MATERIALS AND METHODS
Soil substrate and plant material conditions
In the first experiment, five different plant species were sown in a soil substrate. The soil substrate was composed of ¼ cover crop soil which was collected at the Ag Research Development & Education Center (ARDEC) of Colorado State University located at 4616 NE Frontage Road, Fort Collins, CO 80524 (DMS 40° 38’ 59.172” N and 104° 59’ 44.34” W), where only experiments without agrochemical inputs are performed. Additionally, ¼ Thermorock® non-sterile horticultural vermiculite (#2 grosses 3.5 cubs. ft.), ¼ Promix Bx® peat moss, and ¼ QUIKRETE® Play Sand (Atlanta, GA) non-sterile sand were combined into the final substrate.
Super sweet hybrid corn (Zea mays ‘Sh2’), bean (Phaseolus vulgaris L. ‘Seychelles’), tomato (Solanum lycopersicum‘Rutgers’), and red beet (Beta vulgaris L . ‘Burpee bred’) seeds were purchased from W. Atlee Burpee & Co. (Warminster, PA).Arabidopsis thaliana ecotype Col-0 seeds were obtained from Lehle (Round Rock, TX).
Seeds were sterilized in sodium hypochlorite solution and Milli-Q® water at 2% (V / V), subsequently rinsed four times with sterile distilled water, and planted directly into the soil substrate contained in 1L plastic pots.
Two seeds were placed in each pot to ensure the germination of at least one seedling per pot and incubated in growth chambers with a photoperiod of 16 h light / 8 h night at 25° C +/- 1° C for seven days. After seven days, the additional seedlings that germinated were thinned to leave only one plant per pot for all species. We used a completely randomized design with six replicates per treatment, maintaining one plant per pot. The seedlings of plants grew for four weeks in growth chambers. The plants were watered every two days with 50 mL sterilized water per pot.
Experimental design: herbivory experiment using Trichoplusia ni(Hübner)
In the fifth week of plant growth, each one of the different plant species was divided into two experimental groups: (1) Test Group: Plants with the herbivorous attack (n = 6); (2) Control Group: Plants without an herbivorous attack (n = 6). A population of Trichoplusia nilarvae in the third instar stage was obtained from Frontier Agricultural Sciences (Newark, DE). Four T. ni larvae were placed by each plant in this study.
Herbivory was introduced for three days, and then removed for days four and five. Then, new Trichoplusia ni larvae were introduced again for days six and seven. The insect removal allowed us to establish the effect of herbivory as well as keep enough biomass available for further testing after seven experimental days.
The rhizosphere samples were collected from the immediate soil adjacent to the roots, 2 mm from the root surface, after plant herbivory for seven days. Substrate soil where the plants grew up was kept in a plant growth chamber (photoperiod of 16 h light / 8 h night at 25° C +/- 1° C) for 24 hours, and re-used for re-sowing plants in the following experiment (see next section). The fresh and dry biomass, root height, and root biomass information were collected from all repetitions and experimental groups. One-way ANOVA was used to perform biomass data analyses, and the means between control and insect group, for each one of the plants studied, which presented differences compared by the T-Student test (Bonferroni) at 5% probability level.
Growth conditions of the second plant generation in re-used soil substrate
New seeds of sweet corn, Arabidopsis , beans, tomato, and red beet from the same batch of seeds previously used were sown in two types of substrate: 1) soil substrate from control plants and 2) soil substrate from the plants subjected to insect attack in the previous experiment. The seeds were sterilized and planted according to the protocol previously described. The plants were kept in a growth chamber in the same conditions as the first sowing. No additional fertilizer was applied. After four weeks, the plants were collected, and the dry biomass, fresh biomass, root height, and root biomass for all replicates were measured. The dry biomass data from the second experiment were used to perform data analyses. The means between control and insect group, for each one of the plants studied in the second generation, were compared by the T-Student test (Bonferroni) at 5% probability level. The free software Sisvar 9.5 (Ferreira, 2019, http://www.dex.ufla.br/~danielff/programas/sisvar.html) was used for the statistical analysis, and OriginPro 8.5 software (OriginLab Corporation, MA, https://www.originlab.com/) was used for graphics visualization. The graphics represent the relative values data for each one of the plants studied.
DNA rhizosphere extraction and Illumina Miseq Sequencing
DNA of six replicates was extracted for each treatment (for five different plants) from the conditioned soil (soils from the first generation of plants after insect attack). Each sample was composed of 50 mg of rhizosphere soil. The total DNA was extracted using theMoBio kit (PowerPlant ® DNA Isolation Kit) according to the manufacturer’s instructions. The DNA concentration was measured by a Nanodrop® spectrophotometer (Thermo Fisher Scientific ). The extracted DNA was stored at -20° C until the samples were subjected to amplification and Illumina Miseq .
The sequencing was performed according to Illumina protocols. The 16S rRNA gene hypervariable region V3-V4 was targeted to estimate bacterial communities present in the rhizosphere soil samples.
The sequencing was performed by a two-step PCR: First qPCR was conducted to quantify the 16S rRNA V3-V4 region with Illumina adapter sequences against a standard curve to quantify bacterial biomass in rRNA copies per gram soil. The next round of PCR included the addition of Illumina barcode sequences for multiplexing of 94 samples, including positive (Zymo mock community) and negative (water) control.
