Animal research during this study was conducted in accordance with recommendations of approval No. 216/2012 and following the guidelines provided by the Veterinary Office of the Public Health Department of the Canton of Zurich (Switzerland). The number of dead larvae post infection was determined at various time points visually based on the lack of a heartbeat. Experiments were carried out until 72 h post infection (hpi) and at the end of the experiments embryos that were alive were euthanized with an overdose of 4 g L⁻¹ buffered tricaine. Usually, with the evaluation of distress and pain by behavioral observations, embryos were euthanized by prolonged immersion in overdose of tricaine solution and were left in the solution for at least 10 min after cessation of opercular movement. Since pain sensitivity has not been developed at these earlier stages (4 dpf−7 dpf), this is not contemplated as a painful technique. The maximum age attained by the embryos throughout investigation was 72 hpf and embryos had not yet reached free feeding stage.

The WT and csp deletion mutants L. monocytogenes EGDe used in this study are described in Table 1. All csp mutants were constructed in-frame as previously described (Schmid et al., 2009). Green fluorescent protein (GFP) expressing derivatives of the strains were generated through site specific PSA-integrase mediated single copy integration of the pPL3-eGFP plasmid (Shen and Higgins, 2005) into the tRNA-Arg locus (Lauer et al., 2002). Sixteen hours secondary stationary phase cultures that were confirmed through optical density measurements and viable cell counting were used in this study. Primary cultures were prepared by inoculating 10 mL BHI (BHI; Oxoid, Hampshire, UK) broth and growing for 16 h at 37°C and 150 rpm. Secondary cultures were subsequently prepared by inoculating 10 ml BHI with these primary cultures (1:1,000) and growing for 16 h at 37°C and 150 rpm. Optical density measurement (OD₆₀₀) and viable cell count based growth curves conducted for each strain showed that all secondary cultures grown in this way and subsequently used for experiments were in the stationary growth phase stage. The pPL3-eGFP integrated strains were similarly grown on BHI agar and broth medium supplemented with erythromycin at 5 μg ml⁻¹.

THP-1 cells (ATCC TIB-202) were maintained in T75 tissue culture flasks (TPP- Techno plastic products, Switzerland) and grown to confluence in RPMI 1640 medium (RPMI; Sigma Aldrich, Germany) containing 0.3 g l⁻¹ L-glutamine, 2 g l⁻¹ sodium bicarbonate supplemented with 10 mM HEPES (Sigma Aldrich, Buchs, Switzerland), 1 mM sodium pyruvate (Sigma Aldrich, Buchs, Switzerland), 4.5 g l⁻¹ glucose and 10 % fetal bovine heat-treated (56°C, 30 min) serum (Sigma Aldrich, Buchs, Switzerland Germany) and incubated at 37°C with 5% CO₂.

THP-1 cells were seeded at a density of 10⁵ cells per well in 24-well tissue plates and incubated for at least 24 h (37°C and 5% CO₂) in RPMI 1640 containing 0.1 μg ml⁻¹ of phorbol 12-myristate 13-acetate (PMA; Sigma Aldrich, Buchs, Switzerland) to induce stable differentiation of the THP-1 monocytes into macrophages. The PMA containing medium was removed and the cells washed once using 500 μl RPMI 1640. THP-1 macrophages were infected at ratio of 10 bacteria organisms per cell using Listeria that had been cultivated as described above and diluted in RPMI 1640. After 45 min (37°C and 5% CO₂) of incubation the macrophages were washed twice using DPBS (500 μl) before incubation for 45 min in RPMI 1640 (500 μl) containing 50 μg ml⁻¹ gentamicin to kill extracellular bacteria. The macrophages were washed once using 0.5 ml of antibiotic free RPMI 1640 after this step and then incubated in 0.5 ml RPMI 1640 that contained 10 μg ml⁻¹ gentamicin. Infections immediately stopped at this point represented time point t₀ and the initial macrophage internalized inoculum. For the rest of the samples incubation was continued and infections were stopped after 6 (t₆) and 24 (t₂₄) h of macrophage infection. To stop infections the macrophages were washed twice using 0.5 ml DPBS and then lysed using 0.5 ml of DPBS plus 0.5% Triton X-100. The cell lysates were 10-fold serially diluted, plated out on plate count agar plates (PC agar, Sigma) and incubated for 24 h at 37°C before enumerating the number of viable intracellular bacteria. Intracellular amounts of bacteria (IBC) CFUs (colony forming units) present at 6 (t₆) and 24 (t₂₄) h of THP-1 macrophage infection were determined and expressed as a percentage relative to the initial intracellular inoculum found in macrophages at time point zero (t₀) as follows: (IBCt_n – IBCt₀/IBCt₀) × 100, where IBCt₀ and IBCt_n are the intracellular bacterial CFUs determined at time points zero (t₀) and n (t₆ and t₂₄), respectively.

