Plants and insects
Rice (Oryza sativa) cultivar Minghui63 was used in this study. Rice plants were grown in a greenhouse at 27 ± 3 °C with 75 ± 10% RH (relative humidity) and a photoperiod of 16:8 h L:D (light:dark). The cultivation of rice plants followed the same procedure as described previously27. Plants were used for experiments when they were at the tillering stage, which occurred about 44–49 days after sowing.
C. suppressalis larvae were reared on an artificial diet as described70. Ten percent honey water solution was provided to supply nutrition for the adults. N. lugens were maintained on a BPH-susceptible rice variety Taichung Native 1 (TN1)38. T. japonicum were obtained from Keyun Industry Co., Ltd (Jiyuan, China). Newly emerged adult wasps were maintained in glass tubes (3.5 cm diameter, 20 cm height) and supplied with 10% honey water solution as a food source and were maintained for at least 6 h to ensure free mating, before females were used for the following experiments. All three species were maintained in climatic chambers at 27 ± 1 °C, 75 ± 5% RH, and a photoperiod of 16:8 h L:D.
Performance of caterpillars on insect-infested rice plants
Multiple types of rice plants were prepared: (i) uninfested plants, meaning that potted rice plants remained intact without insect infestation; (ii) SSB-infested plants, each potted rice plant was artificially infested with one 3rd instar SSB larva that had been starved for >3 h for 48 h; (iii) BPH-infested plants, each potted rice plant was artificially infested with a mix of fifteen 3rd and 4th instars BPH nymphs for 48 h; (iv) SSB/BPH-infested plants, each potted rice plant was simultaneously infested with one SSB larva and 15 BPH nymphs for 48 h; (v) SSB → BPH-infested plants, each potted rice plant was artificially infested with one SSB larvae alone for the first 24 h, then 15 BPH nymphs were additionally introduced for another 24 h; (vi) BPH → SSB-infested plants, namely each potted rice plant was artificially infested with 15 BPH nymphs for the first 24 h, then one SSB larvae were additionally introduced for another 24 h. Plant treatments were conducted as described in detail in our previous study27. During herbivory treatment, the uninfested plants were placed in a separate room to avoid possible volatile-mediated interference. During the subsequent bioassays, both SSB caterpillar and BPH nymphs remained in or on the rice plants.
Two bioassays were conducted to test the performance of C. suppressalis larvae feeding on differently treated rice plants. The first bioassay included the plant treatments i, ii, iii, and vi, and the second bioassay included the plant treatments i, ii, v, and vi. Three 2-day-old larvae of C. suppressalis were gently introduced onto the middle stem of each rice plant using a soft brush. The infested rice plants were then placed in climatic chambers at 27 ± 1 °C, 75 ± 5% relative humidity, and a photoperiod of 16:8 h L:D. The C. suppressalis larvae were retrieved from the rice plants after 7 days feeding, and they were weighed on a precision balance (CPA2250, Sartorius AG, Germany; readability = 0.01 mg). The mean weight of the three caterpillars on each plant was considered as one biological replicate. The experiment was repeated four times using different batches of plants and herbivores, resulting in a total of 30–46 biological replicates for each treatment.
Oviposition-preferences of C. suppressalis females choosing among differently infested rice plants
Greenhouse experiment
In the greenhouse, seven choice tests were conducted with C. suppressalis females including (i) SSB-infested plants versus uninfested plants; (ii) BPH-infested plants versus uninfested plants; (iii) SSB/BPH-infested plants versus uninfested plants; (iv) SSB-infested plants versus BPH-infested plants; (v) SSB-infested plants versus SSB/BPH-infested plants; (vi) BPH-infested plants versus SSB/BPH-infested plants; and (vii) the test in which C. suppressalis females were exposed to all four types of rice plants. The experiments were performed as described in detail by Jiao et al.30. In brief, four potted plants were positioned in the four corners of a cage (80 × 80 × 100 cm) made of 80-mesh nylon nets for each test. For paired comparisons, two potted plants belonging to the same treatment were placed in opposite corners of each age, and in the test with four types of rice plants, each type of plant was positioned in one of the four corners of each cage. Five pairs of freshly emerged moths (less than 1 day) were released in each cage, and a clean Petri dish (9 cm diameter) containing a cotton ball soaked with a 10% honey solution was placed in the center of the cage as food source. After 72 h, the number of individual eggs on each plant were determined. The experiment was conducted in a greenhouse at 27 ± 3 °C, 65 ± 10% RH, and a photoperiod of 16:8 h L:D. Each choice test was repeated with 9–11 times (replicates).
