OUP user menu

Analysis of CcpA mutations defective in carbon catabolite repression in Bacillus megaterium

Alexandra Kraus, Wolfgang Hillen
DOI: http://dx.doi.org/10.1111/j.1574-6968.1997.tb10485.x 221-226 First published online: 1 August 1997


Five mutations in ccpA of Bacillus megaterium with impaired functions were analysed for carbon catabolite repression. The phenotypes support the hypothesis that CcpA assumes a PurR/LacI fold. The completely inactive mutants CcpA119GE and CcpA326am cause alterations which are incompatible with that fold. A mutation with reduced activity, CcpA81GE, affects a site that would be partially surface exposed and may interfere with structure formation or cofactor binding. A mutation in the putative hinge α-helix, CcpA52AE, is negative transdominant over wild-type ccpA. The mutant CcpA38am is inactive, although reduced amounts of wild-type size protein are produced.

1 Introduction

Carbon catabolite repression (CCR) in Bacillus subtilis, B. megaterium and Staphylococcus xylosus is mediated by the catabolite responsive element cre and the trans-acting protein CcpA [3, 610]. cre may be located in the promoter region or in the translated sequence of the regulated gene or operon. Deletion or mutation of cre results in reduced CCR. This has been demonstrated for xylA of B. megaterium[6, 24], and for amyE[23], gnt[20], levR[19] and xylA[11, 13], all of B. subtilis.

cre is bound by CcpA, which belongs to the LacI/GalR family of bacterial repressors[27]. The purine repressor PurR of Escherichia coli[25] exhibits high homology to B. megaterium CcpA (33.7% identical residues). This is interesting since PurR needs a co-repressor to bind DNA[1]. Binding of CcpA to cre in the xyl and gnt sequence contexts is enhanced by the presence of HPr-Ser46-P [5, 6, 22], which also binds to CcpA in the absence of cre[2]. Binding of CcpA to cre of amyE was reported without a cofactor[12].

We isolated B. megaterium mutants which failed to display CCR of xyl operon expression and mapped five single mutations in ccpA. The effects of these mutations are consistent with the idea that the protein forms the same peptide fold as PurR[25].

2 Materials and methods

2.1 General methods

Recombinant DNA was carried out as described[10]. β-galactosidase activities were determined after growing cells to an optical density of 0.5 at 600 nm in M9 medium containing the indicated carbon sources[10].

2.2 Bacterial strains and plasmids and culture and growth conditions

All bacterial strains and plasmids used and constructed in this study are listed in Table 1. B. megaterium WH331 was the parental strain for all strain constructions described in this study. For DNA preparations E. coli and B. megaterium were grown in LB. Cells grown in M9 medium supplemented with 0.02% yeast extract and 0.02% casamino acids were used for β-galactosidase assays. 0.5% succinate was added to the M9 medium as a general carbon source. Medium for xylose induction contained 0.25% xylose (M9/X) and for glucose repression 0.25% xylose and 0.25% glucose (M9/XG) respectively. Screening on plates was done on M9/XG with 40 mg/l Xgal. Wild-type ccpA strains (xylA-lacZ) grow to white colonies on these plates. B. megaterium cells carrying plasmids were grown at 30°C, plasmid-free cells were grown at 37°C.

View this table:

