Portal de Periódicos da CAPES

FacB, the Aspergillus nidulans activator of acetate utilization genes, binds dissimilar DNA sequences

1998; Springer Nature; Volume: 17; Issue: 7 Linguagem: Inglês

10.1093/emboj/17.7.2042

1460-2075

R. B. Todd, Alex Andrianopoulos, Meryl A. Davis, Michael J. Hynes,

Plant Disease Resistance and Genetics

Article1 April 1998free access FacB, the Aspergillus nidulans activator of acetate utilization genes, binds dissimilar DNA sequences Richard B. Todd Richard B. Todd Present address: Biotechnology Laboratory and Department of Botany, Room 237 Wesbrook Building, 6174 University Blvd, University of British Columbia, Vancouver, BC, V6T 1Z3 Canada Search for more papers by this author Alex Andrianopoulos Alex Andrianopoulos Department of Genetics, The University of Melbourne, Parkville, Victoria, 3052 Australia Search for more papers by this author Meryl A. Davis Meryl A. Davis Department of Genetics, The University of Melbourne, Parkville, Victoria, 3052 Australia Search for more papers by this author Michael J. Hynes Corresponding Author Michael J. Hynes Department of Genetics, The University of Melbourne, Parkville, Victoria, 3052 Australia Search for more papers by this author Richard B. Todd Richard B. Todd Present address: Biotechnology Laboratory and Department of Botany, Room 237 Wesbrook Building, 6174 University Blvd, University of British Columbia, Vancouver, BC, V6T 1Z3 Canada Search for more papers by this author Alex Andrianopoulos Alex Andrianopoulos Department of Genetics, The University of Melbourne, Parkville, Victoria, 3052 Australia Search for more papers by this author Meryl A. Davis Meryl A. Davis Department of Genetics, The University of Melbourne, Parkville, Victoria, 3052 Australia Search for more papers by this author Michael J. Hynes Corresponding Author Michael J. Hynes Department of Genetics, The University of Melbourne, Parkville, Victoria, 3052 Australia Search for more papers by this author Author Information Richard B. Todd2, Alex Andrianopoulos1, Meryl A. Davis1 and Michael J. Hynes 1 1Department of Genetics, The University of Melbourne, Parkville, Victoria, 3052 Australia 2Present address: Biotechnology Laboratory and Department of Botany, Room 237 Wesbrook Building, 6174 University Blvd, University of British Columbia, Vancouver, BC, V6T 1Z3 Canada *Corresponding author. E-mail: [email protected] The EMBO Journal (1998)17:2042-2054https://doi.org/10.1093/emboj/17.7.2042 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The facB gene is required for acetate induction of acetamidase (amdS) and the acetate utilization enzymes acetyl-CoA synthase (facA), isocitrate lyase (acuD) and malate synthase (acuE) in Aspergillus nidulans. The facB gene encodes a transcriptional activator with a GAL4-type Zn(II)2Cys6 zinc binuclear cluster DNA-binding domain which is shown to be required for DNA binding. In vitro DNA-binding sites for FacB in the 5′ regions of the amdS, facA, acuD and acuE genes have been identified. Mutations in amdS FacB DNA-binding sites affected expression of an amdS–lacZ reporter in vivo and altered the affinity of in vitro DNA binding. This study shows that the FacB Zn(II)2Cys6 cluster binds to dissimilar sites which show similarity in form but not sequence with DNA-binding sites of other Zn(II)2Cys6 proteins. Sequences with homology to FacB sites are found in the 5′ regions of genes regulated by the closely related yeast Zn(II)2Cys6 protein CAT8. Introduction Acetate metabolism has been studied in a number of fungi including Neurospora crassa, Coprinus cinereus, Saccharomyces cerevisiae and Aspergillus nidulans (e.g. Flavell and Fincham, 1968a,b; Casselton and Casselton, 1974; Armitt et al., 1976; Fernández et al., 1992; Schöler and Schüller, 1994). Acetate is activated by acetyl-CoA synthase (ACS) to acetyl-CoA, which is metabolized via the anaplerotic glyoxylate bypass of the tricarboxylic acid cycle comprising isocitrate lyase (ICL), which converts isocitrate to glyoxylate and succinate, and malate synthase (MAS), which converts glyoxylate and acetyl-CoA to malate (Kornberg, 1966). The facB gene of A.nidulans encodes the major regulator of genes involved in acetate utilization. Recessive loss-of-function