Strain-Specific Adaptations of Streptococcus mitis-oralis to Serial In Vitro Passage in Daptomycin (DAP): Genotypic and Phenotypic Characteristics

Viridans group streptococci (VGS), especially the Streptococcus mitis-oralis subgroup, are pivotal pathogens in a variety of invasive endovascular infections, including “toxic shock” in neutropenic cancer patients and infective endocarditis (IE). Previously, we showed that the serial in vitro passage of S. mitis-oralis strains in sublethal daptomycin (DAP) resulted in rapid, high-level and stable DAP-resistance (DAP-R), which is accompanied by distinct changes in several genotypic and phenotypic signatures: (1) the disappearance of two key membrane phospholipids, phosphatidylglycerol (PG) and cardiolipin (CL); (2) increased membrane fluidity; (3) increased positive surface charge; (4) single nucleotide polymorphisms (SNPs) in two loci involved in CL biosynthesis (pgsA; cdsA); and (5) DAP hyperaccumulation. The current study examined these same metrics following in vitro serial DAP passages of a separate well-characterized S. mitis-oralis bloodstream isolate (SF100). Although some metrics seen in prior DAP post-passage strains were recapitulated with SF100 (e.g., pgsA SNPs, enhanced membrane fluidity), we observed the following major differences (comparing the parental versus post-passage variant): (1) no change in PG content; (2) reduced, but not absent, CL, with enhancement in phosphatidic acid (PA) content; (3) an unusual pattern of CL localization; (4) significantly decreased positive surface charge; (5) no difference in DAP accumulation; and (6) no cdsA SNPs. Thus, S. mitis-oralis strains are not “pre-programmed” phenotypically and/or genotypically to adapt in an identical manner during the evolution of the DAP-R.


S. mitis-oralis Strains
The DAP-S parental SF100 S. mitis-oralis strain used in this study was identified by MALDI-TOF, and has a DAP MIC of 1 µg/mL [20]). It is a clinical bloodstream isolate from a patient with IE without prior DAP exposures [19]; moreover, it is highly capable of inducing experimental IE [20]. In order to generate DAP-R derivatives, SF100 was subjected to a 20 d serial in vitro passage using concentrations of DAP (20 µg/mL) as described previously [20]. The DAP-R derivative variant was re-passaged in antibiotic-free brain heart infusion (BHI) media for five days to confirm the stability of the DAP-R phenotype [20]. MICs for the serial passage isolates were determined by microbroth dilution, as well as E test assays [15].

Phospholipid (PL) Profiling
Briefly, late stationary phase cells were used to extract the PLs from SF100 and its DAP-R variants, as described previously [15,17,18]. The major PLs (i.e., CL, PG, and PA, as well as a prominent glycolipid [GL] seen in all VGS) were separated by two-dimensional thin-layer chromatography (2D-TLC) (Silica 60 F254 HPTLC plates [Merck]). Identification of all the PL spots on the TLC plates was confirmed by comparison to known standards following exposure to iodine vapors and spraying with CuSO 4 (100 mg/mL) containing 8% phosphoric acid (v/v) and heated at 180 • C [15,17,18]. The quantification of individual PLs removed directly from TLC plates was performed by digesting with 70% perchloric acid (0.3 mL) at 180 • C for 3 h. The individual PLs were quantified by measuring the optical density at 660 nm using spectrophotometry as described before [15,17,18]. The PL results are expressed as their individual mean proportions (±SD), employing at least three independent experiments performed on different days.

CL Localization
The fluorescent dye 10-N-nonyl arcidine orange (NAO) was used to assess the localization of CM anionic PL-rich microdomains by previously described techniques [15]. Briefly, NAO preferentially binds to anionic PL-enriched microdomains (predominantly CL) in the CM. For fluorescence microscopy, S. mitis-oralis cells were grown in TSB (Tryptic Soy Broth) to an exponential phase (A600 of approximately 0.3). Then, NAO was added to the growth medium to a final concentration of 1 µM, which did not inhibit the growth of S. mitis-oralis strains. Samples were stained for 3.5 h at 37 • C with gentle shaking in the dark (because of light sensitivity). Study strains were washed three times with 0.9% saline post-incubation and immediately immobilized on a coverslip treated with poly-l-lysine (Sigma-Aldrich). Fluorescence imaging and phase-contrast imaging were carried out on an Olympus IX71 microscope with a PlanApo N 100-objective. Green fluorescence from NAO (representing anionic PLs) was detected by using a standard fluorescein isothiocyanate (FITC) filter (excitation at 490 nm and emission at 528 nm). Image acquisition was performed using the SlideBook 5.0 software package. Three independent experiments were performed for each strain on different days.

