**Impact of** *Staphylococcus aureus* **Small Colony Variants on Human Lung Epithelial Cells with Subsequent Influenza Virus Infection**

**Janine J. Wilden <sup>1</sup> , Eike R. Hrincius 1, Silke Niemann 2, Yvonne Boergeling 1, Bettina Lö**ffl**er 3,4, Stephan Ludwig 1,5 and Christina Ehrhardt 6,\***


Received: 13 November 2020; Accepted: 11 December 2020; Published: 15 December 2020 -

**Abstract:** Human beings are exposed to microorganisms every day. Among those, diverse commensals and potential pathogens including *Staphylococcus aureus* (*S. aureus*) compose a significant part of the respiratory tract microbiota. Remarkably, bacterial colonization is supposed to affect the outcome of viral respiratory tract infections, including those caused by influenza viruses (IV). Since 30% of the world's population is already colonized with *S. aureus*that can develop metabolically inactive dormant phenotypes and seasonal IV circulate every year, super-infections are likely to occur. Although IV and *S. aureus* super-infections are widely described in the literature, the interactions of these pathogens with each other and the host cell are only scarcely understood. Especially, the effect of quasi-dormant bacterial subpopulations on IV infections is barely investigated. In the present study, we aimed to investigate the impact of *S. aureus* small colony variants on the cell intrinsic immune response during a subsequent IV infection in vitro. In fact, we observed a significant impact on the regulation of pro-inflammatory factors, contributing to a synergistic effect on cell intrinsic innate immune response and induction of harmful cell death. Interestingly, the cytopathic effect, which was observed in presence of both pathogens, was not due to an increased pathogen load.

**Keywords:** *Staphylococcus aureus*; small colony variants; influenza virus; super-infection; pro-inflammatory response

#### **1. Introduction**

The respiratory tract is a major portal for microorganisms, through which virus infections can cause non-symptomatic, mild, and self-limiting but also severe diseases, sometimes with fatal outcomes [1]. A growing body of evidence shows that the human respiratory tract contains a highly adapted microbiota including commensal and opportunistic pathogens. Among those, *Staphylococcus aureus* (*S. aureus*) is of special importance, forming quasi-dormant subpopulations characterized by increased fitness compared to other phenotypes [2]. Colonization of *S. aureus* could either be persistent or

non-persistent, whereby nasal colonization appears to be the most prominent localization [3]. *S. aureus* as a community-acquired pathogen is already colonized on approximately 30% of the human population, some without causing any symptoms [4]. During long-term colonization or infection, *S. aureus* can change phenotypes to so-called small colony variants (SCVs), which adapt in their metabolic and phenotypic characteristics, allowing them to evade the host's immune system. SCVs can be localized intracellularly and are characterized by a slow growth rate, non-pigmentation, less hemolytic activity, and decreased antibiotic susceptibility [5–7] but often enhanced surface presentation of adhesion molecules [8]. SCVs are often misdiagnosed [9]. Due to their slow growth, they often get overgrown by other bacteria, and an initially effective antibiotic treatment results in the development of resistances accompanied by chronic and relapsing infections [5,6,8,10,11]. The clinical relevance of colonizing SCVs gets obvious in patients with chronic respiratory diseases, such as chronic obstructive pulmonary disease (COPD) or cystic fibrosis (CF) [5]. Patients who are colonized with bacteria are more likely to suffer from recurring infections [12], as the phenotype can revert to the pathogenic phenotype.

Besides, simultaneous occurrence of different pathogens can induce or even exacerbate a pathological effect in the lung. Super-infections with influenza viruses (IV) and with the community-acquired *S. aureus* are known to be harmful and lead to increased inflammatory lung damage [13]. Due to their quick adaptation and genomic changes, both pathogens can evade the host's immune response, causing the tedious development of effective medications. Concerning super-infections, most studies describe infections with a primary viral infection that paves the path for a secondary bacterial infection [14–17]. However, there is evidence that primary bacterial colonization also occurs prior to viral infections [18].

However, the influence of colonizing *S. aureus* SCVs on subsequent IV infection is largely unexplored. Thus, in the present study, we aimed to investigate the effect of the bacterial strain *S. aureus* 3878SCV on cell intrinsic immune responses to a subsequent IV infection, in vitro. Here, we observed that the response of anti-viral gene expression was barely changed. However, pro-inflammatory genes were highly upregulated upon super-infection, resulting in an induction of necrotic cell death. Thus, we were able to show that colonizing SCVs could enhance severity of subsequent viral infection.

#### **2. Materials and Methods**

#### *2.1. Cell Lines, Virus Strains, and Bacteria Strain*

All cell lines were cultivated at 37 ◦C and 5% CO2 under sterile conditions. Human lung epithelial cells A549 (American Type Culture Collection (ATCC), Wesel, Germany) were cultivated in Dulbeccos's modified eagle medium (DMEM; Sigma-Aldrich, St. Louis, MO, USA) and Madin-Darby canine kidney cells II (MDCKII) in minimum essential medium eagle (MEM; Sigma-Aldrich, St. Louis, MO, USA), supplemented with 10% fetal bovine serum (FBS; Biochrom, Berlin, Germany).

The human IV strains A/Puerto Rico/8/34 (H1N1, PR8-M) and A/Panama/2007/99 (H3N2, Panama) were taken from the virus stock of the Institute of Virology Muenster, 48149 Muenster, Germany, subcultured and passaged on MDCKII cells.

The persisting bacterial strain *S. aureus* 3878SCV, wildtype phenotype strain *S. aureus* 3878WT, and the human lung isolate of another SCV phenotype strain *S. aureus* 814SCV (provided by Karsten Becker, Institute of Medical Microbiology, Muenster, Germany) were stored at −80 ◦C in a 30% glycerol/brain-heart infusion (BHI; Merck; Darmstadt, Germany) medium. *S. aureus* 3878SCV and *S. aureus* 3878WT were already characterized and described previously [10,19–21]. Before experiments, bacteria were plated on blood agar plates to take single clones, which were inoculated in BHI medium and incubated for 24 h at 37 ◦C and 5% CO2. For bacterial infection, bacterial suspension was washed with phosphate buffered saline (PBS) (4000 rpm; 4 ◦C; 5 min) and adjusted to an optical density of OD600nm = 1. Growth kinetics were performed to determine a colony forming unit (CFU) of 2 <sup>×</sup> 108 CFU/mL at OD600nm = 1 for each bacterial strain used.

#### *2.2. Super-Infection Protocol*

Human lung epithelial cells were seeded in either 6-well plates (0.5 <sup>×</sup> 106) or 12-well plates (0.2 <sup>×</sup> 106) in 2 mL or 1 mL culture medium 24 h before infection. For bacterial infection, the overnight culture was set to OD600nm = 1 to determine the multiplicity of infection (MOI). Cells were washed with PBS and infected with *S. aureus* 3878SCV in invasion media (DMEMINV: DMEM supplemented with 1% human serum albumin, 25 nmol/L HEPES) for 24 h with a MOI of 0.01. For viral infection, supernatant was aspirated, cells were washed with PBS and incubated with IV PR8-M (MOI = 0.1) or IV Panama (MOI = 0.01) in infection PBS (PBSINF: PBS supplemented with 0.2% bovine serum albumin (BSA), 1 mM MgCl2, 0.9 mM CaCl2, 100 U/mL penicillin, 0.1 mg/mL streptomycin) for 30 min. Viral suspension was aspirated, and cells were washed with PBS and further incubated in infection media (DMEMINF: DMEM supplemented with 0.2% bovine serum albumin (BSA), 1 mM MgCl2, 0.9 mM CaCl2) up to 8 hpvi, 24 hpvi, 32 hpvi, 44 hpvi, or 48 hpvi (hours post-viral infection).

#### *2.3. Transfection Protocol*

For transfection of the 3× NFκB reporter plasmid construct as described elsewhere [22] (0.1 μg/μL) A549 cells were seeded in 12-well plates as described above. Cells were transfected with 0.1 μg/μL of the indicated plasmid for 4 h with Lipofectamine® 2000 (Invitrogen, Carlsbad, CA, USA) corresponding to the manufacturer's protocol. Afterwards, cells were washed with PBS and further incubated in cell culture media up to 24 h. Afterwards transfected cells were infected up to 8 hpvi. Performance of luciferase assay was done as described elsewhere [23].

