trans-Acting Factors and cis Elements Involved in Glucose Repression of Arabinan Degradation in Bacillus subtilis

cis-Regulatory elements in the promoter regions of abnA and xsa.

The involvement of two cres present in the araABDLMNPQ-abfA operon (cre araA and cre araB) (Fig. 1) in CCR was shown in previous studies (12, 17). Both cre araA, located between the promoter region and the araA gene, and cre araB, placed 2 kb downstream within the araB gene, are independently functional, and both contribute to glucose repression (12). Moreover, the level of glucose repression, 21.8-fold, measured in a strain bearing an abfA′-lacZ reporter fusion (IQB450) (23; see below) was only 1.6-fold higher than that observed with a strain bearing an araAB′-lacZ fusion, i.e., 13.4-fold (12), suggesting the absence of other cre active sites in glucose-mediated CCR of abfA expression. Expression of abnA and xsa is also subject to glucose repression (23). Potential cre sequences, namely, cre abnA (+83-TGTAAGCGCTTTCT [SubtiList position 2949135]) (Fig. 1) (17) and cre xsa (+1-TAAAAGCGCTTACA [SubtiList position 2914316]) (Fig. 1) (18), are present in the promoter regions of the genes. To assess the functionality of these putative regulatory sites, we introduced a single-base-pair substitution which destroyed the central symmetry in both cre abnA (+89 C→A) and cre xsa (+7 C→A) (Fig. 1). The two promoter regions, together with the respective mutated cres, were fused to the lacZ reporter gene and introduced into the amyE locus. In the resulting B. subtilis strains, IQB472 and IQB473 (Table 1), respectively, the levels of accumulated β-galactosidase activity in the absence or presence of the inducer arabinose and under repressing conditions (arabinose plus glucose) were determined. Glucose repression was almost abolished in the mutants compared to that in strains bearing the wild-type promoters (IQB410 and IQB405) (23) (Table 2). These results indicate that both cre abnA and cre xsa are functional and that the designed mutation prevented binding of CcpA (see below) in the presence of glucose. Thus, cre abnA and cre xsa are the cis-acting elements involved in glucose-mediated CCR of abnA and xsa expression at the transcriptional level. In addition, the results suggest that only one cre in each promoter is responsible for this phenomenon.

• Open in new tab
• Download powerpoint

FIG. 1.

Organization of the B. subtilis chromosome region comprising the arabinan-degrading enzyme genes abnA, abfA, and xsa. The three genes are represented by dotted arrows pointing in the direction of transcription. abfA belongs to the araABDLMNPQ-abfA metabolic operon, abnA is located immediately upstream, and xsa is positioned 23 kb downstream of the metabolic operon. Hairpin structures denote terminators. The dotted boxes below the physical map represent extensions of the inserts fused to the lacZ reporter gene in the indicated plasmids. Plasmid pSA1 was integrated into the host chromosome by means of a single-crossover (Campbell-type) recombinational event that occurred in the region of homology of the resulting strain (Table 1). Linearized DNAs from plasmids pSN40, pZI26, pZI27, and pRIT1 were used to transform B. subtilis strains (Table 1), and the fusions were integrated into the chromosome via double recombination at the amyE locus. Above the map, the DNA sequences of cre abnA and cre xsa are depicted, and the numbers indicate the positions of the abnA and xsa genes relative to the transcriptional start site. The single nucleotide change (C→A) introduced into each cre is shown.

A major role of CcpA and different contributions of effectors HPr and Crh to CCR.

The arabinan-degrading enzyme genes are subject to temporal regulation, since the levels of expression of abfA, xsa, and abnA increase at early postexponential phase in response to arabinose and arabinan (23). Here we analyzed this effect in the presence of glucose. The expression of abfA′-lacZ, xsa′-lacZ, and abnA′-lacZ transcriptional fusions, integrated into the chromosomes of strains IQB450, IQB405, and IQB410, respectively (Table 1) (23), was analyzed as described above, and samples were collected 2 and 4 h after induction (t ₂ and t ₄, respectively), corresponding to the middle exponential and early postexponential (transitional) growth phases, respectively (Fig. 2A). Although expression of abfA′-lacZ, xsa′-lacZ, and abnA′-lacZ increased at t ₄, the glucose repression indexes, measured as ratios between expression in the presence of arabinose and the values obtained in the presence of arabinose plus glucose, were similar during exponential growth and transitional phase (Fig. 2B, C, and D). These results suggest that the effect of glucose repression under these conditions is not affected over time.

