Transcriptome analysis of Streptococcus gallolyticus subsp. gallolyticus in interaction with THP-1 macrophage-like cells

Background Streptococcus gallolyticus subsp. gallolyticus (S. gallolyticus) is a pathogen of infective endocarditis. It was observed previously that this bacterium survives longer in macrophages than other species and the phagocytic uptake by and survival in THP-1 macrophages is strain-dependent. Methods The phagocytosis assay was performed with THP-1 macrophages. S. gallolyticus specific whole genome microarrays were used for transcriptome analysis. Results Better survival in macrophages was observed for UCN34, BAA-2069 and ATCC43143 than for DSM16831 and LMG17956. S. gallolyticus strains show high resistance to tested bactericidal agents (acid, lysozyme and hydrogen peroxide). S. gallolyticus stimulates significant lower cytokine gene expression and causes less lysis of macrophages compared to the control strain Staphylococcus aureus. S. gallolyticus reacts to oxidative burst with a higher gene expression of NADH oxidase initially at the early phase. Expression of genes involved in D-alanylation of teichoic acid, carbohydrate metabolism and transport systems were upregulated thereafter. Conclusion S. gallolyticus is very resistant to bactericidal agents normally causing degradation of bacteria in phagolysosomes. Additionally, the D-alanylation of teichoic acid is an important factor for survival.

Macrophages play a key role in innate immunity [11]. In the cardiac valve of patients with streptococcal-induced IE, 4.5% of the areas examined were covered by macrophages, whereas no macrophages were found in the cardiac valves without infection [12]. Boleij et al. have already shown that S. gallolyticus subsp. gallolyticus is able to survive much longer in macrophages than other bacteria (e.g. Bacillus subtilis) [13]. Macrophages are able to phagocyte pathogens and trap them in maturing phagosomes, killing them, breaking them down by bactericidal mechanisms and presenting their pathogens' antigens on their surface [14,15]. Key features of phagocytosis are reactive oxygen species (ROS), nitrogen species, hydrolases and the decrease of the pH value [16]. Thereby, pathogenic bacteria have evolved mechanisms to evade killing by macrophages [17]. Some Staphylococcus aureus strains are able to survive the oxidative burst by different mechanisms or modify their peptidoglycan backbone to become lysozyme-resistant. Streptococcus iniae modulates the inflammatory response of macrophages for intracellular replication [18]. Herdt et al. observed that some S. bovis strains were able to evade phagosomes of macrophages from pigeons, replicated intracellularly and lysed all macrophages within 7 h [19].
In the present study, we analyzed the reaction of THP-1 macrophages to the phagocytosis of S. gallolyticus subsp. gallolyticus and vice versa the reaction of S. gallolyticus subsp. gallolyticus to THP-1 macrophages. Additionally, strain dependent survival in presence of different antimicrobial substances was revealed.

Material and methods
Cell culture and bacterial strains THP-1 cells were cultivated in RPMI 1640 medium (RPMI 1640; Thermo Scientific, Waltham, USA) supplemented with 10% fetal calf serum (FCS; Pan Biotech; Aidenbach, Germany) and 1 × antibiotic/antimycotic solution (100 units/ml Penicillin, 0,1 mg/ml Streptomycin, 0,25 μg/ ml Amphotericin B, AB/AM, Pan Biotech) at 37˚C and 5% CO 2 . S. gallolyticus subsp. gallolyticus strains (Table 1) were grown overnight in brain heart infusion broth (BHI; Thermo Scientific, Waltham, USA) at 37˚C and 220 rpm. Overnight cultures were used for phagocytosis assays. Bacterial cultures in the exponential growth phase were generated by inoculating 5 ml BHI medium with 100 μl overnight culture. The exponential growth phase was reached after 2.5 h at 37˚C and 220 rpm. Cultures in the exponential phase were used to analyze the survival of bacteria in differently supplemented BHI. The bacterial titer was determined by serial dilutions in Dulbecco's phosphate-buffered saline (DPBS) and plating 100 μl of an adequate concentration in triplicate on tryptone soya (TS) agar (Thermo Scientific, Waltham, USA). The TS agar plates were incubated at 37˚C and the colonies that grew were counted by the aCOLyte colony counter (Synbiosis, Cambridge, Great Britain).