Field cage experiment
The oviposition preference of SSB females was further assessed in a field near Langfang City (39.58° N, 116.48° E), China. Four choice tests were conducted: (i) SSB-infested plants versus uninfested plants; (ii) BPH-infested plants versus uninfested plants; (iii) SSB/BPH-infested plants versus uninfested plants; and (iv) SSB/BPH-infested plants versus SSB-infested plants. The treated rice plants were prepared as described above and were transplanted into experimental plots (1.5 × 1.5 m). For each pairwise comparison, six plots of rice plants were covered with a screened cage (8 × 5 × 2.5 m) made of 80-mesh nylon net to prevent moths from entering or escaping. Each of the six plots contained nine rice plants of a particular treatment, with three plots per cage representing the same treatment. Plots were separated by a 1 m buffer and they were alternately distributed in a 3 × 2 grid arrangement in each cage (Supplementary Fig. 4). Approximately 50 mating pairs of newly emerged C. suppressalis adults (<24 h) were released into each cage. After 72 h, the number of individual eggs on each plant was determined. The total number of eggs of three plots in each cage was regarded as one replicate, 3–4 replicates were conducted for each pairwise comparisons.
Rice plant response to herbivore infestation
RNA-seq and data analysis
To explore the molecular mechanisms underlying the rice plant-mediated interaction between BPH and SSB, gene expression changes in rice response to infestation by SSB, BPH, or both were analyzed by RNA-seq. The rice plants, uninfested (control) or infested, were prepared as described above. After 48 h, the stems of the plants were harvested and frozen in liquid nitrogen. Samples from five individual plants of the same treatment were pooled together as one biological replicate, and three replicates were collected for each treatment.
RNA-seq analyses were performed as described previously71. In brief, total RNA was extracted using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and treated with RNase-free DNase I (NEB, Ipswich, MA, USA) to remove any genomic DNA. Library preparation and RNA-seq were performed by Novogene (Beijing, China) using an Illumina Hiseq 4000 system, resulting in ~45–55 million raw reads per sample. Raw reads were subjected to quality checking and trimming to remove adapters, poly-N sequences, and low-quality bases (Phred quality score Q < 20). The yield clean data of each sample were aligned to the rice reference genome IRGSP-1.0 (https://rapdb.dna.affrc.go.jp) using HISAT2 (v2.09)72, and the number of reads mapped to each gene was counted with featureCounts (v1.5.0-p3)73. The expression level of each gene was calculated as FPKM (fragments per kilobase of transcript per million per million fragments mapped) according to an established protocol74. Expression differentiation analyses were conducted with the DESeq2 R package (v. 1.18.0)75. Genes with absolute value of log2(fold change) > 0 and P-value < 0.05 were defined as DEGs. The enriched functions of DEGs in RNA-seq datasets were annotated with the Gene Ontology (GO) function using the clusterProfiler R package (v4.0.2)76, and GO terms with Benjamini–Hochberg false discovery rate (FDR) adjusted P-value (Padj) <0.05 were considered significantly enriched. The transcriptional signatures of hormonal responses of rice plant to herbivory relative to gene expression in Arabidopsis induced by diverse phytohormones was analyzed using Hormonometer program77. Since TPIs serve as indicators of induced resistance in rice plants, especially against chewing herbivores such as SSB44,45, the analyses focused on expression profiles of TPIs-related genes among the four plant treatments. The expression of nine selected TPIs genes were validated by quantitative real-time PCR (qRT-PCR) analyses as previously described78. qRT-PCR was conducted on a Bio-RadCFX96 Touch Real-time PCR Detection System instrument (Bio-Rad, Hercules, CA, USA) using TransStart® Top Green qPCR SuperMix (TransGen Biotech, Beijing, China). The rice ubiquitin 5 gene was used as the internal standard to normalize the variations in gene expression. The primers used are listed in Supplementary Table 1.
Quantification of endogenous jasmonic and salicylic acid
JA signaling is well established as the core pathway that regulates chemical defenses in rice plant against herbivores, including SSB and BPH45,47,50,79. SA is commonly reported to be involved in cross-talk with JA36,47,50, and we therefore focused on the analysis of these two major plant hormones. Our RNA-seq results suggested that additional infestation by BPH significantly suppressed the expressions of genes related to JA and SA signaling. Both types of genes are highly upregulated in response to SSB infestation50. To confirm this, we quantified the JA and SA levels in rice plants with two treatments: (i) rice plants that were infested with one third-instar SSB larva alone for 24 h; (ii) rice plants that were first infested with one third-instar SSB larva for 12 h and then also with 15 BPH female adults for another 12 h. Rice stems were harvested at three time points: 0 h (uninfested control plants), as well as 12 h and 24 h after infestation. For each treatment, stems from five individual plants were harvested and pooled together as one biological replicate, and three replicates were collected for each time point.
