The findings reviewed here overturn a major tenet of bacterial physiology, namely that stimuli which switch-on inducible responses are always detected by intracellular sensors, with all other components and stages in induction also being intracellular. Such an induction mechanism even applies to quorum-sensed responses, and some others which involve functioning of extracellular components, and had previously been believed to occur in all cases. In contrast, for the stress responses reviewed here, triggering is by a quite distinct process, pairs of extracellular components being involved, with the stress sensing component (the extracellular sensing component, ESC) and the signalling component, which derives from it and induces the stress (the extracellular induction component, EIC), being extracellular and the stimulus detection occurring in the growth medium. The ESCs and EICs can also be referred to as extracellular sensing and signalling pheromones, since they are not only needed for induction in the stressed culture, but can act as pheromones in the same region activating other organisms which fail to produce the extracellular component (EC) pair. They can also diffuse to other regions and there act as pheromones influencing unstressed organisms or those which fail to produce such ECs. The cross-talk occurring due to such interactions, can then switch-on stress responses in such unstressed organisms and in those which cannot form the ESC/EIC pair. Accordingly, the ESC/EIC pairs can bring about a form of intercellular communication between organisms. If the unstressed organisms, which are induced to stress tolerance by such extracellular components, are facing impending stress challenge, then the pheromonal activities of the ECs provide an early warning system against stress. The specific ESC/EIC pairs switch-on numerous responses; often these pairs are proteins, but non-protein ECs also occur and for a few systems, full induction needs two ESC/EIC pairs. Most of the above ECs needed for response induction are highly resistant to irreversible inactivation by lethal agents and conditions and, accordingly, many killed cultures still contain ESCs or EICs. If these killed cultures come into contact with unstressed living organisms, the ECs again act pheromonally, altering the tolerance to stress of the living organisms. It has been claimed that bacteria sense increased temperature using ribosomes or the DnaK gene product. The work reviewed here shows that, for thermal triggering of thermotolerance and acid tolerance in E. coli, it is ESCs which act as thermometers.

Escheichia 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 DOG, 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.

During growth of Escherichia coli roll in broth at pH 5.0, an extracellular protein termed an extracellular induction component (EIC) appears in the medium, this component being essential for acid tolerance induction. The present study establishes that the EIC arises from an extracellular precursor which is formed during growth at pH 7.0, and that conversion of the precursor to the EIC occurs at pH 5.0 (and other mildly acidic pH values) in the absence of organisms. On the basis that this precursor is produced by non-stressed cells as well as by stressed ones, and that it is converted to the EIC (which in turn induces the tolerance response) by the stress, the precursor can be considered to be a stress sensor, the first extracellular stimulus sensor to be reported. The EIC formed at pH 5.0 was inactivated at pH 9.0. This inactivation probably involved conversion back to the precursor as EIC was reformed if the alkali-inactivated component was incubated at pH 5.0. Both mild heat treatments (exposure to 40-55 degrees C) and u.v. irradiation also activated the precursor; the active induction component formed by the mild heat treatments was reversibly inactivated at pH 9.0 and so it seems likely that the component formed by heat treatment is similar or identical to the EIC produced at acidic pH. In contrast, the EIC produced by u.v. irradiation was not inactivated at pH 9.0, suggesting that it is different in some way to the EICs produced from the precursor by acidity or by heat treatment. It is likely that many responses affecting stress tolerance involve the functioning of such extracellular sensors, as similar components were shown to be involved in the acid tolerance responses induced at pH 7.0 by glucose, L-aspartate and L-glutamate. Extracellular stimulus sensors may also be needed for other inducible responses.

An extracellular induction component (EIC), needed for acid tolerance induction at pH 5.0 in Escherichia coli, arises from an extracellular precursor which senses acid stress and is activated (forming the EIC) by such stress. The precursor, which is a heat-stable protein, was formed by cells which had not been subjected to acid stress, being present in culture media after growth at pH values from 7.0 to 9.0. This stress-sensing molecule was activated to the EIC at pH values from 4.5 to 6.0 but not at pH 6.5 and did not form EIC on incubation at an extremely acidic pH e.g. 2.0. The precursor was not inactivated at pH 2.0. Precursor activation might be reversible, as the EIC lost its ability to induce acid tolerance after incubation at pH 9.0, but regained it if subsequently incubated at pH 5.0. Whereas the sensor formed at pH 7.0 can only be activated at pH 5.0 to 6.0, that synthesized at pH 9.0 can be activated at pH 5.0 to 7.5. Accordingly, this work shows that the acid stress sensor is extracellular, and it is proposed that its presence in the medium rather than in the cells, allows more sensitive and rapid responses to acid stress.

