• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. Dec 2003; 185(24): 7044–7052.
PMCID: PMC296246

The Growth Advantage in Stationary-Phase Phenotype Conferred by rpoS Mutations Is Dependent on the pH and Nutrient Environment


Escherichia coli cells that are aged in batch culture display an increased fitness referred to as the growth advantage in stationary phase, or GASP, phenotype. A common early adaptation to this culture environment is a mutant rpoS allele, such as rpoS819, that results in attenuated RpoS activity. However, it is important to note that during long-term batch culture, environmental conditions are in flux. To date, most studies of the GASP phenotype have focused on identifying alleles that render an advantage in a specific environment, Luria-Bertani broth (LB) batch culture. To determine what role environmental conditions play in rendering relative fitness advantages to E. coli cells carrying either the wild-type or rpoS819 alleles, we performed competitions under a variety of culture conditions in which either the available nutrients, the pH, or both were manipulated. In LB medium, we found that while the rpoS819 allele confers a strong competitive fitness advantage at basic pH, it confers a reduced advantage under neutral conditions, and it is disadvantageous under acidic conditions. Similar results were found using other media. rpoS819 conferred its greatest advantage in basic minimal medium in which either glucose or Casamino Acids were the sole source of carbon and energy. In acidic medium supplemented with either Casamino Acids or glucose, the wild-type allele conferred a slight advantage. In addition, populations were dynamic under all pH conditions tested, with neither the wild-type nor mutant rpoS alleles sweeping a culture. We also found that the strength of the fitness advantage gained during a 10-day incubation is pH dependent.

Escherichia coli occupies numerous habitats under a wide variety of conditions. In some habitats, E. coli cells are often exposed to a variety of rapidly changing conditions, which may include changes in pH, nutrient availability, temperature, oxidative stress, and osmolarity. Most bacteria, including E. coli, have a variety of stress response mechanisms that allow them to survive these changes. RpoS, a sigma factor that is a global regulator of gene expression, plays a particularly important role in the E. coli stress response (20-22). Although stress responses allow E. coli to survive many changes in environmental conditions, over the long term, genetic adaptation, through natural selection of advantageous mutations, will lead to greater fitness in an altered environment.

In a rich medium, such as Luria-Bertani broth (LB), E. coli grown in aerated batch culture for several days at 37°C passes through five distinct phases (17, 39). During exponential phase, which follows an initial lag phase, the population grows rapidly, doubling every 20 to 30 min. During this period of rapid growth, carbohydrates initially serve as the primary carbon and energy source but are quickly spent, and other nutrients, including peptides, amino acids, nucleic acids, nucleotides, and fatty acids, are utilized to sustain growth (7, 14, 29, 52). As these remaining nutrients are depleted, the population enters stationary phase, during which little change is observed in the number of viable cells in the culture for 2 to 3 days. As the culture enters death phase, cell viability abruptly declines and approximately 99% of the cells die. Finally, during long-term stationary phase, there is a long period of slow decline in the remaining viable cell population (12, 13, 17, 18, 39).

Two significant changes that occur in the medium during stationary phase may play important roles in the precipitous loss of viability that follows: the depletion of readily available nutrients and an increase in medium pH. During early exponential phase, as cells consume carbohydrates, acetate is released and the medium becomes acidic (27). As the carbohydrate supply is exhausted, amino acid catabolism predominates and, as cells enter stationary phase, ammonia released into the medium causes it to become basic (29). To survive under these conditions, cells must adapt to a high energy expenditure that is imposed on them by the need to actively acquire protons from the basic medium environment in order to maintain pH homeostasis in the cytoplasm (37). Eventually, readily available nutrients are exhausted and the only nutrients available are those derived from the dead bacteria (52). In order to scavenge nutrients from the bacterial debris, proteins, nucleic acids, and lipids must be broken down into their constituent parts, which imposes another large energy cost on the cell.

The rapid, stressful changes that occur in the medium during the first few days in culture place a heavy selective burden on the E. coli population. Given the mutation rate in batch culture, a large population should contain abundant genotypic diversity upon which selection can act, leading one to expect that E. coli can adapt quickly to significant sublethal changes in the local environment. Indeed, this phenomenon is observed when E. coli cells are incubated for several days in LB batch culture (15, 48, 49). It has been shown that when E. coli cells that have been incubated for at least 10 days in LB batch culture are competed against E. coli cells that have been cultured for a single day, the bacteria from the older culture exhibit a significant fitness advantage and eventually take over the culture (49). The increased fitness displayed by the older culture population is referred to as the growth advantage in stationary phase, or GASP, phenotype (13, 48, 49). The first allele linked to the GASP phenotype, rpoS819, encoded an RpoS with attenuated activity (49). RpoS is an important regulator of stationary-phase gene expression and also plays a key role in the general stress response of E. coli (21). Attenuated RpoS activity has been shown to be a common early adaptation to the LB batch culture environment and typically confers a fitness advantage in competitions against wild-type cells in LB batch culture.

We determined the relative advantage rendered to E. coli by the wild-type or rpoS819 alleles under culture conditions in which either the available nutrients, the pH, or both had been manipulated. We determined that the rpoS819 allele confers a strong competitive fitness advantage under basic pH conditions, shows a reduced advantage under neutral conditions, and is disadvantageous under acidic conditions. Under basic conditions, rpoS819 exhibits its strongest advantage when amino acids are the primary source of nutrients. We have also found that the attenuated RpoS phenotype emerges during stationary phase in cultures grown in acidic, neutral, or alkaline media and that the variability, from culture to culture, in the proportion of the cells having this phenotype is greatest in acidic media and least in neutral media. Finally, competition assay results suggest that E. coli increases its relative fitness at a lower rate under neutral and acidic conditions than under basic ones.