#### *2.4. Intra- and Extracellular Bacterial Titer Measurements*

Extracellular bacterial titers were determined by collecting the supernatant of infected cells including the washing with PBS. Cells were lysed via hypotonic shock with 2 mL ddH2O according to Tuchscherr et al. [7,8] (37 ◦C, 30 min) to determine intracellular bacterial titers, including adherent bacteria at the cells surface. Bacterial suspensions were centrifuged (4000 rpm, 4 ◦C, 10 min), pellets were resuspended in 1 mL PBS, and serial dilutions (1:10) were plated on BHI agar plates and incubated for 32 h at 37 ◦C.

#### *2.5. Standard Plaque Assay*

Infectious virus particles in the supernatant were titrated to determine viral titers. A standard plaque assay was performed as described earlier [24].

#### *2.6. Quantitative Real-Time PCR (qRT-PCR)*

RNA isolation was performed with RNeasy Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Reverse transcription was performed with 2 μg of total RNA with Revert AID H Minus Reverse Transciptase (Thermo Fisher Scientific, Karlsruhe, Germany) and oligo (dT) primers according to the manufacturer's protocol. qRT-PCR was performed using a Roche LightCycler 480 and Brilliant SYBRGreen Mastermix (Agilent, Santa Clara, CA, USA) according to the manufacturer's instructions. The following primers were used: GAPDH: fwd 5 GCAAATTCCATGGCACCGT3 , rev 5 GCCCCACTTGATTTGGAGG3 ; IL-6: fwd 5 AACCTGAACCTTCCAAAGATGG3 , rev 5 TCTGGCTTGTTCCTCACTAGT3 ; IL-8: fwd 5 CTTGTTCCACTGTGCCTTGGTT3 , rev 5 GCTTCCACATGTCCTCACAACAT3 ; TNFα: fwd 5 -ATGAGCACTGAAAGCATGATC-3 , rev 5 -GAGGGCTGATTAGAGAGAGGT-3 ; IL-1β: fwd 5 -CAGCTACGAATCTCCGACCAC-3 , rev 5 -GGCAGGGAACCAGCATCTTC-3 ; IFNγ: fwd 5 AAACGAGATGACTTCGAAAAGCTG3 , rev 5 TGTTTAGCTGCTGGCGACAG3 ; RIG-I: fwd 5 CCTACCTACATCCTGAGCTACAT3 , rev 5 TCTAGGGCATCCAAAAAGCCA3 ; IFNβ: fwd 5 TCTGGCACAACAGGTAGTAGGC3 , rev 5 GAGAAGCACAACAGGAGAGCAA3 ; MxA: fwd 5 GTTTCCGAAGTGGACATCGCA3 , rev 5 GAAGGGCAACTCCTGACAGT3 ; OAS1: fwd

5 GATCTCAGAAATACCCCAGCCA3 , rev 5 AGCTACCTCGGAAGCACCTT3 . Relative changes in expression levels (n-fold) were calculated according to the 2−ΔΔC*<sup>t</sup>* method [25].

Bacterial RNA was isolated with the RNeasy Protect Bacteria Mini Kit (Qiagen, Hilden, Germany), and cDNA synthesis was performed using QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. qRT-PCR was performed using a Roche LightCycler 480 (Basel, Switzerland) and Brilliant SYBRGreen Mastermix (Agilent, Santa Clara, CA, USA) according to the manufacturer's instructions. The primers to determine the gene expression of *gyrB*, *aroE*, *arg*, *hla*, *sarA*, and *sigB* were already described elsewhere [7].

#### *2.7. RT2 Profiler Array Analysis*

For pathway focused gene expression analysis, we used *RT<sup>2</sup> Profiler PCR Arrays* (Qiagen, Hilden, Germany). RNA isolation, cDNA synthesis and procedure were performed according to the manufacturer's protocol and instructions. Analysis of data was accomplished by using the GeneGlobe Data Analysis Center recommended by Qiagen [26].

#### *2.8. FACS Analysis*

Determination of secreted proteins in the supernatant was performed with BioLegend's LEGENDplex™ (San Diego, CA, USA) according to the manufacturer's protocol. The human anti-viral and pro-inflammatory chemokine panels were used. Results were analyzed by BioLegend's cloud-based LEGENDplex™ Data Analysis Software. To analyze apoptotic or necrotic cells, infection was performed as described above until 44 hpvi. Cells were treated with tumor necrosis factor related apoptosis inducing ligand (TRAIL; Enzo Life Sciences, Farmingdale, NY, USA) (150 ng/mL) 4.5 h before harvested and used as a positive control for apoptosis. The supernatant was collected for this purpose, and cells were detached from the wells with trypsin-EDTA and recombined with the supernatant. Cell suspension was centrifuged at 1000× *g* at room temperature (RT) for 5 min, and cells were washed with PBS supplemented with 5% FCS. Afterwards, cells were stained with annexin V FITC (20 μL) (ImmunoTool, Friesoythe, Germany) and 1:2000 eBioscience™ Fixable Viability Dye eFluor™ 660 (Thermo Fisher Scientific, Karlsruhe, Germany) in 100 μL 1× annexin V staining buffer (10× annexin V staining buffer: 0.1 M HEPES, 1.4 M NaCl, and 25 mM CaCl2 (pH 7.5)) for 30 min at RT in the dark. Further 150 μL of staining buffer were added and the supernatant was removed after centrifugation. Cells were fixed with 500 μL PBS containing 4% formaldehyde and 1.25 mM CaCl2 for 20 min at RT in the dark. Cells were finally resuspended in 150 μL staining buffer and stored at 4 ◦C until measurement with the FACSCalibur flow cytometer (BD Biosciences, Heidelberg, Germany), followed by the analysis with FlowJo software (v.10; Flow Jo, Ashland, OR, USA). Three gates were set as the following: annexin V positive cells (early apoptotic cells) and live/dead marker positive cells (cells with a membrane rupture tending to necrosis).

#### *2.9. Recording of Cytopathic E*ff*ect of Infected Cells*

To record the CPE at different time points, cells were visualized with Canon (EOS 500D) by light microscopy (Axiovert 40C, ZEISS, Jena, Germany) with a 10× magnification.

#### *2.10. SDS-PAGE and Western Blot Analysis*

Protein expressions were determined by separating proteins in a polyacrylamide gel and subsequent transfer on nitrocellulose membranes by western blot analysis as described earlier [27]. The following antibodies were used: pMLKL [(S353) #91689 Cell Signaling, Frankfurt, Germany], PARP (#611039 BD, Heidelberg, Germany) and ERK1/2 (#4696 Cell Signaling, Frankfurt, Germany).

#### *2.11. Lactate Dehydrogenase (LDH) Assay*

The lactate dehydrogenase assay (CellBiolabs, San Diego, CA, USA) was used to measure the cell cytotoxicity and was used according to the manufacturer's instructions. Cells were infected as described previously, and 90 μL of the supernatant was mixed with 10 μL of the LDH cytotoxicity reagent in a 96-well plate. This plate was incubated at 37 ◦C and 5% CO2 for 30 min, and the OD450nm was measured on a Spectromax M2 Instrument (Molecular Devices, Munich, Germany). Triton X-100 used according to the manufacturer's instructions served as a positive control.

#### *2.12. Quantification and Statistical Analysis*

All data represent the means + standard deviation (SD) of three independent experiments. Statistical significances were determined by unpaired *t*-test (Figure S4A), one-way ANOVA followed by Tukey's, (Figures 4D,E, 5, 6A,B,D,E, Figure S1B,D,E, S2D,E, S3 and S4C–E) or two-way ANOVA followed by Sidak's (Figure 2, Figures S1C and S4B) or followed by Tukey's (Figure 4A–C,F–I and Figure S2A–C,F–I) multiple comparison test using GraphPad Prism software (v.7.03, GraphPad Prism, Inc., La Jolla, CA, USA).

#### **3. Results**

#### *3.1. Primary S. aureus 3878SCV Infection Provokes a Cytopathic E*ff*ect in Presence of IV*

Cell death mechanisms induced by *S. aureus* or IV alone are very well investigated and described [28–33]. With respect to IV and *S. aureus* super-infection, we recently were able to show a *S. aureus*-mediated switch from IV-induced apoptosis to necrosis [27]. It is known that IV infection paves the path for secondary bacterial infection, resulting in enhanced pathogen-load [15,34], cytokine expression [35,36], and cell death [27]. Since *S. aureus* often persist in humans without any harm, we aimed to investigate the effects of colonizing *S. aureus* SCVs on secondary IV super-infection.

