CO₂- and Anaerobiosis-Induced Changes in Physiology and Gene Expression of Different Listeria monocytogenes Strains

Abstract

Although carbon dioxide (CO₂) is known to inhibit growth of most bacteria, very little is known about the cellular response. The food-borne pathogen Listeria monocytogenes is characterized by its ability to grow in high CO₂ concentrations at refrigeration temperatures. We examined the listerial responses of different strains to growth in air, 100% N₂, and 100% CO₂. The CO₂-induced changes in membrane lipid fatty acid composition and expression of selected genes were strain dependent. The acid-tolerant L. monocytogenes LO28 responded in the same manner to CO₂ as to other anaerobic, slightly acidic environments (100% N₂, pH 5.7). An increase in the expression of the genes encoding glutamate decarboxylase (essential for survival in strong acid) as well as an increased amount of branched-chain fatty acids in the membrane was observed in both atmospheres. In contrast, the acid-sensitive L. monocytogenes strain EGD responded differently to CO₂ and N₂ at the same pH. In a separate experiment with L. monocytogenes 412, an increased isocitrate dehydrogenase activity level was observed for cells grown in CO₂-containing atmospheres. Together, our findings demonstrate that the CO₂-response is a partly strain-dependent complex mechanism. The possible links between the CO₂-dependent changes in isocitrate dehydrogenase activity, glutamate metabolism and branched fatty acid biosynthesis are discussed.

Modified atmosphere packaging (MAP) of foods is used to improve shelf life and safety by inhibition of spoilage organisms and pathogens. The gases used in MAP are most often combinations of oxygen (O₂), nitrogen (N₂), and carbon dioxide (CO₂). In most cases, the bacteriostatic effect is obtained by a combination of decreased O₂ and increased CO₂ concentrations (12). CO₂ is colorless, odorless, noncombustible, and bacteriostatic (21). It is both water and lipid soluble (12), and it does not leave toxic residues on MAP-stored foods (19).

In food preservation MAP is frequently used for fresh muscle foods and ready-to-use vegetables stored and displayed at chill temperatures (12). The solubility of CO₂ increases at low temperatures (29), and it has recently been demonstrated that bacterial growth inhibition is proportional to the concentration of CO₂ dissolved in the water phase of a food matrix (8). The solubility of CO₂ is increased at neutral pH, high water activity, high buffer capacity, and high headspace-to-sample ratio (29).

Listeria monocytogenes is a food-borne pathogen that can cause the severe human infection listeriosis (2). Cases of listeriosis caused by the consumption of contaminated meat or meat products have been reported since the 1990s (13), and several national prevention programs have been initiated (15). The organism is able to grow in the presence of CO₂ at low temperatures (2). CO₂ concentrations above 70% (N₂ balance gas) are required to inhibit growth of L. monocytogenes (11, 17, 20, 27, 43) at chill temperatures (<7°C), but even at such high CO₂ concentrations, the presence of just 5% O₂ may allow growth (43). This extent of adaptability makes L. monocytogenes a good candidate in which to study the CO₂ response.

Despite the many observations on CO₂-induced growth inhibition little is known of the mechanisms involved. It has been hypothesized that CO₂ inhibits growth of bacteria by (i) changing the conformation of cellular enzymes and thereby decreasing the rate of the metabolic reactions (36); (ii) CO₂ product repression of carboxylases and decarboxylases (25); (iii) disrupting cell membrane structural integrity and/or specific functions (40); (iv) decreasing intracellular pH, which demands that ATP is used for reestablishing pH_i rather than for growing (44); or (v) some combination of these mechanisms (9). It was reported recently that glutamate decarboxylase is essential for survival of L. monocytogenes in strong acid (6). If CO₂ growth-inhibition is caused by acidification of the environment, the cells will most likely use this mechanism to reestablish pH_i and maintain growth. As it seems probable that bacteria have a multifaceted response towards CO₂, we conducted several parallel experiments to examine the atmosphere-induced responses in acid-sensitive and acid-tolerant L. monocytogenes strains. The purpose of the present study was to compare the changes in metabolism, membrane lipid fatty acyl (FA) composition, and expression of specific genes in air, 100% N₂, and 100% CO₂.

