Listeria monocytogenes GlmR Is an Accessory Uridyltransferase Essential for Cytosolic Survival and Virulence

ABSTRACT The cytosol of eukaryotic host cells is an intrinsically hostile environment for bacteria. Understanding how cytosolic pathogens adapt to and survive in the cytosol is critical to developing novel therapeutic interventions against these pathogens. The cytosolic pathogen Listeria monocytogenes requires glmR (previously known as yvcK), a gene of unknown function, for resistance to cell-wall stress, cytosolic survival, inflammasome avoidance, and, ultimately, virulence in vivo. In this study, a genetic suppressor screen revealed that blocking utilization of UDP N-acetylglucosamine (UDP-GlcNAc) by a nonessential wall teichoic acid decoration pathway restored resistance to lysozyme and partially restored virulence of ΔglmR mutants. In parallel, metabolomic analysis revealed that ΔglmR mutants are impaired in the production of UDP-GlcNAc, an essential peptidoglycan and wall teichoic acid (WTA) precursor. We next demonstrated that purified GlmR can directly catalyze the synthesis of UDP-GlcNAc from GlcNAc-1P and UTP, suggesting that it is an accessory uridyltransferase. Biochemical analysis of GlmR orthologues suggests that uridyltransferase activity is conserved. Finally, mutational analysis resulting in a GlmR mutant with impaired catalytic activity demonstrated that uridyltransferase activity was essential to facilitate cell-wall stress responses and virulence in vivo. Taken together, these studies indicate that GlmR is an evolutionary conserved accessory uridyltransferase required for cytosolic survival and virulence of L. monocytogenes.

selection using a Himar1 mariner-based transposon mutant library in a DglmR mutant background. Twenty unique transposon insertions disrupting 15 unique genes suppressed the DglmR mutant's lysozyme sensitivity (Table 1), and each of these phenotypes was confirmed by transducing the transposons into a new DglmR mutant. The suppressors represent a diverse set of cellular processes that likely contribute to lysozyme resistance in a variety of ways, including mechanisms that are both generic and GlmR specific. Mutations that generically upregulate stress response pathways may not be useful for understanding GlmR function. Therefore, to prioritize lysozyme suppressor mutants most relevant to the DglmR mutant virulence defect, we assessed the ex vivo virulence of all the transduced DglmR lysozyme suppressor mutants in a plaque assay. The plaquing assay represents the most complete ex vivo assay for virulence of L. monocytogenes requiring cellular invasion, cytosolic survival, intracellular replication, cell-to-cell spread, and secondary vacuole escape (27). In addition to being sensitive to b-lactam antibiotics and lysozyme in vitro, DglmR mutants are unable to form wild-type-sized plaques in fibroblast monolayers (Fig. 1A, B). Only second-site mutations in yfhO, gtcA, and corA statistically significantly rescued the DglmR plaquing defect (Fig. 1B), while second-site mutations in relA, pbpA, and oppA further inhibited plaquing efficiency of DglmR mutants. The yfhO::Tn and gtcA::Tn displayed the most robust suppressor phenotype, so we chose to focus on these mutants for follow-up studies. Both the yfhO::Tn and gtcA::Tn transduced transposons suppress lysozyme sensitivity of a DglmR mutant, consistent with their identification through the lysozyme suppressor screen (Fig. 1C). Furthermore, the double mutants have no impact on growth in the absence of lysozyme, and the lysozyme sensitivity of the suppressor mutants can be restored by expression of yfhO or gtcA in trans, respectively (Fig. S1). In L. monocytogenes 1/2a strains, both YfhO and GtcA are required for modification of the wall teichoic acid (WTA) repeating ribitol subunits with N-acetylglucosamine (GlcNAc) derived from UDP-GlcNAc (15,28,29). We confirmed that the DglmR gtcA::Tn double mutant is defective for GlcNAc