Escherichia coli wild type strains show biphasic "diauxic" growth when bacteria are grown in a medium containing glucose and any one of a variety of sugars. In the present work, it is reported that fructose can substitute for glucose in promoting diauxic growth in fruR stains lacking the fructose repressor FruR, also called Cra, the Catabolite repressor/activator protein. Addition of cAMP to the growth medium can often eliminate the diauxic lag observed before utilization of the second sugar commences, but in fruR strains, cAMP does not totally abolished the diauxic lag period. It is proposed that the pool of unphosphorylated Enzyme IIAGlc, a glucose-specific component of the phosphotransferase system, is higher in fructose-grown fruR strains than in glucose-grown wild-type strains. It is concluded that the key factor in producing diauxie is the ability of a given sugar to regulate indirectly, via its transport, the phosphorylation state of Enzyme IIAGlc, thereby controlling Enzyme IIAGlc secondary regulatory functions, i.e. regulation of adenylate cyclase and the phenomenon of inducer exclusion.
INTRODUCTIONEnteric bacteria respond efficiently to environmental stimuli by regulating gene expression thereby selecting appropriate metabolic pathways. The phenomenon of 'diauxie', extensively studied by Jacques Monod (1942) is exemplary (for reviews see Magasanik, 1970 and Roseman & Meadow, 1990). When bacteria are grown in a chemically defined medium containing two sugars, biphasic 'diauxic' growth is often observed. Monod classified sugars into class 'A' or 'B'. Class A sugars (mannose, fructose, mannitol) do not give rise to diauxic growth when present in the medium with glucose. By contrast, class B sugars (arabinose, maltose, rhamnose, lactose, xylose, galactose) do give rise to diauxic growth when present with glucose.
The discovery of the phosphoenolpyruvate (PEP)-dependent phosphotransferase system (PTS) by Saul Roseman and coworkers (Kundig et al., 1964) provided a starting point for an understanding of diauxie. The primary function of the PTS is to transport and concomitantly phosphorylate sugar substrates (PTS sugars). This process occurs through phosphoryl transfer from PEP and involves Enzyme I, HPr and the sugar-specific Enzyme II complexes. Class A sugars as defined by Jacques Monod generally correspond to PTS sugars, and an explanation for diauxic growth came with the discovery of secondary functions for the PTS, namely, regulation of adenylate cyclase (AC) (Saier & Feucht, 1975) and inducer exclusion (Saier & Roseman, 1976).
As regards AC regulation, Enzyme IIAGlc (IIAGlc), a glucose-specific component of the PTS, plays a major role (Feucht & Saier, 1980; Postma et al., 1981). When IIAGlc is phosphorylated due to the phosphoryl transfer activities of Enzyme I and HPr, it activates AC. However in the presence of exogenous glucose transport via the PTS occurs, leading to the production of cytoplasmic glucose 6-phosphate and a concomitant decrease in the concentration of phosphorylated IIAGlc, which causes a decrease in the rate of cAMP synthesis.
Class B sugars require the Catabolite gene Activator Protein (CAP) (Zubay et al., 1970), also called Cyclic AMP Receptor Protein (CRP) (Emmer et al., 1970), as well as cAMP for the synthesis of catabolic enzymes responsible for their degradation, the CAP-cAMP complex acting as a positive transcriptional regulator. In cyaA strains (lacking AC) or crp strains (lacking CAP), expression of catabolite-sensitive operons is dramatically reduced. In cyaA strains expression can be restored by addition of exogenous cAMP.
Glucose can prevent the expression of catabolite-sensitive operons not only by its effect on the intracellular cAMP level but also by exclusion of the operon-specific inducer from the cell. To exemplify with the E. coli lactose operon, lactose entry is prevented during glucose transport by direct interaction of dephospho-IIAGlc with the lactose permease (Osumi & Saier, 1982). The current belief is that both PTS-mediated regulatory mechanisms are required for full manifestation of diauxie.
