Abstract
Bacterial membrane lipids are critical for membrane bilayer formation, cell division, protein localization, stress responses, and pathogenesis. Despite their critical roles, membrane lipids have not been fully elucidated for many pathogens. Here, we report the discovery of a novel cationic glycolipid, Lysyl-Glucosyl-Diacylglycerol (Lys-Glc-DAG) that is synthesized in high abundance by the bacterium Streptococcus agalactiae (Group B Streptococcus, GBS). To our knowledge, Lys-Glc-DAG is more positively charged than any other known lipids. Lys-Glc-DAG carries two positive net charges per molecule, distinct from the widely described lysylated phospholipid Lysyl-phosphatidylglycerol (Lys-PG) which carries one positive net charge due to the presence of a negatively charged phosphate moiety. We use normal phase liquid chromatography (NPLC) coupled with electrospray ionization (ESI) high-resolution tandem mass spectrometry (HRMS/MS) and genetic approaches to determine that Lys-Glc-DAG is synthesized by the enzyme MprF in GBS, which covalently modifies the neutral glycolipid Glc-DAG with the cationic amino acid lysine. GBS is a leading cause of neonatal meningitis, which requires traversal of the endothelial blood-brain barrier (BBB). We demonstrate that GBS strains lacking mprF exhibit a significant decrease in the ability to invade BBB endothelial cells. Further, mice challenged with a GBSΔmprF mutant developed bacteremia comparably to Wild-Type infected mice yet had less recovered bacteria from brain tissue and a lower incidence of meningitis. Thus, our data suggest that Lys-Glc-DAG may contribute to bacterial uptake into host cells and disease progression. Importantly, our discovery provides a platform for further study of cationic lipids at the host-pathogen interface.
Introduction
Bacterial cellular membranes are dynamic structures that are critical for survival under varying environmental conditions and are essential for host-pathogen interactions. Phospholipids and glycolipids within the membrane have varying chemical properties that alter the physiology of the membrane, which bacteria can modulate in response to environmental stresses such as pH (1), antibiotic treatment (2), and human metabolites (3). Despite their critical roles in the survival and pathogenesis, membrane lipids have not been carefully characterized using modern lipidomic techniques for many important human pathogens, including Streptococcus agalactiae (Group B Streptococcus; GBS). GBS colonizes the lower genital and gastrointestinal tracts of ∼30% of healthy women (4, 5). However, GBS can cause sepsis and pneumonia in newborns and is a leading cause of neonatal meningitis, resulting in long-lasting neurological effects in survivors (6-8). Due to the severity of the resulting diseases, intrapartum antibiotic prophylaxis is prescribed for colonized pregnant women (7, 9). Even with these measures, a more complete understanding of GBS pathogenesis and new therapeutic and preventive measures are needed to mitigate the devastating impact of GBS neonatal infection.
Research on the pathogenesis of the GBS has mainly focused on cell wall-anchored or secreted proteins and polysaccharides that aid in the attachment to and invasion of host cells. The numerous attachment and virulence factors possessed by the GBS are summarized in a recent review by Armistead et al 2019 (10). Comparatively little is known about GBS cellular membrane lipids. To our knowledge, the only characterization of GBS lipids prior to our current study was the identification of the phospholipids phosphatidylglycerol (PG), cardiolipin (CL), and lysyl-phosphatidylglycerol (Lys-PG) in GBS (11, 12). Similarly, investigation into the glycolipids of the GBS membrane has focused on di-glucosyl-diacylglycerol (Glc2-DAG), which is the lipid anchor of the Type I lipoteichoic acid, and its role in pathogenesis (13).
In this study, we utilized normal phase liquid chromatography (NPLC) coupled with electrospray ionization (ESI) high-resolution tandem mass spectrometry (HRMS/MS) to characterize the GBS membrane lipid composition, and identified a novel cationic glycolipid, lysyl-glucosyl-diacylglycerol (Lys-Glc-DAG), which comprises a major portion of the GBS total lipid extract. While Lys-PG has been reported in a range of bacterial species (14), Lys-Glc-DAG represents, to our knowledge, the first example of lysine modification of a neutral glycolipid. By gene deletion and heterologous expression, we show the GBS MprF enzyme is responsible for the biosynthesis of both the novel Lys-Glc-DAG and Lys-PG. Most strikingly, using an in vivo hematogenous murine infection model, we demonstrate that MprF does not contribute to GBS bloodstream survival. This distinguishes the GBS MprF from the well-known Staphylococcus aureus MprF, which synthesizes only Lys-PG (15, 16). Rather, GBS MprF contributes specifically to meningitis and penetration of the blood-brain barrier. These results greatly expand our knowledge of naturally occurring lipids and MprF functionality and reveal insights into the pathogenesis of meningitis caused by GBS.
