ABSTRACT
Isoprenoids are an essential and diverse class of molecules, present in all forms of life, that are synthesized from an essential common precursor derived from either the mevalonate pathway or the nonmevalonate pathway. Most bacteria have one pathway or the other, but the Gram-positive, facultative intracellular pathogen Listeria monocytogenes is unusual because it carries all the genes for both pathways. While the mevalonate pathway has previously been reported to be essential, here we demonstrate that the nonmevalonate pathway can support growth of strains 10403S and EGD-e, but only anaerobically. L. monocytogenes lacking the gene hmgR, encoding the rate-limiting enzyme of the mevalonate pathway, had a doubling time of 4 h under anaerobic conditions, in contrast to the 45 min doubling time of the wild type. In contrast, deleting hmgR in two clinical isolates resulted in mutants that grew significantly faster, doubling in approximately 2 h anaerobically, although they still failed to grow under aerobic conditions without mevalonate. The difference in anaerobic growth rate was traced to three amino acid changes in the nonmevalonate pathway enzyme GcpE, and these changes were sufficient to increase the growth rate of 10403S to the rate observed in the clinical isolates. Despite an increased growth rate, virulence was still dependent on the mevalonate pathway in 10403S strains expressing the more active GcpE allele.
INTRODUCTION
Isoprenoids represent the largest family of compounds present in all living organisms and are used for a wide variety of processes, including cell wall synthesis, electron transport, and maintaining membrane fluidity (1). Isoprenoids are derived from the essential isoprenoid precursor molecules isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP). These precursors are synthesized by two distinct pathways, the mevalonate pathway and the nonmevalonate pathway (Fig. 1A). While mammals all use the mevalonate pathway (2), bacteria usually carry genes for only one of the two pathways (3). In rare instances bacteria lack both pathways, such as the obligate intracellular bacterium Rickettsia parkeri, but these bacteria depend on the host cell as a source for isoprenoid precursors (4). Listeria monocytogenes is a Gram-positive, facultative intracellular pathogen that, unlike most bacteria, has the genes for both the mevalonate and the nonmevalonate pathways (5). Work from others reported that the rate-limiting enzyme of the mevalonate pathway, HmgR, is essential in L. monocytogenes EGD-e, as strains lacking hmgR cannot grow unless supplemented with mevalonate in the growth media (6, 7). Deleting either of the last two enzymes in the nonmevalonate pathway, GcpE or LytB, had no impact on growth in vitro and a negligible effect on virulence (7).
Mevalonate pathway is essential for aerobic growth. (A) Abbreviated diagram of L. monocytogenes isoprenoid precursor pathways. See reference 5 for the full pathway. (B) Disk diffusion assay using 10 μl of 1 M mevalonate on a BHI agar plate spread with the WT or 10403S ΔhmgR strain, demonstrating mevalonate-dependent growth. White bar is 20 mm. (C) Aerobic BHI growth curve of WT, 10403S ΔhmgR strain, and 10403S ΔhmgR strain with hmgR gene on the integrating plasmid pPL2. The 10403S ΔhmgR strain was grown in BHI supplemented with the indicated concentrations of mevalonate (MEV). ***, P < 0.0001.
While it is curious that L. monocytogenes has both isoprenoid precursor pathways, interest in these pathways also stems from observations that an intermediate molecule of the nonmevalonate pathway, (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMBPP), activates Vγ9Vδ2 T cells, an innate-like T cell subset found in humans and nonhuman primates but not in mice (8). Upon activation by microorganisms with the nonmevalonate pathway or synthetic ligands, Vγ9Vδ2 T cells proliferate, produce proinflammatory cytokines, and have cytotoxic activity against cells presenting HMBPP (9). Vγ9Vδ2 T cells have a broad role communicating with innate and adaptive immune cells to coordinate responses to bacterial infections, but it has been difficult to gain a mechanistic understanding of their functions due to the challenge of developing an appropriate model system (10, 11).