WTA decoration based on loss of wheat germ agglutinin staining ( Fig. S2 and Text S1). Finally, disruption of gtcA or yfhO in a DglmR mutant partially restores virulence in a murine model of disseminated listeriosis (Fig. 1D). Taken together, these data suggest that elimination of nonessential decoration of WTA with GlcNAc increases available pools of UDP-GlcNAc, which can rescue DglmR mutant lysozyme sensitivity and virulence ex vivo and in vivo. DglmR mutants have depleted pools of UDP-GlcNAc. Loss of GlcNAc decoration of the WTA restored lysozyme resistance and partial virulence to DglmR-deficient mutants; therefore, we hypothesized that DglmR mutants may have metabolic defects leading to decreased UDP-GlcNAc synthesis. To test this hypothesis, we utilized untargeted metabolomics to identify differentially abundant metabolites in DglmR mutants relative to wild-type (WT) L. monocytogenes. After growth in modified Listeria synthetic media (LSM), metabolites were extracted and untargeted LC-MS was performed. Subsequent analysis using MAVEN software resulted in 1,073 putative metabolites assigned identities according to their m/z and mapped to the Kyoto Encyclopedia of Genes and Genomes (KEGG) (30). Importantly, although not every metabolite was validated by MS-MS or control standards, 37 putative metabolites were identified with .2-fold differences between wild type and the DglmR mutant across three biological replicates ( Fig. 2A, Table S1). The relatively small number of differentially abundant metabolites suggests that GlmR does not have a global regulatory function, at least under the growth conditions tested. Consistent with our hypothesis, UDP-GlcNAc was among the most differentially abundant metabolites in DglmR mutants compared to wild-type L. monocytogenes. UDP-GlcNAc levels were reduced .3-fold in the DglmR mutant ( Fig. 2B) relative to wild-type controls, consistent with the hypothesis from the suppressor screen that UDP-GlcNAc metabolism is disrupted in the DglmR mutant. UDP-N-acetyl-muramic acid (UDP-MurNAc), another peptidoglycan precursor downstream of UDP-GlcNAc (Fig. 2B, C), was similarly decreased in the DglmR mutant (;50% of wild type). UDP-GlcNAc is synthesized by the GlmSMU pathway that converts fructose-6- phosphate into UDP-GlcNAc through a four-step enzymatic process (Fig. 2C). Upstream of UDP-GlcNAc in the GlmSMU pathway, N-acetylglucosamine-1 phosphate (GlcNAc-1P) levels were also significantly reduced in the DglmR mutant; however, UTP levels were unchanged (Fig. 2B, C). We analyzed other metabolites in the GlmSMU pathway but were unable to observe the GlmSMU pathway intermediates glucosamine-1 phosphate (GlcN-1P) and glucosamine-6 phosphate (GlcN-6P) (Fig. 2B, C) via our mass spectrometry (MS) method even when running purified standards. Finally, levels of the glycolytic intermediates fructose-6-phosphate (F6P) and fructose-1,6-bisphosphate (FBP) were unchanged in the DglmR mutant, suggesting that deficits in muropeptide precursors are due specifically to alterations in the UDP-GlcNAc synthesis pathway and not in a more central metabolic pathway (Fig. 2B, Fig. 2C). Consistent with the model that blocking a nonessential UDP-GlcNAc utilizing pathway increases available UDP-GlcNAc for essential peptidoglycan (PG) or WTA synthesis, targeted metabolomics analysis of both the DglmR gtcA::Tn and the DglmR yfhO::Tn suppressor mutants with specific standards for UDP-GlcNAc demonstrated significant rescue of UDP-GlcNAc levels, though not all the way back to wild-type levels (Fig. 2D). Incomplete restoration of UDP-GlcNAc levels in these suppressor mutants could explain the partial complementation phenotype for the suppressors in the context of lysozyme sensitivity and virulence in vivo (Fig. 1C, D). Taken together, these data demonstrate that DglmR mutants have reduced levels of UDP-GlcNAc and suggest that restoration of UDP-GlcNAc pools restores cell-wall stress responses and virulence in vivo.