mutations in facB lead to resistance to fluoroacetate, reduced growth on medium containing acetate as a sole carbon source and reduced levels of activities required for acetate utilization, i.e.ACS (encoded by facA), ICL (encoded by acuD), MAS (encoded by acuE), phosphoenol pyruvate carboxy kinase (PEPCK) and fructose-1,6-bisphosphatase (FBP) (Apirion, 1965; Armitt et al., 1976; Hynes, 1977; Kelly, 1980). In addition, facB mutations affect acetate induction of acetamidase (encoded by amdS) and NADP-linked isocitrate dehydrogenase (NADP-IDH) (Hynes, 1977; Kelly and Hynes, 1982; Todd et al., 1997b). These pleiotropic effects have led to the proposal that facB is a positively acting regulatory gene involved in acetate induction (Hynes, 1977; Kelly, 1980). The cis-dominant amdI9 mutation, a single base pair transition at −210 relative to the translational start point of amdS, results in a 10-fold increased acetate induction of acetamidase (Hynes, 1975; Hynes et al., 1988). facB recessive loss-of-function mutations are epistatic to the amdI9 mutation (Hynes, 1977). Multiple copies of a plasmid carrying the amdS gene with the amdI9 mutation, but not a wild-type amdS plasmid, resulted in reduced growth on acetate medium (Kelly and Hynes, 1987), indicating that the amdI9 mutation results in increased affinity for, and therefore titration of, the facB gene product. Hynes et al. (1988) localized regions of 5′ amdS involved in titration to −219 to −111 by the transformation of an amdI9 strain with plasmids containing subcloned fragments. Multiple copies of the cloned facB gene reversed the titration of the facB gene product (Katz and Hynes, 1989). The facB gene encodes a protein of 867 residues which contains features characteristic of transcriptional activators. Near the N-terminus is a Zn(II)2Cys6 (or C6 zinc) binuclear cluster DNA-binding motif (residues 24–51), followed by a linker region (residues 52–66) and a series of five leucine zipper-like heptad repeats (residues 67–101) possibly involved in dimerization (Todd et al., 1997a). The Zn(II)2Cys6 cluster DNA-binding motif as well as the linker region of FacB shows high similarity with those of the S.cerevisiae CAT8 and SIP4 proteins. The CAT8 protein is thought to be the central regulator of gluconeogenesis (Hedges et al., 1995) while SIP4 is involved in carbon metabolism via its interaction with SNF1 (Lesage et al., 1996). The GAL4 Zn(II)2Cys6 DNA-binding motif consists of six conserved cysteine residues, arranged CX2CX6CX6CX2CX6C, which form two α-helical structures separated by a proline-associated loop and coordinates two zinc(II) ions to form a cloverleaf shaped structure (Pan and Coleman, 1990, 1991; Gardner et al., 1991; Baleja et al., 1992; Kraulis et al., 1992; Marmorstein et al., 1992). Zn(II)2Cys6 DNA-binding domains interact with DNA-binding sites which are similar in both structure and sequence, consisting of conserved terminal trinucleotides, usually in a symmetrical configuration, spaced by an internal variable sequence of defined length, e.g. GAL4 and LAC9 bind to CGGN11CCG, PPR1 and UaY bind to CGGN6CCG, and PUT3 binds to CGGN10CCG (Carey et al., 1989; Siddiqui and Brandriss, 1989; Halvorsen et al., 1990, 1991; Roy et al., 1990; Gödecke et al., 1991; Marmorstein et al., 1992; Suárez et al., 1995; for a review, see Todd and Andrianopoulos, 1997). Here we show that the FacB Zn(II)2Cys6 cluster is required for in vitro DNA binding. We have identified FacB-binding sites 5′ to amdS and genes of acetate utilization and analysed the effects of binding site mutations. The FacB Zn(II)2Cys6 domain binds to two disparate classes of DNA-binding site which are similar in form but dissimilar in sequence to DNA-binding sites of other Zn(II)2Cys6 proteins. Sequences with homology to FacB DNA-binding sites are found in the CAT8-dependent UAS 5′ of the S.cerevisiae FBP1 gene, indicating that the CAT8 Zn(II)2Cys6 cluster may bind DNA-binding sites similar to that of FacB. Results FacB binds to multiple regions of the amdS promoter N-terminal FacB residues 4–417 were expressed in Escherichia coli as a fusion protein with the E.coli maltose-binding protein (MBP) (see Materials and methods). Electrophoretic mobility shift assays (EMSAs) with