In mutant strains lacking the fructose repressor FruR, discovered by Geerse et al. (1986) and Kornberg & Elvin (1987), also known as Cra, the "Catabolite repressor/activator" protein (Saier & Ramseier, 1996), IIAGlc is largely dephosphorylated when cells are grown in fructose-containing medium (Crasnier-Mednansky et al., 1997). In this communication, we analyze the physiological consequences of this phenomenon and report that, in fruR strains, fructose can substitute for glucose in producing diauxie.
MATERIALS and METHODSThe isogenic Escherichia coli K-12 strains used in this study are listed in the following table. The LJ strains can be obtained from Mary Berlyn, Coli Genetic Stock Center, Yale University.
| Strains | Genotype | Reference |
|---|---|---|
| TP2503 | F- xylA7 ilvA215 ΔargH1 | De Reuse & Danchin, 1988 |
| LJ2806 | F- xylA7 ilvA215 ΔargH1 fruR11::Tn10 | Crasnier-Mednansky et al., 1997 |
| TP9500 | F- xylA7 ΔargH1 ΔcyaA854 | Crasnier & Danchin, 1990 |
| LJ2809 | F- xylA7 ΔargH1 ΔcyaA854 fruR11::Tn10 | Crasnier-Mednansky et al., 1997 |
| LJ2807 | F- xylA7 ilvA215 ΔargH1 aroB Δcrr::kan fruR11::Tn10 | Crasnier-Mednansky et al., 1997 |
The growth medium for measuring diauxie was minimal medium 63 (MM63) (Miller, 1992) supplemented with the required amino acids (1 mM), thiamine (5 µg/ml) and a pair of sugars. When present, kanamycin was added at 50 µg/ml. The fruR strains were routinely checked for their sensitivity to 0.2 % xylitol (Reiner, 1977) as well as their inability to grow on lactate and pyruvate (Geerse et al., 1986).
β-Galactosidase assayβ-Galactosidase activity was measured in MM63 as described previously (Pardee et al., 1959). A unit of β-galactosidase activity was defined as the amount of enzyme that hydrolyzes 1 nmol of o-nitrophenyl β-D-galactoside per minute under standard conditions.
RESULTSIsogenic strains TP2503 (wild type) and LJ2806 (fruR) were grown in MM63 supplemented with 1.5 mM fructose and 5 mM lactose. In contrast to the wild-type strain, diauxic growth was observed for the fruR mutant strain, Figure 1. In the case of the wild-type strain, a slightly biphasic growth curve was observed probably due to the more rapid utilization of fructose as compared to lactose. Similar results were obtained with another set of isogenic cyaA and cyaA fruR strains (TP9500 and LJ2809) containing a low copy number plasmid (IncW) carrying the wild type cyaA gene (pDIA1900) (data not shown). These results are in agreement with recently published data showing (i) that cAMP levels are 10-fold reduced in fructose-grown fruR strains relative to fructose-grown wild-type strains, and (ii) that lactose exclusion by fructose occurs in fruR mutants (Crasnier-Mednansky et al., 1997). The explanation for both phenomena is that IIAGlc is largely dephosphorylated in fruR mutants but not in wild-type strains when bacteria are grown in the presence of fructose.
Effect of cAMPWhen wild type E. coli strains are grown in the presence of a mixture of a class A and B sugar, addition of cAMP to the growth medium can eliminate the diauxic lag observed before utilization of the class B sugar commences (Ullmann & Monod, 1968; Ullmann et al., 1969). The effect of cAMP on the diauxic growth behavior of fruR strains was investigated as follows. Strain LJ2809 (cyaA fruR) was transformed with plasmid pDIA1920 that encodes a 48 kDa truncated form of AC (half the size of wild-type AC) which is not subject to regulation by IIAGlc (Crasnier et al., 1994). This truncated form of AC catalyzes synthesis of increased amounts of cAMP as compared to the wild type enzyme, especially when the strains are grown on glucose. It has been reported that diauxic growth in the presence of glucose is abolished in strains containing the 48 kDa AC (Crasnier et al., 1994). As shown in Figure 2, when the fruR strain was grown in fructose plus lactose medium, the increased level of cAMP due to the presence of the 48 kDa AC only partially abolished the lag period, which was shortened from 110 minutes to 30 minutes. A comparable result was obtained when the growth medium of strain LJ2809 (cyaA fruR) was supplemented with 1 mM cAMP (data not shown).