DAP Binding Assays
Bodipy-DAP (BDP-DAP) was used to measure the interaction of DAP with the parental and DAP-R variants, as outlined before [15,17]. In brief, all study strains were grown in BHI broth at 37 • C, pelleted, and incubated with BDP-DAP (concentrations 4 and 64 µg/mL) in LB (Luria-Bertani) broth supplemented with Ca 2+ at 50 mg/liter) for 15 min in the dark. After staining, excess BDP-DAP was removed by washing-centrifugation, the pellet was washed twice in LB broth, and then, it was immobilized on a coverslip pretreated with poly-l-lysine (Sigma-Aldrich St. Louis, MO, USA). Bacterial cells were seen with an Olympus IX71 fluorescence microscope equipped with PlanApo N 100× objective. Fluorescence measurements were carried out by using a standard FITC filter. The fluorescence intensity was quantitated and normalized to the protein concentration of the samples to assess the relative amount of BDP-DAP bound to the S. mitis-oralis isolates, as detailed elsewhere [15,17]. Protein concentrations above were measured by using standard bicinchoninic acid (BCA) protein assay kit (Thermo Scientific, West Hills, CA, USA) as instructed by the manufacturer described before [21,22]

Surface Charge
The relative positive surface charge of SF100 and its DAP-R variants was measured by using the cytochrome C (cyt C) binding assay following a spectrophotometric method [15,17,18]. Briefly, strains were grown in BHI broth, washed with 20 mM MOPS (3-morpholinopropane-1-sulfonic acid) buffer (pH 7.0), and resuspended in the same buffer at OD 578 = 1.0 Cells were incubated with 0.5 mg/mL cyt C for 10 min, and the amount of cyt C (a highly positively charged molecule) remaining in the supernatant after 15 min exposure was measured spectrophotometrically at OD 530 nm. The greater the amount of cyt C detected in the supernatant (i.e., more unbound cyt C) equates to a relatively increased positive surface charge [15,17,18]. The quantified data represented the means (±SD) of the amount of cyt C detected in the supernatant from a minimum of three independent experiments.

CM Fluidity
Fluidity was determined by using the fluorescent probe 1,6-diphenyl-1,3,5-hexatriene (DPH; excitation and emission wavelengths are 360 nm and 426 nm, respectively). The protocol of DPH incorporation into the CM has been described in detail previously [15,17,18]. Spectrofluorimetry was used to quantify fluorescence polarization and define the polarization index (PI) value as done before (BioTek SFM 25) [15]. An inverse correlation exists between PI value and fluidity (i.e., a lower PI value equates to a higher extent of CM fluidity) [15,17,18]. A minimum of three independent experiments were conducted for each isolate on different days.

Whole Genome Sequencing
Genomic DNA from SF100 WT and a DAP-resistant variant isolated on day 20 of passage (strain D20) from the VGS strain-set was isolated by cetyltrimethylammonium bromide (CTAB) extraction as described previously [15,20]. Genomic DNA was further purified for sequencing using a PureLink Genome DNA Minikit (Invitrogen Waltham, MA, USA) according to the manufacturer's instructions [15,20]. PacBio library construction, sequencing, and annotation were conducted by using standard methods [15,20]. The generation of genome assemblies and subsequent identification of single nucleotide polymorphisms (SNPs) and insertion-deletion mutations (indels) were performed as detailed elsewhere [15,20]. All mutations identified by whole genome sequencing were confirmed by PCR and Sanger sequencing.

Statistics
The two-tailed Students t-test was utilized for statistical analysis of all quantitative data. p values of ≤ 0.05 were considered "significant".

Generation of DAP-R S. mitis-oralis Strain SF100
Following a 20 d serial passage in DAP, the DAP MIC of SF100 progressively rose to >256 µg/mL (Table 1) [20]. This phenotype was stable on the passage of this DAP-R variant in antibiotic-free media.