MATERIALS AND METHODS

Bacterial strains.

L. monocytogenes 412 (obtained from the Danish Meat Research Institute, Roskilde, Denmark) was isolated from bacon. L. monocytogenes LO28, a clinical strain (isolated by P. Cossart, Institute Pasteur, Paris, France) was received from J. E. Olsen, The Royal Veterinary and Agricultural University, Frederiksberg, Denmark. L. monocytogenes EGD was obtained from G. B. Mackaness, Saranac Lake, N.Y. (24), also via J. E. Olsen. The genome of a variant of the latter strain (L. monocytogenes EGDe) was recently sequenced and published (14). Our preliminary experiments had shown L. monocytogenes EGD to be more CO₂ sensitive than L. monocytogenes 412 and LO28 (23).

Culture conditions.

Preinoculum was prepared from a frozen stock (−80°C) using a standard procedure with incubation of cells in 10 ml of brain heart infusion (BHI) (Merck, Darmstadt, Germany) at 10°C for 7 days. BHI was used as medium for growth experiments, while a defined medium (IMM [35]) was used for metabolic studies. Cultures were enumerated on tryptone soy agar plates (Oxoid, Basingstoke, United Kingdom). Listeria selective base (Oxford Formula; Oxoid) was used for tentative verification of Listeria.

Examination of CO₂-induced changes in primary metabolism.

L. monocytogenes 412 and EGD were grown in IMM at 10°C from an initial cell density of 10⁵ CFU/ml. Each strain was incubated aerobically in conical flasks shaken at 200 rpm or statically (anaerobically) in jars (Oxoid) filled with 100% CO₂ or 100% N₂. Duplicate experiments were performed. The jars were evacuated and filled with gas three times to ensure that the atmospheres were O₂ free. Samples were harvested once in the lag phase, twice in the exponential growth phase, and three times in the stationary phase. The optical density at 600 nm (OD₆₀₀), pH, and the number of CFU were determined at the time of harvest. A sample volume of 1.5 ml was filter sterilized and stored at −20°C until analysis. Metabolites were identified by high-performance liquid chromatography as described by Holm Hansen et al. (18).

Examination of CO₂-induced changes in isocitrate dehydrogenase (IDH) activity. (i) Growth conditions.

L. monocytogenes 412 was grown in BHI from an initial cell density of 10⁵ CFU/ml at 10°C in six anaerobic atmospheres (100% CO₂, 80% CO₂, 60% CO₂, 40% CO₂, 20% CO₂, or 0% CO₂, with N₂ as the balance gas) and three aerobic atmospheres (air, 79% CO₂-21% O₂, or 79% CO₂-2% O₂, with N₂ as the balance gas). Three test tubes, each containing 10 ml of inoculum, were placed in each of eight anaerobic jars, and three test tubes were incubated in air. Each jar was evacuated three times and then filled with one of the eight atmospheres using CO₂, O₂, and N₂ gas (Hede Nielsen, Horsens, Denmark) and a gas mixer (MAP MIX 9000; PBI Dansensor A/S, Ringsted, Denmark). All samples were incubated at 10°C. The IDH activity level, total cell protein, OD₆₀₀, CFU, medium pH, and cell morphology were examined at the time of sampling.

(ii) Preparation of enzyme.

Late-exponential-phase cultures were diluted to an OD₆₀₀ of 0.250 in a cuvette, transferred to Eppendorf tubes, and centrifuged (17,320 × g, 5 min). The bacterial pellet was washed twice in 1 ml of 0.005 M sodium phosphate buffer, and this was followed by centrifugation (17,320 × g, 5 min). The pellet was resuspended in 0.5 ml of 0.005 M sodium phosphate buffer, and 0.3 g of glass beads (diameter, 106 μm; acid washed; catalog no. G-4649; Sigma Aldrich, Vallensbak Strand, Denmark) was added. The bacterial cells were homogenized with four bursts of 45 s using a Fastprep FP120 (KemEnTech, Copenhagen, Denmark). Between the homogenizations, the cells were placed on ice. The cell lysate was centrifuged (17,320 × g, 10 min) and the supernatant containing the enzyme was transferred immediately to a fresh Eppendorf tube.

