Glycerol Uptake Is Important for L-Form Formation and Persistence in Staphylococcus aureus

S. aureus is a significant human pathogen and has previously been shown to form cell wall deficient forms or L-forms in vitro and in vivo during infection. Despite many previous studies on S. aureus L-forms, the mechanisms of L-form formation in this organism remain unknown. Here we established the L-form model in S. aureus and constructed a transposon mutant library to identify genes involved in L-form formation. Screening of the library for mutants defective in L-form formation identified glpF involved in glycerol uptake being important for L-form formation in S. aureus. Consistent with this observation, glpF was found to be highly expressed in L-form S. aureus but hardly expressed in normal walled form. In addition, glpF mutant was found to be defective in antibiotic persistence. The defect in L-form formation and antibiotic persistence of the glpF mutant could be complemented by the wild type glpF gene. These findings provide new insight into the mechanisms of L-form formation and persistence in S. aureus and may have implications for development of new drugs targeting persisters for improved treatment.


Introduction
L-form bacteria refer to cell wall deficient form of bacteria that were first discovered by Emmy Klienenberger in 1935 [1]. There is a large body of literature on the bacterial L-forms due to their fascinating biology and their potential importance in latent and persistent infections [2,3,4,5]. L-form bacteria do not grow in regular culture medium but require special culture conditions including rich medium, serum, cell wall inhibitors such as penicillin, osmotic protectant such as sucrose or sodium chloride and soft agar. In contrast to normal bacteria with cell wall that divide by binary fission mediated by FtsZ protein, Lform bacteria divide in a FtsZ-independent manner upon cell envelope stress under specialized conditions [6]. Thus L-form bacteria serve as a useful model to study cell division. There is recent interest in the molecular basis of L-form bacteria formation and survival [5,6,7,8,9]. Despite numerous studies, little is known about the mechanisms of L-form formation. Previous studies have mainly identified mutations in genes involved in cell wall synthesis or cell division in stable L-form bacteria being important for L-form formation [6,9,10,11]. Using E. coli unstable L-form bacteria as a model, we recently systematically examined the molecular basis of L-form formation by microarray analysis and mutant screens and identified a network of genes and pathways involved in L-form formation or survival [7]. These include DNA repair and protection (SOS response), energy production, efflux/transporters, iron homeostasis, cell envelope stress, protein degradation such as trans-translation [7]. These pathways share significant similarity to those involved in bacterial persister and biofilm formation [7]. Despite the above progress, the molecular basis of L-form bacteria formation in other bacteria remains largely unknown.
Staphylococcus aureus is the leading cause of wound and nosocomial infections [12]. Methicillin-resistant S. aureus (MRSA) poses a significant threat in different parts of the world. S. aureus is known to form L-form bacteria in vitro [13,14] or in vivo during infection [15,16,17] or after antibiotic treatment [18]. Clinical samples from patients suffering from MRSA infection contained L-form bacteria exhibiting typical ''fried-egg'' morphology [16]. There is the interesting observation that going through L-form stage with initial phenotypic resistance or persistence to beta-lactam cell wall antibiotic led to subsequent stable genetic resistance after reversion to walled normal form in S. aureus [13]. In addition, S. aureus has been demonstrated to form persisters in different studies [19,20]. However, the molecular mechanisms of L-form bacteria formation and persistence in S. aureus are unclear. In this study, we constructed a transposon mutant library of S. aureus and performed a preliminary screen to identify mutants defective in L-form formation. We identified glpF involved in glycerol uptake being critical for L-form formation in S. aureus. In addition, we found that glpF is also important for persistence to antibiotics in S. aureus.

Antibiotics
Penicillin, ampicillin, chloramphenicol, erythromycin, tetracycline, and norfloxacin were obtained from Sigma-Aldrich Co., and their stock solutions were freshly prepared, filter-sterilized and used at appropriate concentrations as indicated.

Bacterial strains and culture conditions
Bacterial strains and plasmids used in this study are listed in Table 1. All of the S. aureus strains except strain RN4220 were derivatives of the Newman strain. S. aureus strains were cultivated in tryptic soy broth (TSB) (BBL, 211768) and tryptic soy agar (TSA) (Difco, 236950) at 37uC. Chloramphenicol, tetracycline and erythromycin were used at concentrations 5, 2.5 and 10 mg/ml respectively for generating random transposon mutant library. Tetracycline was used at 5 mg/ml for complementation of the glpF mutant.