Escherichia coli shifted from broth at external pH (pH0) 7·0 to pH0 7·0 broth plus glucose rapidly induced marked acid tolerance which also appeared, albeit to a lesser extent, plus maltose, sucrose or lactose. Tolerance appeared without the medium pH becoming acidic. Tolerance was most substantial when glucose was added at pH0 7·0 but was also appreciable at pH0 7·5, 8·0 and 8·5. Induction of tolerance by glucose was markedly reduced by cyclic AMP and essentially abolished plus NaCl or sucrose; the induction process was also reduced but not fully inhibited by chloramphenicol, tetracycline and nalidixic acid. Glucose‐induced organisms showed less acid damage to DNA and β‐galactosidase and it is likely that this is because glucose induces a new pH homeostatic mechanism which keeps internal pH close to neutrality at acidic pH0. In conclusion, it is clear that glucose induces a novel acid tolerance response in log‐phase E. coli at pH0 7·0; it is now known that induction of this response involves the functioning of extracellular induction components including an extracellular induction protein.

Several acid tolerance responses of Escherichia coli were associated with secretion into the growth media of components (frequently proteins) which altered acid tolerance of other cultures. First, medium filtrates from cultures induced to acid tolerance by several conditions converted pH 7.0-grown organism to tolerance and, for most such responses, filtrate proteins were needed for full induction. Secondly, filtrates from cultures induced to acid sensitivity at alkaline pH produced sensitisation of resistant cultures. Thirdly, filtrates from inherently tolerant or sensitive strains altered tolerance or sensitivity of normal strains. In many cases, filtrate components were essential for the original response, e.g. acid habituation at pH 5.0. Extracellular components may function as intermediates only in stress tolerance responses, but other adaptive responses must be tested as such components may function in other inducible processes.

N-acyl-L-homoserine lactones are involved as intercellular signalling agents controlling a wide range of physiological responses in Gram-negative bacteria. They function especially in vibrios, pseudomonads and erwinias as well as in Rhizobium and Agrobacterium spp, particularly where the bacteria are in symbiotic or parasitic relationships with higher organisms. Several Gram-negatives, such as Escherichia coli, do not, however, appear to produce or respond to AHLs and they may have other intercellular signalling molecules. The present review reports that several stress related responses in E. coli can be induced by supernatant fluids from cultures which have already induced the response. In some cases at least, the active agents in the supernatant fluids are proteins rather than AHLs.

Escherichia coli grown at pH 5.0 became acid-tolerant (acid-habituated) but, in addition, neutralized medium filtrates from cultures of E. coli grown to log-phase or stationary-phase at pH 5.0 (pH 5.0 filtrates) induced acid tolerance when added to log-phase E. coli growing at pH 7.0. In contrast, filtrates from pH 7.0-grown cultures were ineffective. The pH 5.0 filtrates were inactivated by heating in a boiling water-bath but there was less activity loss at 75 degrees C. Protease also inactivated such filtrates, which suggested that a heat-resistant protein (or proteins) in the filtrates was essential for the induction of acid tolerance. Filtrates from cells grown at pH 5.0 plus phosphate or adenosine 3':5'-cyclic monophosphate (cAMP) were much less effective in inducing acid tolerance, while the conversion of pH 7.0-grown log-phase cells to acid tolerance by pH 5.0 filtrates was inhibited by cAMP and bicarbonate. It seems likely that the acid tolerance response (acid habituation) involved the functioning of the extracellular protein(s) as protease reduces tolerance induction if added during acid habituation. Most inducible responses are believed to involve the functioning of only intracellular reactions and components; the present results suggest that this is not the case for acid habituation, as an extracellular protein (or proteins) is needed for induction.

