Divergent Evolution of the Activity and Regulation of the Glutamate Decarboxylase Systems in Listeria monocytogenes EGD-e and 10403S: Roles in Virulence and Acid Tolerance

The glutamate decarboxylase (GAD) system has been shown to be important for the survival of Listeria monocytogenes in low pH environments. The bacterium can use this faculty to maintain pH homeostasis under acidic conditions. The accepted model for the GAD system proposes that the antiport of glutamate into the bacterial cell in exchange for γ-aminobutyric acid (GABA) is coupled to an intracellular decarboxylation reaction of glutamate into GABA that consumes protons and therefore facilitates pH homeostasis. Most strains of L. monocytogenes possess three decarboxylase genes (gadD1, D2 & D3) and two antiporter genes (gadT1 & gadT2). Here, we confirm that the gadD3 encodes a glutamate decarboxylase dedicated to the intracellular GAD system (GADi), which produces GABA from cytoplasmic glutamate in the absence of antiport activity. We also compare the functionality of the GAD system between two commonly studied reference strains, EGD-e and 10403S with differences in terms of acid resistance. Through functional genomics we show that EGD-e is unable to export GABA and relies exclusively in the GADi system, which is driven primarily by GadD3 in this strain. In contrast 10403S relies upon GadD2 to maintain both an intracellular and extracellular GAD system (GADi/GADe). Through experiments with a murinised variant of EGD-e (EGDm) in mice, we found that the GAD system plays a significant role in the overall virulence of this strain. Double mutants lacking either gadD1D3 or gadD2D3 of the GAD system displayed reduced acid tolerance and were significantly affected in their ability to cause infection following oral inoculation. Since EGDm exploits GADi but not GADe the results indicate that the GADi system makes a contribution to virulence within the mouse. Furthermore, we also provide evidence that there might be a separate line of evolution in the GAD system between two commonly used reference strains.


Introduction
Survival in sometimes harsh environmental conditions is vital for any pathogen en route to infection of the host. The foodborne pathogen Listeria monocytogenes is well noted for an ability to withstand high salt environments [1][2][3], high pressure [4,5], grow at low temperature [6] and within a broad pH range over which it can survive [3,7,8]. This makes it a major concern for the food industry where preservation methods often employ combinations of pH, salinity and temperature controls. In order for L. monocytogenes to survive low pH environments, the bacterium has evolved several mechanisms that allow it to maintain pH homeostasis. These include the arginine deiminase system [9], an F 0 F 1 ATPase [10], the adaptive acid tolerance response (ATR) [11] and the glutamate decarboxylase (GAD) system [12]. The GAD system has been shown in L. monocytogenes to be important for survival in synthetic gastric fluid [12] but not in the presence of organic acids [8] commonly found in foods.
The accepted model for the GAD system ( Fig. 1) involves the combined action of a membrane bound antiporter (GadT) and a cytosolic glutamate decarboxylase (GadD). During exposure to low pH, the bacterium can exchange an extracellular molecule of glutamate for an intracellular molecule of c-aminobutyric acid (GABA) via the GadT antiporter/s. This imported glutamate then undergoes a decarboxylation to form GABA via the GadD enzyme/s. At pH , 4.5, glutamate is imported in a neutral form (Glu 0 ) [13], which allows the removal of intracellular H + when glutamate is converted to GABA. This consumption of intracel-lular protons helps to maintain a tolerable intracellular pH. GABA generated via this reaction is expected to exit the cell via the antiporter in exchange for further glutamate, allowing a cycling process to continue (Fig. 1). In previous work we have shown that GAD activity can take place independently of the antiporter, a finding that prompted a revision of the previous model by introducing the concepts of an extracellular GAD system (GAD e ) i.e. a GAD system relying on the Glu/GABA antiport and an intracellular GAD system (GAD i ) i.e. a GAD system that relies on intracellular pools of glutamate or glutamate possibly imported into the cell via a glutamate transporter [14].
For L. monocytogenes, the glutamate/GABA antiporter can be encoded by one of two genes (gadT1, T2) while it possesses up to three decarboxylase encoding genes (gadD1, D2 & D3) [15]. Previously, gadD3 has been considered as a putative glutamate decarboxylase encoding gene and some lines of evidence suggest that this glutamate decarboxylase may form an integral part of the GAD i [14]. These genes are encoded in three transcriptional units gadD1T1, gadT2D2 and gadD3 [16] (Fig. 1). While all strains possess both gadT2D2 and gadD3, strains from serotype 4 do not possess the gadD1T1 operon [16]. Previously, genetic studies on the GAD system have shown that gadD2 is critical for survival of L. monocytogenes at low pH, while gadD1 is important at a milder pH [12,16]. This work was carried out in the strain LO28 (serotype 1/2c), however recent work has shown that the GAD system can behave differently depending on both strain and media type. Analysis of strain EGD-e has shown that the antiporters are non functional, which results in a GAD system that produces intracellular GABA from the cytoplasmic pool of glutamate [14]. Furthermore, in a chemically defined media, the strain 10403S also relies solely on the activity of GAD i , despite having a functional antiport mechanism [17].
Despite what has been learned so far it is apparent that the accepted model of the GAD system in L. monocytogenes is not complete and certain knowledge gaps remain including elucidating the role of the GadD3 and of the other GAD genes in different environmental conditions as well as the role the GAD system plays in vivo. Therefore, here we compare the GAD system of two important reference strains of L. monocytogenes, EGDm (a murinised form of EGD-e) and 10403S and we describe an extensive genetic and physiological analysis of the GAD system in these strains, including the previously uncharacterised gadD3 gene. Furthermore, while studies have shown that the GAD system is important for survival in porcine gastric juice, the role of this system during pathogenesis in a live animal model remains unknown.

