The bacterial strains, plasmids and primers used in this study are listed in Table S1. E. coli DH5α (TransGen, Beijing, China), S17-1 (Simon et al., 1983) strains used for general cloning conjugal transferring (Chen et al., 2016) were cultured in Luria–Bertani (LB) broth [10 g/l tryptone (Oxoid Ltd., Basingstoke, UK), 5 g/l NaCl (BBI life sciences, Shanghai, China), 5 g/l yeast extract (Oxoid Ltd.), pH 7.0–7.5] or LB agar (LB broth containing 15 g/l agar) under aerobic conditions at 37°C. When needed, the medium was supplemented with tetracycline (50 μg/ml) (BBI life sciences), gentamicin (100 μg/ml) (BBI life sciences), or carbenicillin (150 μg/ml) (BBI life sciences) for P. aeruginosa PA14 (Liberati et al., 2006), and ampicillin (100 μg/ml) (BBI life sciences) or tetracycline (10 μg/ml) for E. coli. When needed, Isopropyl β-D-1-thiogalactopyranoside (IPTG) at indicated concentrations was added to culture mediums.

For DNA manipulation, standard protocols or manufacture instructions of commercial products were followed. For the complementation of the fis gene, the open reading frame of fis and its upstream 486 bp region was amplified by PCR with PA14 chromosomal DNA as the template. The PCR product was cloned into the BamHI-EcoRI sites of pUC18T-mini-Tn7T-Tc (Choi and Schweizer, 2006), resulting in pUC18T-mini-Tn7T-Tc-fis. The pEX18Tc-T0T1 insertion was constructed by cloning the 934 bp upstream and 987 bp downstream fragments of the exsB-exsA intergenic region into the EcoRI-HindIII sites of plasmid pEX18Tc (Hoang et al., 1998), and a 289 bp DNA fragment containing terminators T0T1 was amplified by PCR from pUC18T-mini-Tn7T (Choi and Schweizer, 2006) and inserted in between the two fragments. Chromosomal gene mutations were generated by homologous recombination as described previously (Hoang et al., 1998). To construct the C-terminus His-tagged ExsA driven by its native promoter, a DNA fragment containing the 300 bp fragment upstream of exsA and the exsA coding region was amplified by PCR, the His-tag coding sequence was included in one of the PCR primers. The PCR product was cloned into the XbaI-HindIII sites of a promoterless pUCP20 (Li et al., 2016). To construct the exsA promoter lacZ transcriptional fusion (P_exsA-lacZ), the 500 bp fragment upstream of the exsA coding region was amplified by PCR and cloned into the BamHI-EcoRI sites of pDN19lacZΩ. Sequences of the PCR primers were listed in Table S1.

Bacteria were grown at 37°C under indicated conditions to indicated optical density of 600 nm (OD₆₀₀). Total RNA was isolated with a RNAprep Pure Bacteria Kit (Tiangen Biotech, Beijing, China). The cDNA was synthesized from total RNA using random primers and PrimeScript Reverse Transcriptase (TaKaRa, Dalian, China). Specific Primers (Table S1) were used for reverse transcription (RT) and quantitative PCR. For quantitative PCR, cDNA was mixed with 4 pmol of forward and reverse primers and SYBR Premix Ex Taq™ II (TaKaRa) in a total reaction volume of 20 μl. The 30S ribosomal protein coding gene rpsL was used as an internal control (Li et al., 2013). The results were determined using a CFX Connect Real-Time system (Bio-Rad, USA).

The transcriptome sequencing analysis was performed by Beijing Genomics Institute. Total RNA was isolated from bacteria at OD₆₀₀ of 1.0 with an RNAprep Pure Bacteria Kit (Tiangen). After DNase I (NEB) digestion, rRNA was removed from the total RNA by using Ribo-Zero Magnetic Kit (Bacteria, EPICENTRE). The mRNA was fragmented into short fragments by using fragmentation buffer (Ambion). Then cDNA was synthesized using the mRNA fragments as templates. The purified fragmented cDNA was combined with End Repair Mix and A-Tailing Mix for end reparation and single nucleotide A (adenine) addition. Then, the short fragments were connected with adapters. After agarose gel electrophoresis, the suitable fragments were selected for the PCR amplification as templates. During the quality control steps, Agilent 2100 Bioanaylzer and ABI StepOnePlus Real-Time PCR System were used in quantification and qualification of the sample library. At last, the library was sequenced using Illumina HiSeq™ 2000 or other sequencer when necessary. The RNA expression analysis was based on the predicted genes of strain PA14 (http://www.pseudomonas.com). P-values were calculated referring to “the significance of digital gene expression profiles” (Audic and Claverie, 1997). A strict algorithm to identify differentially expressed genes between two samples is described as follow:

