General Phenotype of NADase Induction by CLI Treatment in Streptococcus pyogenes

The administration of high-dose clindamycin (CLI) along with penicillin is recommended for the treatment of streptococcal toxic-shock syndrome (STSS). However, we previously reported that a “subinhibitory dose” of CLI induced the expression of the NAD-glycohydrolase (NADase) exotoxin in an emm1-type Streptococcus pyogenes 1529 strain isolated from an STSS patient. In this study, we examine NADase induction by CLI treatment using an extracellular NADase activity assay instead of the previous two-dimensional gel electrophoresis assay. The examination revealed that CLI administration can induce NADase expression in a dose-dependent manner. We analyzed 23 CLI-susceptible strains (5 emm1 strains, 6 emm3 strains, 3 emm4 strains, 1 emm6 strain, 3 emm12 strains, 1 emm28 strain, and 4 emm89 strains), and 19 of the 23 strains showed similar NADase induction phenotypes to that shown in strain 1529. These results indicate that NADase induction by CLI treatment is not restricted to specific strains and it could be a standard phenotype among CLI-susceptible S. pyogenes strains. We also analyzed four CLI-resistant strains. All four strains showed increased extracellular NADase activities at high concentrations of CLI that did not inhibit bacterial growth. These results indicated that the subinhibitory dose of CLI was not the critical factor for NADase induction.


Introduction
Streptococcus pyogenes is a Gram-positive bacterium that infects the upper respiratory tract, including the tonsils and pharynx, and it is responsible for postinfection diseases, such as rheumatic fever and glomerulonephritis. S. pyogenes also causes severe invasive diseases, including necrotizing fasciitis and streptococcal toxic-shock syndrome (STSS) [1][2][3][4][5].
e administration of high-dose clindamycin (CLI) along with penicillin is recommended for the treatment of STSS. However, in our previous study, the levels of two exoproteins, streptolysin O (Slo) and NAD (+)-glycohydrolase (Nga or NADase), were increased by CLI treatment (named "CLI-dependent Slo/NADase induction") in three of the five STSS strains analyzed by two-dimensional gel electrophoresis [6]. e NADase, the expression of which is negatively regulated by the CovRS two-component regulatory system [7,8], is an important virulence factor for STSS [9]. We also have studied NADase [10][11][12]. In this study, we examined the CLI-dependent NADase induction using a more quantitative extracellular NADase activity assay than the previous two-dimensional gel electrophoresis assay. Consequently, the examination provided new insights into CLI-dependent NADase induction.

Determination of emm Type.
Typing of the M proteincoding gene (emm) of streptococcal isolates was performed according to the methods by CDC Streptococcus Laboratory (https://www.cdc.gov/streplab/protocol-emm-type.html) as described previously [14].

Quantification of NADase Activity in Bacterial Supernatants.
A 40 μL sample of the overnight culture was added to 4 ml of fresh BHI-YE broth containing CLI at various concentrations, and the samples were cultured for 18 h. After 18 h, the optical density at 660 nm (OD 660 ) was measured. e NADase activity of the culture supernatants was determined using the method of Stevens et al. [15] as described previously [10].