GlmR is an accessory uridyltransferase. Two recent studies suggested that GlmR's function in B. subtilis is to enhance the activity of GlmS through direct GlmR-GlmS interactions, thus increasing levels of UDP-GlcNAc (Fig. 2C). These studies utilized bacterial two-hybrid assays to demonstrate a direct interaction between B. subtilis GlmR and GlmS (26), and a subsequent study demonstrated that this interaction modulates GlmS activity (31). To determine if GlmR-GlmS interactions are conserved in L. monocytogenes, we expressed both B. subtilis and L. monocytogenes GlmS and GlmR constructs in the BACTH bacterial two-hybrid system. Each protein was expressed independently as both N-and C-terminal fusions to both T18 and T25. Four replicates of the bluewhite assay were performed due to variability in the system from a known thresholding effect (32), and quantitative b-galactosidase assays were performed in triplicate. As predicted based on their crystal structures, GlmS (33) and GlmR (PDB 2Q7X and 1HZB) from both B. subtilis and L. monocytogenes homodimerized, demonstrating that the constructs were expressed and functional (Fig. S3). Positive but inconsistent interactions between B. subtilis GlmR and GlmS were observed as previously reported for one set of B. subtilis fusion proteins ( Fig. S4) (26); however, no combination of L. monocytogenes GlmR and GlmS produced an interaction except those for which there was also activity observed in the empty vector controls (Fig. S5). Taken together, these data suggest that GlmR regulation of GlmS through protein-protein interactions may not be evolutionarily conserved among GlmR homologues and that GlmR may function to regulate UDP-GlcNAc levels by a novel mechanism in L. monocytogenes.
A distant homologue of GlmR is CofD, a 2-phospho-L-lactate transferase involved in the synthesis of Coenzyme F420 in actinobacteria (34). This homology to a catalytic protein suggests that perhaps GlmR has direct enzymatic activity, perhaps as an accessory enzyme in muropeptide biosynthesis. In multiple Gram-positive pathogens, MurZ is an accessory enzyme involved in muropeptide synthesis that is upregulated in the context of cell-wall stress (35). We had previously demonstrated that GlmR protein levels are similarly increased in the presence of lysozyme (16), leading to the hypothesis that GlmR could be an accessory enzyme functioning to increase pools of UDP-GlcNAc in the context of cell-wall stress. To test this hypothesis, we cloned and purified GlmR from L. monocytogenes and assessed its potential enzymatic activity in the last two steps of the canonical GlmSMU pathway normally catalyzed by GlmU to produce UDP-GlcNAc (Fig. 2C). GlmU is a bifunctional enzyme that contains both acetytransferase and uridyltransferase activity. Using mass spectrometry to assess the results of each reaction, we found that GlmR catalyzed the synthesis of UDP-GlcNAc from GlcNAc-1P and UTP (Fig. 3), similar to both commercially purchased Escherichia coli GlmU as well as L. monocytogenes GlmU that we expressed and purified (Fig. 3). Importantly, no UDP-GlcNAc was detectable with substrates UTP and GlcNAc-1P alone, indicating that catalysis required either the GlmU or GlmR protein (Fig. 3). In contrast, we detected no acetyltransferase activity associated with GlmR, suggesting that GlmR is not a dual-functional enzyme like GlmU and further demonstrating that the GlmR activity observed was not an artifact of accidental copurification of GlmU (Fig. S6A). Importantly, in these reactions, the substrate GlcN-1P is not detectable via our MS method, consistent with our inability to detect this metabolite in our untargeted MS method (Fig. 2). Finally, the absence of UDP-GlcNAc in a GlmR reaction GlmR is an Accessory Uridyltransferase mBio mixture lacking GlcNAc-1P and UTP as the substrates or after the protein was heated excludes the possibility of UDP-GlcNAc being a copurified artifact with GlmR (Fig. S6B). Taken together, these data suggest that GlmR can act as a uridyltransferase enzyme to directly facilitate increased production of UDP-GlcNAc in response to cell-wall stress. GlmR uridyltransferase activity is conserved. GlmR is the second gene of a highly conserved