Preparation of biofilm and extraction of RNA

In this study, S. Typhi strain ATCC 7251 was obtained from the Bio-bank, Institute for Research in Molecular Medicine, Universiti Sains Malaysia, Malaysia. Standard serological and biochemical testing were done to confirm the S. Typhi organism before subsequent preparations for transcriptome analysis.

S. Typhi planktonic cells were grown in 1% nutrient broth (HiMedia, India), prepared as per the manufacturer’s protocol, and incubated in a shaking incubator at 37 °C until an OD₆₀₀ value of 0.6 was recorded on the UV spectrometer (Spectronic BioMate 3, ThermoFisher Scientific, USA). S. Typhi biofilm tissue was grown in 50 ml Falcon tubes (SPL Life Sciences, Korea) using an optimized method described by Ganjali Dashti et al. [9]. Briefly, to simulate the gallbladder environment, a biofilm media consisting of a combination of 1.22% ox-bile (HiMedia, India), 0.5 M sodium chloride (Bio Basic Inc., Canada), 0.001 M potassium chloride (Merck, USA), 0.56 M glucose (Bio Basic Inc., Canada) and 1% nutrient broth was used to culture the S. Typhi cells in 50 ml Falcon tubes. In the tube, 1 ml of planktonic cells was added to 24 ml biofilm media. The tubes were placed into an incubator shaker for 24 h at 37 °C with a rotation speed of 250 rpm. The presence of bile and application of strong agitation caused the S. Typhi cells to form biofilm tissues to protect themselves from the stressful environment. The sticky layer that formed and attached to the surface of the polypropylene tubes was considered as the mature biofilm and was collected for RNA extraction.

RNA extraction was conducted using TRIzol® (Life Technologies, USA). For RNA extraction of planktonic cells, 15 ml of S. Typhi culture grown in nutrient broth was collected and pelleted down using a centrifuge at 12,000 xg for 15 min at 4 °C. The supernatant was removed and the pellet was washed once with 1× phosphate-buffered saline to remove residual nutrient broth. TRIzol® was added to the pellet, the tissue was then homogenized, followed by nucleic acid extraction according to the manufacturer’s protocol. For biofilm tissues, a Bead beater (Biospec Products, USA) was used to disrupt the polysaccharide matrix of the biofilm tissue. A total of 0.1 g of biofilm tissue was collected in a 2 ml centrifuge tube pre-filled with 500 mg zirconium beads (Benchmark, USA). The mixture was then homogenized at a revolution rate of 4800 rpm for 10 s with 30-s intervals cooling in ice for a total of 30 s of homogenization time. For each sample, three replicates were carried out to ensure complete homogenization of the biofilm tissue in TRIzol®. The quality of extracted RNA for both planktonic and biofilm cells was checked using an Agilent Bioanalyzer 2100 (Agilent Technologies, USA), and the concentrations were determined using a Nanodrop 2000c spectrophotometer system (Thermofisher, USA). Prior to cDNA conversion, the extracted RNA from planktonic and biofilm S. Typhi cells were treated with DNase I to remove genomic DNA contamination.

RNA-sequencing

RNA-sequencing was conducted using an Illumina HiSeq 2500 machine. The extracted RNA samples from both planktonic and biofilm S. Typhi cells were subjected to the Illumina sequencing workflow by the service provider [10]. The workflow included ribosomal RNA depletion, cDNA conversion, library preparation and cluster generation followed by sequencing. Efficient library generation was then assessed using a Bioanalyzer platform (Agilent), and an Illumina MiSeq-QC run was performed. A total of three replicates were sequenced for each sample type. Paired-end transcript reads of 100 bp length with coverage of around one million transcript reads per sample were generated. The raw transcript reads were kept and stored in separate files per sample in “fastq” file format. Data files were up-loaded to at the NCBI Sequence Read Archive (SRA) (GEO Accession Number: GSE93686).

Differential expression analysis

To acquire the counts data for differential expression analysis of the data generated, the two sample types, i.e. planktonic cells and biofilm cells, were mapped to the reference genome S. Typhi CT18 downloaded from the NCBI RefSeq database (Accession number: NC_003198.1). Data processing was done using the Tuxedo Suite program installed on a bioinformatics server at RIKEN. First Bowtie2 (Version 2.2.6) [11] was used to align the transcripts against the reference genome to produce a SAM file. The SAM file was then converted to BAM format and sorted by genome coordinates using SAMtools (Version 0.0.18) [12]. After sorting, Cufflinks (Version 2.2.1) was used for calculating the differential gene expression levels between the planktonic cells and biofilm cells [12]. CummeRbund (Version 2.12.1), an R program written for Cufflinks, was used to view and tabulate the results in an Excel file [13]. Gene annotation was also done using CummeRbund. Gene function information was added to the tabulated Cuffdiff results using an in-house Perl script.

Genes responsible for bacteria membrane matrix

The gene that was most highly up-regulated in the biofilm cells was STY1254, a multiple stress resistance protein, bhsA. It plays a role in the membrane structure of S. Typhi [16]. According to a study by Zhang et al. [17], the bhsA gene increases the stickiness of the membrane protein in E. coli, allowing it to stick to surfaces during biofilm formation. The researchers showed that deletion of the bhsA gene in E. coli caused more biofilm to form as the bacteria was not able to stick to the apparatus surface, causing them to form more biofilm matrix for protection. Gene STY1255 is involved in the peptidoglycan biosynthesis pathway and is part of the cell wall biogenesis [18]. It also plays a part in anchoring the major outer membrane Braun lipoprotein to the peptidoglycan to stabilize the cell wall, giving biofilm cells a thicker cell wall and added stability [19].

