Fig 1.
The redox linchpin connecting three-carbon (glycerol) metabolism and six-carbon (glycolysis) metabolism in B. burgdorferi.
(A) Schematic overview of the intersection of glycerol metabolism and glycolysis, including the conversion of pyruvate to lactate and the use of glycerol-3-phosphate (G3P) for lipid and lipoprotein biosynthesis. Dihydroxyacetone phosphate (DHAP); glycerol uptake facilitator (GlpF, BB0240); glycerol kinase (GlpK, BB0241); glycerol-3-phosphate dehydrogenase (GlpD, BB0243); glycerol-3-phosphate dehydrogenase (GpsA, BB0368); triose phosphate isomerase (TPI); phosphotransferase systems (PTS); lactate permease (LctP). (B) This redox junction consists of two predicted glycerol-3-phosphate dehydrogenases, GlpD and GpsA. GlpD putatively oxidizes G3P to DHAP, while reducing flavin adenine dinucleotide (FAD) or NAD+, to feed glycerol into glycolysis. GpsA reduces DHAP to G3P, using the reducing power of NAD(P)H, to provide carbohydrates for lipoproteins and glycerophospholipids.
Fig 2.
Heterologous complementation of an E. coli gpsA mutant with B. burgdorferi gpsA restores growth in glucose.
The E. coli gpsA null mutant, strain BB20-14 (Ec ΔgpsA, white circles), E. coli gpsA null mutant with the inducible pUC18 expression vector (Ec ΔgpsA pUC18, gray circles) or E. coli gpsA null mutant with pUC18 carrying the B. burgdorferi gpsA gene (Ec ΔgpsA pUC18-gpsABb) were grown in M9 minimal media containing 0.1 mM IPTG either without (A) or with 1% glucose (B) at 37°C. Cell density measurements (OD600) were taken every 17 min. Data are the average from two separate cultures for each strain and error bars represent SEM; the experiment shown is representative of three independent biological replicates. The Ec ΔgpsA pUC18-gpsABb strain had significantly higher (p < 0.05) OD600 values compared to the other two strains from 4 h to 8 h of growth, as determined by one-way ANOVA with a Tukey’s post-hoc test.
Fig 3.
Growth of the gpsA and glpD mutants and survival under nutrient stress without or with different carbon sources in vitro.
(A) B. burgdorferi strains were inoculated at 1 × 105 cells ml-1 in BSK + RS and grown at 35°C. Cells were enumerated every 24 h and cell density plotted over eight days. Data are the means from three independent biological replicates and error bars represent SEM. No significant difference in cell density was determined except at day 1 between the wild-type (WT) and ΔgpsA/ΔglpD strains: * indicates p = 0.046 between the mean cell density of WT and ΔgpsA/ΔglpD strains at day 1 as determined by one-way ANOVA with a Tukey’s post-hoc test. Strains were grown in BSK + RS at 35°C (to a cell density of 5–9 × 107 cells ml-1) before shifting to RPMI alone (B), or RPMI containing 0.4% glycerol (C) or 0.4% N-acetylglucosamine (GlcNAc) (D) and incubated at 35°C for 24 h. Cultures were plated in semi-solid BSK media and allowed to grow at 35°C in 5% CO2 before colony enumeration. Data are presented as the percent survival of each strain before (0 h) shifting to the nutrient stress media. Data are the mean of at least three biological replicates and errors bars represent the SEM. Significance determined by one-way ANOVA with a Tukey’s post-hoc test. (* p < 0.0001; ** p < 0.002; # p < 0.01; † p = 0.0075).
Fig 4.
Decreased reductase activity and increased round body formation in the gpsA mutant.
(A) Wild-type (WT), (B) ΔgpsA mutant, (C) gpsA+ complement, (D) ΔgpsA/ΔglpD double mutant and (E) ΔglpD mutant strains were grown in BSK + RS at 35°C (to a cell density of 5–9 × 107 cells ml-1) before shifting to RPMI for 16 h at 35°C. Bacterial reductase activity was detected by staining with RedoxSensor Green and membrane integrity assessed by staining with propidium iodide (PI). Live cells were imaged by fluorescence microscopy for RedoxSensor Green (cyan) and PI (magenta) and overlaid with white light images. (F) The percentage of cells stained with RedoxSensor Green and the percentage of cells in round body (RB) form was quantified. Data are the mean of three biological replicates and error bars represent the SEM. Asterisks signify a significant difference (p ≤ 0.0001) between the mean percent of RedoxSensor-stained ΔgpsA cells compared to all other strains and the mean percent of RBs in the ΔgpsA strain compared to all other strains as determined by one-way ANOVA with a Tukey’s post-hoc test.
