RNA-seq and two-dimensional gel electrophoresis revealed genomic and proteomic signatures during bacteriophage-host interactions in Pseudomonas aeruginosa

Background Understanding the biological nature of bacteriophage is important in exploring the therapeutic and biotechnological potentials of bacteriophages. However, available information is limited to the infection processes on either model phages infecting Escherichia coli or lytic phages against pathogens. The interplay between lysogenic phage and its host was rarely studied. Results We investigated the interactions between Pseudomonas aeruginosa and a lysogenic bacteriophage PaP3 through RNA-seq and two-dimensional gel electrophoresis (2D-GE). Compared to the uninfected host, a total of 2,891 (51.3%) differentially expressed genes (DGEs) were identified, most of which were repressed by phages, including the changes in metabolic-related and virulence-associated genes. The RT-qPCR results showed consistent directional changes compared with the RNA-seq results. According to 2D-GE, phage structure proteins were detected after phage infection. The host proteins, such as flagella hook-associated proteins, disappeared gradually after phage infection and may be shut off by phage. Conclusions All these indicate that although lysogenic phages do not immediately lyse the host, they play a significant regulatory role in the expression of host genes. Our findings provide an expanded view of the lysogenic phage infection processes and may offer potential targets for therapeutic intervention against P. aeruginosa infections.

3 than two hundred of P. aeruginosa phage genomes have been deposited in NCBI, includes dsDNA, dsRNA, and ssRNA phages [4]. besides, the therapeutic potential of P. aeruginosa phages has also been extensively explored in animal models [5,6]. In 2018, the trial was stopped early due to some obstacles, but it is the hope of an anti-infection route [7].
Recently, the Engineered bacteriophages successfully treated a 15-year-old patient who was infected with mycobacterium [8]. However, a solid understanding of phage-host interactions at the molecular level is essential for future phage therapy applications.
Moreover, these cellular processes that targeted by phages may be considered as potential antimicrobial drug targets [9]. Thus, the molecular interactions between phage and host is valuable for identifying potential antimicrobial target.
P. aeruginosa lytic phage-host interactions have recently been revealed by different research groups. Chevallereau et al. (2016) found that the P. aeruginosa bacteriophage PAK_P3 can hijack RNA processing and significantly deplete bacterial transcripts to facilitate phage replication [12]. Zhao et al. (2017) also found that lytic phage PaP1 infection results in downregulation of 354 differentially expressed genes (DEGs) [18]. De Smet et al. (2016) revealed that metabolic impacts were highly phage-specific and that phage-encoded auxiliary metabolic genes can reprogram host metabolism in a phagespecific manner [11].
Our group has studied the interactions between the lysogenic phage PaP3 and its host through microarray [14]. We further elucidate this process through RNA sequencing (RNAseq) and two-dimensional gel electrophoresis (2D-GE). The transcriptomic data and 2D-GE 4 results allowed us to gain a global view on the hindrances between phage and bacterial host interaction.

Analysis of the PA3 Gene Dynamic Expression after Phage Infection.
According to the one-step growth curve of PaP3 and the dynamic infection cycle between phage and host, we examined the transcriptional changes at five time points after phage infection (5,10,20,30, and 80 min) by using the RNA-seq analysis, and phage-uninfected host cells (0 min) were used as controls. These time points contain the entire cycle of phage infection in the host and thus fully represent host gene expression. The six sequenced reads covered between 70% and 100% of genes, with an average of 4,298 genes, which accounted for 80% of all reference genes.
Relative to the uninfected host, a total of 2,962 DEGs were obtained according to the gene expression analysis; among these DEGs 2,891 and 71 were PA3 and PaP3 genes, respectively. The 2,891 DEGs (fold change ≥2, P < 0.05), included 1,550 up-regulated and 1 368 down-regulated DEGs (Fig. 1). Moreover, most upregulated and downregulated genes are detected at 80 mins, when most cells are lysed and phages are released. Thus, the results from 80 mins might not be accurate.
Host phage regulation mainly occurred in the early logarithmic phase (10-20 min in this experiment). In the DEG analysis, 1,550 genes were upregulated, while 1,368 genes were downregulated. Analysis of the functions of these suppressed PA3 genes indicate that except for PA3 gene itself and a large proportion of undefined gene, the rest of the genes are mainly distributed in the transcriptional regulators (339/3863, 8.7%), amino acid synthesis and metabolism (185/3863, 4.7%), post-translational modification and degradation (166/3863, 4.3%), and energy metabolism (150/3863 3.8%) (Fig. 2). These results indicate that PaP3 exerts a general inhibitory effect on host genome transcription. 5 A total of 71 phage genes were clustered by hierarchical cluster analysis based on the result of RNA-Seq analysis. It gathered into 3 clusters of genes, including 12 early genes (cluster1, at 5-10 min), 21 late genes (cluster 2, at 10-20 min) and 38 middle genes (cluster3, at 30-80 min) (Fig. 3). The results are consistent with microarray analysis results (Zhao et al., 2016b).

