Staphylococcal Protein A Is a Key Factor in Neutrophil Extracellular Traps Formation

Staphylococcus aureus are strong inducers of neutrophil extracellular traps (NETs), a defense mechanism of neutrophils against pathogens. Our aim was to explore the role of Protein A in S. aureus-induced NETosis. We determined the Protein A production of four different S. aureus strains and found a direct relationship between the degree of NETosis induction and Protein A production: strains producing higher concentrations of Protein A evoke significantly more NETs. A S. aureus strain in which Protein A as well as a second binding protein for immunoglobulins (Sbi) have been knocked-out (ΔSpA ΔSbi) induced significantly less NETosis than the wild-type strain. NETosis induction by this knockout strain can be rescued by the addition of purified Protein A. Dead S. aureus did not induce NETosis. In conclusion, Protein A is a determinant for NETosis induction by S. aureus.

regions of the B-cell receptor and act as a B-cell superartigen (22). However, little is known about its direct effect on innate immune cells, particularly neutrophils. Since neutrophils are one of the earliest effector host immune cell against S. aureus invasion and because of their ability to form NETs, we were interested to study whether Protein A is also involved in NETosis. To achieve this, we determined the Protein A production in different S. aureus strains and its relationship with NETosis inducing capacity. Next, we obtained more insight in the role of Protein A in NETosis by studying the rescue of NETosis with Protein A in a S. aureus Protein A knockout strain.

MaTerials anD MeThODs bacterial strains
Bacterial strains used in this study are listed in Table 1. Strains were obtained from the bacterial collection of Department of Medical Microbiology and Infectious Diseases, Erasmus MC Rotterdam.

bacterial growth condition
All strains were cultured on Trypticase™ Soy Agar (TSA) (Becton Dickinson, Breda, The Netherlands) with 5% sheep blood overnight at 37°C. Protein A and the second binding protein for immunoglobulins (Sbi) double knockout Newman strain (ΔSpA ΔSbi) was cultured on TSA containing 5 µg/ml gentamycin and 5 µg/ml tetracycline to maintain its knockout status. After an overnight incubation at 37°C, bacteria were suspended in NaCl 0.9% solution (OD 0.5 at OD600nm), and 200 µl was added to a sterile Erlenmeyer flask containing 100 ml Iscove's Modified Dulbecco's Medium (IMDM) (Gibco, Bleiswijk, The Netherlands). The flask was then incubated for 24 h at 37°C at 150 rpm. The next day, based on OD600nm measurements, the individual strains were concentrated to reach a final concentration of 2 × 10 10 bacteria/ml. Heat killed bacteria were generated by incubating the bacteria at 96°C for 10 min. Bacteria were then harvested and transferred to new IMDM medium. To control for the effectiveness of the heat treatment, the heat-killed bacteria were stained with propidium iodide (PI, diluted 1:400, Sigma Aldrich, Zwijndrecht, The Netherlands) and a sample was cultured to check for growth by plating.

secreted Protein a Measurement
The concentrations of released Protein A by S. aureus strains were measured using a sandwich ELISA type assay specific for Protein A (Enzo, Bruxelles, Belgium) according to the manufacturer's protocol. The detection range of the kit was 15.6-1,000 pg Protein A/ml. S. aureus strains Newman, USA300, RN6390, M116, and Newman ΔSpA ΔSbi were cultured as described above and after overnight culturing, 20 µl of the supernatant was collected, centrifuged at 4,000 g, and then filtered. Supernatant from Newman ΔSpA ΔSbi bacteria was included as a negative control. The optical density at 450 nm was measured using a Biotek plate reader (Biotek) with Gen5 software and used to calculate the protein A concentration.

Facs analysis of surface associated Protein a
Four milliliters of IMDM were inoculated with an overnight culture of S. aureus (Newman, USA300, RN6390, and M116) to obtain OD600nm of 0.05. The individual cultures were incubated for 24 h at 37°C with continuous shaking at 230 rpm. The OD600nm of the bacteria culture was normalized to 0.300, and the bacteria were washed 3 times with PBS, followed by centrifugation for 5 min at 4,000 g. The individual bacterial pellets were suspended in 100 µl PBS. Ten microliters of each bacterial suspension were mixed with either 10 µl of 1:50 dilution of anti-protein A IgY-FITC (FITC-labeled Chicken anti-Protein A), (Gallus Immunotech Inc., Fergus, Canada) or with 10 µl PBS, used as negative control, in an U shape 96-well microplate (Greiner Bio-One, Oberösterreich, Germany). The plate was then incubated at 10°C for 45 min at 800 rpm in the dark. After washing 3 times with 200 µl PBS, bacteria were centrifuged for 5 min at 3,500 g. Bacteria were suspended in 50 µl of PBS and their fluorescence (emission 488-522 nm) was quantified with Accuri C6 Flow Cytometer and analyzed using Accuri C6 Software (version 1.0.264.21) (both BD Bioscience, Breda, The Netherlands). Values are expressed in mean fluorescence intensity (MFI).