Fig 5.
GpsA regulates central metabolism and redox balance.
(A) Polar metabolomics of bacterial strains subjected to principal component (PC) analysis with pareto-scaling. (B) The corresponding metabolite loading distribution for the analysis displayed in (A) for principal components 1 and 2. (C) Metabolites that significantly vary between ΔgpsA and wild type (WT) with a false discovery rate (FDR) less than 5%. Values in the heatmap at left are displayed as the log2(fold change mutant versus WT) and values in the accompanying heatmap at right indicate whether that metabolite passes a 5% FDR filter for the indicated comparison as assessed by a Benjamini-Hochberg correction. (D) Metabolic map of the changes in glycolysis and the glycerol shunt that occur with the loss of GpsA. All measured metabolites in the included pathways are displayed. The log2(fold change ΔgpsA versus WT) is displayed as color of the node and the -log(p-value) is displayed as the size of the node. Enzymes in the glycerol arm of metabolism are displayed as diamonds. Data are from four independent biological replicates.
Fig 6.
GpsA regulates NADH/NAD+ levels.
Wild-type (WT, black circles), ΔgpsA mutant (ΔgpsA, white circles), gpsA complemented (gpsA+, dark gray circles) and ΔgpsA/ΔglpD double mutant (light gray circles) strains were grown in BSK + RS at 35°C to ~5 × 107 cells ml-1 and NAD+ and NADH levels were measured with the NAD/NADH-Glo Assay. Each point represents a single biological replicate and bars represent the means. Asterisks represent a significant difference (p < 0.0001) in the means of NADH/NAD+ molar ratios of both the ΔgpsA mutant and the ΔgpsA/ΔglpD double mutant compared to those of both the wild type and the gpsA+ determined by one-way ANOVA with a Tukey’s post-hoc test.
Fig 7.
GpsA and GlpD levels are independent.
(A) Wild-type (WT), ΔglpD mutant (ΔglpD) and glpD complemented (glpD+) strains were grown in BSK + RS at 35°C or shifted to nutrient stress medium (RPMI) for 24 h before total cell lysates were collected. Samples were separated by SDS-PAGE and analyzed by immunoblot with antibodies against GpsA or FlaB (as a control). GpsA is the lower band of the doublet (see S1H Fig). (B) WT, ΔgpsA mutant (ΔgpsA) and gpsA complemented (gpsA+) strains were grown and analyzed as in (A) except antibodies against GlpD were used for the immunoblots in the upper panels. Three independent experiments were done and representative data are shown.
Table 1.
Infectivity by intradermal needle inoculation.
Fig 8.
Persistence of the gpsA and glpD mutants in ticks.
B. burgdorferi strains were introduced to larval ticks by immersion infection and fed on naïve mice. Acquisition and persistence of wild type (WT, black circles), ΔgpsA mutant (white circles), gpsA complement (gpsA+) (dark gray circles) and ΔgpsA/ΔglpD double mutant (light gray circles) was assessed in ticks one week after larvae fed to repletion (A), following the molt to nymphs, nine weeks later (B) and one week after nymphs fed to repletion on naïve mice (C). DNA from infected ticks was isolated at each stage and B. burgdorferi load in ticks measured by TaqMan qPCR using primers/probe to flaB. Bars represent the mean of data points from a single experiment. The means of both the ΔgpsA and ΔgpsA/ΔglpD mutants were significantly different (* p < 0.0007) from both the wild type and the gpsA+ determined by one-way ANOVA with a Tukey’s post-hoc test in (B). The mean of the gpsA+ was significantly different (** p ≤ 0.03) from both ΔgpsA and ΔgpsA/ΔglpD mutants as determined by one-way ANOVA with a Tukey’s post-hoc test in (C).
Table 2.
Infectivity by nymph bite.