Mutations in Catabolite Control Protein CcpA Separating Growth Effects from Catabolite Repression

Random mutagenesis of ccpA. We randomly mutagenized theccpA gene and set up an in vivo screen for novel or altered functions. The ccpA gene, including its putative promoter, was cloned into M13mp18 (34), and single-stranded DNA of the resulting phage mWH560 was treated with nitrite and amplified by error-prone PCR, as described previously (26). The resulting fragments were cloned into pWH2051 to give a pool of randomly mutatedccpA genes. The 5′ flanking region of ccpA in pWH2051 is shorter than in the otherwise isogenic plasmid pWH2005, which has been used for regulatory studies of CcpA (11, 16). However, CCR mediated by pWH2005 or pWH2051 carrying the wild-type gene is identical (data not shown).

Screening for ccpA alleles that differentially affect CCR and growth.We screened the mutant pool for ccpAalleles that allow a chromosomal ccpA deletion mutant to grow normally on glucose as the sole carbon source without restoring CCR of the xyl operon (ccpA ^c, defective in CCR only). We also looked for strains deficient in growth but active in CCR (ccpA ^g, deficient in growth only). For this purpose, we constructed B. megaterium WH353 [lac ΔccpA gdh2φ(xylA1-spoVG-lacZ) ΔxylR] carrying a xylA-lacZ fusion and chromosomal deletions in ccpA and xylR to observe only the effects exerted by plasmid-encoded ccpA. The bacteria were transformed with the mutant pool of ccpA by protoplast transformation (28). After regeneration, the cells were plated on M9 minimal medium containing glucose (0.2%) as a regulatory carbon source, succinate (0.5%) as a carbon source which is not effective in CCR, and X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) (80 mg/liter). We screened for large blue colonies (ccpA ^c) and small white colonies (ccpA ^g). Plasmid DNA was isolated, passaged through Escherichia coli, and again transformed intoB. megaterium WH353 or WH419 [lac ΔccpA gdh2φ(xylA3-spoVG-lacZ)] (kindly provided by A. Wagner, Erlangen, Germany), and the phenotypes were reconfirmed on plates. Since B. megaterium WH419 carries a functionalxylR gene, all experiments with this strain were carried out in the presence of 0.5% xylose in the growth medium. Results obtained with WH419 and WH353 were identical. The screen yielded sixccpA ^c alleles and oneccpA ^g mutant from about 2,500 primary transformants.

Characterization of ccpA mutations affecting catabolite repression only (ccpA ^c).Sequencing of theccpA alleles revealed that all ccpA ^cmutants carried multiple mutations, as listed in Table1. We separated the mutations by combining ccpA ^c alleles with the wild type at the unique AflII site, followed by screening as described above. The subclones were named for the parental mutation, with the addition of “n” or “c”, indicating an N- or C-terminal mutation with respect to the restriction site. The superscript “c” was retained for those alleles that confer the ccpA ^cphenotype. The relative growth rates for the original mutants and all subclones were estimated from colony sizes on plates and are noted in Table 1. The ccpA ^c mutants were indistinguishable in colony size from those containing wild-typeccpA.

To quantify CCR exerted by ccpA alleles, β-galactosidase activity in M9 medium with succinate or glucose and succinate as sole carbon sources was recorded. As expected from their blue phenotype on plates, the original ccpA ^c mutants show reduced CCR compared to wild-type ccpA, even though the remaining repression efficiency is higher than that of the ΔccpAmutant (Table 1). The mutants were considered CCR negative if they exerted 25% or less wild-type repression. The subclones of two multiply mutated ccpA ^c alleles (ccpA ^c 9 andccpA ^c 12) did not show reduced CCR upon separation of the original mutations (Table 1). ThreeccpA alleles with single mutations (ccpA ^c 10n,ccpA ^c 19n, andccpA ^c 20n) conferred wild-type-like growth along with reduced CCR (Table 1). To ensure that effects were not caused by different levels of CcpA expression, we performed Western blot analyses with extracts of total cell protein, as described previously (16). All of the mutant ccpA alleles expressed CcpA to levels similar to the wild type (data not shown).

For those ccpA ^c mutants caused by a single-amino-acid exchange, the growth phenotype was quantified on M9 minimal medium with glucose (0.5%). Growth on plates was scored by determining the number of CFU per colony (Fig.1). The values reflect the colony sizes given in Table 1. The growth rates in liquid medium (Fig. 1) are consistent with these observations for mutants CcpA^c10n and CcpA^c19n, which grow at the same rate as the wild type. CcpA^c20n exhibits slower growth than the deletion mutant in liquid culture, which is in contrast to its behavior on solid medium. The activity of CcpA in CCR is dependent on the agitation of liquid cultures, with activities at low levels of agitation being comparable to those observed on solid media (2). The growth differences in CcpA^c20n could be due to a similar effect, which may imply that growth regulation by CcpA^c20n is sensitive to environmental conditions.

• Open in new tab
• Download powerpoint

Fig. 1.