The evaluation of acute emesis in a multiple-trauma paediatric patient with severe traumatic brain injury can present a difficult but not unusual problem in the rehabilitation unit. The differential diagnosis will be discussed in a paediatric case in which emesis was related to a urinoma, secondary to ureteropelvic disruption, compressing the duodenum. We present a case of a 6-year-old Caucasian girl with severe brain injury who developed emesis while in rehabilitation. An X-ray of the abdomen revealed a shadow over the right psoas. Subsequently mild compression of the duodenum was diagnosed by computerized tomography, and ultrasound demonstrated a large fluid collection inferior to the right kidney. A large volume of urine was drained and a catheter placed. Pyelogram revealed apparent disruption of the right ureteropelvic junction, which was confirmed intraoperatively. The kidney was removed secondary to severe parenchymal abnormalities. The emesis subsequently resolved. Differential diagnosis of acute onset of emesis in the paediatric brain-injured patient should include a careful search for underlying previously unrecognized problems.

Escherichia coli shifted from external pH (pH(O)) 7.0 to pH(O) 8.5-9.5 rapidly becomes tolerant to pH(O) 10.0-11.5, induction of tolerance (alkali habituation) being dependent on periplasmic or external alkalinization with either NaOH or KOH. Induction needs protein synthesis and makes organisms resistant to DNA damage by alkali and better able to repair any damage that occurs. Induction of tolerance was reduced by glucose (not reversed by cAMP) and by amiloride, was dependent on DNA gyrase and was abolished by fur and himA lesions (the latter suggests IHF involvement). Tolerance induction was not prevented by L-leucine, FeCl3 or FeSO4 nor by hns or relA mutations. Habituation probably involves attachment of IHF upstream of the promoter leading to DNA bending which switches on transcription. Habituation is aberrant in nhaA mutants, so ability to resist alkali damage may only arise if NhaA is induced, with extrusion of Na+ by this antiporter during alkali challenge. In accord with one tolerance component involving NhaA induction, beta-galactosidase formation from nhaA-lacZ fusions at pH(O) 9.0 was inhibited by glucose and amiloride.

NaCl-induced acid sensitivity is not a response to high osmotic pressure but is triggered by a high internal sodium ion concentration, evidenced by the finding that conditions which enhance Na+ influx or reduce Na+ efflux, led to increased internal Na+ and allowed induction at lower external sodium ion concentrations and that induction could occur with no osmotic upshock. NaCl-induced acid sensitivity was not observed in Escherichia coli strain MC4100 and its relA lesion was investigated as a possible cause. Transformation of this strain and two other relA mutants to relA+ allowed sensitization by NaCl and whereas strain CF1648 (relA+) was sensitized, its relA deletion derivative was not. Additionally, transduction of two relA+ strains to relA produced derivatives which were not sensitized by salt. Although MC4100 was not sensitized by NaCl, envZ derivatives of it were sensitized. Lesions in fur and tonB did not prevent sensitization by NaCl, although extent of sensitization was slightly increased by fur and considerably increased by tonB. Acetate strongly inhibited sensitization, which was also subject to glucose repression, reversible by cAMP. The possibility was discussed that (p)ppGpp and cAMP both positively affect transcription of the genes encoding the acid sensitization components; in accord with this, high concentrations of cAMP suppressed the effect of the relA lesion of sensitization. Salt-induced organisms are more sensitive to acid damage to DNA, to acid inhibition of enzyme synthesis and DNA transfer and slightly less able to repair acid-damaged transforming DNA.

Escherichia coli grown in low-salt broth (LSB) showed rapid induction of acid sensitivity when L-leucine was added. Induction was strongly inhibited by nalidixic acid and by tetracycline but only slightly by chloramphenicol and rifampicin. Regulation of the effect appeared to involve H-NS, CysB, the ferric uptake regulator (Fur) and (p)ppGpp; ompA mutants were also unable to mount the response. Acid sensitization was greatly reduced by glucose, ferric chloride and ferrous sulphate but was not affected appreciably by phosphate or amiloride. Lesions abolishing synthesis of integration host factor, the PhoE porin and NhaA antiporter, did not prevent sensitization by L-leucine and lrp lesions only slightly affected the response. Organisms exposed to L-leucine were less susceptible to phage K3. Acid sensitivity depends on appearance of a new outer membrane (OM) pathway for protons and L-leucine-induced organisms may use a new pore formed from a modified OmpA protein.