Effects of CLI on Extracellular NADase Activity in emm1
Strains. In the previous study [6], a subinhibitory dose of CLI induced the expression of NADase in an emm1 Streptococcus pyogenes 1529 strain isolated from an STSS patient. e subinhibitory dose was defined as the maximum antibiotic concentration that did not suppress bacterial growth in the previous study [6]. However, it is unknown whether the subinhibitory dose is the best concentration to induce the expression of NADase because other concentrations were not tested. erefore, we examined the previous study using various CLI concentrations ( Figure 1 and Table 1). e bacterial growth was measured using the OD 660 (see Materials and Methods for detail), which was the method used in the previous study [6]. When strain 1529 was cultured for 18 h without CLI (0 μg/ml), the average OD 660 was 1.03. When cultured with 2 −8 , 2 −7 , 2 −6 , and 2 −5 μg/ml CLI, the average OD 660 values were 1.05, 0.98, 0.88, and 0.31, respectively. e OD 660 values listed were 101%, 95%, 86%, and 30% of the OD 660 without CLI, respectively ( Figure 1). e subinhibitory dose was newly defined as the maximum CLI concentration that did not suppress the bacterial growth to less than 50% of that without CLI. e new subinhibitory dose was 2 −6 (≈1.6 × 10 −2 ) μg/ml in strain 1529 (shown in red and black in Figure 1 and Table 1, respectively). e extracellular NADase activities were increased from 2 −8 to 2 −6 μg/ ml CLI in a dose-dependent manner, and the NADase activity reached the maximum in the 2 −6 μg/ml CLI, the new subinhibitory dose (Figure 1).
Strain 1529 was the only emm1-type strain analyzed in the previous study [6]. In the current study, we analyzed an additional four emm1 strains (11T-3, 10-85, 11-171, and 12-5) isolated from STSS or non-STSS patients, and these strains showed similar NADase induction phenotypes to those shown in strain 1529 ( Figure 1 and Table 1).  Figure 1: Extracellular NADase activity in emm1-type strains cultured with clindamycin (CLI). One unit of NADase activity was defined as the amount (μg) of β-NAD cleaved per hour per μL of culture supernatant, as described previously [10]. e OD 660 of overnight culture with CLI (2 −8 to 2 −4 μg/ml) is shown by a percentage (%) of that without CLI (0 μg/ml). e maximum CLI concentration that did not suppress the bacterial growth to less than 50% of that without CLI is shown in red, which was 2 −6 μg/ml in strains 1529, 11T-3, 10-85, and 11-171, and 2 −5 μg/ml in strain 12-5. At least three independent experiments were performed. Error bars indicate the standard errors of the means. Significant increases in extracellular NADase activity by CLI treatment when compared with that without CLI treatment, shown by the black bar, are marked with asterisks ( * p < 0.05 using Student's t-test).

Effects of CLI on Extracellular
NADase Activity in emm3, emm4, and emm12 Strains. In the previous study [6], not only strain 1529 but also four additional strains (1268-emm3, 1266-emm4, 1547-emm5, and GG01-emm12) were analyzed. e 1547-emm5 and GG01-emm12 strains showed CLI-dependent NADase induction, whereas strains 1268-emm3 and 1266-emm4 did not. In order to investigate how the difference between the former two and the latter two strains in the NADase induction phenotypes is related to their emm-genotypes, an additional twelve strains (six emm3, three emm4, and three emm12) were analyzed, although we could not collect additional emm5 strains. Among the six emm3 strains, two strains (ncu6B and ncu62 A) showed CLI-dependent NADase induction, although the other four strains (SSI-1, 13-O-8, 13-O-10, and 15-T-11) did not ( Figure 2 and Table 1). All emm4 and emm12 strains showed CLI-dependent NADase induction ( Figure 3 and Table 1). ese results indicated that the emm3-and emm4-genotypes were not attributable to "no NADase induction" phenotypes that were previously reported for the 1268-emm3 and 1266-emm4 strains [6].

Effects of CLI on Extracellular
NADase Activity in the emm6, emm28, and emm89 Strains. We analyzed one emm6, one emm28, and four emm89 strains. All six strains showed the CLI-dependent NADase induction phenotype ( Figure 4 and Table 1).