Bacterial strains and growth
All strains used in this study are listed in Table 1. Strains were grown on BHI (Lab M) agar plates at 37uC for 24 to 48 h. Cultures grown overnight were set up by inoculating a single colony into 25 ml BHI broth in 250 ml Erlenmeyer flasks and incubating at 37uC with aeration.

Generation of knockout mutants
Deletion mutants were generated using the splicing by overlap extension (SOEing) PCR method and allelic replacement, as previously described [18]. All primers used in the process are listed in Table 2. Deletion cassettes were cloned into the shuttle vector pKSV7. Double mutants were generated by introducing the deletion cassette into the confirmed single EGDm DgadD3 mutant. PCR amplification of DNA for use in cloning and downstream work was carried out by using the high-fidelity Velocity DNA polymerase (Bioline), while screening was carried out by using Biotaq DNA polymerase (Bioline).

Acid survival assays
Cultures of bacteria were grown to mid-exponential (OD 600 = 0.35) or stationary phase (16 h) at 37uC in BHI medium. The pH of these cultures was lowered with 3 M HCl to pH 3.5 for mid-exponential or pH 2.5 for stationary phase cultures. Samples were taken every 20 min for 1 h and serially diluted in phosphatebuffered saline (PBS). Dilutions were plated in triplicate onto BHI agar and incubated overnight at 37uC. Colonies were counted to determine the number of surviving cells.

GABase assays
Intracellular GABA (GABA i ) and extracellular GABA (GABA e ), were measured as previously described [17,19]. Cultures were grown to mid-exponential phase (OD 600 = 0.35) or stationary phase (16 h) at 37uC with aeration in BHI medium. For exponential phase GABA measurements the cultures were reduced to pH 4.0, while for stationary phase GABA measurements, the pH of the cultures was lowered to 4.0 (EGDm) or 3.5 (10403S) with 3 M HCl. Different pH reductions for each strain were necessary to ensure optimal GABA production for each strain. Extractions were made after 1 h of acid treatment. Non HCltreated cultures were used as negative controls. GABase from Pseudomonas fluorescens (Sigma Aldrich, Steinheim, Germany) was used in the enzymatic assay and increases in OD 340 nm were measured using a Tecan Sunrise absorbance plate reader.