Growth of selected mutants on solid and liquid minimal media with 0.5% glucose as the sole source of carbon. To score growth on solid medium, cells were streaked on plates and incubated at 30°C for 20 h. Single colonies were excised with the surrounding agar and resuspended in liquid medium. The number of CFU was determined by plating aliquots from serial dilutions (hatched bars). Growth rates in liquid medium were determined at 32°C with vigorous agitation (open bars).

To interpret possible defects caused by the mutations, we made use of the common three-dimensional fold of LacI and PurR, which probably holds for the entire LacI/GalR family of bacterial regulators, including CcpA (32), and is supported by limited proteolysis (12) and mutational data (16, 17). Their functional implications are discussed in the light of structure-function relationships in LacI (5, 20) and PurR (30).

The three singular ccpA ^c mutations are located in the N-terminal DNA-binding domain of CcpA, as depicted in Fig.2. The Thr4-to-Ser exchange in CcpA^c10n aligns with Thr3 of PurR, where it is located N-terminally to the positioning α-helix 1 of the helix-turn-helix motif (Fig. 2) and does not contact DNA (30). Mutational analysis of LacI revealed that the equivalent Thr5 residue could be replaced only by Ser, yielding a partially active repressor (15). Our observation of diminished repression ofxylA-lacZ by CcpA^c10n resembles that result.

• Open in new tab
• Download powerpoint

Fig. 2.

Sequence alignment of the N termini of B. megaterium (Bme) and B. subtilis (Bsu) CcpA and the purine and lactose repressors of E. coli (Eco). Identical amino acids are shaded. Numbered boxes below the sequences denote α-helices, as determined for PurR. The positions of amino acid exchanges of mutants isolated in this study are given above the sequences.

Arg47 is mutated to His in CcpA^c19n. The equivalent residues Ser46 of PurR and Ile48 of LacI are involved in anchoring the DNA-binding domain to the core of the protein (30, 31). Any mutation at this position is likely to affect the positioning of the headpieces relative to the protein core and thus their alignment on the operator.

The ccpA ^c20n mutation, Asn49 to Ser, is located next to the hinge helix which connects the DNA-binding domain to the core of the protein and contacts DNA. Asn residues are frequently found at this position in the LacI/GalR family (32). PurR contains a Ser which contacts the DNA at a phosphate residue (30). Thus, the Asn49-to-Ser exchange in CcpA may affect DNA binding. In contrast to this, the corresponding residue in LacI, Asn50, is involved in interdomain contacts between the headpiece and the core (20,31). The LacI Asn50-to-Ser mutant is fully active, while other residues at position 50 yield inactive mutants (15). Thus, the underlying reason for the reduced CCR may be altered DNA binding, but this conclusion is not unambiguously deducible from these data.

Taken together, all three ccpA ^c mutations are at positions likely to directly or indirectly affect DNA binding. This relates the loss of CCR to impaired recognition of cre. The fact that loss of CCR is not accompanied by a growth defect may be explained in two ways. Either DNA sequences recognized for growth regulation are different from the cre mediating CCR inxyl or from the cre consensus sequence in general or the interaction with possibly unknown cofactors leads to a titration effect that is lacking in the ΔccpA mutant but not in theccpA ^c mutants.

Characterization of a ccpA mutant affecting growth only (ccpA ^g).The only CcpA^gmutation, CcpA^g14, limited growth to small colonies on all media tested, but not as small as those of WH353 (ΔccpA). Therefore, the growth phenotype ofccpA ^g 14 is designated “+/−” in Table 1. Quantification of its growth behavior on solid and liquid media led to values similar to those of the deletion mutant. The β-galactosidase activity of the xylA-lacZ fusion shows that the ccpA ^g 14 mutant is even more efficient in repression than is the wild type (Table 1). CcpA^g14 is expressed to a level similar to that of the wild-type protein, as determined by immunoblotting (data not shown).

The ccpA ^g 14 mutation leads to a single-amino-acid exchange, Ala17 to Thr. This residue is located in the recognition helix (Fig. 2). The analogous Thr16 of PurR is involved in protein-DNA interaction and contacts two bases directly (30). The equivalent Gln18 of LacI also determines operator recognition (19, 20). One would therefore assume that theccpA ^g 14 mutation leads to an altered DNA-binding specificity of CcpA. However, because catabolite repression of xylA is even somewhat more effective withccpA ^g 14 than in the wild type,cre of xylA must be efficiently recognized. Thus, the growth rate reduction may not involve DNA binding, although the properties of this mutant would also be consistent with speculation that a different cis-acting sequence mediates growth rate control.

We have established for the first time that growth rate control and CCR are separable functions of CcpA. The underlying molecular basis could be interaction with different signal molecules, differences in signal transfer from the effector binding site to the headpiece, or recognition of different cis-acting sequences. The four single mutants are affected in amino acids involved in DNA recognition or interdomain contacts relevant for recognition of DNA. Although altered cofactor binding cannot be rigorously ruled out, binding of CcpA to as-yet-unidentified DNA sites is likely to be part of its function in growth regulation.