• Open in new tab
• Download powerpoint

FIG. 2.

Effects of ccpA, ptsH1, crh, hprK, and araR mutations on catabolite repression of the xsa, abnA, and abfA genes. (A) Growth curves for wild-type (squares) and ccpA-null mutant (triangles) B. subtilis strains in C minimal medium (21) supplemented with 1% (wt/vol) casein hydrolysate in the presence of 0.4% (wt/vol) arabinose plus 0.4% (wt/vol) glucose. Exponential growth phase (t ₂) and transitional phase (t ₄) are indicated. The strains bearing the xsa′-lacZ (B), abnA′-lacZ (C), and abfA′-lacZ (D) transcriptional fusions were grown in the absence of sugar, in the presence of 0.4% (wt/vol) arabinose, and in the presence of 0.4% (wt/vol) arabinose plus 0.4% (wt/vol) glucose. The levels of accumulated β-galactosidase activity, expressed as Miller units, were measured at hours 2 (t ₂) and 4 (t ₄) after induction (addition of arabinose). The results are represented by the glucose repression index, calculated as the ratio between the level of accumulated Miller units obtained in the presence of arabinose and the value determined in the presence of arabinose plus glucose. Each value is the average of two measurements each from three independent experiments.

To identify the trans-acting factors involved in CCR of the arabinan-degrading enzyme genes, we examined the expression of abfA′-lacZ, xsa′-lacZ, and abnA′-lacZ transcriptional fusions, as described above, in B. subtilis ccpA, pstH1, crh, hprK, and araR mutant backgrounds (Table 1). In a wild-type background, the addition of glucose caused 21.8-, 19.3-, and 6.5-fold repression of abfA′-lacZ, xsa′-lacZ, and abnA′-lacZ expression, respectively (strains IQB450, IQB405, and IQB410, respectively) during exponential growth phase (23). Disruption of the ccpA gene led to a complete loss of glucose-mediated CCR of expression from the abfA′-lacZ (IQB474), xsa′-lacZ (IQB423), and abnA′-lacZ (IQB422) fusions (Fig. 2), revealing its major role in glucose-mediated CCR of the arabinan-degrading enzyme genes. The ptsH1 mutation (HPr Ser46 to Ala), which impaired HPr Ser46 phosphorylation (5), partially abolished glucose repression of abfA′-lacZ (IQB475), xsa′-lacZ (IQB467), and abnA′-lacZ (IQB466) fusion expression (Fig. 2). In contrast, disruption of the crh gene caused almost no effect on glucose repression of expression from the abfA′-lacZ, xsa′-lacZ, and abnA′-lacZ fusions (IQB476, IQB469, and IQB468, respectively) during exponential growth phase (t ₂). Accordingly, in the ptsH1 crh double mutant background, the levels of glucose repression, at t ₂, of abfA′-lacZ (IQB477), xsa′-lacZ (IQB471), and abnA′-lacZ (IQB470) fusions were only slightly lower than those observed in the single ptsH1 mutant (Fig. 2), suggesting no or a small contribution of Crh to CCR mediated by glucose. These results are in agreement with previous observations that both CcpA and HPr, but not Crh, participate in glucose repression of the araE and araAB genes (12). However, in the crh-null mutant background, a significant relief of glucose repression of abfA′-lacZ and xsa′-lacZ expression, of 6.4-fold and 2.1-fold, respectively, was observed at transitional phase (t ₄), suggesting a contribution of Crh to glucose-mediated CCR at late growth phases (discussed below). The disruption of the hprK gene, encoding the bifunctional HPr kinase/phosphorylase, which reversible phosphorylates HPr and Crh (8, 16, 24), also caused a decrease of glucose repression (in IQB478, IQB479, and IQB480) (Fig. 2).