Real-time PCR determination of gadD gene transcription
Real-time PCR was used to determine the relative expression of the genes encoding the GAD system, using the 16S rRNA gene as a reference gene. Cells were grown overnight (16 h) in triplicate (biological replicates) in 20 ml BHI at 37uC. Prior to RNA isolation, 1 ml of culture was mixed with 2 ml RNAlater (R) (Sigma) and incubated at room-temperature for 5 min. Subsequently cells were harvested by centrifugation at 8,0006g for 5 min washed once in 1 ml 1 M Tris buffer (pH 7.4) and disrupted by microwaving for 15 s at 700 W. Pellets were resuspended in 350 ml Buffer RLT (Qiagen) and cell debris removed by centrifuging at 12,0006g for 2 min. Total RNA was obtained from each of the biological replicates using the RNeasy kit (Qiagen, Hilden, Germany), quantified using a NanoDrop (Thermo Scientific, Wilmington, DE), treated with DNase (Turbo DNA-free; Ambion, Austin, TX), and then used to synthesize cDNA. To ensure the absence of DNA contamination in the samples of treated RNA, a PCR reaction was carried out on each RNA sample using 16S primers. To obtain cDNA 15 ml of total RNA (2 mg ml 21 ) was mixed with 1 ml of random primers (3 mg ml 21 ; Invitrogen, Carlsbad, CA), 1 ml of a 10 mM stock of deoxyribonucleotide (dNTP) mix (Invitrogen, Carlsbad, CA). The mixture was heated to 65uC for 5 min, transferred onto ice, and centrifuged briefly before 4 ml of First Strand buffer and 2 ml or 100 mM dithiothreitol (DTT) was added. Following an incubation of 2 min at 25uC, 1 ml of Superscript II (200 U ml 21 ) was added. Subsequently, the mixture was incubated at 25uC for 10 min, at 42uC for 50 min, and finally at 70uC for 15 min before the cDNA was stored at 220uC. Relative quantification of the expression of the genes was carried out by fluorometric real-time PCR using the SYBR green master mix and the Quantitect SYBR green PCR kit (Qiagen, Hilden, Germany). Following the instructions of the kit, 2.5 ml of cDNA was mixed with 5 ml QuantiTect SYBR green PCR master mix (containing fast-start Taq DNA polymerase, SYBR green dye, buffer, and MgCl 2 ), and the appropriate set of primers (Table 1) was added to a final concentration of 0.2 mM each. Each set of primers was designed with specificity toward the genome of L. monocytogenes strain 10403S (accession no. AARZ02000000), producing amplicons in the range of 207 to 268 bp (Table 1). Finally, molecular-biology-grade water was added to the mixture, up to a final volume of 10 ml, and placed in a well of the 96-well microtiter plate (LightCycler 480 multiwell plate 384; Roche Diagnostic GmbH, Manheim, Germany). Once the other PCRs were set up in the other wells, the plate was placed in the LightCycler 480 instrument (Roche Diagnostic GmbH, Manheim, Germany), and the PCR was programmed starting with an initial denaturation at 95uC for 10 min, followed by amplification for 40 cycles at 95uC for 1 s, 5 s at various annealing temperatures, depending on the melting temperature of the set of primers (Table 1), and 72uC for 9 to 11 s. The specificity of amplification for each product was determined by a melting curve analysis at 95uC for 1 s and 65uC for 15 s, followed by a progressive increase of the temperature to 95uC with a ramp rate of 0.11uC s 21 , with continued measurement of fluorescence, and finally cooling of the plate at 40uC for 30 s. Alongside each real-time PCR assay, a control reaction without added cDNA was run as a negative control. Relative expression was calculated as a ratio between expression of target genes (gadD1, gadT1, gadD2, gadT2, and gadD3) and the expression of the 16S rRNA gene, which served as reference gene in each cDNA sample. Calculations were carried out following the ''advanced relative quantification'' settings of the LightCycler 480 software program with PCR efficiency correction.

Animal and infection studies
Eleven week old Balb/c mice were used for infection studies. Animals were housed within the biological services unit, University College Cork and fed 2018S Teklad Global 18% Protein Rodent Diet (Harlan). The EGDm strain used in this study was murinised as previously described [20]. All GAD system mutants were introduced into this strain prior to infection. Overnight cultures of bacteria were grown in BHI medium at 37uC for 16-18 h. Cultures were washed and resuspended in PBS. Mice were inoculated with bacteria to a concentration of 10 10 in 200 ml via direct gastric gavage at 5 mice per strain. Three days post infection, mice were sacrificed and the liver, spleen, mesenteric lymph nodes (MLN) and intestinal faeces were harvested. Liver, spleen and MLN were homogenised in PBS. The weight of the intestinal content was determined and then homogenised in PBS. All samples were serially diluted and plated in duplicate onto either BHI agar (Liver & Spleen) or Listeria Selective agar (LSA; Oxoid) (MLN & intestinal content).