Phagocytosis assay
This phagocytosis assay was based on the assays of Boleji et al. and Kaneko et al. [13,20]. Either 1.5 × 10 6 THP-1 monocytes per well were disseminated in 12-well plates or 1.3 × 10 5 THP-1 monocytes per well were disseminated in 96-well plates with 50 ng/ml PMA-supplemented medium to differentiate the monocytes into macrophages within three days. On the third day, the medium was changed into PMA-free medium after washing the cells twice with DPBS (Thermo Scientific, Waltham, USA

Cytotoxicity assay
The cytotoxicity of different S. gallolyticus subsp. gallolyticus strains and Staphylococcus aureus on macrophages was measured using the commercial Pierce LDH Assay Kit (Thermo Scientific, Waltham, USA), following the manufacturer's instructions. The macrophages were infected with S. gallolyticus subsp. gallolyticus or Staphylococcus aureus as described above. The plates were incubated for 5 h using a cell monolayer completely lysed by the lysis solution provided in the kit as a positive control. The negative control was determined by adding 10 μl water instead of bacteria. The plate was spun down at 250 × g for 3 min and 50 μl of the supernatants from infected cells, positive control and negative control were transferred to a new 96-well plate, 50 μl reaction mix was added to the supernatants and they were incubated for 30 min in the dark. The reaction was stopped by adding 50 μl stop solution. Absorption was measured by λ = 490 nm and λ = 680 nm. OD 680 was measured to subtract the instrumental background. The percentage of macrophages lysed was calculated by the following formula: Results of this assay are shown in Table 2 and Fig B2 in S1 File.

Intracellular relative quantification of reactive oxidative species
For the quantification of reactive oxidative species, the phagocytosis assay was carried out as described above with 12 well plates and a MOI of five. Fluorescence-excitation was carried out at λ = 488 nm and emission was detected at λ = 525 nm. The emission of the negative control (no stimulation) was subtracted from the emission Table 2. Ratio of bacteria to macrophages in microscopic analysis and response of THP-1 macrophages to Streptococcus gallolyticus subsp. gallolyticus strains and Staphylococcus aureus ATCC 25923. 1+2) Microscopic analysis revealed the ratio of bacteria counted (without differentiation between dead and alive bacteria) to macrophages with the distinction between bacteria within and outside macrophages (bacteria/macrophage; counted macrophages are set as 1; Fig A and supplemental method in S1 File; n = 100 macrophages).
3) The oxidative reaction of macrophages to bacteria was detected by fluorescence signaling (DCF) at time point t = 0 h. H 2 O 2 (1 mM) was used as stimulation control (Fig B1 in S1 File). 3) Cytotoxicity of bacteria to THP-1 macrophages compared to control (lysis of THP-1 macrophages by lysis buffer was set as 100%) was analyzed by LDH-assay 5 h after phagocytosis (

Lysozyme resistance
Fresh BHI medium was inoculated to test the lysozyme resistance of different S. gallolyticus subsp. gallolyticus strains. Subsequently, 20 μl of bacterial culture (exponential phase) was added to 980 μl BHI medium in 24-well plates supplemented with lysozyme from chicken egg white (Sigma, Steinheim, Germany) (0, 5, 10 or 20 mg/ml), incubated at 37˚C and 70 rpm. As control bacterial growth in BHI medium without any supplement was taken. After 5 h of growth in this medium, cfu were determined by plating assay as described above.

