Carbon dioxide increases acid resistance in Escherichia coli


Dr Hiroshi Kobayashi, Graduate School of Pharmaceutical Sciences, Chiba University, 1-8-1, Inohana, Chuo-ku, Chiba 260-8675, Japan


Aims:  To investigate how carbon dioxide affects the acid resistance of Escherichia coli.

Methods and Results: Escherichia coli W3110 was grown in minimal EG medium at pH 7·5, and cells were adapted at pH 5·5 at 37°C with and without supply of carbon dioxide and nitrogen gases. The number of colonies grown on LB medium was measured after cells were challenged in minimal EG medium of pH 2·5 at 37°C under various conditions. When carbon dioxide was supplied at both the acid adaptation and challenge stages, 94% of cells survived after the acid challenge for 1 h, while the survival rates were 50 and 67% when nitrogen gas and glutamate were supplied respectively. After the acid challenge for 3 h, the survival rate observed with the carbon dioxide gas supply was again 2·5-fold higher than those with the nitrogen gas supply.

Conclusion:  Carbon dioxide was shown to participate in the maintenance of high viability under acidic conditions.

Significance and Impact of the Study:  This study provides useful information for research into bacterial pathogenesis, fermentation and food preservation.


To infect animals and cause disease, it is important that bacteria overcome acidic stress to survive in the gastric stomach, but this mechanism has not yet been completely clarified. It is known that environmental factors such as nutrient level, temperature and pH affect acid resistance (AR) in Escherichia coli. In addition, some amino acid decarboxylases such as arginine, glutamate, lysine, histidine and ornithine, also participate in the AR of E. coli (Gale 1946; Lin et al. 1995, 1996; Guilfoyle and Hirshfield 1996). The glutamate-dependent system was the predominant AR system in stationary phase cells (Castanie-Cornet and Foster 2001). It was suggested that the arginine- and lysine-dependent systems were induced more strongly in cells grown in an aerobic medium at acidic pH than anaerobic-grown cells (Diez-Gonzalez and Karaibrahimoglu 2004). The global regulator RpoS-dependent AR system was shown to be present in E. coli and Salmonella enterica serovar Typhimurium (Hengge-Aronis 1993; Castanie-Cornet and Foster 2001; Diez-Gonzalez and Karaibrahimoglu 2004).

When E. coli was adapted at pH 5·5, such adapted cells were more resistant to acid and the adaptation treatment was suggested to cause transient synthesis of some key acid shock proteins against extreme acidic pH (Foster 1991, 1993).

In this study, we found that CO2, a product of amino acid decarboxylation, greatly increased the AR of E. coli and the CO2 effect was more significant at the adaptation than the challenge stage.

Materials and methods

Strain and culture media

Escherichia coli wild-type strain W3110 was grown at 37°C in minimal EG medium containing 0·5% glucose (Vogel and Bonner 1956). The pH values of the EG medium were adjusted at pH 2·0, 5·5 and 7·5 with HCl and KOH. When cells were grown without shaking, a tightly capped tube filled with the medium was used. When CO2 and N2 were supplied, the culture tubes were connected with a rubber balloon containing CO2 or N2 gas.

Growth conditions and AR assay

Growth of E. coli was monitored by measuring the absorbance of the medium at 600 nm and the cell density at an absorbance of 0·3 was c. 8 × 107 CFU ml−1. Escherichia coli cells grown overnight at 37°C in EG medium (pH 7·5) with shaking (240 rev min−1) were diluted at 1 : 1000 with EG medium (pH 7·5), and grown to an absorbance of 0·3. For adaptation, the cultured cells were suspended in a twofold volume of EG medium pH 5·5 and grown without shaking under conditions with the supply of CO2 or N2. When stationary phase cells were adapted directly at pH 5·5, they were suspended in EG medium of pH 5·5 at a cell density of 4 × 107 CFU ml−1. A balloon containing c. 2 l of gas was connected to the culture tube containing 5 ml of EG medium. When 2 l of N2 gas were pumped into the balloon, the diameter of the balloon extended to 15 cm. The gas volume was estimated using this result. The cell density increased to 8 × 107 CFU ml−1 (the absorbance of 0·3) after c. 4 h incubation (Fig. 1), and then the culture was mixed with EG medium of pH 2·0 at 1 : 1 with a resulting pH of 2·5 ± 0·1. When cells grown at pH 7·5 were directly challenged at pH 2·5, the pH of the mixture was adjusted quickly to pH 2·5. After cells were incubated without shaking under various conditions for the indicated time, they were plated on LB agar plates at pH 6·9. Viable cells were counted after the plates were incubated at 37°C for 15–20 h. Three to four repetitions were carried out for each experiment.

