Targeted inactivation of Salmonella Agona metabolic genes by group II introns and in vivo assessment of pathogenicity and anti-tumour activity in mouse model

Bacterial strains, growth and storage conditions

A food source strain S. Agona (Ag1) and a genetically modified Ag 1 were used in this study. The food source strain S. Agona was recovered from indigenous vegetables in Selangor state, Malaysia (Khoo et al., 2009). The genetically modified Ag1 was constructed in previous work with double mutations in sopB and sopD genes, termed as ΔsopBΔsopD double knockout (Khoo et al., 2015). This particular ΔsopBΔsopD double knockout was used in this study for further genetic engineering work to further reduce its pathogenicity. The bacteria was cultured in Nutrient Broth or Nutrient agar plate (Merck KGaA, Darmstadt, Germany) at 37 °C and stock stored in 15% glycerol at −80 °C.

Cancer cell line and growth conditions

Mus musculus, mouse colon carcinoma cell line, CT-26 was obtained from American Type Culture Collection (ATCC), USA, and cultured in Roswell Park Memorial Institute 1640 (RPMI 1640) (ScienCell, San Diego, CA, USA) complete medium supplemented with 10% fetal bovine serum (Sigma-Aldrich, Darmstadt, Germany), 1% v/v penicillin/streptomycin (Nacalai Tesque, Japan) and 2% of L-glutamine (Gibco Life Technologies, São Paulo, Brazil) in a humidified incubator with 5% CO₂. Cell line was detached from the culture flask using a 0.25% trypsin-EDTA solution and re-suspended as monolayer cell suspension in RPMI 1640 culture medium.

Group II intron targeted gene disruption

Salmonella Agona leuB and argD metabolic genes were amplified using polymerase chain reaction (PCR) in the present study followed by gene disruption using Targetron Gene knockout System (Sigma-Aldrich, Darmstadt, Germany). A computer algorithm at Sigma-Aldrich Targetron design website (http://www.Sigma-aldrich.com/targetronaccess) was used to identify the potential L1.LtrB insertion sites. Three primers were designed to modify the RNA portion of intron template which base pair optimally to corresponding insertion sites, referred to as IBS, EBS1d, EBS2. The designed primers are listed in . A splicing-by-overlap-extension (SOE) PCR reaction which retargets the RNA portion of the intron by primers-mediated mutation was performed according to the manufacturer’s protocol with slightly modification. The purified 350 bp PCR containing modified intron portion was then digested with HindIII (1U) and BaeGI (0.5U) (Sigma-Aldrich, Darmstadt, Germany) followed by ligation into the pACD4-C intron donor chloramphenicol-resistant linear vector (2 ng) (Sigma-Aldrich, Darmstadt, Germany). The ligation reaction was performed using Quick-link T4 Ligation Kit (Sigma-Aldrich, Darmstadt, Germany) according to the manufacturer’s instruction. The ligated product was first transformed into E.coli QIAGEN EZ competent cells due to low transformation competency of S. Agona strain. The pooled plasmid from E. coli was then co-transformed with plasmid pAR1219 which expressed T7 RNA polymerase into S. Agona via electroporation. The intron insertion in the transformant was identified by colony PCR using gene specific primers which amplifies the entire intron inserted target gene. To adapt the system for multiple gene knockouts, the first constructed leuB knockout was applied to sequentially mutate the second argD metabolic gene for the construction of ΔleuBΔargD double knockouts. The ΔsopBΔsopD double knockout was subjected to the same experiment as mentioned above to construct the quadruple knockout ΔsopBΔsopDΔleuBΔargD.

In vitro growth assessment of knockout strains

Growth curve analysis of wild-type S. Agona (W), double knockout ΔleuB ΔargD (LA), double knockout ΔsopB ΔsopD (BD) and quadruple knockout ΔsopB ΔsopD ΔleuB ΔargD (BDLA) was carried out in M9 minimal media. Each overnight culture in Nutrient broth was washed twice with phosphate-buffered saline (PBS) prior to inoculation in M9 minimal media. M9 minimal media was prepared according to recipe describe by Sambrook (Sambrook & Russell, 2001). Overnight bacterial culture was then inoculated in M9 minimal media at 1:100 ratio. Double knockout LA and quadruple knockout BDLA was cultured in M9 minimal media supplemented with leucine and arginine separately, and both leucine and arginine together in respective flasks. Samples were incubated with shaking (200 rpm) at 37 °C. Bacterial growth was recorded for a total period of 12 h by measuring the OD₆₀₀ of the cultures at intervals of 1 h.