GFP expressing EGDe WT and csp mutant strains (Table 1) were used in macrophage infection assays with subsequent microscopic analysis. In each well of the 24 well plate, a sterile glass coverslip (13 mm in diameter, Menzel-Gläser) was placed and THP-1 cells were seeded on the coverslips. Cell culture, PMA activation and infection assays were carried out as mentioned above. At each indicated time point (t₀, t₆, and t₂₄), THP-1 macrophages adhered to coverslips were gently washed twice with DPBS and fixed with 4% paraformaldehyde (PFA, Sigma) at 4°C for 15 min. After fixation, the PFA was removed, the macrophages were gently washed twice with DPBS to remove residual PFA, and the samples were stained for 1 h at room temperature with Hoechst Dye (Life Technologies, Zug, Switzerland) and Concanavalin A Alexa Fluor® 594 Conjugate (Life Technologies). After staining, the macrophages were washed 3–5 times with DPBS, mounted on glass slide using Fluoromount (Sigma) mounting medium, air dried in the dark and imaged using a Leica TCS SP5 Confocal microscope (63x or 40x oil-immersion objective, excitation at 405 nm for Hoechst, 488 nm for GFP and 594 nm for Concanavalin A).

Zebrafish (Danio rerio) strains used in this study were wik lines. Adult fish were kept at a 14/10 h light/dark cycle at a pH of 7.5 and at 27°C. Eggs were obtained from natural spawning between adult fish which were set up pairwise in individual breeding tanks. Embryos were raised at 28°C in petri dishes containing E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, and 0.33 mM MgSO₄) supplemented with 0.3 μg/ml of methylene blue. From 24 h post fertilization (hpf), 0.003% 1-phenyl-2-thiourea (PTU; Sigma Aldrich, Buchs, Switzerland) was added to prevent melanin synthesis. Staging of the embryos was performed according to Kimmel et al. (1995). All zebrafish experiments were conducted with the approval (no. 216/2012) from the Veterinary Office, Public Health Department, Canton of Zurich (Switzerland).

Injections were performed using borosilicate glass microcapillary injection needles (Science Products, 1210332, 1 mm OD × 0.78 mm ID) and a PV830 Pneumatic PicoPump (World Precision Instruments). Bacteria for microinjection experiments were harvested from stationary phase BHI bacteria cultures by centrifugation at 5,000 × g for 10 min and washed in 10 ml Dulbecco's Phosphate-Buffered Saline (DPBS) before the cells were resuspended in DPBS and appropriate dilutions prepared. Two-day post fertilization (dpf) embryos were manually dechorionated and anesthetized with 200 mg/l buffered tricaine (Sigma Aldrich, Buchs, Switzerland, MS-222) prior to injections. Subsequently, the embryos were aligned on an agar plate and injected with 100 CFU (range: 90–136 CFU) in 1–2 nl volume of a bacterial suspension in DPBS into the yolk sac or blood circulation via the caudal vein close to the urogenital opening. Prior to injections the volume of the injection suspension was adjusted by injecting a droplet into mineral oil and measuring its diameter over a micrometer scale bar. The number of CFU injected was determined by direct microinjection of a DPBS droplet on agar plates and confirmed by disintegrating five embryos individually immediately after microinjection (0 hpi) and plating the lysates on LB agar. Post injection the infected embryos were allowed to recover in a petri dish in fresh E3 medium for 15 min. To follow the infection and mortality kinetics, embryos were transferred into 24-well plates (one embryo per well) in 1 ml E3 medium per well, incubated at 28°C and observed for signs of disease and survival under a stereomicroscope twice a day. For survival assays after infection, the number of dead larvae was determined visually based on the absence of a heartbeat. At each time point, five embryos or larvae were collected, euthanized, and individually treated for bacterial enumeration. For subsequent microscopic analyses larvae were euthanized with an overdose of 4 g/l buffered tricaine and transferred into respective buffers and fixatives.