In a first set of experiments, we focused on the cell morphology of A549 human lung epithelial cells in absence and presence of *S. aureus* 3878SCV and IV. For this reason, A549 human lung epithelial cells were infected with *S. aureus* 3878SCV, which is a well described SCV patient isolate [10,37], for 24 h followed by infection with IV strain A/Puerto Rico/8/34 (PR8-M; H1N1) for the indicated points in time. The morphology of single- and super-infected cells was analyzed by light microscopy in comparison to uninfected control (mock) (Figure 1). While the cell monolayer is still intact in un-, single-, and super-infected cells up to 32 hpvi (hours post-viral infection), first changes in the cell morphology were visible 48 hpvi in single virus-infected and super-infected cells. Pictures of virus-infected cells showed a less confluent cell monolayer compared to uninfected cells, and in super-infected samples a clear cytopathic effect was observed, indicated by cell monolayer disruption and floating cells. To be able to ascribe these findings to the SCV phenotype, we additionally specified the pathological difference between *S. aureus* wildtype phenotype and SCV phenotype (*S. aureus* 3878WT and *S. aureus* 3878SCV) by infecting A549 human lung epithelial cells. Cell morphology was monitored by light-microscopy (Supplementary Figure S1A) and cell viability was quantified by lactate dehydrogenase assay (LDH) assay (Supplementary Figure S1B). Both assays indicate a massive destruction of the cell monolayer 8 h post bacterial infection (hpbi) with *S. aureus* 3878WT in comparison to *S. aureus* 3878SCV. Further, the determination of the expression of distinct bacterial genes, which are involved in the virulence of the pathogens, verified the reduced virulence of *S. aureus* 3878SCV in comparison to the *S. aureus* 3878WT (Supplementary Figure S1C). Based on these results, *S. aureus* 3878WT was not used in the following experiments. The analysis of cell viability at 32 hpvi and 48 hpvi confirmed the cell disturbance in presence of *S. aureus* 3878SCV and IV infection (Supplementary Figure S1D,E).

**Figure 1.** *S. aureus* 3878SCV colonization and subsequent influenza virus infection provokes a cytopathic effect. A549 human lung epithelial cells were infected with *S. aureus* 3878SCV (multiplicity of infection (MOI) = 0.01) for 24 h at 37 ◦C and 5% CO2. Afterwards, cells were infected with influenza viruses (IV) Puerto Rico/8 (PR8)-M (MOI = 0.1) until the indicated points in time (hours post-viral infection, hpvi). Cells were visualized by light microscopy with a 10× magnification. Shown are representative images of three independent experiments (*n* = 3).

These results point to an altered cell culture environment and/or cellular signaling upon super-infection with *S. aureus* 3878SCV and IV, which could be triggered by increased pathogen load or cell intrinsic signaling changes in presence of both pathogens.

#### *3.2. Primary Infection with S. aureus 3878SCV Followed by IV Infection Had No Impact on Bacterial or Viral Titers*

First, we analyzed whether the observed cytotoxicity of co-infected A549 human lung epithelial cells with *S. aureus* 3878SCV and IV was due to increased pathogen load. For this, we infected A549 cells with *S. aureus* 3878SCV for 24 h and super-infected with two different IV strains for the indicated points in time to determine the amount of plaque forming units (PFU) or colony forming units (CFU) of viruses or bacteria, respectively (Figure 2).

In general, titers of IV and SCVs increased with time, but neither viral (Figure 2A,B) nor bacterial titers (Figure 2C–F) were significantly changed upon super-infection compared to single-infected cells, a phenomenon independent of the virus strain used [PR8-M (H1N1), A/Panama/2007/99 (Panama; H3N2)].

Thus, these data indicate that the disruption of the cell monolayer upon super-infection is not induced by increased amounts of pathogens but by a different mechanism that is altered by the presence of both pathogens.

**Figure 2.** Pathogen load is not affected during *S. aureus* 3878SCV colonization and subsequent influenza virus infection. A549 human lung epithelial cells were infected with *S. aureus* 3878SCV (MOI = 0.01) for 24 h and/or super-infected with (**A**,**C**,**E**) IV PR8-M (H1N1; MOI = 0.1) or (**B**,**D**,**F**) IV Panama (H3N2; MOI = 0.01) for 8 hpvi, 24 hpvi, or 32 hpvi. At the indicated times post-viral infection, supernatants were collected to determine viral and extracellular bacterial titers. Afterwards, cells were lysed via hypotonic shock to analyze intracellular bacterial titers. Means + SD of three independent experiments with technical duplicates are shown (*n* = 3). Statistical significance (compared to single-IV infection (**A**,**B**) or single-bacteria infection (**C**–**F**) was analyzed by a two-way ANOVA, followed by Sidak's multiple comparison test; (hpvi = hours post-viral infection; ns = not significant).

#### *3.3. Pro-Inflammatory Gene Expression Is Highly Upregulated after Super-Infection of S. aureus 3878SCV and IV*

Given the observation that super-infection of *S. aureus* 3878SCV and IV PR8-M induced a cytopathic effect (Figure 1), which was not caused by increased pathogen load (Figure 2), we aimed to elucidate if changes of cell intrinsic signaling and inflammatory gene expression might be responsible for this phenomenon. We analyzed the gene expression of 84 different genes, involved in different signaling cascades by use of a *RT<sup>2</sup> profiler Array* (Qiagen, Hilden, Germany) in a single experiment to gain a first insight in the complexity of cellular signaling (Figure 3). This enables a quick analysis of expression levels of different genes that are organized by their function to be able to limit the amount of genes, altering the cell intrinsic signaling. Here, we used the anti-viral immune response panel, including pattern recognition receptors (PRRs), cytokines, and chemokines involved in pathogen recognition and immune responses. The bioinformatic analysis is based on conventional ct-values and was performed with the recommended GeneGlobe online software [26]. A clustergram was generated to visually illustrate all up- and downregulated genes that were analyzed (Figure 3A). To further interpret the results of the *RT*<sup>2</sup> *profiler Array*, we did an in silico clustering of the upregulated genes of the array that

were highly upregulated (difference of an n-fold of 2) in *S. aureus* 3878SCV and PR8-M super-infected cells compared to single-infected cells (*APOBEC3G*, *CASP1*, *CASP10*, *CCL3*, *CCL5*, *CD40*, *CD80*, *CTSS*, *CXCL10*, *CXCL11*, *CYLD*, *IL1B*, *IL6*, *CXCL8*, *MEFV*, *TLR3*, *TNF*, *B2M*) (see Supplementary Table S1), with respect to their linkage to specific signaling pathways (Figure 3B) by using the Kyoto Encyclopedia of Genes and Genomes mapper (KEGG mapper). KEGG mapper is a database resource of collected information about pathways and the involved genes representing a pool of molecular interactions, reactions, and their relation to each other [38–40]. Down-regulated genes were excluded, as the gene expressions were negligible (Supplementary Table S1). Besides gene clusters connected to expected PRR pathways including TLR-(11 genes involved, out of the 18 highly upregulated genes comparing co- and single-infected cells identified (11/18), NLR- (7/18), TNFR- (6/18), RLR- (5/18), and NFκB- (5/18) signaling pathways (Figure 3B), we identified gene clusters belonging to two cell death mechanisms, necroptosis (5/18) and apoptosis (3/18). Furthermore, we identified genes involved in the IL-17 (5/18) and c-type lectin (5/18) signaling pathways. To further classify the activated genes leading to the observed cytopathic effect on human lung epithelial cells, we searched for a specific induction pattern in which super-infected cells led to upregulated genes. We, therefore, compared all upregulated genes of single-infected to super-infected samples in a Venn diagram (Figure 3C,D). We identified 11 genes that were induced in all three infection-scenarios compared to uninfected cells and 9 genes that were upregulated in super-infected cells only. We also compared the upregulated genes for super-infection with IV Panama. Here, all infection scenarios shared the induction of 12 genes, where 7 genes were exclusively induced by the super-infection of *S. aureus* 3878SCV and IV Panama. The upregulated genes of the Venn diagram are listed in Table S2A,B. With respect to the mRNA expression levels shown in Supplementary Table S1 and the cytopathic effect observed in super-infected cells (Figure 1), an induction of pro-inflammatory immune response can be concluded, which was further visualized by graphs, exhibiting the gene expression of the highly upregulated genes (Figure 3E).

To confirm an increased pro-inflammatory status of the human lung epithelial cells upon super-infection, we analyzed the mRNA expression of different representative pro-inflammatory cytokines and chemokines (IL-6, IL-8, TNFα, IL-1β, and IFN-γ) in detail (Figure 4A–E). Furthermore, we analyzed the mRNA expression of molecules that are involved in the induction of the type-I-IFN signaling (RIG-I, IFN-β, MxA, and OAS1) (Figure 4F–I), since it was described that IV-induced type-I-IFN signaling had an impact on bacterial infections [41].