During exponential growth of cells with an araR-null mutant background, we observed a twofold derepression of expression of xsa′-lacZ (IQB406) and abfA′-lacZ (IQB453) in the presence of glucose compared to that of the wild type (Fig. 2B and C). AraR is a negative regulator of the arabinose-inducible genes (6, 7, 20), and this effect was previously observed with araE′-lacZ and araAB′-lacZ fusions (12). The AraR binding sites and cre sequences are located close to each other in the promoter regions of the araABDLMNPQ-abfA, araE, and xsa genes (12, 19, 23). Since under particular physiological conditions both CcpA and AraR might be bound to the DNA in close proximity, a possible interaction (cooperation) between the two proteins in the negative control of these genes was proposed (12, 19). Interestingly, this effect was not observed for the abnA′-lacZ fusion in an araR-null mutant background (strain IQB411) (Fig. 2D; Table 1). This observation is in agreement with a different localization of cre abnA relative to the AraR binding site in the abnA promoter and also to a weak control of abnA expression by AraR as a result of a distinct mechanism of action (23).

Crh expression increases at transitional growth phase.

The contribution of both effectors HPr and Crh to CCR by glucose can depend on the growth conditions, as shown for the hut operon, which responds to HPr in rich LB medium and to both HPr and Crh in minimal medium (36). Crh is the major effector for CCR of the citM gene in a medium containing succinate and glutamate (32). In connection with this, it was recently shown that the nature of the carbon source influences ptsH and crh expression (9). Using ptsH′-lacZ and crh′-lacZ transcription and translational fusion analysis, it was suggested that the different contributions of Crh and HPr to CCR might be due to the drastic differences in their synthesis rates under the conditions that cause CCR. In the presence of PTS carbohydrates, such as glucose, the synthesis level of HPr during exponential growth is 100-fold higher than that of Crh, which may explain the minor role of the latter in certain circumstances (9). Moreover, CcpA binds Crh(Ser-P) more weakly than HPr(Ser-P) (28). We have previously shown the absence of a Crh contribution to glucose repression of the araABDLMNPQ-abfA and araE genes in both C minimal medium supplemented with casein hydrolysate and CSK minimal medium (12). To further investigate the lack of Crh dependency during exponential growth and its possible contribution to glucose repression of the arabinan-degrading enzyme genes during the transitional growth phase, we analyzed the expression of both ptsH′-lacZ and crh′-lacZ transcriptional fusions at different growth stages. The constructed transcriptional fusions are identical to those described by Görke et al. (9). In B. subtilis strains bearing the fusions, the complete ptsH and crh coding regions are restored downstream from the lacZ insertion cassette. Their expression is directed by the IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible Pspac promoter in order to prevent polar effects caused by insertions into their natural chromosomal loci (Table 1). Expression from the crh′-lacZ (IQB495) (Table 1) and ptsH′-lacZ (IQB496) (Table 1) transcriptional fusions during exponential growth phase (t ₂) was comparable in both C minimal medium supplemented with casein hydrolysate and CSK minimal medium (Table 3). The presence of arabinose, a non-PTS sugar, did not influence crh and ptsH expression (Table 3). In addition, the β-galactosidase activities obtained for crh′-lacZ and ptsH′-lacZ in CSK minimal medium plus glucose (29 and 229, respectively) (Table 3) were very similar to those obtained by Görke et al. (9). Interestingly, at transitional growth phase (t ₄), the expression of crh increased about twofold in the presence of glucose and in both media. In contrast, ptsH expression was unaffected by the growth phase. These results suggest that increased expression of crh at the transcriptional level during transitional phase in the presence of glucose may lead to a higher cellular concentration of Crh than of HPr. Thus, this difference in synthesis level may contribute to the glucose repression relief effect on abfA′-lacZ and xsa′-lacZ expression in the crh-null mutant background (Fig. 2). However, due to a still expected higher concentration of HPr than Crh under these conditions, other factors, such as intracellular variation of certain metabolites which may influence Crh(Ser-P) phosphorylation and/or CcpA-Crh(Ser-P) complex formation, are likely to contribute to Crh(Ser-P)-mediated glucose repression at transitional phase.

Concluding remarks.

CCR of the arabinan-degrading enzyme genes is mediated by binding of CcpA to one cre located downstream near the promoter regions of abnA and xsa and to two cres present in the araABDLMNPQ-abfA operon, with cre araA located near the promoter and cre araB positioned within the araB gene. Our data suggest that during exponential growth phase, glucose-mediated CCR is mediated by CcpA associated with the coeffector HPr(Ser-P), and that towards the end of exponential growth, at transitional phase, glucose repression is achieved by both CcpA-HPr(Ser-P) and CcpA-Crh(Ser-P) complexes.