Determination of enzyme activity.

IDH catalyzes the conversion of isocitrate + NADP⁺ (Mn²⁺ excess) to α-ketoglutarate + CO₂ + NADPH. The activity level was measured using an IDH kit (Sigma Aldrich) and total cell protein was determined using a protein assay (Bio-Rad, Hercules, Calif.). The enzyme activity was calculated as the conversion of millimoles of NADP⁺ to NADPH per minute per milligram of total cell protein. Purified IDH (Sigma Aldrich) served as the positive control. Two independent experiments with three replications were conducted.

Examination of CO₂-induced changes in the membrane.

From standard preparations of preinoculum, L. monocytogenes LO28 and EGD at 10⁵ CFU/ml, respectively, were inoculated in BHI or BHI with a pH of 5.7 (adjusted with 2 M HCl). The cultures were incubated in 100% CO₂, air, or 100% N₂ (a total of four combinations: 100% CO₂, air, 100% N₂, and 100% N_{2 pH 5.7}) at 10°C until the late exponential growth phase. Since preliminary experiments showed that constant flushing of CO₂ over BHI lowered the pH of the medium to 5.7, the acidic-anaerobic condition (N_{2 pH 5.7}) was used to mimic the environment created by CO₂.

Bacteria were harvested from 1.5-liter cultures by centrifugation (8,000 × g, 15 min, 4°C). The pellet was washed, centrifuged, resuspended in 16 ml of chloroform-methanol-0.1 N HCl (1:2:0.8, by vol), and 3 g of glass beads (0.10 to 0.15 mm diameter; acid-washed; Sigma Aldrich) was added. The mixture was placed on ice for 60 min, during which time it was mixed frequently. Chloroform and 0.1 N HCl was added to obtain a chloroform-methanol-0.1 N HCl ratio of 1:1:1 (by vol). The suspension was vortexed and centrifuged (2,000 × g, 3 min), and the chloroform phase (lower liquid phase) was transferred to a new tube. Six milliliters of chloroform was added to the original tube. The lower chloroform phase of this second extraction was transferred to the tube containing the chloroform phase from the first extraction. The lipids were dried under nitrogen, resuspended in chloroform-methanol (9:1, vol/vol), dried under nitrogen, and stored at −20°C in a small volume of the solvent mixture. The lipid FA composition was analyzed using capillary gas chromatography as described by Ntougias and Russell (33).

Examination of CO₂-induced changes in gene expression.

L. monocytogenes LO28 and EGD were grown as described for the analysis of membrane composition. At the time of sampling, 16 ml of culture was harvested by centrifugation (14,300 × g, 2 min, 4°C). The tubes were quickly placed on ice, the supernatant was discarded, and the pellet was immediately frozen in liquid nitrogen. The tubes were stored at −80°C. Total RNA for gene expression analyses was extracted from the cultures according to the procedure described by Gravesen et al. (16).

The effects of atmosphere on expression of genes involved in acid stress response (gadA, gadB, and gadC) and virulence (prfA, actA, hly, mpl, plcB, inlA, inlB, inlC, inlC2, inlD, inlE, inlF, and lmaA) were compared for L. monocytogenes LO28 and EGD, respectively, using a DNA microarray. Microarray fabrication, preparation of in vitro transcribed control RNA, preparation of labeled cDNA, microarray hybridization, data analysis, and statistical analysis were all performed as described by Gravesen et al. (16).

Data analysis.

Using SAS software (SAS Institute, Cary, N.C.) linear model analyses were conducted to determine significant effects of atmosphere. In addition, t tests were carried out to determine statistically significant grouping in the data material.

RESULTS

Production of metabolites.

For both strains, the major end product differed between aerobic and anaerobic cultures, but no qualitative effect of CO₂ per se was observed (Fig. 1). The increases in the number of CFU as a function of time were similar for the two strains in the different atmospheres, respectively, whereas the OD was higher for L. monocytogenes 412 than L. monocytogenes EGD. Acetic acid and lactic acid were the major products in aerobic cultures, and acetoin was only produced aerobically. Lactic acid was the major product in anaerobic cultures, and formic acid was only produced anaerobically. There was no significant difference in ethanol production for the different atmospheres.