RT-qPCR Validation of Selected DEGs.
To validate DGEs identified by RNA-seq, we selected nine DEGs for RT-qPCR (Table 1)  KEGG pathway significant enrichment analysis can identify the most important biochemical metabolic and signal transduction pathways ( Table 2). The above DEGs that may be involved in the metabolic pathway were analyzed by combining the KEGG pathway database. As compared to the control group, a total of 2,962 DEGs were observed after phage PaP3 infected the host, and a total of 25 metabolic pathways were identified. The genes involved in differential expression are mainly involved in metabolic pathways, biosynthesis of secondary metabolites, and two-component systems. Pathways with enriched differential genes were mainly observed in the 30-and 80-min samples, further indicating that the regulation of phage PaP3 to the host mainly occurred in the late stage.
Interestingly Analysis of pathways enriched with differential gene expression in the late stage of phage infection show that the Cationic antimicrobial peptide (CAMP) resistance was significantly inhibited. Also, the suppressed the metabolism paths of essential amino acids (such as valine, leucine, isoleucine, tryptophan, lysine, histidine, arginine, and beta alanine), which were critical to bacterial life. In the aliphatic acid metabolic pathways, most of the relevant DEGs showed suppressed expression by inhibiting the gene expression of intermediates catalytic enzymes in aliphatic acid metabolism. This pathway involved two suppressing genes, namely, synthesizing aliphatic acid coenzyme FadD1 and FadD2. Mutations of these two enzymes not only lead to synthesis reduction of lipase, protease, rhamnolipid, and phospholipase in P. aeruginosa but also reduce the use of carbon sources in the environment [19]. The mentioned carbon sources include aliphatic acids and choline phospholipids (lung surface-active substance). Furthermore, bacterial movement and group movement are inhibited. FadD1 and FadD2 are also related to the virulence of P. aeruginosa, and the expression of these two enzymes are detected in cystic 7 fibrosis cases [20,21].

Influence of Phage Infection on Host Virulence.
The VFBD database revealed 251 virulence-related genes annotated in the PAO1 strain genome and 115 genes differentially expressed after the bacteriophage PaP3 infected host PA3. Among these genes, 52 (45.2%) were downregulated, including most flagella genes involved in adhesion (adherence), type IV pili biosynthesis and twitching motility-related genes and Hcp secretion island-1 encoded type VI secretion system (H-T6SS) [22]. Further analysis showed that the transcription factors controlling these virulence genes also 8 presented low expression levels, suggesting that the phage PaP3 can globally manipulate the expression level of the host strain PA3 transcriptome through transcription factors with broad regulation. The upregulated genes in virulence-related genes included those involved in Phenazines biosynthesis with antimicrobial activity and alginate biosynthesis with antiphagocytosis, as well as genes involved in rhamnolipid and pyoverdine biosynthesis. The upregulated virulence-related genes also include the biotypes of P.
aeruginosa type Ⅲ secretion system from biosurfactants. These results suggest that the upregulated virulence genes may be involved in PA3 phage PaP3 immune resistance. In conclusion, phage infection may alter the virulence and resistance of hosts.

Comparison of the Host DEGs from RNA-seq and Microarray Platforms.
Globally  Fig. 4). This comparison suggested that the absolute level of gene expression determined by both methods were correlated.
Interaction of PA3 and Phage PaP3 at the Proteome Level.