L. monocytogenes is being developed as a cancer vaccine platform, using a live-attenuated recombinant bacterial strain to stimulate CD8+ T cell responses (12). Despite impressive results in mice, human clinical trials have shown promising but limited responses to L. monocytogenes-based treatments. Given the potential role of Vγ9Vδ2 T cells in responses to HMBPP-producing L. monocytogenes in humans, we were interested in further understanding isoprenoid precursor pathways in strain 10403S, the bacterial strain used for these studies. Additionally, we were intrigued by the observation that L. monocytogenes contains both pathways, despite no reported function for the nonmevalonate pathway.
In this study, we further examined the role of the mevalonate and nonmevalonate pathways for L. monocytogenes growth and pathogenesis. Additionally, the nonmevalonate pathway was examined in two L. monocytogenes strains, FSL N1-017 (FSL), which was isolated from trout brine (13), and HPB2262 Aureli 1997 (HPB2262), which was isolated from an outbreak of gastroenteritis (14). These isolates were chosen because they are in L. monocytogenes lineage I, organisms which are far more prevalent in human listeriosis cases and are genetically distinct from 10403S and EGD-e, which are in lineage II (15). We found that the mevalonate pathway was not essential for growth when ΔhmgR mutants were cultured anaerobically, as strains with only the nonmevalonate pathway grew in the absence of mevalonate. Additionally, three amino acid differences found in the lineage I strains could significantly alter GcpE function in L. monocytogenes strain 10403S, resulting in improved anaerobic growth compared to that of wild-type (WT) L. monocytogenes. These results demonstrate that either the mevalonate pathway or the nonmevalonate pathway is sufficient for L. monocytogenes growth under anaerobic conditions.
RESULTS
L. monocytogenes 10403S does not require the mevalonate pathway for anaerobic growth.Consistent with previous reports (7), a 10403S strain lacking hmgR (the 10403S ΔhmgR strain) was unable to grow on brain heart infusion (BHI) agar unless supplemented with mevalonate (Fig. 1B). In liquid media the growth of 10403S ΔhmgR was fully restored with 1 mM mevalonate or if hmgR was complemented on an integrating plasmid, pPL2 (Fig. 1C).
The last two enzymes of the nonmevalonate pathway, GcpE and LytB, contain [4Fe-4S] iron-sulfur clusters (FeS clusters) (16, 17). In the presence of oxygen, FeS clusters can be oxidized, leaving a catalytically inactive [3Fe-4S]1+ FeS cluster (18). GcpE and LytB (also referred to as IspG and IspH, respectively) are particularly oxygen labile in other organisms, since the FeS cluster is solvent exposed (19–21). We hypothesized that L. monocytogenes growth using the nonmevalonate pathway was promoted anaerobically due to decreased oxidation of GcpE or LytB FeS clusters. Indeed, strains lacking hmgR grew in the absence of mevalonate, although they exhibited a severe increase in doubling time, increasing from 43 min aerobically with mevalonate to approximately 4 h anaerobically without mevalonate (Fig. 2A). Additionally, the 10403S ΔhmgR strain formed a visible colony overnight on BHI agar supplemented with mevalonate, while it required 4 to 5 days to form similarly sized colonies anaerobically without mevalonate. Growth was dependent on the nonmevalonate pathway, as mutants in both pathways (10403S ΔhmgR ΔgcpE and 10403S ΔhmgR ΔlytB mutants) lost viability over the course of the anaerobic experiment (Fig. 2A) and did not form colonies on BHI agar anaerobically without mevalonate.
Nonmevalonate pathway is sufficient for anaerobic growth. (A) Anaerobic BHI growth curve of 10403S ΔhmgR strain compared to mutants in both pathways (10403S ΔhmgR ΔgcpE and 10403S ΔhmgR ΔlytB mutants). (B) Number of CFU measured after plating aerobic 10403S ΔhmgR cultures under indicated oxygen or mevalonate conditions. (C) Anaerobic growth of 10403S ΔhmgR compared to ΔhmgR mutants from indicated L. monocytogenes strains. (D) Number of CFU measured after plating aerobic FSL ΔhmgR, HPB ΔhmgR, and EGD-e ΔhmgR cultures under indicated oxygen or mevalonate conditions. (E) Anaerobic growth of 10403S ΔhmgR ΔgcpE strain with gcpE from indicated strains on pPL2 compared to that of 10403S ΔhmgR strains with gcpE on the chromosome.