FIG. 1.

Growth (•); consumption of glucose (○); and production of lactic acid □, acetic acid (▵), acetoin (×), formic acid (◊), and ethanol (+) as a function of time for L. monocytogenes 412 (a) and L. monocytogenes EGD (b) grown in IMM (35) at 10°C in three atmospheres.

Effect of CO₂ on IDH.

The IDH activity level varied from 0.20 to 0.45 U/mg protein for L. monocytogenes 412 grown in BHI at 10°C under nine different atmospheres (Fig. 2). The activity level was higher when CO₂ was present. It was significantly higher (P < 0.001) in 79% CO₂-21% O₂ and 79% CO₂-2% O₂ than in air. When the IDH activities were compared for the anaerobic atmospheres, the level was significantly higher (P < 0.001) in 20% CO₂-80% N₂ than in the 100% N₂ control. The activity level was also higher in the anaerobic atmospheres containing 40 to 100% CO₂ than in the 100% N₂ control, but this difference was not statistically significant. If the determined activity levels in all the CO₂-containing atmospheres were pooled (40 to 100% CO₂) this new average based on 30 determinations was significantly higher (P < 0.027) than in the 100% N₂ control (based on six determinations). For both aerobic and anaerobic atmospheres, the differences in the activity levels could not be explained by differences in OD, number of CFU per milliliter, pH, or morphology (results not shown).

FIG. 2.

IDH activities in L. monocytogenes 412 grown in aerobic CO₂-containing atmospheres (light bars) or anaerobic CO₂-containing atmospheres (dark bars) in BHI at 10°C. Each bar represents the average of six determinations.

CO₂-induced changes in the membrane.

The lipid FA composition of L. monocytogenes LO28 grown in air consisted mainly (87.6%) of a15:0, a17:0, i15:0, and n16:0 (Table 1). In CO₂ and N_{2 pH 5.7}, the content of branched-chain FAs decreased whereas the content of unsaturated FAs increased. The FA composition of L. monocytogenes EGD grown in air consisted mainly (73.8%) of a15:0, a17:0, n18:0, n16:0, and i15:0. In contrast to strain LO28, strain EGD responded differently to CO₂ and N_{2 pH 5.7}. In CO₂, the ratio of anteiso-branched to iso-branched FAs increased slightly together with the content of unsaturated FAs compared to growth in air. In N_{2 pH 5.7}, the content of unsaturated FAs decreased slightly, and the content of branched-chain FAs decreased notably, while the ratio of anteiso-branched to iso-branched FAs remained unchanged.

TABLE 1.

Membrane FA composition of two L. monocytogenes strains grown under various conditions^a

Fatty acid	% Composition after growth of strain in:
	L. monocytogenes LO28				L. monocytogenes EGD
	Air	N2	N2 pH 5.7	CO2	Air	N2	N2 pH 5.7	CO2
n12:0	0.2	1.1	1.2	1.6	0.7	0.5	1.5	1.4
n13:0	0.1	0.0	0.4	0.3	0.2	0.4	0.0	1.2
n14:0	2.1	2.5	3.9	3.9	2.1	1.6	3.9	3.9
n15:0	0.1	0.6	0.1	0.4	0.5	0.1	1.1	0.4
n16:0	6.8	10.1	13.7	13.2	7.6	3.9	10.9	6.5
n18:0	1.9	8.3	6.1	8.8	9.0	4.2	7.8	5.0
n20:0	0.2	0.0	0.9	0.0	0.4	0.0	0.0	0.0
n13:1	1.4	1.5	3.1	2.1	1.6	0.1	0.5	3.2
n16:1	1.1	1.0	1.7	1.7	0.4	1.4	2.0	0.7
n17:1	0.4	1.1	1.9	2.0	1.1	0.9	0.0	0.7
n18:1	0.5	1.7	3.1	2.7	1.4	0.7	1.2	1.3
i14	1.5	1.0	1.8	1.9	1.0	1.2	1.6	0.8
i15	0.8	8.0	4.9	5.6	7.6	8.1	5.6	7.0
a15	56.7	43.8	36.6	31.3	47.6	57.3	43.7	54.0
i16	0.8	1.6	2.0	1.6	1.1	0.6	1.0	0.6
i17	0.8	1.3	2.2	1.7	0.8	0.5	0.6	0.7
a17	13.3	8.0	8.3	10.4	11.0	14.8	11.9	5.7
i18	0.4	5.0	2.9	4.1	3.2	2.3	4.1	3.6
a19	0.4	3.0	2.7	3.6	1.8	1.0	0.3	1.6
Others	0.6	0.4	2.5	3.2	0.9	0.5	2.2	1.7
Total	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0
ECL	15.3	15.7	15.4	15.4	15.6	15.5	15.4	15.2
% Unsaturated FA	3.4	5.2	9.9	8.4	4.4	3.2	3.7	5.8
% Branched FA	84.7	71.7	61.4	60.1	74.1	85.7	68.9	74.0
a/ib	4.9	3.2	3.4	3.0	4.4	5.7	4.3	4.8