To date, most studies of the GASP phenotype have focused on the attenuation of RpoS activity, which is the earliest adaptation to long-term LB batch culture that has been identified. This may have given the impression that a mutation in rpoS is always essential for GASP. As we show here, that is not the case. These studies also shed light on an apparent paradox concerning rpoS. Namely, if the reduction of RpoS activity is beneficial in terms of relative fitness, why would an E. coli cell carry the wild-type allele?


Strains, media, and culture conditions.

The bacterial strains used are listed in Table Table11 and are derived from ZK126, a descendant of E. coli K-12 strain W3110 (8). All experiments were performed at 37°C with aeration in a TC-7 rolling drum (New Brunswick Scientific, Edison, N.J.). LB broth Lennox (Difco) was prepared following the manufacturer's instructions. M63 minimal medium was prepared as described elsewhere (30) and was supplemented with 1 μg of thiamine per ml, 1 mM MgSO4, and either 0.2% glucose or 0.2% Casamino Acids. pH (5.0, 7.0, or 9.0) of minimal medium was set by manipulating the concentrations of K2HPO4 and KH2PO4. Total phosphate in the solution was the same regardless of pH. pH of the LB broth was set by adding morpholineethanesulfonic acid (pH 5.0), HEPES (pH 7.0), or N-tris(hydroxymethyl)methyl-3-aminopropane-sulfonic acid (TAPS) (pH 9.0) to 100 mM. Viable counts were determined by serial dilution with plating on either LB plates or, where appropriate, LB plates supplemented with streptomycin (25 μg/ml) or nalidixic acid (20 μg/ml). The limit of detection in all experiments was <1,000 CFU/ml, unless otherwise noted. All chemicals were obtained from Sigma-Aldrich (St. Louis, Mo.).

E. coli strains used in this study

Preparation of heat-killed, conditioned media.

LB cultures were inoculated from frozen glycerol stocks of either ZK1142 or ZK1143 into 6 ml of LB and grown at 37°C for either 5 or 10 days. To kill cells, cultures were then transferred to a shaking water bath and incubated for 2 h at 60°C. One milliliter of each culture was plated on LB agar containing either nalidixic acid (20 μg/ml) or streptomycin (25 μg/ml). No viable counts were seen in any heat-killed culture. Neutral pH heat-killed medium was buffered with 100 mM HEPES.

Construction of bolA::lacZ strains.

Strains SF2054, SF2055, SF2056, and SF2057 were created by P1vir transduction of the λ-bolA::lacZ-Kanr fusion in strain ZK918, a lambda lysogen (3). Transduced E. coli cells were plated on MacConkey plates containing kanamycin (50 μg/ml). Kanamycin-resistant transductants of ZK1142 and ZK1143 that yielded red colonies or of ZK819 and ZK820 that yielded pink colonies were streaked onto MacConkey plates. Single colonies were selected and grown overnight at 37°C in LB containing kanamycin (50 μg/ml) and nalidixic acid (20 μg/ml) or streptomycin (25 μg/ml), as appropriate.

Assays for attenuated rpoS activity.

For plate assays of RpoS activity of E. coli cells containing the λ-bolA::lacZ gene fusion construct, each culture was sampled, appropriately diluted, and then spread onto MacConkey plates. Plates were incubated at 37°C for 16 h. Colonies were scored as RpoS wild type or attenuated based on their color: red for wild type and light pink or white for attenuated. The presence of an attenuated or wild-type rpoS allele was confirmed by performing an assay for catalase activity; the activity of the stationary-phase-specific catalase, KatE, is dependent on RpoS (36). Colonies from MacConkey plates were patched onto LB agar plates and grown overnight at 37°C. Catalase activity was scored by assessing the bubbling generated by the O2 released following the dropwise application of a hydrogen peroxide (30% [wt/vol]) solution. Instant and vigorous bubbling reflected wild-type RpoS activity (24). Delayed and/or reduced bubbling indicated the presence of an attenuated rpoS mutation.

Competition assays.

Competition experiments were performed essentially as described previously (15, 49). Five-milliliter aliquots of overnight cultures were inoculated from frozen glycerol stocks. These were then subcultured and incubated for 24 h before being mixed for the competitions. For GASP competitions, one set of cultures was incubated for 10 days and then inoculated at a 1:1,000 (vol/vol) dilution into a culture grown for 1 day. Cells in the 10-day-old and 1-day-old cultures carried different markers. The cell densities of the two populations were monitored by serial dilution and plating on LB agar containing nalidixic acid (20 μg/ml) or streptomycin (25 μg/ml).


Competitions of wild-type and rpoS819 alleles in heat-killed, conditioned medium.

We have found that when E. coli cells carrying a wild-type rpoS allele are incubated in LB batch culture, rpoS mutants expressing an attenuated phenotype emerge within a few days of inoculation and have a striking fitness advantage during and following death phase. What is less obvious are the environmental factors that are responsible for this fitness advantage. This question was addressed by competing wild-type E. coli against rpoS819 mutants in heat-killed medium obtained from 10-day-old cultures (Fig. (Fig.1).1). Heat-killed medium consists simply of an aged batch culture that has been incubated at 60°C for 2 h, rendering all cells dead. Following the death phase, nutrients must be derived from the cell debris. Therefore, to approximate conditions found in the medium after death phase, we performed competitions in heat-killed medium, rather than filtered conditioned medium, because a substantial portion of the potential nutrients are lost during filtration when bacterial carcasses and cell debris are removed. When grown in 10-day-old heat-killed medium, rpoS819 mutants had a significant fitness advantage (Fig. (Fig.1A).1A). In two of four competitions the wild-type cells were undetectable by day 7, and in the other competitions the attenuated mutants outnumbered the wild type by more than 600-fold by day 10. Similar results were found using 5-day-old heat-killed medium (data not shown).