Denote the number of unambiguous clean tags (which means reads in RNA-Seq) from gene A as x, given every gene's expression occupies only a small part of the library, x yields to the Poisson distribution:

The total clean tag number of the sample 1 is N₁, and total clean tag number of sample 2 is N₂; gene A holds x tags in sample 1 and y tags in sample 2. The probability of gene A expressed equally between two samples can be calculated with:

The full transcriptome sequencing data has been deposited in the NCBI SRA, with accession number SRP099178.

The bacterial cytotoxicity was determined by the lactate dehydrogenase (LDH) release assay as previously described (Anderson et al., 2010; Li et al., 2016). HeLa cells (Li et al., 2016) were cultured in DMEM medium (Corning, USA) with 10% fetal bovine serum (FBS) supplemented with 1% penicillin/streptomycin (Gibco, USA) at 37°C in 5% CO₂. 1.5 × 10⁴ HeLa cells were seeded into each well of a 96-well plate and cultured for 24 h. The medium was replaced with antibiotic and FBS free DMEM before P. aeruginosa infection. Overnight bacterial cultures were subcultured in fresh LB and grown to an OD₆₀₀ of 1.0. Bacteria were washed once and resuspended in 1 × PBS. HeLa cells were infected with bacteria at a multiplicity of infection (MOI) of 50. The plate was centrifuged at 2,000 g for 10 min to synchronize the bacterial infection. After 3 h, LDH present in the supernatant was measured using the LDH cytotoxicity assay kit (Beyotime, China). Cells treated with 1% Triton X-100 were used as positive control of maximum release (100% percentage of cytotoxicity). The percentage of cytotoxicity was calculated following the manufacturer's instruction.

All animal experiments complied with Nankai University and Chinese National Guidelines regarding the use of animals in research. The protocol was approved by the institutional animal care and use committee of the college of life sciences of Nankai University with permit number: NK-04-2012. Overnight bacterial cultures were subcultured in fresh LB at 37°C to an OD₆₀₀ of 1.0. The bacterial cultures were centrifuged at 12 000 rpm for 1 min, and the pellets were resuspended in PBS. Each 6-week old female BALB/c mouse (Academy of Military Medical Sciences, Beijing, China) was anesthetized with 100 μL of 7.5% chloral hydrate by intraperitoneal injection. Then each mouse was intranasally inoculated with 1 × 10⁷ CFU bacteria. Bacterial colonization in the lung was determined as described previously (Sun et al., 2014). Briefly, 14 h post infection (hpi), mice were sacrificed by inhalation of CO₂. Lungs were isolated and homogenized in 1% proteose peptone (Solarbio, Beijing, China). Bacterial loads were determined by plating serial dilutions and counting colonies. For the survival assay, the mice were intranasally inoculated with 2 × 10⁷ CFU bacteria and monitored for 5 days.

Six week old female BALB/c mice were intranasally inoculated with 2 × 10⁷ CFU bacteria as described above. Mice were sacrificed by inhalation of CO₂ at 6 hpi. Bronchoalveolar lavage fluid (BALF) was obtained by annulation of the trachea followed by twice instillations of 1 ml sterile PBS with 0.5 mM EDTA. 200 μl of the BALF was used for bacterial counting. The remaining BALF was centrifuged at 12,000 rpm for 2 min, and the pellets were resuspended in 200 μl TRIzol reagent (Thermo Fisher Scientific, USA). Total RNA was isolated using a Direct-zol RNA Miniprep kit (ZYMO research, USA).

