Abstract
The degradation of many structurally diverse aromatic compounds in Acinetobacter baylyi is accomplished by the β-ketoadipate pathway. In addition to specific induction of expression by certain aromatic compounds, this pathway is regulated by complex mechanisms at multiple levels, which are the topic of this study. Multiple operons feeding into the β-ketoadipate pathway are controlled by carbon catabolite repression (CCR) caused by succinate plus acetate. The pathways under study enable the catabolism of benzoate (ben), catechol (catA), cis,cis-muconate (catB,C,I,J,F,D), vanillate (van), hydroxycinnamates (hca), dicarboxylates (dca), salicylate (sal), anthranilate (ant) and benzyl esters (are). For analysis of CCR at the transcriptional level a luciferase reporter gene cassette was introduced into the operons. The Crc (catabolite repression control) protein is involved in repression of all operons (except for catA), as demonstrated by the analysis of respective crc strains. In addition, cross-regulation was demonstrated for the vanA,B, hca and dca operons. The presence of protocatechuate caused transcriptional repression of the vanA,B- and hca-encoded funnelling pathways (vertical regulation). Thus the results presented extend the understanding both of CCR and of the effects of Crc for all aromatic degradative pathways of A. baylyi and increase the number of operons known to be controlled by two additional mechanisms, cross-regulation and vertical regulation.
-
↵†Present address: Max Planck Institute for Biophysical Chemistry, Theoretical and Computational Biophysics Department, D-37077 Göttingen, Germany.
Edited by: H. L. Drake
INTRODUCTION
The bacterium Acinetobacter baylyi is a soil organism known to be able to use aromatic substances through the β-ketoadipate pathway (Harwood & Parales, 1996). Numerous more complex aromatic compounds can be converted into the two central starting compounds of the β-ketoadipate pathway, protocatechuate (PCA) and catechol, by additional short metabolic pathways (funnelling pathways). The expression of all the respective operons is thoroughly controlled by specific inducers. Furthermore, to cope with an array of environmental changes, the β-ketoadipate pathway and its funnelling pathways are controlled by a regulatory network the complexity of which is only beginning to be elucidated (Vaneechoutte et al. 2006; Gerischer, 2008; Williams & Kay, 2008). One part of this network is carbon catabolite repression (CCR) (Cánovas & Stanier, 1967; Tresguerres et al., 1970; Dal et al., 2005; Fischer et al., 2008). The molecular mechanisms of CCR are well understood in Escherichia coli and Gram-positive bacteria such as Bacillus subtilis, but not in bacteria belonging to the genera Pseudomonas and Acinetobacter. In these bacteria, organic acids such as succinate and acetate, as well as the protein Crc (catabolite repression control), play an important role in CCR (Wolff et al., 1991; Zimmermann et al., 2009). A. baylyi Crc has been shown to be involved in the degradation of the pca-qui transcript, which encodes enzymes dealing with quinate and PCA degradation (Zimmermann et al., 2009). In Pseudomonas putida, Crc affects the expression of genes involved in aromatic compound degradation (ben, cat, pca and pobA) (Morales et al., 2004). Direct binding of Crc to the RNA region directing translation of the regulators BenR and AlkS was demonstrated, indicating translational repression (Moreno et al., 2007; Moreno & Rojo, 2008).
In addition to CCR, cross-regulation becomes effective when mixtures of substrates feeding into both branches of the pathway are presented to an organism. In A. baylyi, cross-regulation results in a dominance of the catechol branch over the PCA branch (Brzostowicz et al., 2003; Siehler et al., 2007). There is evidence that transcriptional regulators BenM and CatM, which can bind regulatory regions upstream of pcaU (Gerischer et al., 1998), are involved in this cross-regulation. Finally, vertical regulation has also been observed in the presence of PCA. For example, the degradation of p-hydroxybenzoate (POB) is repressed by its own reaction product, PCA (Brzostowicz et al., 2003; Siehler et al., 2007).
In this investigation we analysed more operons with respect to CCR, cross-regulation and vertical regulation. Given the indicated global nature of the mechanism it is relevant to know whether multiple operons are affected, and in the future to examine whether the same mechanism is the cause. We used chromosomal luciferase reporter gene fusions in operons encoding enzymes for the degradation of benzoate, benzyl esters, anthranilate, hydroxycinnamates, dicarboxylic acids, vanillate, salicylate and catechol to gain a more comprehensive understanding of gene expression within the aromatic degradative pathways of A. baylyi. We addressed the involvement of Crc by investigating crc strains in parallel. We showed that all the operons analysed are affected by CCR; the withdrawal of Crc derepresses gene expression in most cases. Cross-regulation and/or vertical regulation was observed in all operons investigated (cross-regulation in hca, van and dca; vertical regulation in hca and van).
