Properties of an L-glutamate-induced acid tolerance response which involves the functioning of extracellular induction components


Professor R.J. Rowbury, Department of Biology, University College London, Gower Street, London WC1E 6BT, UK.


Escherichia coli became more acid tolerant following incubation for 60 min in a medium containing l-glutamate at pH 7·0, 7·5 or 8·5. Several agents, including cAMP, NaCl, sucrose, SDS and DOC, prevented tolerance appearing if present with l-glutamate. Lesions in cysB, hns, fur, himA and relA, which frequently affect pH responses, failed to prevent l-glutamate-induced acid tolerance but a lesion in l-glutamate decarboxylase abolished the response. Induction of acid tolerance by l-glutamate was associated with the accumulation in the growth medium of a protein (or proteins) which was able to convert pH 7·0-grown cultures to acid tolerance, and the original l-glutamate-induced tolerance response was dependent on this component(s). Acid tolerance was also induced by l-aspartate at pH 7·0 and induction of such tolerance was dependent on an extracellular protein (or proteins). The l-glutamate and l-aspartate acid tolerance induction processes are further examples of a number of stress tolerance responses which differ from most inductions in that extracellular components, including extracellular sensors, are required.


It has become clear that a major factor influencing the ability of potentially pathogenic enterobacteria, ingested in food and water, to cause disease is their acid tolerance. This property is important because gastric acidity must be resisted by incoming organisms; even if there is successful passage through the stomach, the organism will be exposed to milder acidity in the presence of weak acids in the upper intestine ( Lee & Gemmell 1972) while if it is phagocytosed, it will be subjected to low pH in the phagolysosome ( Jensen & Bainton 1973).

Whether an ingested organism has the acid tolerance to resist the above challenges is governed, first, by the level of inherent acid tolerance ( Humphrey et al. 1995 ) and, secondly, by whether or not any of a range of acid tolerance responses have been induced. Many such responses are known to occur in enterobacteria ( Cooper & Rowbury 1986; Rowbury et al. 1989 ; Foster & Hall 1990; Lee et al. 1994 ) but the majority are induced only under mildly acidic conditions and would not be present in organisms which have been growing at neutral pH prior to ingestion. As most organisms would not have experienced prior exposure to acidity, induced tolerance responses occurring at neutral pH are of particular interest, especially if the responses can be induced by metabolites commonly found in foods. Recently, it has been established that some agents induce acid tolerance at neutral pH when present at concentrations similar to those found in foods. The present study describes acid tolerance induced at neutral pH by l-glutamate; this tolerance is of particular applied importance because the acid (as the sodium salt) is frequently present in a wide range of foods, especially oriental foods, many being of neutral pH.

Where inducible processes are switched on by chemical stimuli, these generally enter the cell or a cell compartment where they interact with a sensor to set in train the series of reactions which lead to the response. Once the sensor has been activated, all the reactions and components involved in induction are believed to be intracellular ( Stock et al. 1989 ; Neidhardt et al. 1990 ). The present study reports that the acid tolerance response induced by l-glutamate, and one induced by l-aspartate, differ from the above inducible systems in that extracellular components are involved in induction, and the sensors are extracellular.

Materials and methods

Bacterial strains and growth conditions

The strains used were all derivatives of Escherichia coli K12 and were as follows: strain 1829 (trp, lac), strain 1829 ColV, I-K94 (1829 ColV), strain 1829-TetR1 (1829 himA), strain 1829-AmpR1 (1829 hns), strain 1829-KanR1 (1829 fur), strain 1829-KanR2 (1829 relA), strain JA199 (trpE5, leu-6, thi, hsdR), strain NK1 (JA199 cysB), strain PA309 (thr-1, leuB6, fhuA2, lacY1, glnV44, gal-6, trp-1, hisG1, rfbD1, galP63, Δ (gltB-gltF) 500, rpsL9, malT1, xylA7, mtlA2, ΔargH1, thi-1) and strain DE70 (PA309 gadA, gadR).

