The cysteine protease ApdS from Streptococcus suis promotes evasion of innate immune defenses by cleaving the antimicrobial peptide cathelicidin LL-37

Streptococcus suis is a globally distributed zoonotic pathogen associated with meningitis and septicemia in humans, posing a serious threat to public health. To successfully invade and disseminate within its host, this bacterium must overcome the innate immune system. The antimicrobial peptide LL-37 impedes invading pathogens by directly perforating bacterial membranes and stimulating the immune function of neutrophils, which are the major effector cells against S. suis. However, little is known about the biological relationship between S. suis and LL-37 and how this bacterium adapts to and evades LL-37–mediated immune responses. In this study by using an array of approaches, including enzyme, chemotaxis, cytokine assays, quantitative RT-PCR, and CD spectroscopy, we found that the cysteine protease ApdS from S. suis cleaves LL-37 and thereby plays a key role in the interaction between S. suis and human neutrophils. S. suis infection stimulated LL-37 production in human neutrophils, and S. suis exposure to LL-37 up-regulated ApdS protease expression in the bacterium. We observed that ApdS targets and rapidly cleaves LL-37, impairing its bactericidal activity against S. suis. We attributed this effect to the decreased helical content of the secondary structure in the truncated peptide. Moreover, ApdS rescued S. suis from killing by human neutrophils and neutrophil extracellular traps because LL-37 truncation attenuated neutrophil chemotaxis and inhibited the formation of extracellular traps and the production of reactive oxygen species. Altogether, our findings reveal an immunosuppressive strategy of S. suis whereby the bacterium blunts the innate host defenses via ApdS protease-mediated LL-37 cleavage.

the principal first line of defense against invading pathogens (1). As a sole member of the human cathelicidin family, the cationic peptide LL-37 is an extremely potent killer against different bacterial infections (1)(2)(3). This peptide can be strongly induced during injury and infection, and it is not only produced by epithelial cells but is also widely expressed in immune cells, such as neutrophils, monocytes, macrophages, and NK cells (4). LL-37 is derived from the C terminus of the precursor hCAP18 protein, which is stored in granules, and it is then released as a mature peptide in the extracellular environment to exercise its bactericidal effect (5). The characteristic positive charge allows LL-37 to associate with the negatively-charged membranes of bacteria. During this interaction, LL-37 converts itself into an ␣-helical conformation, which results in membrane penetration and bacterial lysis (6).
Of note, in addition to its outstanding bactericidal activity, LL-37 directly or indirectly exerts a wide range of immunomodulatory effects on a variety of immune cells, especially neutrophils (4,7). Neutrophils, the most abundant circulating leukocytes in the bloodstream, are considered to be the first cells of the innate immune system to migrate toward the site of infection (8). The modulation of LL-37 on neutrophils is the most diverse and critical, comprising chemotaxis, the production of reactive oxygen species (ROS), 3 the formation of neutrophil extracellular traps (NETs), and the induction of CXCL8 production, etc. (9 -12). Thus, as the main producer of LL-37, neutrophils both express and react to LL-37, and this feedback leads to high concentrations of LL-37 at the site of bacterial invasion. The burst of LL-37 further triggers a more effective immune response of neutrophils, which are capable of rapidly clearing the bacterial infection (13).
Streptococcus suis is an important zoonotic bacterial pathogen that constitutes a significant threat to human health (14). Since the first human case described in Denmark in 1968, over 1,600 human cases of S. suis infection occurred in more than 20 countries and regions, including Asia, North America, and Europe (15,16). Two severe outbreaks of human infections were reported in China in 1998 and 2005, resulting in 229 infections and 52 deaths (17,18). The most frequent clinical manifestations in humans are septicemia and meningitis. S. suis needs to survive well and proliferate in the bloodstream to maintain a high level of bacteremia (14,19). During this process, it is vital that S. suis resists and evades LL-37 and the neutrophil response, which are the first lines of innate immune defenses. However, the S. suis-LL-37 interaction and the mechanism employed by the bacterium to thwart LL-37 and human neutrophils remain incompletely defined.
A previous study identified an aminopeptidase of S. suis that could cleave the arginine residues at the N termini of the substrate and may have a potential contribution in virulence (20). This protease, designated as ApdS here, belongs to the family of cysteine proteases (20). In this study, we characterize the role of the cysteine protease ApdS in the interaction among S. suis, LL-37, and human neutrophils. Our data indicate that the cleavage of LL-37 by ApdS protease is one of the important mechanisms exploited by S. suis to subvert innate host defenses.

S. suis infection stimulates cathelicidin LL-37 production in human neutrophils, whereas S. suis exposure to LL-37 up-regulates the expression of the apdS gene
As a successful intruder into humans, S. suis manages to disarm the innate immune system and to disseminate to the organs through the bloodstream (14,19). To determine the role of the antimicrobial peptide LL-37 in the process of S. suis invasion, the LL-37 production in human neutrophils when exposed to S. suis wildtype (WT) strain 05ZYH33 was investigated using immunofluorescence microscopy with an antibody directed against LL-37. After the infection of neutrophils with S. suis at a multiplicity of infection (m.o.i.) of 0.1, the production of hCAP-18/LL-37 was significantly increased and stored in secondary granules of neutrophils, as indicated by their colocalization with lactoferrin, which is the canonical marker of secondary granules (Fig. 1, A and B). As shown in Fig. 1C, with an increasing number of S. suis cells at the time of infection, the expression of hCAP-18 protein was up-regulated, and the release of LL-37 was also increased. The intensity band ratio of hCAP-18/␤-actin and LL-37//␤-actin is shown in Fig. 1, D and E. These results indicated that S. suis infection could quickly stimulate cathelicidin LL-37 production in human neutrophils, underscoring the importance of LL-37 in the host defense against S. suis infection.
To examine the change of apdS gene expression during the response of S. suis to LL-37, the transcription level of the apdS gene was analyzed using quantitative real-time PCRs (qRT-PCR). When the S. suis strain was treated in vitro with various sublethal concentrations (0.1, 0.2, and 0.4 M) of LL-37, the growth curves and the number of surviving bacteria were similar to those of untreated bacteria (Fig. 1, F and G), but the transcription levels of apdS were different (Fig. 1H). In the presence of increasing concentrations of LL-37, the expression of the apdS gene was up-regulated in a dose-dependent manner. Moreover, with regard to the protein levels, LL-37 also stimulated the expression of the ApdS protein (Fig. 1I). The intensity band ratio of ApdS/GyrB is shown in Fig. 1J. These results suggest that apdS may aid S. suis in adapting to and resisting LL-37 killing.

