Our previous results indicated that the luminescence of Strep. Xen20 did not accurately reflect bacterial load in vivo and suggested this may be due to a D-glucose-mediated inhibitory effect (Davis et al., 2015). To confirm this finding on a static medium, we plated Strep. Xen20 on Todd-Hewitt with yeast extract (THY) plates supplemented with 0 or 50 mM D-glucose. For this experiment, cells were grown overnight and imaged for luminescence (Figure 1). Colonies with D-glucose supplementation showed decreased luminescence. Interestingly, the presence of sheep blood (SB) also significantly reduced the bioluminescence produced by Strep. Xen20. Similar CFUs were also observed by colony counting.

Strep. Xen20 has the luxABCDE cassette inserted in the iol operon used for inositol catabolism. More specifically, the gene is under the iolE promoter (Park et al., 2003), which encodes 2-keto-myo-inositol dehydratase (Yoshida et al., 2004). The IolR repressor controls the entire iol operon (Yoshida et al., 1999). Under high inositol conditions, the repressor is liberated, and genes are expressed. As such, we tested the contribution of inositol to bioluminescence expression in Strep. Xen20. The results (Figure 2A) indicate increased light from bacterial cells grown in the presence of myo-inositol. In contrast, co-supplementation of Strep. Xen20, with both inositol and a low concentration of glucose (3 mM), overcame this inositol-dependent enhancement. All BLI signal returned to baseline in less than 3 h (Figure 2B). To evaluate this inositol benefit in actively growing cells, equal portions of Strep. Xen20 were spread on Luria broth (LB) plates in the absence or the presence of 6 mM myo-inositol, of 50 mM D-glucose, or both additives (Figure 2C). Results paralleled Figure 1 and the tube tests in Figures 2A,B. BLI from bacterial plates at 14, 24, and 38 h in the presence of D-glucose plates showed near-complete loss of light production, and this was only slightly counteracted by the addition of myo-inositol at the 38 h time point (Figure 2C).

To support cross-conditional comparisons and to coordinate timing, we imaged the four distinct conditions simultaneously over 2 days, i.e., the Strep. Xen20 kinetic profile showed a definite 2-reaction trace with the development of a new inositol peak at 16–17 h that was almost, approximately, fourfold greater than the amount of light emitted by the LB control (Figures 2D,E). Despite similar bacterial growth over the time course, the BLI signal for both the 50 mM D-glucose- and 6 mM myo-inositol-supplemented conditions showed markedly lower light production. Conversely, Staph. Xen29 showed no change in light production with 6 mM inositol supplementation alone and an approximately twofold increase when grown in the presence of D-glucose that was inositol-independent (Figures 2F,G). These results agree with previous research that showed the expression of green fluorescent protein (GFP), placed under the iolE promoter, was decreased in glucose-rich conditions under the iolE promoter in Salmonella enterica serovar Typhimurium strain 14028 (Kroger et al., 2011).

Since inositol cannot function as a carbon source in S. pyogenes M49 strains [based on the Kypto Encyclopedia of Genes and Genomes (KEGG) genomic data], we investigated the agonistic or antagonistic effects of central metabolites on the production of light. As previously stated, bacterial bioluminescence is dependent on the production of substrates by the bacterial cell (Meighen, 1993); therefore, we investigated the link of bioluminescence to central metabolism in S. pyogenes (see pathway scheme in Figure 3). Bioluminescence is entirely dependent on the availability of activated acyl donors, which are produced de novo by the bacterial cell and serve as the substrate for the lux operon to create fatty aldehydes. These substrates are then converted to fatty acids in the presence of FMNH₂ and oxygen, thereby releasing light. Activated acyl donors are synthesized in one of two ways, first, coenzyme A (CoA)-containing acyl compounds can be created via the breakdown of fatty acids by β-oxidation (green box, Figure 3); second, CoA-containing fatty acid building blocks can be created via the fatty acid biosynthesis pathway (salmon box, Figure 3). The second pathway is dependent upon the creation of acetyl-CoA by the glycolysis/lactic acid fermentation pathway (blue box, Figure 3).

Although the KEGG pathways are not available for Strep. Xen20, a pathway exists for the closely related M49 serotype ancestor NZ131¹ (Mcshan et al., 2008). The analysis of this pathway revealed important allosteric and feedback mechanisms for bioluminescence production. First, M49 strains lack the β-oxidation path beyond the creation of hexadecanoyl-CoA from hexadentate. In contrast, S. aureus NCTC8325, the parental strain of a frequently manipulated strain RN4220, showed increased capacity for degradation of fatty acids. Therefore, in M49 strains, most activated acyl donors must be created by the fatty acid biosynthesis pathway. The analysis of these pathways revealed M49 to have all necessary enzymes for these pathways, which led us to investigate the contribution of glycolytic compounds on the BLI signal.