To examine whether anaerobic growth was a result of acquired suppressor mutations, the plating efficiency of strains was examined under aerobic and anaerobic conditions after overnight cultures were plated with or without mevalonate or oxygen. Aerobically, no 10403S ΔhmgR colonies grew unless the media contained mevalonate. Anaerobically, equal numbers of bacteria were recovered without mevalonate relative to media with mevalonate (Fig. 2B). Collectively, these data indicate that the nonmevalonate pathway supports growth in the absence of the mevalonate pathway.
Growth phenotype of mevalonate pathway mutants in other L. monocytogenes strains.Previous work on L. monocytogenes isoprenoid precursor pathways was conducted using strain EGD-e, where the mevalonate pathway was found to be essential (7). However, the authors did not examine anaerobic growth of their mevalonate pathway mutant, so we were curious whether any strain-specific growth differences existed between EGD-e and 10403S. The mevalonate pathway was deleted in EGD-e (EGD-e ΔhmgR strain) and two lineage I strains, L. monocytogenes FSL N1-017 (FSL ΔhmgR strain) and L. monocytogenes HPB2262 (HPB ΔhmgR strain). No growth defects were observed when the parental strains (EGD-e, FSL N1-017, and HPB2262) were cultured anaerobically or when ΔhmgR mutants were cultured with mevalonate, either aerobically or anaerobically (data not shown). Both lineage I strains doubled in approximately 2 h when hmgR was deleted, which was significantly faster than the 10403S ΔhmgR strain (Fig. 2C). The EGD-e ΔhmgR strain grew on BHI agar anaerobically without mevalonate (see Fig. S1 in the supplemental material) but did not grow in liquid culture, preventing the measurement of a doubling time. Additionally, when the plating efficiency of these strains was measured, equal numbers of bacteria were recovered anaerobically without mevalonate relative to media with mevalonate (Fig. 2D), similar to what was observed with the 10403S ΔhmgR strain.
To determine whether there was a genetically encoded basis for the observed differences in growth, we compared the 5′ untranslated regions (UTRs) and coding sequences of 10403S, FSL N1-017, and HPB2262 for each gene in the nonmevalonate pathway. Three genes, dxr, ispE, and ispF, were identical among all three strains. The 5′ UTR of ispD differed between strains, but no growth differences were observed when the genes for ispD or lytB from the lineage I strain were introduced into the 10403S ΔhmgR strain. However, strains complemented with gcpE from either FSL N1-017 or HPB2262 grew significantly faster than strains with 10403S gcpE (Fig. 2E).
Three residues in GcpE account for differences in anaerobic growth.GcpE chimeric proteins were constructed to more precisely identify the molecular origin of the growth differences between 10403S and the lineage I strains. First, the 5′ UTRs and coding sequences from each strain were exchanged and introduced into the 10403S ΔhmgR ΔgcpE strain. Only strains complemented with the protein-coding sequence from either lineage I strain grew significantly faster, even if expressed downstream of the 10403S 5′ UTR (Fig. 3A). Twelve amino acid differences exist in GcpE between 10403S and the lineage I strains. A chimeric protein was constructed containing the N-terminal portion of 10403S and the C-terminal portion of the FSL N1-017 protein, from amino acid 251 (Fig. 3B). The C-terminal end was chosen because it contained all four residues involved with FeS cluster coordination and the majority of the catalytic residues. Strains complemented with the chimeric strain grew as well as strains with the entire FSL N1-017 gene (Fig. 3C).
Increased growth of lineage I strains is caused by changes in GcpE coding sequence. (A) Anaerobic growth of 10403S ΔhmgR ΔgcpE strain complemented with the gcpE coding sequence from different strains, all expressed from the 10403S 5′ UTR. (B) Sequence alignment of L. monocytogenes GcpE found in two lineage I strains and two laboratory strains. Listed residues indicate sequence differences relative to 10403S. (C) Anaerobic growth of 10403S ΔhmgR and FSL N1-017 ΔhmgR strains and 10403S ΔhmgR ΔgcpE strain complemented with a GcpE chimera containing the N-terminal domain (amino acids 1 to 250) from 10403S and the C-terminal domain from FSL N1-017 (amino acids 251 to 369). (D) Anaerobic growth of the 10403S ΔhmgR ΔgcpE strain complemented with point mutation GcpEI343V/D344E or GcpEF357Y/V359E.
