Plants, mites and bacteria
Tomato seeds (Solanum lycopersicum cv. Castlemart (CM), defenseless-1 (def-1), and a transgenic line 35S::prosystemin (prosys+), both in the genetic background of CM), as well as S. lycopersicum cv. Moneymaker (MM) and the transgenic line nahG (in the genetic background of MM) were germinated in soil and grown in a greenhouse compartment at a temperature of 25°C and a 15/9 hour light/dark regime. One week prior to each experiment, plants were transferred to a climate room with day/night temperatures of 27ºC/25ºC, a 16/8 h light/dark regime and 60% relative humidity (RH). Def-1 plants are deficient in wound- and systemin-induced JA accumulation and in the expression of downstream defense genes [26]. prosys +plants overexpress the prosystemin gene, resulting in a constitutively activated JA-signaling pathway [27]. nahG plants are transformed with the bacterial gene nahG encoding salicylate hydroxylase which removes endogenous SA by converting it into catechol [34]. Both the prosys +and the nahG gene are under the control of the constitutively expressed CaMV 35S promoter. Experiments with WT and mutant/transgenic lines were always carried out in parallel.
Tomato russet mites (Aculops lycopersici, also referred to as russet mites or abbreviated as RM) (Acari: Eriophyidae) were obtained from Koppert Biological Systems (Berkel en Rodenrijs, the Netherlands), who, in turn, had obtained them in the summer of 2008 from naturally infested plants in a greenhouse in the Westland area (the Netherlands), and were since then reared in insect cages (BugDorm-44590DH, Bug Dorm Store, MegaView Science Co., Taichung, Taiwan) in a climate room on tomato plants (cv. CM) that were between three and five weeks old.
Two-spotted spider mites (T. urticae, also called spider mites or abbreviated as SM) (Acari: Tetranychidae) were originally obtained in 2001 from a single European spindle tree (Euonymus europea L.), in the dunes near Santpoort, the Netherlands (GPS coordinates: 52 26.503 N, 4 36.315 E). The strain we used has been described as a JA-inducing mite genotype and as susceptible to these defenses [10]. Since its collection from the field, the strain has been propagated on detached bean (Phaseolus vulgaris) (cv. 746 Speedy) leaves that were placed with the abaxial surface on wet cotton wool and maintained in a climate room (temperature of 25ºC, a 16/8 hour light/dark regime and 60% RH).
Pseudomonas syringae pv. tomato DC3000 bacteria were grown on King’s Broth (KB) medium [51] agar plates, containing rifampicin (50 μg/ml), and grown at 28°C for two to three days. Subsequently, single colonies were picked and bacterial cultures were grown overnight at 28°C in liquid KB medium with rifampicin (50 μg/ml). Bacterial cells were collected by centrifugation (3,000 rpm for 10 minutes), resuspended in 10 mm MgSO4 and adjusted to the required optical density (OD) before pressure infiltration into the leaflets.
Infestation and sampling of plants used for gene expression and phytohormone analyses
At the start of the experiments, 21-day-old tomato plants were infested with spider mites (SM), russet mites (RM) or with the two species together. RM infestations were done by transferring the mites on small pieces of leaflets (about 0.5ºCm2) to the leaflets of uninfested plants. These leaflet pieces had been cut from leaves picked from a well-infested tomato plant and each piece contained about 250 mobile stages of RM as determined with a stereomicroscope. Plants with spider mites received five spider mites per leaflet on each of three leaflets per plant. Thus, each plant received 15 spider mites in total. RM and SM were introduced at the same time. To prevent mites from dispersing we applied a thin barrier of lanolin (Sigma-Aldrich Chemie B.V., Zwijndrecht, The Netherlands) on the petioles of leaflets that were chosen for infestation. Uninfested control plants received lanolin, but no mites. In total, three leaflets per plant were infested which were pooled at the time of sampling. This sample was taken as one biological replicate. Sampled leaflets were flash-frozen in liquid nitrogen and stored at −80°C until total RNA or phytohormones were extracted. We always picked the same leaflets for infestation, that is, one leaflet of the second compound leaf (counted from the bottom to the top of the plant), one from the third compound leaf and the terminal leaflet of the fourth compound leaf. Sampling was performed seven days after infestation. Generally, starting with a density of about 250 RM per leaflet infection symptoms became visible five to six days after infestation. After seven days of infestation (the time-point of sampling) symptoms of RM-infestation were clearly visible, but leaflets were not necrotic or senesced. The relative transcript levels presented in Figures 3 and 4 represent the mean of 11 to 13 biological replicates, obtained from three independent experiments. The phytohormone levels presented in Figure 2 represent the mean of nine to ten biological replicates, obtained from two independent experiments.