The larvae were transferred to a 1.5 ml Eppendorf tube containing 1 ml DPBS supplemented with 1% Triton X-100 and disintegrated by repeated pipetting and vortexing for 3 min. Subsequently, 100 μl of this mixture was plated onto BHI selective plates (i.e., erythromycin 5 μg l⁻¹ for strains harboring pPL3::GFP) and the plates were incubated up to 48 h at 37°C.

Reverse transcription quantitative-PCR (RT-qPCR) was used to determine transcript levels of selected genes in EGDe WT and the csp mutant strains. Bacteria grown to stationary phase in BHI as described above or recovered from THP-1 macrophages that were infected for 6 h with Listeria, and then washed and lysed as described above were used. L. monocytogenes cells (10⁹ CFUs) in 1 ml of stationary phase BHI broth cultures prepared as described above were harvested in RNA protect Bacteria reagent (Qiagen AG, Hombrechtikon, Switzerland) and resuspended using 0.5 ml of the RNeasy Plus Mini Kit (Qiagen AG, Hombrechtikon, Switzerland) lysis buffer. Bacteria recovered from macrophage lysates by centrifugation were similarly harvested in RNA-protect Bacteria Reagent and subsequently re-suspended in the lysis buffer. L. monocytogenes cells resuspended in the lysis buffer were transferred into beads in MagNA lyser tubes and mechanically disrupted using the MagNA Lyser Instrument (Roche Molecular Diagnostics, Rotkreuz, Switzerland). RNA was subsequently isolated from the lysates following the RNeasy Plus Mini Kit protocol. Purified RNA was quantified using a Nanodrop (Thermo Scientific, USA) and its quality was verified using a BioAnalyzer (Agilent technologies, USA). cDNA was synthesized from 100 ng of RNA samples with RNA integrity numbers (RINs) of 7 or above. The Quantitect Reverse Transcription Kit (Qiagen AG, Hombrechtikon, Switzerland) was used. Residual DNA contamination of RNA samples was ruled out through the inclusion of no RT controls in the analysis. 2.5 ng of the cDNA were used as templates for the RT-qPCR. Primers listed in Table 2 and the SYBR green I kit (Roche Molecular Diagnostics, Penzbrug, Germany) were used for the RT-qPCR in a LC480 instrument (Roche Molecular Diagnostics, Rotkreuz, Switzerland). Relative cDNA quantification was performed using the Light Cycler 480 Relative Quantification Software (Roche Molecular Diagnostics). Amounts of the transcript amounts were normalized to 16S rRNA as reference gene. Using a delta-delta CT approach, transcript levels of each gene were expressed relative to the values of a calibrator mRNA sample, which was derived from stationary phase EGDe WT culture grown in BHI broth.

Virulence protein expression was assessed in both extracellular and intracellular grown L. monocytogenes. To assess protein expression in extracellular bacteria, L. monocytogenes were cultured to stationary phase in BHI (30 ml; 16 h at 37°C and 150 rpm) as described above, standardized to OD600 of 1.0 by dilution in BHI broth and the cells were collected by centrifugation (10,000 g for 2 min) at 4°C. To assess intracellular virulence protein expression THP-1 macrophages infected with Listeria for 6 h. Post infection the THP-1 macrophages were washed and lysed to release intracellular bacteria as described above. L. monocytogenes were recovered from the THP-1 macrophage lysates by centrifugation (10,000 g for 10 min) at 4°C.