The C-terminal end of GcpE has nine differences between the lineage I strains and 10403S, the majority of which are located in three distinct areas; therefore, we focused on those changes. Two pairs of mutations, GcpEI343V/D344E and GcpEF357Y/V359E, had no impact on anaerobic growth (Fig. 3D). Mutating three other residues in combination, GcpEK291T/E293K/V294A (GcpE*), significantly increased the growth rate of the 10403S ΔhmgR ΔgcpE strain, while the single mutants of either GcpEK291T or GcpEV294A were not sufficient to increase growth rate (Fig. 4A). Furthermore, when the same GcpE* mutations were made on the chromosome rather than on a plasmid, the resulting 10403S ΔhmgR gcpE* strain had an anaerobic growth rate identical to that of the FSL ΔhmgR strain (Fig. 4B).
Point mutations fixing GcpE function increase anaerobic growth. (A) Anaerobic growth of ΔhmgR ΔgcpE strain complemented with GcpEK291T/E293K/V294A or point mutations GcpEK291T and GcpEV294A. (B) Anaerobic growth of 10403S ΔhmgR strain, FSL N1-017 ΔhmgR strain, and 10403S ΔhmgR strain with GcpEK291T/E293K/V294A mutation on the chromosome. (C) Predicted GcpE structure generated with Phyre2 protein structure prediction software. Two GcpE dimers, colored in green and cyan, with iron-sulfur cluster coordinating residues colored in purple and residues 291 to 294 colored in red.
Given the distinct growth differences between the 10403S ΔhmgR and 10403S ΔhmgR gcpE* strains, we sought to understand how these mutations impacted protein function. The protein structure prediction program Phyre2 was used to generate a model of GcpE based on the crystal structure from Aquifex aeolicus (strain VF5). The predicted structure placed residues 291 to 294 in a loop distal from both the FeS cluster and the catalytic TIM barrel in the protein (Fig. 4C), providing little insight into the mechanism by which these mutations alter 10403S ΔhmgR gcpE* strain anaerobic growth.
Nonmevalonate pathway function does not alter L. monocytogenes virulence in mice.Previous studies found minimal virulence defects when the nonmevalonate pathway was mutated in EGD-e (6, 7, 22), so we were curious whether 10403S had a similar phenotype. No significant virulence defects were observed in an intravenous (i.v.) infection (Fig. S2A) or a 5-day oral infection (Fig. S2B) with nonmevalonate pathway mutants. Furthermore, there were no differences in virulence between 10403S ΔhmgR and 10403S ΔhmgR gcpE* strains in an i.v. infection (Fig. 5A) and oral infection (Fig. 5B) model, as both had significant virulence defects but were not different from each other.
L. monocytogenes virulence in mice does not change with increased nonmevalonate pathway growth. (A) Number of CFU recovered from indicated organs 1 day after i.v. infection with indicated L. monocytogenes strains. Data show two independent experiments, with ten mice total for each strain, plotting medians and interquartile ranges. (B) Number of CFU recovered from mice 5 days after oral infection with indicated L. monocytogenes strains. Data show two independent experiments, with ten mice total for each strain. (C) Number of CFU recovered from pellets after oral infection with indicated L. monocytogenes strains. Plotted line is median number of CFU from two independent experiments (10403S hly::Tn917, 10403S gcpE* hly::Tn917, n = 20; 10403S ΔgcpE, 10403S ΔlytB, n = 15), and the shaded area is the interquartile range. P.I., postinfection.