Sampling for mites in tomato fields
Tomatoes growing in the field were sampled to determine the (co-)occurrence of plant-eating mites. Samplings were performed in several areas of Italy that are well-known for their tomato production (see Additional file 1: Figure S1). Samplings were performed at 93 different sites in total, in the summers of 1997 and 1998. On each site at least two samples were taken, the first one in July and the second one in September.
The sampling method we employed is commonly used for estimating mite densities in a wide variety of horticultural crops in agricultural fields [52]. Briefly, one leaf (that is, a compound leaf) was picked per plant by walking along the tomato rows. Plants were vertically subdivided in three parts: basal, mid and apical. One-third of the total number of leaves collected at a particular site consisted of leaves from the basal part, one-third from the central part and one-third from the apical part of the plants. Leaves were collected from plants from at least 15 to 18 different rows in each field and plants from which leaves were collected were standing at least one meter (2 to 3 plants) from each other. In total, 75 to 100 leaves were collected for each sample. Subsequently, these leaves were examined under a stereomicroscope and on each leaflet the presence/absence of spider mites as well as that of russet mites was recorded. Numbers of spider mites and russet mites present were counted.
Quantification of gene expression by qRT-PCR
Leaflets were cut at the base and three leaflets per plant were pooled in 50-ml tubes, flash frozen in liquid nitrogen, and stored at -80ºC. Leaflets were ground in liquid nitrogen, and total RNA was extracted using a phenol-LiCl-based method as described [53]. The integrity of RNA was checked on 1% agarose gels and subsequently quantified using a NanoDrop 100 spectrophotometer (Fisher Scientific, Loughborough, UK). DNA was removed with DNAse (Ambion, Huntingdon, UK) according to the manufacturer’s instructions, after which a control PCR was carried out to confirm the absence of genomic contaminations. cDNA was synthesized from 2 μg total RNA using a poly(dT) primer and M-MuLV Reverse Transcriptase (Fermentas, St. Leon-Rot, Germany) according to the manufacturer’s instructions. cDNA dilutions (10x) were used as the template in quantitative reverse-transcriptase PCR (qRT-PCR). Reactions were carried out in a total volume of 20 μl containing 0.25 μm of each primer, 0.1 μl ROX reference dye and 1 μl of cDNA template. Two technical replicates were performed per measurement. qRT-PCR was performed with Platinum SYBR Green qPCRSuperMix (Invitrogen, Paisley, UK) using an ABI 7500 (Applied Biosystems, Foster City, CA, USA) system. The program was set to 2 minutes at 50ºC, 10 minutes at 95ºC, 45ºCycles of 15 seconds at 95°C and 1 minute at 60ºC, followed by a melting curve analysis. Target gene expression levels were normalized to those of actin. The normalized expression (NE) data were calculated by the ΔCt method NE = (1/(PEtargetCt_target))/(1/(PEreferenceCt_reference) (PE = primer efficiency; Ct = cycle threshold). The PEs were determined by fitting a linear regression line on the Ct-values of a standard cDNA dilution series. Specific amplification was ensured by melting curve analyses and generated amplicons were sequenced. For presenting the qRT-PCR data in the graphs we projected the data on a relative scale by dividing all values by the lowest average value (such that the lowest average is always 1). The primers we used are listed in Additional file 8: Table S2.