FIG. 1.
The rpoS819 allele confers a competitive advantage in 10-day-old heat-killed, conditioned medium. Wild-type (solid lines) and rpoS819 (dashed lines) cells from 1-day-old cultures were inoculated at 1:1,000 into unbuffered (A) or pH 7 buffered heat-killed ...

Although nutrient availability has a significant impact on the relative fitness of wild-type and rpoS mutant strains, another factor that may affect the relative fitness conferred by these alleles is the alkaline pH of the stationary-phase medium. To determine whether pH plays an important role in the fitness advantage conferred by the rpoS819 mutation in heat-killed medium, we buffered the medium to pH 7 before competing the wild-type and rpoS819 mutant strains. Once again, the rpoS819 allele conferred a competitive advantage (Fig. (Fig.1B).1B). In this case, the fitness advantage was typically little more than a 10-fold advantage (Fig. (Fig.1B).1B). In heat-killed medium that had not been buffered, the competitive fitness advantage of the rpoS819 mutant was greater than 600-fold, suggesting that pH may play a greater role in the outcome of the competitions than does nutrient content under these conditions. There is also a possibility that nutrient content and pH have a synergistic effect on competitive fitness.

Competition of wild-type and rpoS819 strains in basic, neutral, and acidic LB.

To further investigate the effect of pH on the fitness advantage conferred to E. coli carrying an rpoS819 allele, competitions were performed in basic, neutral, and acidic LB media. In basic medium, as expected, rpoS819 conferred a significant fitness advantage, just as it does in a culture that is not buffered (Fig. (Fig.2A).2A). In basic medium, the rpoS819 mutant had a 30- to 380-fold advantage by day 10 (Fig. (Fig.2A).2A). In neutral medium, the presence of the rpoS819 allele conferred a reduced competitive fitness advantage of 8- to 50-fold (Fig. (Fig.2B).2B). However, in acidic medium, rpoS819 conferred no significant fitness advantage and may have been slightly deleterious (Fig. (Fig.2C2C).

FIG. 2.
Competitive fitness of wild-type and rpoS819 cells in LB batch culture is dependent on pH. Wild-type (solid lines) and rpoS819 (dashed lines) cells from 1-day-old cultures were inoculated at 1:1,000 into LB buffered at pH 9 (A), pH 7 (B), or pH 5 (C). ...

When incubated in monoculture, the attenuated and wild-type strains had identical growth rates and viable cell yields at each pH (data not shown).

RpoS phenotypic diversity during batch culture in basic, neutral, and acidic rich medium.

Since rpoS819 alleles are disadvantageous in acidic medium, we speculated that rpoS819 mutants might occasionally revert back to the wild-type phenotype, increase in number, and become detectable during long-term incubation under acidic conditions. To determine the RpoS phenotypes of cells within a culture population during long-term incubation, we employed a λ-bolA::lacZ promoter fusion whose expression is dependent on RpoS (5). When plated on MacConkey agar, colonies in which the bolA promoter is strongly induced are dark red, indicating the presence of wild-type RpoS. Colonies carrying an rpoS819 allele are pink or white in appearance. Cells were also patched to fresh LB plates and treated with hydrogen peroxide to determine the relative strength of catalase activity using the bubble assay (24). In every case, catalase activity was high in dark red colonies and low in white or pink colonies (data not shown).

In acidic cultures initially inoculated with E. coli carrying an rpoS819 allele, we found that wild-type revertants were seen as early as day 6 (Table (Table2).2). Given that no significant fitness advantage is conferred by either allele under these conditions, it was particularly surprising that no wild-type cells were present after day 7 in culture. In fact, in acidic medium the proportion of the population carrying the wild-type allele peaked on day 7 at 28% and thereafter was undetectable (Table (Table2).2). In contrast, when acidic LB cultures were inoculated with E. coli carrying the wild-type allele, although rpoS mutants were observed by day 4 the wild-type phenotype was seen in 62% of the population on day 10 (Table (Table22).

Frequency of wild-type rpoS allelesa

We then looked for rpoS reversion during incubation in neutral and basic LB cultures (Table (Table2).2). Even under these conditions, rpoS819 alleles occasionally reverted to the wild type, though at a lower frequency. In all instances, the frequencies of wild-type alleles peaked on day 7, at 10% in neutral medium and 7% in basic medium; thereafter, wild-type cells were undetectable (Table (Table22).

We also determined RpoS phenotypes for wild-type cells inoculated into acidic, neutral, and basic cultures. Attenuated RpoS activity was first observed after 3 days of incubation in alkaline medium, 4 days in acidic medium, and 5 days in neutral medium (Table (Table2).2). Although the frequency of the wild-type allele in acidic medium (62%) was double that found in basic medium (30%) on day 10, the wild-type allele frequency was highest in neutral medium (76%) (Table (Table2).2). Given the results of the competition experiments, we had expected that the wild-type frequency would be highest in the acidic medium. We also observed a great deal of variability from culture to culture in the alkaline and acidic media (Table (Table22).

Competition between 10-day-old and 1-day-old cultures in basic, neutral, and acidic LB medium.

Since pH stress appears to play an important role in manifesting the GASP phenotype conferred by the rpoS819 allele, we wanted to determine how well E. coli cells incubated for 10 days at a particular pH might compete against E. coli cells incubated overnight at that same pH and against overnight cultures grown in unbuffered medium. As expected, E. coli grown in basic medium for 10 days competed very well in cultures that were not buffered (Fig. (Fig.3A)3A) and in basic medium (Fig. (Fig.3D).3D). This was not particularly surprising, given that unbuffered LB cultures become basic by the time they reach stationary phase. Somewhat surprising was the fact that cultures grown at neutral pH for 10 days competed well when inoculated as a minority into cultures that were not buffered (Fig. (Fig.3B),3B), as well as in neutral medium (Fig. (Fig.3E).3E). We found that E. coli cells incubated for 10 days under acidic conditions competed well in acidic medium (Fig. (Fig.3F)3F) but competed very poorly in cultures that were not buffered (Fig. (Fig.3C).3C). One notable difference between competitions performed in acidic, neutral, or basic medium is the strength of the competitive fitness advantage of the E. coli cells taken from 10-day-old cultures. The relative competitive fitness of cells from cultures grown for 10 days in basic medium is quite strong, >1,800-fold by day 10 (Fig. (Fig.3D).3D). The advantage of the 10-day-old cultures grown in neutral medium was reduced (Fig. (Fig.3E)3E) to ~75-fold, and 10-day-old cultures grown in acidic medium showed significantly smaller increases in fitness (Fig. (Fig.3F),3F), with the older culture still predominant in the population on day 10.