In *S. aureus* 3878SCV colonized cells subsequently infected with IV PR8-M the mRNA expression of IL-6, IL-8, TNFα, and IL-1β 32 hpvi was induced if compared to uninfected cells or single-infected cells (Figure 4A–D) and IFN-γ showed the same tendency (Figure 4E). Single-infection of *S. aureus* 3878SCV or IV PR8-M resulted in no significant induction of the mRNA expression 8 hpvi, 24 hpvi, or 32 hpvi, except for IL-8, which was significantly induced 8 hpvi in bacteria single-infected cells (Figure 4B). Nevertheless, this induction was abolished over time. Genes, encoding key proteins involved in the recognition, and induction of type-I-IFN signaling were upregulated in IV PR8-M infected cells 8 hpvi (IFN-β by tendency) (Figure 4G) or 24 hpvi (RIG-I, MxA and OAS1) (Figure 4F,H,I). Previous colonization with *S. aureus* 3878SCV had no impact on IV-induced mRNA expression of factors linked to the type-I-IFN response, except for RIG-I at 32 hpvi, which was significantly decreased in super-infected cells. However, the enhanced RIG-I mRNA synthesis did not result in alterations of viral titers. Similar results were obtained upon super-infection with *S. aureus* 3878SCV and IV Panama, indicating a virus-independent effect (Supplementary Figure S2A–I).

**Figure 3.** Gene expression analysis by *RT2 Profiler Array*. (**A**–**E**) A549 lung epithelial cells were infected with *S. aureus* 3878SCV (MOI = 0.01) for 24 h and/or super-infected with IV PR8-M (H1N1; MOI = 0.1) or IV Panama (H3N2; MOI = 0.01) for 32 h. Subsequently, RNA was isolated and further used to perform the *RT*<sup>2</sup> *Profiler Array* (Qiagen, Hilden, Germany). Ct-values were analyzed with the recommended QIAGEN web portal [26]. (**A**) A clustergram is shown, visualizing the up- and downregulated genes of the customized 84-gene array. (**B**) Ten potential signaling pathways are listed, which can be analyzed by the *RT2 Profiler Array* with the correlating count of genes involved. The mapping was done by using the Kyoto Encyclopedia of Genes and Genomes (KEGG) mapper [38–40]. (**C**,**D**) A Venn diagram of the upregulated genes in a super-infection scenario with *S. aureus* 3878SCV and IV PR8-M (**C**) or IV Panama (**D**) is shown. The analysis was performed by use of http://bioinformatics.psb.ugent.be/webtools/Venn/. (**E**) Gene expression of highly induced genes indicating an increased pro-inflammatory cytokine response. Values are shown as n-fold over mock (32 hpvi); (*n* = 1); (hpvi = hours post-viral infection).

**Figure 4.** Pro-inflammatory cytokines and chemokines are enhanced after *S. aureus* 3878SCV colonization and subsequent IV PR8-M infection. (**A**–**I**) A549 human lung epithelial cells were infected with *S. aureus* 3878SCV (MOI = 0.01) for 24 h and/or super-infected with IV PR8-M (H1N1; MOI = 0.1) for 8 hpvi, 24 hpvi and/or 32 hpvi. Afterwards, RNA was isolated and mRNA levels of IL-6, IL-8, TNFα, IL-1β, IFN-γ, RIG-I, IFN-β, MxA, and OAS1 were determined by qRT-PCR. All values were correlated to the representative mock-control 8 hpvi (IL-6, IL-8, TNFα, RIG-I, IFNβ, MxA and OAS1) or 32 hpvi (IL-1β and IFN-γ). Means + SD of three independent experiments including technical duplicates are shown. Statistical significance was analyzed by a two-way (**A**–**C**), (**F**–**I**) or one-way (**D**,**E**) ANOVA, followed by Tukey's multiple comparison test (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001); (hpvi = hours post-viral infection; ns = not significant).

To analyze whether the induction of mRNA synthesis of pro-inflammatory genes could also be detected on protein level and to get further insights into the cell intrinsic innate immune status of super-infected A549 cells, the protein expression of exemplary cytokines and chemokines was monitored by FACS analysis (Figure 5A–H). Remarkably, FACS analysis verified the increased pro-inflammatory response of A549 human lung epithelial cells for the secretion of representative factors. In super-infected cells, protein levels of IL-6, RANTES (CCL5), IP-10, and I-TAC were significantly induced compared to uninfected or single-infected cells with either *S. aureus* 3878SCV or IV PR8-M (Figure 5A,C,E,F). TNFα was also significantly upregulated upon super-infection compared to uninfected and bacteria single-infected cells but not to IV PR8-M single-infected cells. Furthermore, IV PR8-M infection

provoked TNFα protein expression 32 hpvi (Figure 5B). Representative IFN protein concentrations of IFN-γ and IFNβ (Figure 5G,H) showed no alteration in the amount of secreted proteins.

**Figure 5.** Secretion of the pro-inflammatory cytokines and chemokines are enhanced after *S. aureus* 3878SCV colonization and subsequent IV PR8-M infection regulated by TLR2- and RIG-I-mediated NFκB promoter activation. (**A**–**H**) A549 human lung epithelial cells were infected with *S. aureus* 3878SCV (MOI = 0.01) for 24 h and/or super-infected with IV PR8-M (H1N1; MOI = 0.1) for 32 h. Afterwards, supernatants were collected to measure the concentration of secreted proteins via FACS analysis. Means + SD of three independent experiments, including technical duplicates, are shown. (**I**) A549 human lung epithelial cells were transfected with 3× NFκB luciferase promoter reporter construct for 24 h prior to super-infection as described before. Afterwards cells were harvested and analyzed for luciferase activity. (**J**–**L**) A549 human lung epithelial cells were stimulated with LTA (100 ng/mL) for 24 h at 37 ◦C and 5% CO2. Afterwards, cells were stimulated with cellular RNA (cRNA) or viral RNA (vRNA) (100 ng/mL) in the presence or absence of LTA for 4 h at 37 ◦C and 5% CO2. Subsequently, RNA was isolated, and mRNA levels of IL-6, IL-8, and TNFα, were measured by qRT-PCR. All values are correlated to the respective mock-control (*n* = 3). Statistical significance was analyzed by one-way ANOVA followed by Tukey's multiple comparison test (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001; # = \*\*\* *p* < 0.001 compared to LTA + vRNA, except for vRNA); (hpvi = hours post-viral infection; ns = not significant).

These results emphasize the enhanced cell intrinsic pro-inflammatory status of the super-infected cells. We could confirm these results by the use of IV strain Panama, verifying a viral strain independent effect (Supplementary Figure S3A–H). Based on these results and due to the fact that the induction of pro-inflammatory cytokines and chemokines is mainly driven by NFκB activation, we hypothesized an induction of the pro-inflammatory response via specific PRRs, resulting in the activation of the NFκB-signaling cascade [42]. To confirm the induction of NFκB-signaling we transfected A549 cells with an artificial NFκB promoter-dependent luciferase reporter plasmid prior to super-infection with *S. aureus* 3878SCV and subsequent IV PR8-M infection. An increase of NFκB activation was observed in super-infected cells compared to uninfected and IV PR8-M-infected cells, while *S. aureus* 3878SCV infection only resulted in an increase of NFκB activation by trend (Figure 5I). This induction pattern was also confirmed in cells super-infected with IV Panama (Supplementary Figure S3I).

The induction of pro-inflammatory responses via NFκB in epithelial cells after pathogen exposure can be triggered by both pathogens through different factors and their corresponding receptors [43–45], which were further analyzed. To exclude specific pathogen-mediated interference with cellular factors due to differences in protein expression, such as virulence factors or surface proteins, viral RNA (vRNA) and bacterial lipoteichonic acid (LTA, Invivogen, San Diego, CA, USA) were used as pathogen specific molecular stimuli. In human lung epithelial cells vRNA is mainly recognized by RIG-I, leading to a strong induction of the type-I-IFN-signaling cascade [46], while LTA is mainly recognized by TLR-2 [47]. We investigated mRNA expression of IL-6, IL-8, and TNFα after stimulating A549 cells with vRNA and LTA (Figure 5J–L). The results matched our findings obtained from super-infected cells, since significant enhancement of mRNA expression of IL-6, IL-8, and TNFα was observed in presence of both stimuli. Artificial effects caused by RNA transfection could be excluded due to equal cytokine mRNA expression induced by cellular RNA (cRNA) and cRNA + LTA stimulated cells. Stimulation with vRNA tended to induce the expression of IL-6, IL-8, and TNFα, which, however, was not significant compared to unstimulated cells.