S. suis cysteine protease ApdS cleaves cathelicidin LL-37
ApdS protein consists of 607 amino acids (aa), containing an N-terminal signal peptide (1-25 aa), a C69 peptidase domain (26 -474 aa), and a C-terminal LPSTG motif (574 -578 aa, Fig.  S1A). Multiple sequence alignment showed that the ApdS of S. suis exhibited similarity to the homologs of both Streptococcus pyogenes and Lactobacillus farciminis, with 45.8 and 24.6% sequence identity, respectively (Fig. S1B). To explore the structural features of ApdS, the three-dimensional structure (3D) of ApdS was simulated using that of L. farciminis C69-family cysteine peptidase (PDB code 5iau) as a template. Automated molecular modeling of the 3D structure generated a homology model consisting of 14 ␣-helices and 16 ␤-sheets. The N-terminal Cys-26 is predicted to be the catalytic site, which is located in the middle of the 3D structure ( Fig. 2A).
To explore the proteolytic activity of the ApdS protease, S. suis ApdS and its catalytic site-mutated protein ApdS(C26A) were overexpressed in the Escherichia coli BL21 (DE3) strain, and the recombinant His 6 -tag fusion proteins were purified using Ni 2ϩ -affinity chromatography (Fig. 2B). Because the C69 family of cysteine protease has N-terminal aminopeptidase activity, the cleavage specificity of ApdS was determined on four chromogenic peptidase substrates. Two (L-Leu-pNa and Gly-pNa) were rapidly hydrolyzed, although no obvious activity was detected against L-Pro-pNa and Gly-Pro-pNa. The K m and K cat values for the Michaelis-Menten kinetics of ApdS were 0.222 mM and 202 s Ϫ1 on L-Leu-pNa, 0.298 mM and 104 s Ϫ1 on Gly-pNa (Fig. 2C).
To determine the ability of ApdS to cleave LL-37, we incubated LL-37 in the presence or absence of ApdS or ApdS(C26A) and analyzed the samples by MS. A reduction in mass of LL-37 after ApdS treatment was found that is equivalent to the loss of its respective N-terminal LLG amino acid (Fig. 2D). The cleavage product was designated as LLG-cleaved peptide (abbreviated as LLGCP). However, there was no reduction in the mass of LL-37 treated with ApdS(C26A) (Fig. 2D), proving that Cys-26 is the catalytic site of ApdS protease.

ApdS-dependent cleavage of LL-37 impairs its bactericidal activity against S. suis
To determine whether the ApdS protease has a role in assisting S. suis resisting against LL-37 killing, an apdS gene-deleted mutant strain (⌬apdS) and its complemented strain (C⌬apdS) were constructed and confirmed by PCR and DNA sequencing (Fig. S2). The results of qRT-PCR showed that the transcription levels of the downstream genes of apdS in WT and ⌬apdS were similar, confirming that the deletion of apdS was nonpolar (Fig.  S3A). Furthermore, ApdS transcription and protein expression in ⌬apdS were virtually undetectable and restored in C⌬apdS (Fig. S3, B and C), and the growth of the three strains was similar (Fig. S3D). The survivability of the WT, ⌬apdS, and C⌬apdS strains upon exposure to the LL-37 was evaluated with a bacte-ricidal assay. As shown in Fig. 3A, the ⌬apdS mutant was significantly more sensitive than the WT strain to LL-37 over a concentration range of 1-4 M. Additionally, the sensitivity of ⌬apdS to LL-37 was increased with the extension of incubation time (Fig. 3B). trans-Complementation with apdS expressed on a pSET2 plasmid restored resistance to LL-37.
Subsequently, we compared the bactericidal ability of LL-37 and LLGCP on S. suis. Because the WT strain can express the ApdS protease to cleave LL-37, the bactericidal ability of LL-37 and LLGCP was evaluated in the ⌬apdS mutant, thus excluding the interference of the ApdS protease. As shown in Fig. 3C, LLGCP had a weaker bactericidal activity compared with LL-37, as indicated by the number of surviving bacteria (CFUs/ ml) following incubation in the presence of each peptide. To compare the effect of these peptides in the bacterial membrane integrity, the permeability of the ⌬apdS mutant after LL-37 and LLGCP treatment was confirmed using the SYTOX Green dye. SYTOX Green is a cationic cyanine dye that can only penetrate compromised membranes, and then it binds to the nucleic acids of bacteria causing a large increase in fluorescence intensity. SYTOX Green fluorescence was measured after the addition of each peptide to S. suis cells suspended in PBS buffer. Treatment with LL-37 showed accelerating membrane permeabilization, as indicated by a concentration-dependent increase in fluorescence intensity (Fig. 3D). Although LLGCP treatment also showed the SYTOX Green influx, the degree of permeabilization was significantly lower than that of LL-37, suggesting their different bactericidal abilities against S. suis (Fig. 3D). These findings indicate that the ApdS protease is crucial for protecting S. suis from the bactericidal activity of LL-37.