EMSA was performed as previously described with minor modification (Sun et al., 2014). Briefly, DNA fragments corresponding to sequence upstream of exsA and exsC were synthesized. DNA fragments (200 ng) were incubated with 0 to 6 mM purified recombinant Fis protein at 25°C for 30 min in a 20-μl reaction [10 mM Tris-HCl, pH 7.5, 5 mM MgCl₂, 5 mM KCl, 0.1% (v/v) NP-40 (Solarbio), and 1 mM dithiothreitol]. Samples were loaded onto an 8% native polyacrylamide gel in 0.5 × Tris-borate-EDTA (TBE) buffer (0.044 M Tris base, 0.044 M boric acid, 0.001 M EDTA, pH 8.0) that had been prerun for 1 h, electrophoresed on ice at 100 V for 1.5 h, followed by DNA staining in 0.5 × TBE containing 0.5 μg/ml ethidium bromide. Bands were visualized with a molecular imager ChemiDoc™ XRS+ (Bio-Rad).

Overnight bacterial cultures were subcultured in fresh LB with or without 5 mM EGTA at 37°C for 3 h. The pellets from 1 ml cultures were then resuspended in 100 μl loading buffer (50 mM Tris-HCl, pH 6.8, 2% (w/v) SDS, 0.1% (w/v) BPB, 10% (v/v) Glycerol, 1% (v/v) 2-ME). Protein samples from equivalent amounts of protein were loaded onto a 15% SDS-PAGE gel. Proteins were separated by electrophoresis followed by transferring to a polyvinylidene difluoride (PVDF) membrane (Millipore, USA). The target proteins were hybridized with a rabbit monoclonal anti-His antibody (CST, USA) or a mouse monoclonal anti-Flag antibody (Sigma, USA). Signals were detected with an Immobilon™ Western kit (Millipore).

Overnight bacterial cultures were subcultured in fresh LB with or without 5 mM EGTA at 37°C with shaking, and samples were harvested when OD₆₀₀ of the cultures reached 1.0. The β-galactosidase activity was measured with substrate ortho-nitrophenyl-galactopyranoside (ONPG) (BBI life sciences) as previously described (Ha et al., 2004).

The statistical analyses were performed with the Prism software (Graphpad Software). The real time PCR and β-galactosidase assay results were analyzed by the Student's t-test (two-tailed). Bacterial colonization results were analyzed with the Mann-Whitney test. Survival data were analyzed with the Log-rank (Mantel-Cox) test.

To examine the role of Fis in the virulence of P. aeruginosa, we infected mice with wild type PA14 or a fis::Tn mutant from the nonredundant library of PA14 transposon mutants (Liberati et al., 2006) in an acute pneumonia model. 14 h post infection, lungs were isolated and the bacterial numbers were determined by serial dilution and plating. Compared to the wild type strain, the number of the fis::Tn mutant was significantly lower (Figure 1A). Complementation with a fis gene driven by its native promoter partially restored the bacterial load (Figure 1A). To confirm the role of Fis in virulence, we monitored the survival rate in the acute pneumonia model. Infection with wild type PA14 or the complemented strain caused death in all the infected mice, whereas infection with the fis::Tn resulted in 40% survival rate (Figure 1B).

In the PA14 fis::Tn mutant, the Tn was inserted right before the stop codon of the fis gene (Liberati et al., 2006). To examine the effect of Tn insertion on the expression of Fis, we determined the mRNA level of Fis by real time PCR. Compared to wild type PA14, the mRNA level of Fis was lower in the fis::Tn mutant, but higher in the complemented strain in bacteria at early-, mid-log and stationary growth phases (Figure 1C). Thus, the excessive expression of Fis in the complemented strain might affect bacterial virulence, resulting in partial restoration of bacterial number in lung and slower killing of infected mice (Figures 1A,B).