Results
Identification of Lys-Glc-DAG, a novel cationic glycolipid in GBS
The membrane lipids of three GBS clinical isolates of representative serotypes were characterized: COH1 (17), A909 (18), CNCTC 10/84 and CJB111 (serotypes III, 1a, and V, respectively) (19). Common Gram-positive bacterial lipids were identified by normal phase LC coupled with negative ion ESI/MS/MS, including diacylglycerol (DAG), monohexosyldiacylglycerol (MHDAG), dihexosyldiacylglycerol (DHDAG), phosphatidylglycerol (PG), and lysyl-phosphatidylglycerol (Lys-PG), as shown by the negative total ion chromatogram (TIC) (Fig. 1A).
Surprisingly, the positive TIC (Fig. 1B, Supplemental Figure S1) shows highly abundant peaks of unknown identity at the retention time ∼25-29 min. The mass spectra (Fig. 1C) and LC retention times of this lipid do not match with any other bacterial lipids we have analyzed or exact masses in lipidomic databases. Tandem MS (MS/MS) in the positive ion mode (Fig. 1D), negative ion mode (Fig. 1E), and high-resolution mass measurement (Fig. 1C) allowed us to propose lysyl-glucosyl-diacylglycerol (Lys-Glc-DAG) (Fig. 1F) as the structure of this unknown lipid. Observed and exact masses of Lys-Glc-DAG are shown in Table S1. The assignment of glucose was based on the observation that glucosyl-diacylglycerol (Glc-DAG) is a major membrane component of GBS and other streptococci (13), and results from an isotopic labeling experiment using 13C-labeled glucose (Glucose-13C6). The assignment of lysine modification was supported by an isotopic labeling experiment with deuterated lysine (lysine-d4). The expected mass shifts (+4 Da) were observed in both molecular ions and MS/MS product ions (Supplemental Figure S2). Comparison of both MS/MS spectra of labeled (Glucose-13C6) and unlabeled Lys-Glc-DAG indicates the lysine residue is linked to the 6-position of glucose (Supplemental Figure S2). Lys-Glc-DAG consists of several molecular species with different fatty acyl compositions resulting in different retention times and multiple, unresolved TIC peaks (∼25-29 min).
GBS MprF synthesizes Lys-Glc-DAG
The enzyme MprF (multiple peptide resistance factor) catalyzes the aminoacylation of PG with lysine in some Gram-positive pathogens (15, 20). We determined that GBS MprF is responsible and sufficient for synthesizing Lys-Glc-DAG as well as Lys-PG. Deletion of mprF from both COH1 and CJB111 abolishes Lys-Glc-DAG and Lys-PG synthesis, which are restored by complementation (Fig. 1G, Supplemental Figure S3). Deletion of GBS mprF does not confer a growth defect in Todd-Hewitt broth or tissue culture medium. The oral colonizer Streptococcus mitis does not encode mprF or synthesize Lys-PG but synthesizes Glc-DAG and PG (2, 3). Heterologous expression of GBS mprF in S. mitis results in Lys-Glc-DAG and Lys-PG production (Fig. 1H), while expression of Enterococcus faecium mprF results in only Lys-PG production (Fig. 1H), as expected (1). Biosynthetic pathways involving MprF are shown in Fig. 1I.
MprF contributes to GBS pathogenesis
We investigated whether MprF contributes to GBS invasion into brain endothelial cells and development of meningitis. To mimic the human blood-brain barrier (BBB), we utilized the human cerebral microvascular endothelial cell line hCMEC/D3. In vitro assays for adhesion and invasion were performed as described previously (13, 21, 22). There was no significant difference in the ability of ΔmprF compared to WT and complement cells to attach to hCMEC/D3 cells (Fig. 2A). However, we observed a significant decrease in the amount of ΔmprF recovered from the intracellular compartment of hCMEC/D3 cells (Fig. 2A). The reduced invasion phenotype was confirmed in the hypervirulent serotype V strain, CJB111 (23, 24) (Supplemental Figure S4). Intracellular survival requires GBS to survive low pH conditions in lysosomes (pH 4.5 – 5.5) (25), and ΔmprF is unable to survive low pH conditions (Fig. 2B). This suggests that MprF promotes GBS invasion and possibly intracellular survival in brain endothelial cells.