Overall, these data suggest an induction of pro-inflammatory gene expression responses through the detection of bacterial and viral components via the pathogen-associated molecular pattern receptors (PAMP) RIG-I and TLR-2, followed by the induction of NFκB. To exclude bacterial strain-specific effects, another SCV strain (*S. aureus* 814SCV) was used to determine pathogen loads and pro-inflammatory gene expression in IV super-infection (Supplementary Figure S4). While neither viral titers nor intraand extracellular bacterial load were increased in presence of both pathogens, pro-inflammatory cytokine expression was enhanced, verifying the former observations.

#### *3.4. S. aureus 3878SCV Provoke Enhanced Necrotic Cell Death in Presence of IV Infection*

The observed disruption of the cell monolayer (Figure 1) could be induced by a variety of mechanisms. Besides the involvement of pro-inflammatory cytokines in the innate immune response, these factors are also involved in the induction of cell death mechanisms, like apoptosis and necrosis. As the results shown in Figure 3 indicate, an upregulation of pro-inflammatory cytokines, the cell death mechanisms might be triggered by TLRs or cell death receptors through PAMPs or cytokines, like TNFα, among others [48–50]. Therefore, we further investigated the induction of apoptosis and necrosis, correlating to the cell death mechanisms identified in the *RT2 Profiler Array* analysis (Figure 3B).

As the results of the LDH assay led to the hypothesis of an induced necrotic cell death mechanism, we performed FACS analysis to determine the number of early apoptotic cells by detecting phosphatidylserine which switches to the cells' surface in early apoptotic cells and can be labeled with annexin V. Cells with a membrane rupture tending to necrosis were detected by using a viability marker comparable to 7-aminoactinomycin D and propidium iodide staining. Therefore, we performed the infection up to 44 hpvi to be able to still distinguish between early apoptosis and necrotic-like cells and stained the cells accordingly. The amount of necrotic cells significantly increased comparing un- or single-infected with super-infected cells, probably indicating necrosis (Figure 6A). Furthermore, the amount of apoptotic cells was significantly higher in IV-infected cells compared to un-, bacteria-, or super-infected cells 44 hpvi (Figure 6B).

**Figure 6.** *S. aureus* 3878SCV colonization and subsequent IV infection inhibits IV-induced apoptosis but results in the induction of necrosis. (**A**–**E**) A549 human lung epithelial cells were infected with *S. aureus* 3878SCV (MOI = 0.01) for 24 h and/or super-infected with IV PR8-M (H1N1; MOI = 0.1) for 44 hpvi (**A**,**B**) or 32 hpvi (**C**–**E**). At the indicated times post-viral infection, total amount of cells was collected to perform FACS analysis to determine the relative amount of viability marker positive cells (**A**) or annexin V positive cells (**B**). Furthermore, whole cell lysates were subjected to western blot analysis (**C**). (**D**,**E**) Densitometrical analysis of three independent western blot experiments of cleaved pMLKL (**D**) and PARP (**E**) 32 hpvi are shown. Equal protein amounts were calculated by correlating the signal intensities to their corresponding ERK1/2 signals. Means + SD of three independent experiments are shown (*n* = 3). Statistical significance was analyzed by a one-way ANOVA, followed by Tukey's multiple comparison test (**A**,**B**,**D**,**E**); (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001); (hpvi = hours post-viral infection; ns = not significant).

As cell death mechanisms like necrosis can be further defined in specific mechanisms and to compare our findings to previously described inductions of cell death mechanisms upon infection with SCV [51] or in co-infection scenarios [27], we performed western blot analysis to be able to differentiate between necroptosis and apoptosis.

Necroptosis is an inflammatory programmed form of necrosis, which was already described in a recent publication reporting its induction by *S. aureus* SCV in single-infected human primary keratinocytes [51]. Necroptosis is induced via a receptor-interacting protein (RIP) kinase-mediated activation, resulting in the phosphorylation and oligomerization of mixed lineage kinase domain like pseudokinase (MLKL) and pore formation, leading to the release of inflammatory cytokines. To distinguish necroptosis from apoptosis induction, we monitored the induction of phosphorylated MLKL and PARP cleavage, which are indications for both cell death mechanisms [52]. We infected A549 human lung epithelial cells with *S. aureus* 3878SCV for 24 h, followed by IV infection with PR8-M for 32 h (Figure 6C–E).

In super-infected cells, induction of pMLKL 32 hpvi was observed in comparison to uninfected, bacteria-, or virus single-infected cells, respectively. PARP cleavage was more likely to be induced in IV PR8-M-infected cells, and was slightly decreased in super-infected samples 32 hpvi (Figure 6C). However, the densitometrical analysis of three independent experiments could only confirm a trend of activated MLKL due to induced phosphorylation in super-infected cells (Figure 6D), whereas an induction of apoptosis upon IV infection could be verified (Figure 6E). Additionally, pyroptosis as another form of regulated necrotic cell death mechanism, which is activated via the induction of the inflammasome resulting in the cleavage of gasdermin D, could not be detected by cleaved gasdermin D (Supplementary Figure S5). The original blots are shown in the Supplementary Figure S6.

Thus, our results indicate a necrotic cell death induction, most likely induced by increased pro-inflammatory gene expression response after super-infection with *S. aureus* 3878SCV, followed by secondary IV infection with PR8-M and Panama.

#### **4. Discussion**

The first occurrence of persisting bacteria or SCVs was already described about 100 years ago [53]. Even though they are known for such a long time, not many studies were undertaken to elucidate their impact on cellular responses or their impact on additional infections with other pathogens. Our aim was to investigate the interaction of *S. aureus* SCVs with a subsequent IV infection in respect to epithelial cell responses, which built the first cellular barrier for pathogens in the lung. Here, we demonstrate that invasive *S. aureus* SCVs do have an impact on the cell intrinsic response in human lung epithelial cells, as indicated by highly secreted pro-inflammatory cytokines and chemokines (Figure 5 and Supplementary Figure S3) and, furthermore, an induction of necrotic cell death of super-infected compared to single-infected cells (Figure 6). This was somehow surprising, since the majority of SCVs are not described to significantly induce cell intrinsic responses, due to decreased secretion of virulence factors [54], a feature that would match their dormant status. In particular, not much is known about the impact of SCVs on lung tissue responses and nothing so far about their impact on a secondary IV infection. In this study, we were able to show a cytopathic effect accompanied by increased pro-inflammatory cytokine and chemokine release and necrotic cell death through colonizing *S. aureus* 3878SCV and different IV strains, such as PR8-M and Panama.

Typically, super-infections with pathogenic *S. aureus* strains and IV led to increased pathogen loads accompanied with the induction of pro-inflammatory responses [13,35]. However, this could not be confirmed within the present study in the SCV and IV super-infection scenario. It was shown previously that super-infection with pathogenic *S. aureus* leads to the enhancement of viral titers due to the inhibition of STAT1 and STAT2 dimerization, resulting in decreased production of anti-viral factors [15]. This inhibitory effect could be excluded since mRNA expression of RIG-I, IFNβ, MxA, or OAS1 in super-infected cells compared to IV-infected cells were not altered (Figure 4 and Supplementary Figure S2). Based on these results, we could exclude an effect of the anti-viral

response and the involvement of an altered pathogen load. Nevertheless, we identified a clear induction of pro-inflammatory cytokines and chemokines in human lung epithelial cells. Besides the attraction of immune cells and the induction of an anti-pathogen status of the cell, pro-inflammatory cytokines induce a stress response leading to the induction of cell death mechanisms via TLRs or death receptors [55–57].

To proof the impact of two main cell death mechanisms we performed FACS analysis to monitor early apoptotic and necrotic-like cells. Correlating to the LDH assays, we could confirm an increase in necrosis during super-infection (Figure 6A). Besides, the amount of apoptotic cells was decreased in super- compared to IV-infected cells. Further specifications of cell-death mechanisms by western blot analysis revealed the tendency for an increase of phosphorylated MLKL in super-infected cells, giving the hit of probably induced necroptosis. Concomitantly, IV-induced PARP cleavage was reduced in super-infected cells compared to IV PR8-M-infected cells by trend.