The fis gene is in the same operon with PA4852. Thus, the expression of PA4852 might be affected by the Tn insertion. A real time PCR assay revealed similar mRNA levels of PA4852 in wild type PA14 and the fis::Tn mutant (Figure S1), suggesting that the Tn insertion did not affect the expression of PA4852. It has been demonstrated that the translation of Fis is repressed by the small RNA RgsA (Lu et al., 2016). The RgsA level in the fis::Tn mutant was similar as that in the wild type PA14 (Figure S2), suggesting that the RgsA mediated regulation on Fis might be normal in the fis::Tn mutant. In combination, these results suggest that the lower expression of Fis might result in attenuation of the virulence of PA14. Since fis has been indicated as an essential gene in P. aeruginosa, we used the fis::Tn mutant in our further studies (Liberati et al., 2006; Jakovleva et al., 2012).

To understand how Fis affects bacterial virulence, we compared transcriptome profiles between the fis::Tn mutant and wild type PA14. Interestingly, the T3SS genes were down regulated in the fis::Tn mutant (Table 1). To confirm the expression levels of T3SS genes, bacteria were cultured under T3SS non-inducing and inducing conditions. Real time PCR assay revealed lower mRNA levels of exoU, pcrV, exsC, and exsA in the fis::Tn mutant, which were partially restored in the complemented strain (Figure 2A). By utilizing a transcriptional fusion of the exsC promoter and a lacZ gene (P_exsC-lacZ), we found the exsC promoter activity was lower in the fis::Tn mutant (Figure S3). To confirm the expression level of T3SS genes, a His-tagged exoU (ExoU-His) driven by its native promoter was transferred into the bacteria. As shown in Figure 2B, the ExoU-His level was reduced in the fis::Tn mutant under T3SS inducing condition. Since T3SS plays a major role in bacterial cytotoxicity (Yahr and Wolfgang, 2006; Tan et al., 2016), we infected HeLa cells with PA14, the fis::Tn mutant and the complemented strain. Consistent with the mRNA levels of T3SS genes, the fis::Tn mutant displayed reduced cytotoxicity to HeLa cells (Figure 2C). We then examined the expression levels of T3SS genes in the murine acute pneumonia model. Bacteria were isolated from BALF at 6 hpi. mRNA levels of exoU, pcrV, exsC and exsA were lower in the fis::Tn mutant, which was partly restored in the complemented strain (Figure 2D).

To further confirm the role of Fis in the regulation of T3SS, we cloned a C-terminus His-tagged fis gene (fis-His) into pMMB67EH (Fürste et al., 1986), where the expression of fis-His is driven by an inducible tac promoter. In the presence of this plasmid, we deleted the chromosomal fis gene, resulting in the strain Δfis/pMMB67EH-fis-His. In the absence of IPTG, the strain could not grow up in LB, confirming the essential role of Fis in P. aeruginosa. For an unknown reason, EGTA significantly repressed the growth of the strain Δfis/pMMB67EH-fis-His in the presence of various concentrations of IPTG. Thus, we only examined the expression of T3SS genes in LB medium. In the presence of 0.02 mM IPTG, the RNA levels of T3SS genes were lower in the Δfis/pMMB67EH-fis-His strain. Increasing amount of IPTG resulted in higher amount of Fis-His protein and up regulation of the T3SS genes (Figures 3A,B). Next, we examined the bacterial cytotoxicity with IPTG in the tissue culture medium. At low concentration of IPTG, the bacteria displayed minimal cytotoxicity, whereas the presence of IPTG increased the bacterial cytotoxicity (Figure 3C). In combination, these results suggest that Fis is required for the activation of the T3SS.

In LB broth, the growth rate of the fis::Tn mutant was similar as that of the wild type PA14 (Figure S4). However, it has been reported that the fis::Tn mutant grows more slowly in BM2 swarming medium (Yeung et al., 2009), indicating a role of Fis in bacterial growth under certain conditions. Thus, the lower bacterial loads in the lungs from the fis::Tn mutant infected mice might be due to slower bacteria growth rate. To verify the role of T3SS in the reduced colonization of the fis::Tn mutant, we overexpressed exsA in the fis::Tn mutant and examined bacterial cytotoxicity and colonization of lungs. As shown in Figure S5, overexpression of exsA did not affect the growth of the fis::Tn mutant. Overexpression of exsA in the fis::Tn mutant restored bacterial cytotoxicity (Figure 3D) and colonization in vivo (Figure 3E). Therefore, the defective T3SS indeed attributed to the attenuated virulence of the fis::Tn mutant.