Endogenous measurements of JA and SA were performed by the plant hormone platform at the Institute of Genetics and Developmental Biology, Chinese Academy of Sciences as previously described80.
TFs regulation prediction
As the key regulators of transcription, TFs are important in mediating plant responses to herbivory. To gain more insight into the mechanistic links between hormonal regulation and gene expression changes, we further analyzed the potential regulatory interactions between TFs and the five JA signaling genes (Supplementary Data 4) that were found to be upregulated by SSB infestation but were downregulated by dual infestation. Gene IDs of TFs in rice were retrieved from platform PlantRegMap81 and were then compared to the downregulated DEGs between dual infestation and SSB-infestation treatments to identify potential regulated TFs. To identify upstream regulatory TFs, the 2000-bp sequences upstream of the start codon of the JA signaling genes were extracted from the Phytozome database (https://phytozome-next.jgi.doe.gov/). Then, these sequences were subjected to the PlantRegMap with a threshold P-value ≤ 1e-5 for TF binding site prediction. The motifs of TFs that have the potential to bind to the promoters of the JA signaling genes were screened, and a possible interaction was assigned if there were one or more binding sites of a TF on the promoter of a JA signaling gene. The predicted TFs were matched with the identified TFs that were downregulated in dual infestation as compared to SSB infestation, in order to obtain the potentially co-expressed TFs.
Quantification of TPIs
We further measured the accumulation of TPIs in rice plants subjected to insect infestation. These experiments were prompted by the RNA-seq results indicating that the upregulation of TPIs-related genes in response to SSB infestation is significantly suppressed after co-infestation with BPH. The same plant treatments were included as used for RNA-seq but with new batches of plants. Stem samples were collected at 48 h and 72 h of insect infestation. Intact rice plants that served as controls were also sampled at the same time points. Samples from five individual plants were pooled together as one biological replicate, and three replicates were collected for each treatment. All samples were immediately frozen in liquid nitrogen and stored at −80 °C until further analyses.
TPIs contents were determined using enzyme linked immunosorbent assay (ELISA) kits (J&L Biological, Shanghai, China). The stem samples were ground into a fine power in liquid nitrogen using a mortar and pestle, and each sample (0.1 g) was homogenized in 0.01 M Phosphate Buffered Saline (PBS) buffer (pH = 7.4) (Sigma–Aldrich, St. Louis, MO, USA) with a sample-PBS proportion of 1:9 (1 g plant sample/9 ml of PBS). Samples were centrifuged at 4000 × g for 15 min at 4 °C, and the supernatant was collected. The ELISA experiments were performed following the protocols provided with the kits. The optical density values were recorded at 450 nm using a microplate spectrophotometer (PowerWave XS2, BioTek, Winooski, VT, USA). The protein concentrations in plant samples were measured using a bicinchoninic acid (BCA) protein assay kit (Aidlab Biotechnologies Co., Ltd., Beijing, China) according to the manufacturer’s instructions. The amount of protease inhibitor was calculated based on a standard curve, and results were expressed as µg protease inhibitor per mg protein.
Performance of C. suppressalis larvae feeding on TPI mutant and overexpression lines
To confirm that TPIs is the main defense compound in rice plants that provide resistance against SSB caterpillars, we assessed the performance of SSB larvae on rice lines apip4-5 with APIP4 gene (Os01g0124200) deletion and APIP4-OX-16-2 plants with APIP4 gene overexpression, as well as on their corresponding wild-type (WT) Nipponbare (NPB) rice plants. These rice lines were generated by CRISPR/Cas9 technology (for details to see Zhang et al.82). They were used because our RNA-seq analyses suggested that the APIP4 gene is an important regulator of TPIs production (Fig. 4). TPIs contents were first measured in stems of each rice line. Rice stem samples from two individual plants were pooled together as one biological replicate, and five replicates were collected for each line. The measurement of TPIs in these rice plants were conducted using ELISA as described above. For the bioassay, 2-day-old SSB larvae were individually fed on plants of each rice line. Larval mass was determined after 5 days of feeding, using a precision balance (CPA2250, Sartorius AG, Germany; readability = 0.01 mg). For each treatment, 36–53 insects were tested. This experiment was conducted in a climatic chamber at 27 ± 1 °C, 75 ± 5% relative humidity, and a photoperiod of 16:8 h L:D.