Strains were grown throughout in Oxoid no. 2 broth (25 g l−1) at 37 °C and mid log-phase cultures (grown to 7–10 × 107 ml−1) were used in all experiments.

Induction of acid tolerance by L-glutamate and other agents

For induction of tolerance, such log-phase cultures were incubated at the stated pH with l-glutamate (250 mg l−1) present or absent; in a few experiments, the inducer was l-aspartate (250 mg l−1). Incubation was for 60 min with potential induction inhibitors present where stated. In the studies of the role of extracellular components in induction, as many of the components so far demonstrated are proteins, some incubations had protease (2 units ml−1 of Sigma P4531 protease) added to destroy extracellular proteins.

Preparation and use of medium filtrates

Mid log-phase organisms of the stated strains were grown in pH 7·0 broth and diluted into fresh pH 7·0 broth with l-glutamate, l-aspartate and any other additions present where stated. After 120 min incubation, cultures were filtered through Acrodisc 0·2 μm pore size syringe filters (Gelman Sciences, Ann Arbor, MI, USA) and dialysed overnight against 50 volumes of 0·065 mmol l−1 phosphate buffer pH 7·0 to remove the inducing metabolites. Prior to use, some filtrates were treated with protease (Sigma P4531 protease on beaded agarose, 2 units ml−1 for 30 min at 30 °C), with excess protease being removed by filtration through glass fibre filters (Gelman Sciences). Dialysed filtrates (one part filtrate to one part culture) were mixed with log-phase pH 7·0-grown strain 1829 ColV cultures (approximately 3 × 107 organisms ml−1) and incubated for 45 min at pH 7·0 prior to washing of the cultures and acid challenge.

Acid challenge

After washing with 3 volumes of pH 7·0 broth, organisms suspended at approximately 3 × 107 ml−1 were incubated for the stated time period in broth at pH 3·0. Challenged cultures were neutralized, plated on nutrient agar (no. 2 broth plus Difco granulated agar 2% w/v) and colonies enumerated after 20 h at 37 °C.


Log-phase strain 1829 ColV became markedly acid tolerant on exposure for 60 min to l-glutamate at pH 7·0. Analysis of the results of seven experiments indicated that organisms grown in broth without additions showed 0·83 ± 0·05 S.E.M.% survival after acid challenge for 7 min at pH 3·0, whereas after the same challenge, organisms pre-incubated in broth at pH 7·0 for 60 min with l-glutamate, 250 mg l−1, showed 23·5 ± 1·3% survival, i.e. l-glutamate significantly increased the acid tolerance of strain 1829 ColV (99·9% confidence). The extent of tolerance induction was greatest at pH 7·0, but increased acid tolerance also appeared at pH 7·5, 8·0 and 8·5 ( Table 1).

Table 1. Properties of the l-glutamate-induced acid tolerance response
Presence of inducerPotential inhibitor added during incubation Incubation pH Acid tolerance of cells (% survival after challenge)
  1. Log-phase cultures of strain 1829 ColV were transferred to fresh broth at the stated pH and with any indicated additions. Incubation was for 60 min, after which organisms were washed and acid challenged (pH 3·0 for 7 min). l-glutamate, where present, was added at 250 mg l−1.

  2. Dieth, diethanolamine.

  3. Results are shown for a single representative experiment but all experiments were repeated three times with consistent results.

NoNone7·0 1·2
Yes ( l-glutamate) None7·027·4
NoNone7·5 0·8
Yes ( l-glutamate) None7·522·0
NoNone8·0 0·8
Yes ( l-glutamate) None8·025·0
NoNone8·5 1·6
Yes ( l-glutamate) None8·516·4
Yes ( l-glutamate) Nalidixic acid (15 mg l−1) 7·026·3
Yes ( d-glucose) None7·032·5
Yes ( d-glucose) Nalidixic acid (15 mg l−1) 7·0 5·4
Yes ( l-glutamate) Dieth (30 mmol l−1) 7·029·0