We hypothesized that these in vitro phenotypes of ΔmprF would translate into a diminished ability to penetrate the BBB and produce meningitis in vivo. Using our standard model of GBS hematogenous meningitis (13, 21) mice were challenged with either WT GBS or ΔmprF. Mice were sacrificed at 72 h to determine bacterial loads in blood and brain tissue. We recovered significantly less CFU in the brains of ΔmprF-infected mice compared to the WT-infected mice (Fig. 2C). However, there was no significant difference in CFU recovered from the bloodstream (Fig. 2D), demonstrating that ΔmprF does not have a general in vivo growth defect. Furthermore, mice challenged with WT GBS had significantly more leukocyte infiltration, meningeal thickening and neutrophil chemokine, KC, in brain homogenates compared to ΔmprF mutant-infected animals (Fig. 2E-G). Taken together, mprF contributes to GBS penetration into the brain and to the pathogenesis of meningitis in vivo.
Discussion
Lipid nanoparticles are essential for mRNA vaccine delivery. Engineered cationic lipids are utilized in lipid nanoparticles for vaccine and drug delivery and are required for uptake of particles into cells (26, 27). Substantial effort has been dedicated to the synthesis of cationic lipids with low toxicity and efficient delivery properties. Here, we report the discovery of Lys-Glc-DAG, a naturally occurring cationic glycolipid synthesized in high abundance by GBS, which in conjunction with Lys-PG aids in invasion of human endothelial cells. We discovered that GBS MprF uniquely synthesizes the novel and highly abundant Lys-Glc-DAG, as well as Lys-PG. This establishes that GBS capitalizes on MprF to modulate charges of both glycolipids and phospholipids at the membrane.
MprF catalyzes the aminoacylation of the anionic phospholipid PG in a range of Gram-positive and Gram-negative bacteria (15, 20). MprF is a membrane-bound enzyme comprised of a N-terminal lipid flippase domain (28) and a C-terminal catalytic domain that catalyzes the aminoacylation of the glycerol group of PG by using aminoacyl-tRNAs as the amino acid donors (29-31). An important function of PG aminoacylation is proposed to be lowering the net negative charge of the cellular envelope to confer protection from cationic antimicrobial peptides (CAMPs) produced by host immune systems and bacteriocins produced by competitor bacteria (15, 20). However, a previous study observed no contribution of mprF to GBS in vitro susceptibility to commonly studied CAMPs, which is unlike the well-characterized S. aureus mprF (32), thus highlighting the unique differences between the extracellular surface of these bacteria.
Based on our tissue culture and mouse infection experiments, we propose that GBS have an MprF enzyme and corresponding cellular lipid properties that are adapted for efficient invasion of mammalian cells. Deletion of mprF impacts GBS virulence, but only for meningitis, and not for bacteremia. This demonstrates that MprF plays a specific role in BBB penetration, but not in vivo survival in general. It is unknown how lysinylated lipids in the GBS membrane, which is covered by a layer of peptidoglycan, mechanistically impact invasion. Because Lys-Glc-DAG is abundantly synthesized by GBS MprF, with Lys-PG a comparatively minor product, it is likely that Lys-Glc-DAG is the most relevant lipid for meningitis pathogenesis. Speculatively, Lys-Glc-DAG may contribute to membrane vesicle (MV) formation by GBS. MVs have previously been shown to be pro-inflammatory and result in preterm birth and fetal death in mice (33), but have not been studied during meningitis progression. In future studies, it will be key to investigate this, as well as the specific host inflammatory and signaling responses to the GBS mprF mutant.