Based on the functions of the genes STY1254 (bhsA) and STY1255, and the results from our transcriptome study, the two genes may be responsible for the binding of S. Typhi to the surface of the polypropylene tubes, thus allowing biofilm formation and maintenance. Not only were the two genes responsible for adhesion, they also affected the properties of the surface membrane, which presumably allows the bacteria to survive the acidic environment of the biofilm culture media [16–19]. It was also found in a study by Salazar et al. [20] that ycfR/bhsA promotes the attachment of S. Typhimurium to the surface of glass, polystyrene, spinach leaves and tomato fruit.

The gene STY1389, which codes for cyclic-di-GMP-binding biofilm dispersal mediator protein, was also found to be highly up-regulated in mature biofilm (log₂-fold change = 1.79). In a review paper by Valentinin and Filloux [21], it was shown that cyclic-di-GMP regulated biofilm formation in P. aeruginosa by affecting various pathways such as flagella rotation to Type IV pili retraction, exopolysaccharide production, surface adhesin expression, antimicrobial resistance, and biofilm dispersion. Our result suggests that the gene STY1389 is responsible for mediating biofilm dispersion and propagation of infection at the mature stage of the biofilm cycle.

Genes responsible for antibiotic resistance

Based on the transcriptome data, the Mar regulon genes, marA and marR, were significantly up-regulated in the biofilm cells as compared to the planktonic cells. Perera and Grove [22] indicated that the Mar regulon was associated with antibiotic resistance and stress responses. The Mar regulon also regulates bacterial virulence factors, such as cell wall proteins and surface adhesins, as reported by Prieto and colleagues [23] which showed that the genes marA and marB were up-regulated when exposed to bile. Thus, it can be hypothesized that the Mar regulon prepares the S. Typhi cells for potential antibiotic resistance, or against stress factors, such as the bile used in growing the biofilm.

Yet another reason could be that some proteins serve several functions, such as the efflux pump in pathogenic organisms which serve to pump toxic chemicals out from the inside of the cells to the external environment, but can also serve as a method to extricate antibiotics from the cell cytoplasm to the external environment [22, 23]. Thus, the Mar regulon may serve two functions, ie. extrication of bile and antibiotics.

Genes associated with general metabolic mechanisms

The cold shock-like protein family, cspE, cspC and cspB, was significantly down-regulated in biofilm cells. Also, it was shown that the DNA protection and DNA binding genes, STY4154 and hupA, respectively, were down-regulated. Many membrane transport proteins, such as tatE, were also down-regulated in the biofilm cells. These results showed that while the S. Typhi cells remained in the mature biofilm, the cells entered a state of stasis and reduce the energy requirements.

Ribosomal Modulation Factor (rmf) was the most down-regulated gene in the mature biofilm cells. Based on an article by Niven and El-Sharoud [24], rmf was used for stabilizing ribosomes during stress conditions; that is, it would be expected for rmf to be highly up-regulated during the early stage of biofilm formation. However, it can also be theorized that the biofilm cells collected in this study at 24 h post-incubation were already matured and have entered the stasis stage. Thus, to reduce or inhibit general protein translation in the cells, there was a need to down-regulate rmf. This hypothesis is supported by a paper published by Yamagishi et al. [25] that showed a reduction in rmf caused cells to lose their viability and ability to form ribosome dimers as well as when the cells transitioned from the growth to stationary state. It was also hypothesized in the paper that rmf is down-regulated under certain growth conditions.

This study found that ribosomal RNA (rRNA), such as multiple ribosomal subunit genes rplD, rplW, rpsS and rplU, were significantly up-regulated in the biofilm cells. This suggests that general protein translation continues even in the sessile biofilm tissue.

Genes involved in biofilm regulation

STY0893, the bssR gene for biofilm regulation, was significantly down-regulated in the biofilm cells. According to the paper by Domka et al., [26], the bssR gene, also called yliH in E. coli cells, was up-regulated during biofilm formation. However, deletion of the yliH gene in E. coli caused the biofilm mass to increase by as much as 290-fold, but did not affect growth of the cells under normal conditions. However, this transcriptome study on S. Typhi biofilm cells showed that the bssR gene (STY0893) was significantly down-regulated as compared to the planktonic cells. BLAST analysis comparison between E. coli K-12 and S. Typhi CT18 showed that while the gene bssR is present in both species, it only has a 75% nucleotide similarity. Also since the base pair length is different between the two species; it can be assumed that the two proteins are different. The difference in results between this study (on S. Typhi) and the studies on E. coli [26] could be due to other parameters, such as different growth conditions, media and surfaces used for biofilm culture.

YjfO is a gene that encodes for biofilm stress and motility protein A. In a study done on E. coli, yjfO was found to be associated with cell survival by formation of biofilm in response to peroxide stress [27]. It was shown that yjfO was up-regulated in E. coli during the micro-colony formation and biofilm maturation processes. However, it was also shown to be unnecessary for planktonic cell growth or initial surface adhesion. Comparison of the results of the study by Weber et al. [27] with our data suggests that in S. Typhi there could possibly be a different function for the gene yjfO, as it was down-regulated in S. Typhi for biofilm formation. However, this can also be explained that since yjfO is responsible for micro-colony formation, it is also possible that the transcription rate of the gene in mature biofilm tissue was suppressed as there is no further need for the gene in the static environment of the biofilm tissue.