Cleavage of LL-37 influences its ␣-helical conformation
The permeabilizing effect of the antimicrobial peptide on bacterial membranes is frequently associated with the presence of a specific secondary structure (21). Previous studies have indicated that the interaction of LL-37 with the bacterial membrane components should result in conformational changes of this peptide (22). To investigate the impact of the cleavage of LL-37 by ApdS on the ␣-helical secondary structure, the structural features of LL-37 and LLGCP in different environments were examined, and the circular dichroism (CD) spectra are illustrated in Fig. 4. In water, the spectra for both LL-37 and LLGCP exhibited a typical random-coil conformation (Fig. 4A). However, the spectra for LL-37 in PBS (pH 7.4) exhibited an ␣-helical structure, which is consistent with the results reported by Zelezetsky et al. (23). The spectrum for LLGCP also showed a helical character, but with only 23.4% helical content compared with the 37.1% helical content of LL-37 (Fig. 4B). Moreover, 50% trifluoroethanol (TFE) has a strong helix-promoting property and has been frequently used as a membranelike environment. In this condition, the spectra for both LL-37 and LLGCP showed a greater helical character, 53.3 and 43.1%, respectively (Fig. 4C). In addition, 30 mM SDS has been frequently used for mimicking the negatively-charged bacterial membrane environment. In this condition, the helical content of LL-37 increased to 63.0%, whereas LLGCP had a helical content of 38.5% (Fig. 4D). These results revealed that the helical contents of LL-37 are significantly higher than those of LLGCP in the PBS and membrane-mimicking environments. Compared with LLGCP, the more stable helical character of LL-37 when interacting with the bacterial membrane may contribute to its superior antimicrobial activity.

ApdS aids S. suis to resist the killing by neutrophils and NETs
Being produced by phagocytes, mainly neutrophils and macrophages, cathelicidin LL-37 significantly promotes the immune defenses against invasive infections (4,7). It has been found that LL-37 is produced at a high concentration of 0.2-1.5 M in most bodily fluids, but at infection sites the concentration can increase more than 6 -10 times (24 -26, 28). For instance, LL-37 is present in the bronchoalveolar lavage fluid at 3.3 M in healthy individuals but can increase up to 13.1 and 25.5 M in patients with pulmonary and systemic infections (26,27). To investigate the role of the ApdS protease in the evasion of innate immune responses by S. suis, the ability of the WT, ⌬apdS, and C⌬apdS strains to survive in human blood, macrophages, neutrophils, and NETs was examined. The concentrations of LL-37 in these different environments were determined (Fig. 5, A-D). Compared with the WT strain, the viability of ⌬apdS after exposure to human blood was markedly reduced, and C⌬apdS rescued the survival defect of the ⌬apdS mutant in the blood (Fig. 5E). In addition, compared with the WT strain, the ⌬apdS mutant became more susceptible to killing by THP-1 macrophages when LL-37 expression was induced with vitamin D (Fig. 5F). The survival ability of the WT, ⌬apdS, and C⌬apdS strains was further evaluated in bactericidal assays using human neutrophils and NETs (which were induced with phorbol 12-myristate 13-acetate (PMA) from human neutrophils). The

Streptococcus suis evades LL-37 and human neutrophils
results showed that the WT strain was more resistant to human neutrophil killing than the ⌬apdS mutant, and it also survived better within NETs (Fig. 5, G and H). Taken together, in accordance with the role of ApdS in evasion from LL-37 killing, ApdS also protects S. suis from killing by LL-37-producing phagocytes, which show the crucial role of ApdS in S. suis resisting and escaping from innate immune-mediated killing.

ApdS-dependent cleavage of LL-37 attenuates its chemotactic ability to attract human neutrophils
How could ApdS mediate neutrophil evasion? An attractive hypothesis is that ApdS-dependent cleavage of LL-37 not only impairs its bactericidal activity, but it also destroys its immunomodulatory properties for neutrophils. Previous studies have found that LL-37 enhanced lysosome formation in phagocytes    (10). Thus, we first assessed the ability of LL-37 and LLGCP to induce the lysosome formation in neutrophils. Unexpectedly, LL-37 and LLGCP induced lysosome formation to similar extents (Fig. S4). These data indicated that ApdS-dependent cleavage of LL-37 did not damage the immunomodulatory properties of LL-37 for lysosome formation.

Streptococcus suis evades LL-37 and human neutrophils
A notable feature of LL-37 is that it has excellent chemotactic properties for neutrophil migration toward sites of infection (4,9,27). To examine whether the cleavage of LL-37 by ApdS affects its chemotactic ability, the neutrophils were pre-labeled with CellTracker, and their migration promoted by LL-37 and LLGCP was monitored using a transwell assay. The results indicate that LL-37 stimulated remarkable chemotaxis of neutrophils in a dose-dependent manner up to a concentration of 2 M. However, in the presence of LLGCP, a significant reduction in neutrophil migration was observed compared with that of LL-37 induction (Fig. 6A). It was also observed that LL-37 could promote chemotaxis through the induction of chemokine CXCL8 production (12). In line with the results of chemotaxis, compared with the untreated group, both LL-37 and LLGCP stimulated the CXCL8 production in a dose-dependent manner. However, LLGCP exhibited a reduced CXCL8 immunostimulatory effect compared with the potent ability of LL-37 (Fig. 6B). Collectively, these data demonstrated that ApdS protease could impair the chemotactic property of LL-37 required for the migration of human neutrophils to the local site of S. suis infection, which can prevent the subsequent functional activation of human neutrophil for phagocytosis.

