The antimicrobial activity of zinc against group B Streptococcus is strain-dependent across diverse sequence types, capsular serotypes, and invasive versus colonizing isolates

Background Streptococcus agalactiae or Group B Streptococcus (GBS) is an encapsulated gram-positive bacterial pathobiont that commonly colonizes the lower gastrointestinal tract and reproductive tract of human hosts. This bacterium can infect the gravid reproductive tract and cause invasive infections of pregnant patients and neonates. Upon colonizing the reproductive tract, the bacterial cell is presented with numerous nutritional challenges imposed by the host. One strategy employed by the host innate immune system is intoxication of bacterial invaders with certain transition metals such as zinc. Methodology Previous work has demonstrated that GBS must employ elegant strategies to circumnavigate zinc stress in order to survive in the vertebrate host. We assessed 30 strains of GBS from diverse isolation sources, capsular serotypes, and sequence types for susceptibility or resistance to zinc intoxication. Results Invasive strains, such as those isolated from early onset disease manifestations of GBS infection were significantly less susceptible to zinc toxicity than colonizing strains isolated from rectovaginal swabs of pregnant patients. Additionally, capsular type III (cpsIII) strains and the ST-17 and ST-19 strains exhibited the greatest resilience to zinc stress, whereas ST-1 and ST-12 strains as well as those possessing capsular type Ib (cpsIb) were more sensitive to zinc intoxication. Thus, this study demonstrates that the transition metal zinc possesses antimicrobial properties against a wide range of GBS strains, with isolation source, capsular serotype, and sequence type contributing to susceptibility or resistance to zinc stress.


Introduction
Group B Streptococcus (GBS), or Streptococcus agalactiae, infections are one of the top five leading causes of neonatal mortality. GBS infection induces chorioamnionitis, preterm prelabor rupture of the gestational membranes (PPROM), preterm birth, and both maternal and neonatal sepsis [1]. GBS disease in neonates often dissemination in a mouse model of invasive disease in the blood, heart, liver, and bladder. Together, these results indicate that zinc is an important innate immune antimicrobial strategy employed against GBS, and that GBS resistance to zinc toxicity is critical for full virulence.
While zinc has been implicated in modulating the immune response to infection [15], the role zinc plays in the context of perinatal-related GBS infections has not yet been elucidated. In this study, we advanced previous findings by analyzing the antimicrobial effects of zinc against a panel of clinical GBS strains that vary by capsular serotype, sequence type (ST), isolation source, and clinical presentation. We observed strain-specific variation in susceptibility to zinc intoxication with specific differences between invasive and colonizing strains and across different capsular serotypes and STs.

Bacterial strains and culture conditions
A bank of 27 GBS clinical strains were provided by Dr. H. Dele Davies following recovery from neonates with invasive disease [16] and colonized mothers sampled before and after childbirth [17] for use in this study (Table 1). These strains were all isolated from separate patients and were previously characterized using multilocus sequence typing and cps typing [18,19]. Three common laboratory reference strains (A909, NEM316, and COH1; American Type Culture Collection) were also evaluated. Bacterial strains were cultured on tryptic soy agar (TSA) plates supplemented with 5% sheep blood at 37 °C overnight followed by inoculation into brain-heart infusion broth (BHI) and incubation in aerobic conditions (ambient air, shaking at 200 rpm) at 37 °C. After 24 h, bacterial density was measured spectrophotometrically to determine the optical density at 600 nm (OD 600 ). These bacterial cultures were used for growth and viability assays.

Evaluation of bacterial growth
GBS growth was determined by a spectrophotometric reading of optical density (OD) at OD 600 as previously described [20]. Briefly, GBS cultures were grown overnight and diluted at 1:10 in fresh BHI medium; 100 μL of 1:10 diluted cultures were added to each well in a 96-well plate. Increasing concentrations of divalent zinc ions (Zn 2+ ) in the form of zinc chloride (ZnCl 2 ) (0, 125, 250, 500, 750, 1000, 2500, 5000 μM) were added to the culture media. These concentrations were chosen because they represent a range of physiologically relevant concentrations often encountered in the host-pathogen environment in vivo [11]. The plates were incubated statically in 5% CO 2 at 37 °C overnight. The following day, bacterial density was estimated via OD 600 . Three fresh biological replicates were assessed with 1-3 technical replicates within each biological replicate, and the OD 600 values were normalized to a blank control of sterile, uninoculated bacteriological medium (BHI).

Statistical analyses
Statistical analyses were performed using Mann-Whitney U for MIC studies, and either Student's t-test or one-way ANOVA with either Tukey's or Dunnett's post hoc correction for multiple comparisons for bacterial growth assays. All reported P values were adjusted to account for multiple comparisons. P values of ≤0.05 were considered significant. All data analyzed in this work were derived from at least three biological replicates. Statistical analyses were performed using GraphPad Prism 9 software (GraphPad Prism Software Inc., La Jolla, California).