Acid tolerance
Survival of S. gallolyticus subsp. gallolyticus strains at pH 4 was analyzed as follows. Bacterial cultures in the exponential phase were centrifuged (5,000 × g, 5 min) and the supernatant was discarded. The pellets were resuspended in BHI medium (pH = 7.4) or in BHI medium with a pH-value of 4 adjusted by hydrochloric acid. After 5 h of incubation at 37˚C and 220 rpm, the cfu were determined by plating assay. The pH-value was checked after 5 h by pH indicator strips and no change in pH through inoculation was detected.
RNA extraction of THP-1 macrophages with following cDNA synthesis Isolation of RNA was performed with the NucleoSpin RNA II kit (Machery and Nagel, Düren, Germany), following the manufacturer's instruction. The RNA was eluted with 30 μl RNasefree water and quantified using the NanoDrop 2000 (VWR, Radnor, USA). The synthesis of cDNA was carried out by using the Superscript II Reverse Transcriptase Kit (Invitrogen, Carlsbad, USA). An amount of 2 μg of RNA was diluted in 20 μl RNase-free water and cDNA was generated within three steps, following the manufacturer's instructions. The cDNA was diluted with water at a ratio of 1:5 for real-time PCR.
RNA extraction of S. gallolyticus subsp. gallolyticus with following cDNA synthesis for real-time PCR Gene expression was analyzed from the two S. gallolyticus subsp. gallolyticus isolates BAA-2069 and UCN34. The RNA from these strains was isolated at three different time points. The initial time point was directly after centrifugation (400 × g, 5 min) of the 96-well plates (-1 h). The second RNA extraction was at the time point 0 h and the third at 5 h. The RNA was extracted with the peqGOLD Bacterial RNA Kit (VWR, Radnor, USA). Bacterial cells were resuspended in TE buffer and lysis buffer T and transferred to Lysing Matrix B tubes (MP Biomedicals, Santa Ana, USA). Bacterial cells were disrupted by using Vortex-Genie 2 (Scientific Industries, New York, USA) for 3 min at full speed. Further RNA extraction was carried out following the manufacturer's instructions. The RNA was eluted with 30 μl RNase-free water and quantified using the NanoDrop 2000. RNA quality was determined with the Agilent RNA 6000 Pico Kit (Agilent). The RNA samples had a RIN value >8. The synthesis of cDNA was carried out with the High-Capacity cDNA Reverse Transcription Kit (Thermo Scientific, Waltham, USA), following the manufacturer's instructions. The cDNA was generated from 500 ng RNA by a onestep PCR. The cDNA was diluted in water at a ratio of 1:10 for real-time PCR.

Relative quantitative real-time PCR
Gene expression analysis of the cytokines in THP-1 macrophages and verification of S. gallolyticus subsp. gallolyticus microarray results were performed by real-time PCR using the LightCycler 480 (Roche, Basel, Switzerland). The reaction volume was 10 μl containing 2.5 μl cDNA, 0.25 μl each Primer (20 μM), 5.0 μl LightCycler 480 SYBR Green I Master-Kit and 2.0 μl water. The denaturation of the reaction mix took place initially at 95˚C (10 min), followed by 45 cycles consisting of denaturation for 10 s at 95˚C, annealing at 63 or 65˚C, respectively (see Table A  . Intron-spanning primers were used for the human gene expression analysis. All oligonucleotides, their sequences and associated annealing temperatures are listed in Table A in S1 File.
Gene expression analysis of S. gallolyticus subsp. gallolyticus using fullgenome microarray Phagocytosis assay was carried out as described above. Three biological replicates for each strain and time point were generated. To conclude false positive signals, RNA of THP-1 macrophages without bacteria was also processed and hybridized. Microarrays had a customized design (MyArray; OakLabs GmbH, Hennigsdorf, Germany) which was generated out of the sequences of four different S. gallolyticus subsp. gallolyticus genomes [7][8][9]. The microarrays were synthesized by Agilent and present 4,382 genes by 10,607 oligonucleotides consisting of 60 bases. The cDNA and cRNA synthesis, including Cy3-labeling, and the microarray hybridization was carried out with the Quick Amp WT Labeling Kit, one-color (Agilent, Santa Clara, USA), following the manufacturer's recommendations, but with the following exceptions. Due to mixed RNA (prokaryotic and eukaryotic), more RNA (75 ng) was used for labeling and 900 ng labeled cRNA was used for hybridization.

Microscopic analysis
The method of the microscopic analysis is described in the supplement in S1 File.

Statistics
Experimental data were analyzed by Mann-Whitney U test using GraphPad Prism 6.0 (Graph-Pad Software). P values less than 0.05 were considered as statistically significant. Mean with standard error is displayed in figures. For microarray analysis raw data were quantile-normalized, statistical t-test (Welch's t-test) gene expression data were generated by the Direct Array software (OakLabs, Hennigsdorf, Germany). Thereby, all log 2 values between -1 and 1 were ignored and only statistically significant values (p < 0.05) are displayed.

