Introduction

Escherichia coli is able to utilize a wide range of carbon and nitrogen sources. Many of these compounds are metabolized by inducible pathways. Catabolite repression in bacteria is a phenomenon by which glucose and its analogs decrease the intracellular level of both cAMP and its receptor protein, CRP. This in turn causes repression of a set of catabolite-sensitive operons such as lac, ara, mal and others. As glucose is consumed, cAMP and CRP levels increase and the cAMP–CRP complex activates a set of operons whose products are necessary for the recruitment and metabolism of alternative carbon sources (Ullmann and Danchin, 1983; Ishizuka et al., 1993; Kolb et al., 1993).

In addition, the cAMP–CRP complex participates in the regulation of a large number of other processes in E.coli. It mediates activation of genes whose expression is necessary for the metabolism of some amino acids (tnaA, ilvB, etc., Friden et al., 1982; Ullmann and Danchin, 1983), for the metabolism of nucleic acids (deo, cdd, etc., Ullmann and Danchin, 1983) and the synthesis of some membrane proteins (ompA and ompF, Scott and Harwood, 1980). Furthermore, cAMP–CRP is reported to be a repressor of its own transcription (Aiba, 1983; Cossart and Gicquel-Sanzey, 1985; Okamoto and Freundlich, 1986; Hanamura and Aiba, 1991), of genes for adenylate cyclase (Majerfeld et al., 1981; Mori and Aiba, 1985), of genes encoding enzymes participating in glutamine metabolism (Prusiner et al., 1972) and of genes for the outer membrane protein III (Mallick and Herrlich, 1979).

CRP is composed of two identical subunits comprising 209 amino acids (Aiba et al., 1982; Cossart and Gicquel-Sanzey, 1982). When complexed with its allosteric effector, cAMP, it undergoes a conformational transition and binds to a 22 bp target site within, or close to, promoters. At its site of interaction, cAMP–CRP induces sharp bends in DNA ranging from 90 to 180° (Wu and Crothers, 1984; Liu-Johnson et al., 1986). X-ray crystallography and in vitro binding experiments have revealed that the C-terminal domain of the protein interacts with DNA, while the large N-terminal domain binds cAMP (Ogden et al., 1980; Lee et al., 1981; McKay and Steitz, 1981).

FIS is a small, homodimeric DNA-binding protein which, similarly to cAMP–CRP, performs a variety of different roles in E.coli. FIS stimulates stable RNA synthesis upon binding to the upstream activating sequences (UAS) of the corresponding promoters (Nilsson et al., 1990; Ross et al, 1990; Lazarus and Travers, 1993; Emilsson and Nilsson, 1995). FIS influences chromosomal replication by binding to oriC (Gille et al., 1991; Filutowicz et al., 1992; Hiasa and Marians, 1994; Wold et al., 1996), and is a repressor of its own transcription (Koch et al., 1991; Ninnemann et al., 1992). More recently, FIS has been shown to modulate expression of some RpoS-regulated genes (Xu and Johnson, 1995a,b,c), to regulate the promoters of the hupA and hupB genes encoding HU (Claret and Rouviere-Yaniv, 1996) and of the hns gene encoding H-NS (Falconi et al., 1996), and to regulate expression of several operons involved in the catabolism of sugars and nucleic acids (González-Gil et al., 1996). The intracellular FIS concentration varies during cell growth, being highest after nutritional upshift and leveling off during exponential growth (Thompson et al., 1987; Ball et al., 1992).

FIS and CRP exhibit related structural features which lead to similar modes of action. Both proteins contain a helix–turn–helix DNA-binding domain located at the C-terminus (McKay and Steitz, 1981; Kostrewa et al., 1991). Both proteins are homodimers in solution and bend the DNA upon binding (Liu-Johnson et al., 1986; Thompson and Landy, 1988). Their regulatory properties strongly depend on helical phasing (Gaston et al., 1990; Newlands et al., 1992). Additionally, both FIS and CRP bind to highly degenerate sequences (Ebright et al., 1984; Hübner and Arber, 1989), suggesting that not only sequence specificity, but other features like DNA topology and the nucleoprotein context in which a binding site is located (Pan et al., 1996) determine their interaction with DNA.

Transcription of crp is both activated and repressed by its own product. Activation occurs through binding of cAMP–CRP to the CRP II site located upstream of the crp promoter (Hanamura and Aiba, 1992), while repression depends on the CRP I site located downstream of the crp promoter (Aiba, 1983). It has been proposed that repression is accomplished through synthesis of a divergent RNA molecule whose transcription starts 2 bp upstream from the crp mRNA, runs in the opposite direction and is strongly dependent on cAMP–CRP (Okamoto and Freundlich, 1986; Okamoto et al., 1988). This scenario has been questioned by Hanamura and Aiba (1991) who reported that the divergent RNA does not repress crp transcription. They demonstrated that in the presence of cAMP–CRP the RNA polymerase preferentially binds to the divergent RNA promoter, while in its absence it binds to the crp promoter.

In a recent publication, we showed that FIS regulates expression of several operons involved in catabolism of sugars and nucleic acids (González-Gil et al., 1996). In most cases, the observed FIS regulation of the structural genes was indirect and was mediated through effects of FIS on the transcription of the respective repressor genes. A common feature of these FIS-regulated operons was that their transcription is activated by cAMP–CRP. This led us to investigate whether FIS also affects crp expression. In this study, we demonstrate that crp transcription is regulated by FIS and that both FIS and cAMP–CRP are required to keep the crp promoter in a repressed state. Furthermore, we present evidence that the transcriptional regulation of crp during cellular growth is achieved by changes in the composition of nucleoprotein complexes formed at the crp promoter.