High concentrations of zinc suppress bacterial growth in many clinical GBS isolates
Previous reports indicate that zinc has antimicrobial activities against GBS [14]. To enhance the generalizability of these findings, we sought to test a larger number of GBS strains, thereby capturing more isolates (both colonizing and invasive) across diverse capsular serotypes and genetic STs. We also investigated the effects of increasing zinc concentration exposure. Out of the 30 GBS strains screened, 4 strains (GB0083, GB0561, GB0651, and A909) exhibited significant inhibition of bacterial growth when treated with 125 μM zinc (Table 1; P < 0.05, Student's t test, compared to medium alone control cultures). All of these strains were classified as  colonizing strains except the single laboratory strain (A909). At 250 μM zinc, an additional colonizing strain (GB0571) exhibited significant inhibition of bacterial growth compared to cultures grown in medium alone (Table 1; P < 0.05, Student's t test). Growth of three additional strains, two colonizing strains and one laboratory strain (GB0012, GB0112, and NEM316, respectively), was inhibited when treated with a concentration of 500 μM zinc (Table 1; P < 0.05, Student's t test, compared to medium alone control cultures). At 750 μM zinc, strains GB0411 (invasive) and COH1 (laboratory) exhibited significant inhibition of bacterial growth compared to the medium only control (Table 1; P < 0.05, Student's t test). The growth of 5 additional strains (GB0037, GB0397, GB0438, GB0241, GB0653), including three invasive and two colonizing isolates, respectively, was inhibited when treated with 1000 μM zinc (Table 1; P < 0.05, Student's t test, compared to medium alone control cultures). The growth of 9 additional strains (GB002, GB0115, GB0285, GB0291, GB0654, GB0064, GB0069, GB0374, GB0418), including 5 colonizing and 4 invasive strains, respectively, was significantly inhibited when treated with 2500 μM zinc (Table 1; P < 0.05, Student's t test, compared to medium alone control cultures). Finally, the growth of six strains (GB0066, GB0079, GB0377, GB0390, GB0590, and GB0663), including four invasive strains and two colonizing strains, respectively, was unaffected when treated with concentrations of zinc up to 5000 μM (Table 1; P > 0.05, Student's t test, compared to medium alone control cultures).

GBS colonizing and invasive strain types differ in susceptibility to zinc intoxication
Because zinc has been shown to be a crucial antimicrobial strategy deployed by the innate immune system [10][11][12], we hypothesized that there could be differences in susceptibility to zinc intoxication between colonizing and invasive GBS strains. To test this, we stratified the GBS strains by clinical phenotype. Strains were classified as "colonizing" if they were recovered from asymptomatic women sampled before or after childbirth, whereas "invasive" strains were isolated from babies with GBS disease. Strains were exposed to increasing concentrations of zinc chloride and growth was measured after 24 h of static incubation in 5% CO 2 at 37 °C (Fig. 1). In medium alone, the mean OD 600 measurement was calculated as 0.34 for colonizing strains and 0.33 for invasive strains, results that were statistically indistinguishable (P = 0.1627, Mann-Whitney U test). Similarly, no significant difference was noted in raw OD 600 values between colonizing and invasive strains at 125, 250, 500, 750, 1000, 2500, or 5000 μM zinc exposure. However, calculation of percent growth of each strain (comparing growth of a specific strain at each zinc concentration compared to growth of that strain in medium alone) revealed that invasive strains exhibited a significantly enhanced mean growth when treated with 250, 500, 1000, and 2500 μM zinc (12,12,14,19, and 36%, respectively) compared to colonizing strains (P < 0.05, Mann-Whitney U test, Fig. 2 (Fig. 3), a 68% increase which was statistically significant (P = 0.0402, Mann-Whitney U test).