Results
Phagocytosis of five S. gallolyticus subsp. gallolyticus strains More bacterial cells of LMG17956 and BAA-2069 and the S. aureus strain ATCC25923 were phagocytized (about 50%), whereas 39% of the inoculum of the strain ATCC43143 given was found intracellularly in THP-1 macrophages. After eight hours of incubation the two strains DSM16831 and LMG17956 were killed significantly more compared to other strains ( Fig 1B). Microscopic analysis revealed how many bacterial cells per macrophage are present at t = 1 h (Table 2; row: 2-3). Thereby, distinction between live and dead cells was not possible. It was observed that the S. gallolyticus subsp. gallolyticus strains were generally located intracellularly, whereas the S. aureus strain ATCC25923 has been found extracellular of THP-1 macrophages (Fig A  in S1 File; supplementary methods).
The following aspects of the macrophages' response to bacteria were analyzed to determine the source of these strain-dependent differences by the survival within THP-1 macrophages: 1) oxidative burst (row: 4), Table 2) cytokine expression of macrophages (rows: 6-8) and 3) cell lysis (row: 5). The THP-1 macrophages developed the same oxidative burst after stimulation with S. gallolyticus subsp. gallolyticus or S. aureus. Lysis of the macrophages did not differ between the S. gallolyticus subsp. gallolyticus strains after 5 h of stimulation. Indeed, cell lysis was significantly lower compared to the lysis caused by S. aureus (p < 0.05; Table 2). After 5 h of stimulation with S. gallolyticus subsp. gallolyticus, the gene expression of IL8 and IL-1B was not considerably increased. Thereby, gene expression of all three cytokines increased much more highly through phagocytosis of the S. aureus strain ATCC25923 (Table 2).

Survival and growth of S. gallolyticus subsp. gallolyticus under different conditions
Specific characteristics of bacteria can influence their death or survival in macrophages, such as resistances to reactive oxygen species, acid or lysozyme. Therefore, we examined straindependent differences under these different bactericidal conditions (Fig 2). All Figs show the percentage of vital bacteria in the presence of the bactericidal agent compared to the respective control (100%). It was shown that about 25% of the S. gallolyticus subsp. gallolyticus cells survived in acid medium (pH 4). The isolate BAA-2069 survived slightly better (32%; p < 0.05) than the other S. gallolyticus subsp. gallolyticus isolates. The S. aureus isolate ATCC25923 was significantly more susceptible to the acid medium than the S. gallolyticus subsp. gallolyticus isolates (p < 0.0005; Fig 2A).
Resistance to H 2 O 2 was tested in BHI medium with different concentrations of this reactive oxygen agent (Fig 2B). A concentration of 15 mM led to a survival of 24% for the isolate LMG17956.
Altogether, the five S. gallolyticus subsp. gallolyticus isolates analyzed were more resistant to 15 mM H 2 O 2 than the S. aureus isolate ATCC25923 tested (Fig 2B; p < 0.0001), which showed only viable bacteria of 8% compared to the control. No living bacterial cells were detected in medium supplemented with 20 mM H 2 O 2 .
Furthermore, S. gallolyticus subsp. gallolyticus showed strain-dependent growth in the presence of different lysozyme concentrations (Fig 2C). The isolate UCN34 survived as well as the S. aureus isolate ATCC25923 in the presence of lysozyme concentrations up to 20 mg/ml. These strains showed even an increase in growth, UCN 34 with 5 mg/ml and the S. aureus control strain with any lysozyme supplement (5 mg/ml; 10 mg/ml; 20 mg/ml). By contrast, the  Table B in S1 File. *: p < 0.05; **: p < 0.005; ****: p < 0.0001; n = 4. The standard error is marked with error bars. SGG = Streptococcus gallolyticus subsp. gallolyticus; SA = Staphylococcus aureus.