We hypothesized that the nonmevalonate pathway makes a minor contribution to growth, and if so, differences may become apparent in a long-term oral infection model. During an oral infection, systemic bacterial dissemination adds an additional layer of complexity for understanding bacterial survival in the gut. Bacteria that spread from the gut can colonize the gallbladder and reseed the intestinal tract (23–25). This reservoir of bacteria is then the principal source of bacteria shed in the feces, rather than bacteria that have only survived in the anaerobic environment of the gut. Therefore, to prevent systemic dissemination and more closely examine long-term survival, a transposon (26) was used to disrupt the essential virulence factor listeriolysin O in our WT strain (10403S hly::Tn917), both nonmevalonate pathway deletions (10403S ΔgcpE hly::Tn917 and 10403S ΔlytB hly::Tn917) and the fast-growing GcpE mutant (10403S gcpE* hly::Tn917). After oral infection, the number of L. monocytogenes CFU in the feces decreased over time, but mice continued to shed detectable amounts of bacteria 31 days postinfection (Fig. 5C). Previous studies show that mice stop shedding WT bacteria 2 weeks postinfection in a streptomycin pretreatment model (27), so it was unexpected that all 10403S hly::Tn917 strains continued to be shed 4 weeks postinfection. While small differences in the median number of CFU shed from mice were noted, these changes did not rise to the level of statistical significance. These data suggest that the increased in vitro anaerobic growth with a GcpE* mutation does not impact bacterial survival when the bacteria are confined to the gut of mice.
DISCUSSION
The results of this study showed that the L. monocytogenes mevalonate pathway was essential for growth aerobically and supports growth anaerobically, while the nonmevalonate pathway was sufficient for growth only anaerobically. This growth phenotype was observed in four L. monocytogenes strains, although strains FSL N1-017 and HPB2262 lacking hmgR grew significantly faster than 10403S and EGD-e. Genetic approaches were used to identify three amino acid residues in the nonmevalonate pathway enzyme GcpE that were sufficient to eliminate differences in anaerobic growth between strains. A strain with GcpEK291T/E293K/V294A mutations grew more rapidly in pure culture but did not have significantly altered growth phenotypes in mice.
As a facultative anaerobe, L. monocytogenes substantially reprograms its metabolism in the absence of oxygen (28). We show that the nonmevalonate pathway functions anaerobically but note that L. monocytogenes still grows faster using the mevalonate pathway. This implies that there may be unidentified anaerobic growth conditions where the nonmevalonate pathway is required or beneficial for growth. Aerobically, L. monocytogenes uses oxygen as a terminal electron acceptor to maintain NAD+/NADH ratios while pyruvate is converted to a variety of reduced (lactate and ethanol) and oxidized (acetate and acetoin) fermentation products. Anaerobically, pyruvate represents the primary electron acceptor, and consequently, the reduced fermentation product lactate predominates (29). Reflecting these metabolic changes, pyruvate dehydrogenase is downregulated anaerobically and, presumably, less acetyl coenzyme A (acetyl-CoA) is generated (28). The mevalonate and nonmevalonate pathways start with distinct molecules to generate isoprenoid precursors. The mevalonate pathway uses two molecules of acetyl-CoA, while the nonmevalonate pathway uses one molecule of pyruvate and one molecule of glyceraldehyde 3-phosphate. Under anaerobic conditions, the metabolic shift toward lactate fermentation may limit acetyl-CoA levels, or the acetyl-CoA that is present may need to be used for NAD+ regeneration. As a result, L. monocytogenes may use the nonmevalonate pathway anaerobically as an acetyl-CoA-independent means of producing isoprenoid precursors. However, under conditions used in this study, the mevalonate pathway was sufficient.
We hypothesize that the nonmevalonate pathway fails to function aerobically in L. monocytogenes because GcpE and LytB are oxygen-labile, FeS cluster-containing enzymes. FeS clusters are essential metal cofactors in enzymes but can be oxidized and lost in the presence of oxygen. As a result, organisms have developed mechanisms for protecting FeS clusters from oxidation, including multiple biosynthetic pathways to produce replacement FeS clusters for those damaged by oxidative stresses (30). Three distinct pathways for FeS cluster synthesis have been identified and designated the Nif, Suf, and Isc systems (31). The Gram-negative bacterium Escherichia coli has both the Isc and Suf systems (32, 33), while the Gram-positive model organism Bacillus subtilis and L. monocytogenes only have the Suf system. Isc and Suf genes are essential, but recent studies show that they are essential for synthesizing the FeS clusters in GcpE and LytB (34, 35). B. subtilis only encodes the nonmevalonate pathway but grows aerobically, which suggests that L. monocytogenes is fundamentally different from B. subtilis and may have a defect synthesizing FeS clusters, which prevents it from using the nonmevalonate pathway aerobically.