Our identification of the novel Lys-Glc-DAG glycolipid rationalizes further study of the lipidomes of human pathogens. First, lipids contribute to virulence, and understanding these virulence mechanisms and the mechanisms for lipid synthesis may identify novel antimicrobial drug targets. The decreased in vivo pathogenicity of the ΔmprF mutant identifies GBS MprF as a candidate for targeting by antimicrobial strategies. Moreover, Lys-Glc-DAG could be utilized as a specific molecular biomarker for GBS diagnostics. Second, pathogens use lipids to modulate their surface charges and interact with their hosts, and they innovate lipids to meet their specific needs. Lys-Glc-DAG is a naturally occurring, strongly cationic lipid with potential for use in lipid nanoparticles for vaccine and drug delivery. Importantly, our discovery suggests that lipidome analysis of human pathogens is likely to reveal novel lipids of biotechnological utility.
Materials and methods
Bacterial strains, media, and growth conditions
See Table S2 for strains used in this study. GBS strains were grown statically at 37°C in Todd-Hewitt Broth (THB) and S. mitis strains were grown statically at 37°C and 5% CO2, unless otherwise stated. Streptococcal chemically defined medium (34) was diluted from stock as described (35) with 1% w/v glucose (referred to as DM), slightly modified from (36), unless otherwise stated. Escherichia coli strains were grown in Lysogeny Broth (LB) at 37°C with rotation at 225 rpm. Kanamycin and erythromycin (Sigma-Aldrich) were supplemented to media at 50 μg/mL and 300 μg/mL for E. coli, respectively, or 300 μg/mL and 5 μg/mL, respectively, for streptococcal strains.
Routine molecular biology techniques
All PCR reactions utilized Phusion polymerase (Thermo Fisher). PCR products and restriction digest products were purified using GeneJET PCR purification kit (Thermo Fisher) per manufacturer protocols. See Table S3 for primers. Plasmids were extracted using GeneJET plasmid miniprep kit (Thermo Fisher) per manufacturer protocols. Restriction enzyme digests utilized XbaI, XhoI, and PstI (New England Biolabs) for 3 h at 37°C in a water bath. Ligations utilized T4 DNA ligase (New England Biolabs) at 16°C overnight or Gibson Assembly Master Mix (New England Biolabs) per manufacturer protocols where stated. All plasmid constructs were sequence confirmed by Sanger sequencing (Massachusetts General Hospital DNA Core or CU Anschutz Molecular Biology Core).
Deuterated lysine and 13C6-D-glucose isotope tracking
A GBS COH1 colony was inoculated into 15 mL of DM containing 450 μM lysine-d4 (Cambridge Isotopes Laboratories) or a single COH1 colony was inoculated into 10 mL DM supplemented with 0.5% w/v 13C6D-glucose (U-13C6, Cambridge Isotopes Laboratories) for overnight growth for lipidomic analysis described below.
Construction of MprF expression plasmids
Genomic DNA was isolated using the Qiagen DNeasy Blood and Tissue kit per the manufacturer’s protocol with the exception that cells were pre-treated with 180 μL 50 mg/mL lysozyme, 25 μL 2500 U/mL mutanolysin, and 15 μL 20 mg/mL pre-boiled RNase A and incubated at 37°C for 2 h. The mprF genes from GBS COH1, (GBSCOH1_1931), GBS CJB111 (ID870_10050), and E. faecium 1,231,410 (EFTG_00601) were amplified and either Gibson ligated into pABG5ΔphoZ (37) or ligated into pDCErm (38). Plasmid constructs were transformed into chemical competent E. coli. Briefly, chemically competent cells were incubated for 10 min on ice with 5 μL of Gibson reaction before heat shock at 42°C for 70 sec, then placed on ice for 2 min before 900 μL of cold SOC medium was added. Outgrowth was performed at 37°C, with shaking at 225 rpm, for 1 h. Cultures were plated on LB agar plates containing 50 μg/mL kanamycin. Colonies were screened by PCR for presence of the mprF insert.
Expression of mprF in S. mitis
Natural transformation was performed as previously described (3). Briefly, precultures were thawed at room temperature, diluted in 900 μL of THB, further diluted 1:50 in prewarmed 5 mL THB, and incubated for 45 min at 37°C. 500 μL of culture was then aliquoted with 1 μL of 1 mg/ml competence-stimulating peptide (EIRQTHNIFFNFFKRR) and 1 μg/mL plasmid. Transformation reaction mixtures were cultured for 2 h at 37°C in microcentrifuge tubes before being plated on THB agar supplemented with 300 μg/mL kanamycin. Single transformant colonies were cultured in 15 mL THB overnight. PCR was used to confirm the presence of the mprF insert on the plasmid. Plasmids were extracted and sequence confirmed as described above. Lipidomics was performed as described below in biological triplicate.