FIG. 3.
pH determines the strength of the GASP phenotype. Wild-type strain ZK1143 was incubated for 10 days in LB buffered at pH 9 (A and D), pH 7 (B and E), or pH 5 (C and F). Aliquots of these cells were then inoculated as a minority (dashed lines) into 1-day-old ...

Competition in minimal medium containing Casamino Acids.

During stationary phase, one of the primary nutrients available is amino acids derived from dead bacteria (52). As shown above, another important factor affecting the fitness of E. coli in culture is the pH of the medium. In order to directly determine the impact of pH on the fitness advantage conferred by the rpoS819 allele when amino acids are the primary nutrient, we performed competition experiments in minimal medium supplemented with Casamino Acids, buffered to either basic, neutral, or acidic pH. In basic medium, the rpoS819 allele conferred up to a 1,000-fold competitive advantage (Fig. (Fig.4A).4A). This was not surprising given that, during long-term stationary phase, amino acids and small peptides probably provide a substantial portion of the nutrients available and that the medium is basic. In neutral medium, E. coli cells carrying an rpoS819 allele initially had a slight fitness advantage over the wild type, but the advantage was lost by day 10 (Fig. (Fig.4B).4B). In acidic medium, the wild-type allele conferred up to a 45-fold advantage over cells harboring rpoS819 (Fig. (Fig.4C4C).

FIG. 4.
Competitive fitness of wild-type and rpoS819 cells incubated in minimal medium supplemented with Casamino Acids is dependent on pH. Wild-type (solid lines) and rpoS819 (dashed lines) cells from 1-day-old cultures were inoculated at 1:1,000 into minimal ...

In monocultures, the rpoS819 and wild-type strains showed equal relative fitness at each pH, as determined by growth rates and viable cell counts (data not shown).

Competition in glucose minimal medium.

To determine whether the impact of pH on E. coli carrying an rpoS819 allele was nutrient dependent, we competed rpoS819 mutants against wild type in glucose minimal medium at basic, neutral, or acidic pH (Fig. (Fig.5).5). In basic glucose minimal medium, the rpoS819 allele conferred a small but consistent fitness advantage (Fig. (Fig.5A).5A). Neither strain had a clear advantage in neutral glucose minimal medium (Fig. (Fig.5B).5B). In acidic glucose minimal medium, wild-type cells initially had a significant competitive fitness advantage of approximately 100-fold, but much of that advantage was lost by day 10 (Fig. (Fig.5C).5C). For the most part, the wild-type advantage in acidic glucose minimal medium appeared to be conferred by a lower rate of death, while in basic medium the rpoS819 allele conferred a consistent competitive fitness advantage.

FIG. 5.
Competitive fitness of wild-type and rpoS819 cells in glucose minimal medium is dependent on pH. Wild-type (solid lines) and rpoS819 (dashed lines) cells from 1-day-old cultures were inoculated at 1:1,000 into minimal medium supplemented with glucose ...

In monocultures, no difference in growth or survival was observed between the wild type and attenuated mutant strains (data not shown).


E. coli is found in numerous niches, under varying, often rapidly changing conditions of nutrient availability, pH, osmolarity, oxidative stress, and temperature. To survive under a variety of environmental conditions, bacteria must be able to adapt quickly to alterations in any combination of these factors. E. coli has a number of stress response mechanisms that allow it to adapt to sublethal changes in the environment, but extended exposure to a new environment will result in selective pressure favoring specific new alleles that confer a fitness advantage in the new environment.

Given the relatively short generation time and the large populations that can be maintained under laboratory conditions, it is possible to study, in a limited way, the dynamics of genetic adaptation to a changing environment. For instance, in LB batch culture, two significant changes occur in the medium during stationary phase that may play important roles in the precipitous loss of viability that follows: the depletion of readily available nutrients and an increase in medium pH.

It has been well documented that in long-term batch culture E. coli rapidly adapts to a changing environment (13, 15, 48, 49, 52, 53). As these cultures age, changes in population density, nutrient availability, and pH place the bacteria under significant selective pressure. Therefore, it is not surprising that E. coli cells taken from 10-day-old cultures out-compete those from overnight cultures, and those from 20-day-old cultures out-compete those from 10-day-old cultures (15, 48, 49). An early adaptation commonly selected for under these conditions is a mutation in rpoS that results in an attenuation of RpoS activity (48, 49). RpoS is an important regulator of stationary-phase gene expression and also plays a key role in the E. coli general stress response (21, 22).

While RpoS phenotypes are known to be diverse on day 10 in an LB batch culture, phenotypic diversity and variability prior to day 10, during early stationary phase and the first few days in long-term stationary phase, had not been quantitated. Under other culture conditions, sweeps of new alleles that confer a particularly strong fitness advantage have been observed. For instance, in chemostats containing minimal glucose medium at neutral pH, rpoS null mutations quickly take over the population (32). However, when the medium is maintained at acidic pH a sweep does not occur. One possible explanation for this difference is the fact that the cells must adapt to two conflicting stresses. Since glucose scavenging is regulated by RpoD, it seems plausible that the elimination of RpoS activity may reduce competition with RpoD for the core RNA polymerase, allowing a higher level of expression of those genes involved in glucose scavenging. However, in acidic medium, the stress response requires wild-type RpoS activity. The result may be that some cells that survive are better adapted to acid stress and others are better adapted to nutrient stress. Each cell must balance its need to secure nutrients with its need to resist acid stress, and there are probably multiple ways to balance these needs that result in equal fitness.