Construction of S. aureus transposon mutant library
S. aureus transposon mutant library was constructed as described [12]. Briefly, S. aureus RN4220 and Newman strain electrocompetent cells were prepared from log phase cultures grown in TSB medium. pBursa and pFA545 plasmid DNA were transformed into S. aureus competent cells by electroporation (voltage = 2.5 kV, resistance = 100 V, capacity = 25 mF) using MicroPulser Electroporation Apparatus (Bio-Rad). pFA545 plasmid DNA was transformed into S. aureus strain RN4220 and then was isolated from RN4220 for introduction into S. aureus Newman strain. pBursa was transformed into S. aureus Newman strain carrying pFA545. The cells were spread on TSA plates containing 2.5 mg/ml tetracycline and 5 mg/ml chloramphenicol followed by incubation at 30uC and then transferred onto a 43uC prewarmed TSA containing 10 mg/ml erythromycin and incubated at 43uC. About 6000 mutant clones were picked and cultured in 96 wells and then stored as transposon mutant library at 280uC until use.
Induction of S. aureus L-form colonies S. aureus was grown in brain heart infusion (BHI) broth (Becton Dickinson, BD) overnight to stationary phase. Undiluted cultures were spotted onto L-form induction media (LIM) which consisted of BHI supplemented with 1% agar (BD), 10% horse serum (Sigma), 3.5% sodium chloride, 20% sucrose, 0.125% magnesium sulfate, and 600 mg (1000 units)/ml of Penicillin G (Sigma). After the inoculum was absorbed into the agar, the plates were inverted and incubated at 33uC for 7,10 days. The bacterial colonies were detected by inverted microscope (Nikon GM3) and the typical Lform colonies appeared as ''fried egg''.

Microscopy
S. aureus L-form bacteria were examined using a Nikon GM3 inverted microscope for ''fried egg'' colonies grown on LIM agar along with normal growth on a control medium without penicillin. The typical colonies were fixed by glutaraldehyde before being processed and examined by electron microscopy (EM). Scanning EM and transmission EM were performed with scanning electron microscope (JSM-6380Lv) and transmission electron microscope (JEM-1230), respectively, using procedures as described [7,21].
Library screen to identify mutants with defect in L-form colony formation The library screening procedure was similar to that as we previously described [7]. Briefly, the mutant library consisting of 6076 transposon mutants of S. aureus Newman was grown in 200 ml BHI medium at 37uC overnight in 96-well plates without shaking. Stationary phase culture of the mutant library was transferred onto L-form medium LIM plates (150 mm) by a 96well replicator. Plates were allowed to dry before being inverted and incubated at 33uC for up to 7 days before mutants were scored for defect in forming L-form colonies on LIM plates.

Inverse PCR and DNA sequencing of PCR products from L-form mutants
Overnight cultures of L-form deficient mutants and the S. aureus parent strain Newman were centrifuged and bacterial cells were collected for DNA extraction. The genome DNA was isolated by using lysostaphin (Sigma), glass beads (0.1 mm), RNase solution (4 mg/ml), followed by phenol/chloroform extraction and ethanol DNA precipitation. The purified chromosomal DNA was digested by restriction enzyme AciI (New England Biolabs) and DNA restriction fragments were then circularized using T4 DNA ligase (New England Biolabs). The ligated DNA (5 ml) was used as template for inverse PCR reaction in a 25 ml reaction volume with primers ermF and ermR ( Table 2). The PCR cycling parameters were 10 min at 96uC, followed by 40 cycles of 30 s at 94uC, 30 s at 63uC, and 3 min at 72uC. The PCR products were subjected to DNA sequencing with primer ermF. The identity of the DNA sequences was searched in the NCBI database using the BLAST algorithm to identify the gene of interest.

Real-time reverse transcription PCR
Real-time RT-PCR was used to assess the level of expression of glpF transcription in S. aureus L-form versus classical form bacteria. Cultures of S. aureus strain Newman grown overnight in TSB were inoculated onto LIM and BHI media with sucrose control respectively. These plates were incubated at 33uC for 7 days. The colonies grown on LIM and BHI with sucrose were collected for RNA isolation. The culture samples were washed once with DEPC-H 2 O and centrifuged at 8,000 g at 4uC for 5 min. RNA was isolated according to manufacturer's instruction (Sangon Biotech Co., Ltd., Shanghai). Primers corresponding to the genes of interest were designed using Primer Express software (Version 2.0, Applied Biosystems) ( Table 2). Total RNA was converted to cDNA using Super-Script III First-Strand Synthesis (Takara Bio) as described by the manufacturer. The cDNA was used as template to perform real-time RT-PCR per instruction of the reagent kit SYBR Premix Ex Taq II (Takara Bio). The expression of 16S rRNA was used as the control for estimating the fold changes of genes of interest. Cycling parameters were 95uC for 30 s and followed by 40 cycles of 5 s at 95uC, 30 s at 60uC. Relative expression levels were determined by the comparative threshold cycle (ggCt) method.