To extract proteins, L. monocytogenes pellets were resuspended in RIPA lysis buffer (1 ml) containing a protease inhibitor cocktail (Cell Biolabs Inc, San Diego, USA) and mechanically disrupted (2 × 6,500 rpm for 1 min) in the MaGNA lyser instrument (Roche Molecular Diagnostics). The resulting L. monocytogenes lysates were cleared by centrifugation (10,000 g for 10 min at 4°C) and protein concentration in the supernatants was determined. Equal amounts of proteins (20 μg) from each sample were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE; 4–20%) and transferred onto PVDF membranes. The membranes were probed with primary rabbit polyclonal antibodies raised against PrfA (1:5,000), Mpl (1:5,000) and LLO (Diatheva, Fano PU, Italy; 1:1,000), as well as primary mouse monoclonal antibodies directed against ActA (Abanova, Taipei, Taiwan, 1:1,000) and P60 (Adipogen, San Diego, USA; 1:2,000). A HRP-conjugated anti-rabbit IgG (Sigma-Aldrich, Buchs, Switzerland; 1:2,000) secondary antibody was subsequently used to probe all blots except the ActA and P60 blots, which were probed using an HRP-conjugated anti-mouse IgG (Sigma-Aldrich, Buchs, Switzerland; 1:100,000) secondary antibody. The primary and secondary antibodies were diluted in 1% skimmed milk solution prepared in tris buffered saline with tween 20. The densitometric quantification of the proteins was carried out using the ImageJ software (National Institute of Health, USA. The intensities of PrfA, Mpl, LLO, and ActA bands obtained by chemiluminescence were normalized using the intensity of P60 from the corresponding sample.

To compare PI-PLC activities, the colonies of L. monocytogenes strains were streaked on ALOA (Agar Listeria according to Ottaviani and Agosti) plates, incubated at 37°C and visually examined after 48 h for the zone of opacity. Aggregation of Listeria strains was compared as previously described (46). In short, 16 h stationary phase BHI cultures of WT and csp mutant L. monocytogenes strains were adjusted to OD₆₀₀ of 3.5 in BHI. The adjusted cultures were statically incubated at 37°C for 24 h and monitored through visual and microscopic examination for aggregation. Aggregation was indicated by decreasing cultural supernatant optical density and the development of bacterial sediments. To determine the optical density, aliquots were collected ~1 cm from the top of each sample at defined time intervals to measure OD₆₀₀.

Swarming motility was compared by spotting 16-h stationary phase BHI cultures from each strain (5 μl) onto the surfaces of 0.25% BHI agar plates and incubating 48 h at 25°C. For electron microscopic examination bacteria colonies grown overnight at 25°C on a 0.25% BHI agar plate were picked up using an inoculation loop and transferred to an Eppendorf tube containing 100 μl of 2.5% glutaraldehyde and incubated for at least 30 min at room temperature to fix. Further, 1:1 suspensions were prepared using 0.1 M sodium phosphate buffer and bacteria-glutaraldehyde mix. These suspensions were then transferred onto 150 mesh copper grids with Formvar carbon film (Electron Microscopy Sciences, USA) and incubated at room temperature for 2 min for adhesion. Following adhesion, bacterial flagella were negatively stained with uranyl acetate for 10 s and with 1:1 suspension of uranyl acetate and water for 20 s, and observed in an electron microscope.

Statistical analyses were carried out using JMP software (Version 12.1.0, SAS Institute Inc., NC, USA). All experiments presented were performed independently at least three times. One way ANOVA with post-hoc Tukey HSD tests were used to assess statistical significance of differences relative to the WT as well as between the different csp mutant strains in macrophage and zebra infection as well as RT-qPCR assays. Kaplan Meier survival analysis and statistics for experiments with zebrafish were done with GraphPad Prism 7 software (GraphPad Software, United States). p < 0.05 were considered to be statistically significant.