METHODS
Bacterial strains and growth conditions.
Strains of A. baylyi were grown on minimal medium at 30 °C as described earlier (Trautwein & Gerischer, 2001). Carbon sources were used at the following concentrations unless indicated otherwise: pyruvate, 20 mM; lactate, 20 mM; gluconate, 20 mM; acetate, 15 mM; succinate, 30 mM; succinate and acetate, 15 mM each. The following concentrations were used for induction: benzoate, 0.5 mM; p-coumarate, 1 μM; adipate, 1 mM; vanillate, 0.5 mM; salicylate, 0.5 mM; anthranilate, 1 mM; benzyl alcohol, 2 mM. Benzyl alcohol, p-coumarate, vanillate and adipate were dissolved in DMSO. Antibiotics for A. baylyi strains were used at the following concentrations: 100 μg spectinomycin ml−1; 20 μg streptomycin ml−1.
Strains of E. coli were grown in LB medium at 37 °C. Antibiotics were used in the following concentrations: 100 μg ampicillin ml−1; 100 μg spectinomycin ml−1; 20 μg streptomycin ml−1.
For growth experiments, A. baylyi strains (Table 1⇓) with luciferase transcriptional gene fusions were precultured on minimal medium complemented with the carbon source that would later be used in the experiment (except for the aromatic compound).
Bacterial strains and plasmids used in this study
Plasmid and strain construction.
To integrate the luciferase reporter gene into specific genes, PCRs were performed with primers listed in Table 2⇓, using chromosomal DNA from A. baylyi as a template. The fragments were cleaved with restriction enzymes and cloned (Table 1⇑). After integration of the luc cassette, the fusion constructs were cleaved with the indicated enzymes and used for transformation of A. baylyi. The aad9-mediated spectinomycin resistance was used for selection. The restriction sites used for the plasmid and strain construction were native sites in all cases. Standard methods were used for plasmid isolation, DNA purification, restriction endonuclease cleavage, ligation and transformation. Transformation of A. baylyi was done as described by Fischer et al. (2008).
Primers used in this study
To verify that luciferase fusions were integrated into the genome of A. baylyi strain ADP1 at the correct position, PCR analysis was employed, using a gene-specific primer (catA5, catB1, vanB2, vanK1 and salA1) and the luc primer (Table 2⇑), specific for the luc-aad9 cassette. The gene-specific primer targeted a sequence outside the DNA that was used for transformation.
Plasmid pAC57 was used to disrupt the crc gene in all strains containing the chromosomal luciferase reporter gene fusion (Zimmermann et al., 2009). pAC57 carries a crc gene that was rendered non-functional by the insertion of an Ω cassette, which carries a spectinomycin and streptomycin resistance gene. This construct was cleaved from the vector backbone by the restriction endonucleases XbaI and PstI and used for transformation of strains containing a luciferase reporter gene fusion to create the respective crc strain (Table 1⇑). Growth in the presence of spectinomycin and streptomycin was used to identify candidates with the desired modification. PCR analysis with primers crc3 and crc4 (Table 2⇑) was employed to confirm the correct integration of the construct into the corresponding region on the chromosome. Again, primers targeted loci outside the DNA that had been used for transformation.
PCR.
Cells of an overnight culture were suspended in water, boiled for 10 min, cooled on ice and centrifuged. The supernatant contained the chromosomal DNA and was used as template. The conditions using Taq DNA polymerase were 95 °C for 3 min, followed by 30 cycles of denaturation for 1 min at 95 °C, annealing for 1 min at 53 to 68 °C (depending on the primers) and elongation at 72 °C for a variable period (depending on the length of the amplified fragment).
Determination of luciferase activity.
At different time points during growth, samples of A. baylyi cells were taken and mixed with d-luciferin. The luciferase activity was detected as described earlier (Siehler et al., 2007). The measured relative light units (RLU) were divided by the respective OD600 to normalize the results. A value from the mid-exponential growth phase was read as characteristic for the strain and condition. Each growth experiment was carried out at least three times independently. The luciferase activity on different carbon sources was normalized to the corresponding activity on pyruvate (set to 100 %), or with the corresponding activity of the parental strain (crc+) on succinate and/or acetate (set to 100 %) for the crc strains. Error bars indicate standard deviation.