T3SS genes are directly regulated by ExsA (Hauser, 2009; Diaz et al., 2011) and we have found that the mRNA level of exsA was lower in the fis::Tn mutant (Figures 2A,D). Since Fis functions as a transcriptional factor (TF) (Muskhelishvili et al., 1995; Schneider et al., 2001; Aiyar et al., 2002; Cho et al., 2008; Kahramanoglou et al., 2011; Prigent-Combaret et al., 2012), we suspected Fis might directly affect the transcription of exsA. By carefully searching for the consensus Fis binding motif (Cho et al., 2008; Shao et al., 2008; Kahramanoglou et al., 2011; Hancock et al., 2016), we identified a possible Fis binding site at the -10 box of the exsA promoter (Figure 4A). An EMSA assay demonstrated an interaction between Fis and this fragment (Figure 4B). As a control, the fragments up stream of the binding site could not bind to Fis at the concentration of 2 mM. Of note, at higher concentrations of Fis, DNA retardation was observed with all the tested probes (Figure 4B), which might be attributed to the nonspecific coating of DNA by Fis, resulting in the formation of Fis-DNA filament referred to as a “low mobility complex (LMC)” (Skoko et al., 2006). To further verify the specific binding between Fis and the -10 box of the P_exsA, we mutated the conserved nucleotides of the Fis binding site based on the study in E. coli. The replaced nucleotides were highlighted in Figure 4C. Indeed, alternation of the highly conserved nucleotides abolished the interaction between Fis and the fragment (Figure 4C). Since ExsC positively regulates T3SS (Hauser, 2009; Diaz et al., 2011), we also examined whether Fis specifically binds to the exsC promoter region. However, no specific binding between Fis and the test fragments were observed (Figure S6). These results indicated that Fis directly binds to the -10 box of P_exsA.

The specific binding between Fis and the P_exsA region raised a possibility that Fis affects the activity of P_exsA. To test this possibility, we constructed a P_exsA-lacZ transcriptional fusion (Figure 5A). The β-galactosidase levels were similar between wild type PA14 and the fis::Tn mutant under either T3SS non-inducing or inducing condition (Figure 5B). To confirm this observation, we constructed a C-terminus His-tagged ExsA driven by the P_exsA (Figure 5C). Consistent with the P_exsA-lacZ reporter result, similar ExsA-His protein levels were observed in wild type PA14 and the fis::Tn mutant (Figure 5D). These results suggest that Fis might not directly affect the P_exsA activity.

Previously studies demonstrated that the P_exsA transcriptional activity is much weaker than P_exsC (Yahr and Frank, 1994; Marsden et al., 2016). We observed the same results under both T3SS non-inducing and inducing conditions (Figure 6B). It has been speculated that transcription of exsA might be driven by the exsC promoter under T3SS inducing condition (Hauser, 2009; Diaz et al., 2011). If this is the case, Fis might be involved in the regulation of exsA transcription initiated from the exsC promoter. To test this hypothesis, we firstly examined whether exsCEBA are in one transcript. We constructed a P_exsC-A-lacZ transcriptional fusion, where the lacZ gene was cloned downstream of a fragment ranging from P_exsC to exsA coding region (Figure 6A). In wild type PA14, the β-galactosidase levels were much higher than that driven by P_exsA (P_exsA-lacZ), indicating a continuous transcription from P_exsC to exsA coding region (Figure 6B). We then designed a pair of primers annealing to the coding regions of exsB and exsA, thus the PCR product spans the intergenic region between exsB and exsA (Figure 7A). Total RNA was isolated from PA14 grown under T3SS non-inducing and inducing conditions, followed by RT-PCR. PCR products were observed at both conditions, with higher amount under T3SS inducing condition (Figure 7B), indicating a transcript from exsB to exsA. To further confirm the continuous transcription from exsB to exsA, we performed real time PCR with primers annealing to the coding regions of exsC, exsB, exsA as well as the exsB-exsA intergenic region (Figure 7A). The mRNA levels of all the tested genes and the RNA level of the exsB-exsA intergenic region were induced by EGTA in wild type PA14 (Figure 7C). These results suggest that exsA and exsB are in one transcript and the transcription level responses to the T3SS inducing signal.