Since the survival and multiplication of L. monocytogenes within target host cells is crucial step for successful host infections, we initially examined the functional contribution of Csps in this bacterium during human macrophage infection in cell culture. The survival and growth of WT and csp gene deletion mutant strains of L. monocytogenes EGDe during the infection of human derived THP-1 macrophages was compared. Based on comparison of viable bacteria CFUs that were recovered inside macrophages at infection time point zero (t₀) we determined that there are similar levels of bacteria that were initially internalized into THP-1 macrophages for the WT strain as well as all the csp mutants (Figure S1). Subsequent monitoring of changes in intracellular bacterial CFUs conducted after 6 (t₆) and 24 (t₂₄) h of macrophage infection however showed that the complete deletion of all the three csp genes (ΔcspABD), as well as the presence of only single csp genes (ΔcspBD, ΔcspAD, and ΔcspAB) is associated with impaired survival and growth of L. monocytogenes in THP-1 macrophages (Figure 1A). In samples examined at 6 h (t₆) post infection, the WT strain bacteria levels inside macrophages had increased by an average of 33.3% when compared to the inoculum found inside macrophages at 0 (t₀) h of infection. But a ΔcspABD mutant deleted in all the three csp genes only displayed limited growth with its levels inside macrophages having increased by an average of only 2.8% at this time point. At 24 h (t₂₄) post infection, the WT strain levels in macrophages showed further increase (40.5% increase relative to the t₀ levels), whereas the ΔcspABD mutant could not be detected in THP-1 macrophage infected for 24 h. This suggests that both long-term growth and survival of L. monocytogenes inside human macrophages are significantly impaired without Csps (Figure 1A).

A caveat in assessing the functional role of individual csp genes in L. monocytogenes using single gene deletion mutant backgrounds is that some phenotypes might be masked due to functional redundancy that exist between the three csp genes found in this bacterium. To overcome this caveat and allow us to confirm as well as assess the level of phenotypic contributions of the individual csp genes during THP-1 macrophage infections, three mutant strains, which are deleted in two leaving only one of the three csp genes intact were analyzed. To this end the ΔcspBD, ΔcspAD, and ΔcspAB mutant strains expressing CspA, CspB, and CspD, respectively, were compared with respect to growth and survival in human macrophages. We found that the survival and growth phenotypes achieved by the three single Csp producing strains in THP-1 macrophages although variable were all significantly above those observed for ΔcspABD mutant, which lacks all the three csp genes (Figure 1A). These observations thus did not only confirm that all three csp genes were functionally relevant but that they also contribute to varying extents toward L. monocytogenes survival and growth in human macrophages. We found that none of the three Csps when produced alone is capable of restoring the WT phenotypic levels without the other two csp genes (Figure 1A). Overall there was a clear hierarchical trend of WT > ΔcspAD (CspB) > ΔcspAB (CspD) > ΔcspBD (CspA) > ΔcspABD (Figure 1A; P < 0.05) observed with respect to the intracellular bacteria CFU levels reached inside the THP-1 macrophages at 6 and 24 h post infection (Figure 1A).