Effects of CLI on Extracellular NADase Activity in emm89
Strains Having an Inactive CovS Allele. Recently, we simultaneously isolated five S. pyogenes strains from the pharynx, sputum, knee joint, cerebrospinal fluid, and blood of a single STSS patient [18]. All five strains of S. pyogenes were derived from a single emm89 clone. e three strains from the knee joint, cerebrospinal fluid, and blood, but not the other two strains, contained a mutation in the covS gene to lose its function. As shown in Figure 5, the two strains from the pharynx and sputum showed the CLI-dependent NADase induction phenotype, whereas the other three strains, which had an inactivated covS allele, did not. Basal (in the absence of CLI) levels of NADase activities. Bacteria were cultured without CLI. U: Units, one unit of NADase activity is defined as the amount (μg) of β-NAD cleaved per hour per μl culture supernatant, as described previously [10]. b Whether the extracellular NADase activity of each strain was increased with CLI treatment (+) or not (−) was determined based on the experimental results shown in Figure 1 to 6. CLI-dependent NADase activity induction was expressed as values of maximum fold-induction by CLI in parentheses. N/A, not applicant.  Figure 2: Extracellular NADase activity in S. pyogenes emm3-type strains cultured with CLI. NADase activity was determined based on the criteria described in Figure 1. e OD 660 of overnight culture with CLI (2 −8 to 2 −5 μg/ml) is shown by a percentage (%) of that without CLI (0 μg/ml). e maximum CLI concentration that did not suppress the bacterial growth to less than 50% of that without CLI is shown in red, which was 2 −6 μg/ml in all the strains. At least three independent experiments were performed. Error bars indicate the standard errors of the means. Significant increases in extracellular NADase activity by CLI treatment when compared with that without CLI treatment, which is shown by the black bar, are marked with asterisks ( * p < 0.05 using Student's t-test).  Figure 3: Extracellular NADase activity in emm4-or emm12-type strains cultured with CLI. NADase activity was determined based on the criteria described in Figure 1. e OD 660 of overnight culture with CLI (2 −10 to 2 −5 μg/ml) is shown by a percentage (%) of that without CLI (0 μg/ml). e maximum CLI concentration that did not suppress the bacterial growth to less than 50% of that without CLI is shown in red, which was 2 −6 μg/ml in all three strains. At least three independent experiments were performed. Error bars indicate the standard errors of the means. Significant increases in extracellular NADase activity by CLI treatment when compared with that without CLI treatment, which is shown by the black bar, are marked with asterisks ( * p < 0.05 using Student's t-test).  Figure 4: Extracellular NADase activity in emm6, emm28, or emm89 strains cultured with CLI. NADase activity was determined based on the criteria described in Figure 1. e OD 660 of overnight culture with CLI (2 −9 to 2 −4 μg/ml) is shown by a percentage (%) of that without CLI (0 μg/ml). e maximum CLI concentration that did not suppress the bacterial growth to less than 50% of that without CLI is shown in red, which was 2 −7 μg/ml in strain NU-syoni, 2 −6 μg/ml in strains ncu28A and Seibu, 2 −7 μg/ml in strain ncu52 A, 2 −6 μg/ml in strain ncu60 A, 2 −5 μg/ml in strain ncu61A, respectively. At least three independent experiments were performed. Error bars indicate the standard errors of the means. Significant increases in extracellular NADase activity by CLI treatment when compared with that without CLI treatment, which is shown by the black bar, are marked with asterisks ( * p < 0.05 using Student's t-test). and (E) knee joint of a single patient. Bacteria were cultured with CLI (2 −9 to 2 −5 μg/ml). NADase activity was determined based on the criteria described in Figure 1. e OD 660 of overnight culture with CLI is shown by a percentage (%) of that without CLI. e maximum CLI concentration that did not suppress the bacterial growth to less than 50% of that without CLI is shown in red, which was 2 −7 μg/ml in all five strains. At least three independent experiments were performed. Error bars indicate the standard errors of the means. Significant increases in extracellular NADase activity by CLI treatment when compared with that without CLI treatment, which is shown by the black bar, are marked with asterisks ( * p < 0.05 using Student's t-test).

Effects of CLI on Extracellular NADase Activity in a CLI-Resistant emm1
Strain. In S. pyogenes, CLI-resistant strains have been identified worldwide. We reported that a CLIresistant strain, D2TY, induced the expression of NADase by treatment with 1 (�2 0 ) μg/ml CLI [17]. However, it is unknown whether the dose (2 0 μg/ml) of CLI was the best concentration to induce the expression of the NADase because other concentrations were not tested. erefore, we examined the previous study using various CLI concentrations. As shown in Figure 6, extracellular NADase activities were increased from 2 −6 to 2 2 μg/ml of CLI in a dose-dependent manner, and the NADase activity reached the maximum at a concentration of 2 2 μg/ml CLI, which was not the subinhibitory dose. Furthermore, surprisingly, extracellular NADase activities were decreased from 2 2 to 2 9 μg/ml CLI in a dose-dependent manner. In the 2 8 μg/ml CLI, the subinhibitory dose for this strain, the extracellular NADase activity was equivalent to that without  92  91  96  93  92  91  91  90  92  92  90  92  73  78  91  92  8  53 NADase activity (units) * * * * * * * * * * * * * * * * * Figure 6: Extracellular NADase activity in the four CLI-resistant emm1-or emm12-type strains. Bacteria were cultured with CLI (2 −6 to 2 9 μg/ml). NADase activity was determined based on the criteria described in Figure 1. e OD 660 of overnight culture with CLI is shown by a percentage (%) of that without CLI. e maximum CLI concentration that did not suppress the bacterial growth to less than 50% of that without CLI is shown in red, which was 2 8 μg/ml in all four strains. At least three independent experiments were performed. Error bars indicate the standard errors of the means. Significant increases in extracellular NADase activity by CLI treatment when compared with that without CLI treatment, which is shown by the black bar, are marked with asterisks ( * p < 0.05 using Student's t-test). 6 International Journal of Microbiology CLI treatment (p � 0.88).
ese results indicated that the subinhibitory CLI dose itself is not critical to increase the extracellular NADase activity in at least this D2TY strain.