Complementation of S. aureus mutants
The wild type glpF gene from S. aureus Newman was amplified by PCR using primers glpfF and glpfR (see Table 2). The PCR primers were taken from 2134 bp and 99 bp upstream and downstream of the glpF gene and contained restriction sites BamHI and PstI respectively. The PCR parameters were: 94uC 15 min, followed by 35 cycles of 94uC 30 s, 55uC 30 s, and 72uC 2 min, and a final extension at 72uC for 10 min. The PCR products were digested with BamHI and PstI and then cloned into plasmid pT181 cut with the same enzymes. The ligation products were electroporated into S. aureus strain RN4220 and the positive clones were identified by restriction digestion and PCR. The  For complementation of the glpF mutant in L-form formation, S. aureus Newman strain was grown in BHI broth, and both glpF mutant-pT181 vector control and the glpF pT181-glpF complemented strain were grown in BHI broth with 5 mg/ml tetracycline overnight to stationary phase. The strains were spotted onto LIM and incubated aerobically at 33uC for 5 days followed by detection of typical L-form colonies which appeared as ''fried egg'' by inverted microscope as described above.

Effect of glycerol on L-form formation
The stationary phase cultures of wild type S. aureus Newman strain grown in BHI broth and glpF mutant and the complemented strain in BHI broth containing 5 mg/ml tetracycline were spotted onto LIM without glycerol and with varying concentrations (0, 0.1%, 1%, and 10%) of glycerol, respectively. After incubation aerobically at 33uC for 5 days, the ''fried egg'' colonies were detected to confirm the effects of glycerol on glpF mutant Lform formation.

Persister assays
For antibiotic exposure, S. aureus glpF mutant, the glpF mutant complemented strain, and S. aureus parent strain Newman were grown to stationary phase overnight when antibiotics ampicillin (50 mg/ml) and norfloxacin (40 mg/ml) were added to undiluted cultures and incubated without shaking for various times up to 7 or 8 days. Aliquots of bacterial cultures exposed to antibiotics were taken at different time points and washed and then plated onto TSA plates for CFU counting.

