Distinct Signatures of Host Defense Suppression by Plant-Feeding Mites

Abstract

Tomato plants are attacked by diverse herbivorous arthropods, including by cell-content-feeding mites, such as the extreme generalist Tetranychus urticae and specialists like Tetranychus evansi and Aculops lycopersici. Mite feeding induces plant defense responses that reduce mite performance. However, T. evansi and A. lycopersici suppress plant defenses via poorly understood mechanisms and, consequently, maintain a high performance on tomato. On a shared host, T. urticae can be facilitated by either of the specialist mites, likely due to the suppression of plant defenses. To better understand defense suppression and indirect plant-mediated interactions between herbivorous mites, we used gene-expression microarrays to analyze the transcriptomic changes in tomato after attack by either a single mite species (T. urticae, T. evansi, A. lycopersici) or two species simultaneously (T. urticae plus T. evansi or T. urticae plus A. lycopersici). Additionally, we assessed mite-induced changes in defense-associated phytohormones using LC-MS/MS. Compared to non-infested controls, jasmonates (JAs) and salicylate (SA) accumulated to higher amounts upon all mite-infestation treatments, but the response was attenuated after single infestations with defense-suppressors. Strikingly, whereas 8 to 10% of tomato genes were differentially expressed upon single infestations with T. urticae or A. lycopersici, respectively, only 0.1% was altered in T. evansi-infested plants. Transcriptome analysis of dual-infested leaves revealed that A. lycopersici primarily suppressed T. urticae-induced JA defenses, while T. evansi dampened T. urticae-triggered host responses on a transcriptome-wide scale. The latter suggests that T. evansi not solely down-regulates plant gene expression, but rather directs it back towards housekeeping levels. Our results provide valuable new insights into the mechanisms underlying host defense suppression and the plant-mediated facilitation of competing herbivores.

Keywords: comparative transcriptomics; defense suppression; dual infestation; facilitation; herbivore; plant defense; plant-mediated interactions; tomato red spider mite (Tetranychus evansi); tomato russet mite (Aculops lycopersici); two-spotted spider mite (Tetranychus urticae).

Figure 1

Phytohormone concentrations in tomato (Solanum lycopersicum) leaflets after seven days of infestation with herbivorous mites. Tomato leaves were infested with either a single mite species (Tetranychus urticae (Tu), Tetranychus evansi (Te), or Aculops lycopersici (Al)) or two species simultaneously (T. urticae plus T. evansi (Tu+Te), or T. urticae plus A. lycopersici (Tu+Al)). Non-infested plants served as controls (C). The figure shows the average (+ SEM) amounts of: (A) 12-oxo-phytodienoic acid (OPDA); (B) jasmonic acid (JA); (C) jasmonic acid-isoleucine (JA-Ile), and; (D) salicylic acid (SA). Multivariate analysis of variance indicated that the factor “mite-infestation treatment” had a significant effect on the phytohormonal profile (F₂₀ = 8187; p < 0.001). Different letters above the bars indicate significant differences at a level of p ≤ 0.05, after applying a generalized linear model followed by Fisher’s Least Significant Difference test. Phytohormone concentrations are presented as nanogram per gram fresh leaf material (ng·g⁻¹·FW).

Figure 2

General overview of transcriptional responses in tomato (Solanum lycopersicum) leaves after seven days of infestation with herbivorous mites. Tomato leaves were infested with either a single mite species (Tetranychus urticae (Tu), Tetranychus evansi (Te), or Aculops lycopersici (Al)) or two species simultaneously (T. urticae plus T. evansi (Tu+Te), or T. urticae plus A. lycopersici (Tu+Al)). Non-infested plants served as controls (C) and were used as a common reference in the transcriptional comparisons. (A) The total numbers of differentially expressed genes (DEGs) across the five mite feeding regimes. Bars represent the number of up- or down-regulated tomato genes identified based on a Benjamini and Hochberg false discovery rate adjusted p ≤ 0.05. The black-lined sections within the bars indicate the number of up- or down-regulated tomato genes with an absolute fold change (FC) ≥ 1.5 (i.e., Log₂FC ≥ 0.585). (B) Principal component analysis plot of the tomato transcriptomic responses to the five mite feeding regimes. (C) Venn-diagrams showing the overlap of the tomato transcriptomic responses to the five mite feeding regimes for up-regulated and down-regulated genes. (D) Transcriptional patterns of the five clusters of tomato DEGs across the five mite feeding regimes. Colored diamonds in each plot represent the average (±SD) of the transcript levels per feeding regime. The Log₂FC cutoff value of 0.585 is depicted by dashed lines. The five clusters were identified using a k-means clustering approach.