Figure 1.

Growth curve at the adaptation stage. Cells cultured at pH 7·5 were inoculated into the adaptation medium at pH 5·5 as described in Materials and methods. Growth was monitored by measuring the absorbance of the medium at 600 nm. (bsl00066) No addition; (bsl00072) 5 mmol l−1 glutamate was added; (•) CO2 was supplied; (bsl00001) N2 was supplied


When CO2 was supplied at the adaptation stage, the survival rate increased (Table 1, Exps A and B). It was shown that the AR systems were also induced under anaerobiosis (Diez-Gonzalez and Russell 1999; Diez-Gonzalez and Karaibrahimoglu 2004). However, the effect of CO2 was unlikely due to anaerobiosis because the survival rate was fourfold higher than that with N2 supply (Table 1, Exps B and C). The medium pH was decreased <0·2 pH units during adaptation in an EG medium of pH 5·5 under all conditions used, making it unlikely that the effect of CO2 was because of the decrease in medium pH during adaptation stage. When CO2 was supplied at the acid challenge stage, the survival rates were slightly higher than that with the N2 supply in both cells adapted with CO2 and N2 supply (Table 1, Exps D–G). From these results, it was clear that the adaptation stage is more important for acquiring CO2-dependent AR.

Table 1.  Effect of carbon dioxide, nitrogen and glutamate on the induction of acid resistance
Exp.Adaptation conditionsAcid challenge conditionsSurvival rate (%) ± SD after acid challenge for 1 h
  1. *Glutamate was added at 5 mmol l−1.

ANo additionNo addition2·4 ± 0·7
BWith CO2No addition55 ± 16
CWith N2No addition14 ± 1
DWith CO2With CO294 ± 17
EWith CO2With N286 ± 9
FWith N2With CO256 ± 10
GWith N2With N250 ± 3
HWith Glu*With Glu*67 ± 10
INo adaptationWith N20·0049 ± 0·0012
JNo adaptationWith CO20·038 ± 0·018

It was reported that glutamate-protected cells against acidic environments (Arnold and Kaspar 1995; Castanie-Cornet and Foster 2001). The survival rate enhanced by the addition of 5 mmol l−1 glutamate was lower than that with CO2 supply under our conditions (Table 1, Exp. H). The addition of glutamate had no significant effect on the growth at pH 5·5 (Fig. 1) and no growth was observed at pH 2·5. Next, cells grown at pH 7·5 were challenged in EG medium at pH 2·5 without adaptation at pH 5·5. The survival rate of the nonadapted cells increased again when the acid challenge was performed with the CO2 supply (Table 1, Exps I and J). When stationary phase cells were directly adapted at pH 5·5 and challenged in EG medium of pH 2·5 with CO2 supply, the survival rate was 0·030 ± 0·015%.

To further confirm that CO2 induced the AR of E. coli, the AR was challenged for different time periods. The survival rate of cells adapted with the CO2 supply decreased to 21% after the acid challenge for two h, but a further decrease in the survival rate was slight after the acid challenge for 3 h (Table 2). The rates observed with the CO2 supply were c. 2·5-fold higher than those with the N2 supply at any challenge time tested (Table 2). The changes in medium pH during the acid challenge for 3 h were <0·2 pH units in both cases. These results suggested that CO2 was able to protect cells against cell death under acidic conditions in E. coli.