Macrophage survival assay
Permanent stocks of a human derived THP-1 macrophage cell line [21] were maintained at a concentration of 1610 6 in 1 ml Roswell Park Memorial Institute (RPMI) 1640 medium (Sigma Aldrich) supplemented with 10% (v/v) heat-inactivated foetal bovine serum (FBS; Lonza) and 10% (v/v) DMSO and stored at 2 80uC. These were recovered by resuspension in 3 ml RPMI supplemented with 10% (v/v) FBS and 100 U ml 21 penicillin/ streptomycin (PenStrep; Sigma Aldrich) and incubated at 37uC in 5% CO 2 for 2 days. The culture was then centrifuged at 4006g for 5 min and the pellet washed in 5 ml phosphate buffered saline and centrifuged again. The pellet was then resuspended in 5 ml RPMI with 10% FBS and PenStrep and incubated as above for 5-7 days. For survival assays, the macrophages were centrifuged at 4006g for 5 min and washed twice in 5 ml PBS and resuspended in RPMI and 10% FBS supplemented with 0.16 mM phorbol Table 2. Primers used in this study. myristate acetate to a concentration of 1610 5 cells ml 21 . One millilitre was then seeded per well of a 24 well tissue culture plate and the plate was incubated for a further 24 h at 37uC with 5% CO 2 to allow the cells to adhere to the surface of the well. After incubation, the media was removed from each well and 1 ml of PBS was added to wash the cells. The plate was shaken gently and the PBS removed. In parallel, L. monocytogenes cultures were grown to stationary phase in BHI where 1 ml of culture was then centrifuged at 10,0006g for 5 min, washed once in 1 ml PBS and resuspended in 1 ml PBS to give an estimated concentration of 1610 9 cfu ml 21 . Bacterial cells were then diluted to 1610 6 in RPMI and 10% FBS. One millilitre of this bacterial suspension was added to each washed well of THP-1 cells to give a multiplicity of infection (MOI) of 10. Plates were incubated at 37uC with 5% CO 2 for 1 h to allow for invasion. After the 1 h coincubation period, the media was removed from the wells and the cells were washed with 1 ml PBS and then 1 ml of RPMI and 10% FBS supplemented with 30 mg ml 21 Gentamicin was added to kill remaining extracellular bacteria. Plates were returned to incubate for 1 h at 37uC with 5% CO 2 . Bacterial counts were performed by removing the media from each well, washing once in 1 ml PBS and resuspending the cells with 1 ml ice cold sterile dH 2 O for 1 min. Serial dilutions were performed and the suspension was spread plated in duplicate onto BHI agar plates. All agar plates were incubated for 24 h at 37uC before counting.

Statistical analysis of results
Experiments were carried out with three biological replicates and at least two technical replicates for each sample. Significant differences between samples tested were determined by using either a paired Student t test or one-way ANOVA. Results were considered significant when they possessed a P value of ,0.05. Error bars indicating standard deviations from the means are displayed on graphs.

Ethical statement
All animal procedures were approved by the University Animal Experimental Ethics Committee (AEEC) in University College Cork (approval ID 2011/017) and were carried out in a specialized facility. Mice were euthanized by cervical dislocation. Work was carried out under license from the Irish Department of Health.

Deletion of gadD1 or gadD2 increases acid survival in EGDm
To investigate the role of the GAD system in acid tolerance and virulence each of the three decarboxylase genes (gadD1, gadD2, gadD3) were deleted individually from a murinised strain of EGDe (designated EGDm), generating three deletion mutants. The murinisation process as explained and carried out previously [22], involved the targeted mutagenesis of two key amino acids (Ser192Asn and Tyr369Ser) in the InlA protein of EGD-e. Strains lacking a single gadD gene were grown to stationary phase (at which point the GAD system is known to play a crucial role in the survival of acidic conditions [12,23]) and tested for survival at low pH. After 20 min the numbers of surviving wild-type EGDm cells began to rapidly reduce as did those of the strain with a deletion in gadD3. Sixty minutes post treatment both the parental EGDm and EGDm DgadD3 had reduced by over 4.5 log-cycles (Fig. 2a). In contrast the numbers of both EGDm DgadD1 and EGDm DgadD2 strains reduced by only ,2 log-cycles after 60 min. When the survival of mid-exponential phase cells was tested the gadD1 and gadD2 deletions did not confer any survival advantage compared to the parental control (Fig. S1a). Measurements of the stationary phase transcript levels of each gadD gene prior to acid exposure showed that gadD3 was expressed by almost 2 orders of magnitude greater than either gadD1 or gadD2 (Fig. 2c). Deletion of any of the gadD genes did not appear to significantly affect the base levels of any of the remaining two gadD genes although gadD1 in the DgadD3 background was reduced but this difference displayed a p-value of 0.05. Thus the increased stationary phase acid resistance observed in the EGDm DgadD1 and EGDm DgadD2 strains cannot be explained by an increase in the transcription of the remaining gadD genes.