We similarly also used GFP expressing bacteria and fluorescent microscopy to qualitatively follow the intracellular fates of WT and csp mutant strains during THP-1 macrophage infections. This approach despite being qualitative revealed that intracellular multiplication and survival of the ΔcspABD mutant without Csps was impaired compared to the WT strain during human macrophage infection (Figure 1B). Samples collected immediately post macrophage infection at the t₀ time point showed no microscopically visible differences with respect to macrophage internalized bacteria observed inside infected THP-1 macrophages between the WT strain and the different csp mutants (Figure 1B). Such observations were therefore consistent with viable bacterial counts results that had quantitatively shown no significant differences in the initial macrophage entry between WT and the csp mutant strains (Figure S1). Significant differences were however observed between the WT and the csp mutant strains upon examination of samples collected 6 (t₆) and 24 (t₂₄) h after infection (Figure 1B). At both these time points samples infected with the WT strain displayed relatively higher amounts of bacteria inside macrophages compared to their t₀ samples as well as to samples of macrophages that were infected with different csp mutants. The WT bacteria found inside the macrophages at these points occurred in large clumps consistent with an ongoing multiplication inside the THP-1 macrophages. In contrast, there were relatively fewer and visibly smaller intracellular bacteria clumps observed for the ΔcspABD and the three single Csp strains at 6 h (t₆) post infection compared to WT. This was indicative of slower intracellular replication of these mutants compared to WT inside the human macrophages. Moreover, consistent with the quantitative viable bacterial count results there were no ΔcspABD detected inside THP-1 macrophages infected for 24 (t₂₄) h with this mutant (Figure 1B). Although all three single Csp strains could be detected and also displayed microscopic evidence of ongoing intracellular multiplication at 6 (t₆) as well as 24 (t₂₄) h post macrophage infection, the levels of intracellular replicating bacteria observed inside the macrophages for all were visibly lower than the WT strain but more than those observed for the ΔcspABD mutant.

Overall our analysis of the different csp mutants using this in vitro cell culture based virulence model thus indicates that all three Csps are functionally relevant but to varying degrees in promoting L. monocytogenes survival and growth during human macrophage infection. Moreover, the fact that none of the three Csps when expressed alone can restore these phenotypes to WT levels indicates that the functional activities of all the three csp genes is necessary for maximal survival and growth of this bacterium during human macrophage infection.

The impaired survival and growth observed for the different csp mutants in human macrophages prompted us to further examine the functional relevance of Csps to L. monocytogenes virulence using a zebrafish embryo based multicellular in vivo infection model. Zebrafish embryos were infected with WT and csp mutants of L. monocytogenes EGDe strains by injecting into the blood stream and the capability of the different strains to induce mortality of the infected zebra fish embryos was monitored over a 3 day (72 h) infection period. Survival curves generated showed that while all WT strain infected zebrafish embryos succumbed to the infection within 2 days of infection none of those infected with ΔcspABD infected embryos had died even after 3 days post injection (dpi) (Figure 2A). All embryos infected with the three strains harboring single csp genes also eventually succumbed but 100% mortality level in all the three strains was delayed occurring a day later than WT at 3 dpi. A comparison of the level of mortality associated with all strains at 1 and 2 dpi showed a hierarchical trend of WT > ΔcspAD (cspB) > ΔcspAB (cspD) > ΔcspBD (cspA) > ΔcspABD; P < 0.05), which was similar to the survival and growth efficiency trends observed using the THP-1 human macrophage cell infection model (Figure 2A). To additionally assess for differences observed in virulence we monitored for trends in the bacterial loads associated with the WT and csp mutant strains during the course of the zebra fish embryo infections. Bacterial CFU levels within infected zebrafish embryos at 0, 1, and 2 dpi were determined. At 0 dpi similar levels of bacterial loads were determined in embryos infected with the WT strain and all the csp mutants. In 1 and 2 dpi samples we detected similar progressive increase in the bacterial loads relative to the initial bacterial loads found at 0 dpi (Figure 2B). More importantly there were no significant differences in overall bacterial loads achieved between the WT strain and the three strains possessing single csp genes. In contrast, the bacterial loads in ΔcspABD mutant infected embryos also increased on 1 dpi relative to 0 dpi, but the bacterial load increases were lower compared to those observed for WT and single Csp strain infected embryos. Furthermore, in contrast to the bacterial loads of the other strains that increased, the ΔcspABD bacterial loads showed a significant reduction between 1 and 2 dpi, indicating that the long-term survival of the ΔcspABD mutant was also compromised within infected zebrafish embryos. Overall the ΔcspABD bacterial loads achieved at 1 and 2 dpi were significantly lower than those of the WT and single Csp strains. Overall our results thus indicated that the three Csps while all functionally relevant also showed variable capacities in enabling full L. monocytogenes virulence expression using this in vivo multicellular infection model based on zebrafish embryos.