operon of three genes found in firmicutes and actinobacteria. In S. aureus, the GlmR homologue YvcK is predicted to be essential (21), while in B.subtilis it is found to be important for synthesis of UDP-GlcNAc (18,22). The S. aureus and B. subtilis GlmR homologues exhibit high homology to L. monocytogenes GlmR, with 46% identity, 69% similarity and 47% identity, 70% similarity, respectively, and are best conserved near the putative N-terminal active site (Fig. 4A). To determine whether GlmR uridyltransfer-  GlmR is an Accessory Uridyltransferase mBio ase enzymatic function is conserved among firmicute homologues, we cloned and purified GlmR from S. aureus and B. subtilis and assessed enzymatic activity. Each protein exhibited uridyltransferase activity similar to L. monocytogenes GlmR (Fig. 4B). To test for functional conservation of GlmR function in vivo, we complemented the L. monocytogenes DglmR mutant with codon-optimized glmR homologues from S. aureus and B. subtilis. As hypothesized based on their conserved enzymatic activity, both the S. aureus and B.subtilis homologues rescued lysozyme sensitivity of a L. monocytogenes DglmR mutant when expressed in trans (Fig. 4C). Taken together, these data suggest that the uridyltransferase enzymatic function of GlmR is conserved in diverse firmicutes, including both pathogens and nonpathogens. GlmR uridyltransferase activity is required for cell-wall stress responses and virulence in vivo. Our data suggest that GlmR can act as an accessory uridyltransferase; however, whether this activity is required for cell-wall stress and virulence is unknown. To determine if GlmR uridyltransferase activity is important for L. monocytogenes lysozyme resistance and virulence, we aimed to create a catalytically inactive GlmR. The amino acid sequence of L. monocytogenes GlmR is highly similar to the Bacillus halodurans GlmR homolog (;47% sequence identity), for which the crystal structure (PDB 2O2Z) has been solved. Based on this similarity, we used Phyre2 to generate an L. monocytogenes GlmR structural model, using the 2O2Z structure as a template, and found the two structures to be superimposable (Fig. 5A). We then overlaid the L. monocytogenes GlmR structural model on the N-terminal uridyltransferase domain of Haemophillus influenzae GlmU (2V0I), which is structurally and biochemically well characterized (36). This analysis revealed several similar structural elements between L. monocytogenes GlmR and H. influenzae GlmU. For instance, both structures harbor a core set of seven b-sheets sandwiched by a-helices. In the H. influenzae GlmU uridyltransferase active site, residues K25, Q76, and D105 coordinate UTP binding and are absolutely required for enzymatic activity. In this proximity in the GlmR model, we identified D40, D41, and N198 residues that are highly conserved among GlmR homologs (Fig. 4A) and predicted that they are important for catalytic activity. This predicted active site is distinct from the site demonstrated to bind UDP-GlcNAc in B. subitilis GlmR, where Y265 and R301 are essential for UDP-GlcNAc binding (31). Notably, Y265 is absent in L. monocytogenes GlmR (Fig. 4A).
To test the hypothesis that uridyltransferase activity is necessary for virulence, we created a D40A, D41A, N198A mutant GlmR (GlmR3), purified the mutant protein, and assessed uridyltransferase activity. Activity of the GlmR3 mutant was ;100-fold reduced in an in vitro biochemical assay compared to wild-type GlmR (Fig. 3F, G). Complementation of a DglmR mutant with glmR3 was unable to rescue lysozyme sensitivity (Fig. 5B) despite equal or even increased levels of expression compared to the wild-type GlmR complement (Fig. S7). Finally, to test the hypothesis that uridyltransferase activity is important for virulence, we infected mice and quantified bacterial burdens in an in vivo model of disseminated listeriosis. In contrast to complementation with wild-type GlmR, the GlmR3 mutant was unable to rescue the virulence defect of the DglmR mutant (Fig. 5C, D). Taken together, these data suggest that the uridyltransferase activity of GlmR is essential for mediating cell-wall stress responses during infection to facilitate virulence of L. monocytogenes.