the −268 to +22 region of the amdS promoter (Figure 1A), which is required for facB-dependent activation of amdS expression (Hynes et al., 1988), showed that the MBP–FacB(4–417) fusion protein but not MBP alone can bind to this DNA fragment (Figure 1B). A single major DNA–protein complex was evident at all concentrations of MBP–FacB tested. A minor faster migrating complex was also evident at all concentrations, while at higher concentrations of MBP–FacB an additional, lower mobility, complex was observed. The fixed stoichiometry between the fast migrating complex and the major complex suggests that the former most likely represents a minor degradation product of MBP–FacB. Figure 1.FacB binds to multiple regions of the amdS promoter. (A) The amdS 5′ regulatory region showing relevant restriction sites and their nucleotide coordinates relative to the translational initiation site (+1). Probes used for EMSAs are shown below with their end coordinates. (B) EMSA using the amdS promoter region from −268 to +22, encompassing at least two FacB-binding sites, with increasing concentrations of protein extract. Reactions contained no added protein extract (lane 1), extract expressing the MBP protein (lanes 2–3 with 6 and 60 μg, respectively) or extract expressing MBP–FacB fusion protein (lanes 4–10 with 0.4, 0.75, 1.5, 3, 6 and 60 μg, respectively). (C) Promoter-scanning EMSA using fragments encompassing the amdS 5′ regulatory region. Each probe is denoted by its end coordinates shown below each set of three lanes, which represent no added protein extract (no extract), extract containing the MBP protein and extract containing the MBP–FacB fusion protein. The amount of extract used was 6 μg, except for reactions with the −647 to −578 probe where 30 μg of extract was used. Download figure Download PowerPoint Extracts from cells expressing the shorter MBP–FacB(4–142) fusion protein (see Materials and methods) exhibited the same DNA-binding specificity for representative DNA probes used in EMSAs as those from cells expressing MBP–FacB(4–417) (data not shown). Thus, residues 4–142, which contain the Zn(II)Cys6 cluster and adjacent heptad repeat region, are sufficient for DNA binding. Consistent with this, the in vitro-generated Zn(II)Cys6 cluster mutation facBR40G41, which abolishes in vivo facB function (Todd et al., 1997a), was introduced into the MBP–FacB fusion protein construct and found to result in loss of MBP–FacB binding to both 5:6 and 31:32 oligonucleotides in EMSAs (Figure 2C). Figure 2.FacB binds to a single binding site in a sequence-specific manner and requires the Zn(II)Cys6 binuclear cluster. (A) Specific competition of FacB binding. EMSA using the amdS 5′ regulatory region from −268 to +22 as a probe. Binding was performed with 6 μg of MBP–FacB fusion protein extract in the presence of increasing amounts of unlabelled competitor DNAs. Specific competitor DNAs were the unlabelled probe (−268 to +22), the wild-type amdS region identified by the amdI18 mutation (31:32) and the amdI9 mutant region (5:6). Non-specific competitor control was the oligonucleotide pair spanning the CCAAT motif in the amdS promoter (10:11). Reactions contained no competitor, 6.25-, 12.5-, 25-, 50- and 75-fold molar excess of competitor to probe (lanes 1–6), or no competitor, 11-, 22-, 43- and 86-fold molar excess of competitor to probe (lanes 7–11, 12–16 and 17–21). (B) EMSA using the oligonucleotide pairs 5:6, 31:32 and 10:11 as probes with either no added protein extract (no extract), 30 μg of extract containing the MBP protein or 30 μg of extract containing the MBP–FacB protein. (C) EMSA using the wild-type MBP–FacB fusion protein or a mutated MBP–FacB fusion (MBP–FacB R40G41). The latter has the two adjacent cysteine residues in the fourth cysteine position of the Zn(II)2Cys6 binuclear cluster DNA-binding motif mutated. EMSA probes were the 5:6 and 31:32 oligonucleotide pairs, and reactions contained no extract or 6 μg of the MBP–FacB fusion extracts. Download figure Download PowerPoint In three non-overlapping regions of the amdS promoter, −647 to −578, −268 to −151 and −147 to +22, specific DNA binding by the FacB fusion protein was observed (Figure 1C). Binding to the −647 to −578 fragment was weaker