Results

crp transcription is regulated by FIS

In order to assess whether FIS influences crp transcription, we prepared RNA from the isogenic strains CSH50 and CSH50 fis at different time points during growth. Interestingly, Northern analysis revealed that crp was transcribed more actively in the absence than in the presence of FIS, and this held true for all time points analyzed, spanning early logarithmic to stationary phase. During logarithmic phase, up to 5-fold more crp RNA was detected in CSH50 fis than in CSH50 (Figure 1).

The crp hybridization signal corresponding to RNA isolated from CSH50 fis always consisted of two bands, the faster migrating band being more abundant than the slower migrating band. In RNA prepared from CSH50, the higher molecular weight crp signal was present at all time points, while the faster migrating RNA species was only detectable in late logarithmic and stationary phase cells (Figure 1). This result suggested that in the absence of FIS (CSH50 fis) or at low FIS concentration (CSH50 at late logarithmic or stationary phases), two transcripts of different length originate from the crp promoter.

Since transcription of crp has been postulated to be autoregulated by activation of divergent RNA synthesis, one might expect that transcription from the divergent promoter would also be affected by FIS. Northern analysis revealed that in CSH50 the divergent RNA was most abundant at mid-log phase (200 Klett units in Figure 2). Since, in this strain, the amount of crp transcript was fairly constant during growth (Figure 1), divergent RNA synthesis apparently does not reduce crp transcription. Northern analysis of RNA isolated from CSH50 fis demonstrated that the divergent RNA could not be detected at any time during growth (Figure 2). These results suggest that FIS is required for activation of the divergent promoter. Thus, FIS and cAMP–CRP regulate transcription from both the crp and divergent RNA promoters.

FIS prevents crp transcription initiation

To discriminate whether the two crp mRNA species detected in CSH50 and CSH50 fis result from a processing of the major transcript or from transcription initiating at different sites, we analyzed the products of crp transcription in vitro using supercoiled templates. The initiation sites were detected by primer extension. Four transcripts were observed, the longest one (crp1) mapping to the start site at +1 (Figure 3). This band corresponded to the transcription start site previously mapped for the crp gene (Aiba, 1983; Okamoto and Freundlich, 1986). The shortest and most abundant message (crp2) comprised two transcripts differing by 1 bp in length and mapping to positions +79 and +80 respectively. A fourth transcript initiating at position +73 was also detected. The initiation frequency at position +73 was similar to that at crp1 and noticeably lower than at crp2. Initiation at crp2 and +73 sites was reduced in the presence of low FIS concentrations, while initiation at the crp1 site was inhibited only at high FIS concentrations (Figure 3). Since all crp1, crp2 and crp +73 messages are readily detectable in a purified in vitro system, we conclude that the crp promoter contains three transcription initiation sites.

Analysis of CRP expression in vivo

To elucidate further how repression of crp transcription takes place, we cloned the crp promoter (-161 to +200) upstream of a lacZ gene in the low copy number expression vector pRS415. The resulting plasmid (pGClac) harboring the crp promoter–lacZ fusion was introduced in the isogenic strains CSH50, CSH50 fis, CSH50 crp and CSH50 crp fis. -Galactosidase activity of cell extracts prepared from overnight cultures of the respective strains revealed that the crp promoter was 2-fold more active in the absence than in the presence of FIS (Figure 4B). As expected, deletion of crp also led to an increased activity of the crp promoter (Figure 4B).

It has been described that CRP lowers the intracellular level of cAMP (Botsford and Drexler, 1978; Ishizuka et al., 1993). In order to determine whether the increased activity of the crp promoter in fis cells is the result of a reduced cAMP level such that no active cAMP–CRP complexes are formed, we grew the cultures in the presence of 1 mM cAMP. -Galactosidase activity was determined at early and late logarithmic phase and from overnight cultures. As expected, addition of cAMP caused repression of the crp promoter in CSH50 and remained without effect in CSH50 crp and CSH50 crp fis (Figure 4A). In CSH50 fis, cAMP addition did not repress the crp promoter, indicating that FIS is essential for repression of crp gene transcription. In the presence of 1 mM cAMP, -galactosidase levels were up to 10-fold higher in CSH50 crp than in CSH50 and up to 30-fold higher in the absence than in the presence of FIS. In addition, there was no difference in -galactosidase activity between CSH50 crp and CSH50 crp fis, except for early logarithmic growth. These results are consistent with the Northern analyses and indicate that both FIS and cAMP–CRP are required throughout the growth phase to keep the crp promoter in a repressed state.

FIS and CRP compete for binding to overlapping binding sites at the crp promoter

To investigate the mechanism by which FIS represses crp transcription, we mapped by DNase I footprinting the FIS- and CRP-binding sites on a crp promoter fragment encompassing sequences from -126 to +119. In vitro, FIS bound the crp promoter at sequences both upstream and downstream of the transcriptional start point, crp1 (Figure 5A and B). The downstream protected region encompasses at least three adjacent FIS-binding sites which differ slightly in affinity. The site centered at +68 is occupied first, and the sites centered at +48 and +30 are protected only at higher FIS concentrations (Figure 5B, compare lanes 2 and 3; sites matching the consensus FIS-binding site are underlined in Figure 5A). Consistent with this observation, the site centered at +68 has the best match to the consensus FIS-binding site (Figure 5C). The far upstream site centered at -92 was occupied simultaneously with the site centered at +68 and showed a good fit to the consensus sequence (Figure 5B, lane 2, and Figure 5C). The FIS site centered at +30 overlapped the -35 element of the divergent RNA promoter. Binding of FIS at this site caused strong DNase I hypersensitivity at position +33, suggesting a substantial DNA distortion in this region. The high affinity FIS-binding site centered at +68 overlapped the transcription initiation site at +73 and almost overlapped the crp2 site, suggesting that binding of FIS may prevent transcription initiation from these sites by steric hindrance.