DISCUSSION
GlmR is a highly conserved protein that is essential for virulence in L. monocytogenes and M. tuberculosis, but whose function remains largely unknown (16,19,20). In this study, we discovered that GlmR has conserved uridyltransferase activity that facilitates cell-wall stress responses during infection. Our findings are also consistent with a recent study utilizing trans-Cinnamaldehyde (t-Cin) hypersensitive L. monocytogenes glmR:Himar1 mutants, which identified suppressor mutations in genes involved in the biosynthesis of UDP-GlcNAc (37). When the glmR:Himar1 mutant was engineered to overexpress glmU, growth in t-Cin was fully restored, whereas overexpression of glmS or glmM only partially restored resistance to t-Cin, further supporting the idea that GlmR is involved in the biosynthesis of UDP-GlcNAc and that the terminal step of the canonical GlmSMU pathway is rate limiting (23,37). Deciphering the activities of proteins of unknown function, such as GlmR, is a major challenge not only in microbial pathogenesis but in biology at large. Indeed, 25% of predicted biochemical reactions do not have an assigned enzyme, suggesting that many proteins of unknown function have enzymatic activity (38,39). Recent metabolomics approaches such as activity-based metabolomics have shown great promise in identifying these functions (39,40). Combining parallel screening approaches such as genetics, transcriptomics, proteomics, and metabolomics generates targeted hypotheses about the roles of proteins of unknown function in physiological processes. In this study, an untargeted metabolomics approach combined with a classical bacterial genetics suppressor screen allowed us to discover the uridyltransferase activity possessed by GlmR. GlmR mutants in L. monocytogenes and other organisms demonstrate both cell-wall stress response defects and defects during growth on gluconeogenic substrates (18,19,22). Although we cannot rule out that GlmR has potentially separable functions in gluconeogenic metabolism and cell-wall precursor metabolism (22), our identification of suppressor mutations that rescue virulence through restoration of UDP-GlcNAc levels suggests that GlmR's role in mediating cell-wall homeostasis via UDP-GlcNAc is critical during infection. GlmR's function in promoting cytosolic survival further suggests that bacteria experience cell-wall stress in the cytosol; however, the cytosolic CAD responsible for imparting cell-wall stress is unknown. Guanylate binding proteins (GBPs) and lysozyme are not responsible for the cytosolic cell-wall stress as GlmR is required for cytosolic survival even in Gbp Chr32/2 and LysM 2/2 macrophages (16,41). Furthermore, our group has previously published that the PASTA (penicillin binding protein and serine/threonine kinase associated) kinase cell-wall stress sensor PrkA is essential for survival in the cytosol, consistent with a direct cell-wall-acting stressor in the macrophage cytosol (42). Alternatively, the GlmR function in promoting cytosolic survival may be due to metabolic stress tied to its potential role in growth on gluconeogenic substrates such as glycerol during cytosolic replication. Importantly, a role for GlmR in direct cell-wall stress responses versus dealing with metabolic stress in promoting cytosolic survival and virulence are not mutually exclusive. Future identification of the cytosolic CADs targeting the bacterial cell wall will illuminate novel host defense pathways, not only against L. monocytogenes, but also other bacteria that invade the cytosol, including both canonical and noncanonical cytosolic pathogens such as M. tuberculosis and S. aureus. Furthermore, other bacterial pathogens that require GlmR for survival and virulence, such as S. aureus (21) and M. tuberculosis (19,20), likely require GlmR to deal with cell-wall stress in their conventional replication niches.
We found that GlmR uridyltransferase activity is conserved in S. aureus, B. subtilis, and M. tuberculosis (data not shown), representatives of the Firmicutes and Actinobacteria phyla. This conservation, combined with its essential role in virulence of a number of important pathogens, suggests that it may be an attractive drug candidate. Indeed, both the acetyl-and uridyltransferase activities of M. tuberculosis GlmU have been targeted by small molecules as a novel antibiotic strategy (43). Whether uridyltransferase inhibitors of GlmU could also bind and inhibit GlmR will need to be assessed. Among GlmR homologues, the N-terminal putative nucleotide binding region is most highly conserved. This raises important questions not only about the design of GlmR small molecule inhibitors, but also about substrate specificity of GlmR homologues and whether different GlmR proteins may have flexibility to catalyze different reactions with regard to the sugar component. Indeed, this may explain why GlmR appears to have a role in cell-wall homeostasis and alternatively a role in gluconeogenic metabolism. Crystal structures of GlmR homologues in complex with their substrates will be critical both for antibiotic development and an understanding of the potential promiscuity of these enzymes.