Strep. Xen20 processes sugars by homolactic fermentation, a process that utilizes D-glucose as the preferred carbon source (Levering et al., 2012). Based on KEGG genome available data, glycine is converted to pyruvate via conversion to serine by L-serine dehydratase in reactions essential to the formation of one-carbon metabolites. Therefore, we utilized D-glucose as a glycolytic carbon source and glycine as a non-glycolytic carbon source. D-glucose is present at physiological concentrations ranging between 3.9 and 5 mM in healthy adult mice (Lee and Bressler, 1981). Therefore, we tested exogenous D-glucose and glycine at either 3 or 6 mM (Figures 4A,B). The addition of either exogenous D-glucose and glycine to Strep. Xen20 decreased the BLI signal (Figures 4C,D). In contrary to this, the addition of D-glucose or glycine increased the luminescence expression in Staph. Xen29 (Figures 4E,F).

To test tissue-derived carbon sources, we perturbed the carbon source equilibrium of both light-producing pathogens (Strep. Xen20 and Staph. Xen29) by characterizing the glucose dependency of BLI signal. The dose-dependent effect of D-glucose on light production was monitored following incubation with M9 minimal media supplemented only with casein hydrolysate and yeast extract (Figure 5). The supplements meant to provide the microbe with amino acids and cofactors essential for protein and DNA synthesis. In this moderately enhanced M9 medium, we found that Strep. Xen20 produced detectable levels of luminescence. Providing an additional glycolytic carbon source, in the form of D-glucose, did not significantly increase radiance. In contrast, Staph. Xen29 produced only moderate levels of luminescence in the absence of further carbon sources. The addition of 6 mM D-glucose significantly increased the production of light in Staph. Xen29. Similar to D-glucose, Strep. Xen20 grown in modified M9 alternatively supplemented with glycine (3 mM) produced no significant increase in BLI signal. In contrast, all supplementations increased bioluminescent expression in Staph. Xen29 cells with a maximal BLI signal when glycine was present (Figure 5).

Homolactic fermentation produces pyruvate that is either converted to acetyl-CoA by pyruvate dehydrogenase or to lactate by L-lactate dehydrogenase (LDH, blue box, Figure 3). Pyruvate conversion by pyruvate dehydrogenase would prepare substrates relevant to light production. In contrast, conversion by LDH would scavenge pyruvate away from this potential pool. To test the effect of pyruvate conversion to lactate and the availability of the acetyl-CoA pool, various concentrations (0–64 mM) of oxamate, an inhibitor of LDH, were added to Strep. Xen20 cells (Figure 6; Feldman-Salit et al., 2013). However, luminescence was constant up to 32 mM oxamate (Figure 6A). At 64 mM oxamate, the radiance increased for the glucose-containing sample, but there was also an observed decrease in cell density. A diffusion assay tested the effect that drug eluted into agar has on light production (Figure 6B). Equal CFUs of Strep. Xen20 cells were added to the quadrants of an agar plate, and an oxamate or relevant control discs were added (Figure 6B). Discs were then imbued with equivalent volumes applied to them (20 μl), and the solutions were sterile water, glucose alone (0.022 mg), oxamate (0.14 mg), and both glucose and oxamate. Dried discs were placed on the agar plates. Bioluminescence production was higher in quadrants with oxamate (Figure 6C).

Given the dependence of Strep. Xen20 bioluminescence on glucose homeostasis, we sought to correlate glucose concentration with the relative light production of Strep. Xen20. Light production of Strep. Xen20 decreased at relatively similar rates when added to an equal volume of glucose in phosphate-buffered saline (PBS). However, each sample displayed a different time to initial decrease and terminal luminescence (Figure 7A). Glucose concentration in the PBS and mouse plasma samples was confirmed using a commercial blood glucose meter. Based on the linear relationship, the mouse plasma had a glucose concentration of 3.3 mM (Figure 7B). After the D-glucose standards were allowed to incubate with the Strep. Xen20 culture overnight, an exponentially decreasing relationship was found to exist between the amount of glucose and Strep. Xen20 BLI (Figure 7C). The data were analyzed using a single exponential and used to predict glucose concentration in mouse plasma. Our predicted concentration of glucose in the mouse plasma matched actual glucometer values determined independently.