If the P_exsC driven transcription of exsA is required for the activation of the T3SS, interception of the transcription between exsB and exsA coding regions should diminish the expression of exsA and consequently other T3SS genes. To test this hypothesis, we inserted two tandem transcription terminators (T0T1) up stream of the exsA promoter (Figure 7A). RT-PCR confirmed the break between exsB and exsA transcript (Figure 7B, lanes 5–8). Indeed, the exsA mRNA levels were significantly reduced by the insertion of T0T1 under both T3SS inducing and non-inducing conditions (Figure 8A). Consequently, the levels of exsC mRNA and the ExoU protein were reduced by the T0T1 insertion (Figures 8B,C). Therefore, the P_exsC driven transcription of exsA plays an essential role in the expression of T3SS genes.

We have demonstrated that the P_exsA activity in the fis::Tn mutant was similar to that in the wild type PA14 (Figure 5B). Consistently, insertion of T0T1 up stream of P_exsA in the fis::Tn mutant resulted in similar mRNA level of exsA as that in PA14 with the T0T1 insertion (Figure 8A), confirming that Fis is not involved in the regulation of P_exsA activity. Thus, we suspected that Fis might affect the transcription of exsA driven by P_exsC and subsequent expression of T3SS genes. To test this possibility, we transferred the P_exsC-A-lacZ transcriptional fusion into the fis::Tn mutant. The β-galactosidase level was significantly lower in the fis::Tn mutant than that in the wild type PA14 (Figure 9A). To further verify that Fis affects the transcription of exsA driven by P_exsC by directly binding to the exsB-exsA intergenic region, the conserved nucleotides inside the Fis binding sequence were mutated (Figure 6A), which had been shown to significantly reduced the binding between Fis and the fragment (Figure 4C). The mutated P_exsC-A-lacZ fusions were designated as P_exsC-Am1-lacZ and P_exsC-Am2-lacZ, respectively (Figure 6A). Compared to the original P_exsC-A-lacZ, the P_exsC-Am1-lacZ, and P_exsC-Am2-lacZ yielded lower levels of β-galactosidase in wild type PA14 (Figure 9A). In combination, these results suggest that Fis affects the P_exsC driven exsA transcription through direct binding to the exsB-exsA intergenic region.

We then isolated RNA from the fis::Tn mutant under T3SS inducing or non-inducing conditions and performed RT-PCR with the primers annealing to the coding regions of exsB and exsA. The PCR product level was much lower than that in the wild type PA14 (Figure 7B). Consistently, real time PCR results revealed that the RNA spanning the exsB-exsA intergenic region was less in the fis::Tn mutant under T3SS inducing condition (Figure 7C). These results suggest a role of Fis in the continuous transcription from exsB to exsA. Since ExsA activates P_exsC (Diaz et al., 2011), our observation might be due to defective translation of the ExsA indirectly affected by Fis. To test this possibility, we utilized two previously constructed C-terminus FLAG-tagged exsA (exsA-Flag) driven by an exogenous tac promoter (Li et al., 2013). In one of the construct, namely exsA-Flag-A, the exsA coding region was directly fused with the tac promoter, whereas in the other construct (exsA-Flag-S), the exsB-exsA intergenic region was included (Figure 9B). Without the exsB-exsA intergenic region, similar levels of ExsA-FLAG were observed in wild type PA14 and the fis::Tn mutant. However, presence of the exsB-exsA intergenic region resulted in less ExsA-FLAG in the fis::Tn mutant (Figure 9C), thus confirming the role of Fis in the continuous transcription from exsB to exsA. Consistently, in the fis::Tn mutant the exsA-Flag-A resulted in higher mRNA levels of exoU, exsC, and pcrV than the exsA-Flag-S (Figure 9D). In wild type PA14, the exsA-Flag-A resulted in higher mRNA levels of exoU, exsC, and pcrV than those in the fis::Tn mutant, which might be due to higher expression of the chromosomal exsA. In combination, these results suggest that Fis is required for the P_exsC–dependent transcription of exsA, which is required for the full activation of the T3SS.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.