Deletion of gadD2 and gadD3 reduces acid tolerance in 10403S
To determine if the relative contribution of the gadD genes to acid tolerance was conserved in another well-studied reference strain the same gene deletions were introduced into L. monocytogenes 10403S, which belongs to the same lineage (II) as EGD-e, but carries over 30,000 SNPs [24]. Once more the effect of acid treatment was tested on stationary phase cultures for 1 h in BHI. After 60 min wild-type 10403S showed a reduction in cell numbers of over 1.5 log-cycles (Fig. 2b), substantially less killing than was observed in EGDm over this period (Fig. 2a). Deletion of gadD1 did not affect survival at this pH compared to the wild-type. For 10403S a steady decrease in cell numbers was observed over 60 min with a final reduction of about 3 log-cycles, 100-fold greater than the parent strain. The increased sensitivity of the DgadD2 mutant was also observed in exponentially growing cells (Fig. S1b). 10403S DgadD3 did not show a significant reduction in cell numbers compared to 10403S until after 40 min of acid treatment and reached a final reduction of about 3 log-cycles after 60 min (Fig. 2b). A profile of the gadD gene transcript levels prior to acid exposure in 10403S showed that gadD2 is the dominant transcript while gadD1 is transcribed at the lowest levels (Fig. 2d). As in EGDm, deletion of one gadD gene did not significantly affect the expression of the remaining genes apart from gadD2 in the DgadD1 background. Together these results suggest that the two strains have evolved different transcriptional controls over the gadD genes and further indicate that gadD gene transcript levels per se don't determine the intrinsic level of acid tolerance.
EGDm relies on gadD3 for GABA production while 10403S requires gadD2 GABA is the main product of the GAD system and therefore measurement of its production could give insights into how the systems compare in these strains. To this end stationary phase cultures of both strains were challenged to mild acidic pH for one hour and both the extracellular and intracellular GABA (GABA e / GABA i ) production was measured. Previous work has indicated that EGD-e and therefore, EGDm does not produce GABA e while 10403S produces both GABA e and GABA i [14]. As expected, GABA e was not produced here by any EGDm strain in response to low pH (Fig. 3a). After 1 h at pH 4.0 however EGDm produced over 5 mM GABA i . A finding worth documenting was that EGDm produced higher concentrations of GABA i in response to acidification than our previously studied EGD-e strain (data not shown). Since the parent EGD-e strains came from separate laboratories it may be that some mutation had arisen in one of the lines that impacts on the GAD system. Deletion of either gadD1 or gadD2 however did not affect GABA i production compared to the parent strain (Fig. 3b). GABA production in EGDm DgadD3 however was almost abolished, with production just above detectable levels for the assay (0.45 mM; Fig. 3a). In exponential phase no differences in GABA levels were detected between EGDm and the corresponding gadD mutants; indeed GABA levels were at or below the detection limit of the assay in these cultures (Fig. S2). Unlike EGDm, stationary phase cultures of 10403S produced both GABA i (2.72 mM) and GABA e (3.35 mM) after acid treatment. Deletion of gadD1 or gadD3 did not significantly affect the acid-induced GABA production. 10403S DgadD2 did have a significant decrease in both GABA i (0.64 mM; Fig. 3b) and GABA e (0.75 mM; Fig. 3a). A similar effect on GABA production was observed for the DgadD2 mutant (Fig. S2). Thus the gadD gene with the highest transcript levels in each strain also contributed the most to GABA production in that strain; gadD3 for EGDm and gadD2 for 10403S.