Binding of cAMP–CRP alone protected the region between positions +29 and +49 and between positions -50 and -70, corresponding to CRP sites I and II respectively (Figure 5B, lanes 5–8). These CRP-binding sites were mapped previously by Aiba (1983) and Hanamura and Aiba (1992). Binding of cAMP–CRP at site I increased DNase I cleavage at positions +35 and +45 (Figure 5B, lanes 8 and 13). Likewise, binding of cAMP–CRP at site II increased DNase I cleavage at two positions within the -50 to -70 region (Figure 5B, lanes 8 and 13).

The FIS-binding sites centered at +48 and +30 overlap CRP-binding site I (Figure 5B, compare lanes 4 and 8). Addition of FIS (10 nM) together with cAMP–CRP (250 nM) resulted in the simultaneous occupation of CRP site I and the non-overlapping FIS site centered at +68, producing an extensive protected region downstream of the crp promoter. Similarly, CRP site II and the FIS site centered at -92 were occupied simultaneously (Figure 5B, lane 14). Increasing FIS concentrations relative to cAMP–CRP prevented binding of cAMP–CRP at site I, but had no effect on cAMP–CRP binding at site II, as evidenced from both the protection pattern by cAMP–CRP and the DNase I-hypersensitive sites (Figure 5B, lane 15). At a FIS/cAMP–CRP ratio of 100:250 nM, CRP-binding site II, as well as the -10 and -35 regions of the crp promoter, became occupied by FIS (Figure 5B, lane 16). In accordance with our primer extension analysis, this result suggests that occupation of secondary binding sites in the upstream region of the crp promoter at high FIS concentrations may both displace cAMP–CRP from binding to site II and prevent transcription initiation from the crp1 site.

Analysis of crp transcription on linear templates in vitro

In order to determine the effect of the distinct nucleoprotein complexes formed at the crp promoter on transcriptional activity, we performed multiple round runoff assays using linear DNA templates encompassing the crp promoter sequences from -126 to +119. In contrast to the result obtained using supercoiled DNA templates, in the absence of other proteins, RNA polymerase (RNAP) initiated transcription from crp1, but not from crp2 or +73 sites (Figure 6, lane 1). In line with previous reports (Hanamura and Aiba, 1992), addition of cAMP–CRP activated transcription from the divergent promoter and markedly reduced initiation at crp1 (Figure 6, lanes 2–5). At high cAMP–CRP concentrations, the repression of the crp1 promoter was relieved (Figure 6, lane 5). In contrast, FIS repressed crp1 transcription in a concentration-dependent manner (Figure 6, lanes 6–9) without activating the divergent RNA promoter. Furthermore, FIS prevented activation of the divergent promoter even in the presence of cAMP–CRP (Figure 6, lanes 10–13). Addition of cAMP–CRP in combination with FIS slightly alleviated the repression of the crp1 promoter, but not of the divergent promoter, indicating that the divergent promoter is repressed by FIS even more efficiently than is the crp1 promoter. These results are consistent with the DNase I protection experiments which show that FIS displaces cAMP–CRP more efficiently from binding site I than from site II (see Figure 5B). Taken together, these results indicate that the nucleoprotein complexes formed at the crp promoter differentially affect promoter activity.

Discussion

Regulation of crp transcription

Here we have shown that in addition to cAMP–CRP, another trans-acting factor, FIS, is indispensable for repression of crp. In CSH50 fis cells, the endogenous crp transcription was strongly enhanced during all phases of growth. Moreover, the crp promoter on a low copy number plasmid was derepressed in CSH50 fis after overnight culture (measured by -galactosidase activity), although these cells overproduced CRP and cAMP was added to the culture. These in vivo results clearly indicate that both proteins, FIS and cAMP–CRP, are required for crp repression, since the lack of either transcriptional regulator leads to a derepressed crp promoter.

In CSH50, a fairly constant level of crp mRNA was transcribed during growth (see Figure 1, upper band). The transcription start site of this mRNA coincides with the one already mapped by two different groups (Aiba, 1983; Okamoto and Freundlich, 1986). From late logarithmic to stationary phase in CSH50, as well as in CSH50 fis, a second mRNA species of smaller size was also detected. Transcription of this smaller mRNA is initiated 73, 79 and 80 bp downstream from the first transcriptional start site. These RNA species were probably not detected by the aforementioned groups because they used RNA prepared from fis⁺ cells at early logarithmic growth. Our results indicate that binding of FIS to high affinity sites in the downstream region of the crp promoter prevents transcription initiation from +73 and crp2 sites. Interestingly, the downstream sites were used very inefficiently when transcription was analyzed on linear DNA templates in vitro. It is possible that specific template topology, e.g. a high level of negative supercoiling of DNA, is required to activate these sites for transcription. We note that the crp2 site is used very efficiently in fis cells, which are characterized by high levels of negative supercoiling (Schneider et al., 1997). In line with this notion, transcription initiation at crp2 is very efficient on negatively supercoiled templates in vitro (Figure 3). The putative RNAP-binding sequence identified within this region contains features characteristic of supercoiling-dependent promoters such as suboptimal -35 and -10 hexamers, a 16 bp spacer and a GC-rich discriminator sequence between the -10 hexamer and the initiation site (Figure 5A; Borowiec and Gralla, 1987; Giladi et al., 1992; Jordi et al., 1995). However, additional experiments are required to define unequivocally the RNAP-binding elements of the crp2 promoter as well as those giving rise to transcription initiation at position +73.