PL Content of DAP-R S. mitis-oralis Strains
As noted for other S. mitis-oralis strains [15,17,18], 2D-TLC of parental DAP-S strain SF100 revealed four major lipid species: three PLs (PG, CL, and PA) and a large glycolipid (GL) spot. The day 20 (D20) DAP-R passage variant exhibited a significant reduction of CL content as compared to the SF100 parental strain (Table 2). Furthermore, the amount of PG was not altered significantly in the SF100 D20 strain versus the DAP-S parental strain. The quantitative PL assays paralleled the qualitative 2D-TLC plate assays above ( Figure S2). Of interest, the current dataset was somewhat distinct from our prior study in which a prototypic DAP-R S. mitis-oralis strain showed a virtual absence of both PG and CL in the CM [15].

Distribution of Anionic PL Microdomains
The parental SF100 DAP-S strain ( Figure 1A) exhibited anionic PL microdomains mainly located at the septa similar to other DAP-S Gram-positive bacteria [15]. In addition, non-septal discrete fluorescent foci were also observed, which is reminiscent of the ExPortal found in S. pyogenes [23,24]. However, in contrast to prior studies [15], septal anionic PL staining was lost in DAP-R derivatives of SF100 ( Figure 1B-D); these findings suggested a redistribution of anionic PL microdomains (e.g., CL) away from the septum; this redistribution may well play an important role in DAP-R in S. mitis-oralis strains (mirroring similar data in DAP-R E. faecalis) [22].
Antibiotics 2020, 9, x 5 of 10 the SF100 parental strain (Table 2). Furthermore, the amount of PG was not altered significantly in the SF100 D20 strain versus the DAP-S parental strain. The quantitative PL assays paralleled the qualitative 2D-TLC plate assays above ( Figure S2). Of interest, the current dataset was somewhat distinct from our prior study in which a prototypic DAP-R S. mitis-oralis strain showed a virtual absence of both PG and CL in the CM [15].

Distribution of Anionic PL Microdomains.
The parental SF100 DAP-S strain ( Figure 1A) exhibited anionic PL microdomains mainly located at the septa similar to other DAP-S Gram-positive bacteria [15]. In addition, non-septal discrete fluorescent foci were also observed, which is reminiscent of the ExPortal found in S. pyogenes [23,24]. However, in contrast to prior studies [15], septal anionic PL staining was lost in DAP-R derivatives of SF100 ( Figure 1B-D); these findings suggested a redistribution of anionic PL microdomains (e.g., CL) away from the septum; this redistribution may well play an important role in DAP-R in S. mitisoralis strains (mirroring similar data in DAP-R E. faecalis) [22].

Cell Surface Positive Charge
It has been well-chronicled that many DAP-R S. aureus, enterococci, and some prior S. mitis-oralis strains exhibit increases in relative positive surface charge versus their respective DAP-S parental strains [15,17,18,21,23]. This DAP-R adaptation has been proposed to generate a "charge-repulsive milieu" that contributes to a reduced ability of calcium-DAP to bind to target CMs to initiate its bactericidal effects. Surprisingly, and in contradistinction to a previous study [15], as SF100 evolved stable DAP-R upon in vitro passage, its surface became substantially less positively charged (Table  3).

Cell Surface Positive Charge
It has been well-chronicled that many DAP-R S. aureus, enterococci, and some prior S. mitis-oralis strains exhibit increases in relative positive surface charge versus their respective DAP-S parental strains [15,17,18,21,23]. This DAP-R adaptation has been proposed to generate a "charge-repulsive milieu" that contributes to a reduced ability of calcium-DAP to bind to target CMs to initiate its bactericidal effects. Surprisingly, and in contradistinction to a previous study [15], as SF100 evolved stable DAP-R upon in vitro passage, its surface became substantially less positively charged (Table 3).

DAP CM Binding Profiles
As opposed to a prior study [15], in comparing the parental SF100 strain with DAP-R SF100 D20, no quantitative differences in overall DAP binding were detected at either DAP exposure concentration ( Figure 2).

DAP CM Binding Profiles
As opposed to a prior study [15], in comparing the parental SF100 strain with DAP-R SF100 D20, no quantitative differences in overall DAP binding were detected at either DAP exposure concentration ( Figure 2).

CM Fluidity
It is well known that the relative state of the CM "order" has a profound impact on the ability of calcium-DAP and other cationic peptides to interact with the target CM. Interestingly, enhancements in fluidity were, indeed, observed in DAP-R variants; of interest, this phenotype was somewhat delayed during the evolution of DAP-R, becoming evident only in the D20 DAP-R passage variant (Table 3).