Counteracting L-glutamate-induced acid tolerance

A range of metabolites and other agents was tested for their effects on tolerance induction by l-glutamate. Cyclic AMP, NaCl, sucrose, sodium dodecyl sulphate (SDS), sodium deoxycholate (DOC) and phosphate markedly reduced or abolished the tolerance induced by glutamate. Several other tested compounds (e.g. diethanolamine) had little or no effect ( Table 2). Also, whereas sodium benzoate added during induction at 15 mmol l−1 had no effect, this acid abolished the acid tolerance shown by l-glutamate-induced cells when it was present during acid challenge, i.e when challenged plus benzoate (15 mmol l−1), l-glutamate-induced cells showed no better acid tolerance than did uninduced cells.

Table 2. Agents which reverse or counteract acid tolerance induction by l-glutamate or l-asparate
Acid tolerance (% survival after pH 3·0 for 7 min) for organisms induced with
Potential inhibitor present during tolerance induction l-glutamate l-aspartate
  1. Log-phase organisms of strain 1829 ColV were transferred to fresh broth with l-glutamate (250 mg l−1) or l-aspartate (250 mg l−1) and any other additions where indicated. Incubation was for 60 min followed by washing and acid challenge of cultures. Results are shown for a single representative experiment but all experiments were repeated three times with consistent results. None of the agents added affected viability at the concentrations used.

cAMP (3 mmol l−1) 1·7 1·0
NaCl (300 mmol l−1) 0·9 1·2
Ethanol (10% v/v)20·020·0
Sucrose (365 mmol l−1) 1·0 0·6
Urea (300 mmol l−1) 20·0 6·5
Phosphate (10 mmol l−1) 8·112·6
SDS (90 mg l−1) 0·9 2·3
DOC (270 mg l−1) 1·218·0

Nalidixic acid and acid tolerance induction

Many inducible stress responses are inhibited by the DNA gyrase inhibitor, nalidixic acid, and this is often taken to indicate that induction is dependent on changes in DNA supercoiling ( Ni Bhriain et al. 1989 ). Such inhibition occurs for the recently reported acid tolerance response induced at pH 7·0 by glucose ( Rowbury & Goodson 1998b). In contrast, nalidixic acid had no inhibitory effect on induction of the l-glutamate-induced acid tolerance ( Table 1).

Regulatory components and tolerance induction by L-glutamate

Several pH responses have been shown to require the ferric uptake regulator (Fur; Foster & Hall 1990), the integration host factor (IHF, encoded by the him genes), H-NS, CysB ( Shi et al. 1993 ; Shi & Bennett 1994; Hassani et al. 1995 ) and the RelA gene product ( Rowbury 1997). Accordingly, mutants altered in the genes encoding these regulatory components were tested for ability to show acid tolerance induction by l-glutamate. None of the lesions prevented such induction ( Table 3).

Table 3. Regulatory components and l-glutamate-induced acid tolerance
Percentage survival afterchallenge (pH 3·0 for 7 min) for organisms incubated for tolerance induction
Strain tested for acid tolerance minus l-glutamate plus l-glutamate
Strain 18291·322·5
Strain 1829-AmpR1 (hns) 2·016·6
Strain 1829-TetR1 (himA) 0·910·5
Strain 1829-KanR1 (fur) 1·527·0
Strain 1829-KanR2 (relA) 0·7 9·4
Log-phase cultures of the stated strains were incubated with l-glutamate (250 mg l−1) present or absent. After 60 min incubation, organisms were washed and acid challenged.