ApdS-dependent cleavage of LL-37 inhibits its ability in promoting NETs formation and ROS production
An interesting feature of LL-37 is that it can facilitate NET formation, which is one of the fundamental innate immune defense mechanisms (11). To investigate whether ApdS has an effect on the ability of LL-37 in NET formation, LL-37 and LLGCP were used to stimulate human blood-derived neutrophils. The NETs were observed using immunofluorescence microscopy with an antibody directed against DNA/histone 1 complexes. As shown in Fig. 7A, neutrophils alone without peptide incubation did not produce the NET structure. As a positive control, PMA could effectively activate neutrophils to form a characteristic NET structure. In addition, extensive web-like structures into the extracellular space were released when neutrophils were incubated with 8 M LL-37. Neutrophils stimulated with the same concentration of LLGCP also released extracellular DNA-containing fibers, but the area of NET structure was significantly reduced compared with that of LL-37. However, when neutrophils were treated with 2 M peptides, neither LL-37 nor LLGCP could trigger NET formation (Fig.  7B). To further define whether the cleavage of LL-37 by ApdS affects ROS production, a fluorescent ROS dye was used to evaluate LL-37 and LLGCP for intracellular ROS production. As shown in Fig. 7C, LL-37 induced significantly more ROS production at the concentrations of 2-8 M than LLGCP, suggesting that the cleavage of LL-37 by ApdS affects its ability for ROS production. Taken together, a relatively high concentration of LL-37 could significantly stimulate NETs formation, which is consistent with the results of a previous study (11), and

Streptococcus suis evades LL-37 and human neutrophils
ApdS protease could attenuate the effect of LL-37 on NET formation and ROS production.

Discussion
The host innate immune system plays a critical role in the control of invading microbial pathogens. A key feature of this defense is the production of antimicrobial peptides. The only human cathelicidin, LL-37, shows a broad-spectrum bactericidal activity against both Gram-negative and Gram-positive bacteria, various viruses, and fungi (1). The concentration of LL-37 can increase more than 6 -10 times in patients during infection (24 -26, 28), which means that there is a surge of LL-37 surrounding the activated neutrophils at a site of infection. Similarly, in the model of the human blood-cerebrospinal fluid barrier, S. suis infection significantly induced the transcript expression of LL-37 (29). During the process of infection, S. suis is always exposed to the killing effect of LL-37, and how this bacterium deals with LL-37 has remained enigmatic so far. Here, we demonstrated for the first time that S. suis targets and cleaves LL-37 through the activity of the ApdS protease, resulting in impaired bactericidal activity toward S. suis. Additionally, the ApdS-dependent cleavage of LL-37 also impaired its immunomodulatory properties for human neutrophils. Because neutrophils are the major effector cells in the innate immune defense against S. suis (29,30), ApdS protease is expected to be a novel immune evasion factor involved in S. suis pathogenesis.
S. suis commonly causes acute clinical symptoms in humans (31), suggesting that this bacterium is able to quickly and effectively thwart the innate immune response before the activation of the adaptive immune response. Our study demonstrated that LL-37, as the first line of the innate immune system, could be stimulated within a few minutes after S. suis infection. Intriguingly, the increased concentration of LL-37 further up-regulated the transcript expression of the apdS gene, implying that S. suis could efficiently adapt to the host environment in response to LL-37, and it may employ the ApdS protease to resist the killing by LL-37. Notably, although the action of LL-37 is fast and only takes a few minutes to kill bacteria (32), the proteolytic activity of ApdS protease is even more rapid, and the cleavage of LL-37 is completed within merely a few seconds, further illustrating the importance of ApdS for S. suis in combating the early immune response.
Our results showed that the cleavage of LL-37 by ApdS protease results in the removal of three amino acids (LLG) from LL-37, and this truncated peptide (LLGCP) has been found to be less active than LL-37 against S. suis. Like other cationic