While cells that have attenuated RpoS activity do not entirely sweep an LB batch culture, they usually have a clear fitness advantage during the first 10 days of incubation. We have found that pH plays a particularly important role in determining the rpoS allele that confers a fitness advantage. In LB batch culture and in basic minimal medium containing either Casamino Acids or glucose, the advantage of the rpoS819 allele first manifests itself during death phase when selective pressures, such as alkaline pH and a dramatic change in the profile of available nutrients, may be greatest. In acidic medium, the wild-type allele confers a significant fitness advantage. Considering the clear advantage that the wild-type allele confers in acidic medium, it is surprising that when acidic LB is inoculated with E. coli cells carrying the wild-type allele, after 6 days of incubation a significant portion of the culture carries an rpoS allele that results in attenuated activity. In addition, the variability from culture to culture in the proportion of the cells having wild-type RpoS activity is quite dramatic in acidic media. A similar, though less dramatic, variability is seen between the alkaline cultures, while little variability is seen in the neutral cultures. Perhaps the degree of variability from culture to culture is dependent on the number of stresses present and the degree to which those stresses are in conflict. It has recently been shown that mutagenesis in aging colonies is strongest where selective pressures are most diverse (4). Perhaps this higher mutation rate, along with the probability that conflicting stresses can be adapted to in a number of ways that render equal fitness, explains the difference in variability found between the neutral and acidic cultures.

Although rpoS819 confers a significant fitness advantage in heat-killed medium obtained from aged LB batch cultures, the fitness advantage decreases significantly when this medium is buffered to neutral pH. In fact, the advantage conferred by rpoS819 at neutral pH appears to be limited exclusively to exponential growth in heat-killed medium, and it is probably due to its more efficient catabolism of the available nutrients. However, in heat-killed medium that is not buffered, rpoS819 continues to confer a fitness advantage in stationary phase that ultimately results in wild-type cells being undetectable. Thus, although an advantage is conferred by rpoS819 when the nutrients available are derived from LB batch cultures that have passed through death phase, it is the additional stress caused by the alkaline pH of the medium that has the greater impact. This is supported by the fact that rpoS819 also confers a fitness advantage in glucose minimal medium at alkaline pH.

The question arises as to the conditions that render a fitness advantage to cells expressing wild-type RpoS. While rpoS819 is advantageous for catabolizing amino acids in basic medium, we found that in acidic minimal medium containing Casamino Acids, rpoS+ E. coli cells have a competitive advantage. This is not surprising, given the importance of RpoS for acid resistance (6, 16, 38). It is interesting that while wild-type cells are only slightly affected by a short exposure to a severe acid stress in LB (pH 2.5 for 2 h), the same stress is lethal to a strain having attenuated RpoS activity (44). In acidic glucose minimal medium, the wild type has a small fitness advantage over the attenuated strain. It may be that E. coli acid resistance systems work less efficiently in acidic media containing glucose than in LB or in media containing Casamino Acids. The RpoS-dependent acid resistance system is repressed by glucose (6), and two other acid resistance systems require amino acids, either glutamate or arginine, to be activated (23, 25, 26). In addition, rpoS mutants are known to scavenge glucose more efficiently than wild-type cells (11).

Our results also suggest that E. coli populations increase their relative competitive fitness more slowly in neutral and acidic environments than under basic conditions. This is not surprising, since E. coli is often exposed to acidic and neutral conditions in natural environments and, therefore, may already be well-adapted to this pH range. In the stomach, E. coli must survive high levels of inorganic acid and both inorganic and organic acids in the small intestine. In the stomach of a healthy human who fasts, the pH is typically about 2.0 (43). Under acidic conditions, the need to maintain cytoplasmic pH homeostasis may be mitigated by the pH gradient that exists across the cytoplasmic membrane serving as a source of energy, decreasing the fitness load caused by limited nutrient availability (19). In neutral medium, there is no need to actively maintain cytoplasmic pH, relieving a significant stress that is found in a basic environment (37). Since a population should adapt to the conditions under which it is aged, we expected that E. coli incubated at one pH would compete less well at another. However, cells aged in neutral media adapt to nutrient availability and so, since the 1-day-old cultures have not adapted to either pH or nutrient availability, it is reasonable that these aged cells would out-compete 1-day-old cells.

Since wild-type cells have a competitive advantage in acidic LB, we wanted to determine whether occasional revertants might increase in number and become detectable during long-term incubation in this medium. In acidic LB medium inoculated with E. coli cells carrying rpoS819, we observed wild-type revertants within 6 days in culture. We also observed revertants in basic and neutral pH cultures, although at much lower frequencies. However, under all conditions tested, the wild-type revertants that emerged in the cultures were ultimately lost. Since rpoS819 renders a strong fitness advantage under basic conditions, the revertants observed in alkaline medium were unexpected. The fact that revertants were only observed on days 6 and 7 suggests that the fitness advantage conferred by the wild-type allele is limited to, or at least most potent, during early long-term stationary phase. It is interesting that a burst of synthesis of Dps, a gene regulated by RpoS during stationary phase, peaks on day 6 to 7 during long-term stationary-phase incubation (S.Nair and S. E. Finkel, unpublished data).