neutrophil isolation
Neutrophils were isolated as described previously (23), under endotoxin-free conditions. Briefly, medium Lymphoprep™ (Stemcell Technologies) was used to isolate neutrophils from blood derived from healthy donors within the age range of 25-50. Donors did not use any medication (e.g., anti-inflammatory drugs) that could influence study results. Red blood cells were lysed using Erythrolysis buffer (155 mM NH4Cl, 10 mM KHCO3, 1 mM EDTA, pH 7.4) and then neutrophils were washed two times with HEPES buffer. A final neutrophil concentration of 2 × 10 7 cells/ml was used. All experiments were approved by the Medical Ethics Committee of the Erasmus MC. Each experiment was performed with neutrophils derived from a different donor.

neTosis induction and imaging
Neutrophils were stained for DNA with Hoechst 34580 (diluted 1:10,000, Life Technologies) and for extracellular DNA with PI (diluted 1:400, Sigma Aldrich) in 500 µl DMEM culture medium (Biowhittaker, Lonza). The cells were allowed to attach to gelatincoated coverslips at 37°C for at least 1 h.
To induce NETosis, 500 µl 2 × 10 10 bacteria/ml were added to 500 µl 2 × 10 7 neutrophils/ml in a Attofluor Cell Chamber  neTs Quantification z-Stack images were taken from randomly distributed fields of view within the cell chamber every 3-5 min, in a time frame of 5-40 min. Every image had the same X, Y, and Z dimensions. In each image, NETs were manually traced in every z-stack by using ImageJ (Version 1.49, National Institutes of Health, USA). The total volume (cubic micrometer) of NETs was calculated from 10 to 15 images per strain. The average percentage of NETs coverage was then calculated: total volume of NETs/total volume.

statistical analyses
Data are presented as mean ± SEM. A statistical unpaired, twotailed t-test with Statistical Package for the Social Sciences (SPSS, IBM, version 21) was used to analyze differences between groups. Results were considered significant when p < 0.05.
The amount of Protein A associated to the bacterial surface was higher in Newman and USA300 (3,703 ± 404 MFI and  3,767 ± 163 MFI, respectively) compared to M116 and RN6390 (2,336 ± 288 MFI and 1,649 ± 254 MFI, respectively) ( Figure 1B). In literature, it was described that Protein A is in particular expressed in classical bacterial growth media during the exponential growth phase. To determine if this is the case in IMDM, we as well made a growth curve and determined at three time points the amount of surface associated Protein A. Interestingly, the level of the surface associated Protein A during stationary phase (2.5 and 3.5 h) is higher compared to its exponential phase (24 h, Figure 1C), except for strain RN6390. When we tested these S. aureus strains to see whether the amount of Protein A plays a role in NETosis, we observed a positive correlation between Protein A levels and NETosis. High Protein A-producing strains Newman and USA300 induced significantly more NETosis (Newman 10.7 ± 1.9% and USA300 13.5 ± 3.8% of the total volume) compared to the low Protein A-producing strains M116 (0.7 ± 0.2%, p < 0.001 and p = 0.003, respectively) and RN6390 (3.4 ± 1.4%, p = 0.005 and p = 0.02, respectively) (Figure 2). Corresponding NETs volumes are 3,750 ± 675 µm 3 (Newman), 4,738 ± 1347 µm 3 (USA300), 245 ± 72 µm 3 (M116), and 1,208 ± 482 µm 3 (RN6390). NETosis in all strains started within 5 min and within 40 min all neutrophils had formed NETs.
To further determine the contribution of Protein A to NETosis induction, we used a SpA Sbi double knockout strain of S. aureus Newman (ΔSpA ΔSbi) to induce NETosis. Only very modest NETosis was observed (2.3 ± 0.9%), which was significantly lower than of the WT Newman strain (p < 0.001). We could recover NETosis by the knockout strain by adding purified Protein A. Addition of 0.9, 9, or 90 µg/ml of purified Protein A to the knockout strain, prior to NETosis induction, rescued NETosis induction (8.1 ± 1.9%, p = 0.01; 7.4 ± 1.2%, p = 0.005; 7.2 ± 1.7%, p = 0.02, respectively) (Figure 3) to comparable levels that were observed for WT Newman strain (p = 0.36, p = 0.26, and p = 0.19, respectively).  Additionally, we also were able to rescue NETosis formation to the level induced by WT Newman when Protein A was added to the modestly NETosis inducing strain M116 (0.7 ± 0.2% in M116 vs 9.1 ± 2.0% in M116 plus Protein A, p < 0.001), giving it the same NETosis induction rate compared to WT Newman (p = 0.63, Figure 4).
No NETosis was observed when 0.9, 9, or 90 µg/ml of purified Protein A was added to the neutrophils without bacteria (0.0 ± 0.0, Figure 3). To explore the effect of bacterial viability on NETosis induction, 0.9 and 9 µg/ml Protein A were added to dead bacteria (WT Newman). No NETosis was observed (Figure 4) which indicates that living bacteria are needed in order to induce NETosis. We observed that dead bacteria were phagocytosed by neutrophils, however, when 0.9-90 µg/ml Protein A was added, bacteria were not cleared by the neutrophils (Figure 5).