Streptococcus suis evades LL-37 and human neutrophils
peptides, the mechanism exploited by LL-37 to kill bacterial cells is the perforation of their cytoplasmic membranes. The initial LL-37-membrane interactions are mediated through electrostatic properties that the net positive charge facilitates LL-37 to interact with the net negative charge of bacterial surfaces, containing lipopolysaccharides or lipoteichoic acids (1). Thus, the changes in net charge of a peptide can influence its electrostatic affinity for bacterial cell surfaces, which further influences its bactericidal activity. At neutral pH, LL-37 is highly charged with net charge of ϩ6, but after cleavage by ApdS, the net charge of LLGCP remains ϩ6, suggesting that the reduced bactericidal activity of LLGCP is not related to the net charge of this peptide. In addition to charge, the structural conformation of peptide is also a major factor for its bactericidal activity (33,34). LL-37 presents a disordered conformation in aqueous solutions but forms an ␣-helical conformation in the presence of physiological salt concentrations (22). Our data of the CD spectra revealed that LLGCP experienced the same structural transition as LL-37 from the disordered to helical structure. In the membrane-mimetic environments, this conformation is further stabilized as indicated by an increased helical content. However, in either the physiological environment or the membrane-mimetic environments, the helical content of LLGCP is always lower than that of LL-37. These results adequately explain why the bactericidal activity of LL-37 is reduced after the cleavage by ApdS. In line with this, it has been reported that the extent of helicity correlates with the antibacterial activity of LL-37 (22,23). The truncated peptide of LL-37, which contains four residue deletions at the N terminus, has been showed to be less helical, leading to a decrease in bactericidal activity when compared with that of LL-37 (22). Taken together, with the help of ApdS protease, S. suis avoids and resists the LL-37-mediated killing, and the decreased bactericidal activity of the cleaved peptide LLGCP is attributed to the diminished helical content of its secondary structure.
Until now, many pathogens have been found to successfully evade the LL-37-mediated killing by producing surface-associated or secreted proteases. A metalloprotease of Staphylococcus aureus, aureolysin, inactivated the antimicrobial peptide LL-37 by cleaving the C-terminal peptide bonds and promoted the bacterial survival within the LL-37-rich environment of macrophage phagolysosomes (35). Similarly, a metalloprotease of Tannerella forsythia, karilysin, also cleaved LL-37 and significantly reduced its bactericidal activity (36). Rapala-Kozik et al. (37) had found that the pathogenic Candida albicans could effectively use aspartic proteases to destroy the antimicrobial and immunomodulatory properties of LL-37, thus enabling the pathogen to survive and propagate. In S. pyogenes, the proteolytic inactivation of LL-37 by the cysteine protease SpeB was also reported (38). Additionally, pathogenic bacteria also have other LL-37 resistance mechanisms. The M1 surface protein from group A Streptococcus was found to contribute to the LL-37 resistance of this bacterium through sequestering and neutralizing LL-37 (39). In addition, efflux pumps are used by many bacterial pathogens to resist LL-37 by extruding it from the cell membrane site of action to the extracellular environment. The MefE/Mel efflux pump contributed to LL-37 resistance in Streptococcus pneumoniae (40), whereas the Pmt ABC transporter of S. aureus was shown to be essential for defending the bacteria from killing by LL-37 (41).
The antimicrobial peptide resistance mechanisms of S. suis have remained unclear until now. Only the study of Fittipaldi et al. (42) reported that knockout of the dltA gene in S. suis gives rise to the absence of lipoteichoic acid in D-alanylation, leading to increased susceptibility to the bacterium-derived peptide colistin, polymyxin B, and the frog-derived peptide magainin II. The lipoteichoic acid D-alanylation has been found to reduce the net negative charge of lipoteichoic acids and to relieve the electrostatic attraction between cationic antimicrobial peptide and the bacterial cell envelope (43). Although not described in the study of Fittipaldi et al. (42), we speculate that D-alanylation of lipoteichoic acids also blunts the access of LL-37 to the surface of S. suis. Nonetheless, the lavish amounts of LL-37 released by neutrophils still surround S. suis and further attract more and more neutrophils toward the infection site to engulf the bacteria. Interestingly, our study revealed a novel ApdS protease-driven mechanism exploited by S. suis to impede the chemotaxis of neutrophils. The cleavage of LL-37 by the ApdS protease not only directly impairs its chemotactic activity, but it also indirectly hinders the migration through the down-regulation of the CXCL8 production, which is a key chemokine for neutrophils.
Of note, LL-37 is also able to facilitate the formation of NETs, which is a novel anti-microbial defense strategy against pathogenic bacteria (11,44,45). Although the underlying mechanism of LL-37 for formation of NETs is still unknown, we found that a high concentration of LL-37 indeed promotes the formation of NETs, consistent with previous research (11). In fact, the concentrations of LL-37 released at infection sites through neutrophil degranulation are much higher than the concentration in our study used for formation of NETs, and NETs are likely to be found at places with a high microbial burden (26,27,45). Thus, we surmise that the neutrophils can estimate the quantity of bacteria by sensing the concentration of LL-37 stimulated by bacteria. Once bacteria reach the quantity that LL-37 can no longer control, neutrophils may initiate NETs' formation through an unknown mechanism to imprison the bacteria. A recent study (46) also demonstrated that LL-37 could act as a second signal to induce inflammasome activation in airway epithelial cells and promote an inflammasome-mediated altruistic cell death of Pseudomonas aeruginosa-infected epithelial cells. It is clear that LL-37 acts as a "fire alarm" to enhance rapid escalation of inflammatory responses to an uncontrolled bacterial infection (46). Furthermore, different LL-37-derived truncated peptides have been demonstrated to exhibit significant differences in the NET formation (11). Our study found that the cleavage of LL-37 by ApdS protease weakened the ability of LL-37 for NETs' formation, and the apdS deletion increased the sensitivity of S. suis to NETs' killing. Since it has been reported that the WT S. suis and different nuclease mutants did not show significant differences in bacterial survival within NETs (29), it seems that the nucleases cannot effectively help S. suis to escape from NETs. Thus, NETs may be protected by LL-37 against bacterial nucleases, and the survival of S. suis within NETs may result from its ability to resist the killing by LL-37.

Streptococcus suis evades LL-37 and human neutrophils
At present, antibiotic resistance of pathogenic bacteria to conventional drugs is already a global health concern. To cope with this grim situation, antimicrobial peptides are being considered as important candidates for developing novel antibiotics (47), and many of them now undergo preclinical or clinical development, including LL-37 (48). Additionally, designing and improving the antimicrobial peptides for the optimal therapeutic application is an ongoing task (49). Amino acids Arg and Lys, both of which carry positive charges at neutral pH values, are commonly used to design antimicrobial peptides (50). We found that the ApdS protease of S. suis can also cleave Arg and Lys from the N terminus of the peptide (data not shown), thus reminding us that these amino acids need to be arranged to the suitable location or modified when designing antimicrobial peptides for therapy of the S. suis infection.
In summary, this study provides evidence of a novel strategy of immune evasion by which S. suis manipulates LL-37 and neutrophils, thwarting the innate immunity. The key event is targeting and cleaving the LL-37 to prejudice its bactericidal activity and immunomodulatory properties for human neutrophils. These findings also inspire the consideration for the current design strategies of antimicrobial peptides needs to be improved for boosting its therapeutic application to avoid being damaged by microorganisms, as well as to enhance its immunomodulatory function.