When glucose-lactose instead of fructose-lactose growth was studied, the lag period of the fruR minus strain was substantially longer than that observed for the fruR plus strain, Figure 3-A and B, gray open squares. An increased level of cAMP (due to the presence of the 48 kDa AC) completely abolished diauxie in the fruR plus strain but not in the fruR minus strain, Figure 3-A and B, black open squares. In the fruR strain, the lag period was reduced from 90 minutes to 30 minutes. This observation suggests that loss of FruR function enhances glucose- as well as fructose-promoted repression of lac operon expression.
Lactose operon expression during diauxieIn order to estimate the levels of lac operon expression, β-galactosidase activities were measured during fructose-lactose growth of the wild type and fruR strains. As can be seen in Figure 4-A, β-galactosidase activity in the fruR mutant strain was not detectable until fructose had been exhausted from the growth medium and cells had entered into the diauxic lag phase. β-Galactosidase activity became appreciable only during the second phase of growth when lactose utilization occurred. Growth on lactose commenced when β-galactosidase activity reached about 500 units per mg dry weight of bacteria. In the case of the wild type strain, β-galactosidase activity increased linearly during growth (data not shown). Appreciable induction of the lactose operon in the wild type cells occurred even though fructose was used preferentially to lactose (as suggested by the slightly biphasic growth). Maximal β-galactosidase activities of wild type and fruR strains were similar (about 1700 units per mg).
β-Galactosidase activities were also measured during growth of the fruR strain in a fructose-lactose medium supplemented with cAMP (data not shown). Addition of cAMP to the growth medium did not appreciably promote synthesis of β-galactosidase during the first growth phase. Interestingly, a short lag remained and growth on lactose began when β-galactosidase activity approached about 500 units per mg dry weight, the same as reported above for the fruR mutant growing in the absence of exogenous cAMP. This is in accordance with expectation since the occurrence of a lag period is believed to correspond to the time necessary for induction of lactose operon expression.
In order to investigate the role of IIAGlc in promoting the lag period, a fruR crr strain (lacking both FruR and IIAGlc) was grown in fructose plus lactose medium supplemented with 1 mM cAMP. As can be seen in Figure 4-B, no lag was observed under such conditions indicating that IIAGlc is responsible for the occurrence of the lag period observed during growth of the fruR strain. In addition, β-galactosidase activities increased during the entire period of growth.
Diauxie in fruR strains with different pairs of sugarsMutant fruR strains were examined for their ability to generate diauxic growth in response to different pairs of sugars. MM63 was supplemented with a mixture of fructose and maltose or fructose and galactose. Not surprisingly, the fructose/maltose pair gave rise to the same biphasic, diauxic growth curve that was observed for the fructose/lactose pair shown in Figure 1 (data not shown). In the case of the fructose/galactose pair, diauxic growth was also observed, but cAMP exerted little or no effect on the growth curve (data not shown). This latter observation is in agreement with a previous report showing that cAMP does not seem to antagonize the glucose effects on galactose operon expression (Joseph et al., 1981).
Kornberg and Elvin (1987) reported that fruR strains, unlike wild type strains, preferentially utilize fructose over glucose. When growing on an equimolar mixture of these hexoses, fruR strains utilize approximately seven times more fructose than glucose. When fruR strains were analyzed for growth in the presence of 1.5 mM fructose plus 1.5 mM glucose, neither diauxic nor biphasic growth was observed, and both wild-type and fruR strains showed similar behavior (data not shown). Re-phosphorylation of IIAGlc after fructose has been utilized is probably fast enough not to be a limiting factor for glucose transport.
DISCUSSIONThe work described here is consistent with the conclusion that two conditions are required for diauxic growth: (1) a low level of cAMP and (2) inducer exclusion. Both conditions occur upon dephosphorylation of IIAGlc. Fructose-grown fruR strains show a low level of cAMP, slightly lower than that of glucose-grown wild type strains. Also inducer exclusion occurs in fructose-grown fruR strains (Crasnier-Mednansky et al., 1997). Consequently, fructose elicits diauxic growth in fruR strains.