Susceptibility to zinc toxicity differs across GBS strains from varying isolation sources
Because differences were observed between invasive and colonizing strains, we hypothesized that the source of bacterial strain isolation could contribute to zinc intoxication susceptibility due to GBS adaptation to the ecology of the specific host niche. To test this, we stratified strains into categories based on source of strain isolation (Fig. 4) such as blood from neonatal early onset disease (EOD), late onset disease (LOD), rectal or vaginal swabs (Vaginal/Rectal), or adult blood from septic patients (Blood/Sepsis). Results indicate that at concentrations of 250, 500, 750, 1000, and 2500 μM zinc, isolates from EOD have significantly enhanced growth (12,13,13,27, and 39%, respectively) compared to vaginal/rectal isolates (P < 0.05, one-way ANOVA). At concentration of 250 μM zinc, LOD isolates were significantly more susceptible to zinc toxicity than those isolated from blood/sepsis (P < 0.05, one-way ANOVA). At 500 μM zinc, blood/sepsis isolates were significantly more tolerant of zinc stress than their vaginal/rectal or LOD isolate counterparts (P < 0.05, oneway ANOVA). At 750 μM zinc, blood/sepsis and EOD isolates were significantly less susceptible to zinc stress than rectovaginal isolates. At 1000 μM zinc, EOD and blood/sepsis isolates were significantly less susceptible than rectovaginal isolates. However, GBS isolated from LOD were significantly more susceptible than strains from the other three isolation sites.

GBS capsular serotype confers varying susceptibility to zinc intoxication
Capsular serotypes have been implicated as an important virulence factor that aids in evasion of the innate immune response [21]. Additionally, capsular serotype III strains are associated with higher rates of invasive neonatal disease [22] and account for the majority of late-onset meningitis cases in neonates [23]. Because invasive strains and isolates from EOD were less susceptible to zinc toxicity than rectovaginal colonizing strains, we hypothesized that capsular serotype variation could contribute to alterations in susceptibility to zinc toxicity.
To test this, we stratified strains based on capsular serotype and analyzed growth in cultures exposed to increasing concentration of zinc (Fig. 5). At concentrations as low as 125 μM of zinc, capsular serotype III (cpsIII) strains emerged as having enhanced growth compared to the cpsIb and cpsII strains. Notably, the cpsIII strains also exhibited enhanced growth at 250, 500, 750, 1000, and 2500 μM zinc compared to the cpsIb strains (P < 0.05, one-way ANOVA). At 2500 μM zinc, cpsIII strains exhibited the highest mean growth (mean OD 600 = 0.267), followed by strains with cpsIa (mean OD 600 = 0.246), cpsV (mean OD 600 = 0.200), cpsII (mean OD 600 = 0.126), and cpsIb (mean OD 600 = 0.107). The cpsVI strains were most sensitive to zinc toxicity at a concentration of 2500 μM (mean OD 600 = 0.100) compared to other capsular serotypes. However, at concentrations of 125, 250, 500, 750, and 1000 μM zinc, cpsIb isolates consistently exhibited the lowest growth, or greatest level of inhibition, among all capsular serotypes tested, underscoring their susceptibility to zinc intoxication (P < 0.05, one-way ANOVA). At a concentration of 5000 μM zinc, no statistically significant differences were observed between capsular types, a result that is likely due to a threshold effect of all strains experiencing significant growth inhibition. Because our results have shown that colonizing strains may be more susceptible to zinc toxicity, and the cpsIII strains exhibited low susceptibility to zinc toxicity, we stratified the cpsIII strains into invasive versus colonizing strains to ascertain if there were any differences in these two cohorts (Fig. 6). Mann-Whitney U analyses revealed no statistically significant difference between invasive and colonizing cpsIII strains (P = 0.4401).

GBS susceptibility to zinc toxicity varies across sequence types
Because different GBS sequence types (STs) are associated with maternal colonization and neonatal disease [24], it is possible that strains of different STs have variable mechanisms to facilitate metal homeostasis. To test this hypothesis, we stratified GBS strains by ST and analyzed the growth under varying concentrations of zinc (Fig. 7). Interestingly, the ST-19 isolates that exhibited the highest resistance to zinc intoxication were all classified as cpsIII strains.