Effect of insect-induced volatiles on the oviposition behavior of SSB moths
Collection and analysis of rice plant volatiles
Individual rice plants were either uninfested or infested with SSB larvae alone, BPH nymphs alone, or both species simultaneously for 48 h using the method described above. The emitted volatiles were trapped using a dynamic headspace collection system in a climate chamber at 27 ± 3 °C, 75 ± 10% RH, and then analyzed and identified as described27. To collect the volatiles, two plants were placed into a glass bottle (3142 ml) connected to an air flow. Air was purified through activated charcoal, molecular sieves (5 Å, beads, 8–12 mesh, Sigma–Aldrich), and silica gel Rubin (cobalt-free drying agent, Sigma–Aldrich) before entering the glass bottle. After 30 min, a volatile collection was started by pushing and pulling air out of the glass bottle at a rate of 400 ml min−1 through a glass tube (5 mm diameter, 8 cm height) filled with 30 mg Super Q traps (80/100 mesh, ANPEL Laboratory Technologies (Shanghai) lnc, China) for 3 h (21:00–24:00, the time period when SSB lay their eggs). Volatiles collected on the Super Q traps were extracted with 200 μl of methylene chloride, and 500 ng of nonyl acetate in 10 μl of methylene chloride was added to the samples as an internal standard. The extracts were stored at −30 °C until further analyses. For each treatment, collections were repeated 7–9 times.
A gas chromatograph-mass spectrometer (GCMS QP-2010SE, Shimadzu Ltd., Kyoto, Japan) was used to separate, quantify and identify the collected volatiles. A 1 μL volume of each sample was injected into an RTX-5 MS fused silica capillary column (30 m × 0.25 mm ID × 0.25 μm film thickness; Restek Co., Bellefonte, PA, USA). The inlet was operated in split-less injection mode, and the injector, was maintained at 250 °C. Helium was used as the carrier gas with a flow of 1.0 ml min−1 in constant flow mode. The gas chromatograph oven was initially set at a temperature of 40 °C for 2 min before being raised by 6 °C a minute until it reached 250 °C, at which it was kept for 2 min. The mass spectrometer was operated in scan mode with a mass range of 33–300 amu at 5.24 scans s−1 and the spectra were recorded in electron impact ionization (EI) at 70 eV. The ion source and mass quadrupole were set at 230 and 150 °C, respectively.
The chemicals were first identified by mass spectral matches to library spectra as well as by matching observed retention time with that of available authentic standards. If standards were unavailable, tentative identifications were made based on referenced mass spectra data available from the National Institute of Standards and Technology (NIST) library (Scientific Instrument Services, Inc, Ringoes, NJ, USA) or based on previous studies27,40. Relative quantification of compounds was based on its integrated area relative to the internal standard27,51.
Odor preferences of SSB females
The response of SSB females to volatiles released from differently treated rice plants were investigated to better understand the mechanism underlying the moth’s oviposition preferences. The total volatiles emitted from uninfested plants, SSB-infested plants, BPH-infested plants and SSB/BPH-infested plants were collected for this experiment. Plant treatments and volatiles collections were the same as described above but without the addition of the internal standard. The collected volatiles were diluted in paraffin oil (purity 99%; Sigma–Aldrich, St. Louis, MO, USA) at 1:4 (v/v) and were stored at −80 °C before use.
One milliliter of each of the four types of volatile solutions were separately pipetted on the center of a filter paper strip (4 × 21 cm), which were then hung from the four corners of a cage (45 × 45 × 45 cm) made of 80-mesh nylon net. Five pairs of freshly emerged SSB moths (<24 h) were released in each cage. After 72 h, the number of eggs deposited on the filter paper strips and the surface of the nylon nets near each paper strip were determined. This oviposition choice test was repeated eight times.
Response of the egg parasitoid T. japonicum to herbivore-infested rice plants
Multiple types of herbivore-infested rice plants were prepared: (i) uninfested plants (control); (ii) SSB-infested plants; (iii) BPH-infested plants; (iv) SSB/BPH-infested plants; (v) plants infested with SSB eggs (referred to as egg-infested plants); (vi) plant infested with SSB larvae and their eggs (referred to as SSB/egg-infested plants); (vii) plants infested with BPH nymphs and SSB eggs (BPH/egg-infested plants); and (viii) plants infested with both SSB larvae, BPH nymphs, and SSB eggs (referred to as SSB/BPH/egg-infested plants). To prepare these treatments, plants were first artificially infested with herbivores for 48 h as described above, then some of them were subjected to SSB eggs deposition. For that, two potted rice plants of the same type were placed in a cage (45 × 45 × 45 cm) made of 80-mesh nylon nets, then 30 pairs of freshly emerged moths (<24 h) were released in each cage to mate and lay eggs. After 24 h, the plants were removed from the cage and those that carried 200–250 eggs were used as odor sources. During the period of egg deposition and the subsequent olfactometer experiments with the parasitoid, all insects remained in or on the rice plants.