^aStrains were grown in BHI at 10°C in 100% CO₂. Controls were also carried out under aerobic and anaerobic conditions.

^bRatio of anteiso- and iso-branched fatty acids.

CO₂-induced changes in gene expression.

Expression levels of genes are shown as average ratios of the CO₂ versus air and CO₂ versus N_{2 pH 5.7} (Table 2). The three genes coding for the glutamate decarboxylase A, B, and C subunits (gadA, gadB, and gadC) were all, compared to growth in air, significantly (P < 0.00001) overexpressed—by 4.4-, 8.9-, and 10.6-fold, respectively—in L. monocytogenes LO28 grown in CO₂. The levels of expression of these genes, respectively, were similar in CO₂ and N_{2 pH 5.7}. The gadB gene was expressed weakly in L. monocytogenes EGD in all atmospheres, whereas the expression of gadA and gadC was undetectable.

TABLE 2.

Expression of genes associated with acid sensitivity and virulence in L. monocytogenes grown in BHI at 10°C to the late exponential growth phase in different atmospheres^a

Gene	Mean expression ratio ± SD (P) for strain:
	LO28		EGD
	CO2/air	CO2/N2 pH 5.7	CO2/air	CO2/N2 pH 5.7
Positive control	0.98 ± 0.10	0.95 ± 0.12	0.99 ± 0.13	1.01 ± 0.13
Acid response
gadA	4.4 ± 2.1 (9.3 × 10−6)	1.02 ± 0.54 (0.65)	NDb	ND
gadB	8.9 ± 4.1 (1.3 × 10−6)	1.07 ± 0.19 (0.04)	0.90 ± 0.61 (0.71)	1.04 ± 0.42 (0.87)
gadC	10.6 ± 4.6 (5.7 × 10−7)	1.04 ± 0.29 (0.25)	ND	ND
Virulence
inlA	9.4 ± 6.2 (6.4 × 10−4)	0.77 ± 0.54 (0.28)	4.0 ± 4.9 (0.03)	0.91 ± 0.74 (0.59)
lmaA	ND	ND	ND	0.02 ± 0.01 (7.3 × 10−30)

^aRatios and P values in boldface type represent genes with statistically significant difference in expression (P < 0.0008, corresponding to less than 5% probability of false-positive regulation with a Bonferroni correction [see Materials and Methods]).

^bND, not detectable.

Of the 13 virulence-associated genes examined, the expression was only altered significantly in two, namely, inlA and lmaA. The expression of the virulence factor inlA was significantly higher in CO₂ than in air for L. monocytogenes LO28. In strain EGD a nonsignificant increase was observed. For both strains, the levels of expression of inlA were similar in CO₂ and N_{2 pH 5.7}. In strain EGD, the expression of lmaA was significantly lower in CO₂ than in N_{2 pH 5.7}.

DISCUSSION

Primary metabolism.