In the first qPCR run the set primers 341F/785R were used with Illumina Miseq adapter sequences (Klindworth et al., 2012). The primer 341F is correspondent to sequence F5’-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGAGGCAGCAG-3’and the primer 785R is correspondent to sequence R5’-GTCTCGTGGGCTCGGAGATGTGTATAAGA
GACAGGACTACHVGGGTATCTAATCC-3’.
The quantitative polymerase chain reaction (qPCR) consisted of 20 µl reaction volumes containing 2 µl of template DNA (5ng/ µl) and 18 µl of the master mix. The master mix consisted of 10 µL of 2X Maxima SYBR Green (Thermo Scientific, Waltham, MA, USA), and 1 µL each (10 µM) of forward and reverse primers, with the addition of 6 µL of molecular grade water. The PCR cycling conditions were: 95º C for 5 minutes, 25 cycles of 95º C for 40 seconds, 55º C for 30 seconds, 72º C for 60 seconds, and final annealing at 72º C for 5 minutes. The amplicons from the last PCR cycling were purified using an in-house preparation of solid-phase reversible immobilization (SPRI) magnetic beads based on the methodology of Glenn (2011) with modifications and original protocol of Rohland and Reich (2012). For the standard curve, purifiedPseudomonas putida KT2440 was used in the same cycle run with the amplicons’ samples to quantify the starting rRNA copies g-1 soil fresh weight. A positive control of Zymo (ZymoBIOMICS, Irvine, CA) mock community DNA, and a negative control (water), were included in duplicate and carried through the rest of the protocol.
The Illumina Nextera XT index sequences were attached for each sample by a new qPCR run. The qPCR conditions were: 5 µL of first-round PCR product, 25 µL of 2X Maxima SYBR Green (Thermo Scientific, Waltham, MA, USA), 10 µL water, and 5 µL each of forward and reverse indices, all combined for a total of 50µL and amplified at 95º C for 3 minutes, 8 cycles of 95º C for 30 seconds, 55º C for 30 seconds and 72º C for 30 seconds, followed by final annealing of 72º C for 5 minutes. The PCR product was bead-cleaned using SPRI beads and quantified using Qubit fluorometer (Thermo Scientific, Waltham, MA, USA) before normalization and pooling. The equality pool was run on a TapeStation system (Agilent Technologies, Santa Clara, CA, USA) to determine amplicon size and purity, and Kapa Biosystems (Sigma-Aldrich, St Louis, MO, USA). The last qPCR run was performed according to the manufacturers’ instructions to determine the concentration, and MiSeq libraries were quantified. PhiX (15%) was used as an internal library control. The library was then subjected to 250 base pair paired-end multiplex sequencing (MiSeq V3-V4 reagent kit) on an Illumina MiSeq at CSU’s Next Generation Sequencing Laboratory (Fort Collins, CO).
16S rRNA V3-V4 sequence analysis
De-multiplexed raw fastq files were processed with the R Bioconductor package (Callahan et al., 2016), and primers were removed from each sequence using the open-source Python program Cutadapt (Martin, 2011). The amplicon sequence variants (ASV) were inferred with the default pipeline in DADA2. Each ASV identified in DADA2 was classified to the closest reference sequence contained in the GreenGenes reference database (version 13_5_99) using the usearch_global option (minimum identify of 97%) contained in the open source program VSEARCH (Rognes et al., 2016). The analyses were conducted using the open source bioinformatic software myPhyloDB (https://nrrc.ars.usda.gov/myPhyloDB/home/), version 1.2.0 (Manter et al., 2016). A total of 55,484 high quality reads per sample were obtained. Samples with fewer than 3,000 reads were removed from the analysis. Four samples out of the 94 were removed. The average read length of bacteria for the 16S rRNA subunit was 601 bp.
Each taxonomic profile was used to determine bacterial phyla-specific abundances. The Shannon diversity index was used to assess initial bacterial alpha diversity. For beta diversity, microbial community composition was analyzed using principal coordinates analysis (PCoA) based on the Bray-Curtis dissimilarity index. A complementary non-parametric multivariate statistical test, including permutational analysis of variance (PERMANOVA) and non-parametric univariate ANCOVA analyses, was used to test the differences in microbial communities with the Bray-Curtis distance and 999 permutations with myPhyloDB (Manter et al., 2016).
An ASV (Amplicon Sequence Variant) table was used as output from myPhyloDB at the family and genus level presented between each experimental group. The shared and unique ASVs among treatments were counted, and their distributions are shown by a Venn diagram from the packet ‘jvenn’, a plugin for the ‘jQuery’ (http://jvenn.toulouse.inra.fr/app/index.html) Javascript library (Bardou et al., 2014).
RESULTS
1. Insect herbivory preference experiment
Herbivory significantly reduced the biomass of all plants after feeding for a period of seven days (p <0.05) (Figure 1). Sweet corn and beans were consumed by the pest insects at rates of 50% and 46%, respectively (Figure 1). A. thaliana biomass was also significantly consumed (p <0.05) at a rate of 57%. In tomato and red beet, T. ni larvae consumption was around 79% and 86%, respectively. The T. ni insect demonstrated the lowest preference for beans and the highest preference for tomato and red beet, the most consumed plants in the experiment.