DiscUssiOn
In this study, we demonstrated that Protein A secretion is positively correlated to NETosis induction by S. aureus. Previously, it has been described that different bacterial species induce different NETosis responses (10,12,23). We showed that between different S. aureus strains, differences in NETosis response are seen, which is correlated to the amount of Protein A present. Furthermore, NETosis induction by S. aureus with little to no Protein A present can be rescued by adding Protein A.
In our study, the bacterial cell wall-associated protein A level is higher during the stationary phase than during the exponential phase. In a previous study by Gao et al., using trypticase soy broth (TSB) medium, the spa gene, that encodes Protein A in S. aureus, is upregulated in the exponential growth phase of the bacteria and the expression could be regulated via various factors (24). In our experiments, the use of IMDM over TSB is preferred, since we work with neutrophils and the composition of culture media for mammalian cells, such as IMDM, is more closely resembling the composition of human conditions. S. aureus behaves differently in IMDM than when cultured in TSB medium (25), which is an interesting observation.
Protein A is one of the important virulence factors that S. aureus uses to evade the immune system. By Protein A binding Protein A Determines NETs Formation Frontiers in Immunology | www.frontiersin.org February 2018 | Volume 9 | Article 165 <1% of NETs were formed when either dead bacteria or Protein A alone was added to the neutrophils. We now observed that phagocytosis was inhibited when Protein A was added to dead bacteria, although in our experiment, to our knowledge, no IgG was present. Protein A has also been described to bind to the TNFR1 of immune cells, causing production of pro-inflammatory cytokines, such as IL-6 and TNF-α (27). Both of these cytokines have been described as inducers of NETosis (28,29 (16,32). These studies speculated that the ability of NETosis induction by bacteria might be a function of bacteria to kill human neutrophils, due to the high survival of bacteria that were able to escape the NETs. Our findings contribute to this hypothesis and suggest that S. aureus is not only able to evade phagocytosis by  to the Fcγ region of immunoglobulin IgG, bacteria can avoid being opsonized and phagocytosed by neutrophils and other immune cells, as it inhibits binding to the neutrophils Fc receptor (26).
In previous experiments, we already reported that dead bacteria are phagocytosed by neutrophils and do not induce NETosis (23). Also, in line with previous experiments (23), no spontaneous NETs were formed within 1 h of induction, as we showed that secreting Protein A, but also to eliminate the neutrophils by evoking NETosis, from which they can escape by producing nucleases.
In this study, we used the double knockout strain S. aureus Newman for SpA and Sbi. The NETosis inducing capacity of this strain also might have been affected by Sbi. Sbi has a similar function and structure as Protein A, and it can interact with complement protein C3, thereby inhibiting opsonization and, therefore, phagocytosis (32,33). However, in our experiments, we were able to fully rescue NETosis induction by only adding purified Protein A, indicating that Protein A is the important determinant of NETosis in this experimental setting.
In chronic wound infections such as osteomyelitis and ulcers, S. aureus is one of the most frequently isolated bacteria (3,34). Also in other diseases where S. aureus infection plays a major role, such as sepsis (7) and COPD (35)(36)(37)(38), more NETs have been associated with a higher severity of the disease. By understanding more about direct interaction between the human immune system, in this case neutrophils, and S. aureus, we find more ways to efficiently target the bacteria by interfering with the bacterial products that influence and inhibit our immune system. In conclusion, the amount of Protein A present is an important determinant of NETosis induction by S. aureus, and NETosis inducing capacities of strains with little to no Protein A present can be enhanced by addition of commercial Protein A.

aUThOr cOnTribUTiOns
All authors contributed to the experimental design. Experiments were performed by TH, AS, TA, NT, and SM. Data were processed by TH, AS, TA, and WC. All authors contributed to data interpretation. Figures were generated by TH and were interpreted by all authors. Statistical analyses were performed by TH and AS. The manuscript was written by TH and AS and revised by all authors.