Involvement of l-glutamate decarboxylase in L-glutamate-induced acid tolerance

Strain PA309, which contains l-glutamate decarboxylase, was able to induce the l-glutamate acid tolerance response. Thus, this strain grown in pH 7·0 broth to mid log-phase showed 0·4% survival after 7 min at pH 3·0 but after induction by l-glutamate (250 mg l−1) for 60 min, survival of organisms with the same challenge was 23·0%. In contrast, the PA309 derivative, strain DE70, which lacks one form of this decarboxylase (encoded by the gadA gene), was inherently more acid-sensitive than strain PA309, pH 7·0 broth-grown organisms of the mutant showing 0·03% survival after exposure to pH 3·0 for 7 min. Also, strain DE70 failed to show appreciable acid tolerance induction by l-glutamate, organisms grown with this amino acid (250 mg l−1), giving only 0·08% survival after the same acid challenge.

The role of extracellular induction components in L-glutamate-induced acid tolerance

It has recently been established that induction of several stress tolerance responses involves the functioning of extracellular components ( Hussain et al. 1998 ; Rowbury & Goodson 1998a, b), i.e. components in medium filtrates. Accordingly, the possible role of medium filtrates in l-glutamate-induced acid tolerance has been studied. For these tests, log-phase organisms of strain 1829 ColV were incubated at pH 7·0 with l-glutamate present or absent for 120 min; medium filtrates were then prepared and dialysed to remove glutamate. When the resulting filtrates were added to log-phase pH 7·0-grown strain 1829 ColV with incubation for 45 min, organisms incubated with the filtrates from l-glutamate-exposed cultures ( l-glutamate filtrates) became acid tolerant whereas organisms with filtrates from control cultures (pH 7·0-grown without l-glutamate) did not ( Table 4). The active component from the l-glutamate filtrates is a protein (or proteins), as protease almost completely inactivated such filtrates ( Table 4). The sensor for this response is extracellular.

Table 4. Filtrates from l-glutamate- and l-aspartate-induced cultures convert pH 7·0-grown organisms to acid tolerance at pH 7·0
Source of filtrate added to pH 7·0-grown organisms Filtrate treated with protease Percentage survival after challenge for organisms incubated ± filtrate
  1. Log-phase strain 1829 ColV cultures were transferred to fresh pH 7·0 broth with l-glutamate (250 mg l−1) or l-aspartate (250 mg l−1) present where stated. After 120 min incubation, filtrates were prepared (see Materials and methods) and where stated, treated with protease (2 units ml−1) for 30 min at 30 °C followed by removal of excess protease by filtration. Log-phase strain 1829 ColV culture (about 3 × 107ml −1) was mixed with dialysed filtrate (one part filtrate to one part culture) and the mixture incubated for 45 min at pH 7·0 prior to washing and acid challenge (pH 3·0 for 7 min).

  2. N.A., not applicable.

  3. Dialysis removed glutamate or aspartate from filtrates as control incubations containing these amino acids plus broth (but no cells) failed to induce any acid tolerance after dialysis.

No filtrateN.A. 1·0
From cells incubated with  l-glutamate No23·5
From cells incubated with  l-glutamate Yes 3·1
From cells incubated with  l-aspartate No15·6
From cells incubated with  l-aspartate Yes 1·1

The l-glutamate filtrate protein(s) which induces acid tolerance in pH 7·0-grown cells is needed for the original l-glutamate-induced acid tolerance response as protease added to organisms exposed to l-glutamate reduced tolerance induction ( Table 5).

Table 5. Involvement of extracellular components in l-glutamate- and l-aspartate-induced acid tolerance
Organisms induced to acid tolerance by additions of Inhibitory agent  present during  tolerance induction Percentage survival after acid challenge (pH 3·0 for 7 min)
  1. Log-phase organisms of strain 1829 ColV grown at pH 7·0 were transferred to fresh pH 7·0 broth with the stated inducer and with protease (2 units ml−1) present where stated. Some incubations were in a two compartment vessel (VectaSpin 20 polysulphone unit) with the culture plus inducer in the inner compartment and seven volumes of pH 7·0 broth plus inducer in the outer compartment (the compartments being separated by a 0·2 μm pore filter so that any extracellular components produced by the cultures were removed by filtration). Incubation was for 60 min and cultures were then washed and acid challenged.