Invasion assay

The CT-26 cells at number of 1 × 10⁵ were seeded in 96-well plate for 24 h at 37 °C. Cell lines were washed twice with sterile PBS prior to the infection with different Salmonella strains. Salmonella strains in late-log phase growth were diluted in 100 µl of RPMI media and added at the 100 multiplicity of infection (MOI). After infection, cells were washed three times with PBS and then treated with 100 µg/mL of gentamicin (Bio-Basic, Markham, Ontario, Canada) to kill the external bacteria, for 1 h at 37 °C. After the gentamicin incubation, the cells were washed for three times again and then incubated with 0.1% Triton X-100 (Merck, Darmstadt, Germany) for 10 min at 37 °C. The number of internalized bacteria was determined by plating ten-fold serial dilutions of the cell lysates on xylose lysine deoxycholate (XLD) agar plates.

In vivo toxicity profile of attenuated Salmonella

Survival test analysis

Mice used in this study were purchased from a local supplier (Saintifi Enterprise, Kuala Lumpur, Malaysia) and adapted for one week prior to the experiment (see ‘Ethics Statement’ below). All mice were maintained in an enriched environment and in accordance with institutional animal care and guidelines. Consent for experiments involving animal was obtained from the appropriate University Putra Malaysia (UPM) Ethics Committee. Six-to-eight week old male, BALB/c mice were randomly divided into five groups and treated with W, BD, LA, BDLA and PBS as negative control. The inoculums were administered intraperitoneally with different concentrations of bacteria ranging from 10¹ to 10⁷ cfu in 100 µl PBS. Animal deaths were recorded for 30 days.

Ethics statement

All animal experiments conformed to the Institutional Animal Care and Use Committee of Universiti Putra Malaysia using the protocol UPM/IACUC/AUP-R056/2013.

Evaluation of potential toxicity of bacterial infection

Six-to-8 week-old male, BALB/c mice were divided into five groups and treated with W, LA, BD, and BDLA at concentration of 10⁴ cfu in 100 µl of PBS. PBS was used as negative control. Mice were sacrificed at day-1, day-3, day-10, day-21 and day-36. The potential side effects and toxicities induced by the treatments were evaluated in the aspect of bacterial distribution, histopathology and full blood count analysis.

Analysis of bacterial distribution

The isolation and titration of bacteria from mice were performed as described with slight modification (Jia et al., 2007). Briefly, five organs include liver, spleen, kidney, intestine, and lung were removed aseptically and homogenized with PBS at a ratio of 10:1 (PBS volume (mL): tissue weight (g)). The homogenates were then serial diluted in ten-fold dilution, and plated onto XLD agar and incubated overnight at 37 °C. The titer of bacteria in each organ was determined on the next day by counting colonies and dividing them by weight of tissue (c.f.u/tissue (g)).

Histological analysis

Portion of liver, spleen, lung, kidney and intestine were fixed immediately with 10% paraformaldehyde/0.1 M phosphate buffer (PB) with fixative volume of 15–20 times greater than tissue volume. The subsequent embedding and staining processes were proceeding according to standard histological procedures. Inflammation and infectious foci of the tissues were observed under a light microscope to evaluate the severity of inflammation induced by bacteria.

Full blood count analysis

Blood samples were collected from mice at day-1, 3, 10, 21, 36 days post infection. Blood samples were collected through cardiac puncture and were stored immediately in EDTA blood collection tubes. The level of neutrophils, lymphocytes and white blood cells were determined subsequently by Haematology and Clinical Biochemistry Laboratory, Faculty of Veterinary, Universiti Putra Malaysia.

In vivo tumour growth suppression

Tumour growth inhibition assay was conducted in mice with large and small tumour model separately and each tumour model was subjected to single and multiple administrations of the treatments. For tumour implantation, 6-to 8-week-old male, BALB/c mice were implanted subcutaneously on the mid-right flank with CT-26 colon carcinoma cells at concentration of 10⁶ cells in 100 µl of PBS. Tumour was allowed to grow for 2 weeks before treatment. Mice were examined daily until the tumour size was approximately 250 mm³ and 450 mm³ for small and large tumour respectively. Different Salmonella strains, W, BD, LA and BDLA were administered intraperitoneally at concentration of 10⁴ cfu in 100 µl of PBS. Anti-tumour activity of the treatments was evaluated by measuring tumour size every 2 days. Tumours were measured individually with a digital caliper. Tumour volumes were determined by the formula: tumour volume = length × width² × 0.52 (Yoon et al., 2011). The experiments was repeated as above experimental groups, the mice were re-treated 1 week after the first treatment and were followed up to 30 days for tumour size analysis.