Strains and plasmids

Strain/plasmidRelevant genotypeReference
B. megaterium WH331gdh(xylA-lacZ)[10]
B. megaterium WH356WH331, ΔccpA[10]
B. megaterium WH398WH331, xylA-catThis work
B. megaterium WH331CcpA52AEWH331, ccpA 155 C→AThis work
B. megaterium WH331CcpA81GEWH331, ccpA 242 G→AThis work
B. megaterium WH331CcpA119GEWH331, ccpA 356 G→AThis work
B. megaterium WH398CcpA38amWH398, ccpA 108 A missingThis work
B. megaterium WH398 CcpA326amWH398, ccpA 976 G→TThis work
pWH341neo, cat[14]
pWH1509Kamp, tet, neo, orits[24]
pWH1510amp, tet, xylA-lacZ[24]
pWH1510catamp, tet, xylA-cat-lacZThis work
pWH1518neo, tet, orits, xylA-lacZ[24]
pWH1518catneo, tet, orits, xylA-cat-lacZThis work
pWH1522amp, neo, orits, xylA-catThis work
pWH1564amp, his6-ccpA[15]
pWH1565amp, his6-ccpA 976 G→TThis work
pWH2005amp, neo, orits, ccpA[10]
pWH2041amp, neo, orits, ΔccpA[10]
pWH2060amp, neo, orits, ccpA 155 C→AThis work
pWH2061amp, neo, orits, ccpA 242 G→AThis work
pWH2062amp, neo, orits, ccpA 356 G→AThis work
pWH2063amp, neo, orits, ccpA 108 A missingThis work
pWH2064amp, neo, orits, ccpA 976 G→TThis work

2.3 Construction of B. megaterium WH398

Plasmid pWH1510 (xylA-lacZ) was used to construct a xylA-cat fusion by cloning a 780 bp Sau 3A fragment from pWH341 (carrying a promoterless cat-gene) into the Bam HI site of pWH1510, yielding pWH1510cat. Plasmid pWH1510cat was restricted with Sca I and Bss HII to obtain a 2520 bp xylA-cat-lacZ′ fragment, which was ligated into the integration vector pWH1518/Sca I/Bss HII resulting in plasmid pWH1518cat. The lacZ-gene of pWH1518cat was removed by digestion with Bst EII and Sau I, filling in the protruding ends and religation. The resulting plasmid, designated pWH1522, was integrated into the xyl operon of WH331, leading to strain WH398. Southern blot analysis confirmed the desired construction. The strain tolerates 5 mg/l chloramphenicol under xylose-induced conditions, but is sensitive to the drug when glucose is added.

2.4 Construction of plasmids pWH2061–pWH2064

Chromosomal DNA of the mutant strains WH331CcpA81GE, WH331CcpA119GE and WH398CcpA38am was used to amplify mutant ccpA fragments by PCR. Primer PHind (GGAAGCTTGATTTATATACG) originates from a sequence upstream of the ccpA promoter and creates a Hin dIII restriction site. Primer Pm2 pairs downstream of the singular Sac II site in ccpA (TAATTTCTAGATCGTTTG). PCR products were restricted with Hin dIII and Sac II and the desired 864 bp fragments ligated into pWH2005/Hin dIII/Sac II. The resulting plasmids were designated pWH2061, pWH2062 and pWH2063, respectively. Chromosomal DNA of WH398CcpA326am was amplified with primer Pp4 (CAGATATTCGCTATCAT, upstream of Sac II) and primer Pm6 (CCGTTTTACCAGC, downstream of an Ase I site within the ccpB sequence). The 786 bp Sac II/Ase I fragment was ligated into pWH1564/Sac II/Nde I. The resulting plasmid pWH1565 was restricted with Sac II and Nhe I, and the resulting 627 bp fragment was ligated into pWH2005/Sac II/Nhe I yielding pWH2064.

3 Results and discussion

3.1 Isolation and sequence analysis of ccpA mutants

Two different approaches were taken to randomly isolate mutants defective in CCR: The first involved mutagenesis of strain WH331 (xylA-lacZ) with EMS followed by screening for blue colonies on plates containing xylose and glucose as previously described[10]. Three independent candidates were isolated, two of which turned out to be identical. The second approach made use of chloramphenicol sensitivity of strain WH398 (xylA-cat in addition to xylA-lacZ) grown in the presence of glucose. We plated overnight cultures on minimal medium containing xylose, glucose and chloramphenicol and isolated two spontaneous CmR mutants which failed to display glucose repression.

