The strains and plasmids used in this study are listed in Table 1. E. coli TOP10 (Invitrogen, Paisley, United Kingdom) were used as a cloning host for the creation of pPL2CBR^opt. E. coli was grown aerobically at 37°C in Luria-Bertani broth (LB) medium. L. monocytogenes EGDe (Glaser et al., 2001) and murinized L. monocytogenes EGDe^m (Monk et al., 2010) were grown aerobically at 30°C in brain heart infusion (BHI) medium (Oxoid, Basingstoke, United Kingdom). Clinical strains of L. monocytogenes strains 130029, 140030, and 150008 were obtained from clinical cases of listeriosis in Ireland (manuscript in preparation). The pH of the LB and BHI medium was buffered with 100 mM 3-(N-morpholino)propanesulfonic acid (pH 7.4) where outlined in the text. When required, antibiotics were added to the media as follows: for E. coli; chloramphenicol (30 μg/ml), ampicillin (100 μg/ml) and for L. monocytogenes; chloramphenicol (7.5 μg/ml). For measurement of growth curves, L. monocytogenes strains were grown overnight and diluted in fresh medium until the OD₆₀₀ reached 0.05. Subsequently strains were grown shaking at 30°C and readings were taken using a spectrophotometer. To measure bioluminescence during growth over time, overnight cultures were diluted in fresh medium to OD₆₀₀ 0.05 and subsequently grown at 30°C. At selected time points, 50 μl of culture was mixed with 50 μl luciferin (PerkinElmer) with a final concentration of 150 μg/ml as described by the manufacturer.

CBR was codon-optimized (CBR^opt) for expression in Listeria and synthesized by GenScript (GenScript United States Inc., Piscataway, NJ, United States). The CBR^opt gene (Gene Accession No. KY628997) was synthesized with an intact upstream constitutive Listerial promoter, highly expressed Listerial promoter (P_HELP) (Riedel et al., 2007) and the fragment was sub-cloned into the pPL2 backbone (Lauer et al., 2002) in E. coli using Xho1 and Pst1 restriction enzymes resulting in pPL2CBR^opt (see Figure 1). Plasmid DNA was isolated from E. coli using a QIAprep Spin Miniprep kit according to the manufacturer’s instructions (QIAGEN, Crawley, United Kingdom). The sequences were verified by sequencing through MWG Biotech AG (Ebersberg, Germany). Transformation of L. monocytogenes was performed as described (Park and Stewart, 1990). Successful integrants were screened for site-specific integration by PCR using template DNA of L. monocytogenes as described previously (Bron et al., 2006). Restriction endonucleases (Roche Diagnostic, Mannheim, Germany), T4 DNA ligase (Roche) and 2× PCR mixture (NEB, United Kingdom) were used as advised by the manufacturer. From each successful transformant one colony carrying the anticipated genotype was chosen for further bioluminescence analysis using an IVIS Lumina II system (PerkinElmer).

Listeria monocytogenes EGDe::pPL2CBR^opt, L. monocytogenes EGDe::pPL2lux (silent vector) or L. monocytogenes EGDe::pPL2 luxP_HELP (expressing bacterial lux operon) (Table 1) were grown over night in BHI medium. Subsequently, cultures were grown in fresh BHI medium to a specified OD₆₀₀. At specified OD₆₀₀ (OD₆₀₀ 1 or 1.5) cells were transferred to 96-well round bottom plates for imaging. Luciferin (PerkinElmer) was added to L. monocytogenes EGDe::pPL2CBR^opt (for in vitro 150 μg/ml, for in vivo 150 mg/kg body weight) according to the manufacturer’s instructions 15 min prior to imaging. Measurements of bioluminescence were repeated with intervening layers of uniformly cut meat slices (one to five layers, thickness varied from 2 to 10 mm depending on meat type) tested over the surface of the plate. Three different meat types (to reflect different physical and chemical parameters), chicken, serrano ham, and fresh sliced raw beef steak were tested. Images were acquired with up to 3 min exposure and processed by Living Image Software v4.2.