CO₂-induced changes in metabolism have not previously been investigated for L. monocytogenes. We observed no differences in the metabolites produced in CO₂ or N₂, but growth and metabolite production was faster in N₂ than in CO₂. Lack of acetoin production in the anaerobic cultures showed that exposure to air during sampling was negligible. The differences between metabolites from aerobically and anaerobically (N₂) grown cultures are in agreement with the findings of Romick et al. (37). In the case of the related, nonpathogenic food spoilage organism Brochothrix thermosphacta, the metabolites produced from glucose in N₂ and CO₂ are also similar and differ from those produced in air (3).

Enzyme activity.

The IDH enzyme was chosen for several reasons: (i) it catalyzes a CO₂-producing reaction and is therefore assumed to be affected by CO₂ (22, 25), (ii) it only catalyzes one reaction, (iii) the enzyme was known to exist in the organism (14, 41), and (iv) an IDH kit was commercially available. The slight but significant increase in IDH activity in the presence of CO₂ and oxygen is in contrast to the findings in the gram-negative Pseudomonas aeruginosa (25) but in accordance with those in the yeast Sclerotium rolfsii (28). Although the observed differences were not statistically significant for the anaerobic cultures, the observed changes were consistent in replicate experiments (i.e., the lack of statistical significance is not the result of some positive and some negative differences), and they are therefore regarded as being trends worth noting. Based on the examination of this one enzyme which catalyzes a CO₂-producing reaction, our results cannot necessarily be extrapolated to carboxylases and decarboxylases in general (22, 25). The results can, however, possibly be linked to the observed changes in the gene expression and cell membrane as suggested at the end of this discussion.

Membrane changes.

Comparing aerobic growth with growth in CO₂, L. monocytogenes LO28 responded to in the latter by producing a membrane lipid FA composition containing less branched-chain and more straight-chain components. This should render the membrane less fluid; however, there are also changes in minor components, particularly the unsaturated fatty acids, which more than double in their proportion, and no overall change in the FA equivalent chain length (ECL). It is not possible to say whether the reduction in proportion of anteiso-branched relative to iso-branched fatty acids (fluidity decreasing change) would counteract the increase in unsaturated fatty acids (fluidity increasing change), without biophysical measurements of fluidity parameters. Comparison of the data for growth in air or CO₂ with the controls in N₂ (neutral pH) or N_{2 pH 5.7} shows that the fatty acid changes are likely to be largely but not entirely due to the reduction in pH caused by CO₂ rather than the CO₂ per se. This observation was confirmed by growing cultures aerobically in air with the pH adjusted to 5.7, when similar fatty acid compositional changes were obtained to those in CO₂ or in N₂ at pH 5.7 (data not shown).

Like Nilsson et al. (32), we also observed an increase in the shorter-chain fatty acids 12:0 and 14:0 during growth in CO₂, and our results indicate that this is a response to acidification of the culture medium. The increases we observed were much smaller than those of Nilsson et al., but it is difficult to make further comparisons since the method employed in our study allows us to identify a larger percentage (>97%) of the fatty acids. It is well documented that a15:0, i15:0, and a17:0 are the major fatty acids in the Scott A strain and other strains of Listeria (31, 42) and that a15:0 has a particular role in growth at low temperature (1, 31) and regulation of membrane fluidity (10). Our data for the relative percentages of the major anteiso- and iso-branched 15:0 and 17:0 fatty acids in strain LO28 are comparable to those found by Nichols et al. (31) in the Scott A strain.

Although the EGD strain of L. monocytogenes had a very similar membrane lipid FA composition to that of L. monocytogenes LO28, the pattern in variation differed. There were no significant differences (P > 0.05) in the proportion of total branched-chain or unsaturated fatty acids, their mean chain length (ECL) (Table 1), or the relative proportions of anteiso- and iso-branched fatty acids. However, the response to CO₂ of the major fatty acid, a15:0, was an increase in contrast to L. monocytogenes LO28, where the relative amount decreased. Variations in fatty acid composition of different strains of Listeria and their response to stress factors such as temperature and salt have been noted by other researchers (38).

In summary, our results on CO₂-induced changes in membrane composition indicate that these are associated with the acidification created by the gas. We have not found a clear link between the action of CO₂ and changes in membrane fluidity. If such a connection exists it has to be detected by direct biophysical measurements.

