4 dpi as ‘intermediate’; and those with a peak at 7 dpi as ‘late’. Compared with T. urticae Santpoort-2, both Tetranychus evansi Viçosa-1 and T. evansi Algarrobo-1 mites suppressed genes from all three groups, while T. urticae DeLier-1 mites only moderately suppressed the ‘late’ defense genes. Uninfested leaflets were used as controls. The bars represent the means (+ SEM) of the normalized transcript abundances scaled to the lowest mean value per 7 d gene panel. Transcript abundances were normalized to actin. Numbers above the bars indicate the mean value represented by the bar. Expression data were statistically evaluated per day and bars annotated with different letters were significantly different according to Fisher's LSD test (P ≤ 0.05) after ANOVA. Gene identifiers and corresponding references can be found in Table S2. GAME-1,Glycoalkaloid Metabolism-1;JIP-21,Jasmonate-inducible protein-21;LX,RNase Lycopersicon extravacuolar;PI-IIc,Proteinase Inhibitor IIc;PPO-D,Polyphenol-oxidase-D;PPO-F,Polyphenol-oxidase-F;PR-1a,Pathogenesis-related protein 1a;PR-P6,Pathogenesis-related protein 6;TD-2,Threonine Deaminase-2;THM27,Tomato Hypocotyl Myb 17.

Fig 4

The amounts of jasmonic acid-isoleucine (JA-Ile) and salicylic acid (SA), along with transcript abundances of Proteinase Inhibitor IIc (PI-IIc) and Pathogenesis-related 1a (PR-1a) in tomato (Solanum lycopersicum) leaflets after 7 d of infestation with inducer Tetranychus urticae Santpoort-2, suppressor T. evansi Viçosa-1, and suppressor T. urticae DeLier-1, or a combination of inducer and either of the suppressors. The figure shows the amounts of JA-Ile and PI-IIc transcript (a, c) and the amounts of free SA and PR-1a transcript (b, d). Leaflets were infested with T. urticae Santpoort-2 (TuSP-2), T. evansi Viçosa-1 (TeV-1), or T. urticae DeLier-1 (TuDL-1), or simultaneously with TuSP-2 and either TeV-1 or TuDL-1 (both). Uninfested leaflets were used as controls. The numbers below the x-axis indicate the number of adult female mites used to infest the leaflets. The bars show the means (+ SEM), which are given as numbers within or above the bars. Transcript abundances were normalized to actin and scaled to the lowest mean per panel. Bars annotated with different letters (lowercase for JA-Ile and SA; uppercase for PI-IIc and PR-1a) were significantly different according to Fisher's least significant difference (LSD) test (P ≤ 0.05) after ANOVA.