On solid medium, streaks of ColV(+) Escherichia coli strains produced much larger inhibitory zones on sensitive indicators at pH 7.0 and 9.0 than at pH 5.0. Such an acid pH effect also occurred in liquid medium, with sensitive strains being killed by ColV(+) ones at pH 7.0 or 9.0 but not at pH 5.0. Colicin V was responsible for the killing effect. Culture filtrates from PAP1401 (ColV(+)) were used to examine the basis for the pH effect and it was found that acid pH interfered directly or indirectly with colicin V production but that colicin V was not irreversibly inactivated at pH 5.0. Additionally, colicin V formed at pH 7.0 was less active when tested on sensitive strains at pH 5.0 than at pH 7.0. Neither reduced colicin synthesis nor its reduced activity can be attributed to increased availability of iron at pH 5.0.

Escherichia coli previously grown in low-salt broth, pH 7.0, produced organisms which were markedly more acid sensitive when subsequently cultured in the same broth with 200 mM or more salt (NaCl) added. Induction of acid sensitivity occurred rapidly at both 37 and 30 degrees C, with a substantial effect within 15 min. Sensitization was partially inhibited by chloramphenicol and tetracycline and may depend on both protein synthesis-dependent and -independent physiological changes in the NaCl-induced organisms; sensitization did not result from osmotic shocking on transfer to challenge medium. Induction of acid sensitivity was affected by neither the sodium ion pore inhibitor amiloride nor the DNA synthesis inhibitor nalidixic acid; rifampin had a small effect, similar to that of chloramphenicol. Chlorides of other monovalent cations, especially Li+ and NH4+, also produced sensitization to acid, although CsCl was ineffective but did not interfere with sensitization by NaCl. Other sodium salts were also active as sensitizers, as were chlorides of divalent cations, but although sucrose (but not glycerol) was a good inducer, the results were not fully in accord with triggering of induction solely by the NaCl-associated increase in osmotic pressure. Sensitization was not prevented by deletion of the nhaA, nhaR, or nhaB gene. Acid sensitivity of NaCl-induced cells was slightly reduced after 90 min of growth at 37 degrees C in low-salt broth but was completely lost after 240 min. For NaCl-induced cells, acid killing in challenge media was not inhibited by amiloride. The NaCl-induced sensitization is distinct from the phenomenon of acid sensitivity induction in E. coli at alkaline external pH.

The lethal effects of inorganic acid on phoE+ Escherichia coli strains, grown at neutral pH(o), were enhanced by chloramphenicol, apparently because some organisms acquire acid tolerance (habituate) during challenge and chloramphenicol stops this. Phosphate (and/or polyphosphate) present during challenge prevented killing and damage by acid to outer membranes, DNA and cellular enzymes but did not prevent acid pH(o) enhancing novobiocin activity. To reverse acid effects, phosphate must interact with or cross the outer membrane but need not enter the cytoplasm; it is probable that it competes with H+ (or protonated anions) for passage through the PhoE pore. Phosphate also prevented induction of beta-galactosidase in a strain with the cadA promoter fused to lacZ. Four unc mutants showed essentially normal acid sensitivity and habituation; the same was true for strains with lesions in fur, oxyR, katF, phoP, cadA and hycB. In contrast, deletion of rpoH led to slightly increased acid sensitivity for cells grown at pH(o) 7.0, although habituation was relatively normal.

Transfer of pH 7.0-grown Escherichia coli to pH 9.0 led to rapid acid sensitivity induction (ASI), the response being fully accomplished within 15 min at 37-degrees-C in broth. Only a slight increase in acid sensitivity occurred at pH 8.2 but the response was substantial at pH 8.4 and complete at pH 9.0 with no further sensitization at pH 9-5-10-5. ASI was not prevented by lesions in rpoH, katF, ompR, relA, spoT, fur, phoU, phoM (CreC), phoBIR, unc(atp), phoP or cadA and was unaffected by nalidixic acid, L-leucine or iron starvation or excess. Full acid sensitivity was maintained for at least 2 h after a shift from pH 9.0 back to pH 7-0. ASI did not depend to a major extent on PhoE derepression and increased acid sensitivity of alkali-induced strain C75a (phoE+) probably did not involve use of a new outer membrane proton pore.