It has been observed that some, but not all, E. coli isolates have attenuated RpoS activity after 10 days of incubation in LB batch culture (49). However, this diversity in RpoS phenotypes has not been quantitated over the course of a 10-day incubation. In addition, the effect of pH on this diversity has not been addressed. Therefore, we inoculated acidic, basic, and neutral LB cultures and then determined the proportion of cells having attenuated RpoS activity on each day for 10 days. We first observed cells having an attenuated RpoS activity on day 3 in basic medium, day 4 in acidic medium, and day 5 in neutral medium. Although the proportion of cells having the wild-type phenotype in acidic medium was double that found in basic medium on day 10, the highest proportion was found in neutral medium. Given the results of the competition experiments, we had expected that the wild-type frequency would be highest in the acidic medium. The detrimental effect of the wild-type RpoS phenotype on scavenging nutrients may explain the discrepancy between our expectations and the results that were obtained. These experiments were done in triplicate at each pH, and a great deal of variability was observed from culture to culture in the alkaline and acidic media, but little in the neutral medium. The presence of multiple stresses in the acidic and basic cultures may be the cause of this difference in variability.

Although RpoS regulates the E. coli response to low nutrient conditions and to a variety of other stresses (20-22), other sigma factors play important roles in stress responses as well (20). RpoE functions overlap those of RpoS, adding a degree of flexibility to the regulation of the general stress response (42). When E. coli is exposed to multiple stresses, competition among sigma factors for access to RNA polymerase may determine how strongly various regulons are expressed (20). For example, RpoD induces the expression of many important stress defense genes during stationary phase, including uspA (10). uspA encodes an integral membrane protein that is one of the most abundant proteins found in stationary-phase cells. This gene is expressed in response to growth arrest and can be superinduced by increased RpoD binding to RNA polymerase in an rpoS null mutant (10). In general, decreased levels of RpoS lead to the superinduction of RpoD-dependent stationary-phase-inducible genes, suggesting that these two proteins compete for RNA polymerase and for access to promoters. An rpoS-activated gene, uspB, which causes cell death when expressed at high levels, lies immediately upstream of uspA and also encodes an integral membrane protein. Perhaps cells having an attenuated RpoS activity express lower levels of this protein, which may confer a fitness advantage to them during death phase in alkaline medium. In a similar vein, it has been suggested that the RpoS-regulated entericidin locus plays a role in programmed cell death (3). It has also been suggested that senescence plays a major role in cell death, but this does not explain the long-term survival of E. coli populations of greater than 106 CFU/ml that follows the rapid loss of viability during the death phase (12).

RpoN activates the transcription of genes that are needed to efficiently assimilate ammonia and to facilitate the uptake of amino acids, especially glutamine (34, 51). During stationary phase, a large portion of the nutrients available are in the form of amino acids and peptides, making it possible that RpoN also plays an important regulatory role during this time. During stationary phase, RpoN also activates the psp operon, which has been shown to protect cells from alkaline stress (31, 46).

There are RpoS-induced genes that are required for competitive fitness during stationary phase. In fact, rpoS null mutants are quickly out-competed by wild-type cells during long-term stationary-phase incubation (M. J. Farrell and S. E. Finkel, unpublished data). Two RpoS-regulated genes that are known to be required for either competitive fitness or survival in stationary phase are dps and nhaA (1, 9, 28). dps encodes a nonspecific DNA binding protein involved in protecting the cell from many stresses (1, 50). nhaA encodes a Na+/H+ antiporter that is essential for survival in alkaline media (33). Expression of nhaA is induced by nhaR, which stimulates RpoD binding to the nhaA promoter (9). In alkaline environments up to pH 7.9, this antiporter helps maintain pH homeostasis (33, 37). Since nhaA expression can also be driven by RpoD, any decreased expression of nhaA due to attenuated RpoS activity can be compensated for. This type of dual regulation may be crucial for allowing E. coli to adapt to a wide range of environments.

Since a number of sigma factors are expressed during stationary phase and many genes are regulated by multiple promoters, sigma factors must compete both for access to RNA polymerase and to these promoters. The diversity of rpoS alleles found in LB batch cultures suggests that the activities of various sigma factors can be modulated in a variety of ways to efficiently promote protection from conflicting stresses that arise during stationary phase. The phenotypic diversity we observed in batch culture correlates well with the heterogeneity of rpoS alleles of E. coli and Salmonella found in natural, laboratory, and clinical environments, implying that the shifting, sometimes conflicting stresses found in many environments often lead to the emergence of rpoS mutants in many populations via natural selection (2, 35, 40, 41, 45, 47, 49). Furthermore, the fact that both wild-type and attenuated rpoS alleles are found in clinical and other natural settings suggests that the LB batch culture model mirrors important aspects of the evolution of bacteria in nature.


This work was supported by the USC/Norris Comprehensive Cancer Center, the James A. Zumberge Fund, and the W. M. Keck Foundation.

We thank Vyacheslav Palchevskiy, Evan Pepper, George O'Toole, Erik Zinser, and William Rosche for helpful comments.