In seeking for a possible explanation of the virulence impairment observed in the csp mutants we next considered the impact Csp loss on the expression of some key virulence proteins in L. monocytogenes. In our previous studies we could show that a strong reduction of LLO protein expression and hemolytic activity was imposed by deletion of csp genes in L. monocytogenes (Schärer et al., 2013). To examine the impacts of Csp loss on the expression of the other virulence proteins we combined Western blot analysis and densitometry based protein band quantification and compared PrfA, LLO, Mpl, and ActA production between the different csp mutants and the parental WT strain (Figure 3, Figures S2–S4). BHI cultured and THP-1 macrophage grown (6 h post macrophage infection) bacteria were used for this analysis. To provide a control and reference protein for the analysis we also determined and quantified the amounts of the secreted P60 protein. Despite all strains showing similar P60 levels, significant differences in PrfA, LLO, Mpl, and ActA levels were detected among the different csp mutants as well as in comparison to the WT strain. A ΔcspABD mutant lacking all three csp genes in particular produced significantly lower amounts of all the examined virulence proteins compared to the WT strain and all three single Csp strains (Figures 3B,C). The three single Csp producing strains in turn all displayed lower levels of these virulence proteins relative to the WT strain. Comparing the single Csp producing strains to each other, we found that PrfA and LLO levels were highest in ΔcspAB (cspD) followed by ΔcspAD (cspB) strains, whereas the ΔcspBD (cspA) consistently showed lowest amounts for these proteins. On the other hand, Mpl levels were similar (BHI cultured bacteria) to slightly different (THP-1 grown bacteria) between the ΔcspBD (cspA) and ΔcspAD (cspB) strains, but the levels in these two strains were significantly below those of the ΔcspAB (cspD) strain. No ActA proteins were detected in the ΔcspABD and ΔcspAB (cspD) strains, whereas ActA amounts found in ΔcspBD (cspA) and ΔcspAD (cspB) strains were below WT levels. Among the single Csp producing strains the ΔcspAD (cspB) strain showed significantly higher ActA levels than the ΔcspBD (cspA) strain (Figures 3B,C).

The combined results from qualitative PlcA (PI-PLC) activity analysis and Western blot quantification of PrfA, LLO, Mpl, and ActA proteins indicated that the loss csp genes significantly reduces virulence protein production. In order to access if reduced virulence protein production is also reflected at transcript level within the different csp mutants we used RT-qPCR and compared the prfA, hly (LLO), mpl, and actA mRNA levels between the WT and the csp mutants. In RNA isolated from both BHI and THP-1 macrophage grown bacteria, we found that consistent with low PrfA, LLO and Mpl levels observed in ΔcspABD, the prfA, hly, and mpl mRNA levels in this mutant were also significantly lower compared to the WT strain (Figure 4). A comparison with the single Csp strains also showed that the prfA, hly, and mpl mRNA levels in ΔcspABD are generally also lower compared to these strains although in some instances the differences were not statistically significant (Figure 4). The prfA, hly, and mpl mRNA levels in the single Csp producing strains on the other hand were for most cases also significantly lower than the WT strain, which is consistent with their reduced PrfA, LLO, and Mpl protein levels compared to the WT strain. An exception to such trends was however observed in case of hly and mpl mRNA levels, which were not significantly different compared to the WT strain in RNA derived from BHI cultured bacteria of the cspB harboring ΔcspAD mutant (Figure 4). Although containing lower PrfA, LLO, and Mpl protein amounts compared to the cspD harboring ΔcspAB mutant, the cspB harboring ΔcspAD mutant however contained higher prfA, hly, and mpl mRNA amounts (Figures 3, 4). More surprising was the fact that all the csp mutants despite presence of lower (ΔcspBD and ΔcspAD) or no (ΔcspABD and ΔcspAB) ActA proteins contained significantly more actA mRNA compared to the WT strain (Figure 4). As such the reduced amounts and absence of ActA in the csp mutants was not associated with a reduction or lack of actA mRNA in these mutants (Figures 3, 4). Overall our observations thus showed that the complete removal or the production of only single Csps leads to either a reduction of proteins as well as their corresponding transcripts (PrfA, LLO, and Mpl) or reduced protein levels while the encoding transcripts are present at elevated (ActA) levels.