All the cell culture procedures were performed as described previously (Bron et al., 2006). C2Bbe1 cells (CRC-2012; American Type Culture Collection), a clone of the Caco2 human adeno carcinoma cell line, were used for the cell invasion assay. Briefly, C2Bbe1 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) with 4.5 g/l Glutamax (Gibco Laboratories, Grand Island, NY, United States), 10% fetal bovine serum (Gibco), 1% (vol/vol) non-essential amino acids (Gibco), 1% (vol/vol) penicillin–streptomycin (Gibco), and 0.01 mg/ml human transferrin (Calbiochem) at 37°C in a 5% CO₂ environment. To perform the cell invasion assay, C2Bbe1 cells were trypsinized (Gibco), harvested by slow centrifugation at 500 × g for 5 min followed by resuspension in antibiotic free DMEM containing 10% fetal calf serum. Viable cells were counted using a standard trypan blue exclusion procedure. A known concentration of 1 × 10⁵ viable cells per well were seeded onto a 24- or 12-well flat bottom tissue culture plate (Sarstedt, Leicester, United Kingdom). Subsequently, the plates were incubated for 48–72 h at 37°C in a 5% CO₂ atmosphere until confluent. Listeria strains were grown overnight in BHI medium, re-inoculated into BHI medium and grown until the OD600 reached 1.0 (1 × 10⁹ CFU/ml). Subsequently, bacterial cells were washed with antibiotic free DMEM medium twice and added to the C2Bbe1 cells at a multiplicity of infection of 100:1 followed by incubation at 37°C for 1 h in 5% CO₂. The infection of C2Bbe1 cells was stopped after 1 h by adding gentamicin to a final concentration of 100 μg/ml, and cells were incubated further for 30 min to kill extracellular bacteria (Kim et al., 2004). Subsequently, the cells were washed three times with DMEM, and the levels of bioluminescence were measured using the IVIS Lumina II system (PerkinElmer). Immediately after imaging, the monolayers were lysed with 1 ml of ice-cold distilled water containing 0.01% triton (Sigma-Aldrich, St. Louis, MO, United States). The lysate was serially diluted in PBS and bacteria were plated onto BHI agar plates to determine the number of viable intracellular bacteria as CFU/ml.

For in vivo experiment, 8- to 10-week-old female BALB/c mice (Envigo, United Kingdom) weighing approximately 18–21 g were housed in pathogen free conditions at 22°C with a 12/12 light cycle. Animals were randomly assigned to groups (n = 4) and provided with mouse chow and water ad libitum. Before experiments, mice were afforded an adaptation period of at least 7 days. For bacteria preparation and infection, overnight cultures of L. monocytogenes were pelleted by centrifugation (7000 × g for 5 min), washed twice with PBS, and used to infect mice (n = 24) intraperitoneally at 2 × 10⁵ CFU in a total of 200 μl of PBS (Rea et al., 2004). The mice were followed for 72 h for imaging with 24 h intervals as described. After 24-, 48-, and 72-h post-infection, the infected mice were imaged following injection of luciferin (PerkinElmer, Inc., Boston, MA, United States) according to the manufacturer’s instructions at a dose of 150 mg/kg body weight administered via IP route as appropriate or imaged directly (mice infected with L. monocytogenes EGDe::pPL2luxP_HELP). After whole body imaging, mice were sacrificed, and post-mortem resected organs were imaged using the IVIS system. Images were acquired with up to 5 min exposure and processed by Living Image Software v4.2 or v4.3.1.

All animal procedures were performed in accordance with the national ethical guidelines prescribed by the Health Products Regulatory Authority (HPRA). Protocols and all the murine experiments were approved by the animal ethics committee of University College Cork (AERR #2010/003 and #2012/015).

All experiments were performed in triplicate at least two times. Data shown are for representative experiments. Statistical significance was based on either the Student’s t-test or ANOVA for comparison of groups (see Figures 2B, 6C). GraphPad Prism 5 or IBM-SPSS v22 was used for graphs and statistical analysis. Linear regressions were carried out using GraphPad Prism 5. In all analysis, p < 0.05 was considered as significant.

The main purpose of the work reported herein was the construction and functional analysis of the pPL2CBR^opt bioluminescence vector (Figure 1A) using the backbone of the pPL2 integrating plasmid. pPL2 contains a PSA listeriophage integrase gene and attachment site that directs site-specific single copy integration into tRNA^Arg locus on the L. monocytogenes chromosome (Lauer et al., 2002). In our lab, we have previously successfully created pPL2-derivatives, pPL2lux and pPL2luxP_HELP that allow BLI through expression of the bacterial lux operon (Bron et al., 2006; Riedel et al., 2007). Here we used a similar approach to create a vector expressing a CBR-luc with potential applications in deep tissue whole body imaging.