1. Almiron, M., A. J. Link, D. Furlong, and R. Kolter. 1992. A novel DNA-binding protein with regulatory and protective roles in starved Escherichia coli. Genes Dev. 6:2646-2654. [PubMed]
2. Atlung, T., H. V. Nielsen, and F. G. Hansen. 2002. Characterization of the allelic variation in the rpoS gene in thirteen K12 and six other non-pathogenic Escherichia coli strains. Mol. Genet. Genomics 266:873-881. [PubMed]
3. Bishop, R. E., B. K. Leskiw, R. S. Hodges, C. M. Kay, and J. H. Weiner. 1998. The entericidin locus of Escherichia coli and its implications for programmed bacterial cell death. J. Mol. Biol. 280:583-596. [PubMed]
4. Bjedov, I., O. Tenaillon, B. Gerard, V. Souza, E. Denamur, M. Radman, F. Taddei, and I. Matic. 2003. Stress-induced mutagenesis in bacteria. Science 300:1404-1409. [PubMed]
5. Bohannon, D. E., N. Connell, J. Keener, A. Tormo, M. Espinosa-Urgel, M. M. Zambrano, and R. Kolter. 1991. Stationary-phase-inducible “gearbox” promoters: differential effects of katF mutations and role of sigma 70. J. Bacteriol. 173:4482-4492. [PMC free article] [PubMed]
6. Castanie-Cornet, M. P., T. A. Penfound, D. Smith, J. F. Elliott, and J. W. Foster. 1999. Control of acid resistance in Escherichia coli. J. Bacteriol. 181:3525-3535. [PMC free article] [PubMed]
7. Clark, D. P., and J. E. Cronan, Jr. 1996. Two-carbon compounds and fatty acids as carbon sources, p. 343-357. In F. C. Neidhardt et al. (ed.), Escherichia coli and Salmonella: cellular and molecular biology, 2nd ed. ASM Press, Washington, D.C.
8. Connell, N., Z. Han, F. Moreno, and R. Kolter. 1987. An E. coli promoter induced by the cessation of growth. Mol. Microbiol. 1:195-201. [PubMed]
9. Dover, N., and E. Padan. 2001. Transcription of nhaA, the main Na+/H+ antiporter of Escherichia coli, is regulated by Na+ and growth phase. J. Bacteriol. 183:644-653. [PMC free article] [PubMed]
10. Farewell, A., K. Kvint, and T. Nystrom. 1998. Negative regulation by RpoS: a case of sigma factor competition. Mol. Microbiol. 29:1039-1051. [PubMed]
11. Ferenci, T. 2001. Hungry bacteria—definition and properties of a nutritional state. Environ. Microbiol. 3:605-611. [PubMed]
12. Finkel, S. E., E. Zinser, S. Gupta, and R. Kolter. 1998. Life and death in stationary phase, p. 3-16. In S. J. W. Busby, C. M. Thomas, and N. L. Brown (ed.), Molecular microbiology, vol. H 103. Springer-Verlag, Berlin, Germany.
13. Finkel, S. E., E. R. Zinser, and R. Kolter. 2000. Long-term survival and evolution in the stationary phase, p. 231-238. In G. Storz and R. Hengge-Aronis (ed.), Bacterial stress responses. ASM Press, Washington, D.C.
14. Finkel, S. E., and R. Kolter. 2001. DNA as a nutrient: novel role for bacterial competence gene homologs. J. Bacteriol. 183:6288-6293. [PMC free article] [PubMed]
15. Finkel, S. E., and R. Kolter. 1999. Evolution of microbial diversity during prolonged starvation. Proc. Natl. Acad. Sci. USA 96:4023-4027. [PMC free article] [PubMed]
16. Foster, J. W., and M. Moreno. 1999. Inducible acid tolerance mechanisms in enteric bacteria. Novartis Found. Symp. 221:55-69. [PubMed]
17. Gay, F. P. 1935. Bacterial growth and reproduction. Charles C. Thomas, Springfield, Ill.
18. Gupta, S. 1997. Microbial survival: mutations that confer a competitive advantage during starvation. M.A. thesis. Harvard University, Cambridge, Mass.
19. Harold, F. M., and P. C. Maloney. 1996. Energy transduction by ion currents, p. 283-306. In F. C. Neidhardt et al. (ed.), Escherichia coli and Salmonella: cellular and molecular biology, 2nd ed. ASM Press, Washington, D.C.
20. Hengge-Aronis, R. 1999. Interplay of global regulators and cell physiology in the general stress response of Escherichia coli. Curr. Opin. Microbiol. 2:148-152. [PubMed]
21. Hengge-Aronis, R. 2002. Recent insights into the general stress response regulatory network in Escherichia coli. J. Mol. Microbiol. Biotechnol. 4:341-346. [PubMed]
22. Hengge-Aronis, R. 1993. Survival of hunger and stress: the role of rpoS in early stationary phase gene regulation in E. coli. Cell 72:165-168. [PubMed]
23. Hersh, B. M., F. T. Farooq, D. N. Barstad, D. L. Blankenhorn, and J. L. Slonczewski. 1996. A glutamate-dependent acid resistance gene in Escherichia coli. J. Bacteriol. 178:3978-3981. [PMC free article] [PubMed]
24. Lange, R., and R. Hengge-Aronis. 1991. Identification of a central regulator of stationary-phase gene expression in Escherichia coli. Mol. Microbiol. 5:49-59. [PubMed]
25. Lin, J., I. S. Lee, J. Frey, J. L. Slonczewski, and J. W. Foster. 1995. Comparative analysis of extreme acid survival in Salmonella typhimurium, Shigella flexneri, and Escherichia coli. J. Bacteriol. 177:4097-4104. [PMC free article] [PubMed]
26. Lin, J., M. P. Smith, K. C. Chapin, H. S. Baik, G. N. Bennett, and J. W. Foster. 1996. Mechanisms of acid resistance in enterohemorrhagic Escherichia coli. Appl. Environ. Microbiol. 62:3094-3100. [PMC free article] [PubMed]
27. Luli, G. W., and W. R. Strohl. 1990. Comparison of growth, acetate production, and acetate inhibition of Escherichia coli strains in batch and fed-batch fermentations. Appl. Environ. Microbiol. 56:1004-1011. [PMC free article] [PubMed]
28. Martinez, A., and R. Kolter. 1997. Protection of DNA during oxidative stress by the nonspecific DNA-binding protein Dps. J. Bacteriol. 179:5188-5194. [PMC free article] [PubMed]