Bacterial strains and growth conditions
The bacterial strains and plasmids used for this study are described in Table 1. The S. suis WT strain used in this study was serotype 2 strain 05ZYH33 (GenBank TM accession no. CP000407.1), which was isolated from a deceased patient in Sichuan Province, China, in 2005. The S. suis strains were cultured in a Todd-Hewitt broth (THB, Difco Laboratories, Detroit, MI) medium or plated on THB agar. For the preparation of competent cells, S. suis strains were grown in THB supplemented with 2% yeast extract (THY). E. coli strains were cultured in Luria-Bertani medium (LB, Difco Laboratories) or plated on LB agar. When necessary, antibiotics were added to the plate or broth at the following concentrations: chloramphenicol (Sigma), 5 g/ml for S. suis and 10 g/ml for E. coli; spectinomycin (Sigma), 100 g/ml for S. suis and 50 g/ml for E. coli; erythromycin (Sigma), 5 g/ml for S. suis and 180 g/ml for E. coli; and ampicillin (Sigma), 100 g/ml for E. coli. All of the strains were routinely grown at 37°C.

Cell and culture conditions
THP-1 macrophages (human leukemic cell line) were grown in RPMI 1640 medium (Gibco). Human neutrophils were isolated from freshly-collected whole-blood samples (EDTAtreated) obtained from healthy volunteers using the Polymor-phPrep system (Axis Shield), as described previously (39), and resuspended in RPMI 1640 medium. For isolation of human neutrophils, written informed consents were obtained from the donors in accordance with the declaration of Helsinki (2013) of the World Medical Association. The medical ethics committee of the First Affiliated Hospital of Dalian Medical University approved the use of human blood for this study (No. LCKY2016-35(X)).

LL-37 production analysis
Neutrophils were seeded into the glass-bottom culture dish at a density of 1 ϫ 10 5 cells per dish, then infected with S. suis at a m.o.i. of 0.1, 1, and 5, and incubated at 37°C for 40 min. After washing with PBS, the cells were fixed with 4% paraformaldehyde in PBS and permeabilized with 0.2% Triton X-100. Nonspecific binding was blocked by incubation with 3% BSA in PBS for 60 min. Then cells were incubated with rabbit anti-LL-37 (Anaspec) and mouse anti-lactoferrin (Abcam) for 60 min at 37°C. After washing three times, the cells were stained with Alexa Fluor 488 -conjugated goat anti-rabbit IgG, Alexa Fluor 647-conjugated goat anti-mouse IgG and co-stained with 4Ј,6diamidino-2-phenylindole (DAPI) (Sigma) to visualize nuclei. Fluorescence images were obtained using the LSM 800 confocal laser-scanning microscope (Carl Zeiss, Germany). Quantification of area stained for LL-37 was performed using ImageJ software (National Institute of Health, Bethesda).

In vitro growth assays
The S. suis strain 05ZYH33 was grown to mid-logarithmic phase in THB medium and then diluted to an optical density at 600 nm (OD 600 ) of 0.1 in 4 ml of THB medium supplemented with LL-37 (GL peptide Inc., purity Ͼ98%, concentration range 0 -1 M). Bacterial growth at 37°C was monitored by measur-

Streptococcus suis evades LL-37 and human neutrophils
ing the OD 600 values at an interval of 1 h using an Eppendorf BioPhotometer (Eppendorf, Germany).

qRT-PCR
S. suis strain 05ZYH33 was grown in 4 ml of THB medium supplemented with LL-37 (concentration range 0 -0.4 M). The cultures were harvested at OD 600 1.0 by centrifugation at 8,000 ϫ g at 4°C for 10 min. Total RNA was extracted from each sample using the RNeasy mini kit (Qiagen), and cDNAs were synthesized using the PrimeScript RT reagent kit (TaKaRa, Japan). Transcript amounts of apdS were determined in a Stratagene Mx3000P system (Agilent Technologies, Germany) by quantification relative to the housekeeping gene gyrB (Table 2), as described previously (51). qRT-PCRs were performed in a MicroAmp Optical 96-well reaction plate using SYBR Green I (TaKaRa) according to the manufacturer's instructions. All of the reactions were performed in triplicate with three independent biological replicates. Relative expression levels were determined using the comparative threshold cycle (⌬⌬Ct) method to calculate the fold change in gene expression (52).

Bioinformatics analysis
The signal peptide cleavage site was predicted using SignalP 4.0 program (53). Multiple protein sequences were aligned using ClustalW software (54) and presented by ESPript program (55). The active catalytic site was predicted was using MEROPS database (56). Protein structure homology modeling was constructed by SWISS-MODEL (57) and presented by PyMOL (58).

Protein expression and purification
The coding sequence of S. suis ApdS residues Cys-26 -Leu-574 was PCR-amplified from the 05ZYH33 genomic DNA using primers apdSF/R ( Table 2). The DNA fragments and pET22b vector (Novagen) were digested with NdeI and XhoI (Takara) for ligation to each other. The construction of recombinant plasmid pET22b-apdS was confirmed by sequencing and transformed into competent E. coli BL21 (DE3) cells. The protein expression was induced for 15 h at 16°C by the addition of 0.5 mM isopropyl 1-thio-␤-D-galactopyranoside in LB broth containing 50 g/ml ampicillin. The His 6 -tag fusion proteins were purified by Ni-Sepharose TM 6 Fast Flow column (GE Healthcare). Protein concentrations were determined using a bicinchoninic acid protein assay kit (Beyotime, China).
For site-directed mutagenesis, the DNA segment, including the mutation of Cys-26 to Ala, was amplified from the plasmid pET22b-apdS using primers C26AF and apdSR ( Table 2). The digested PCR product was ligated with NdeI/XhoI-digested pET22b, and the construct was confirmed by DNA sequencing. The mutated protein ApdS(C26A) was expressed and purified as described above.