Construction of mprF deletion plasmids
Regions ∼2 kb upstream and downstream of the GBS COH1 mprF (GBSCOH1_1931) or CJB111 (ID870_10050) were amplified using PCR. Plasmid, pMBSacB (39), and the PCR products were digested using appropriate restriction enzymes and ligated overnight. 7 μL of the ligation reaction was transformed into chemically competent E. coli DH5α as described above, except that outgrowth was performed at 28°C with shaking at 225 rpm for 90 min prior to plating on LB agar supplemented with 300 μg/mL erythromycin. Plates were incubated at 28°C for 72 h. Colonies were screened by PCR for correct plasmid construction. Positive colonies were inoculated into 50 mL LB media containing antibiotic and incubated at 28°C with rotation at 225 rpm for 72 h. Cultures were pelleted using a Sorvall RC6+ centrifuge at 4,280 × g for 6 min at room temperature. Plasmid was extracted as described above except the cell pellet was split into 5 columns to prevent overloading and serial eluted into 50 μL. Plasmid construction was confirmed via restriction digest using XhoI and XbaI, and the insert was PCR amplified and sequence-verified.
Generation of electrocompetent GBS cells for mprF knockout
Electrocompetent cells were generated as described (39) with minor modifications. Briefly, a GBS COH1 or CJB111 colony was inoculated in 5 mL M17 medium (BD Bacto) with 0.5% glucose and grown overnight at 37°C. The 5 mL was used to inoculate a second overnight culture of 50 mL pre-warmed filter-sterilized M17 medium containing 0.5% glucose, 0.6% glycine, and 25% PEG 8000. The second overnight was added to 130 mL of the same medium and grown for 1 h at 37°C. Cells were pelleted at 3,200 × g in a Sorvall RC6+ at 4°C for 10 min. Cells were washed twice with 25 mL cold filter-sterilized GBS wash buffer containing 25% PEG 8000 and 10% glycerol in water, and pelleted as above. Cell pellets were re-suspended in 1 mL GBS wash buffer and either used immediately for transformation or stored in 100 μL aliquots at −80°C until use.
Deletion of GBS COH1 and CJB111 mprF
Electrocompetent cells were generated as described (39) with minor modifications. The double crossover homologous recombination knockout strategy was performed as described previously (22, 39, 40) with minor modifications. 1 μg of plasmid was added to electrocompetent GBS cells and transferred to a cold 1 mm cuvette (Fisher or BioRad). Electroporation was carried out at 2.5 kV on an Eppendorf eporator. 1 mL of THB containing 12.5% PEG 8000, 20 mM MgCl2, and 2 mM CaCl2 was immediately added and then the entire reaction was transferred to a glass culture tube. Outgrowth was at 28°C for 2 h followed by plating on THB agar supplemented with 5 μg/mL erythromycin. Plates were incubated for 48 h at 28°C. A single colony was cultured overnight in 5 mL THB with 5 μg/mL erythromycin at 28°C. The culture was screened via PCR for the plasmid insert with the initial denaturing step extended to 10 min. The overnight culture was diluted 1:1000 THB containing 5 μg/mL erythromycin and incubated overnight at 37°C to promote single cross over events. The culture was then serial diluted and plated on THB agar plates with antibiotic and incubated at 37°C overnight. Colonies were screened for single crossover events by PCR. Single crossover colonies were inoculated in 5 mL THB at 28°C to promote double crossover events. Overnight cultures were diluted 1:1000 into 5 mL THB containing sterile 0.75 M sucrose and incubated at 37°C. Overnight cultures were serial diluted and plated on THB agar and incubated at 37°C overnight. Colonies were patched onto THB agar with and without 5 μg/mL erythromycin to confirm loss of plasmid. Colonies were also screened by PCR for the loss of mprF. Colonies positive for the loss of mprF were inoculated into 5 mL THB at 37°C. Cultures were stocked and gDNA extracted as described above, with minor modifications. Sequence confirmation of the mprF knockout was done via Sanger sequencing (Massachusetts General Hospital DNA Core or CU Anschutz Molecular Biology Core). The mutant was grown overnight in 15 mL THB at 37°C and pelleted at 6,150 × g for 5 min in a Sorvall RC6+ centrifuge at room temperature for lipid extraction as described. Genomic DNA of COH1ΔmprF was isolated as described above and whole genome sequencing was performed in paired-end reads (2 by 150 bp) on the Illumina NextSeq 550 platform at the Microbial Genome Sequencing Center (Pittsburgh, PA). Illumina sequence reads are deposited in the Sequence Read Archive, accession PRJNA675025.