than that for the other amdS probes as five times the amount of extract was required to achieve a similar proportion of retarded probe. For the remaining two regions, −577 to −405 and −405 to −219, no binding by MBP–FacB could be detected, although it is possible but unlikely that this was obscured by the FacB-independent binding noted for the −577 to −405 probe. These data, together with the results from Figure 1B, clearly demonstrate the presence of at least three independent FacB DNA-binding sites in 5′ amdS. The −647 to −578 fragment contains a sequence with homology to a sequence within the −268 to −151 fragment that has been noted previously (Hynes et al., 1988). Weak binding to this fragment contributes little to amdS expression since deletions lacking this site but retaining sequences within the −268 to −151 and −147 to +22 fragments retain facB-dependent regulation (T.G.Littlejohn and M.J.Hynes, unpublished data). To localize further the binding sites for MBP–FacB protein in the −268 to +22 region of 5′ amdS, specific competition mobility shift assays were performed (Figure 2A). Competition for DNA binding was observed for the double-stranded oligonucleotides 5:6 and 31:32 which span −219 to −199 and −134 to −105, respectively. The double-stranded oligonucleotide 10:11 spanning −185 to −151 did not compete for FacB binding. EMSAs using the 5:6 and 31:32 double-stranded oligonucleotides as probes clearly showed that these sequences contained FacB-binding sites (Figure 2B). In vivo and in vitro analysis of wild-type and mutant FacB DNA-binding sites in 5′ amdS Cis-acting 5′ amdS mutations (Table I) were investigated for their effects on amdS–lacZ levels in vivo and on in vitro MBP–FacB binding. The amdI9 mutation, a T→C transition at −210, has been found to result in increased responses to acetate induction (Hynes, 1975, 1977; Hynes et al., 1988) and is contained within the 5:6 oligonucleotide which binds MBP–FacB (see above). The amdI9 mutation resulted in 4-fold acetate induction compared with 1.3-fold induction for the wild-type amdS promoter, and this increased induction was FacB dependent (Figure 3A). Comparative EMSAs using wild-type and amdI9-carrying −268 to −115 fragments as probes showed that a greater proportion of probe was retarded by MBP–FacB when the amdI9 mutation was present than with wild-type sequences (Figure 4). Densitometry showed that this difference in affinity was ∼7.5-fold over a range of protein concentrations (data not shown). Thus, the amdI9 mutation resulted in an increased affinity of binding by MBP–FacB in vitro. Figure 3.Cis-acting mutations in 5′ amdS define two distinct FacB sites of action in vivo. (A) β-Galactosidase activities of the amdS–lacZ fusion gene with the wild-type or mutant promoter under uninduced and induced growth conditions (see Materials and methods). The amdS promoter of these constructs is depicted showing the amdI9 (I9), amdI18 (I18), amdI15 (I15), amdI41(I41) and amdI93 (I93) mutations and their nucleotide coordinates relative to the translation initiation site. The strain designation for each of these gene replacement transformants is indicated, as is the facB genotype. β-Galactosidase-specific activities as A420 nm/103/min/mg of soluble protein are shown for each strain under uninduced (0.1% glucose) and induced (0.1% glucose, 50 mM sodium acetate) conditions, with standard errors indicated in parentheses. (B) β-Galactosidase activities of the amdS–lacZ fusion gene as descibed in (A). The amdS promoter of these constructs is depicted showing the mutant deleted for sequences from −647 to −117 and two replacement constructs containing the 5:6 or 31:32 oligonucleotide pair in this deleted promoter. Download figure Download PowerPoint Figure 4.FacB binds to mutant amdS promoters with different affinities. (A) EMSA comparing the amdI9 mutation in the upstream FacB-binding site and the amdI18, amdI15 and amdI41 mutations in the downstream FacB-binding site with the wild-type sites. The probes used are indicated beneath each set of three lanes, with the coordinates relative to the translation start of amdS shown below this. The three lanes represent no added protein extract (no extract), 6 μg of extract containing the MBP protein and 6 μg of extract containing the MBP–FacB fusion protein. (B) The amdS 5′ region showing the translation initiation site (ATG) and the location of the amdI9 and amdI18 single base substitutions and the amdI15 and amdI41 insertion mutations. The coordinates relative to the translational initiation site are shown below these sites. The two regions used as probes in (A) are shown with their end coordinates marked. Download figure Download PowerPoint Table 1. 