The crystal structure of L. monocytogenes GcpE has not been solved, but structures from other organisms point toward a mechanism by which mutations in GcpE may alter enzyme function (36–39). GcpE has two major domains, consisting of a catalytic TIM barrel at the N-terminal end and an FeS cluster coordinated by three cysteine residues and one glutamic acid residue at the C-terminal end. The protein forms a homodimer consisting of two subunits aligned head-to-tail. This allows the FeS cluster of one subunit to catalyze reactions in the active site of the opposite subunit, but this mechanism requires the C-terminal domain to rotate significantly while closing on a ligand (38, 40). We hypothesize that all three amino acid changes identified are necessary to alter enzyme kinetics and allow catalysis to proceed more quickly.
Dozens of preclinical studies and clinical trials have used L. monocytogenes 10403S as a vaccine vector for cancer immunotherapy (12, 41–45). All the strains used in these studies produce HMBPP based on the observation that they stimulate Vγ9Vδ2 T cells (43). Based on the results of this study, we predict that 10403S makes less HMBPP than the lineage I strains, although absolute HMBPP concentrations have not been directly measured. The role of Vγ9Vδ2 T cells is still unclear, so it is difficult to determine whether cellular activation in response to HMBPP enhances or diminishes the efficacy of an L. monocytogenes-based vaccine in primates. However, characterizing HMBPP production in the presence of the mevalonate pathway and metabolically engineering 10403S to produce more HMBPP may be useful to improve our understanding of Vγ9Vδ2 T cells and their role in vaccine development (46).
The two L. monocytogenes isolates used in this study raise interesting questions about the relationship between the nonmevalonate pathway and L. monocytogenes pathogenesis in humans. HPB2262 was isolated from an outbreak of febrile gastroenteritis caused by contaminated corn salad (14), but the symptoms of the outbreak were unique because the disease was almost exclusively noninvasive. In contrast, FSL N1-017 was isolated from trout in brine and is not associated with any human outbreaks. However, it is closely related to strain FSL R2-503, which is a clinical isolate from a different outbreak of gastroenteritis (13, 47). It is possible that having a functional nonmevalonate pathway provides L. monocytogenes strains with a selective growth advantage in the human gut. Alternatively, they may have a greater capacity to produce HMBPP and stimulate Vγ9Vδ2 T cells, which may trigger an immune response that contributes to disease (48). However, separating causative versus correlative factors related to Vγ9Vδ2 T cells would be challenging, given the absence of animal models and likelihood that other genes in lineage I L. monocytogenes strains influence pathogenesis.
A large number of clinically significant bacterial and protozoan pathogens, such as Vibrio cholerae, Pseudomonas aeruginosa, Clostridium difficile, Mycobacterium tuberculosis, and Plasmodium species, have the nonmevalonate pathway and produce HMBPP. However, the nonmevalonate pathway is essential for producing isoprenoid precursors in all of these organisms, making it extremely difficult to separate growth defects from virulence defects. In bacteria that encode both pathways, including Mycobacterium marinum (49) and several Streptomyces and Nocardia species (50), even less is known about isoprenoid pathways. By understanding the role of the nonmevalonate pathway in L. monocytogenes, it may be possible to metabolically engineer other bacteria to intentionally manipulate Vγ9Vδ2 T cell activation levels and improve adaptive immune responses to pathogens.
MATERIALS AND METHODS
Construction of L. monocytogenes strains.The L. monocytogenes strains used in this study are all derived from wild-type 10403S (DP-L6253), unless otherwise noted, and are listed in Table 1. Gene deletions were generated by allelic exchange using the plasmid pKSV7 (51). All L. monocytogenes strains were grown in brain heart infusion (BHI) broth supplemented with 5 g/liter yeast extract, 1 g/liter l-cysteine hydrochloride, and 0.001 g/liter resazurin sodium salt, and medium was supplemented with 1 mM mevalonate as needed to support the growth of auxotrophic ΔhmgR strains. Mevalonate was produced by hydrolyzing dl-mevalonolactone (CAS number 674-26-0; Sigma-Aldrich) with 1 N NaOH at 37°C for 1 h according to previously reported methods (6). Agar plates were incubated anaerobically using a BD GasPak EZ anaerobic pouch system (no. 260683), and liquid cultures were incubated in an anaerobic chamber (Coy Laboratory Products) with an environment of 2% H2 balanced with N2.