Quantification of phytohormones by means of liquid chromatography-mass spectrometry
Phytohormones were extracted by homogenizing frozen leaf material (approximately 250 mg) in screw cap tubes containing 1 ml of ethyl acetate spiked with 100 ng of D6-SA and D5-JA (C/D/N Isotopes Inc., Pointe-Claire, Quebec, Canada) as internal standards. Samples were ground twice, using a GenoGrinder (Precellys24 Tissue Homogenizer, Bertin Technologies, Aix-en-Provence, France), at 6,500 rpm for 45 seconds and centrifuged at 13,000 rpm (8 g) for 20 minutes at 4ºC. Supernatants from two extraction steps were pooled and evaporated until dryness in a vacuum concentrator (CentriVap Centrifugal Concentrator, Labconco, Kansas City, MO, USA) at 30ºC. The dried residue was dissolved in 250 μl 70% methanol, vortexed and centrifuged and the supernatant was transferred to liquid chromatography-mass spectrometry (LC-MS) vials (Fisher Scientific, Hampton, NH, USA). Phytohormone measurements were conducted on a liquid chromatography tandem mass spectrometry system (Varian 320 Triple Quad LC/MS/MS, Agilent Technologies, Santa Clara, CA, USA). Twenty microliters of each sample was injected onto a Pursuit 5 column (C18; 50x2.0 mm). The mobile phase comprised solvent A (0.05% formic acid in water; Sigma-Aldrich, Zwijndrecht, the Netherlands) and solvent B (0.05% formic acid in methanol; Sigma-Aldrich). The program was set as follows: 95% solvent A for 1 minute 30 seconds (flow rate 0.4 ml/minute), followed by 6 minutes in which solvent B increased till 98% (0.2 ml/min) which continued for 2 minutes 30 seconds at the same flow rate, followed by 1 minute 30 seconds at an increased flow rate (0.4 ml/min), subsequently returning to 95% solvent A for 1 minute until the end of the run. Compounds were detected in the electrospray ionization-negative mode. Molecular ions [M-H]- at m/z 137 and 209 and 141 and 213 generated from endogenous SA and JA and their internal standards, respectively, were fragmented under 12 V collision energy. The ratios of ion intensities of their respective daughter ions, m/z 93 and 97 and m/z 59 and 61, were used to quantify endogenous SA and JA, respectively. A standard dilution series of pure compounds of JA-Ile (OlChemIm Ltd, Olomouc, Czech Republic), JA and SA (DUCHEFA Biochemie B.V., Haarlem, the Netherlands) was used to estimate the phytohormone concentrations and the retention time. The amounts were corrected for losses occurring during the extraction with a recovery rate, using the JA and SA internal standards.
Spider mite reproductive performance assays
To obtain RM-infested leaflets before the start of the SM-performance experiments, 21-day-old tomato plants were infested with RMs as described earlier.
Subsequently, seven days after plants had been infested with RM, four young adult female SM were placed on the adaxial surface of the RM-infested leaflets and on leaflets of the same age and position of non-infested control plants, using a soft bristle paintbrush. After four days (that is, eleven days after the start of the experiment), infested leaflets were detached and SM adults and their eggs were counted using a stereomicroscope. The results presented in Figure 1B represent the mean number of eggs per mite per day of 24 leaflets that were obtained from 8 plants.
In order to obtain SM of the same age, egg waves were generated in a climate room (temperature of 25ºC, a 16/8 hours light/dark regime and 60% RH), as previously described [33]. In short, 20 to 30 random adult female spider mites were selected from a rearing colony and allowed to produce eggs for a period of 48 hours on detached bean leaflets on wet cotton wool. After this period, the adults were removed but the eggs were maintained. After 14 days, the 2 ± 2 days young adult females were collected and these were used for the oviposition assays.
Pst DC3000 growth assays
At the start of the assays to assess Pst growth (Figure 6), 21-day-old plants were infested with RM as described earlier. Experiments were performed on WT (cv. MM) and SA-deficient nahG plants. After seven days of infestation with RM, the left halves (bordering the midrib) of RM-infested leaflets and similar leaflets from non-infested control plants were pressure infiltrated with Pst DC3000 (OD600 = 0.001) in 10 mm MgSO4 using blunt 1 ml syringes (BD Plastipak, Franklin Lakes, NJ, USA). Leaflets from the third compound leaf (counted from the bottom of the plant) were infected. In total, seven plants were infected per treatment and the experiment was repeated a second time with a similar result (see Additional file 9: Figure S8).
Three days after infection with Pst, leaflets were detached and two 1-cm2 circular leaf discs were punched out from the infiltrated leaflet halves and ground in 500 μl 10 mm MgSO4. Serial dilutions were prepared by taking 20 μl of the leaf disc solution and diluting it in 180 μl 10 mm MgSO4. Twenty μl of each serial dilution was plated on KB medium + rifampicin (50 μg/ml) plates. The number of colony forming units (CFU) was counted two days after incubation at room temperature.
Russet mite population growth experiments
For the RM population growth experiments, 21-day-old tomato plants were infested by transferring 20 RM to each of three leaflets per plant. Thus, each plant was infested with 60 RM in total. To prevent mites from dispersing we applied a thin barrier of lanolin on the petioles of leaflets that were chosen for infestation. Uninfested control plants received lanolin, but no mites. Leaflets with a similar position as those used for the gene expression and phytohormone measurements were used for infection. To assess RM performance, infested leaflets were detached and mites (all stages) were washed off by rinsing the leaflets one by one for 20 seconds in 25 ml 100% ethanol. Infested leaflets that came from the same plant were washed in the same solution. RM were counted by running 2 ml of the leaf washes through a particle counting system (PAMAS SVSS, PAMAS, Rutesheim, Germany). Leaf washes were counted 20 seconds after mixing them, to avoid having air bubbles enter the system. Adult RM are around 120 to 150 μm in size while their eggs are around 20 μm. Therefore, the number of particles measured in the range of 50 to 200 μm was used to quantify the number of adult mites. The number of mites per plant was calculated by multiplying the mean number of particles per ml with the total volume of 25 ml. This counting method was validated by means of a dose–response experiment. In this experiment, we infested leaflets on intact plants with 2, 4, 8, 16, 32 and 64 mites per leaflet on each of three leaflets per plant and after 14 days the mites were washed off, as described earlier. The numbers counted implied exponential population growth, corresponding to the starting conditions (see Additional file 10: Figure S9).