GlmR uridyltransferase activity is conserved, but the mechanism(s) of regulation of GlmR expression and/or activity remains unknown. In L. monocytogenes, GlmR is upregulated at the protein level by cell-wall stress (16). Additionally, GlmR is phosphorylated by PASTA kinases in L. monocytogenes, B. subtilis, and M. tuberculosis; however, the phosphorylation sites differ, and what effect phosphorylation may have on the enzymatic activity is similarly unknown (16,19,44). Subcellular localization of GlmR may also contribute to its regulation as GlmR localization patterns in B. subtilis and M. tuberculosis are dissimilar but consistent with localization of peptidoglycan synthesis in these organisms (19,22,45,46). Finally, recent studies have suggested that GlmR may also act allosterically to regulate the function of GlmS in B. subtilis (26,31). Although we were unable to observe this interaction in L. monocytogenes, GlmR functioning as an allosteric regulator of GlmS and as a functional uridyltransferase are not mutually exclusive and indeed could act synergistically. Identification of mutations that abolish GlmS-GlmR interaction but not enzymatic activity and vice versa are necessary to separate and test these ideas.
This study identified that GlmR, a protein required for L. monocytogenes and M. tuberculosis virulence, is an accessory uridyltransferase necessary for UDP-GlcNAc synthesis in the context of cell-wall stress. Similar to MurA and MurZ in S. aureus (35), this highlights that virulence determinants can be redundant with essential housekeeping enzymes. Often these accessory enzymes are upregulated in the context of stress, such as during infection or antibiotic treatment as is the case with GlmR and MurZ, respectively (35). Indeed, GlmR's enzymatic activity may have gone previously undiscovered despite its importance, because of the protein's low expression during normal laboratory culture with rich media. Additionally, with a potential exception in S. aureus (21), GlmR is likely not essential under laboratory conditions, because of sufficient uridyltransferase activity of GlmU. Conversely, even in a situation where GlmR complemented GlmU uridyltransferase activity, GlmU would still be essential due to its acetyl-transferase function. Future analysis of virulence determinants of unknown function through parallel screening approaches may reveal this redundancy to be even more pervasive.

MATERIALS AND METHODS
Listeria monocytogenes strains and culture. All L. monocytogenes strains used for experiments in this study were 10403S background. The DglmR mutant was described previously (14). L. monocytogenes was grown overnight in brain heart infusion (BHI) at 30°C for all experiments except as described for metabolomic analysis.
Construction of L. monocytogenes strains. Homologue complementation genes used in Fig. 4 were created with gBlocks (IDT) that were codon-optimized for L. monocytogenes and inserted into pIMK2 (47) under the control of the constitutive pHelp promoter. Constructs were cloned in XL1-Blue E. coli with 30 mg/mL kanamycin for pIMK2 and shuttled into L. monocytogenes through conjugation with SM10 or S17 E. coli.
Suppressor selection. A Himar 1 Tn mutant library was created in a DglmR mutant background as described previously (48). Aliquots of the library were thawed, diluted 1:1,000-10,000 in PBS, and inoculated 1:50 into 1 mL of Luria broth (LB) with 1 mg/mL lysozyme and 0.1 mM staurosporine in pentaplicate. Fifty microliters of cultures were plated at 0 h on LB and 6 h on LB 1 mg/mL lysozyme. This selection was carried out four times, and 313 out of 476 resulting colonies were secondarily screened in BHI with lysozyme 1 mg/mL and staurosporine 0.1 mM. Transposon mutations in the remaining suppressors were identified by 2-step PCR using transposon-specific and degenerate primers followed by sanger sequencing and were confirmed by PCR with diagnostic primers (49). To determine whether identified transposon mutations were causative, all unique transposons were transduced into a fresh DglmR background and reconfirmed with diagnostic PCR, sequencing, and rescue of the DglmR mutant lysozyme sensitivity with overnight growth in 1 mg/mL lysozyme in BHI.