5′ amdS mutations affecting FacB binding Mutation Sequencea Wild-type −222 GTTCTGCAGCTTTCCTTGGCCCGTAA −197 amdI9 −222 GTTCTGCAGCTTcCCTTGGCCCGTAA −197 Wild-type −130 CACCATCCGCTCCCCCGGGATCAATG −105 amdI18 −130 CACCATCCGCTCaCCCGGGATCAATG −105 amdI15 −130 CACCATCCGCTCCCCCcgggatcccgGGGATCAATG −105 amdI41 −130 CACCATCCGCTCCCatctggatcctaGtGATCAATG −105 aMutations affecting FacB binding are shown with coordinates relative to the translational startpoint. Substitution and insertion mutations are in lower case and bold. The amdI9 mutation is a single base pair change of T to C at −210 (Hynes et al., 1988). The amdI18 mutation is a single base change of C to A at −118 (Corrick et al., 1987) and the amdI15 mutation is a 10 bp insertion of a BamHI linker between bases −115 and −114 (Littlejohn and Hynes, 1992). The amdI41 mutation is a 10 bp insertion between bases −115 and −114 and base substitutions at −116, −115 and −113. Two in vitro-generated mutations, amdI15 and amdI41, with 10 bp insertions of different sequence between −115 and −114 (Table I) in an amdI9-containing promoter, resulted in reduced levels of reporter gene expression (Figure 3A). The amdI41 mutation, but not the amdI15 mutation, also reduced acetate induction to <2-fold. Binding by MBP–FacB to −147 to +22 fragments from amdI15 and amdI41 was reduced compared with that for wild-type (Figure 4), indicating that these mutations reduced affinity for MBP–FacB in vitro. The amdI18 mutation, a C→A substitution at −118, results in elevated amdS expression under a number of growth conditions but has not been found to be dependent on the facB gene (Hynes, 1978; Corrick et al., 1987). Since it is located within the 31:32 oligonucleotide which binds to MBP–FacB, it was tested in the in vivo assay. The amdI18 mutation increased amdS–lacZ expression by <2-fold under limiting glucose plus acetate-inducing conditions even in the presence of the facB101 mutation, indicating that the effects of this mutation were independent of facB (Figure 3A). In vitro binding of MBP–FacB to wild-type and amdI18-containing −147 to +22 fragments revealed similar levels of binding (Figure 4). The amdI93 deletion of −182 to −151 decreases the spacing between the FacB-binding sites in the −219 to −199 and −134 to −105 regions by 31 nucleotides (Hynes et al., 1988). The amdI93 deletion results in greatly reduced amdS–lacZ levels due to removal of the sites of action for the positively acting proteins AnCF and AmdR (Hynes et al., 1983; Littlejohn and Hynes, 1992; Papagiannopoulos et al., 1996). However, increased acetate induction of 9.7-fold was observed (Figure 3A). Thus, the relative proximity of FacB-binding sites in the amdS promoter may be important for acetate induction. To test whether a single FacB-binding site was sufficient for acetate induction of amdS–lacZ expression, the oligonucleotide pairs 5:6 (−219 to −199) and 31:32 (−134 to −105) were used to replace −646 to −118 sequences in amdS–lacZ gene replacement derivatives. Each of these oligonucleotides failed to confer acetate induction on the amdS–lacZ reporter (Figure 3B), suggesting that two binding sites are required for acetate induction by FacB of the amdS gene. FacB binds to dissimilar sequences in 5′ amdS DNase I footprinting of the wild-type promoter from −268 to +22 identified a single protected region from −140 to −105 on the coding strand and −143 to −104 on the non-coding strand (Figure 5A) overlapping sequences defined by the 31:32 oligonucleotide (−134 to −105). Footprinting of a −268 to +22 fragment derived from the amdI18 mutant promoter identified exactly the same protected regions as those for the wild-type (Figure 5C). In contrast, footprinting of the amdS promoter derived from the amdI9 mutant identified a protected