29. McFall, E., and E. B. Newman. 1996. Amino acids as carbon sources, p. 358-379. In F. C. Neidhardt et al. (ed.), Escherichia coli and Salmonella: cellular and molecular biology, 2nd ed., vol. 1. ASM Press, Washington, D.C.
30. Miller, J. H. 1992. A short course in bacterial genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
31. Model, P., G. Jovanovic, and J. Dworkin. 1997. The Escherichia coli phage-shock-protein (psp) operon. Mol. Microbiol. 24:255-261. [PubMed]
32. Notley-McRobb, L., T. King, and T. Ferenci. 2002. rpoS mutations and loss of general stress resistance in Escherichia coli populations as a consequence of conflict between competing stress responses. J. Bacteriol. 184:806-811. [PMC free article] [PubMed]
33. Padan, E., and S. Schuldiner. 1994. Molecular physiology of the Na+/H+ antiporter in Escherichia coli. J. Exp. Biol. 196:443-456. [PubMed]
34. Reitzer, L., and B. L. Schneider. 2001. Metabolic context and possible physiological themes of σ54-dependent genes in Escherichia coli. Microbiol. Mol. Biol. Rev. 65:422-444. [PMC free article] [PubMed]
35. Robey, M., A. Benito, R. H. Hutson, C. Pascual, S. F. Park, and B. M. Mackey. 2001. Variation in resistance to high hydrostatic pressure and rpoS heterogeneity in natural isolates of Escherichia coli O157:H7. Appl. Environ. Microbiol. 67:4901-4907. [PMC free article] [PubMed]
36. Schellhorn, H. E., and H. M. Hassan. 1988. Transcriptional regulation of katE in Escherichia coli K-12. J. Bacteriol. 170:4286-4292. [PMC free article] [PubMed]
37. Slonczewski, J. L., B. P. Rosen, J. R. Alger, and R. M. Macnab. 1981. pH homeostasis in Escherichia coli: measurement by 31P nuclear magnetic resonance of methylphosphonate and phosphate. Proc. Natl. Acad. Sci. USA 78:6271-6275. [PMC free article] [PubMed]
38. Small, P., D. Blankenhorn, D. Welty, E. Zinser, and J. L. Slonczewski. 1994. Acid and base resistance in Escherichia coli and Shigella flexneri: role of rpoS and growth pH. J. Bacteriol. 176:1729-1737. [PMC free article] [PubMed]
39. Steinhaus, E. A., and J. M. Birkeland. 1939. Studies on the life and death of bacteria. I. The senescent phase in aging cultures and the probable mechanisms involved. J. Bacteriol. 38:249-261. [PMC free article] [PubMed]
40. Sutton, A., R. Buencamino, and A. Eisenstark. 2000. rpoS mutants in archival cultures of Salmonella enterica serovar Typhimurium. J. Bacteriol. 182:4375-4379. [PMC free article] [PubMed]
41. Swords, W. E., B. M. Cannon, and W. H. Benjamin, Jr. 1997. Avirulence of LT2 strains of Salmonella typhimurium results from a defective rpoS gene. Infect. Immun. 65:2451-2453. [PMC free article] [PubMed]
42. Testerman, T. L., A. Vazquez-Torres, Y. Xu, J. Jones-Carson, S. J. Libby, and F. C. Fang. 2002. The alternative sigma factor sigma E controls antioxidant defenses required for Salmonella virulence and stationary-phase survival. Mol. Microbiol. 43:771-782. [PubMed]
43. Verdu, E., F. Viani, D. Armstrong, R. Fraser, H. H. Siegrist, B. Pignatelli, J. P. Idstrom, C. Cederberg, A. L. Blum, and M. Fried. 1994. Effect of omeprazole on intragastric bacterial counts, nitrates, nitrites, and N-nitroso compounds. Gut 35:455-460. [PMC free article] [PubMed]
44. Vulic, M., and R. Kolter. 2001. Evolutionary cheating in Escherichia coli stationary phase cultures. Genetics 158:519-526. [PMC free article] [PubMed]
45. Waterman, S. R., and P. L. Small. 1996. Characterization of the acid resistance phenotype and rpoS alleles of shiga-like toxin-producing Escherichia coli. Infect. Immun. 64:2808-2811. [PMC free article] [PubMed]
46. Weiner, L., and P. Model. 1994. Role of an Escherichia coli stress-response operon in stationary-phase survival. Proc. Natl. Acad. Sci. USA 91:2191-2195. [PMC free article] [PubMed]
47. Wilmes-Riesenberg, M. R., J. W. Foster, and R. Curtiss III. 1997. An altered rpoS allele contributes to the avirulence of Salmonella typhimurium LT2. Infect. Immun. 65:203-210. [PMC free article] [PubMed]
48. Zambrano, M. M., and R. Kolter. 1996. GASPing for life in stationary phase. Cell 86:181-184. [PubMed]
49. Zambrano, M. M., D. A. Siegele, M. Almiron, A. Tormo, and R. Kolter. 1993. Microbial competition: Escherichia coli mutants that take over stationary phase cultures. Science 259:1757-1760. [PubMed]
50. Zhao, G., P. Ceci, A. Ilari, L. Giangiacomo, T. M. Laue, E. Chiancone, and N. D. Chasteen. 2002. Iron and hydrogen peroxide detoxification properties of DNA-binding protein from starved cells. A ferritin-like DNA-binding protein of Escherichia coli. J. Biol. Chem. 277:27689-27696. [PubMed]
51. Zimmer, D. P., E. Soupene, H. L. Lee, V. F. Wendisch, A. B. Khodursky, B. J. Peter, R. A. Bender, and S. Kustu. 2000. Nitrogen regulatory protein C-controlled genes of Escherichia coli: scavenging as a defense against nitrogen limitation. Proc. Natl. Acad. Sci. USA 97:14674-14679. [PMC free article] [PubMed]
52. Zinser, E. R., and R. Kolter. 1999. Mutations enhancing amino acid catabolism confer a growth advantage in stationary phase. J. Bacteriol. 181:5800-5807. [PMC free article] [PubMed]
53. Zinser, E. R., and R. Kolter. 2000. Prolonged stationary-phase incubation selects for lrp mutations in Escherichia coli K-12. J. Bacteriol. 182:4361-4365. [PMC free article] [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...