The divergent RNA could only be detected in fis⁺ cells in vivo. However, it is unlikely that FIS directly activates the divergent RNA promoter for the following reasons. First, we did not see any activation of the divergent RNA promoter by FIS in vitro. Furthermore, as shown in our footprint analyses, binding of FIS at a site contained in the divergent RNA promoter caused a strong DNase I hypersensitivity at position +33, suggesting a substantial distortion of DNA in the -35 region of this promoter. Additionally, maximal expression of the divergent RNA was only detected in vivo at late log phase, when the level of FIS decreases (Ninnemann et al., 1992). Importantly, the crp2 initiation site is the stronger site in the absence of FIS in vivo and on supercoiled templates in vitro, as demonstrated by our Northern and primer extension analyses (Figures 1 and 3). We suggest that the absence of the divergent RNA message in fis cells is due to the efficient utilization of the downstream crp initiation site, which would preclude transcription initiation at the divergent promoter. At late log phase in fis⁺ cells, the FIS concentration might be too low to repress the divergent RNA promoter, but still high enough to prevent initiation at the downstream crp sites. Optimal cAMP–CRP levels might then activate divergent RNA synthesis. Indeed, at low FIS concentrations, binding of cAMP–CRP to site I occurs simultaneously with binding of FIS to its high affinity site centered at +68 (Figure 5) which overlaps the -10 region of the putative crp2 promoter and thus may occlude binding of RNAP at this site.

Our results suggest the following model for regulation of crp transcription during the growth phase (Figure 7). After dilution of an overnight culture into fresh medium (nutritional upshift), a large amount of FIS protein is synthesized immediately (Thompson et al., 1987; Ball et al., 1992; Ninnemann et al., 1992). Under these circumstances, it is very likely that the specific FIS-binding sites downstream of the crp promoter are occupied by FIS and transcription can only initiate at the crp1 site. Complete repression of the upstream initiation site by binding of FIS to secondary sites would be prevented by binding of cAMP–CRP at site II, which relieves repression by FIS and activates crp1 transcription (Hanamura and Aiba, 1992; this study). However, binding of FIS to its downstream sites might also limit transcription initiated at crp1. As growth progresses, the intracellular FIS concentration decreases, whereas the cAMP level increases. cAMP–CRP complexes are formed which, at lowered FIS levels, activate divergent RNA synthesis. At this stage, initiation at the crp1 site would not become limited by FIS, but by the concurrent transcription from the divergent promoter, whereas crp transcription from the downstream initiation sites would be repressed completely. From late logarithmic growth onwards, the intracellular FIS concentration is too low to allow saturation of binding sites. Transcription of the divergent RNA would be reduced to its basal level due to transcription of crp RNA from the downstream promoter, which becomes activated in the absence of FIS. Under these conditions, crp transcription would initiate at crp1 and downstream sites.

Biological implications

The cAMP–CRP complex is often found to be involved in the formation of multiprotein regulatory assemblies. For example, to bind to and achieve repression of the deoP2, cdd and udp promoters, the CytR repressor has to interact with cAMP–CRP (Holst et al., 1992; Sogaard-Andersen and Valentin-Hansen, 1993; Brikun et al., 1996). The proximity of CRP- and FIS-binding sites in the crp promoter suggests that specific nucleoprotein complexes containing FIS and CRP are formed and act as transcriptional barriers for RNAP. The levels of FIS in the cell vary with growth and nutritional supply, and it has been proposed that FIS may serve as an indicator for environmental conditions (Ninnemann et al., 1992; González-Gil et al., 1996). We show that the competition between FIS and CRP for occupation of binding sites in the crp promoter may be the key mechanism for this sensing to occur. The composition of nucleoprotein complexes formed at this promoter could oscillate in response to the nutritional conditions of the cell. Moreover, sequential loss of FIS from binding sites having different affinities during the growth phase could coordinate such transitions with the phase of the growth cycle. This situation closely resembles oscillations in the composition of regulatory nucleoprotein complexes formed at the origin of chromosomal replication in E.coli where FIS is also involved (Cassler et al., 1995), and may reflect a general strategy used by the bacterial cell for rapid adaptation to changing growth conditions.

The results presented unravel one more function of the small DNA-binding protein FIS and further substantiate the importance of this protein in transcriptional control in E.coli.

Materials and methods

Strains and plasmids

Bacterial strains used in this study were E.coli K12 derivatives. CSH50 is ara (lac pro) thi (Miller, 1972). CA8445-1 (crp-45; Sabourin and Beckwith, 1975) was kindly provided by B.Bachmann. Construction of CSH50 fis was as described elsewhere (González-Gil et al., 1996). CSH50 crp was constructed by phage P1 transduction of the crp-45 mutation from CA8445-1 into CSH50. crp mutants were screened by co-transduction of rpsL located at 73 min on the E.coli map (crp is at min 74). CSH50 crp fis was created by phage P1 transduction of the fis mutation from CSH50 fis into CSH50 crp and selection for chloramphenicol resistance.