Complementation of mprF in COH1ΔmprF and CJB111ΔmprF
Electrocompetent GBS strains were generated as previously described (41). Briefly, GBSΔmprF was inoculated into 5 mL THB with 0.6% glycine and grown overnight. The culture was expanded to 50 mL in pre-warmed THB with 0.6% glycine and grown to an OD600 nm of 0.3 and pelleted for 10 min at 3200 × g at 4°C in a Sorvall RC6+ floor centrifuge. The pellet was kept on ice through the remainder of the protocol. The pellet was washed twice with 25 mL and once with 10 mL of cold 0.625 M sucrose and pelleted as above. The cell pellet was resuspended in 400 μL of cold 20% glycerol, aliquoted in 50 μL aliquots, and used immediately or stored at −80°C until use. Electroporation was performed as described above, with recovery in THB supplemented with 0.25 M sucrose, and plated on THB agar with kanamycin at 300 μg/mL.
Acidic Bligh-Dyer extractions
Centrifugation was performed using a Sorvall RC6+ centrifuge. Cultures were pelleted at 4,280 × g for 5 min at room temperature unless otherwise stated. The supernatants were removed, and cell pellets were stored at −80°C until acidic Bligh-Dyer lipid extractions were performed as described (3). Briefly, cell pellets were resuspended in 1X PBS (Sigma-Aldrich) and transferred to Coring Pyrex glass tubes with PTFR-lined caps (VWR), followed by 1:2 vol:vol chloroform:methanol addition. Single phase extractions were vortexed periodically and incubated at room temperature for 15 minutes before 500 × g centrifugation for 10 min. A two-phase Bligh-Dyer was achieved by addition of 100 μL 37% HCl, 1 mL CHCl3, and 900 μl of 1X PBS, which was then vortexed and centrifuged for 5 min at 500 × g. The lower phase was removed to a new tube and dried under nitrogen before being stored at −80°C prior to lipidomic analysis.
Liquid Chromatography/Electrospray Ionization Mass Spectrometry
Normal phase LC was performed on an Agilent 1200 quaternary LC system equipped with an Ascentis silica HPLC column (5 μm; 25 cm by 2.1 mm; Sigma-Aldrich) as described previously (42, 43). Briefly, mobile phase A consisted of chloroform-methanol-aqueous ammonium hydroxide (800:195:5, vol/vol), mobile phase B consisted of chloroform-methanol-water-aqueous ammonium hydroxide (600:340:50:5, vol/vol), and mobile phase C consisted of chloroform-methanol-water-aqueous ammonium hydroxide (450:450:95:5, vol/vol). The elution program consisted of the following: 100% mobile phase A was held isocratically for 2 min, then linearly increased to 100% mobile phase B over 14 min, and held at 100% mobile phase B for 11 min. The LC gradient was then changed to 100% mobile phase C over 3 min, held at 100% mobile phase C for 3 min, and, finally, returned to 100% mobile phase A over 0.5 min and held at 100% mobile phase A for 5 min. The LC eluent (with a total flow rate of 300 μl/min) was introduced into the ESI source of a high-resolution TripleTOF5600 mass spectrometer (Sciex, Framingham, MA). Instrumental settings for negative-ion ESI and MS/MS analysis of lipid species were: IS = −4,500 V, CUR = 20 psi, GSI = 20 psi, DP = −55 V, and FP = −150V. Settings for positive-ion ESI and MS/MS analysis were: IS = +5,000 V, CUR = 20 psi, GSI = 20 psi, DP = +55 V, and FP = +50V. The MS/MS analysis used nitrogen as the collision gas. Data analysis was performed using Analyst TF1.5 software (Sciex, Framingham, MA).