Discussion
Regarding the CLI-susceptible S. pyogenes strains in this study, we updated the information about the CLI-dependent NADase induction phenotypes presented in the previous study [6], in the following regards: (i) we confirmed that the increases in extracellular NADase activity by CLI treatment were statistically significant (Table 1 and Figures 1-5); such statistical analysis was not performed in the previous study [6] because the previous method was not appropriate for statistical analysis (See Introduction section). (ii) We increased the number of strains tested from 5 in the previous study to 23 (Table 1); for example, the number of emm1 strains tested was increased from one to five. is enabled us to reveal that the CLI-dependent NADase induction could be a general phenotype in emm1-type S. pyogenes because all five emm1 strains tested showed the CLI-dependent NADase induction (Table 1 and Figure 1). e number of both emm4 and emm12 strains tested increased from one to three. Because all the strains tested showed the CLI-dependent NADase induction, it may be a general phenotype in emm4 and emm12 strains as well as the emm1 strain (Table 1 and Figure 3). (iii) We analyzed four emm89 strains, and all emm89 strains showed CLI-dependent NADase induction (Table 1 and Figure 4); emm89 strains were not analyzed in the previous study [6] because emm89-type strains were not prevalent at that time.
e four strains, SSI-1, 13-O-8, 13-O-10, and 15-T-11, that did not show the CLI-dependent NADase induction phenotype were all emm3 genotypes (Table 1 and Figure 2). However, we do not think the emm3 genotype itself is a critical factor for the no NADase induction phenotype, because the two emm3 strains, ncu6B and ncu62A, showed the NADase induction phenotype (Table 1 and Figure 2). e four strains having the no NADase induction phenotype showed higher NADase activities than the remaining 19 strains (especially, SSI-1 was highest) when they were cultured without CLI (Table 1 and Figure 2). is higher NADase activity observed in the SSI-1 is caused by a rocA gene mutation [19,20]. Because RocA represses the production of NADase through its ability to enhance CovR/S system function [21], the phenotype caused by a rocA mutation becomes similar to that caused by a covS mutation. In fact, the three emm89 strains with covS mutations showed similar phenotypes to that shown in SSI-1 (Figures 2 and 5).
Regarding the CLI-resistant S. pyogenes strains, we also updated the information about the CLI-dependent NADase induction presented in the previous study [17] in the following regards: (i) We confirmed that the increases in extracellular NADase activity by CLI treatment were statistically significant ( Figure 6); such statistical analysis was not performed in the previous study because the previous method was not appropriate for statistical analysis (see Introduction section). (ii) We increased the number of strains analyzed from one in the previous study to four, and all four strains showed the CLI-dependent NADase induction phenotype ( Figure 6). is enabled us to suggest that the CLI-dependent NADase induction could be a general phenotype in CLI-resistant S. pyogenes strains as well as in CLI-susceptible S. pyogenes strains. (iii) e subinhibitory doses of CLI themselves did not appear to be critical to increase the extracellular NADase activity. Especially, in strain D2TY, but not in the other three strains (11-174, ncu21A, and ncu55A) the treatment with the subinhibitory dose of CLI did not induce the extracellular NADase activity ( Figure 6).
is difference between strain D2TY and the other three strains could be related to their emm-types, because D2TY has the emm1-genotype and the other three strains have the emm12-genotype. In order to investigate this question, other emm1-type CLI-resistant strains should be investigated.
In this study, 19 of the 23 CLI-susceptible strains tested showed similar NADase induction phenotypes (Table 1 and Figures 1-4). ese results indicate that the NADase induced by CLI treatment is not restricted to specific strains and it could be a standard phenotype in S. pyogenes. Although the CovRS system is related to the mechanism for NADase induction, as described above (see also Figure 5), we do not believe that the CovS sensor protein has a direct interaction with CLI. For example, CLI may repress the translation of CovS and/or CovR to cause the derepression of nga. In our future work, we plan to provide evidence for this hypothesis.
Data Availability e data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.
International Journal of Microbiology 7