L. monocytogenes strains
E. coli strains used in this study are listed in Table 2. For vector construction, plasmids were introduced into TOP10 E. coli (Invitrogen) or XL1 Blue E. coli (Stratagene). Plasmids were then transformed into SM10 E. coli and conjugated into L. monocytogenes. PCR was performed with KAPA HiFi DNA polymerase (Kapa Biosystems) or Q5 DNA polymerase (NEB). Positive clones were identified by performing colony PCRs using SapphireAmp fast PCR master mix (TaKaRa Bio) and verified by Sanger sequencing.
E. coli strains
Aerobic growth curves.Strains were grown overnight at 37°C in filter-sterilized BHI and were supplemented with 500 μM mevalonate as necessary. Bacteria were washed with phosphate-buffered saline (PBS) and diluted in 20 ml fresh BHI to an optical density at 600 nm (OD600) of 0.01. Cells were cultured at 37°C with shaking, and growth was measured spectrophotometrically hourly.
Anaerobic growth curves.Medium was degassed overnight in an anaerobic chamber to allow residual oxygen to diffuse out of the medium. Overnight cultures of cells were first grown aerobically and were back diluted into anaerobic medium to a concentration of 103 CFU/ml. Samples were removed from the chamber daily and plated by 10-fold serial dilutions to enumerate CFU.
Structure prediction.The Phyre2 protein modeling web portal was used to generate a predicted model of L. monocytogenes 10403S GcpE (52).
i.v. infections.Intravenous (i.v.) infections were adapted from previously reported methods (53, 54). Briefly, 8-week-old female CD-1 (Charles River) mice were infected i.v. with 105 CFU in 200 μl of PBS, and organs were harvested 48 h postinfection. When mice were infected with ΔhmgR mutants, the inoculum was increased to 106 CFU, and organs were harvested 24 h postinfection due to the growth defect of this strain. To measure organ CFU, mice were euthanized and spleens and livers were removed, homogenized in 5 or 10 ml, respectively, of 0.1% IGEPAL CA-630 (Sigma), and plated to enumerate bacteria.
Oral infections.Oral infections were adapted from previously reported methods (27, 55, 56). Briefly, 5 mg/ml of streptomycin sulfate was added to the drinking water of 8-week-old female CD-1 (Charles River) mice 2 days prior to infection. One day prior to infection, mice were transferred to clean cages and chow was removed to fast the mice overnight. Mice were fed a small piece of white bread inoculated with 107 CFU of L. monocytogenes in 5 μl of PBS, and 3 μl of butter was overlaid on the bread. When mice were infected with ΔhmgR mutants, the inoculum was increased to 108 CFU due to the growth defect of this strain. After infection, mice were returned to cages with standard drinking water and chow. To measure infection burden, pellets were collected from each mouse, weighed, homogenized in 1 ml PBS, and plated by serial dilution to enumerate CFU. Homogenates were plated on selective BHI agar supplemented with 6 g/liter lithium chloride, 6 g/liter glycine, 50 mg/liter nalidixic acid, and 200 mg/liter streptomycin (57). Pellets were collected daily for the first 5 days postinfection and every 2 days for the remainder of the experiment. These studies were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (58). All protocols were reviewed and approved by the Animal Care and Use Committee at the University of California, Berkeley (AUP-2016-05-8811).
ACKNOWLEDGMENTS
We thank J. D. Keasling and J. Kirby for suggesting we examine our mutant anaerobically. We also thank S. H. Light and A. Louie for critical reading of the manuscript and S. H. Light for help with structure modeling.
This work was supported by the National Institutes of Health grants 1P01 AI063302 (D.A.P.) and 1R01 AI027655 (D.A.P.).
FOOTNOTES
- Received 3 October 2019.
- Returned for modification 15 November 2019.
- Accepted 21 November 2019.
- Accepted manuscript posted online 2 December 2019.
Supplemental material is available online only.
- Copyright © 2020 American Society for Microbiology.