A PCR-derived fragment comprising the crp promoter from -161 to +200 flanked by EcoRI and SphI restriction sites was digested with the same enzymes and cloned in pSP72 (Promega), giving rise to pGC1. Plasmid pGC2 was obtained by restriction of pGC1 with HindIII (the crp sequence from +148 to +200 is eliminated) and religation. pRS415 is a pBR322 derivative designed to measure promoter strength as an operon fusion with lacZ (Simons et al., 1987). To construct pGClac, the EcoRI–PvuII crp fragment (-161 to +200) from pGC1 was cloned into pRS415 previously digested with BamHI, the overhang filled with Klenow polymerase and redigested with EcoRI.

Proteins

FIS and RNAP were purified as described previously (Koch and Kahmann, 1985; Metzger et al., 1993). Purified CRP was kindly provided by A.Kolb.

Growth of strains

Overnight cultures were diluted 1:200 in dYT medium (Miller, 1972) and grown with vigorous shaking at 37°C until the indicated cell densities. For CSH50 and CSH50 fis, 60 Klett units correspond to early logarithmic growth, 100–200 Klett units represent logarithmic growth, 300 Klett units are reached at late logarithmic growth and 400 Klett units are stationary growth cells. Although the time taken by CSH50, CSH50 fis, CSH50 crp and CSH50 crp fis to reach stationary phase was the same, the last two strains showed reduced growth in dYT medium, reaching the stationary growth phase at 170 and 100 Klett units respectively.

RNA isolation and Northern analysis

Samples of 10¹⁰ cells were chilled on ice and collected by centrifugation. The pellet was resuspended in 0.6 ml of LETS buffer (100 mM LiCl, 10 mM EDTA, 10 mM Tris–HCl pH 7.5, 0.2% SDS). An equal volume of phenol/chloroform was added and vortexed for 1 min. After centrifugation, the RNA was precipitated by adding LiCl to a final concentration of 0.5 M and 2 vols of ethanol. The RNA was resuspended in 100 l of H₂O. For Northern analysis, 20 g of RNA per lane was separated on a 1% agarose gel containing formaldehyde (Sambrook et al., 1989) and transferred to GeneScreen Plus filters (DuPont). Filters were hybridized overnight with digoxigenin-labeled oligonucleotides (Boehringer) which were complementary to the RNA. Luminescent detection of the hybrid bands was performed according to instructions given by the supplier.

The following oligonucleotides were used: eno, 5'-TGACGGAGCAGCTGCCATACCGACGA-3'; crp, 5'-CTGATTCAGATAGGAGAGGATCAT-3'; and divergent RNA, 5'-GAGTACGCGTACTAACCAAATCGC-3'.

DNase I footprinting

DNase I footprinting was performed with a 245 bp crp promoter fragment end-labeled on the bottom strand. The conditions of footprinting were essentially as described earlier (Muskhelishvili et al., 1997). The crp promoter region (-126 to +119) was PCR amplified using the primers CAP3 (5'-CTGTCTCTGGATTGCCGAAATATG-3') and CAP5 (5'-CTCCACTGCGTCAATTTTCCTG-3') and the pGC1 DNA as template. The primer CAP3 was uniquely end-labeled by using [-³²P]ATP and T4 polynucleotide kinase. The obtained fragment was purified by PAGE using a neutral 0.5 TBE gel. The incubation mixture contained 10 mM Tris–HCl, pH 7.9, 75 mM NaCl, 1 mM dithiothreitol (DTT), 0.2 mM cAMP, and CRP and FIS as indicated in a 20 l volume. After incubation for 60 min at 37°C, DNase I and MgCl₂ were added to 2 g/ml and 10 mM final concentrations respectively. The reaction was terminated after 10 s by adding 80 l of a solution containing 0.5% SDS and 50 mM EDTA. After digestion by proteinase K for 45 min at 45°C and phenol extraction, the aqueous phase was precipitated with ethanol. The pellets were washed with 70% ethanol, dried, dissolved in loading dye and analyzed on a 6% denaturing polyacrylamide gel. Protected and hypersensitive bands were identified by using the Maxam–Gilbert G-ladder (Maxam and Gilbert, 1977) of the same DNA fragment as reference.

In vitro transcription

Supercoiled templates. In vitro transcription was carried out using 2 g of supercoiled pGC1 DNA in a buffer containing 10 mM Tris–HCl pH 8, 2 mM DTT, 100 mM NaCl, 10 mM MgCl₂, 2.5 mM each ATP, GTP, CTP and UTP, 200 nM RNAP, and FIS as indicated for 30 min at 30°C in a 50 l reaction volume. The reaction was stopped by adding 250 l of 1 mM EDTA, 50 nM NaOAc, 0.2% SDS, 10 mM Tris–HCl pH 7.4 and 10 g/ml proteinase K. After incubation for 30 min at 42°C, the reactions were extracted with phenol followed by two rounds of precipitation with ethanol in the presence of 0.3 M NaOAc and 10 g of carrier tRNA. Half of the in vitro reaction products were used for primer extension.