Enzymatic assay
Amidolytic activities of ApdS protease was measured using the substrate L-Leu-pNa, Gly-pNa, L-Pro-pNa, and Gly-Pro-pNa (Sigma, 1 mM final concentration) in PBS buffer (pH 7.4) at 37°C. Assays were performed in 100 l on 96-well plates using a thermostated microplate reader, and the release of p-nitroaniline was measured at 405 nm (BioTek). Kinetic values were measured using L-Leu-pNa and Gly-pNa at various concentrations ranging from 31 M to 2 mM, with a fixed enzyme concentration at 30 nM, in PBS buffer (pH 7.4) at 37°C. V max and K m values were obtained through hyperbolic regression analysis.
For proteolytic degradation of LL-37, 2 M LL-37 was incubated with 20 nM ApdS and ApdS(C26A) in 100 l of PBS buffer. Control reaction lacked ApdS to obtain uncleaved   Reactions were incubated at 37°C for 20 min and submitted for MS analysis using an AB SCIEX 5800 MALDI-TOF/ TOF mass spectrometer (AB Sciex, Canada).

Construction of gene deletion mutant
The primers used for the construction of the deletion mutant ⌬apdS are listed in Table 2. The DNA segments corresponding to the upstream and downstream regions of the apdS gene were amplified from the genomic DNA of WT strain 05ZYH33 using primers UF/UR and DF/DR carrying PstI/XhoI and BamHI/ EcoRI restriction enzyme sites, respectively. The Spc r gene cassette was amplified from plasmid pSET2 (59) with primers spcF/R carrying XhoI/BamHI restriction enzyme sites. The three digested DNA fragments were ligated by incubation with T4 DNA ligase for 6 h at 22°C, and PCR was carried out with primers UF and DR. The DNA fragment was digested with PstI/ EcoRI and cloned into the plasmid pSET6S to produce the plasmid pSET6S⌬apdS. To obtain the isogenic mutant ⌬apdS, the pSET6S⌬apdS was then electroporated into the WT strain 05ZYH33, as described previously (60). After homologous recombination, the apdS gene was replaced by the Spc r cassette, and the S. suis ⌬apdS mutant was verified by sequencing and PCR assays with a series of specific primers listed in Table 2. The transcription levels of the downstream genes of apdS in WT and ⌬apdS were analyzed using qRT-PCR with specific primers ( Table 2).

Complementation of the S. suis ⌬apdS mutant
The 1,974-bp DNA fragment containing the entire apdS open reading frame (ORF) and 150 bp of the upstream promoter were amplified with the primers CF/CR ( Table 2). The 1,185-bp DNA fragment containing the entire erm open reading frame (ORF) and the upstream promoter were amplified with the primers ErmF/R. Both amplicons were subsequently cloned into E. coli-S. suis shuttle plasmid pSET2 (51), yielding the recombinant plasmid pSET2-erm-apdS. This plasmid was confirmed by DNA sequencing (Comate Bioscience Co., Ltd.) and transformed into the ⌬apdS mutant for trans-complementation. The complemented strain C⌬apdS was selected on THB agar supplemented with 5 g/ml erythromycin and verified by PCR and DNA sequencing using the primers CF/CR and ErmF/R. The transcription and protein expression levels of ApdS in WT, ⌬apdS, and C⌬apdS were analyzed using qRT-PCR with specific primers (Table 2) and Western blotting.

Western blotting
Protein samples fractionated by SDS-PAGE were transferred onto nitrocellulose membranes (GE Healthcare). Membranes blocked with 5% BSA in PBST were incubated overnight at 4°C with an appropriately diluted primary antibody in PBST containing 2% BSA. After incubation with DyLight 800 goat antirabbit IgG (HϩL), blots were visualized by using an Odyssey IR imaging system (Li-Cor BioSciences, Lincoln, NE).

LL-37 bactericidal assays
LL-37 bactericidal assays were performed as described previously (62). The S. suis WT strain 05ZYH33, ⌬apdS mutant, and the complemented strain C⌬apdS were grown in THB medium at 37°C to OD 600 0.8. Each strain from the broth cultures was harvested and diluted in PBS (pH 7.4) to a concentration of 1 ϫ 10 4 CFU/ml. The wells of a sterile polystyrene 96-well microtiter plate (Costar 3599) were filled with 90 l of the various concentrations (0 -4 M) of LL-37. Ten microliters of the bacterial suspension were added to each well, and the plate was incubated for 0 -60 min at 37°C. Bacteria incubated with PBS served as controls. Serial dilutions of the bacteria were plated on THB agar for CFU counting.

SYTOX Green uptake assay
The S. suis ⌬apdS mutant was grown to OD 600 0.8. The strain from the broth culture was washed three times and diluted to 1 ϫ 10 5 CFU/ml in PBS (pH 7.4). The bacteria were incubated with SYTOX Green (Invitrogen) for 15 min in the dark, after which they were mixed with the appropriate concentrations (0 -4 M) of LL-37 and LLGCP (GL peptide Inc., purity Ͼ98%) for 30 min. Fluorescence was measured using an EnSpire Multilabel Reader (PerkinElmer Life Sciences, Singapore), with excitation at 485 nm and emission at 520 nm.

Circular dichroism spectroscopy
The secondary structures of LL-37 and LLGCP in different environments were measured using CD spectroscopy. The LL-37 and LLGCP peptides were respectively analyzed at an 8 M concentration in deionized water, PBS (pH 7.4), 50% TFE, and 30 mM SDS. The CD spectra were recorded as the average of three scans at 1-nm intervals between 198 and 250 nm using a CD spectrometer (Chirascan, Applied Photophysics, UK) with a 1-cm path length quartz cuvette. The helical content of these peptides in different environments was analyzed using the BeStSel tool (63).

