The S. Rissen strain RW (serotype 6; antisera were from Staten Serum, Copenhagen, Denmark) was isolated from a food matrix and characterized by Istituto Zooprofilattico Sperimentale Del Mezzogiorno (Portici, Naples, Italy). The S. Rissen strain RR was derived - in the course of this research - from the RW strain following selection for resistance to phage ϕ1 as described in this study. RR cells can spontaneously lose the prophage and thus occur with (RRϕ1+) or without ϕ1 (RSϕ1-) (the superscript S indicates that loss of ϕ1 causes loss of phage resistance). All the bacterial strains were analyzed for cellulose production and LPS phenotype and stored at − 20 °C in LB (Sigma-Aldrich, Milan, Italy) supplemented with glycerol (10%; Sigma-Aldrich, Milan, Italy).
Isolation of the phage ϕ1
Phage ϕ1 excision was induced by incubating RW cells (2 × 108 CFU/5 mL) in LB broth containing 1 μg/mL mitomycin C (Sigma-Aldrich, Milan, Italy) for 1 h at 37 °C. Following centrifugation (5.7 × 103 g), the supernatant was stored at + 4 °C, and the pellet resuspended in 5 mL of LB broth and incubated again at 37 °C for 4 h and then centrifuged. The pellet was discarded, while the supernatants from the two centrifugations were pooled and filtered (filter pore size: 0.22 μm; MF-Millipore, Darmstadt, Germany) .
The titer of phage, expressed as plaque forming units (PFU), was evaluated by using the DLA technique as reported by Sambrook et al.  Phage ϕ1 was stored in SM buffer at − 20 °C. The aliquot in use was kept at + 4 °C.
Isolation of the phage ϕ1-resistant strain RR
RW bacteria in early exponential growth phase were mixed with warm soft agar. The mixture was poured on LB agar (Sigma-Aldrich, Milan, Italy) plates and allowed to solidify. Phage ϕ1 was then spotted (10 μL/spot) and the plates were incubated overnight at 37 °C. The following morning, the colonies grown inside the lysis plaques were picked with a sterile loop and streaked on LB agar plate. This procedure was repeated 3 times. Phage-resistant bacteria were further tested for phage ϕ1 resistance by the spot test. Plaque absence after overnight incubation confirmed that bacteria were phage ϕ1-resistant (RRϕ1+). Phage ϕ1-resistant colonies were detected after about 24 h of selection.
RSϕ1- bacteria (108 CFU in 500 μL LB) were incubated with ϕ1 isolated from RW bacteria (108 PFU/mL) for 72 h. The suspension was mixed with soft agar (4 mL) and then poured on a solid agar. Phage ϕ1was spotted on soft agar (10 μL/spot) and plates were incubated at + 37 °C and inspected daily for plaque formation.
Analysis of cellulose production
Cellulose production was detected by growing bacteria on LB agar supplemented with 200 μg/mL calcofluor (Sigma-Aldrich, Milan, Italy). Plates were incubated at 37 °C for 2–4 days. Colonies were visualized under a 366-nm light source . Congo red binding was detected by growing bacteria on LB agar supplemented with Congo red (40 μg/mL; Sigma-Aldrich, Milan, Italy).
Biofilm thickness determined by confocal laser scanning microscopy
Biofilms were formed on polystyrene Chamber Slides (n° 177,445; Thermo Scientific, Ottawa, Canada). For this purpose, overnight cultures of RRϕ1+, RSϕ1-, and RW strains grown in LB medium were diluted to a final concentration of 0.001 and seeded into a chamber slide at 37 °C for 36 h to assess biofilm thickness and cell viability. The biofilm cell viability was determined with the FilmTracer™ LIVE/DEAD® Biofilm Viability Kit (Molecular Probes, Invitrogen, Carlsbad, California, USA) following the manufacturer’s instructions. Microscopic observations and image acquisitions were performed as described .