Linear templates. Multiple round runoff transcription reactions were performed with a 245 bp crp promoter fragment (-126 to +119) (10 nM) in a buffer containing 10 mM Tris–HCl pH 7.9, 75 mM NaCl, 1 mM DTT, 0.2 mM cAMP, CRP and FIS as indicated, 66 nM RNAP, 1 mM each GTP, CTP and ATP, and 0.1 mM [-³²P]UTP, in a 20 l volume at 37°C. The reaction was terminated in 15 min by addition of an equal volume of formaldehyde loading dye and heating to 94°C. The samples were loaded on 6% denaturing polyacrylamide gels and analyzed by phosphorimaging (PhosphorImager Storm 840, Molecular Dynamics).

Primer extension

In vitro transcription products were annealed with 1 ng of 5'-end-labeled synthetic primer CAP3 (see above) after heating to 70°C for 5 min, quick chilling on ice and subsequent incubation at 42°C for 5 min in a 10 l volume containing 40 U of RNase inhibitor (Boehringer Mannheim). Primer extension was carried out using 200 U of SuperScript™II reverse transcriptase (Gibco-BRL) in a buffer supplied by the manufacturer. After incubation for 1 h at 37°C, the reverse transcriptase was inactivated by heating at 70°C for 15 min. After addition of an equal volume of formaldehyde loading dye, the reaction products were analyzed on 6% denaturing polyacrylamide gels as described above.

-Galactosidase determinations

Overnight cultures were diluted 1:200 in fresh dYT medium which was supplemented with 1 mM cAMP where indicated. Samples taken at the indicated times were assayed for -galactosidase activity following the protocol of Sadler and Novick (1965). -Galactosidase units were multiplied by 1000 to make them equivalent to those of Miller (1972).

Acknowledgements

We thank O.Ninnemann and C.Koch for stimulating discussions. We thank A.Kolb for providing purified CRP. This work was supported by grant SC1^*-CT89-0143 from the EEC and by the Deutsche Forschungsgemeinschaft through SFB 190.

References

Aiba H (1983) Autoregulation of the Escherichia coli crp gene: CRP is a transcriptional repressor for its own gene. Cell, 32, 141–149. | PubMed | ChemPort |

Aiba H, Fujimoto S and Ozaki N (1982) Molecular cloning and nucleotide sequencing of the gene for E.coli cAMP receptor protein. Nucleic Acids Res, 10, 1345–1361. | PubMed | ISI | ChemPort |

Ball CA, Osuna R, Ferguson KC and Johnson RC (1992) Dramatic changes in FIS levels upon nutrient upshift in Escherichia coli. J Bacteriol, 174, 8043–8056. | PubMed | ChemPort |

Borowiec JA and Gralla JD (1987) All three elements of the lac p^s promoter mediate its transcriptional response to DNA supercoiling. J Mol Biol, 195, 85–97.

Botsford JL and Drexler M (1978) CRP and the regulation of cAMP synthesis in Escherichia coli. Mol Gen Genet, 165, 47–56. | PubMed | ChemPort |

Brikun I, Suziedelis K, Stemmann O, Zhong R, Alikhanian L, Linkova E, Mironov A and Berg DE (1996) Analysis of CRP–CytR interactions at the Escherichia coli udp promoter. J Bacteriol, 178, 1614–1622. | PubMed | ChemPort |

Cassler MR, Grimwade J and Leonard AC (1995) Cell cycle-specific changes in nucleoprotein complexes at a chromosomal replication origin. EMBO J, 14, 5833–5841. | PubMed | ChemPort |

Claret L and Rouviere-Yaniv J (1996) Regulation of HU and HU by CRP and FIS in Escherichia coli. J Mol Biol, 263, 126–139. | Article | PubMed | ChemPort |

Cossart P and Gicquel-Sanzey B (1982) Cloning and sequencing of the crp gene of Escherichia coli K-12. Nucleic Acids Res, 10, 1363–1378. | PubMed | ChemPort |

Cossart P and Gicquel-Sanzey B (1985) Regulation of expression of the crp gene of Escherichia coli K-12: in vivo study. J Bacteriol, 161, 454–457. | PubMed | ChemPort |

Ebright RH, Cossart P, Gicquel-Sanzey B and Beckwith J (1984) Mutations that alter the DNA sequence specificity of the catabolite gene activator protein of E.coli. Nature, 311, 232–235. | PubMed | ChemPort |

Emilsson V and Nilsson L (1995) Factor for inversion stimulation-dependent growth rate regulation of serine and threonine tRNA species. J Biol Chem, 270, 16610–16614. | Article | PubMed | ChemPort |

Falconi M, Brandi A, La Teana A, Gualerzi CO and Pon CL (1996) Antagonistic involvement of FIS and H-NS proteins in the transcriptional control of hns expression. Mol Microbiol, 19, 965–975. | Article | PubMed | ChemPort |

Filutowicz M, Ross W, Wild J and Gourse RL (1992) Involvement of FIS protein in replication of the Escherichia coli chromosome. J Bacteriol, 174, 398–407. | PubMed | ISI | ChemPort |

Friden P, Newman T and Freundlich M (1982) Nucleotide sequence of the ilvB promoter-regulatory region: a biosynthetic operon controlled by attenuation and cyclic AMP. Proc Natl Acad Sci USA, 79, 6156–6160. | PubMed | ChemPort |

Gaston K, Bell A, Kolb A, Buc H and Busby S (1990) Stringent spacing requirements for transcription activation by CRP. Cell, 62, 733–743. | PubMed | ISI | ChemPort |

Giladi H, Koby S, Gottesman ME and Oppenheim AB (1992) Supercoiling, integration host factor, and a dual promoter system, participate in the control of the bacteriophage pL promoter. J Mol Biol, 224, 937–948. | PubMed | ChemPort |