Phage transduction. Phage transductions were performed as previously described (50). Briefly, U153 phage stocks were propagated with MACK L. monocytogenes grown overnight in LB at 30°C. MACK cultures were pelleted and resuspended in LB with 10 mM MgCl 2 and 10 mM CaSO 4 , mixed with 0.7% LB agar 10 mM MgCl 2 10 mM CaSO 4 at 42°C, and immediately poured on LB plates and incubated overnight at 30°C. Plaque lysate was soaked out with 10 mM Tris pH 7.5 10 mM MgCl 2 10 mM CaSO 4 buffer and sterilized by 0.2 mm filtration or addition of 1:3 volume chloroform. Donor plaque lysates were prepared using the same conditions and used to infect recipient DglmR cultures for 1 h at room temperature before being plated on erythromycin selection at 37°C.
Lysozyme sensitivity. Overnight 30°C static BHI cultures were backdiluted 1:50 into 96-well plates containing BHI or BHI with lysozyme at 1 mg/mL. Plates were grown at 37°C with continuous shaking for 12 h in an Eon or Synergy HT Microplate Spectrophotometer (BioTek Instruments, Inc., Winooski, VT), and OD 600 was read every 15 min.
Plaque assay. The plaque assay was performed as described (27) except that the multiplicity of infection (MOI) was adjusted for optimal plaque number and an additional agarose media plug was added to wells at 3 days to facilitate an additional 3 days of plaque growth. At 6 days, wells were stained with 0.3% crystal violet and washed with water. After staining, the dishes were scanned and plaque areas were quantified with ImageJ. All strains were assayed in biological triplicate, and the plaque areas of each strain were normalized to wild-type plaque size within each replicate.
Metabolite extraction. Overnight 30°C static BHI cultures were washed with PBS and backdiluted 1:50 into 50 mL of Listeria synthetic media (LSM) baffled flasks at 37°C, shook, and grown to an OD600 of ;0.4. LSM is a derivative of Improved Minimal Media developed by Phan-Thanh and Gorman (51) with several component changes (52). For metabolomic experiments, we reduced the level of morpholinepropanesulfonic acid (MOPS) to 1/5th the normal amount to reduce background MS signal. Five milliliters of culture were deposited by vacuum filtration onto a 0.2-mm nylon membrane (47 mm diameter) GlmR is an Accessory Uridyltransferase mBio in duplicate. The membrane was then placed (cells down) into 1.5 mL cold (220°C or on dry ice) extraction solvent (20:20:10 vol/vol/vol acetonitrile, methanol, water) in a 60-mm petri dish and swirled. After a few moments, the filter was inverted (cells up) and solvent was passed over the surface of the membrane several times to maximize extraction. Finally, the cell extract was stored at 280°C. Extracts were pelleted at 21,000 rcf (relative centrifugal force) at 4°C for 10 min. We dried ;200 mL of extract normalized to OD, with N 2 gas. Extracts were resuspended in 70 mL of high-pressure liquid chromatography (HPLC)-grade water and pelleted at 2,1000 rcf at 4°C for 10 min to remove particulates. All cultures were extracted in biological triplicate or quadruplicate and in technical duplicate.
Metabolite quantification and analysis. Metabolite quantification and analysis was performed with the same instrument, and chromatography set up as previously described (53). Briefly, samples were run through an ACQUITY UPLC BEH C 18 column in an 18-min gradient with Solvent A consisting of 97% water, 3% methanol, 10 mM tributylamine (TBA), 9.8 mM acetic acid, pH 8.2, and Solvent B being 100% methanol. Gradient was 5% Solvent B for 2.5 min, gradually increased to 95% Solvent B at 18 min, held at 95% Solvent B until 20.5 min, returned to 5% Solvent B over 0.5 min, and held at 5% Solvent B for the remaining 4 min. Ions were generated by heated electrospray ionization (HESI; negative mode) and quantified by a hybridquadrupole high-resolution mass spectrometer (Q Exactive Orbitrap, Thermo Scientific). MS scans consisted of full MS scanning for 70 to 1,000 m/z from time zero to18 min, except that MOPS m/z of 208 to 210 was excluded from 1.5 to 3 min. Metabolite peaks were identified using the Metabolomics Analysis and Visualization Engine (MAVEN) (30,54).