pH-adjusted THB growth
Approximately 30 mL of fresh THB were adjusted to different pH values, measured using a Mettler Toledo FiveEasy pH/MV meter, and sterile-filtered using 0.22 μM syringe filters. A final volume of 200 μL culture medium was aliquoted per well in a flat-bottom 96 well plate (Falcon); culture media were not supplemented with antibiotics. Overnight cultures of GBS strains were used to inoculate the wells to a starting OD600nm 0.02 per well. Plates were incubated for 24 h at 37°C before OD600nm was read using a BioTek MX Synergy 2 plate reader. This experiment was performed in biological triplicate.
hCMEC cell adherence and invasion assays
Human Cerebral Microvascular Endothelial cells hCMEC/D3 (obtained from Millipore) were grown in EndoGRO-MV complete media (Millipore, SCME004) supplemented with 5% fetal bovine serum (FBS) and 1 ng/ml fibroblast growth factor-2 (FGF-2; Millipore). Cells were grown in tissue culture treated 24 well plates and 5% CO2 at 37°C.
Assays to determine the total number of bacteria adhered to host cells or intracellular bacteria were performed as described previously (21, 22). Briefly, bacteria were grown to mid log phase (OD600nm 0.4-0.5) and normalized to 1 × 108 to infect cell monolayers at a multiplicity of infection (MOI) of 1 (1 × 105 CFU per well). The total cell-associated GBS were recovered after 30 min incubation. Cells were washed slowly five times with 500 μL 1X PBS (Sigma) and detached by addition of 100 μL of 0.25% trypsin-EDTA solution (Gibco) and incubation for 5 min before lysing the eukaryotic cells with the addition of 400 μL of 0.025% Triton X-100 (Sigma) and vigorous pipetting. The lysates were then serially diluted and plated on THB agar and incubated overnight to determine CFU. Bacterial invasion assays were performed as described above except infection plates were incubated for 2 h before incubation with 100 μg gentamicin (Sigma) and 5 μg penicillin (Sigma) supplemented media for an additional 2 h to kill all extracellular bacteria, prior to being trypsinized, lysed, and plated as described. Experiments were performed in biological triplicate with four technical replicates per experiment.
Murine model of GBS hematogenous meningitis
All animal experiments were conducted under the approval of the Institutional Animal Care and Use Committee (#00316) at the University of Colorado Anschutz Medical Campus and performed using accepted veterinary standards. The murine meningitis model was performed as previously described (22, 44, 45). Briefly, 7-week-old male CD1 (Charles River) mice were challenged intravenously with 1 × 109 CFU of WT COH1 or the isogenic ΔmprF mutant. At 72 h post-infection, mice were euthanized and blood and brain tissue were harvested, homogenized, and serially diluted on THB agar plates to determine bacterial CFU.
Histology and ELISA
Mouse brain tissue was frozen in OCT compound (Sakura) and sectioned using a CM1950 cryostat (Leica). Sections were stained using hematoxylin and eosin (Sigma) and images were taken using a BZ-X710 microscope (Keyence). Images were analyzed using ImageJ software. Meningeal thickening was quantified from sections taken from three different mice per group, and six images per slide. Meningeal thickening was quantified across two points per image. KC protein from mouse brain homogenates was detected by enzyme-linked immunosorbent assay according to the manufacturer’s instructions (R&D systems).
Conflict of interest
The authors have declared that no conflict of interest exists.
Supplemental Figures, and Tables
Acknowledgements
We thank Kathryn Patras at the University of California San Diego for the CNCTC 10/84 strain and Moutusee Islam in Kelli Palmer’s lab at The University of Texas at Dallas for E. faecium 1,231,410 DNA.
The work was supported in part by T32 5T32AI052066-18 and F31 AI164674 for H.S.M, the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Brazil (CAPES; finance code 001 to J.D.C.M), by grants R01NS116716 and R01AI153332 from the National Institutes of Health (NIH) to K.S.D and associated NIH/NINDS supplement from R01NS116716 to R.V, NIH grant R21AI130666 and the Cecil H. and Ida Green Chair in Systems Biology Science to K.P, NIH grant R56AI139105 to K.P and Z.G, and NIH grant U54GM069338 to Z.G.
References
References
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.