Results and Discussion
Generation of antibiotic induced unstable L-forms of S. aureus L-form induction media (LIM) was tested for its ability to induce S. aureus (Newman) to grow as L-form colonies. Stationary phase cells of S. aureus Newman strain produced L-form colonies when plated directly onto LIM. The minimum bacterial inoculum  Role of Glycerol Uptake in L-Form Formation PLOS ONE | www.plosone.org required for L-form colony formation was approximately 10 6 -10 7 bacteria (the maximum dilution for 10 8 -10 9 S. aureus bacteria to form L-form colony was 1:100). This frequency is significantly lower than the L-form formation frequency of E. coli which is 10 4 -10 5 [7]. The typical S. aureus L-form colonies had ''fried egg'' morphology ( Fig. 1A) under inverted microscope in contrast to smooth colony of the normal form of S. aureus (Fig. 1B). The ''fried-egg'' L-form colonies had typical embedded growth into the soft agar and could not be scraped off in contrast to the normal classical forms which did not show embedded growth and could be scraped off easily from agar surface. Transmission electron microscopy (TEM) indicated that the S. aureus L-form bacteria had complete or partial loss of cell wall and contained a large number of intracellular vesicles (Fig. 1C) in contrast to the normal forms with cell wall without obvious vesicles (Fig. 1D). S. aureus Lform (TEM 6 10,000) showed polymorphic sizes and shapes (Fig. 1E) in contrast to normal classical form showing morphology with homogeneous size and round shape with clear and smooth cell boundary (Fig. 1F). Scanning electron microscopy (SEM) indicated that S. aureus L-form colony exhibited rough surface morphology (Fig. 1G) while in the inside of the L-form colony the bacterial cells exhibited polymorphic morphologies of varying sizes ( Fig. 1H) with the L-form colony structure showing similarity to biofilm structure. This is consistent with previous observation that L-form bacteria secrete exopolysaccharide (EPS) to the surface to prevent desiccation similar to biofilms and that defect in genes involved in EPS synthesis can cause lack of L-form growth [7].  Having established the S. aureus L-form conditions, we wanted to identify genes that are involved in L-form formation. To do this, we first constructed a S. aureus transposon mutant library and then grew the library at 37uC overnight in 96-well plates followed by transfer of the mutants onto LIM plates as described in the Methods. Plates were incubated at 33uC for 7 days when mutants were scored for defect in forming L-form colonies.
To identify the genes whose mutation led to defect in L-form formation, we performed inverse PCR as described in the Methods. Using inverse PCR and DNA sequencing we were able to identify 12 genes from 15 mutants, 3 of which mapped to glpF and NWMN_0623 each, 1 mapped to glpK, 1 mapped to gluconate kinase (gntK), NWMN_0623, NWMN_0872 (GTP pyrophosphokinase), NWMN_1269 (sodium:alanine symporter family protein), 2 mapped to hypothetical proteins NWMN_0333 and NWMN_0843 of unknown function, 3 mapped to intergenic region. Because we found 4 mutants (3 in glpF and 1 in glpK) mapped to glycerol metabolism genes which predominate among these identified genes, we therefore focused and further characterized the role of glycerol metabolism genes in this study. To determine if the mutated glpF is indeed responsible for defective L-form formation, we attempted to complement the glpF mutant with the wild type glpF gene using the plasmid vector pT181. However, the initial attempt was unsuccessful when the complemented glpF mutant was plated directly on LIM plates. Since our previous work with E. coli L-form complementation indicated that inducible expression of the gene involved in L-form formation is critical for successful complementation of L-form defect of the mutants [7], we therefore induced the complemented glpF S. aureus strain containing tetracycline inducible vector pT181 with tetracycline in liquid culture prior to plating on LIM plates. This led to successful complementation of the glpF mutant with the wild type glpF gene. However, the effect of the complementation was partial (Fig. 2C) compared with the parent strain Newman ( Fig. 2A), while the glpF mutant transformed with the pT181 vector control did not form any L-form colonies (Fig. 2B).
glpF and glpK were overexpressed in S. aureus L-form bacteria but not in normal bacteria Since we identified mutations in glpF and glpK caused defect in L-form growth, we wanted to know if glpF and glpK are overexpressed in S. aureus L-form bacteria compared with normal classical form S. aureus. To confirm this, we prepared S. aureus Lform bacteria from L-form media LIM and normal S. aureus growth as a control on media without penicillin and isolated RNA from both types of the bacterial cells. The isolated RNA samples were then subjected to RT-PCR. glpF and glpK were found to be expressed at very low levels in normal growth but were significantly induced to 144-fold and 68-fold higher respectively in L-forms than in the normal control growth (P,0.05) (Fig. 3).