RESULTS
CCR of operons encoding funnelling pathways for aromatic compound degradation in A. baylyi
To investigate CCR, growth experiments were performed with the strains carrying catA-luc, catB,C,I-luc, vanA,B-luc, salA-luc and vanK-luc transcriptional fusions (Fig. 1⇓). Prior to this study, nothing was known about CCR of these operons by succinate and/or acetate. In all approaches, strains were grown on succinate and/or acetate with the specific aromatic inducer added (catA-luc, catB,C,I-luc, 0.5 mM benzoate; vanA,B, vanK-luc, 0.5 mM vanillate; salA-luc, 0.5 mM salicylate). We evaluated the effect of the carbon sources alone or in combination since it is known for the pca-qui operon that they have a much stronger effect in combination (Dal et al., 2002). The resulting luciferase activities were compared with the activity after growth on the non-repressing carbon source pyruvate with the specific aromatic inducer (Table 3⇓). The presence of pyruvate in addition to the aromatic substrate is known to have no effect on the pca-qui expression level compared to growth solely on the aromatic substrate (Dal et al., 2002). While the presence of succinate and acetate in combination caused strong repression of all operons analysed, succinate alone was able to repress the catA, catB,C,I,J,F,D and sal operons to a similar degree, while acetate failed to repress the catB,C,I,J,F,D and sal operons to that extent. However, the vanA,B and vanK operons showed only a slight increase on either succinate or acetate compared to both acids in combination. The strongest repression of promoter activity (up to 98 %) was observed for the catA operon. The activity under CCR conditions was comparable with the uninduced activity of this operon (data not shown). The catB,C,I,J,F,D, the vanA,B and the vanK operons showed slightly lower repression, while the sal operon showed only a moderate repression of promoter activity (up to 42 %).
Transcriptional fusions between A. baylyi genes and a luciferase reporter cassette (luc-aad9) made and used in this study. Arrangement and nomenclature of open reading frames is from the published and annotated genome sequence ().
CCR of A. baylyi operons involved in aromatic compound degradation by succinate and acetate
Expression pattern in the presence of lactate and gluconate
Prior to this study, pyruvate (in addition to an inducing carbon source) was shown to be a non-repressing carbon source for the ben, hca, dca, are and ant operons of the β-ketoadipate pathway with regard to CCR (Fischer et al., 2008). Lactate forms the precursor of pyruvate in the reaction catalysed by lactate/pyruvate dehydrogenase, suggesting that these two carbon sources might have related effects. Glucose is a repressing carbon source in E. coli and other bacteria and it was thus interesting to investigate its effect in A. baylyi CCR. While A. baylyi is able to use glucose, it first converts it to gluconate in the periplasm (Young et al., 2005). Gluconate is a much better growth substrate than glucose and was thus chosen instead. Therefore, the effects of lactate and gluconate on the expression control of the operons named above were tested. Growth experiments were performed using three different carbon sources (pyruvate, lactate and gluconate) to characterize the expression of the catA, catB,C,I,J,F,D, vanA,B, vanK and sal operons (Table 4⇓). Furthermore, expression levels of the are and ant operons were also analysed in the presence of gluconate as carbon source with addition of the aromatic inducer (areA-luc, 2.0 mM benzyl alcohol; antA-luc, 1.0 mM anthranilate). To summarize the observations: while the carbon source lactate turned out to be non-repressing for the expression of the catB,C,I,J,F,D and sal operons, a slight repressing effect on vanA,B and vanK expression was observed, as previously shown for the ant and are operons (Fischer et al., 2008). Gluconate is non-repressing for the expression of the sal genes, but has a slight repressing effect on the are operon. Furthermore, gluconate has a moderate repressing effect on the expression of the vanA,B, vanK, catB,C,I,J,F,D and ant operons. The most remarkable result was obtained for the catA operon. Here, lactate is a repressing carbon source whereas the presence of gluconate causes an increased expression (more than twofold higher than in the presence of pyruvate). This is not an effect of Crc since the same expression pattern was seen in the crc strain (data not shown). It should be noted in particular that neither compound significantly repressed the sal operon.