Expression of the GAD system is pH-independent
To determine the transcriptional response of EGDm and 10403S to both low pH exposure and GAD system mutation, the expression levels of all three decarboxylase genes was compared to the expression of 16S rRNA in each strain. Overall expression of the gadD genes did not change in EGDm in response to low pH (Fig. 4a). The expression of gadD3 remained higher than both gadD1 and gadD2 in the parent EGDm strain and EGDm DgadD1 and EGDmDgadD2 throughout the course of the acid challenge. In 10403S, the relative expression of gadD2 was significantly greater than the other genes at stationary phase (Fig 2. d). Deletion of any of the gadD genes did not affect the expression of the remaining two in stationary phase. The expression of the gadD genes did not appear to change in response to acid challenge for any of the four 10403S strains (Fig. 4b). s B activity, which was measured indirectly by recording the levels of the known s B -dependent gene lmo2230 (which encodes a putative arsenate reductase) [2], appeared to remain stable in all strains throughout the acid challenge (Fig. 4a bottom  right). Overall it appeared that regardless of pH gadD3 was the Deletion of gadD3 together with either gadD1 or gadD2 reduces acid tolerance seen in single mutants and abolishes GABA production To account for the increase in acid survival seen in both EGDm DgadD1 and EGDm DgadD2 (Fig. 2a) double knockout mutant were constructed and tested for survival to exposure in pH 2.5 in BHI. EGDm DgadD1D3 counts reduced at a faster rate than EGDm and was about 1-log cycle lower than the wt after 60 min (Fig. 5a). The presence of gadD2 on its own did not appear to protect the strain as well as gadD1 alone. Deletion of both gadD2 and gadD3 however followed a pattern of survival similar to the wild-type but reduced compared to the single knockout of gadD1 (Fig. 5a). Overall it appeared that removing the gadD3 gene from EGDm DgadD1 or EGDm DgadD2 negatively impacted on the increased acid survival seen with these strains in stationary phase. In exponential phase the DgadD1D3 mutant was also found to have increased sensitivity to acid compared to either single deletion alone (Fig. S1a), further highlighting the importance of gadD3 to this strain.
GABA i was measured in stationary phase for each strain after exposure to pH 4.0. The mutants carrying double deletions both failed to produce GABA i in response to exposure to the low pH (Fig. 5b). Possession of either gadD1 or gadD2 alone did not bestow an ability to produce GABA. As EGDm is unable to utilise the GADe system no GABA e was produced by any of these strains in response to acid treatment (data not shown).

EGDm gadD mutants survive gastric passage in mice
In order to analyse the role that the GAD i system plays in vivo, survival through a live animal gastric passage was carried out. Three days post oral inoculation of female Balb/C mice with each of the EGDm gad mutants, the animals were sacrificed and dissemination of the strains was analysed. Counts ranged from 6.50610 3 to 3.80610 7 cfu ml 21 and were similar in the liver, spleen and intestinal content for all single deletion strains and the respective wild-type. The counts for EGDm DgadD1 and EGDm DgadD3 however were significantly lower in the mesenteric lymph node (MLN) compared to the wild-type (1-log; Fig. 6). Double deletion of either gadD1 with gadD3 or gadD2 with gadD3 resulted in reduced counts from both the liver and spleens of mice 3 days post infection. These data suggest that the GAD i system plays a role within the host during the development of an infection and is essential for full virulence potential following oral infection of mice.

Growth of L. monocytogenes GAD system mutants in human THP-1 macrophages
In order to cause infection L. monocytogenes must be able to survive inside phagocytic macrophages. The environment that the bacteria encounter inside these cells is reported to be acidic [25] and thus we investigated the role that the GAD system may play in survival. None of the EGDm mutants showed a significant difference in uptake rate by THP-1 cells compared to the wild type after 2 h (Fig. 7a). Furthermore, there was no significant difference seen in the ability of any strain to grow inside the THP-1 cells over a 7 h time-course. EGDm D gadD1 and EGDm DgadD2 did however show a significant lag until 5 h but eventually reached similar numbers to both EGDm and the gadD3 mutant (Fig. 7a). Similarly, 10403S and its isogenic GAD system mutants grew inside THP-1 macrophages with only an apparent lag seen for 10403S DgadD2 after 3 h. Otherwise no significant differences were seen between the wild-type and mutants for the duration of the experiment.