An unexplained feature of diauxie is that, in wild type strains, some but not all PTS sugars can substitute for glucose in producing diauxie (Roseman & Meadow, 1990). For example, mannitol but not fructose can substitute for glucose in producing diauxie. Growth on mannitol produces a relatively low level of cAMP (Epstein et al., 1975; see also Crasnier-Mednansky et al., 1997 for β-galactosidase activities). Considering the role of cAMP in diauxie, the fact that mannitol but not fructose can substitute for glucose in producing diauxie is probably related to the relatively low level of cAMP produced by wild type strains growing on mannitol. It can be argued that the level of cAMP correlates with the level of phosphorylation of IIAGlc. A decrease in cAMP level is likely to correspond to a decrease in the amount of phosphorylated IIAglc, and consequently to stronger inducer exclusion. Therefore the ability of a given PTS sugar to promote diauxie is related to its ability to indirectly control the phosphorylation of IIAGlc, a feature which had been attributed largely to glucose.
When growth occurs in fructose minimal medium, net cellular amounts of IIAGlc (IIAGlc + P-IIAGlc) have been shown to be independent of the fruR mutation. However in fructose-grown fruR strains, the genes of the fructose operon are over-expressed as compared to fructose-grown wild-type strains, (Crasnier-Mednansky et al., 1997). The over-expressed fructose Enzyme II complex may prevent phosphorylation of IIAGlc by competitively inhibiting the phosphorylation of HPr by Enzyme I. In the case of mannitol, dephosphorylation of IIAGlc leading to a relatively low level of cAMP may result from competition between IIAGlc and IICBAMtl for phosphorylation by HPr. Therefore the level of expression of a PTS enzyme II complex may determine the level of phosphorylation of IIAGlc and consequently sensitivity to diauxie. If this is the case, any PTS sugar should be able to substitute for glucose in producing diauxie if its Enzyme II complex is sufficiently expressed.
However, the rate of phosphoryl transfer between PTS proteins is also a key factor in preventing IIAGlc phosphorylation and thereby producing diauxie. In this context, it is worth noting that amino acid sequence analysis have shown that the fructose and the mannitol Enzyme IIA domains are phylogenetically related, an observation which can be linked to the fact that both fructose and mannitol uptake rates are sufficiently high enough to interfere with phosphorylation of IIAGlc.
From the present work, it can be further concluded that the amount of phosphorylated IIAGlc is lower in glucose-grown fruR strains than in glucose-grown wild-type strains. This conclusion is based on the fact that exogenous cAMP does not totally abolish the diauxic lag of fruR strains growing with the glucose/lactose pair although it abolishes this lag when wild type strains are grown with the same pair of sugar. In the fruR strain, the effect of exogenous cAMP is likely to be counteracted by a stronger inhibition of lactose permease. This inhibition lowers inducer uptake and hence lowers lactose operon expression. In agreement with this proposal, the fructose/lactose diauxic lag was totally abolished when a strain lacking both FruR and IIAGlc was examined for growth in the presence of exogenous cAMP.
Curiously, why is the amount of phosphorylated IIAGlc lower in fruR strains as compared to wild-type strains during glucose transport? In fruR strains, re-phosphorylation of IIAGlc during glucose transport may be slower than in wild-type strains. However, this is unlikely because fruR strains do not exhibit biphasic growth when growing on the fructose and glucose pair of sugars. One possibility is that a fraction of IIAGlc is not involved in glucose transport and remains unphosphorylated. In fruR strains, overproduction of FruB which is homologous to and can substitute for HPr (Geerse et al., 1989) may indirectly prevent phosphorylation of this fraction of IIAGlc by inhibiting phosphorylation of HPr by Enzyme I.
ACKNOWLEDGMENTSThe experimental work was performed in Milton H. Saier, Jr.'s laboratory, University of California at San Diego, Department of Biology, La Jolla, California 92093-0116, USA. This work was supported by Public Health Service grant 2R01 AI14176 from the National Institutes of Allergy and Infectious Diseases.
REFERENCES