Whole-Genome Sequencing (WGS)
WGS of SF100D20 versus the SF100 strain pair revealed a relatively limited number of nonsynonymous SNPs and deletions emerging in the DAP-R SF100 D20 passage variant. Of note, these mutations were broadly grouped into those impacting (1) lipid metabolism; (2) cell wall and cell surface homeostasis; or (3) ion transport and the maintenance of CM potential. As underscored above, each of these latter phenotypes are plausibly critical in the interaction of DAP with target CMs [15,17,18,21,22].
As compared with its parent strain, the DAP-R variant, SF100 D20, contains 13 SNPs and 2 deletions, involving a total of 17 ORFs After eliminating genes of unknown function, as well as those with no apparent functional linkage to DAP-R, we identified four candidate genes containing nonsynonymous SNPs that could likely be linked to DAP-R (Table 4). Similar to a previous study [17], an SNP in pgsA (ORF853) was identified in the SF100 D20 strain; this enzyme catalyzes an important step in CL biosynthesis ( Figure S1). In addition, an SNP was found in ORF1433, which is predicted

CM Fluidity
It is well known that the relative state of the CM "order" has a profound impact on the ability of calcium-DAP and other cationic peptides to interact with the target CM. Interestingly, enhancements in fluidity were, indeed, observed in DAP-R variants; of interest, this phenotype was somewhat delayed during the evolution of DAP-R, becoming evident only in the D20 DAP-R passage variant (Table 3).

Whole-Genome Sequencing (WGS)
WGS of SF100D20 versus the SF100 strain pair revealed a relatively limited number of non-synonymous SNPs and deletions emerging in the DAP-R SF100 D20 passage variant. Of note, these mutations were broadly grouped into those impacting (1) lipid metabolism; (2) cell wall and cell surface homeostasis; or (3) ion transport and the maintenance of CM potential. As underscored above, each of these latter phenotypes are plausibly critical in the interaction of DAP with target CMs [15,17,18,21,22].
As compared with its parent strain, the DAP-R variant, SF100 D20, contains 13 SNPs and 2 deletions, involving a total of 17 ORFs After eliminating genes of unknown function, as well as those with no apparent functional linkage to DAP-R, we identified four candidate genes containing non-synonymous SNPs that could likely be linked to DAP-R (Table 4). Similar to a previous study [17], an SNP in pgsA (ORF853) was identified in the SF100 D20 strain; this enzyme catalyzes an important step in CL biosynthesis ( Figure S1). In addition, an SNP was found in ORF1433, which is predicted to encode an acetyl-coA-acetyl transferase; this is a key enzyme in the mevalonate pathway, which, in turn, is important for fatty acid metabolism and CM order maintenance, especially under stress conditions (e.g., 'cold shock') [25]. ORF1632 (rpoC) encodes the β subunit of RNA polymerase. Mutations in this protein have been linked to DAP-R [25]. Of interest, RpoBC in S. aureus can repress dlt expression, the latter being a key regulator of surface charge maintenance [26]. Thus, an SNP in rpoC could impact the surface positive charge and potentially perturb calcium-DAP surface binding.
SNPs were also found in ORF1590, encoding an alanine dehydrogenase (ald) homologue. Of note, mutations in these latter two genes in S. aureus have each been associated with cationic antimicrobial peptide resistance [27,28].