Brain heart infusion (BHI), sheep blood agar (SBA), and THY were from BD Biosciences (San Jose, CA) or RPI. Kanamycin and D-glucose were purchased from Research Products International (RPI; Mt. Prospect, IL). Glycine was purchased from AMERSCO (Solon, OH). Unless otherwise specified, all reagents were purchased from Sigma Aldrich. M9 minimal medium was prepared as previously described (Miller, 2010).

Strains Strep. Xen20 and Staph. Xen29 (PerkinElmer Inc., Waltham, MA) were grown to stationary phase (OD_600nm > 1) in BHI broth for 18 h at 37°C, and approximate concentration was determined by light scattering at OD_{600 nm} per manufacturer’s instructions. Cultivation conditions were altered either in the liquid media or on the solid media and supplemented with additives, as indicated. Plates and tubes were imaged using the IVIS Lumina XRMS or Lumina II system (PerkinElmer Inc.). BLI was reported as calibrated units of radiance (p/sec/cm²/sr) using the LivingImage software version 4.7 (PerkinElmer Inc.), allowing for comparisons between detectors.

To test if the lux operon was, in fact, under iolE (Park et al., 2003), we diluted bacteria in PBS containing myo-inositol, D-glucose, or both in a final 1 ml volume at equivalent OD_{600 nm.} Tubes were incubated and imaged in the IVIS Lumina XRMS on a 37°C stage and imaged every 20 min for 160 min. Strep. Xen20 was also streaked on a secondary set of plates with LB (RPI) alone or supplemented with either 6 mM myo-inositol, 50 mM D-glucose, or both myo-inositol and D-glucose. Plates were incubated at 37°C and imaged at 14, 24, and 38 h. Catabolite repression was assessed by supplementation of agar with myo-inositol or D-glucose or both in 6-well, flat-bottomed culture plates (Costar, Corning, NY) and imaged every 1.5 h with the IVIS Lumina II for 48 h.

The effect of carbon source availability was determined by incubating cells in sterile PBS (i.e., control) or PBS supplemented with either 3 mM or 6 mM of D-glucose or glycine. Sample tubes were diluted and imaged as before. To determine the contribution of growth rate to Strep. Xen20 expression, this process was repeated for M9 minimal medium supplemented with 1% casein hydrolysate and 0.3% yeast extract. This media provides necessary cofactors for growth in the absence of confounding carbon sources.

To determine the contribution of LDH on Strep. Xen20 bioluminescence, Strep. Xen20 cells in the stationary phase were added to PBS containing a final concentration of 6 mM D-glucose in a 96-well plate. Varying amounts of oxamate were added for final concentrations of 0–256 mM. OD_{600 nm} and luminescence were measured using a Thermo Fischer VarioSkan plate at 20 min intervals over 160 min at a constant 37°C. For comparison purposes, cells treated with oxamate were compared with those untreated in both the control and glucose-supplemented groups. In addition, Strep. Xen20 was grown and imaged for bioluminescence over 48 h on an LB agar plate containing discs saturated with solutions containing 6 mM oxamate, 6 mM D-glucose, or both. The discs were prepared by the addition of 4 equal volume solutions (20 μl) corresponding to sterile water, glucose alone (0.022 mg), oxamate (0.14 mg), and both glucose and oxamate. The discs were dried before placing them on the agar plates.

To evaluate Strep. Xen20 as a potential monitor of local plasma glucose concentrations, Strep. Xen20 grown to the stationary phase in BHI was combined in a 1:1 ratio with glucose in PBS at concentrations between 0 and 10 mM. Strep. Xen20 was also combined with C57BL6 mouse plasma (Innovative Research, Novi, MI). The light production was then monitored at 20 min intervals at 37°C for 2 h using the Glomax 20/20 Luminometer (Promega, Madison, WI). A final measurement was also taken after the samples had been allowed to incubate overnight at room temperature. The glucose level in the glucose standards and mouse plasma were also measured directly using a commercially available ReliOn™ PRIME blood glucose meter (Wal-Mart Stores, Inc., Bentonville, AR) ReliOn PRIME blood glucose test strips. The meter was unable to determine the glucose levels at < 2 mM.

Statistical means of each group were analyzed using two-way, repeated-measures analysis of variance (ANOVA) on R 3.1.2.²

Time, substrate, and the interaction of time × substrate were analyzed for each medium and each bacterium. Results were reported for statistical values p < 0.05 or no significant difference for p > 0.05.