Gene expression analysis of S. gallolyticus subsp. gallolyticus using a full genome microarray
The isolates BAA-2069 and UCN34 were used for microarray experiments because they have shown a better survival in THP-1 macrophages. The gene expression is normed on the time point -1 h, directly after bacteria were attached to the macrophages through centrifugation. The gene expression was normalized to this time point (-1 h), which is set as 0 (log2 ratio) or 1 (fold change). Changes in gene expression are shown for the time points 0 h (20 min incubation with antibiotic-supplemented medium) and 5 h later. Changes between 0 h and 5 h do not differ from the data presented (-1 h and 5 h) and are therefore not displayed. Microarray results, sorted by function, including the log2 ratio, the fold change and the corresponding pvalue are shown in Table 3. Microarray analysis revealed 70 differentially regulated genes in BAA-2069, whereas only 15 genes were detected in UCN34 (Fig 3). According to the microarray data, most genes in BAA-2069 were regulated after 5 h of incubation and only a few genes were regulated prior at 0 h (Table 3). Regulated genes belong mainly to transport systems,  such as the phosphotransferase (PTS) system transporter and ATP transporter, and the metabolisms of carbohydrates, for example, the glg (glgBCDP) operon and the dlt (dltABCD) operon, whose products influence the cell wall. At the earlier time point (0 h), gene expression was

Verification of microarray analysis by real-time PCR
Changes in gene expression were further analyzed by performing relative quantitative realtime PCR with more replicates to confirm the microarray results. Representative genes were selected from especially interesting ones. These genes were analyzed in both S. gallolyticus subsp. gallolyticus strains BAA-2069 and UCN34 to detect strain-dependent differences ( Fig  4). It was shown that the genes of the dlt and glg operon were highly expressed in both isolates after 5 h compared to the control (-1 h). Changes of expression of the genes of the dlt and the glg operon which are found by microarray analysis in BAA-2069 were also found in UCN34 by relative quantitative real-time PCR. Additionally, the changes in gene expression of nox (NADH oxidase), which was found in UCN34 at time point 0 h, was also found in BAA-2069 by real-time PCR at this time point. Changes in gene expression were detected by real-time PCR much more highly, through the higher sensitivity of this method and another normalization method compared to the analysis by whole genome microarray. Downregulated genes found by microarray analysis often showed a less strong change in gene expression in realtime PCR experiments (Fig 4).