Gille H, Egan JB, Roth A and Messer W (1991) The FIS protein binds and bends the origin of chromosomal replication, oriC, of Escherichia coli. Nucleic Acids Res, 19, 4167–4172. | PubMed | ISI | ChemPort |

González-Gil G, Bringmann P and Kahmann R (1996) FIS is a regulator of metabolism in Escherichia coli. Mol Microbiol, 22, 21–29. | PubMed  |

Hanamura A and Aiba H (1991) Molecular mechanism of negative autoregulation of Escherichia coli crp gene. Nucleic Acids Res, 19, 4413–4419. | PubMed | ChemPort |

Hanamura A and Aiba H (1992) A new aspect of transcriptional control of the Escherichia coli crp gene positive autoregulation. Mol Microbiol, 6, 2489–2497. | PubMed | ChemPort |

Hiasa H and Marians KJ (1994) FIS cannot support oriC DNA replication in vitro. J Biol Chem, 269, 24999–25003. | PubMed | ChemPort |

Holst B, Sogaard-Andersen L, Pedersen H and Valentin-Hansen P (1992) The cAMP–CRP/CytR nucleoprotein complex in Escherichia coli: two pairs of closely linked binding sites for the cAMP–CRP activator complex are involved in combinatorial regulation of the cdd promoter. EMBO J, 11, 3635–3643. | PubMed | ChemPort |

Hübner P and Arber W (1989) Mutational analysis of a procaryotic recombinational enhancer element with two functions. EMBO J, 8, 577–585. | PubMed | ISI | ChemPort |

Ishizuka H, Hanamura A, Kunimura T and Aiba H (1993) A lowered concentration of cAMP receptor protein caused by glucose is an important determinant for catabolite repression in Escherichia coli. Mol Microbiol, 10, 341–350. | PubMed | ISI | ChemPort |

Jordi BJAM, Owen-Hughes TA, Hulton CSJ and Higgins CF (1995) DNA twist, flexibility, and transcription of the osmoregulated proU promoter of Salmonella typhimurium. EMBO J, 14, 5690–5700. | PubMed | ChemPort |

Koch C and Kahmann R (1985) Purification and properties of the Escherichia coli host factor required for the inversion of the G segment in bacteriophage Mu. J Biol Chem, 261, 15673–15678.

Koch C, Ninnemann O, Fuss H and Kahmann R (1991) The N-terminal part of the E.coli DNA binding protein FIS is essential for stimulating site-specific DNA inversion but is not required for specific DNA binding. Nucleic Acids Res, 19, 5915–5922. | PubMed | ChemPort |

Kolb A, Busby S, Buc H, Garges S and Adhya S (1993) Transcriptional regulation by cAMP and its receptor protein. Annu Rev Biochem, 62, 749–795. | Article | PubMed | ISI | ChemPort |

Kostrewa D, Granzin J, Koch C, Choe HW, Labahn J, Kahmann R and Saenger W (1991) Three-dimensional structure of the E.coli DNA binding protein FIS. Nature, 349, 178–180. | Article | PubMed | ISI | ChemPort |

Lazarus LR and Travers AA (1993) The E.coli FIS protein is not required for the activation of tyrT transcription on simple nutritional upshift. EMBO J, 12, 2483–2494. | PubMed | ISI | ChemPort |

Lee NL, Gielow WO and Wallace RG (1981) Mechanism of araC autoregulation and the domains of two overlapping promoters, P_C and P_BAD in the L-arabinose regulatory region of Escherichia coli. Proc Natl Acad Sci USA, 78, 752–756. | PubMed | ChemPort |

Liu-Johnson H-N, Gartenberg MR and Crothers DM (1986) The DNA binding domain and bending angle of E.coli CAP protein. Cell, 47, 995–1005. | PubMed | ChemPort |

Majerfeld IH, Miller D, Spitz E and Rickenberg HV (1981) Regulation of the synthesis of adenylate cyclase in Escherichia coli by the cAMP–cAMP receptor protein complex. Mol Gen Genet, 181, 470–475. | PubMed | ChemPort |

Mallick U and Herrlich P (1979) Regulation of synthesis of a major outer membrane protein: cyclic AMP represses Escherichia coli protein III synthesis. Proc Natl Acad Sci USA, 76, 5520–5523. | PubMed | ChemPort |

Maxam A and Gilbert W (1977) A new method for sequencing DNA. Proc Natl Acad Sci USA, 74, 560–564. | PubMed | ChemPort |

McKay DB and Steitz TA (1981) Structure of catabolite gene activator protein at 2.9 Å resolution suggests binding to left-handed B-DNA. Nature, 290, 744–749. | PubMed | ISI | ChemPort |

Metzger W, Schickor P, Meier T, Werel W and Heumann H (1993) Nucleation of RNA chain formation by Escherichia coli DNA-dependent RNA polymerase. J Mol Biol, 232, 35–49. | Article | PubMed | ChemPort |

Miller JH (1972) Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

Mori K and Aiba H (1985) Evidence for negative control of cya transcription by cAMP and cAMP receptor protein in intact Escherichia coli cells. J Biol Chem, 260, 14838–14843. | PubMed | ChemPort |

Muskhelishvili G, Buckle M, Heumann H, Kahmann R and Travers AA (1997) FIS activates sequential steps during transcription initiation at a stable RNA promoter. EMBO J, 16, 3655–3665. | Article | PubMed | ISI | ChemPort |