NoneNone 1·1
l-glutamate (250 mg l−1) None29·2
l-glutamate (250 mg l−1) Protease 7·8
l-aspartate (250 mg l−1) None27·7
l-aspartate (250 mg l−1) Protease 1·8

Acid tolerance induced by L-aspartate and the role of extracellular induction components

Another amino acid that induces acid tolerance at pH 7·0 is l-aspartate. Thus, analysis of the results of seven experiments showed that organisms of strain 1829 ColV grown in broth at pH 7·0 gave 0·73 ± s. e. m. 0·13% survival after exposure to pH 3·0 for 7 min. In contrast, organisms incubated in pH 7·0 broth with l-aspartate 250 mg l−1 for 60 min showed 15·2 ± 0·76% survival after the same challenge. The aspartate-induced acid tolerance was counteracted by several agents ( Table 2), although deoxycholate had no effect.

Experiments analogous to those described above for l-glutamate showed that medium filtrates from l-aspartate-exposed cultures can induce acid tolerance, a protein induction component(s) being involved ( Table 4). This protein(s) is also needed for the original l-aspartate-induced acid tolerance response as protease almost abolished this response ( Table 5). The aspartate sensor is extracellular.


The l-glutamate-induced acid tolerance described here is one of a number of such tolerance responses known to be induced in log-phase organisms at neutral pH. That induced by glucose (the glucose response) has been studied in some detail ( Rowbury & Goodson 1998b) but the present response, and that induced by glucose, are quite distinct in that the former response is markedly more sensitive to inhibition by cAMP than is the latter, while the glucose response is sensitive to inhibition by nalidixic acid (Nal) whereas the l-glutamate response is not ( Table 1). The absence of any effect of Nal on the l-glutamate response suggests that induction is not appreciably affected by the extent of supercoiling in the region of the DNA encoding the tolerance genes. Many inducible responses are affected by Nal ( Ni Bhriain et al. 1989 ) and this applies to some pH responses ( Rowbury et al. 1996 ), but transcription of the l-glutamate response must be regulated differently from these.

It is of interest that metabolites and other agents which can be found in the environment, in foods or in the body can prevent the appearance of the acid tolerance responses ( Table 2) induced by l-glutamate or l-aspartate. Of particular interest is the finding that NaCl which can occur, often at high concentrations, in any of the above locations, was able to prevent tolerance induction. This could be of importance in oriental foods; potentially pathogenic contaminating organisms in any such foods which contain NaCl as well as l-glutamate might not be induced to tolerance.

It is clear that the l-glutamate-induced response can only be mounted if the gadA encoded form of l-glutamate decarboxylase is present, the absence of this enzyme from strain DE70 virtually abolishing induction. One possibility is that cells become acid tolerant plus l-glutamate because the products of the activity of the decarboxylase (which is induced plus l-glutamate) allow them to partially neutralize potentially lethal levels of acidity. The idea that amino acid decarboxylases can, in this way, partially neutralize low pH effects was originally proposed by Gale & Epps (1942) while recently, Meng & Bennett (1992) have studied the involvement of l-lysine decarboxylase and l-arginine decarboxylase in relieving the effects of acidity on growth.

The most striking characteristic of the l-glutamate-induced response (and of that induced by l-aspartate) is that induction is associated with the appearance in the growth medium of components able to convert organisms to acid tolerance at pH 7·0, i.e. under normally non-inducing conditions. Most strikingly, these extracellular components are essential for the original responses as if they are destroyed (using protease) during induction, the responses fail to appear. Such behaviour is quite different to that shown by classical inducible processes in that the latter supposedly involve exclusively intracellular reactions and components ( Neidhardt et al. 1990 ). In contrast, several stress tolerance responses are now known to involve the obligate functioning of extracellular induction components and extracellular sensors ( Rowbury & Goodson 1998a, b; Rowbury & Hussain 1998). Stress tolerance responses are possibly an unusual group of inducible processes, but if not, ideas on how inducible processes in general are switched on may have to be re-examined.