Table 2.  Effect of carbon dioxide on survival during acid challenge for different time periods
Adaptation conditionsAcid challenge conditionsSurvival rate (%) ± SD
1 h2 h3 h
With CO2With CO294 ± 1721 ± 115 ± 5
With N2With N250 ± 39·0 ± 3·96·0 ± 2·4


It has been determined that the bacterial acid resistance mechanism is regulated by four systems including oxidative, glutamate-, arginine- and lysine-dependent systems (Lin et al. 1995; Diez-Gonzalez and Karaibrahimoglu 2004). The survival rates of adapted cells were higher than those of nonadapted cells (Foster and Hall 1990). Aeration increased AR in the presence of lysine and arginine, but there was no difference in the glutamate-dependent AR system between aerobic and anaerobic cells (Diez-Gonzalez and Karaibrahimoglu 2004). In addition, it has been reported that the RpoS-dependent AR system is present in E. coli and Salmonella enterica serovar Typhimurium (Hengge-Aronis 1993; Castanie-Cornet and Foster 2001; Diez-Gonzalez and Karaibrahimoglu 2004).

Our data suggested a novel idea concerning the AR system. When CO2 was supplied to both media of pH 5·5 and 2·5, the AR of E. coli was enhanced, although the growth rate at pH 5·5 was decreased slightly by the addition of this gas (Fig. 1). No growth was observed under all acid challenge conditions. It is unlikely that this effect was because of anaerobiosis with the CO2 supply because the survival rates were higher than that with the N2 supply. The CO2 supply was more effective when this gas was supplied at the adaptation stage, while the survival increased only slightly with CO2 supplied during the acidic challenge.

The volume of CO2 gas in the balloon decreased from c. 2 to 0·5 L after 4 h culture at pH 5·5 under our experimental conditions, and a similar decrease in volume was observed at pH 2·5, but the decrease in the medium pH was <0·2 pH units at both pH 5·5 and 2·5. As pKa values of H2CO3 are 6·37 and 10·2, no detectable change in the medium pH was observed with the dissolution of CO2 in a buffered acidic medium. The gas volume of N2 did not decrease significantly, and the N2 supply had no significant effect on the medium pH.

When the AR was challenged for different time periods, we found that there was a decrease in survival rate between 1 and 2 h, but the rate decreased slightly between 2 and 3 h. At any challenge time tested, the survival rates with the CO2 supply were higher than those with the N2 supply (Table 2). A 15- to 20-h incubation on LB plates may not be sufficient for the recovery of some subinjured cells. As the increase in the survival rate was not marked in the incubation for more than 1 day, the survival was measured without counting such subinjured cells in the present study.

How CO2 protects cells against cell death at extreme low pH remains to be clarified. CO2 is used metabolically via carboxylation enzymes and biosynthesis of carbamoyl phosphate, which is essential for nucleotide synthesis. In fact, it was reported that carbonate was essential for bacterial growth (Guilloton et al. 1993; Kusian et al. 2002), while a high level of CO2 repressed carboxylases and decarboxylases (King and Nagel 1975). Escherichia coli has multiple carbonic anhydrases and these enzymes were essential for E. coli growth (Kusian et al. 2002; Hashimoto and Kato 2003).

Escherichia coli exhibits amino acid decarboxylases such as ornithine, lysine, glutamate and arginine decarboxylases, and amino acid decarboxylase-dependent AR was demonstrated previously. (Lin et al. 1995; Guilfoyle and Hirshfield 1996; Diez-Gonzalez and Karaibrahimoglu 2004). It is still unclear how these enzymes protect bacteria against cell death at extreme low pH. Amino acid decarboxylases produce CO2 and their physiological role was proposed as supplying CO2 (Takayama et al. 1994). Our data suggested that CO2 increased AR at both the adaptation and acid challenge stages, and that the effect of CO2 was much stronger than that of glutamate. Carbon dioxide-related metabolism may be essential for AR systems in E. coli.

Bacterial AR is one of the important factors for bacterial pathogenesis. The concentration of CO2 supplied in this study was much higher than the level in air. The CO2 level can be argued to be high in the stomach, because carbonate is secreted in the human oesophagus and from the salivary glands (Brown et al. 1995), resulting in an increase in the CO2 level in the stomach. CO2 is also supplied from food and bacteria. For example, Helicobacter pylori produces CO2 from urea (Dunn et al. 1997). Thus, the effect of CO2 on AR may be significant in bacterial pathogenesis. In addition, our data may be useful for developing bacterial fermentation and food preservation techniques at the industrial level.