Discussion
Circulating S. gallolyticus subsp. gallolyticus cells in the host must be able to survive human defense mechanisms, such as phagocytosis, to colonize at the human endocardium. Therefore, we analyzed the interaction of S. gallolyticus subsp. gallolyticus with THP-1 macrophages.  go without stimuli into apoptosis, THP-1 monocytes are able to grow in vitro [28]. Therefore, the experiments were always performed under the same conditions without the influence of individual host factors. We focused on strain-dependencies and the gene regulation of S. gallolyticus subsp. gallolyticus by binding (-1 h), internalization and first reaction to macrophages (0 h), and the survival within phagolysosomes (5 h). Boleij et al. revealed that S. gallolyticus subsp. gallolyticus is able to survive longer in THP-1 macrophages than Bacillus subtilis or Lactobacillus plantarum [13]. We observed that the phagocytic uptake and the survival of S. gallolyticus subsp. gallolyticus in macrophages depends on the isolate. The strains BAA-2069 and UCN34, for example, showed a persistent titer in macrophages within a distinct time period (I. Grimm, M. Weinstock, I. Birschmann, J. Dreier, C. Knabbe, and T. Vollmer, submitted). This study revealed that S. gallolyticus subsp. gallolyticus was not able to lyse macrophages actively, and hardly any differences in stimulation of the oxidative burst and cytokine gene expression were observed. Cytokine gene expression was induced by the S. aureus strain ATCC25923 significantly more highly. A potential explanation might be that S. aureus stimulates the cytokine expression higher through the lysis of macrophages, another that S. gallolyticus subsp. gallolyticus has mechanisms to alter the inflammatory response [29].
Phagolysosomes have different bactericidal mechanisms to kill and degrade microorganisms. We analyzed three of these to examine how resistant S. gallolyticus subsp. gallolyticus is and whether the resistance can influence its survival within macrophages. It was shown that S. gallolyticus subsp. gallolyticus survives better in an acidic environment than S. aureus, which could be an advantage in survival in phagolysosomes [17,30]. According to the microarray results, protection against reactive oxygen species is enabled by higher expression of nox. It is known that no more than 30 μM H 2 O 2 would accumulate in phagosomes in addition to other reactive oxygen species [31]. Therefore, all the strains tested in this study would survive the bactericidal effectivity of H 2 O 2 of phagosomes. Thereby, lysozyme resistance can support the establishment of an infection by, for example, better survival in phagocytes [32]. It was shown for Streptococcus bovis that lysozyme enables differentiation between human and bovine isolates because human isolates were more resistant to lysozyme [33]. We could not confirm these distinctions between the S. gallolyticus subsp. gallolyticus strains in this study. Although S. gallolyticus subsp. gallolyticus is a lysozyme-resistant bacterial species, no known gene for peptidoglycan modification was found in these S. gallolyticus subsp. gallolyticus strains by in silico analysis [6][7][8][9]32]. Interestingly, we recognized an increase of bacterial growth of the S. gallolyticus subsp. gallolyticus strain UCN34 and the S. aureus strain ATCC 25923 under different concentrations of lysozyme. This phenomenon has not been recognized for bacteria yet, however at the moment an explanation would be too speculative, since to our knowledge there is no literature providing supplying explanations regarding this aspect. It was shown overall that S. gallolyticus subsp. gallolyticus is quite resistant to bactericidal agents which allow killing and degrading pathogens in phagolysosomes. This could explain how S. gallolyticus subsp. gallolyticus is able to survive longer in THP-1 macrophages than Bacillus subtilis and Lactobacillus plantarum [13]. Additionally, the more rapid reduction of vital cells of the isolates DSM16831 and LMG17956 from animals may be explained by the susceptibility of DSM16831 to lysozyme and the susceptibility of LMG17956 to H 2 O 2 .
The isolate DSM16831 presented a remarkable phenotype which showed less virulent characteristics compared to the other strains tested. This phenotype was further confirmed in this study, since the S. gallolyticus subsp. gallolyticus strain DSM16831 was constantly killed in THP-1 macrophages and was less lysozyme-resistant compared to the other strains ( [10], I. Grimm, M. Weinstock, I. Birschmann, J. Dreier, C. Knabbe, and T. Vollmer, submitted). This is the first study using a full genome microarray to analyze the gene expression of S. gallolyticus subsp. gallolyticus. The first reaction (0 h) to phagocytic uptake is the reaction to the oxidative burst by expressing nox, which is a ROS-metabolizing enzyme [34]. Later (5 h), S. gallolyticus subsp. gallolyticus changes mainly the gene expression of proteins which are involved in cell shape/cell wall synthesis and different metabolisms, such as carbohydrates and lipids. The highest change in gene expression in the S. gallolyticus subsp. gallolyticus BAA-2069 and UCN34 through phagocytotic uptake was found in the dlt operon. The products of this operon are essential for the D-alanine substitution of teichoic acids (TAs), which leads to a resistance to cationic antimicrobial peptides [35]. It was shown for Streptococcus agalactiae that deficiency in D-alanyl-TA leads to a higher susceptibility in phagocytes and contributes to virulence [36]. This could be one of the virulence factors of S. gallolyticus subsp. gallolyticus to survive bactericidal agents in phagolysosomes. Another highly expressed gene after phagocytosis was levA, in addition to other genes which belong to PTS-systems. These transport systems are synthesized when the preferred carbon source is exhausted, which could be explained through the phagocytosis from nutrient-rich medium into phagosomes [37]. Another interesting regulated gene is SGGB AA2069_c13580, which codes for an autolysin. The in silico analysis of this gene showed similarities to atlA of Streptococcus mutans. It was observed that the product of this gene increased fibronectin-binding and contributed to bacterial survival in blood and resistance to phagocytosis; therefore, it was hypothesized that AtlA is a virulence factor in IE [38].
In conclusion, it was shown in the present study that S. gallolyticus subsp. gallolyticus is located intracellularly and hardly stimulates cytokine gene expression in THP-1 macrophages without lysing these cells. S. gallolyticus subsp. gallolyticus is resistant to most bactericidal agents which are characteristic for phagolysosomes. Microarray results revealed that D-alanylation of TA is an important factor for survival. Therefore, we conclude that S. gallolyticus subsp. gallolyticus survives in macrophages through changes in the cell wall of this bacterium.