9
A total of 6 2D-GE maps were created using the samples from uninfected PaP3 cultures (0 min) and from infected PaP3 cultures at the time points of 5, 10, 20, 30, and 80 min.
These maps were individually matched with a reference gel map of uninfected PA3 cultures (Fig. 5A) min. Furthermore, high sequence identity (E = 3e − 109 , identity = 62%) was observed between ORF16 and the phage particle protein of Pseudomonas phage TL by NCBI BlastP analysis. Both ORF06 and ORF16 were classified as phage structure proteins. Thus, the structural proteins of phage PaP3 were detected in the early phase of infection. The expression level of these proteins was listed (Fig. 5B and 5C). The trend is overall fit between RNA-seq and the proteomic content, except phage protein Orf16, which disappeared in the gel but was upregulated in the RNA-seq data after 10 min.

Discussion
In this paper, we describe a P. aeruginosa lysogenic phage infection process through RNAseq technology and 2D-GE-based whole-cell proteomics. In recent years, investigation on the transcriptional response of the host cell to phage infection is an area of extensive concern in transcriptome research by microarray or RNA-seq. A total of 105 references containing the words "transcriptome" and "bacteriophage" in the PubMed Central literature database are available to date, and most of the references describe host response to phage infection. In both the previous study and this work, a large number of observed. In the present study, we also observed three significantly downregulated proteins, FlgL (PA1087), FlgK (PA1086), and GroEL (PA4385) (Fig. 5). These data revealed a significant inhibition on the motility and attachment of the host, as agreed with the RNAseq data. However, the mechanism on how phage protein inhibits the motility of the host needs further investigation.

Conclusions
In this study, the interaction between PaP3 and host P. aeruginosa PA3 was analyzed by transcriptome sequencing which is verified at mRNA and protein levels (RT-PCR and two-dimensional protein electrophoresis). In addition, we compared and analyzed differences between microarray and RNA-seq at the transcriptome level, which provides new evidence for the verification of the two methods. The results showed that although PaP3 was a lysogenic phage, it had a great influence on the host. It could significantly down-regulate the genes of the host bacteria (especially in the middle and late stage), and plays an important role in the host metabolism and the expression of virulence genes. Our research not only enriches our biological understanding of lysogenic phages, but also may provide new targets for prevention and control of P. aeruginosa infections in the future.

Bacterial Strains and Growth Conditions.
The P. aeruginosa PA3 strain and the lysogenic phage PaP3 were maintained in our laboratory. P. aeruginosa strains were grown in Luria-Bertani (LB) medium (for broth culture) or 1.5% (wt/vol.) agar LB plates at 37 °C. The PaP3 particles were collected and purified using CsCl gradient ultracentrifugation.

One-Step Growth Curve.
To determine the PaP3 one-step growth curve, we followed the methods described by Lu et al [28]. Briefly, PA3 logarithmic phase cultures (optical density at 600nm [OD600] of 0.5) were infected with PaP3 (MOI [multiplicity of infection] of 10). After incubation at 37 °C for 5 min to allow adsorption, the mixture was centrifuged for 30 s at 13 000 g. Unabsorbed phages were removed from supernatants by washing twice with LB medium. Sediments were suspended in 5 mL LB, and cultures were grown at 37 °C. The total time of adsorption and washing was approximately 10 min. A total of 50 μL of the sample was withdrawn every 10 min, and the PaP3 particle count was determined using the doublelayer agar plaque method. Burst time and size were calculated based on the one-step growth curve.

RNA-seq and Quantitative Qeal-Time PCR (RT-qPCR).
The samples for RNA-seq and RT-qPCR were prepared as previously described [14].  Table 3. 16S rRNA was selected as the reference gene for normalization. The RT-qPCR results were normalized using 16S rRNA and expressed as fold change (log2 scale) by the comparative Ct method. Control (0 min) is normalized as 0. Table 3. Primers used in this study

2D-GE for the Proteomic Analysis.
The proteomic analysis was conducted by obtaining samples at 10-min intervals after PaP3 infection and performing two-dimensional electrophoresis (2D-GE) separations and image analyses on clear cell lysates [30]. 2D-GE gels were stained with silver, and images were acquired using the ImageScanner combined with LabScan software. The 2D-GE maps were analyzed, and spot data were generated using the standard spot detection parameters in the ImageMaster 2D platinum software. Differential spots were selected and identified using ESI-MS-MS[31].

Ethics approval and consent to participate
Not applicable

Availability of data and materials
The reference genome and RNA-seq data used in this study can be found on NCBI. PaP3 and PA3 were isolated and stored in our laboratory, which are available from the corresponding author based on reasonable requests.