For this purpose, a codon optimized CBR-luc (CBR^opt) gene for optimal expression in L. monocytogenes was synthesized under the influence of a constitutive synthetic promoter, the highly expressed Listeria promoter P_HELP (Riedel et al., 2007). The synthesized construct was then sub-cloned into pPL2 using Xho1 and Pst1 restriction enzymes (Figure 1A). The sequence of the insert was verified and the novel vector was then transformed and integrated into the L. monocytogenes EGDe wild type strain and an engineered L. monocytogenes variant (EGDe^m) expressing a murine-specific InlA (Monk et al., 2010). The pPL2 vector contains a PSA listeriophage integrase gene and attachment site that directs site-specific single copy integration into the tRNA^Arg locus on the L. monocytogenes chromosome (Lauer et al., 2002). The successful integrants were then examined for integration using PCR and positive clones were screened phenotypically for light emission. A total of 10 candidate integrants were picked and screened for integration with results indicating that 9 (90%) were positive integrants. In parallel, we used the pPL2luxP_HELP and pPL2lux (silent vector) (Riedel et al., 2007) as controls for integration and subsequent analysis of luminescence.

We tested the stability of the L. monocytogenes EGDe::pPL2 CBR^opt and EGDe^m::pPL2CBR^opt strains in the absence of chloramphenicol. L. monocytogenes EGDe::pPL2CBR^opt and EGDe^m::pPL2CBR^opt were sub-cultured for at least 50 generations in BHI medium without antibiotic selection. Dilutions of the last passage of the cultures were plated and the presence of the integrated plasmids in the resulting colonies was evaluated by scoring chloramphenicol resistant colonies. Our results indicated that all 50 colonies tested after 50 generations were found to retain chloramphenicol resistance indicating that the integrating plasmid was stably maintained without antibiotic selection pressure. Furthermore, at least 10 random chloramphenicol-resistant colonies were also tested phenotypically for light emission and the results indicated that all 10 tested colonies were bioluminescent.

Previous work has indicated that integration of lux-expressing pPL2 constructs does not influence bacterial growth rate relative to wild type bacteria (Riedel et al., 2007). In the current study we evaluated the growth behavior of L. monocytogenes EGDe::pPL2CBR^opt relative to EGDe::pPL2luxP_HELP (Riedel et al., 2007) and EGDe::pPL2 (Bron et al., 2006). Strains were grown in parallel under standard conditions (Figure 1B). Our data indicated that the growth of L. monocytogenes EGDe::pPL2CBR^opt was normal relative to the other strains tested (Figure 1B). Furthermore, the growth rates of L. monocytogenes EGDe::pPL2CBR^opt were also not affected in BHI (in which pH was adjusted to 7.4) or LB medium (data not shown).

Listeria monocytogenes EGDe is the most commonly used laboratory strain for the analysis of listerial pathogenesis (Glaser et al., 2001). However, there is increasing interest in the analysis of wild type clinical isolates in order to determine strain specific differences in pathogenesis (Maury et al., 2016). For this purpose, three clinical strains (Table 1), all originating from patients who suffered meningitis and isolated from cerebrospinal fluid (CSF) were transformed with pPL2CBR^opt. Our results indicated that all three wild type clinical strains could be successfully tagged with our newly developed plasmid (Supplementary Figure S1). After 30 passages of these strains in vitro 50 colonies were tested and all retained chloramphenicol resistance indicating that the system is as stable as observed in the EGDe laboratory strain. We also determined that the in vitro growth rate of these wild type field isolates in BHI broth was not affected by the presence of the integrated plasmid (data not shown). We transformed the clinical isolates with pPL2luxP_HELP in order to provide a comparison with the CBR^opt system. Results indicated that bioluminescent profile followed the same light emission pattern as observed for the EGDe strain as described below (Supplementary Figure S1).