Figure 3

Gene-expression heat map depicting the relative transcript levels of tomato (Solanum lycopersicum) genes that encode proteins with a (predicted) function in the jasmonic acid (JA) pathway and that were differentially expressed in leaves after seven days of infestation with herbivorous mites. Tomato leaves were infested with either a single mite species (Tetranychus urticae (Tu), Tetranychus evansi (Te), or Aculops lycopersici (Al)) or two species simultaneously (T. urticae plus T. evansi (Tu+Te), or T. urticae plus A. lycopersici (Tu+Al). Non-infested plants served as controls and were used as a common reference in the transcriptional comparisons. Presented genes were differentially expressed (Benjamini and Hochberg false discovery rate adjusted p ≤ 0.05; Log₂ fold change (FC) ≥ 0.585) in at least one of the mite-infestation treatments. The different (sub)sections of the pathway are specified on the left. Dark green squares in the leftmost column denote that transcription of the respective gene was found to be significantly induced in the tomato JA-biosynthesis mutant def-1, 24 h after exogenous application of JA (for details see [39]).

Figure 4

Gene-expression heat map depicting the relative transcript levels of tomato (Solanum lycopersicum) genes that encode proteins with a (predicted) function in the salicylic acid (SA) pathway and that were differentially expressed in leaves after seven days of infestation with herbivorous mites. Tomato leaves were infested with either a single mite species (Tetranychus urticae (Tu), Tetranychus evansi (Te), or Aculops lycopersici (Al)) or two species simultaneously (T. urticae plus T. evansi (Tu+Te), or T. urticae plus A. lycopersici (Tu+Al)). Non-infested plants served as controls and were used as a common reference in the transcriptional comparisons. Presented genes were differentially expressed (Benjamini and Hochberg false discovery rate adjusted p ≤ 0.05; Log₂ fold change (FC) ≥ 0.585) in at least one of the mite-infestation treatments. The different (sub)sections of the pathway are specified on the left. Dark green squares in the leftmost column denote that transcription of the respective gene was found to be significantly induced in the tomato jasmonic acid (JA)-biosynthesis mutant def-1, 24 h after exogenous application of JA (for details see [39]). The light green square denotes JA-inducibility of the respective gene according to [75].

Figure 5

Gene-expression heat map depicting the relative transcript levels of tomato (Solanum lycopersicum) genes that encode proteins with a (predicted) function in the phenylpropanoid pathway and that were differentially expressed in leaves after seven days of infestation with herbivorous mites. Tomato leaves were infested with either a single mite species (Tetranychus urticae (Tu), Tetranychus evansi (Te), or Aculops lycopersici (Al)) or two species simultaneously (T. urticae plus T. evansi (Tu+Te), or T. urticae plus A. lycopersici (Tu+Al)). Non-infested plants served as controls and were used as a common reference in the transcriptional comparisons. Presented genes were differentially expressed (Benjamini and Hochberg false discovery rate adjusted p ≤ 0.05; Log₂ fold change (FC) ≥ 0.585) in at least one of the mite-infestation treatments. The different (sub)sections of the pathway are specified on the left. Dark green squares in the leftmost column denote that transcription of the respective gene was found to be significantly induced in the tomato jasmonic acid (JA)-biosynthesis mutant def-1, 24 h after exogenous application of JA (for details see [39]).

Figure 6

Gene-expression heat map depicting the relative transcript levels of tomato (Solanum lycopersicum) genes that encode proteins with a (predicted) function in terpenoid biosynthesis and that were differentially expressed in leaves after seven days of infestation with herbivorous mites. Tomato leaves were infested with either a single mite species (Tetranychus urticae (Tu), Tetranychus evansi (Te), or Aculops lycopersici (Al)) or two species simultaneously (T. urticae plus T. evansi (Tu+Te), or T. urticae plus A. lycopersici (Tu+Al)). Non-infested plants served as controls and were used as a common reference in the transcriptional comparisons. Presented genes were differentially expressed (Benjamini and Hochberg false discovery rate adjusted p ≤ 0.05; Log₂ fold change (FC) ≥ 0.585) in at least one of the mite-infestation treatments. The different (sub)sections of the pathway are specified on the left. Dark green squares in the leftmost column denote that transcription of the respective gene was found to be significantly induced in the tomato jasmonic acid (JA)-biosynthesis mutant def-1, 24 h after exogenous application of JA (for details see [39]). The light green square denotes JA-inducibility of the respective gene according to [97].

Figure 7

Gene-expression heat map depicting the relative transcript levels of tomato (Solanum lycopersicum) genes that encode proteins with a (predicted) function in cholesterol and steroidal (glyco)alkaloid biosynthesis and that were differentially expressed in leaves after seven days of infestation with herbivorous mites. Tomato leaves were infested with either a single mite species (Tetranychus urticae (Tu), Tetranychus evansi (Te), or Aculops lycopersici (Al)) or two species simultaneously (T. urticae plus T. evansi (Tu+Te), or T. urticae plus A. lycopersici (Tu+Al)). Non-infested plants served as controls and were used as a common reference in the transcriptional comparisons. Presented genes were differentially expressed (Benjamini and Hochberg false discovery rate adjusted p ≤ 0.05; Log₂ fold change (FC) ≥ 0.585) in at least one of the mite-infestation treatments. The different (sub)sections of the pathway are specified on the left. The light green square in the leftmost column denotes JA-inducibility of the respective gene according to [114]. The asterisk indicates that the enzyme encoded by this GAME gene has not been characterized in tomato, i.e., only in potato [115].