Gene expression.

The glutamate decarboxylase complex is essential for survival of L. monocytogenes in strong acid (6), and the expression of gadA, gadB and gadC in strains LO28 and EGD, respectively, grown in CO₂ is in accordance with the recently reported natural gad content of the two strains (6). The similarity in response towards CO₂ and N_{2 pH 5.7} showed that L. monocytogenes LO28 responded to CO₂ as it would to acid using the gad system. Connections between acid tolerance and glutamate decarboxylase have previously been discussed for Clostridium perfringens (7) and shown for Escherichia coli (4, 5, 45) and Lactococcus lactis (39). The lack of gadA and gadC expression in L. monocytogenes EGD indicates that this strain maintains growth in CO₂ in another and less efficient manner. The fact that there was no difference in end product formation or the expression of gad in cultures grown in CO₂ or N₂, indicates overlapping between the CO₂ response and the response to acidity and anaerobiosis, respectively.

Increased virulence has been reported for an acid-insensitive strain of L. monocytogenes (34), and diminished virulence has been reported for a very acid-sensitive mutant derived from L. monocytogenes LO28 (30). There is, however, no information on the pathogenicity of CO₂-adapted cells. We have shown that the individual CO₂-stressed cell may have a higher expression of internalin A, while the expression of other virulence-associated genes were unaffected. Since a recent study by King et al. (26) indicated that CO₂ stress causes a decrease in survival in vitro in bile and acid systems, it seems unlikely that CO₂ will increase the virulence of the cell.

Hypothesis for the CO₂ responses in acid-tolerant strains.

A strain-dependent CO₂ response was observed. In the acid-tolerant strains LO28 (6) and strain 412 (A. Gravesen and S. Knøchel, personal communication), it was possible to link some of the changes in order to suggest how growth is maintained in CO₂. The acid-tolerant strain LO28 presumably used glutamate decarboxylase (glutamate→γ-amino butyric acid [GABA]) to reestablish pH_i (overexpression of gadB) in CO₂. It is currently unknown if glutamate is supplied similarly in different strains. We have evidence of increased glutamate import in strain LO28 (overexpression of gadC) and indications of increased glutamate production in strain 412 since the activity level of IDH (isocitrate→α-ketoglutarate) was increased. If the flux through isocitrate→α-ketoglutarate and glutamate→GABA increases, it seems logical to presume that the flux through the whole isocitrate→α-ketoglutarate→glutamate→GABA pathway increases. The production of branched-chain fatty acids in the membrane is linked to the turnover of α-ketoglutarate→glutamate by glutamate dehydrogenase and we would expect them to alter in concert. However, the amount of branched-chain fatty acids decreased, indicating that the glutamate dehydrogenase activity was low. The situation seems therefore more complex. The cell can use other sources for the supply of glutamate, and if glutamate is partly supplied via production, it may happen through alternative pathways, e.g., isocitrate→α-ketoglutarate→2-oxoglutaramate→l-glutamine→glutamate. The enzyme glutamine synthetase, which catalyzes l-glutamine→glutamate, is present in L. monocytogenes (www.genome.ad.jp).

Strain EGD has previously been observed to respond differently to CO₂ (23) than do the strains with high acid tolerance. No changes were observed in the amount of branched-chain fatty acids and the expression of genes encoding GAD in the acid-sensitive L. monocytogenes EGD. This shows that L. monocytogenes EGD maintains growth in CO₂ by other means, e.g., a more general stress response that includes filamentation. The genome of L. monocytogenes EGDe indicates the presence of gadB′ and gadC′ homologues, but these have not been investigated specifically and their potential involvement is currently not known.

In conclusion, our investigations have shown that it is necessary to investigate more strains in order to characterize the CO₂ response in L. monocytogenes since it is to some extent strain dependent. The acid-tolerant strain LO28 showed the same response to CO₂ and CO₂-mimicking environments, whereas the acid-sensitive strain EGD showed different responses to these two atmospheres. A genome-wide analysis would be a significant step towards understanding the diversity of the response.