Besides intracellular motility roles, ActA has also been described as one of the major cell aggregation determinants in L. monocytogenes. Our observations of reduced amounts and complete lack of ActA amongst the csp mutants led us to hypothesize that cell aggregation in L. monocytogenes might have also been indirectly compromised due to Csp deficiency. To examine this hypothesis, we compared cellular aggregation capacities between WT and csp mutant strains of L. monocytogenes EGDe. To do this, stationary phase cultures that were statically incubated for 24 h were monitored for aggregation visually and through optical density measurement (Figure 5). Under these conditions the WT strain displayed complete aggregation whereas no significant aggregation was observed for the ΔcspABD (no csp genes), ΔcspBD (cspA), and ΔcspAB (cspD) strains. The ΔcspAD (cspB) strain although able to aggregate the levels of cellular aggregation it exhibited were significantly reduced compared to WT. Trends observed in aggregation phenotypes displayed by the different csp mutants interestingly also appear to correlate with their ActA levels as determined by Western blot analysis in BHI cultured bacteria of these mutants (Figure 3A). The ΔcspABD and ΔcspAB strains that produced no ActA as well as the ΔcspBD displaying the lowest detected ActA amounts were all unable to aggregate. In contrast the ΔcspAD, which produced the highest ActA amounts among the mutants was the only csp mutant capable of aggregation. Overall, these results indicated that Csps are also an important determinant of cellular aggregation in L. monocytogenes and this might involve Csp-dependent expression of cell surface proteins including ActA.

The production of flagella and expression of motility contributes to virulence and cellular aggregation phenotypes in L. monocytogenes. Assessing for swarming motility and surface flagellation we also found striking differences between csp mutants and the WT strain of L. monocytogenes EGDe. Although the WT strain showed swarming motility at 25°C, the ΔcspABD (without Csps) and ΔcspAB (cspD) were not motile (Figure 6). The ΔcspBD (cspA) and ΔcspAD (cspB) strains although capable of swarming motility did so to visibly reduced extents when compared to the WT strain. Electron microscopic examination showed that while the motile WT strain exhibited peritrichous flagellation there were no flagella observed on the surfaces of the non-motile ΔcspABD and ΔcspAB mutants (Figure 6). In the weakly motile ΔcspBD and ΔcspAD strains flagellation was detected but at a low frequency compared to the WT strain. While almost all examined WT cells showed peritrichous flagellation, surface flagella expression could only be observed in 20–30% of all cells from these two csp mutants examined. Moreover, in those cells that were flagellated there was relatively fewer flagella per cell observed compared to the WT strain (Figure 6). To further explore the basis for these flagellation differences we also quantified flaA mRNA, which encodes flagellin, the main structural component for flagella, in the WT and csp mutant strains grown in BHI cultures grown at 25°C. Unexpectedly we found that flaA mRNA levels were significantly higher (P < 0.05) than those in the WT strain for all of the csp mutants including the ΔcspABD and ΔcspAB strains, which completely lack flagella (Figure 6). This observation indicated that although flaA mRNA was present in the csp mutants, it might be that it is not being efficiently translated to make the flagellin proteins. Such a scenario might lead to the reduced or lack of flagellation as observed in the csp mutants. Overall these observations therefore indicated that csp genes, particularly cspA and cspB, also influence flagella production and extracellular motility in L. monocytogenes.

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.