We determined the bioluminescence profile of L. monocytogenes EGDe::pPL2CBR^opt over time. L. monocytogenes EGDe::pPL2 CBR^opt was grown in BHI broth for 24 h to monitor bioluminescence (Figures 2A,B). No bioluminescence was detected at the initial stage of growth. The bioluminescence profile increased as bacteria grew and bioluminescence reached a peak following 6 h of bacterial growth and remained relatively high during stationary phase. We directly compared the bioluminescence of L. monocytogenes EGDe::pPL2CBR^opt with the L. monocytogenes EGDe::pPL2luxP_HELP strain as both bioluminescent vectors are constructed using the pPL2 backbone with expression of luciferase under the same P_HELP promoter system (Riedel et al., 2007). Bioluminescence signals of bacterial luciferase (L. monocytogenes EGDe::pPL2luxP_HELP) were higher during initial growth as compared to the isolate expressing CBR^opt. However, the bioluminescence signal from L. monocytogenes EGDe::pPL2luxP_HELP diminished quickly as bacteria entered the stationary phase and the overall bioluminescence expression pattern for this strain was similar to our previously reported data (Riedel et al., 2007). A decrease in the bioluminescence signal during stationary phase may be a feature of light expression from the luxABCDE system as the phenomenon has also been described for other Gram-positive bacteria including Streptococcus pneumonia (Francis et al., 2001) and Lactobacillus casei (Oozeer et al., 2005). It is possible that the diminished Lux signal during stationary phase may be related to decreased metabolic recycling of FMN to FMNH₂ (an essential component for bacterial Lux activity), and ATP production (Francis et al., 2001; Andreu et al., 2010) This hypothesis is partially supported by a previous study which demonstrated a 100-fold increase in signal after providing a gene (frp) encoding a NADPH-FMN oxidoreductase in-trans with luxAB enzymes in a yeast model system (Szittner et al., 2003). However, this limitation would not influence the activity of CBR, which in our system exhibits relatively robust activity in late exponential phase and stationary phase.

In order to measure light intensity per cell, bioluminescence signals were measured in parallel with determination of the total number of viable bacterial CFU. Our results indicated that CBR^opt signals were directly correlated with CFU (r² = 0.9379; Figure 2C). A similar positive correlation was observed between the OD₆₀₀ and total flux during growth in medium (Figure 2D). The strong positive correlation between CFU and total flux indicates that luminescence from CBR-luc tagged bacteria can successfully be used for accurate quantification of bacterial numbers during growth.

In order to directly compare signal intensity each luminescent strain was serially diluted until no luminescence signals were detectable (Figures 3A,B) and the CFU/ml of each representative dilution was calculated in order to determine relationship of total flux/CFU. The bioluminescence profile of both lux and CBR-luc luciferase reporters were indeed dependent on the number of the luminescent bacteria. Bioluminescence intensity was correlated with the number of bioluminescent bacteria both for EGDe::pPL2luxP_HELP and EGDe::pPL2CBR^opt during growth over time in BHI medium (Figure 3B) and the overall analysis indicated that CBR^opt produces greater flux/CFU when compared with bacterial luciferase during growth in medium (p > 0.05) (Figure 3C).

CBR-luc produces light of a wavelength (∼600 nm) that is within the red region of the visible light spectrum. As red light has greater penetrating power in animal tissues, we predicted that light generated by CBR-luc would penetrate mammalian tissues more effectively than that of lux (in the blue spectrum at around 480 nm). We compared the ability of the CBR^opt and Lux luciferase signals to penetrate through varying depths of tissue using a previously described in vitro assay (Mason et al., 2016). Layers of cut meat slices were used to mimic varying levels of tissue depth. An aliquot of 50 μl of cells were deposited into a round bottom 96-well plate and light measurements were repeated following the addition of layers of meat slices to generate tissue depth as indicated in Figure 4. As expected a decrease in light detection was associated with an increase in tissue depth. Signal transmission and total flux of bacterial luciferase (L. monocytogenes EGDe::pPL2luxP_HELP) quickly declined as the thickness of the tissue increased as compared to CBR^opt (L. monocytogenes EGDe::pPL2CBR^opt). In particular L. monocytogenes EGDe::pPL2CBR^opt was clearly visible through multiple layers of raw beef steak whereas the penetrating power of light generated in L. monocytogenes EGDe::pPL2luxP_HELP was poor and below detectable limits (Figures 4A,B).

The pPL2luxP_HELP vector has previously proved effective for the imaging of L. monocytogenes infection in cultured mammalian cells (Riedel et al., 2007). In the current study, we investigated the bioluminescence profile of L. monocytogenes EGDe::pPL2CBR^opt during a cell invasion assay in order to determine the possibility of using this vector for monitoring and detection of EGDe in cell culture. For this purpose, L. monocytogenes EGDe::pPL2CBR^opt, L. monocytogenes EGDe::pPL2luxP_HELP and L. monocytogenes EGDe::pPL2lux were used to infect C2Bbe1 cells for 1 h and the levels of bioluminescence and number of the intracellular bacteria were determined (Figures 5A,B). The efficacy of invasion did not differ between the strains and bacteria were isolated at similar levels from each well. However, the bioluminescence intensity indicated that L. monocytogenes EGDe::pPL2CBR^opt produced a consistently brighter signal and significantly higher flux than L. monocytogenes EGDe::pPL2luxP_HELP in the context of in vitro infection (Figures 5A,B). This suggests that the pPL2CBR^opt vector may have applications for BLI of Listeria during infection assays, perhaps with the potential for use in high throughput assay systems.

