Impact of phage therapy on Pseudomonas syringae pv. syringae and plant microbiome dynamics through coevolution and field experiment - Field trial data set

This entry contains raw data tables, metadata and R scripts used in the analysis of field trial data.

 

Method excerpts:

A field trial with phages was conducted between 19^th June 2023 and 19^th July 2023 and then three months after in September, at a cherry orchard research plot in NIAB, East Malling (N 51° 17' 31.7’’, E 0° 26' 52.1’’). The trees used were 15-year old Prunus avium cultivar Sweetheart on Gisela 5 rootstock. The distance between trees was 1.5 m, with 3 m distance between each row. There were two untreated guard trees between every experimental tree.

A fully randomised block design with two factors (each with two levels), over five blocks with four plots per block (one tree per plot) was used. The two factors were ‘phages’ (cocktail 5C yes/no) and ‘Pss’ (Pss bacteria yes/no) for four total treatment groups: ‘Control’ consisting of PBS buffer; ‘cocktail 5C’ (10⁸ PFU ml^-1); ‘Pss’ (2 x 10⁸ CFU ml^-1); and a combination of ‘Pss+ cocktail 5C’, where Pss was applied first followed by cocktail 5C within 12 h.

Subsequent sampling was done in the morning of day 1-, 2-, 3-, 4- and 30-day post treatment (DPT) to quantify Pss and phage populations. Three (out of four treated) random leaves from two shoots per tree were collected and combined into a single sample. Similarly, two random shoot sections (0.5 - 0.8 cm thick and 4 cm long) were collected from each tree at each time point and processed as described below to assess both Pss and phage populations sizes and collect representative strains.

Using samples obtained from the field experiment, DNA extraction was performed on the remaining 5 ml of leaf washes, from all samples at 3 DPT and 30 DPT, for a total of 40 samples. 3 DPT was chosen as the changes in bacterial CFU and phage PFU were detectable from this time point onward. Leaf washes were centrifuged at 5000 g, for 20 min, at 4°C, resuspended in 500 µl of sterile PBS and stored in sterile 2 ml tubes at -20°C. All samples were centrifuged (16,000 g, 10 min) and the supernatant were carefully discarded. Each pellet (ca. 5 mg) was resuspended in 400 µl of lysis buffer and the DNA extracted according to the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany) manufacturer’s protocol, including the optional RNase A digestion step after lysis. Yields were analysed using a Nanodrop spectrophotometer (Thermo Scientific, Waltham, MA, US). The total bacterial (16S) and fungal (ITS) microbiome abundance was quantified using quantitative PCR (qPCR) using the same primer pairs as used in amplicon sequencing below (Papp-Rupar et al., 2022). The effect of phages on the total 16S and ITS microbiome sizes (as log 10 copy number) was analysed using the same statistical approach as in analysis of CFU and PFU data.

The samples were sent to Novogene UK (Cambridge, UK) for PCR, library prep and amplicon sequencing of fungal ITS1 amplicon using ITS1-1F (5’-CTTGGTCATTTAGAGGAAGTAA-3’ (Gardes & Bruns, 1993) and ITS2 primer (5’-GCTGCGTTCTTCATCGATGC-‘3 (White et al., 1990); and bacterial 16S V5-V7 amplicon using 799F (5’-AACMGGATTAGATACCCKG-3’ (Chelius & Triplett, 2001) and 1193R (5’-ACGTCATCCCCACCTTCC-3’ (Bodenhausen et al., 2013)). Samples were sequenced on an Illumina NovaSeq platform in paired end mode with read length of 250 nt.

Amplicon Sequence Variants (ASVs) were generated from a combined set of leaf and shoot samples (across both 3 DPT and 30 DPT) but analysed separately, using a previously published pipeline (Papp-Rupar et al., 2022). The following raw sequence reads were discarded in quality control step: reads with incorrect bases in the barcode or primer regions; and reads containing adapter contamination. Forward and reverse reads used in the ASV generation step were merged using the UPARSE pipeline V. 11.0 (Edgar, 2013) with stringent criteria: minimum read length of 250 nt, zero differences in read overlap region, maximum expected error threshold of 0.2 (16S) and 0.1 (ITS) per sequence (Edgar and Flyvbjerg, 2015) and minimal merged read length of 400 (16S) or 185 (ITS).  Reads were then dereplicated, chimeric sequences removed and sequences with less than eight replicates discarded before generation of denoised ASVs.  For frequency table generation, reads were merged using “differences in read overlap region” set to 100 to ensure effectively all reads were merged. These unfiltered merged reads were aligned to the ASV representative sequences at the level of 97% similarity to produce an ASV frequency table. Finally, the SINTAX algorithm (https://www.drive5.com/usearch/manual/sintax_algo.html) was used to assign taxonomic ranks to each ASV with the Unite V8.3 (2021-05-10) fungal database (Kõljalg et al., 2013) and “the RDP training set V18” database for the 16S rRNA gene (Cole et al., 2014). The SINTAX algorithm only resolves bacterial ASVs to the genus level, but may resolve fungal ASVs to the species level. Taxonomy assignment confidence was at the 80% level.

 

 

 

 

 

Bact and fungal ASV taxonomy:

zBAC.sintax

zBAC.sintax.taxa

zFUN.sintax

zFUN.sintax.taxa

 

R scripts used to analyse the Bact and fungal ASV data