All five candidates were complemented by pWH2005 (wild-type ccpA), suggesting that they contain mutations in ccpA. The mutant ccpA alleles were isolated by PCR and sequenced. Three candidates from EMS mutagenesis contained a single G→A transition in the Gly codon GGA to the Glu codon GAA at amino acid positions 81 (two clones) and 119 of CcpA. The corresponding mutant strains were named WH331CcpA81GE and WH331CcpA119GE, respectively. One spontaneous ccpA mutant designated WH398CcpA38am contained a frameshift caused by deletion of an A-residue in a run of seven As, leading to an amber codon at position 38 and replacement of Lys37 by Tyr. G→T transversion yielded an amber codon at position 326 in the other spontaneous mutant, called WH398CcpA326am. This results in a truncated CcpA lacking seven amino acids at the C-terminus.

Another mutant was obtained when strain WH398 harbouring plasmid pWH2005 was mutagenized with EMS and screened for blue colonies. It showed partial relief of glucose repression. Curing the plasmid from the cells fully restored glucose repression. Thus, the phenotype was dependent on a plasmid-encoded ccpA mutation, as was confirmed by retransformation into WH356 (ΔccpA) and WH331 (wild-type). The mutant ccpA did not complement WH356 and reduced glucose repression in WH331. This indicated that the plasmid encodes a non-functional CcpA which is negative transdominant over wild-type. The mutation causes an exchange of Ala52 to Glu, called CcpA52AE, and the plasmid was designated pWH2060. ccpA on the chromosome of WH331 was replaced by the ccpA 52AE allele by marker exchange with pWH2060. A candidate giving blue colonies on M9/XG and sensitivity to neomycin was designated WH331CcpA52AE. Bands of similar intensities and CcpA molecular masses were found in Western blots for the wild-type and the mutants 52AE, 81GE, 119GE and 326am (data not shown). The solubilities of the CcpA mutants were determined after the cells were disrupted by sonication and the insoluble proteins were pelleted by centrifugation. The Western blot analyses of both fractions revealed some insoluble protein for the 52AE, 81GE, 119GE and 326am mutants, whereas CcpA119GE is about evenly distributed between the soluble and insoluble fractions.

3.2 Quantification of glucose repression exerted by the ccpA mutants

CCR exerted by the CcpA mutants was scored in B. megaterium strains with a single copy xylA-lacZ fusion integrated downstream from the gdh transcriptional terminator in the chromosome. β-galactosidase expression levels of the five chromosomally located CcpA mutants and the control strains WH331, WH398 (both wild-type ccpA) and WH356 (ccpAΔ 95–261) were determined after growth in medium containing xylose or xylose and glucose and are shown in Fig. 1. β-galactosidase expression levels on xylose were nearly the same in all strains. Full glucose repression as defined in the wild-type strains is severely impaired in all ccpA mutants. Except for WH331CcpA81GE with a residual 2.2-fold repression, all other mutants completely lack glucose repression.

Figure 1

Regulation of xylA-lacZ expression in B. megaterium containing wild-type and mutant ccpA alleles. Hatched bars indicate the β-galactosidase activities obtained under induced conditions (X), solid bars represent the β-galactosidase activities obtained in the presence of xylose and glucose (XG). The strain designations and CcpA mutations are indicated at the bottom.

3.3 Negative transdominance of the ccpA mutants

The mutant ccpA alleles 81GE, 119GE, 38am and 326am were cloned into pWH2005, yielding pWH2061 to pWH2064, which were transformed into WH331, as was done for pWH2060 (CcpA52AE). The transformants contain multiple copies of the plasmid encoded ccpA mutants and a single copy of endogenous wild-type ccpA. Transformants with pWH1509K and pWH2005 and pWH2041 served as controls. The results of the β-galactosidase assays are presented in Table 2. Only CcpA52AE on plasmid pWH2060 affected glucose repression significantly by increasing the β-galactosidase activity about 2-fold. Thus, CcpA52AE is negative transdominant over wild-type indicating an oligomeric structure of the active protein in vivo. This agrees with the dimeric structure found for B. subtilis CcpA in vitro[21].