Discussion
The battle for essential nutrient metals between the vertebrate host and invading pathogen has been closely linked to virulence [25]. Starvation of transition metals is detrimental to bacteria, however, high concentrations of metals like zinc also have antimicrobial effects [10,12]. Zinc plays an important antimicrobial role in innate immune defense against several pathogens, including a variety of Streptococcus spp. [23]. Bacteria that have the capability to survive and replicate inside the phagosome of macrophages employ and regulate metal transport machinery to maintain metal homeostasis. For example, the notorious intracellular pathogen, Mycobacterium tuberculosis, has been shown to upregulate P-type ATPases which act in heavy metal efflux to counteract the toxic, high extracellular zinc levels of the phagosome [26]. In fact, metal intoxication survival is a critical virulence trait in S. pneumoniae and S. pyogenes [27][28][29][30] as well as GBS [14]. In this study, we sought to determine the differences in susceptibility or resistance to zinc toxicity in 30 strains of GBS, spanning diverse capsular serotypes, STs, isolation source, and disease manifestation. Our work demonstrates that GBS colonizing and invasive strain types differ in susceptibility to zinc intoxication. Specifically, invasive strains exhibit diminished susceptibility to zinc toxicity compared to colonizing strains, indicating invasive strains may have acquired adaptations to survive metal intoxication strategies imposed within the vertebrate host during invasive infections. GBS strains from varying isolation sources also exhibit varying susceptibility to zinc intoxication. In particular, strains isolated from early onset disease manifestations have significantly enhanced tolerance of zinc intoxication compared to rectovaginal colonizing isolates. Again, similar to the results observed with colonizing versus invasive isolates, this could reflect strain-specific adaptation to the ecological niche of the host. The human vaginal environment is rich in S100A-family proteins, especially during infectious processes, and these proteins bind and sequester nutrient metals, such as zinc [31][32][33][34]. It is possible that the GBS-colonized vaginal mucosa represents an environment with low zinc availability, thus there is no selective pressure for colonizing strains to develop strategies to circumnavigate high zinc concentrations. Conversely, numerous studies have shown that circulating innate immune cells such as macrophages and neutrophils use zinc intoxication as a strategy to inhibit invading microbes [11,[35][36][37]. It is possible that invasive strains have undergone selection for survival in high zinc environments, yielding strains with greater resistance to the toxicity of this transition metal.
Our results indicate that GBS capsular serotype also confers varying susceptibility to zinc intoxication. Specifically, cpsIII isolates were less susceptible to zinc intoxication, whereas cpsIb isolates were much more susceptible to zinc toxicity. This is interesting because cpsIII is the predominant capsular serotype responsible for invasive neonatal infections. Additionally, there is growing evidence that bacterial exopolysaccharides have strong binding properties for metal sorption [38,39]. Additionally, production of specific exopolysaccharides can promote cellular survival of metal stress [40][41][42]. Thus, it remains possible that the variations in capsular polysaccharide production responsible for alterations in capsular serotype of the surveyed GBS strains could contribute to alternate binding of excess metals, thereby altering GBS susceptibility to intoxication with metals, such as zinc.
Our study also revealed that different sequence types of GBS had varying susceptibilities to zinc intoxication. Specifically, ST-1 and ST-12 were highly susceptible to zinc stress, while ST-17, ST-19, and ST-23 were much more resistant to zinc intoxication. A cross-continental study revealed that GBS ST-1 and ST-19 are associated with asymptomatic colonization, while ST-17 is predominantly associated with invasive neonatal disease [24]. ST-23 was associated with both rectovaginal carriage and invasive GBS disease [24]. By contrast, STs 1, 17, 19 and 23, were all found to colonize pregnant women at higher rates in different patient populations [18]. Additionally, ST-17 (specifically capsular serotype III strains in this clade) was linked to EOD and strongly associated with LOD and meningitis [19,43,44]. This finding further supports a model in which invasive strains are likely undergoing positive selection for zinc resistance as a critical virulence factor to overcome innate immune defenses which employ zinc intoxication as an antimicrobial strategy [45]. Interestingly, in our study, ST-19 strains (largely colonizing strains) that were most resistant to zinc intoxication were all cpsIII strains, underscoring the relationship between capsular polysaccharide production and zinc resistance in GBS.

Limitations of the study
There are several limitations of our study including the clinical definitions of "colonizing" versus "invasive" strains which can be imperfect. Isolating strains from invasive neonatal infections proves that such strains are capable of causing perinatal infection. However, simply because a "colonizing" strain was isolated from a rectal or vaginal swab does not mean it would be incapable of causing invasive disease under different circumstances. Thus, some "colonizing" strains might be misclassified because they have unrecognized invasive potential. Such a misclassification could bias towards a null hypothesis (contributing to a type II error). Additional genomics studies are also warranted to identify genetic traits linked to zinc resistance, particularly in the more virulent lineages. An additional limitation includes the medium sample size which should be expanded in future studies to include more representation in other capsular serotypes and sequence types to draw broader conclusions across a larger number of GBS strains.

Conclusions
In conclusion, we report strain variations within a cohort of GBS strains with respect to susceptibility to zinc intoxication across STs, capsular serotypes, isolation source, and invasive versus colonizing strains. Invasive isolates demonstrated greater resistance to zinc toxicity compared to colonizing strains. Additionally, ST-1 and ST-12 were highly susceptible to zinc stress, while ST-17, ST-19, and ST-23 were much more resistant to zinc intoxication. cpsIII isolates were less susceptible to zinc intoxication, whereas cpsIb isolates were much more susceptible to zinc toxicity. Our study is a pilot study that is hamstrung by the relatively small number of strains. Future studies will require an expansion to include genetic studies and a larger number of strains and diverse capsular and sequence types, as well as GBS strains from non-perinatal sources and distinct geographic locations.