Discussion
The current study with in vitro-derived DAP-R variants emerging from strain SF100 was compelling for a number of reasons. For example, several of the genetic mutations we observed evolving in the SF100 DAP-R variants have been previously linked to DAP-R in other S. mitis-oralis (e.g., pgsA) [17]. Moreover, other documented mutations we observed were in genes involved in metabolic pathways important for CM biogenesis, integrity, or transmembrane potential maintenance, making them plausible candidates for a role in DAP-R. In addition, several of these genes are important in terms of other metabolic parameters that are likely involved in DAP-R (e.g., fatty acid metabolism; Na + /H + antiporters).
Of note, we did not observe non-synonymous mutations in the cdsA gene in SF100 D20 DAP-R variants as documented in other DAP-R S. miti-oralis [15]. However, it should be emphasized that (1) we did identify SNPs in the pgsA gene that catalyze the production of PG, the PL that eventually dimerizes to form CL; and (2) many additional regulatory and metabolic pathways can impact CL synthesis, other than the primary pathway shown in Figure S1 [15]. In this context, WGS of DAP-R in S. aureus and B. subtilis [25,26] has also identified mutations in a number of genes responsible for PL biosynthesis, including pgsA. Thus, it is likely that the mutations identified above by WGS synergize to yield the DAP-R phenotype in SF100.
To assess the phenotypic correlation of mutations in pgsA that were found on WGS, 2D TLC and NAO microscopy were performed to quantify proportional PL content, as well as the localization and distribution of anionic PLs, respectively. We recently documented a unique phenotype in other DAP-R S. mitis-oralis strains, i.e., an apparent complete disappearance of PG and CL, with the accumulation of the PL precursor, PA [15]. Since negatively charged PG is essential for interaction with, and the oligomerization of DAP within target CMs [29], this phenotype was logically linked to DAP-R. In addition, negatively charged PLs (such as PG and CL) are crucial in DAP interactions within the cell envelope at its major site of activity, the septal division plane. In addition, mutations in cdsA and cls (leading to alterations in CL content) in DAP-R S. mitis-oralis and enterococci have been associated with the diversion and mislocalization of DAP away from the septum [15,22]. Moreover, the combined "microdomains" of PG and CL are known to act as a negatively charged CM "docking site" for DAP and other cationic molecules after initial interaction with and penetration of the CM.
In contrast to these prior investigations [15], we did not observe a "PG-CL disappearance phenotype" among DAP-R variants derived from the in vitro DAP passaging of strain SF100. Of note, DAP-R strains exhibited a similar pattern of PG content versus its DAP-S SF100 parental strain; however, a significant reduction of CL content was noted among the DAP-R variants. It is possible that the mutation in pgsA noted above in the DAP-R strain impacted CL content by reducing the amount of PG available for ultimate dimerization into CL.
In addition, NAO staining assays indicated that anionic PL-enriched microdomains were localized at the septum in the DAP-S parental strain as expected. However, in the DAP-R variant, anionic PLs were diverted away from the division septum, likely contributing to DAP-R, as seen in DAP-R enterococci [22].
Another notable feature of prior S. mitis-oralis and other DAP-R organisms is increased positive surface charge [15,17,18]. In contrast, in the current study, the net surface charge of the DAP-R variant became less positive during the evolution of DAP-R in vitro. These data underscore that a "charge-repulsive cell surface" is only one of a number of factors that can potentially account for the DAP-R phenotype. In addition, it is theoretically possible that a more negatively charged surface can "trap" more calcium-DAP and prevent it from interfacing with the CM PLs required for its ultimate mechanisms of action.
The DAP binding assays for the DAP-R SF100 variant yielded data with very different outcomes as compared to other DAP-R S. mitis-oralis derivatives [15]. Thus, prior studies showed that DAP binding was surprisingly increased overall in DAP-R variants that harbor mutations in cdsA, which is associated with the selective hyperaccumulation of DAP in a minority of cocci within individual VGS chains. This supported the notion that in those DAP-R strains, such DAP-hyperaccumulating cells were, in essence, "protecting" the majority of population from exposure to DAP (an example of "altruism") [15]. In contrast, we found no difference in the quantitative amount of DAP binding when comparing our parent versus DAP-R S. mitis-oralis variant.
Alterations in CM fluidity have been linked to DAP-R and cross-resistance to other cationic antimicrobial peptides among other DAP-R Gram-positive pathogens [13,15,17,[30][31][32]. It is believed that there exists a fluidity optimum for the interaction of specific cationic antimicrobial peptides (such as calcium-DAP) with target bacterial CMs. The SF100 DAP-R strains exhibited significant increases in CM fluidity versus its DAP-S parental strain, similar to previous observations, suggesting that this biophysical metric may be contributing to DAP-R [15].

Conclusions
It is likely that S. mitis-oralis can develop DAP-R via several mechanisms, as has been shown in studies of DAP-R in S. aureus and enterococci [30][31][32]. However, the data presented here, in combination with previous work [15][16][17]33] indicate that altered membrane composition and some other metabolic modifications [34] may be a conserved mechanism for DAP-R. Undoubtedly, specific S. mitis-oralis strains are not "pre-programmed" to develop DAP-R by any single pathway.
Supplementary Materials: The following are available online at http://www.mdpi.com/2079-6382/9/8/520/s1, Figure S1: Major steps in the CL biosynthetic pathways; Figure S2. Major CM PLs of S. mitis-oralis strains. Proportions of PG, CL, and PA in CM of SF100 and its DAP-R derivative SF100 D20 were determined by 2D thin-layer chromatography. GL is a major CM glycolipid found in all VGS.