Interestingly, there are two different mechanisms described, how pathogenic *S. aureus* and *S. aureus* SCVs are able to induce necroptosis [27,51]. During the critical phase of *S. aureus* infection the virulence factor *agr* is induced [8], resulting in possible secretion of different toxins, which induces necroptosis [27,33]. In SCV-infected keratinocytes, necroptosis was driven by the activation of glycolysis [51]. *S. aureus* adopts its whole metabolism to persist within the host. The metabolic changes of *S. aureus* were already described elsewhere [58]. As the utilization of the tricarboxylic acid cycle for the host cell and the persisting bacteria is decreased, the glycolysis is stronger induced to generate adenosine triphosphate (ATP). As we observed a disruption of cell monolayer and a possible induction of phosphorylated MLKL upon super-infection, we linked our findings more to necroptosis (Figures 1 and 6). Even though *S. aureus*-induced necroptosis might be independent of TLR stimulation [59], our data indicate a synergistic effect of *S. aureus* 3878SCV and IV inducing cell death, which can be related to TLR2- and RIG-I-mediated pro-inflammatory response induction. In addition, our data show that the superinfection could be imitated with the stimuli LTA and vRNA. This underlines that the initial induction of the pro-inflammatory response and the subsequent cell death must be different from that of pathogenic bacterial strains that induce cell death much more quickly. In case of SCV, this indicates a lower virulence probably due to the decreased secretion of virulence factors. Nonetheless, dormant SCVs can work synergistically and affect the virus-induced immune response. As we performed pure ligand experiments, inhibitory effects of molecules of this pro-inflammatory cell intrinsic response is supposed to trigger cellular stress in the form of reactive oxygen species [60].

So far, the impact of *S. aureus* SCVs with subsequent IV infection had not been investigated. Interestingly, we could give first insights in this super-infection scenario and unravel one extraordinary role of a SCV patients' isolate *S. aureus* 3878SCV with subsequent IV infection. We observed an induction of pro-inflammatory cytokines and chemokines, which underlines the severity of the coincident occurrence of *S. aureus* SCVs and IV. These data point to a cross-interaction of necrotic cell death and pro-inflammatory cell intrinsic response, as the pathogens alone can induce an inflammatory response through PAMPs and secreted cell damage-associated molecular patterns (DAMPs). Upon necrotic cell death induction, further pro-inflammatory responses are induced via DAMP receptors [50], leading to an enhancement of pro-inflammatory cytokines and chemokines seen on transcriptional and translational level.

In summary, we were able to show that persistent *S. aureus* SCV and subsequent IV infection affects cell-internal immune response by inducing the release of pro-inflammatory cytokines and chemokines, resulting in cell death induction.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-2607/8/12/1998/s1, Table S1: List of ct-values analyzed by GeneGlobe online software of the *RT<sup>2</sup> Profiler Array* plate. Table S2: (A,B) Listed gene names of the venn diagrams shown in Figure 3. Figure S1: Wildtype phenotype *S. aureus* 3878 is more virulent compared to *S. aureus* 3878SCV, but *S. aureus* 3878SCV induces LDH release upon super-infection with PR8-M. Figure S2: Pro-inflammatory cytokines and chemokines are enhanced after *S. aureus* 3878SCV colonization and subsequent IV Panama infection. Figure S3: Secretion of the pro-inflammatory cytokines and chemokines are enhanced after *S. aureus* 3878SCV colonization and subsequent IV Panama infection regulated by TLR2- and

RIG-I-mediated NFκB promoter activation. Figure S4: Pathogen load and pro-inflammatory cytokines and chemokines are enhanced after super-infection with the SCV strain *S. aureus* 814SCV. Figure S5: *S. aureus* 3878SCV colonization and subsequent influenza virus infection has no effect on the induction of pyroptosis. Figure S6: Original western blots of Figure 6C and S5A.

**Author Contributions:** Conceptualization, J.J.W., E.R.H., S.L. and C.E.; methodology, J.J.W.; software, J.J.W.; validation, J.J.W., Y.B. and E.R.H.; formal analysis, J.J.W., E.R.H., Y.B., S.L., S.N. and C.E.; investigation, J.J.W.; resources, S.L. and C.E.; data curation, S.L. and C.E.; writing—original draft preparation, J.J.W. and C.E.; writing—review and editing, J.J.W., E.R.H., S.N., Y.B., B.L., S.L. and C.E.; visualization, J.J.W.; supervision, C.E.; project administration, C.E.; funding acquisition, J.J.W., S.L. and C.E. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Deutsche Forschungsgemeinschaft (SFB 1009, project B01 and B02, CRU342 P06 and under Germany´s Excellence Strategy—EXC 2051—Project-ID 390713860). We acknowledge support by the German Research Foundation and the Open Access Publication Fund of the Thueringer Universitaetsund Landesbibliothek Jena Projekt-Nr. 433052568.

**Acknowledgments:** We would like to thank Karsten Becker for providing us with the bacterial isolate.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Communication* **Mild Lactic Acid Stress Causes Strain-Dependent Reduction in SEC Protein Levels**

**Danai Etter 1,2 , Céline Jenni 2, Taurai Tasara <sup>1</sup> and Sophia Johler 1,\***


**Abstract:** Staphylococcal enterotoxin C (SEC) is a major cause of staphylococcal food poisoning in humans and plays a role in bovine mastitis. *Staphylococcus aureus* (*S. aureus*) benefits from a competitive growth advantage under stress conditions encountered in foods such as a low pH. Therefore, understanding the role of stressors such as lactic acid on SEC production is of pivotal relevance to food safety. However, stress-dependent cues and their effects on enterotoxin expression are still poorly understood. In this study, we used human and animal strains harboring different SEC variants in order to evaluate the influence of mild lactic acid stress (pH 6.0) on SEC expression both on transcriptional and translational level. Although only a modest decrease in *sec* mRNA levels was observed under lactic acid stress, protein levels showed a significant decrease in SEC levels for some strains. These findings indicate that post-transcriptional modifications can act in SEC expression under lactic acid stress.

**Keywords:** superantigen; mastitis; food intoxication; regulation; sec variants

**Citation:** Etter, D.; Jenni, C.; Tasara, T.; Johler, S. Mild Lactic Acid Stress Causes Strain-Dependent Reduction in SEC Protein Levels. *Microorganisms* **2021**, *9*, 1014. https://doi.org/ 10.3390/microorganisms9051014

Academic Editor: Rajan P. Adhikari

Received: 24 March 2021 Accepted: 5 May 2021 Published: 8 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

*Staphylococcus aureus* (*S. aureus*) is of major relevance in food intoxications and infectious diseases of humans and animals [1,2]. *S. aureus* employs a plethora of virulence factors including secreted enterotoxins (SEs). They lead to an emetic response when ingested and act as superantigens [3,4]. The consumption of one or several preformed enterotoxins produced by *S. aureus* causes staphylococcal food poisoning (SFP). SFP symptoms include nausea, vomiting, and abdominal pain, followed by diarrhea [5]. It is one of the most common causes of foodborne intoxications worldwide [6]. The European Food Safety Authority (EFSA) reported 393 SFP outbreaks in 2014 and an increase in cases with 434 outbreaks in 2016 [7,8]. Moreover, in the United States, the Centers for Disease Control and Prevention (CDC) reported 17 SFP outbreaks and 566 cases in 2014. The most common contributing factors in these outbreaks were improper maintenance of the cold chain and inadequate food preparation practices, leading to the proliferation of pathogens and the concomitant production of enterotoxins [9].

Of the currently 25 known SEs, SEC is of particular interest since various, often host-specific variants have been reported [10–13]. SEC plays a crucial role in SFP and the development of atopic dermatitis [14]. SEC is also frequently found in milk and milk products [15–24] and represents a key driver of the inflammatory response in bovine mastitis in dairy cattle [25–28]. We recently provided a comprehensive review of SEC variants and their role in SFP, as well as their structure and properties [10].

To reduce the burden of *S. aureus* in the dairy value chain, lactic acid bacteria (LAB) have been used as an intervention in mastitis treatment and as starter cultures in cheese production. For instance, LAB were successfully used to alleviate mastitis symptoms in cows [29] and to inhibit mastitis-causing pathogens including *S. aureus* [30]. The use of

*Weissella paramesenteroides* GIR16L4 and *Lactobacillus rhamnosus* D1 as starter cultures was shown to decrease SEC expression in several *S. aureus* strains [31]. Another study showed decreased expression of *sec* and SEC in co-cultures with LAB compared with pure *S. aureus* cultures by up to 331-fold in TSB and milk [32]. Metabolites of LAB, in particular lactic acid, might present an additional metabolic burden to *S. aureus* and therefore interfere with toxin expression.