region from −221 to −192 on the coding strand and −227 to −191 on the non-coding strand, encompassing the amdI9 T→C transition at −210 (Figure 5B). Thus, in vitro, MBP–FacB binds preferentially at −221 to −192 in the presence of the amdI9 mutation. The reason for this preferential binding in vitro is not clear but may reflect a lack of sensitivity in the footprinting such that only the strongest sites are detected. Figure 5.DNase I protection footprinting of wild-type and mutant FacB-binding sites. DNase I protection footprints of the amdS wild-type (A), amdI9 mutant (B) and amdI18 mutant (C) promoters from −268 to +22 for the coding and non-coding strands are shown. Footprinting using 75 μg of extract containing only the MBP protein or containing the MBP–FacB fusion protein are shown for each probe. A diagrammatic representation of the promoter is shown flanking the footprinted pair of strands such that coordinates denote the two FacB-binding sites in these probes while open rectangles represent the sites evident in the footprinting reactions. The mutations defining amdI9 and amdI18 are presented. All coordinates are relative to the translation initiation start site. Download figure Download PowerPoint Missing contact interference footprinting using the wild-type amdS probe showed that removal of bases from −128 to −111 on the coding strand and −130 to −117 on the non-coding strand significantly affected the ability of the fusion protein to bind DNA (Figure 6A). Similarly, when the amdI9 probe was used, bases in the region from −213 to −200 on the coding strand and −213 to −199 on the non-coding strand were important for binding (Figure 6B). Immediately outside these two regions, small effects on binding were evident. These results are consistent with results of the DNase I footprinting and EMSAs, defining these two dissimilar sequences as bona fide FacB DNA-binding sites. Moreover, both of these sites are functional in vivo (see above). Figure 6.Missing contact footprinting of the two FacB-binding sites. Missing contact footprints of the amdS wild-type (A) and amdI9 mutant (B) promoters from −268 to +22 for the coding and non-coding strands using the MBP–FacB fusion protein are shown at the top. Sequences and nucleotide coordinates encompassing the footprinted region are shown to the right of each footprint, with bases having a major effect on FacB binding marked by a filled circle and those with a minor effect by an open circle. Lanes corresponding to reactions using DNA modified by depyrimidation (Y) or depurination (R) are marked. Lanes from the FacB bound (B) and free (F) fractions are also marked. Densitometric plots derived from the missing contact footprints are shown below these footprints. The sequence of the footprinted region is shown along the inner x-axes for both strands, while nucleotide coordinates are shown on the outer x-axes. The y-axis (ordinate) represents a relative scale of the ratio of free probe to bound probe, corrected for loading differences, such that modified nucleotides which reduce binding by FacB to this probe will have a larger positive value. All coordinates are relative to the translation initiation start site of amdS. Download figure Download PowerPoint Competition experiments (data not shown) revealed that the MBP–FacB binding to labelled 5:6 could be competed by unlabelled 31:32 and, conversely, binding to labelled 31:32 was competed by unlabelled 5:6. These data and the requirement for an intact Zn(II)2Cys6 cluster for binding to the 5:6 and 31:32 oligonucleotides (see above; Figure 2C) indicate that the Zn(II)2Cys6 DNA-binding domain binds both types of dissimilar FacB DNA-binding site. FacB binds multiple regions in the promoters of the co-regulated facA, acuD and acuE genes To identify additional FacB DNA-binding sites, promoter-scanning EMSAs on the acuD, facA and acuE 5′ regions were performed (Figure 7). For 5′ acuD, two overlapping fragments and an adjacent fragment showed strong binding by FacB (−844 to −414, −414 to −202 and −465 to −176), thereby identifying at least two FacB sites (Figure 7B). In addition, the −465 to −176 acuD probe resulted in a major fast migrating complex as well as a slower migrating complex suggestive of at