To test the behavioral responses of T. japonicum to differently treated rice plants, they were offered the following pairs of odor sources: (i) uninfested plants versus egg-infested plants; (ii) uninfested plants versus SSB-infested plants; (iii) uninfested plants versus BPH-infested plants; (iv) egg-infested plants versus SSB/egg-infested plants; (v) egg-infested plants versus BPH/eggs infested plants; (vi) SSB/egg-infested plants versus BPH/egg-infested plants; (vii) egg-infested plants versus SSB/BPH/egg-infested plants; (viii) SSB/egg-infested plants versus SSB/BPH/egg-infested plants; and (ix) BPH/egg-infested plants versus SSB/BPH/egg-infested plants.
Responses of T. japonicum females to these odor sources were investigated in a Y-tube olfactometer as described27. Newly emerged adult wasps were maintained in glass tubes (3.5 cm diameter, 20 cm height) for at least 6 h to ensure that they would mate, before females were used for the experiments. Two rice plants of the same treatment were enclosed in a glass bottle and used as one odor source, and each pair of odor sources was replaced after ten parasitic wasps were tested. For each treatment, a total of 64–88 female wasps were tested. The experiments were conducted between 10:00 and 16:00 on several consecutive days.
Parasitism rates of C. suppressalis eggs by T. japonicum wasps
In a cage experiment, we further tested if the differences in parasitoid attraction observed in the olfactometer for the differentially infested plants can result in differences in parasitism rates of SSB eggs under realistic conditions. The following herbivore-treated plants were prepared as described above: SSB eggs on uninfested, SSB-infested, BPH-infested and SSB/BPH-infested plants. The four types of plants were placed in the four corners of a cage (60 × 60 × 60 cm) made of 80-mesh nylon nets, respectively. Subsequently, 40 pairs of newly emerged wasps (<1 day old) were released into the cage. After 48 h, the rice leaves with SSB eggs were collected, and the total number of SSB eggs on each plant was counted, and their parasitization status was determined under a microscope two days later; the eggs turned black 3 days after being parasitized. The experiment was replicated 12 times. The experiment was performed in a greenhouse at 27 ± 3 °C and with 75 ± 10% RH and a photoperiod of 16:8 h L:D.
Statistical analyses
Statistical analyses were conducted using SPSS 22.0 (IBM SPSS, Somers, NY, USA), R (version 4.0.4, https://www.r-project.org), Microsoft Excel 2019, and SIMCA 14.1 software (Umetrics, Umeå, Sweden). All data were checked for normality and equality of variances prior to statistical analysis. Datasets that did not fit assumptions were square-root (sqrt) transformed to meet the requirements of equal variance and normality. Likelihood ratio test (LR test) applied to a generalized lineal mixed model (GLMM) for overdispersion and grouped design were conducted to compare the number of eggs laid by SSB females on rice plants (Poisson error structure with log link function) using R package lme4 (v1.1-27.1). We used cage ID as random factor to account for the nonindependence between observations within the same cage. Whenever the dispersion value was higher than 2 in the Poisson GLMM, overdispersion was accounted for by using an observation level factor as a random factor. Thus, cage ID and an observation level factor were the two random factors in the mixed models. LR test applied to a generalized lineal model (GLM) were conducted to compare the parasitism rates of SSB eggs by T. japonicum (binomial distribution error with logit link function). Two-way and one-way analysis of variance (ANOVA) followed by least significant difference (LSD) test were used to compare the body weight increases of the SSB larvae on different plant treatments. The contents of JA, SA, and TPIs in different samples were analyzed using one-way ANOVA followed by Tukey’s honest significant difference (HSD) test or two-sided Student’s t-test. Behavioral responses of T. japonicum in Y-tube assays were analyzed using binomial test with an expected response of 50% for either olfactometer arm; parasitoids that did not make a choice were excluded from the analysis. Differences in volatile emission and in gene expression were analyzed by partial least squares-discriminant analysis (PLS-DA)27,67 using SIMCA 14.1 software. The omics data were normalized by medians, log-transformed, and then auto-scaled (mean centered and divided by the standard deviation of each variable) using Metaboanalyst 4.0 software83 before they were subjected to PLS-DA. The significance of treatment differences in PLS-DA was assessed using a permutation analysis (999 repetitions) implemented in the MVA.test from the RVAideMemoire package84.
Source: Ecology - nature.com