Newlands JT, Josaitis CA, Ross W and Gourse LR (1992) Both FIS-dependent and factor-independent upstream activation of the rrnB P1 promoter are face of the helix dependent. Nucleic Acids Res, 20, 719–726. | PubMed | ISI | ChemPort |

Nilsson L, Vanet A, Vijgenboom E and Bosch L (1990) The role of FIS in transactivation of stable RNA operons in E.coli. EMBO J, 9, 727–734. | PubMed | ISI | ChemPort |

Ninnemann O, Koch C and Kahmann R (1992) The E.coli fis promoter is subject to stringent control and autoregulation. EMBO J, 11, 1075–1083. | PubMed | ChemPort |

Ogden S, Haggerty D, Stoner C, Kolodrubetz D and Schleif R (1980) The Escherichia coli L-arabinose operon: binding sites of the regulatory proteins and a mechanism of positive and negative regulation. Proc Natl Acad Sci USA, 77, 3346–3350. | PubMed | ChemPort |

Okamoto K and Freundlich M (1986) Mechanism for the autogenous control of the crp operon: transcriptional inhibition by a divergent RNA transcript. Proc Natl Acad Sci USA, 83, 5000–5004. | PubMed | ChemPort |

Okamoto K, Hara S, Bhasin R and Freundlich M (1988) Evidence in vivo for the autogenous control of the cyclic AMP receptor protein gene (crp) in Escherichia coli by divergent RNA. J Bacteriol, 170, 5076–5079. | PubMed | ChemPort |

Pan CQ, Finkel SE, Cramton SE, Feng J, Sigman DS and Johnson RC (1996) Variable structures of FIS–DNA complexes determined by flanking DNA–protein contacts. J Mol Biol, 264, 675–695. | Article | PubMed | ISI | ChemPort |

Prusiner S, Miller RE and Valentine RC (1972) Adenosine 3':5'-cyclic monophosphate control of the enzymes of glutamine metabolism in Escherichia coli. Proc Natl Acad Sci USA, 69, 2922–2926. | PubMed | ChemPort |

Ross W, Thompson JF, Newlands JT and Gourse RL (1990) E.coli FIS protein activates ribosomal RNA transcription in vitro and in vivo. EMBO J, 9, 3733–3742. | PubMed | ISI | ChemPort |

Sabourin D and Beckwith J (1975) Deletion of the Escherichia coli crp gene. J Bacteriol, 122, 338–340. | PubMed | ISI |

Sadler JR and Novick A (1965) The properties of repressor and the kinetics of each action. J Mol Biol, 12, 305–327. | PubMed | ChemPort |

Sambrook J, Fritsch EF and Maniatis T (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

Schneider R, Travers AA and Muskhelishvili G (1997) FIS modulates growth-phase dependent topological transitions of DNA in E.coli. Mol Microbiol, 26, 519–530. | Article | PubMed | ISI | ChemPort |

Scott NW and Harwood CR (1980) Studies on the influence of the cyclic AMP system on major outer membrane proteins of Escherichia coli. FEMS Microbiol Lett, 9, 95–98. | Article | ChemPort |

Simons RW, Houman F and Kleckner N (1987) Improved single and multicopy lac-based cloning vectors for protein and operon fusions. Gene, 53, 85–96. | Article | PubMed | ISI | ChemPort |

Sogaard-Andersen L and Valentin-Hansen P (1993) Protein–protein interactions in gene regulation: the cAMP–CRP complex sets the specificity for a second DNA binding protein, the CytR repressor. Cell, 75, 557–566. | PubMed | ChemPort |

Thompson JF and Landy A (1988) Empirical estimation of protein-induced DNA bending angles: application to lambda site specific recombination complexes. Nucleic Acids Res, 16, 9687–9705. | PubMed | ISI | ChemPort |

Thompson JF, de Vargas LM, Koch C, Kahmann R and Landy A (1987) Cellular factors couple recombination with growth phase: characterization of a new component in the site-specific recombination pathway. Cell, 50, 901–908. | PubMed | ISI | ChemPort |

Ullmann A and Danchin A (1983) Role of cyclic AMP in bacteria. Adv Cyclic Nucleotide Res, 15, 1–53. | ChemPort |

Weng M, Makaroff CA and Zalkin H (1986) Nucleotide sequence of Escherichia coli pyrG encoding CTP synthetase. J Biol Chem, 261, 5568–5574. | PubMed | ChemPort |

Wold S, Crooke E and Skarstad K (1996) The Escherichia coli FIS protein prevents initiation of DNA replication from oriC in vitro. Nucleic Acids Res, 24, 3527–3532. | Article | PubMed | ChemPort |

Wu H-M and Crothers DM (1984) The locus of sequence-directed and protein-induced DNA bending. Nature, 308, 509–513. | PubMed | ChemPort |

Xu J and Johnson RC (1995a) Identification of genes negatively regulated by FIS: FIS and RpoS comodulate growth-phase dependent gene expression in Escherichia coli. J Bacteriol, 177, 938–947. | ISI | ChemPort |

Xu J and Johnson RC (1995b) aldB an RpoS-dependent gene in Escherichia coli encoding an aldehyde dehydrogenase that is repressed by FIS and activated by CRP. J Bacteriol, 177, 3166–3175. | ChemPort |

Xu J and Johnson RC (1995c) FIS activates the RpoS-dependent stationary-phase expression of proP in Escherichia coli. J Bacteriol, 177, 5222–5231. | ISI | ChemPort |