least two binding sites in this region. DNase I footprinting of this fragment identified three protected regions from −449 to −422, −333 to −305 and −295 to −269 on the coding strand, and −452 to −423, −333 to −305 and −292 to −266 on the non-coding strand (Figure 8). The central protected site showed homology to the −213 to −199 amdS site, while the other two sites showed homology to the −130 to −111 amdS site. Most of the facB-dependent acetate induction has been localized to the −414 to −202 region of the acuD promoter (Bowyer et al., 1994; De Lucas et al., 1994). Two of the protected regions in 5′ acuD, −333 to −305 and −292 to −269, were affected by the deletion of bases −347 to −285, which resulted in an 80% reduction of acetate-induced levels of an acuD–lacZ fusion gene (Bowyer et al., 1994). Thus, at least one of these FacB DNA-binding sites is functional in vivo. Figure 7.FacB binds to distinct sites in the promoters of the genes of the glyoxylate bypass. Promoter-scanning EMSA of the facA (A), acuD (B) and acuE (C) 5′ regulatory regions. End-labelled probes were generated by PCR or as restriction fragments (see Materials and methods). Probes used are indicated beneath each set of three lanes by the end coordinates relative to the translational start site. Each set of three lanes represents no added protein extract (no extract), 6 μg of extract containing the MBP protein and 6 μg of extract containing the MBP–FacB fusion protein. A diagrammatic representation of the facA, acuE and acuD 5′ regulatory regions with the relevant restriction sites and their nucleotide coordinates relative to the translational initiation site (ATG) is shown. The position of probes used for EMSAs is shown, with FacB-specific DNA binding indicated as strong (+), weak (+/−) and undetectable (−). Download figure Download PowerPoint Figure 8.DNase I protection footprints of representative regions of the (A) acuE and (B) acuD promoters for the coding and non-coding strands are shown. Probes represent the acuD promoter region from −465 to −176 and the acuE promoter region from −684 to −404. Footprinting reactions used 75 μg of extract containing only the MBP protein or containing the MBP–FacB fusion protein. A representation of each promoter is shown flanking the footprinted pair of strands such that open rectangles denote the sites evident in the footprinting reactions. All coordinates are relative to the translation initiation start sites of acuD and acuE. Download figure Download PowerPoint For 5′ facA, two partially overlapping fragments showed strong binding by FacB (−1265 to −962 and −1127 to −677) while a third exhibited weak binding (−481 to −168) in EMSAs, thereby identifying at least two FacB sites (Figure 7A). For 5′ acuE, two partially overlapping fragments showed strong binding by FacB (−1010 to −659 and −684 to −404) while a third exhibited weak binding (−404 to −228), thereby identifying at least two FacB sites (Figure 7C). DNase I footprinting was performed on the −684 to −404 region of 5′ acuE (Figure 8). A single protected region was evident from −645 to −614 on the coding strand and from −650 to −612 on the non-coding strand. The site showed homology to the −130 to −111 amdS site. These DNA-binding analyses of 5′ acuD, facA and acuE clearly support the proposal of direct transcriptional control by FacB, and the sites of action for these genes fall into the two distinct classes of FacB-binding sites defined for the amdS promoter. Discussion FacB DNA-binding sites identified by DNase I footprinting with MBP–FacB fall into two classes according to sequence (Table II). Class A FacB-binding sites show sequence homology to the binding site at amdS −125 to −111, while class B FacB sites show sequence homology to the −219 to −203 amdI9 sequence, both of which contact MBP–FacB in missing contact interference assays. The class A and class B consensus sequences derived for footprinted sites are TCC/ GN8–10C/GGA and GCAGNTNNCCN1–2GGC, respectively. Both class A and class B sites show imperfect rotational symmetry. Sequences of fragments which showed MBP–FacB-specific binding in EMSAs, but were not subjected to footprinting analysis, were scanned for regions of homology to class A and class B sit