Statistical analysis

Statistical significance of experimental results was determined by Student’s T-test to compare the differences among two groups. For multiple group comparison, one-way ANOVA analysis of variance was performed followed by post-hoc analysis. All data are expressed as the mean ± SEM; a p-value less than 0.05 was considered to be statistically significant. The statistical analysis was performed using SPSS software (version 16.0; SPSS, Chicago, IL, USA).

leuB and argD metabolic gene knockout using group II intron technology

Group II intron technology was utilized to construct genetically modified S. Agona. The potential targets sites for leuB (9 sites) and argD (5 sites) were predicted by Sigma-Aldrich Targetron. Targetron integration was successful for the 6th design for leuB and 2nd design for argD (Fig. S1). Splicing overlap extension polymerase chain reaction (SOE PCR) was then performed to modify the intron sequence. As shown in Fig. 1A, SOE PCR gave a 350 bp of retargeted intron for leuB and argD gene which is necessary to cause gene disruption. The knockout mutant strains were confirmed by colony PCR using gene-specific primers to flank their respectively inserted intron. The results (Figs. 1B & 1C) show that the intron has been successfully inserted resulting in a significantly increased PCR product sizes approximately 1,000 bp as compared with wild-type S. Agona. Analysis of 92 S. Agona colonies for leuB insertion showed intron insertion occurred at a frequency of 15% where 14 colonies contained intron-inserted gene, that produced expected 2,043 bp PCR product (Fig. 1B). For argD gene, three out of 92 colonies analysed showed gene insertions, indicating gene targeting at the frequency of 3%, which produced a PCR product of around 1,887 bp (Fig. 1C). Stability of intron insertion and purity of clone isolated was evaluated by passaging the bacterial culture in medium for 30 days consecutively and no reversion to wild-type was observed (Fig. S2A). The inserted intron were further verified by PCR amplified and sequenced using the gene-specific primers (Figs. S2B & S2C).

Growth of auxotrophic mutant is defective but invasiveness is not affected in vitro

The in vitro phenotypic characteristic of genetically modified S. Agona was assessed by evaluation of growth in M9 minimal media. Evaluation of growth kinetics revealed that the genetic engineered mutant strains that carried the leuB and argD insertional mutations in S. Agona caused remarkable changes in growth characteristics. Double knockout ΔleuB ΔargD (LA) and quadruple knockout ΔsopB ΔsopD ΔleuB ΔargD (BDLA) were unable to grow in M9 minimal media throughout 12 h. Both of the metabolic gene knockout strains were able to grow partially in M9 minimal media supplemented with leucine only reaching final cell densities OD₆₀₀ ∼ 0.4 compared with OD₆₀₀ ∼ 2.0 for the parental strain after incubation period of 12 h. However, both genetically modified LA and BDLA strains were still unable to grow in M9 minimal media supplemented with arginine only. Growth inhibition of LA and BDLA mutant strains were able to be fully reversed with the supplementation of both leucine and arginine in M9 minimal media. The corresponding cultures reached final cell densities (OD₆₀₀ ∼ 1.9) that were similar to the parental strain S. Agona. Growth ability of double knockout ΔsopB ΔsopD (BD) is not affected as it reaches similar cell density as the parental strain throughout 12 h of incubation period. Salmonella is an intracellular bacteria and invasion is essential to cause the infection. We next examined the invasiveness of these modified S. Agona strains and found no significant difference in the recovered bacterial colony forming unit (Fig. 1E). Together, these data indicate that the modification of leuB and argD genes affect bacterial growth without altering its invasion property.

BDLA mutant replicate slower in internal organ and induced lower inflammatory response

Our previous study has shown that sopB and sopD genes are not required for full pathogenicity in C. elegans model as there is no virulence difference observed between wild-type and mutant strains (Khoo et al., 2015). In order to provide a better understanding of these virulence and the link with metabolic genes in Salmonella, the mutants S. Agona were subjected to pathogenicity analysis by following mice survivability, bacterial colonization and histolopathological analysis.

To investigate if the genetically modified Salmonella strains could possibly lead to lower severity and lethality, BALB/c mice were infected with 1 ×10³ to 1 ×10⁷ cfu of wild-type and genetically modified strains and monitored for survival over 30 days. Wild-type and genetically modified S. Agona-infected mice at the dosage of 1 ×10⁵ −1 ×10⁷/cfu died on the next day after infection. However, all of the mice survived up to 30 days with the lower dosage 10³/cfu for each of the treatment groups and dosage at 10⁴/cfu contributed to the difference in mice survivability between all of the treated groups. As shown in Fig. 2A, the survival rate for wild-type (WT) infected mice is 33.33%, 83.33% for BD, 50% for LD and 83.33% for BDLA. Overall, slightly improvement of mice survivability was noted after genetic modification.