Murine infection is widely used for the analysis of factors contributing to the pathogenesis of L. monocytogenes (Lecuit, 2007; Cossart and Toledo-Arana, 2008). A number of pathogens such as Trypanosoma brucei (Van Reet et al., 2014), Mycobacterium tuberculosis (Chang et al., 2014), and Lactobacillus species (Daniel et al., 2015) have been previously tagged with CBR reporters that provided enhanced whole body images in murine infection model. Given the improved signal intensity mediated by CBR during in vitro growth, in particular during late exponential and stationary-phase, we examined the signal intensity from L. monocytogenes EGDe::pPL2CBR^opt during murine infection. For this purpose, three groups of 12 mice were infected with murinized L. monocytogenes EGDe^m::pPL2CBR^opt, murinized L. monocytogenes EGDe^m::pPL2luxP_HELP, and murinized EGDe^m::pPL2lux with a standard inoculum (2 × 10⁶ CFU/ml) via the intraperitoneal route. Four mice were left uninfected to serve as the control group. Four mice from each group were than imaged at 24 h intervals post-infection until 72 h. For the mice infected with L. monocytogenes EGDe^m::pPL2CBR^opt luciferin was injected intraperitoneally 10 min prior to imaging. After imaging, mice were sacrificed and the resected organs (liver, spleen, and kidneys) were imaged for bioluminescence and further processed for counting the number of viable bacteria. Results of the whole body imaging of mice indicated that no bioluminescence was observed 24 h post-infection in any of the strains. No bioluminescence signals were detectable after dissecting the mice and imaging individual organs at this time point. This lack of bioluminescence correlated well with the lower number of viable bacteria recovered from the resected organs of the infected mice. Of these, the highest number of CFU was recovered from the spleen where CFU/ml were in the range of 1 × 10³.

Notably, mice infected with L. monocytogenes EGDe^m::pPL2CBR^opt produced detectable bioluminescence signals at 48 h post-infection, while no signals were detected from mice infected with L. monocytogenes EGDe^m::pPL2luxP_HELP or the L. monocytogenes EGDe^m::pPL2lux control. We acknowledge that there was some variability in observable signals between individual mice (when imaged at the specified sensitivity) as noted in previous studies by other groups (Poulsen et al., 2011). The number of CFU in these animals reached approximately 1 × 10⁶ CFU/ml in the spleen and liver. Importantly, there was no significant difference in CFU/ml among the infected strains recovered from the infected mice post-mortem in resected organs (results not shown). Interestingly, at 48 h infection bioluminescent signals were detected in the resected liver and spleen of mice infected with L. monocytogenes EGDe^m::pPL2luxP_HELP, but no signals were detectable during whole body imaging, indicating that these signals were attenuated during whole body imaging, potentially as a consequence of the low penetrating power of the shorter wavelength light emitted by Lux. In contrast, bioluminescence signals were observed during whole body imaging of mice infected with L. monocytogenes EGDe^m::pPL2CBR^opt at 48 h post-infection in addition to a clear signal from post-mortem resected organs (Figure 6).

In a tandem study using infection with wild type (non-murinized) EGDe we also noted a clearly enhanced signal from mice infected with CBR^opt labeled bacteria relative to mice infected with Lux-tagged EGDe (Supplementary Figure S2). In this experiment, bacterial burdens in the internal organs were comparable between the groups though the infectious load in the organs was noticeably lower than mice infected with EGDe^m (not shown). Again this suggests that the signal from CBR^opt may provide greater sensitivity at low bacterial loads.

Overall, the ability to detect luminescence signals in mice infected with L. monocytogenes EGDe^m::pPL2CBR^opt at 48 h suggested that pPL2CBR^opt provides for increased sensitivity in the context of whole body imaging. At higher bacterial numbers on day 3 of infection the difference in signal intensity in whole body imaging between the Lux and CBR-labeled bacteria was less evident (Figure 6).