Lactic acid is a lipophilic weak organic acid that freely dissociates through the bacterial membrane. Inside the cell, lactic acid dissociates and thereby releases protons that acidify the cytoplasm. Additional energy is required to maintain internal pH, leading to adaptations in bacterial metabolic processes [33,34]. It can therefore alter toxin regulation by interfering with *S. aureus* regulatory systems such as the accessory gene regulator (*agr*) [35]. The *agr* regulon employs a multicomponent system that is activated by autoinducing peptides (AIP). Upon activation, RNAIII is transcribed, which deactivates the repressor of toxins (*rot*) [36]. The activity of *agr* may be influenced by other regulatory elements that react to changes in the microenvironment or external stressors.

External stressors that have been shown to influence SE production include NaCl [37], low pH [38], nitrite [39], and others [40,41]. Although previous studies demonstrated the substantial resilience of *S. aureus* against low pH, with growth being observed at pH 4 [42–44], pH stress can still influence toxin expression. Lactic acid was previously shown to affect SEA and SED production [45–47]. In the presence of LAB, *agrA*, *sarA*, and *sigB* are typically downregulated, while *rot* is upregulated [48]. However, it remains unclear how external stressors affect the complex regulatory network and whether the presence of lactic acid influences SEC expression. Therefore, we investigated the role of lactic acid in SEC production on mRNA and protein level.

#### **2. Materials and Methods**

#### *2.1. Bacterial Strains, Growth Conditions, and Sample Collection for sec mRNA and SEC Protein Quantification*

All *S. aureus* strains used in this study are listed in Table 1. The strains were grown in LB medium (non-stress control conditions) and in LB supplemented with lactic acid. Mild acidic stress conditions encountered in food were mimicked by adjusting the LB medium (BD, France) to pH 6.0 using ~0.6 mL 90% (*v*/*v*) lactic acid (Merck, Darmstadt, Germany). The medium was buffered using 19.52 g 2-(N-Morpholino) ethanesulfonic acidhydrate (MES hydrate, Sigma-Aldrich, Switzerland) per 1000 mL LB. pH was monitored for 2 representative strains and remained unchanged over the course of the experiment (Supplementary Figure S2). All media were sterile filtered and stored at 4 ◦C.

**Table 1.** Overview of *S. aureus* strains used in this study including their SEC variants, origin, and clonal complex. <sup>1</sup> Medical Department of the German Federal Armed Forces, Germany. <sup>2</sup> Bavarian State Office of Health and Food Safety, Germany.


Single colonies from each strain were transferred from 5% sheep blood agar to 5 mL LB broth and grown overnight (37 ◦C, 125 rpm). Overnight cultures were centrifuged (5000× *g*, 2 min) and washed twice with 0.85% NaCl, before resuspending the pellet in 0.85% saline solution. We adjusted 50 mL of medium (LB and LB + lactic acid) to an OD600 of 0.05 using the washed bacteria. The culture was incubated at 37 ◦C at 125 rpm and harvested after 4, 10, and 24 h during exponential, early stationary, and late stationary growth phase, respectively. Three independent biological replicates were collected.

Growth curves were evaluated by plating serial dilutions on plate count agar (Oxoid, Pratteln, Switzerland) as previously described [37] with minor modifications. Namely, the culture was adjusted to a final OD of 0.05 in 50 mL medium in 250 mL Erlenmeyer flasks. Two independent replicates were assessed.

For *sec* mRNA quantification, 1 mL of sample was added to 3 mL RNAprotect®®Tissue Reagent (Qiagen, Hilden, Germany) and processed according to manufacturer's instructions. The cell pellets were stored at −20 ◦C until further processing. For SEC protein quantification, 1 mL of sample was collected in low protein binding micro-centrifuge tubes (Thermo Scientific, Waltham, MA, USA) and stored at −20 ◦C until further processing.

#### *2.2. RNA Extraction*

RNA extraction was performed with the RNeasy mini Kit Plus (Qiagen, Hilden, Germany) as previously described [52]. All RNA samples were quantified using QuantusFluorometer (Promega, Dübendorf, Switzerland) instruments. Quality control was performed by the Agilent 2100 Bioanalyzer (Agilent Technologies, Waldbronn, Germany) instrument using Agilent RNA 6000 PicoReagents according to the manufacturer's instructions. Samples were included in the study if they met the inclusion criteria of RNA integrity number > 6. RNA integrity numbers ranged from 6.18.5.

#### *2.3. Reverse-Transcription and Quantitative Real-Time PCR*

All RNA samples were diluted to 40 ng/μL in RNase-free water. A total of 480 ng was converted to cDNA using the QuantiTect®®Reverse Transcription Kit (Qiagen, Germany) according to the manufacturer's instructions. A no-RT control and a negative control were included in every run. The final cDNA was diluted 1:10 with DNase-free water (Promega, Madison, WI, USA) and stored at −20 ◦C. The following primers sequences were used for real-time qPCR: forward 5 TAA CGG CAA TAC TTT TTG GT3 and reverse primer 5 AGG TGG ACT TCT ATC TTC AC3 .A5 μL template was added to 10 μL LightCycler®® 480 SYBR Green I Master, 2 μL of each primer (5 μM), and 1 μL nuclease-free water in LightCycler®®480 Mulitwell Plate 96white (Roche, Basel, Germany). The plate was centrifuged for 2 min at 1500× *<sup>g</sup>*. Quantification was performed on the Lightcycler®®96 Instrument (Roche, Basel, Switzerland) as previously described [52]. The relative expression of the target gene *sec* was normalized using the housekeeping genes *rho* and *rplD* [52]. Ct values were determined using the Lightcycler®®Software v. 1.1.0.1320 (Roche). The influence of lactic acid stress on *sec* expression in each strain is expressed as Δct values and relative expression (2-ΔΔct). The following formula was used to calculate expression values: 2-(ref control-sec control)-(ref lactic acid-sec lactic acid). Statistical analysis was performed with RStudio 1.3.1093 and GraphPad Prism 9.0.0. For RNA analysis a mixed-effect linear model was fitted on the fold change, with a full three-way interaction between reference gene, strain, and time effects. Fold change was log10-transformed to ensure normal distribution. To determine whether individual mRNA levels were increased or decreased (indicated by a fold change significantly larger than 1), lsmeans was used to perform a two-sided effect test, with Holm–Bonferroni-corrected *p*-values. The results were regarded as significant if *p* < 0.05.

#### *2.4. Protein Quantification*

An enzyme-linked immunosorbent assay (ELISA) was developed to quantify the effect of mild lactic acid stress on SEC protein levels. The protocol was based on [53] with some modifications according to [41]. Sheep Anti-SEC IgG (Toxin Technology Inc., Sarasota, FL, USA) was used to measure SEC concentrations. For quantification, a standard curve was obtained using SEC2 (Toxin Technology, Inc., USA). Absorbance was measured at 405 nm in a Synergy HT plate reader (BioTek, Sursee, Switzerland). Absorbance values were plotted against toxin concentrations, and values were determined from linear regression

in Excel (version 16.44). ELISA measurements were performed in duplicates. Statistical analysis was performed with RStudio 1.3.1093 and GraphPad Prism 9.0.0. Protein data were analyzed via two-way ANOVA and post hoc Tukey's multiple comparisons. Results were regarded as significant if *p* < 0.05.

#### **3. Results**

#### *3.1. Effect of Mild Lactic Acid Stress (pH 6.0) on Bacterial Growth and sec mRNA Levels*

The bacterial growth of the seven *S. aureus* strains was compared in LB and LB supplemented with lactic acid by plate counting. The growth behavior was similar for all strains under control and stress conditions. Growth was not impaired by lactic acid in any of the investigated strains (Supplementary Figures S1 and S2).

Under lactic acid stress, we observed a trend toward decreased *sec* expression (Figure 1) compared with control conditions. However, the reduction in *sec* expression was straindependent, being only significant in strains BW10 (10 and 24 h), NB6 (24 h), SAR1 (4 h), SAR38 (4 and 10 h), and OV20 (10 h).

**Figure 1.** Effect of lactic acid stress on *sec* mRNA levels in *Staphylococcus aureus* strains BW10, NB6, SAI3, SAI48, SAR1, SAR38, and OV20 in exponential (4 h), early stationary (10 h), and late stationary (24 h) phase. *sec* mRNA levels are expressed as relative quantification values. *sec* mRNA expression was normalized to reference genes *rho* and *rplD* [52]. Replicates are shown as single data points; horizontal grey lines indicate means. Timepoints are signified by fill color (4 h, light blue; 10 h, medium blue; 24 h, dark blue). Significant differences in *sec* mRNA levels in LB compared with LB + lactic acid are marked by asterisks (*p* < 0.05).