Protein purification. (i) GST tagged protein expression and purification scheme. GlmR, GlmR3, and GlmU were cloned into pGex6P in XL1-Blue E. coli. Following confirmation by sequencing, pGex6P vectors were subsequently transformed into Rosetta E. coli. IPTG was added to 500 mM to induce expression, and 3 h postinduction cells were pelleted, resuspended in PBS, and frozen at -80C. Cell suspensions were thawed and lysed by sonication in the presence of protease inhibitors. Cell debris was pelleted, and cell lysate was filtered with a 0.2-mM filter and loaded onto a prepacked glutathione resin column at 4°C. The column was washed two times with 10-column volumes of cleavage buffer (25 mM Tris, pH 8, 100 mM NaCl, 1 mM DTT) before elution. The column was loaded with 80 units of PreScission Protease in 960 mL of cleavage buffer and incubated overnight at 4°C. Elution was collected the next day by adding 3 mL of cleavage buffer to the column, and concentrated between 15 mM and 23 mM. Protein was stored at 4°C, purity was assessed by SDS-Page, and protein was quantified by bicinchoninic acid BCA assay.
His-tagged protein expression and purification scheme. GlmR homologues were cloned into pET20b in XL1-Blue and Rosettas with pLysS except for CuvA, which was expressed from BL21(DE3) from pET23. IPTG was added to 500 mM to induce expression, and 3 h postinduction, cells were pelleted, resuspended in PBS, and frozen at 280°C. Cell suspensions were thawed and lysed by sonication in the presence of protease inhibitors. Cell debris was pelleted, and cell lysate was filtered with a 0.2-mm filter and loaded onto a HisTrap Ni column (GE Healthcare) at 4°C. The column was washed with PBS and PBS 25 mM imidizole before elution with 250 mM imidizole. Elutions were dialyzed overnight at 4°C into 10 mM Tris pH 7.4 100 mM NaCl, which was prepared at 25°C and concentrated to between 6 and 22 mM. Protein was stored at 4°C, purity was assessed by SDS-PAGE, and protein was quantified by bicinchoninic acid (BCA) assay.
Enzymatic activity. Reactions were carried out in 10 mM Tris pH 7.4, 100 mM NaCl, and 1 mM MgCl 2 buffer. Substrates (GlcNAc-1-P, UTP, or Acetyl CoA) were added at 100 mM, and purified E. coli GlmU (Galen Laboratory Supplies, GL01012), L. monocytogenes GlmU, L. monocytogenes GlmR, GlmR homologues, and heat-inactivated (HI) L. monocytogenes GlmR or GlmU were added at 1 mM and incubated at 37°C for 10 min. Protein was removed with a 3-kDa molecular weight cutoff (MWCO) filter, and resulting reaction mixtures were diluted 1 to 10 in solvent A and analyzed by tandem HPLC-MS and Maven software.
Bacterial two-hybrid. GlmR and GlmS from both L. monocytogenes and B. subtilis were cloned inframe into vectors pU18, pU18C, pKT25, and pKNT25 from the BACTH System kit (Euromedex) using XbaI and KpnI. Constructs were made originally in TAM1 or XL1-Blue E. coli and then moved to BTH101 E. coli for testing. Both blue/white screening on X-gal plates and b-galactosidase assays were carried out as previously described (32).
Mouse infection. Infections were performed as previously described (16). Briefly, 6-to 8-week-old female and male C57BL/6 mice were infected IV with 1 Â 10 5 CFU. Forty-eight hours postinfection, livers and spleens were harvested, homogenized in PBS with 0.1% NP-40, and plated for CFU. Two independent replicates of each experiment with 5 mice per group were performed.
Ethics statement. Mice were cared for according to the recommendations of the NIH, published in the Guide for the Care and Use of Laboratory Animals. All techniques used were reviewed and approved by the University of Wisconsin-Madison Institutional Animal Care and Use Committee (IACUC) under the protocol M005916.
Statistical analysis. Prism 6 (GraphPad Software) was used for statistical analysis of data. Means from two groups were compared with unpaired two-tailed Student's t test. Means from more than two groups were analyzed by one-way ANOVA with a post hoc LSD (least significant difference) test. A Mann-Whitney test was used to analyze nonnormal data from animal experiments. * indicates a statistically significant difference (P is less than 0.05).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, DOCX file, 0.01 MB. We declare no competing interests.