View this table:

Test for negative transdominance of ccpA mutants over wild-type ccpA in WH331

PlasmidPlasmid encoded CcpAß-gal. activity (units)
InducedRepressedFold repression
pWH2041Δ 95–2612100±120220±89.5

3.4 Structural interpretation of CcpA-mutants

Extensive mutational analysis of LacI has helped to outline structurally and functionally important residues in that protein, and the alignment with other members of the family showed that structurally important residues are conserved[18]. The importance of such residues was understood after the crystal structures of LacI [4, 16] and PurR[25] had been solved. We take here the opposite approach by asking if the mutations in CcpA actively interfere with a LacI/PurR-like structure, or if they have functional implications. Gly81 and Gly119 are located in the N-terminal portion of the co-repressor binding domain. Both residues are highly conserved, or only similar amino acids like Ala are found in other proteins of the LacI/GalR family. Replacing the hydrogen by the larger Glu side chain in the PurR fold results in sterical interference with α-helix IX. CcpA 81GE exhibits residual repression, whereas CcpA119GE is completely non-functional. Gly81 is located on the surface of the protein where the side chain in the mutant may be able to assume a position with less sterical hindrance. In LacI, Ala82 (corresponding to Gly81 in CcpA) is solvent exposed and can therefore tolerate most substitutions, however, some replacements lead also to an I phenotype[18], which corresponds to the lack of repression for CcpA. Gly121 (corresponding to Gly119 in CcpA) is buried in the interior of LacI and all mutations are non-functional[26], resembling the effect of CcpA119GE.

Although CcpA326am lacks only the last seven amino acids, it is non-functional. These missing amino acids are located on the surface of the C-terminal domain, assuming that the PurR structure applies to CcpA. Their removal exposes hydrophobic residues, such as Ile272, which are otherwise secluded in the interior. The deleted region contains a conserved Arg residue (326 in LacI, 328 in CcpA), substitution of which resulted in unstable LacI-proteins with reduced affinities for IPTG and operator. Deletion of C-terminal amino acids including Arg326 in LacI yielded inactive polypeptides[17]. Nevertheless, the truncated CcpA protein is soluble; therefore, we cannot exclude that the missing residues might play functional rather than structural roles.

Negative transdominant mutants of regulatory proteins are often located in the DNA binding domain, as is the case for CcpA52AE. Position 52 of CcpA corresponds to position 51 of PurR which is located in the so-called hinge helix between the helix-turn-helix motif and the core of the protein[25]. It binds into the minor groove of DNA and bends it open by intercalation of Leu54 into the central CG of the operator. Both residues (Ala52 and Leu55 for CcpA) are highly conserved in the LacI/GalR family[27]. When Ala 51 is replaced by Glu in the PurR structure, the bulky Glu residue (in CcpA52AE) cannot get into close proximity of the DNA, thus interfering with intercalation of Leu55. Furthermore, the negative charge is repulsive to the DNA phosphates. Consequently, this variant should be incapable of binding DNA, but should still form oligomers, which is in agreement with the phenotype of CcpA52AE. The majority of amino acid replacements in the hinge helix of PurR also interfered with regulation[1].

The phenotypes of the CcpA mutants discussed in this study can be explained by assuming that CcpA and PurR or LacI, respectively, share a common fold and therefore support the hypothesis that CcpA belongs to the LacI/GalR family of bacterial regulators.


We thank Drs. Maria Schumacher and Richard G. Brennan for providing the PurR coordinates and many fruitful discussions. We thank Dr. Jörg Stülke for fruitful discussions. This work was supported by the EC through the Biotech program, the Deutsche Forschungsgemeinschaft and the Fonds der chemischen Industrie.


View Abstract