Discussion
Previous studies with L. monocytogenes have shown that strains can vary greatly in terms of their ability to cope with low pH [12,14]. Here the GAD system, one of the major mechanisms for An asterix represents signifcant difference of less than 0.05 between a given mutant and respective wild-type as determined by a student's ttest. doi:10.1371/journal.pone.0112649.g003 acid tolerance was compared between two commonly used reference strains EGDm (murinised EGD-e) and 10403S. The survival of 10403S during exposure to low pH is significantly greater than EGDm (Fig. 2) and this correlates with an ability of 10403S to utilise both GAD i and GAD e in response to low pH. Despite possessing the genes for a putative glutamate/GABA antiporter, EGDm does not export GABA. This may be due to the fact that the amino acid sequences of GadT2 are different at four  10403S (b)). Error bars represent the standard error in the mean of 3 independent biological repeats. Differences found to be significant between the genes at any timepoint for each strain are shown with * Significance was determined where p ,0.05 as determined by a Student's t-test. doi:10.1371/journal.pone.0112649.g004 positions between 10403S and EGDm (V409I, V419I, M438T, I441M, respectively) unlike GadT1 which varies by only a single amino acid (T196K). However, efforts to clone the gadT2 gene from 10403S into EGDm were unsuccessful in inducing GABA e production in EGDm (data not shown). Further studies will be required to test whether expression of gadT2 together with gadD2 from 10403S (i.e. the entire gadT2D2 operon) can induce GABA e production in an EGD-e background.
As EGDm appears to be solely reliant upon intracellular decarboxylation (GAD i ), the three genes encoding glutamate decarboxylases (gadD1, gadD2 and gadD3) were deleted. Somewhat surprisingly, deletion of gadD1 or gadD2 improved the ability of the strain to survive low pH during stationary phase (Fig. 2a), as there was no apparent increase in GABA production, the increased acid resistance may be due to the action of other acid tolerance mechanisms that were not investigated as part of this study. The increase in acid tolerance was comparable to the levels of resistance seen for 10403S, which is one of the most resistant strains of L. monocytogenes [26] (Fig. 2b). None of the single deletions were seen to reduce the bacterium's ability to survive at low pH, showing that each is dispensable for wild-type acid tolerance levels in this strain background. However, deletion of both gadD1 and gadD3 together did prevent the increase in survival seen for EGDm DgadD1 and this double mutant strain, EGDm DgadD1D3, was more acid sensitive than EGDm (Fig. 5a  & Fig. S1a). The deletion of gadD3 along with gadD2 also appeared to prevent the increase in stationary phase acid tolerance seen in the DgadD2 mutant however this strain was not as sensitive as EGDm DgadD1D3 suggesting that gadD1 has a greater role to play in acid tolerance than gadD2 in this strain. Measurements of GABA i across the strains in response to acid, indicated that only EGDm DgadD3 was impaired in an ability to produce GABA i during stationary phase. Taken together with the acid survival data, it would appear that this decarboxylase plays a more important role in the GAD system of this strain compared with the remaining two. Interestingly, a different effect was observed in 104033S after the deletion of the gadD genes. Here deletion of gadD2 negatively impacted on survival during both stationary phase and mid-exponential phase growth (Fig. 2b & Fig. S1b). In 10403S the gadD2 deletion was also accompanied by a reduction in both GABA i and GABA e (Fig. 3 & Fig. S2). It would appear that failure to produce GABA in this strain impacted negatively on its ability to survive at low pH. As gadD2 gadT2 are part of the same operon it is not surprising that both GABA i and GABAe production was affected. The remaining GadT1D1 system apparently could not compensate for the loss of GadT2D2 activity in this strain.
An examination of the transcriptional response for each of the strains in response to acid confirms the differential importance of either gadD2 or gadD3 for 10403S and EGDm, respectively. The gadD2 transcript was the most abundant of the three in 10403S in stationary phase cultures, whereas gadD3 was dominant in the EGDm background (Fig. 4). Overall, neither strain displayed an alteration in gadD gene expression in response to the pH treatment. This may be due to the fact that the cultures have already reached stationary phase and therefore were expressing each gene to a maximal level. Although little is still known about the transcriptional regulation of the GAD system in L. monocytogenes, it is known that both the gadT2D2 operon and gadD3 are at least partially under the control of s B . Using lmo2230 (encoding a putative arsenate reductase) as a reporter of s B activity [2,27], there was clearly no change in s B activity over the course of the acid treatment or as a result of the deletions in the gadD genes. As shown previously [2,27] s B is fully active in stationary phase and additional stress doesn't enhance its activity beyond that level. This might indicate that the cells already possess functional GAD system proteins prior to acid shock and that regulation of the GAD system occurs post transcriptionally. While much of the previous work undertaken on the GAD system in L. monocytogenes has focused on in vitro models and synthetic gastric fluid, the role of the system in vivo has not been addressed previously. In fact there is very little evidence in any bacterial pathogen showing that the GAD system is important for survival in the host. Work with Brucella microti has however shown that deletion of the GAD system reduced counts in mice after oral inoculation [28]. In our study we also found a significant role for the GAD system in the virulence of EGDm in a mouse infection model.