Salmonella genome sequencing, assembly and annotation
The RW or RR strains were expanded in LB broth starting from a single colony. Genomic DNA was then extracted by the phenol-chloroform method, purified with Agencourt AMPure XP beads (Beckman Coulter; beads to DNA ratio 1.8:1), and quantified by the Qubit dsDNA BR Assay Kit (Thermo Fisher, MA, USA). DNA size and purity were measured by the 2200 Tape Station Instrument (Agilent Genomics) and Nanodrop (Thermo Fisher), respectively. Illumina libraries were obtained from 1 μg of genomic DNA, and sequenced with the NextSeq500 instrument using the 150 nt paired-end protocol (Illumina, San Diego, CA). Illumina reads were quality filtered, trimmed using Sickle and finally quality corrected with BayesHammer. Genomes were assembled de novo from Illumina reads using SPAdes 2.9.0 with multiple k-mer combinations: from 101 to 125 with 2 nt steps for the 202 RR genome, with 95, 97, 111, 113 for the RW genome and 101, 105, 109, 113, 117, 121,125 for the ϕ1 genome. The resultant contigs were scaffolded using SSPACE 3.0. Five μg high-molecular-weight genomic DNA (peak >60Kb) were used to prepare ~ 20 Kb-insert SMRT-bell libraries (Pacific Biosciences, CA, USA). The library templates were sequenced using the single molecule real time (SMRT) Sequencing technology on a PacBio RSII sequencer (Pacific Biosciences, Macrogen Inc., Korea). PacBio subreads were extracted using Bash5tools (version 0.8.0), filtered and assembled de-novo with Falcon-Integrate and the settings suggested for bacterial genome. The assembled genome sequence was polished by Quiver v 0.9.2. and gene annotation performed using RAST web service (http://rast.nmpdr.org/) . The ϕ1 insertion site was identified by mapping PacBio reads from RRϕ1+ bacteria against the phage genome assembly and soft-clipped bases were retrieved.
Variant SNP calling
SNP calling was carried out using MUMmer 3.23 tool . Single-nucleotide polymorphisms (SNPs) were identified by Show-SNPs, a script associated with MUMmer 3.23.
The output was filtered by BUFF > 50 with the Show- SNPs flags ClIrx 25 and the SNP position was assembly by quality > 80. The RW Salmonella genome was used as reference. Each assembly was queried with each SNP context from the MUMmer output using BLAST + , retaining only SNPs for which exactly one occurrence of either of the two genomes was found in all assemblies.
PAGE was performed using the system of Laemmli  with sodium deoxycholate (DOC; Thermo Scientific, Waltham, MA USA) as detergent as described . Glycosyl analysis was performed as reported .
Real time PCR
Total RNA was extracted from individual bacterial strains according to the Allprep Bacterial DNA/RNA/Protein Kit protocol (Quiagen) and then reverse-transcribed using the high-capacity cDNA reverse transcription kit (Applied Biosystem). Real-time PCR was carried out using the Step One Real-Time PCR Systems machine (Thermo-Fisher scientific). Reactions were carried out in a 20 μl of Master SYBR Green I mix (Roche Diagnostics Ltd., Lewes, UK). The amplification protocol included 10 min at 95 °C and 40 cycles, each consisting of 10 s at 95 °C for denaturation, 120 min at 57 °C for annealing, and 60 s at 60 °C for extension; the final step was at 4 °C. PCR reactions were carried out in triplicate. Expression values were normalized versus the RW strain. The reference gene was the housekeeping InvA. The relative gene expression was carried out using the Delta Delta ct Method .
Following the genome sequencing experiment, we designed the primers for the 5 kb region using primer 3 as primer design tool. We used the following overlapping eight primer pairs. Primers are designed to amplify regions within a size range of 400–600 bp.
The thermal shock of RW or RRϕ1+ cells was carried out by exposing the cells at − 20 °C for 1 h and + 40 °C for 2 h. The cells were then tested for loss of phage resistance. The phylogenetic tree was constructed using the maximum likelihood method ; for data alignment were used the Blosum 65 (gap open penalty = 11; gap extension penalty = 3), Jukes-Cantor, and UPGMA models. Biofilm production was measured by the crystal violet assay .