To establish whether the bacteremia resulted in colonization of internal organs, we assayed the number of bacteria in spleen, liver, lung, intestine and kidney at five different time points. Systemic infection could be detected starting from day 1 post-infection in both the wild-type and genetically modified S. Agona-infected mice as indicated by the presence of bacteria in all vital organs tested (Fig. 2B). The bacterial load in the examined tissues increased on day-3 post-infection and started to decline from day 10 post-infection, except for the lung tissue which only started to reduce by day 21 post-infection (Fig. 2B). Remarkably, BDLA-infected mice demonstrated the lowest bacterial load amongst all the genetically modified strains in spleen (∼5-fold lower than WT on day-3 and day-10 post-infection) and liver (7.4-fold and 2.5-fold lower than WT on day-3 and day-10 post-infection respectively).

Histology was performed to examine the inflammatory response of spleen and liver to Salmonella infection. The spleen from the PBS-control mouse showed normal architechture of splenocytes, with clear distinction between the red (RP) and white pulp (WP) (Fig. 2C). As for the mouse challenged with wild-type S. Agona (Fig. 2D), the white pulps are fused. Besides that, marginal zone (MZ) were not clearly observed and number of extramedullary hematopoiesis was increased. However, disruption of normal splenic architechture and inflammation was less severe in the mouse infected with BDLA mutant S. Agona (Fig. 2E) as compared to wild-type. We next quantified the number of infection foci within liver to assess the severity of inflammation. As shown in Fig. 2F, liver from the mouse injected with PBS showed normal histology and no infection foci were seen. Liver from the mouse challenged with wild-type S. Agona showed higher number of infection foci per field (Fig. 2G) compared to liver of mouse infected with BDLA mutant S. Agona (Fig. 2H). This phenomenon indicates that BDLA may have lower pathogenicity.

We then assessed the immunological response evoked by wild-type and mutant strains by determining the titer of white blood cell, neutrophil and lymphocyte at 5 different time points. Administration of S. Agona into BALB/c mice led to significant increment in total white blood cell concentration at day-10 post-infection as shown in Fig. 3A. The treated group showed significant difference compared with the control group (PBS) on day-10 and day-21 post-infection. At day-10 post-infection, LA and BDLA treated group showed 1.2-fold and 1.3-fold lower level of induction respectively when compared to WT group. Salmonella infection also induced a significant increase in neutrophil level in mice on day 1 post-infection and subsequently declined on day-3 post-infection followed by normal level on day-21 post-infection (Fig. 3B). Salmonella treated group demonstrated significant difference in neutrophil level compared with PBS control group from day-1 to day-10. LA and BDLA led to modest changes in the increment of neutrophil level, with 1.7-fold and 2-fold lower than WT on day 10 post-infection. Furtherpore, Salmonella infection also caused suppression of lymphocytes on day 1 post-infection and subsequently increased on day-3 post-infection (Fig. 3C). Lymphocyte level returned to normal level on day-36 post-infection. Salmonella treated groups showed significant difference in lymphocyte level compared to PBS control group from day 1 to day 10 (Fig. 3C). At day-10, treatment with LA and BDLA led to modest suppression of lymphocyte level with 1.3-fold and 1.6-fold lower level as compared to WT group. No significant differences were observed between treated group and control group for the rest of the indicated time points. The results from all the tests indicate that BDLA has the lowest virulence as compared with the others.

Auxotrophic strain LA and BDLA show higher tumour growth suppression efficacy

In order to evaluate the potential effect of Leu,,Arg auxotrophy of S. Agona on tumour growth in vitro, CT-26 colon carcinoma-bearing mice were treated with 1 × 10⁴ cfu of mutant strains. The bacteria in liver and tumour were quantified in vitro and the number of colony forming units (cfu) per gram of bacteria in tumour was found to be 2 ×10⁶-fold higher than liver, indicating that genetically modified S. Agona auxotrophic strain prefer to colonize in tumour tissue (Fig. 3B). Assessment of anti-tumour activity mediated by the genetically-modified strains using T/C% (treated over control tumour volume ratio) revealed a modest decreased in the tumour size compared to WT (Fig. 3C). In large tumour model, LA and BDLA able to show lower T/C% when compared to WT treated group (Fig. 3C) (W = 70.36 ± 1.6922; BD = 65.98 ± 1.6636; LA = 59.09 ± 3.806; BDLA = 60.5 ± 2.4758). However, in small tumour model there was no significant tumour growth suppression observed for all the treatments. We also investigated if multiple administration of bacteria dosage could enhance the effectiveness of tumour suppression but found no improvement in the therapy (Fig. 3C). We also examined the side effects of these genetically-modified strains in tumour-bearing mice. The bacterial load in the spleen of tumour-bearing mice was lower compared to tumour-free mice (Fig. S3A). No significant difference was observed in the body weights of the treated and untreated groups (Fig. S4). Additionally, lower systemic neutrophil levels were found in tumour-bearing mice compared to tumour-free mice (Fig. S3). These data suggest that accumulation of Salmonella in tumour tissue induced less side effects.