Phagocytes bactericidal assays
The bactericidal assays were used to compare the growth of WT, ⌬apdS mutant, and C⌬apdS in human blood, THP-1 macrophages, and a neutrophil killing assay as described previously (39,64). Briefly, diluted cultures of strains (50 l, 5 ϫ 10 4 CFU/ml) were combined with 450 l of fresh human blood, and the mixtures were incubated for 0 -1.5 h at 37°C. The viable cell counts were determined by plating diluted samples onto THB agar. The percent of live bacteria was subsequently calculated using the CFU of bacteria in the original inoculum as 100%.
For THP-1 killing, THP-1 cells were treated with 50 nM vitamin D (active form, 1,25-dihydroxyvitamin D 3 ) for 36 h and then infected with S. suis strains at m.o.i. of 1:50 for 3 h at 37°C under 5% CO 2 . In neutrophil-killing assays, neutrophils were infected with S. suis strains at m.o.i. of 1:50 for 3 h. For NETs killing, neutrophils were stimulated with 25 nM PMA (Sigma) 4 h pre-infection. Serial dilutions of the bacteria were plated on THB agar. Colonies were counted, and the percentage of surviving bacteria was calculated using the CFU of bacteria incubated in RPMI 1640 medium as 100%.

Dot blot
To determine the concentrations of LL-37 in human plasma, THP-1 macrophages, neutrophils, and NETs during S. suis infection, dot-blot analysis was performed as described previ-

Streptococcus suis evades LL-37 and human neutrophils
ously (26,28). Briefly, after treatment, the cells were pelleted by centrifugation at 800 ϫ g for 20 min. 3 l of each supernatant sample were dotted onto a nitrocellulose membrane, immunolabeled with rabbit anti-LL-37 (Anaspec), followed by DyLight 800 goat anti-rabbit IgG (HϩL). Signal intensity of each blot was quantified using an Odyssey IR imaging system (Li-Cor BioSciences, Lincoln, NE). Standard peptide concentrations were used to calculate the concentrations in the biological samples.

Lysosome formation analysis
Neutrophils were seeded into the glass-bottom culture dish at a density of 5 ϫ 10 4 cells per dish and then incubated with 0 -2 M LL-37 and LLGCP for 4 h at 37°C under 5% CO 2 . The procedures of fixation, permeabilization, and blocking were the same as described in the experiment for LL-37 production analysis. The cells were incubated with mouse anti-LAMP1 (Abcam) for 60 min at 37°C and then stained with Alexa Fluor 647-conjugated goat anti-mouse IgG and DAPI. Fluorescence images were obtained using the LSM 800 confocal laser-scanning microscope (Carl Zeiss, Germany). Quantification of area stained for lysosome was performed using ImageJ software.

Neutrophil chemotaxis assay
Neutrophil migration was evaluated by the 24-well microchemotaxis chamber technique as described previously (37,65). Briefly, neutrophils were labeled with CellTracker Green (Invitrogen) and washed three times. Then, 1 ϫ 10 6 cells were seeded in the upper chambers of 5-m pore Transwell inserts (Corning Life Sciences). Chemotactic samples of the indicated concentrations of LL-37 and LLGCP were placed in RPMI 1640 medium in the lower chambers, with the 5 M fMLP (Sigma) used as a positive control. Following incubation for 40 min at 37°C under 5% CO 2 , the neutrophil migration was monitored by fluorescence microscopy (Zeiss).

CXCL8 production assay
Neutrophils (1 ϫ 10 6 cells/well) were suspended with 500 l of RPMI medium in 24-well plates. The cells were incubated with various concentrations (0 -8 M) of the LL-37 and LLGCP for 6 h at 37°C under 5% CO 2 . Supernatants were collected and placed in a sterile 96-well plate for ELISA. CXCL8 production was determined by an ELISA kit (Sigma) according to the manufacturer's instructions.

NETs' formation assay
Neutrophils were seeded into the glass-bottom culture dish at a density of 2 ϫ 10 5 cells per dish pre-coated with poly-Llysine and then incubated with 0 -8 M LL-37 and LLGCP for 5 h at 37°C under 5% CO 2 . As a positive control, 25 nM PMA was used to trigger the formation of NETs. The procedure of fixation, permeabilization, and blocking was the same as described in the experiment for LL-37 production analysis. The cells were incubated with mouse anti-DNA/histone 1 (Merck Millipore) for 60 min at 37°C and then stained with Alexa Fluor 488 -conjugated goat anti-mouse IgG and DAPI. Fluorescence images were obtained using the LSM 800 confocal laser-scan-ning microscope (Carl Zeiss, Germany). Quantification of area stained for NETs was performed using ImageJ software.

ROS production assay
The level of cellular ROS was measured using the fluorescence probe DCFDA (Sigma), as described previously (10). Briefly, human neutrophil cells (2 ϫ 10 5 ) were treated with 0 -8 M LL-37 and LLGCP at 37°C under 5% CO 2 . After treatment, the cells were washed with PBS three times and stained with 5 M DCFDA for 15 min in the dark. Cells were washed three times with PBS, and the fluorescence intensity of the cells was read using an EnVision Multilabel Reader (PerkinElmer Life Sciences, UK), with an excitation wavelength of 485 nm and an emission wavelength of 520 nm to represent ROS activity.

Statistical analysis
Statistical analyses were performed using GraphPad Prism version 7.0 (GraphPad Software, La Jolla, CA). Data were analyzed by performing an unpaired two-tailed Student's t test or one-way analysis of variance. p values less than 0.05 were considered statistically significant.