Effect of glycerol on restoring L-form formation in glpF mutant
Since GlpF is involved in glycerol transport, we wanted to address the role of exogenously added glycerol in L-form formation in S. aureus. To do so, we incorporated varying concentrations of glycerol 0, 0.1%, 1% and 10% glycerol into the LIM. In LIM media with 0 and 0.1% glycerol, only wild type S. aureus Newman strain and the glpF complemented strain grew but glpF mutant failed to grow (Table 3). However, at 1% glycerol, the glpF mutant formed 1.03610 4 L-form colonies but its efficiency was much lower than the wild type (7.75610 5 L-form colonies) and the glpF complemented strain (2.90610 5 L-form colonies) (Fig. 4, Table 3). The 1% glycerol only marginally increased the number of L-form colonies of the wild type strain by about 2 fold (Table 3). In contrast, at 10% glycerol, none of the strains Newman, the glpF mutant, or the glpF complemented strain formed L-form colonies. To address if the role of 1% glycerol is to serve as an osmoprotectant in facilitating L-form formation in the glpF mutant, we added known osmoprotectant NaCl at 4.5% to the LIM and assessed if it could allow the glpF mutant to grow as Lforms. However, supplementation of NaCl while allowing L-form formation in wild type and complemented glpF mutant, failed to facilitate L-form formation in the glpF mutant (data not shown).
Defective persistence of S. aureus glpF mutant in antibiotic exposure assays In our previous study with E. coli L-form bacteria, we found that genes involved in L-form formation overlapped with those involved in antibiotic persistence [7]. Exposure of the stationary phase culture of glpF mutant to ampicillin (50 mg/ml) revealed that the mutant began to show defect in persistence at day 3 but the defect was more obvious after 5 days. After 7 days, no CFU was detectable in the glpF mutant transformed with the vector control pT181; in contrast, the complemented strain (transformed with pT181+glpF) and the parent strain Newman had 100-1000 and 10,000 CFUs remaining (Fig. 5) (Table 4). For exposure to norfloxacin (40 mg/ml), the glpF mutant had a similar trend as ampicillin treatment, where it reached zero CFU by day 8.
Complementation of the glpF mutant with the wild type glpF gene only partially restored the persister levels compared with parent strain for ampicillin exposure (Fig. 5) but caused full restoration of persister levels for norfloxacin exposure (Fig. 6) (Table 4).
Although many previous studies have demonstrated the formation of L-forms by S. aureus [13,14,17], the mechanisms involved in L-form formation has remained unknown in this organism. This study provided the first molecular insight into the mechanism of L-form formation in S. aureus by demonstrating the important role of glycerol uptake in the L-form formation. GlpF is involved in glycerol uptake (Fig. 7), and the observation that glpF mutation causes defect in L-form growth in S. aureus suggests uptake of glycerol is essential for L-form formation. The mechanism by which GlpF is involved in L-form formation in S. aureus is likely mediated through its role in production of energy via pyruvate entry into TCA cycle and glycolysis or alternatively through cell membrane synthesis via lysophosphatidic acid (LPA) to phospholipids (Fig. 7) to strengthen membrane integrity required for L-form growth in the absence of cell wall. This study represents the first effort at identifying the mechanisms of L-form formation in S. aureus, and further studies are needed to explore other possible mechanisms of L-form formation besides glycerol uptake in future studies. It is interesting to note that supplementation of appropriate concentration of glycerol (1%) partially compensated for the loss of L-form formation to glpF mutant (Fig. 4). GlpF is known to be a member of the aquaporin (AQP) channel family. Some of AQPs channels that conduct water and also conduct glycerol (aquaglyceroporins) [22]. When the environmental glycerol is increased, the aquaporin (AQP) channel may partially compensate for the loss of GlpF and help the glpF mutant to form a small number of L-form colonies (Fig. 4). We believe the role of glycerol in L-form formation is mediated through its uptake and metabolism rather than an osmoprotectant role in L-form formation for the following reasons. First, we demonstrated in the new experiment that adding 4.5% NaCl as an osmoprotectant did not help the glpF mutant to form L-forms while it allowed wild type and the complemented glpF mutant strains to form L-form colonies. This indicates the defect in glpF mutant cannot be complemented by other osmoprotectant like NaCl. Second, if the role of glycerol were to serve as osmoprotectant, we would expect that higher concentration of glycerol would facilitate L-form formation. However, we found higher glycerol content, e.g., 10% glycerol did not allow Lforms to form. Third, our finding that mutation in glpK which is glycerol kinase involved in glycerol metabolism caused defect in Lform formation also does not support glycerol serving as an osmoprotectant in facilitating L-form formation. This is because GlpK is glycerol kinase that converts glycerol to glycerol-3phosphate which can be used for synthesis of cell membrane phospholipid and also for energy production to facilitate L-form formation. However, high glycerol concentration (10%) inhibited the L-form growth for wild type as well as the mutant S. aureus, presumably because high concentrations of glycerol produce toxic metabolites thus preventing L-form growth.
It is noteworthy that in addition to its role in L-form formation, glpF is also involved in tolerance or persistence to antibiotics in S. aureus as demonstrated by a defect in persistence in glpF mutant upon exposure to ampicillin or norfloxacin ( Fig. 5 and Fig. 6). This finding is consistent with the previous observation that genes involved in glycerol metabolism such as glpD encoding sn-glycerol-3-phosphate dehydrogenase and plsB encoding sn-glycerol-3-phosphate acyltransferase (Fig. 7), have been found to be involved in persister formation [23]. Our findings that glycerol uptake is important for both L-form and persistence of S. aureus and provide further support for the close relationship between the two entities as has been observed for E. coli [7]. This finding is consistent with the recent proposal that L-form bacteria are related to persisters and are part of the heterogeneous persister continuum [24]. The only difference is that frequency of L-form forming bacteria is 2-3 orders of magnitude lower than persister frequency [24] and can be considered ''deep'' persisters [25]. Although previous study with E. coli L-form bacteria did not identify glpF being critical for L-form formation as in S. aureus, microarray analysis indicated that glycerol metabolism gene glpD encoding sn-glycerol-3-phosphate dehydrogenase is upregulated in E. coli L-form bacteria [7], suggesting glycerol metabolism could be important for L-form bacteria in different bacterial species.
In conclusion, we established an L-form model in S. aureus, characterized their morphologies and optima conditions of their formation, and identified genes involved in glycerol uptake and metabolism being important for L-form formation and persistence in S. aureus. These findings shed new light on the mechanisms of L-form formation and persister biology in S. aureus and may have implications for development of new drugs targeting persisters for improved treatment of persistent bacterial infections.