For the RM-SM co-infestation experiment (Figure 5), plants were infested by transferring 20 RM to each of the three leaflets per plant. After seven days of RM infestation, half of the plants were subsequently infested with five adult SM on the RM-infested leaflets. Thus, dual-infested plants were infested in total with 15 SM per plant. The population growth of RM was assessed after fourteen days of infestation with RM (and hence after seven days of infestation with SM) by counting the number of RM as described above. The values presented in Figure 5 represent the mean of 13 to 15 plants, obtained in two independent experiments.
For the RM-Pst co-infestation experiment (see Additional file 5: Figure S5), plants were infested by transferring 20 RM to each of three leaflets per plant. After seven days of RM infestation, three leaflets of half of the RM-infested plants were infiltrated with Pst in 10 mm MgSO4 and three leaflets of the other half of the RM-infested plants were infiltrated with mock 10 mm MgSO4. The same three leaflets were chosen for infiltration from each plant. Since older leaflets are more susceptible to Pst compared to the younger leaflets (JJ Glas, personal observation), the leaflets from the third and fourth compound leaf were infiltrated with a bacterial suspension which had an OD (OD600) of 0.0001 while the oldest leaflet (that is, the leaflet on the second fully expanded leaf) was infiltrated with a lower OD600 of 0.00005. On each leaflet, approximately ¼ of the leaf area was infiltrated with bacteria. The population growth of RM was assessed after 14 days of infestation with RM (and hence after seven days of infestation with Pst) by counting the number of RM as described above. The values presented in Additional file 5: Figure S5 represent the mean of 15 plants, obtained from two independent experiments. Symptoms of Pst infection were visible at the time of sampling, with sometimes (minor) parts of the leaflet being senesced and/or necrotic (see Additional file 5: Figure S5).
For assessing RM performance on WT and def-1 plants (see Additional file 7: Figure S7), plants were infested by transferring 20 RM to each of the three leaflets per plant. RM population growth was assessed after 8, 12 and 16 days by counting the number of RM as described above. The values presented in Additional file 7: Figure S7 represent the means of five to ten plants per time-point, obtained from two independent experiments.
Statistical analyses
Gene expression data were statistically analyzed using a nested analysis of variance (ANOVA). NE values were compared among treatments using `Treatment’ (with the levels `Control’, `SM’, `RM’ or `Both mites’) as fixed factor and with the factors `Experimental replicate’, `Biological replicate’ and `Technical replicate’ included as random factors in the model. The factor `Technical replicate’ (with levels 1 and 2) was nested to the factor `Biological replicate’. The means of each group were compared using Fisher’s LSD (least significant difference) post hoc test.
Spider mite oviposition data (Figure 1B; Additional file 3: Figure S3) were analyzed with ANOVA using `Treatment’ as fixed factor and including `Plant replicate’ as random factor. Means of each group were compared using Fisher’s LSD post hoc test.
Phytohormone data (Figure 2) were log-transformed and analyzed using ANOVA, with `Treatment’ as fixed factor and `Experimental replicate’ included as random factor in the model. Means of each group were compared using Fisher’s LSD post hoc test.
Pst performance data (Figure 6; Additional file 9: Figure S8) were analyzed with the Student’s t-test. Russet mite population growth data (Figure 5, Additional file 5: Figure S5 and Additional file 7: Figure S7) were log-transformed before statistical analysis. Mite density data (Additional file 2) were log-transformed and analyzed with the Student’s t-test for both sampling points independently. Values presented in the graphs represent untransformed data.
Two-sided Student’s t-tests were performed in Excel (Microsoft Corporation, Redmond, WA, USA) and ANOVA followed by Fisher’s LSD tests in SPSS 20.0 (SPSS Inc., Chicago, IL, USA).
Availability of supporting data
Raw data can be accessed via: http://dx.doi.org/10.6084/m9.figshare.1232104.