#### *3.2. SEC Protein Levels under Lactic Acid Stress*

SEC concentrations under lactic acid stress and control conditions were assessed by ELISA at 4, 10, and 24 h (Figure 2, Table 2). Under control conditions, two different expression levels were observed. BW10 and SAI48 were classified as SEC over-producers (>1000 ng/mL) with concentrations ranging from 3410 ng/mL to 9867 ng/mL after 24 h, respectively. NB6, SAI3, SAR1, SAR38, and OV20 were classified as low to moderate level SEC producers (<1000 ng/mL) with concentrations from 54.5 to 344.9 ng/mL after 24 h (Table 2). Expression levels were lowest after 4 h and highest after 24 h for all strains.

**Figure 2.** SEC concentration in log10 ng/mL under lactic acid stress (pH 6.0) compared with control conditions in *Staphylococcus aureus* strains BW10, NB6, SAI3, SAI48, SAR1, SAR38, and OV20 in exponential (4 h), early stationary (10 h), and late stationary (24 h) phase. SEC levels under control conditions are shown in grey; pH stress levels are shown in blue. Replicates are shown as single data points; horizontal lines indicate means. Darkening fill colors indicate progressing time points. Statistically significant differences between conditions are indicated by lines with asterisks (\* *p* < 0.05).

SEC concentrations under lactic acid stress were generally lower than those under non-stress control conditions, with the exception of SAI48, SAR1, and SAR38 (Figure 2). BW10 and SAI48 again showed the highest expression levels with 740 ± 69 and 4887 ± 1027 ng/mL at the late exponential phase, respectively. The low-to-moderate SEC producers NB6, SAI3, SAR1, SAR38, and OV20 ranged from 1 ng/mL (OV20) to 142 ng/mL (OV20) over all time points. SEC production in the human infection isolate SAI48 was the least impaired by lactic acid stress (−29%) while the ovine mastitis isolate OV20 was affected the most (−255%) (Table 2).


**Table 2.** Effect of pH stress on SEC expression. Absolute values in ng/mL including standard deviation. Effect is shown as a percentage difference under lactic acid stress (pH 6.0) compared with non-stress control conditions.

#### **4. Discussion**

Since *S. aureus* possesses a competitive growth advantage in many food matrices, it is crucial to identify compounds that interfere with SE production. Here, we used lactic acid and pH 6.0 to mimic conditions comparable to food matrices such as ham, cheese, and other fermented products [47]. LB medium was used to ensure reproducibility and to allow observation of the effect of lactic acid as an individual constituent. Experiments were performed at 37 ◦C to provide optimal growth conditions for *S. aureus*.

In the investigated strains, lactic acid stress resulted in a trend toward lower *sec* transcription, although the results were not significant for all strains. Significant reductions in mRNA levels were observed for strains isolated from food and bovine mastitis, whereas human infection isolates did not alter *sec* transcription levels. The influence of a complex food matrix containing lactic acid such as milk was demonstrated to reduce *sec* expression, especially in late stationary phase after 48 h [54]. Other SEs, for instance *sed,* were not significantly altered under lactic acid stress [47]. How lactic acid influences toxin transcription likely depends on the genetic background of a strain. We could, however, not observe any correlation between transcription levels and factors such as clonal complex or the SEC toxin variant of the respective strain. Possibly, a larger strain set may provide further insights.

SEC protein data only partly correlated with the *sec* transcriptional patterns. Whereas BW10 showed a significant reduction in mRNA and protein levels after 10 and 24 h, SAI3 did not display any reduced transcriptional activity, but exhibited significantly reduced SEC concentrations. Especially under certain environmental stress situations such as NaCl, sorbic acid, or complex food environments, mRNA levels do not always reflect protein levels, indicating that post-transcriptional modification might be at play [55–58]. Overall, lactic acid stress led to decreased levels of SEC in all strains. Since growth under lactic acid stress was similar to control conditions, factors independent of growth rates must be at play. Again, we did not observe any correlation between toxin production and clonal complex, source, or toxin variant. Still, SEC overproducers BW10 and SAI48 were also the highest SEC producers under lactic acid stress. Interestingly, previous studies reported substantially lower toxin concentrations of 1–70 ng/mL in milk and even under optimal growth conditions [55]. For SEA, an ingested amount of 60 ng was sufficient to reach SFP attack rates of almost 100%, highlighting the threat of overproducer strains such as BW10 and SAI48 [59]. Another study demonstrated even more pronounced reductions in SEA, SEB, SEC, and SED caused by lactic acid. This pronounced reduction could be expected, since higher concentrations (pH < 5) than in the present study were used [60]. In contrast, mild lactic acid stress was also reported to increase the formation of SEA [45]. It was suggested that *S. aureus* strains from human and food sources produce higher SEC levels in contrast with strains from animal sources [11]. The SEC overproducers in our study, BW10 and SAI48, originated from cases of SFP and human infection, respectively, and showed high SEC concentrations under control conditions and lactic acid stress. However, the moderate SEC producers SAR1, SAR38, and OV20 originating from bovine and ovine mastitis milk showed higher SEC concentrations than NB6 and SAI3, which originated from cases of SFP and human infection, respectively. This underlines the importance of investigating multiple strains for *S. aureus* enterotoxin analysis as postulated by previous studies [39,61–63]. The observed reduction in toxin expression may be magnified at lower temperatures as found in food environments.

Several SEs including SEC are regulated by the quorum-sensing system *agr*. The activity of *agr* may be influenced by other regulatory elements that react to changes in the microenvironment or external stressors. Such additional regulators may include, but are not limited to, *sigB*, *sarA*, *saeRS*, *srrAB*, *arlRS*, and *mgrA* [64,65]. pH stress was shown to affect regulatory elements in *S. aureus* such as *rot*, *agr*, and *sarA* [66]. In the presence of LAB, *agrA*, *sarA*, and *sigB* are typically downregulated, while *rot* is upregulated [48]. Lactic acid can influence toxin expression by interacting with DNA, enzymes, or structural proteins (Figure 3). Since the current data suggest a stronger decrease in SEC protein levels compared with mRNA, post-transcriptional events may be involved. The shift in internal

pH may influence mRNA degradation susceptibility or interfere with translational events. For example, elongation factors were shown to be influenced by acid stress [67]. An overall decreased enzyme activity in the cytoplasm due to lower pH may impact SE handling after translation. In addition, folding and transport of SEs may be impaired by a lower pH, leading to less efficient excretion [68]. Which regulatory elements are affected remains unclear at this time.

**Figure 3.** Mechanism of weak organic acids such as lactic acid on bacterial cells. The undissociated form of organic acids (HA) can cross cell membranes when the pH of the surroundings is lower than that of the cellular cytoplasm. Inside the cell, HA can dissociate and acidify the cytoplasm. Acidic pH damages or modifies internal structures such as enzymes, structural proteins, or DNA. In order to control internal pH, energy is required for active export of protons. The figure was adapted from [34] and created in BioRender.com.

In conclusion, our study demonstrated that lactic acid slightly decreases *sec* transcription and has a more pronounced effect on SEC protein levels. It can therefore be a useful tool in minimizing SEC synthesis during food production and preservation. Since transcription patterns and SEC concentrations are highly variable, it is of utmost importance to investigate several strains with different genomic backgrounds and isolated from different sources. We did not find any correlation between the observed data and any strain-specific properties such as toxin variant or clonal complex. Therefore, further research is needed to determine biomarkers associated with the toxicity of a strain. Future studies should also investigate the mechanistic action of lactic acid on the reduction in SEs to provide further insights into the regulation of SE production. Our research highlights the importance of food composition in mitigating SFP. Compounds such as lactic acid may be used as natural preservatives to minimize toxin production and alleviate the burden of staphylococcal intoxications.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/microorganisms9051014/s1: Figure S1: Growth curves under lactic acid stress and control conditions. Figure S2: pH values for 2 representative strains.

**Author Contributions:** Conceptualization, S.J. and D.E.; methodology, S.J., T.T. and D.E.; formal analysis, D.E. and C.J.; investigation, D.E. and C.J.; resources, S.J.; data curation, D.E. and C.J.; writing—original draft preparation, D.E. and C.J.; writing—review and editing, D.E., S.J. and T.T.; visualization, D.E. and C.J.; supervision, D.E. and S.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