Acid sensitivity induction (ASI) at alkaline pH in Escherichia coli 1157 (phoE) is only fully mounted if protein synthesis occurs for the first 5 min of the 12-15 min induction period, but appreciable sensitization occurs in the presence of chloramphenicol indicating that there are protein synthesis-independent and -dependent components (components 1 and 2). Component 1 sensitization is in place after 10 min at pH(o) 9.0 but the dependent process is induced slightly more slowly. Collapsing Delta pH at pH(o) 9.0 did not prevent full ASI, indicating that component 1 and 2 induction does not depend on increased Delta pH. The two induction (or activation) processes responded differently to intermediate levels of external alkalinization; component 1 was induced at pH(o) 8.3 but component 2 needed pH(o) 8.4 or greater. Collapsing Delta pH at pH(o) 8.0 led to full induction of component 1 and appreciable induction of component 2, indicating that both processes are triggered if pH(i) rises to 8.0 or greater (presumably at pH(o) 8.3-8.4). Component 1 appears less important for ASI in 1157-4(phoE(+)) than in 1157(phoE).

Escherichia coli K12 becomes resistant to killing by acid (habituates to acid) in a few minutes at pH 5.0. Habituation involves protein synthesis-dependent and -independent stages; both must occur at an habituating pH. The habituation sensor does not detect increased DELTA-pH (or decreased DELTA-psi) nor an increased difference between pH(o) and periplasmic pH but probably detects a fall in either external or periplasmic pH. Phosphate ions inhibit habituation, at any stage, probably by interfering with outer membrane passage of hydrogen ions. Most outer membrane components tested are not required for habituation but phoE deletion mutants habituated poorly and are acid-resistant. Strains derepressed for phoE, in contrast, showed increased acid sensitivity. These and other results suggest that habituation involves hydrogen ions or protonated carriers crossing the outer membrane preferentially via the PhoE pore, a process inhibited by phosphate and other anions. Stimulation by phosphate of the poor growth of E. coli at pH 5.0 is in accord with the above. Acetate did not enhance acid killing of pH 5.0 cells, suggesting that their resistance does not depend on maintaining pH(i) near to neutrality at an acidic pH(o) level.

Induction of acid resistance (habituation) in Escherichia coli at pH 5.0 took ca 5 min in broth at 37 degrees C and 30-60 min in minimal medium. Induction occurred at a range of pH values from 4.0 to 6.0; it was dependent on continuing protein and RNA synthesis but substantial acid resistance appeared in the presence of nalidixic acid. Acid resistance was long-lasting; organisms grown at pH 5.0 retained most of their resistance after 2 h growth at pH 7.0. Organisms grown at pH 5.0 showed increased synthesis of a number of cytoplasmic proteins compared with the level in cells grown at pH 7.0. DNA repair-deficient strains carrying recA, uvrA or polA1 mutations were more acid-sensitive than the repair-proficient parents but were able to habituate at pH 5.0. Organisms grown at pH 5.0 transferred the ColV plasmid much more effectively at acid pH than did those grown at pH 7.0 and habituated recipients appeared better able to repair incoming acid-damaged plasmid DNA than did those that were non-habituated. Induction of acid resistance at pH 5.0 may be significant for the survival of organisms exposed to periodic discharges of acid effluent in the aquatic environment and habituation may also allow plasmid transfer and repair of acid-damaged plasmid DNA during or after such exposure.

Growth of Escherichia coli 1829 ColV, I-K94 at pH 5.0 led to an increase in u.v. resistance compared with cells grown at pH 7.0. This was due to a phenotypic change, since organisms grown at pH 7.0 showed increased resistance after only 2.5-5.0 min incubation at the mildly acid pH. Other E. coli K12 derivatives became more u.v.-resistant at pH 5.0 including uvrA, recA and polA1 mutants. Organisms grown at pH 5.0 also showed increased Weigle reactivation of u.v.-irradiated lambda phage and this applied to the repair-deficient mutants as well as the parent strains. Both the increased u.v. resistance of acid-habituated cells and their increased ability to bring about Weigle reactivation appear to involve RecA-independent processes and are presumably, therefore, independent of the SOS response.