Due to the extensive use of the EGD-e strain as a model intracellular pathogen [29][30][31], we focused upon the role of the GAD i system in an oral infection model that employs murinised EGDm strain to enhance the progression of invasive disease via the GI tract [20]. A previous study identifying genes expressed by L. monocytogenes in response to Lactobacilli within the mouse gut showed that all three gene systems (gadD1T1, gadD2T2 and gadD3) had significantly increased expression in this environment [32]. This suggests that the GAD system is actively induced within the host GI tract. In our functional study single gadD deletion mutants were not affected in their ability to infect the liver and spleen of mice following oral infection. However the two double mutants displayed significantly reduced levels of infection, indicating that a complete GAD system is required for full infection. This correlates with in vitro survival assays with these mutants, particularly EGDm DgadD1D3, which displays a greater sensitivity to low pH than the single mutants and exhibits a significant virulence defect in our model system. It is important to note that the pH in the stomach of Balb/C mice can be as high as pH 4.04 [33] and the gastric acidity of this particular animal host may have implications for the survival of these gad mutant strains. However, the GAD system is likely to play a more important role in hosts that possess stomachs with higher acidity. It will also be important to investigate the role of the GAD e system in virulence by using a L. monocytogenes strain capable of producing GABA e .
Subsequently, we investigated another stage in the virulence process of L. monocytogenes; the intracellular cycle. In both EGDm and 10403S, the GAD system does not appear to play a major role in survival within human derived macrophages. This was largely expected since the pH inside bone marrow derived macrophages after compartmentalisation of L. monocytogenes is about pH 5.5 [25] while induction of the GAD system in L. monocytogenes normally occurs below pH 4.5 [17]. Interestingly, although GadD1T1 is known to play a role in growth at mild acidic conditions [16] similar to those occurring within macrophages no major effect of these proteins was documented within macrophages.
Overall, the GAD system in L. monocytogenes appears to show a clear divergence in functionality between these two well-studied strains. The strains appear to have adopted the use of either a GAD i system or a combined GAD i /GADe system, which likely reflects their unique evolutionary histories. The strain which possessed both GAD i and GAD e , 10403S, displayed a more acid resistant phenotype suggesting that a functional antiport is highly beneficial. In contrast, the strain which utilised only GAD i did not appear to have a sole reliance on any of its three isoforms of decarboxylases for acid tolerance, indicating a more robust system. From comparing these two strains it is clear that functional divergence of physiologically important pathways can readily occur and further highlights the importance of inter-strain comparisons in addressing the biological significance of any pathway. Figure 6. Infection of Balb/C mice with EGDm GAD system mutants. Plate counts of surviving EGDm GAD system mutants 3 days post infection from female Balb/C mice (n = 5). Isolated from the liver, spleen, mesenteric lymph node (MLN) and faeces. Significant differences (*) between wild-type and mutants were determined using one-way ANOVA. doi:10.1371/journal.pone.0112649.g006 Figure 7. Growth of GAD system mutants in THP-1 macrophages. Growth of EGDm (a) and 10403S (b) GAD system mutants inside THP-1 macrophages over 7 h. Counts are recorded 2 h post coincubation of THP-1 with bacteria at an MOI of 10 (10 6 bacteria; black arrow). Error bars represent the standard deviation from the mean of at least 4 biological replicates for each strain and time-point. Significant differences (*; p ,0.05) were determined using one-way ANOVA. doi:10.1371/journal.pone.0112649.g007 Diversity in Functionality of the GAD System in Listeria monocytogenes Supporting Information Figure S1 Acid survival of L. monocytogenes gad mutants. Mid-log phase EGDm (a) and 10403S (b) Dgad mutants were challenged at pH 3.0. Cell counts were taken every 20 min. Values are the means of data from three individual cultures, with the cell counts for each culture being the means of counts from three platings. Error bars represent the standard error from the mean value of three individual biological repeats. (TIF) Figure S2 GABA production from L. monocytogenes gad mutants. (a) Production of GABA e by EGDm and 10403S gad mutants with (grey) or without (black) 1 h exposure to acid at pH 4.0 (EGDm) or pH 3.5 (10403S). (B) Production of GABA i by EGDm and 10403S gadD mutants with (grey) or without (black) 1 h exposure to acid at pH 4.0 (EGDm) or pH 3.5 (10403S). Dashed horizontal lines indicate the detection limits for GABA in each experiment. Error bars represent the standard deviation from the mean of three individual biological repeats for each sample. (TIF)