Effect of lactate or gluconate on the expression of the indicated operons
Connection between CCR and Crc
A. baylyi Crc has been shown to strongly affect the stability of the pca-qui transcript but there is also a transcriptional effect (significantly increased expression under all conditions in the absence of Crc: Zimmermann et al., 2009). To determine whether the Crc protein is involved in CCR of additional operons involved in aromatic compound degradation at the transcriptional level, crc was disrupted in strains carrying a luciferase reporter gene fusion (Table 1⇑). Growth experiments were performed with crc strains containing a benA-luc, hcaA-luc, dcaA-luc, catA-luc, catB,C,I-luc, salA-luc, vanA,B-luc, vanK-luc, areA-luc or antA-luc transcriptional fusion and luciferase activity was determined (Table 5⇓). Inducers were added as follows: benzoate, 0.5 mM (benA-luc, catA-luc, catB,C,I-luc); p-coumarate, 1 μM (hcaA-luc); adipate, 1.0 mM (dcaA-luc); salicylate, 0.5 mM (salA-luc); vanillate, 0.5 mM (vanA,B-luc, vanK-luc); benzyl alcohol, 2.0 mM (areA-luc); anthranilate, 1.0 mM (antA-luc). Almost all the operons (with the exception of catA) responded with a derepression of promoter activity on succinate plus acetate to various degrees (from 3-fold for the sal operon, up to 28-fold for the ben operon) in comparison with the crc+ strain. On succinate or acetate alone, the results were comparable to those on succinate and acetate except for a few cases, the most notable of which being the ben operon. Here, the absence of crc caused a repression on succinate but no significant effect on acetate. Obviously, Crc negatively affects transcription at most promoters tested. The observation of the derepression in the absence of Crc first made for the pca-qui operon can now be extended to almost all additional operons investigated here.
Effect of crc deletion on the expression of the indicated operons under CCR conditions
Cross-regulation
It was shown earlier that pobA, which encodes the enzyme for the degradation of POB to PCA, is strongly repressed when benzoate is present in addition to POB, although POB is the specific inducer for pobA expression (Brzostowicz et al., 2003). This interaction between the two branches of the β-ketoadipate pathway was named cross-regulation. Here, the effect of different benzoate concentrations (0.1–5.0 mM) on the transcriptional activity of the hca, vanA,B and dca operons was determined (Fig. 2a⇓). Probably due to the toxicity of benzoate, strains grown in the presence of higher amounts of benzoate (3.0 and 5.0 mM) showed a decrease in the growth rate, but reached the same final optical density as the culture grown without benzoate in the medium (data not shown). Since the effects were observed at much lower benzoate concentrations, the growth inhibition at higher benzoate concentration is not disturbing. For the hca and vanA,B operons, a strong decrease in transcriptional activity was observed with increasing benzoate concentrations. This repression is even stronger than CCR by succinate and acetate in combination (90 % for vanA,B; Table 3⇑, and 93 % for hca (Fischer et al., 2008)). For the dca operon, a slightly different behaviour was observed: benzoate concentrations lower than 1.0 mM led to increased promoter activity, while higher concentrations caused repression which was not as strong as in the case of van and hca.
Repression of operons in the presence of benzoate (cross-regulation) or PCA (vertical regulation). Luciferase activity of A. baylyi strains containing the indicated transcriptional luc fusions in the presence of the specific inducer and increasing concentrations of (a) benzoate or (b) PCA. The activity of each operon in the absence of benzoate or PCA was set to 100 %. Each growth experiment was carried out at least three times independently. Error bars indicate standard deviation.
Vertical regulation
It was shown earlier that PCA inhibits pobA expression in the presence of the specific inducer POB (Brzostowicz et al., 2003). Here, we investigated the effect of PCA (0.1–5.0 mM) on promoter activity of the hca and vanA,B operons in the presence of the specific aromatic inducer. In contrast to cultures grown in the presence of higher concentrations of benzoate, no decrease in growth rate was observed for cultures grown in the presence of higher amounts of PCA. Both operons are affected by vertical regulation as well (shown in Fig. 2b⇑). At higher concentrations of PCA, the expression of both operons decreased, as observed for the pobA gene. We refer to this type of regulation as vertical regulation, since a downstream metabolite (with respect to the catabolic pathway, PCA) affects the expression of an operon encoding a pathway for its generation from various precursors.
DISCUSSION
Strong repression of all investigated operons
Earlier studies analysing the ben, dca, hca, are and ant operons (Fischer et al., 2008) and the pca-qui operon (Dal et al., 2005; Siehler et al., 2007) revealed a strong repression of promoter activity when the organic acids succinate and acetate were present in addition to the specific aromatic inducer of the respective pathway. Here, we extended these observations by analysing additional operons all connected to the degradation of aromatic compounds (catA, catB,C,I,J,F,D, vanA,B, sal, vanK, repression between 42 and 98 %; Tables 3⇑ and 5⇑). Furthermore, the investigation included the additional carbon sources lactate and gluconate. While pyruvate generally can be regarded as neutral for gene expression (Dal et al., 2002), the presence of lactate or gluconate has different effects on the expression of the operons analysed, ranging from no effect to a slight repression in most cases (Table 4⇑). A surprising result is the induction of catA by gluconate whereas all other tested carbon sources caused strong repression at this operon. Thus in this single case, gluconate has an even smaller negative effect on induction than pyruvate. In contrast, pyruvate has a small repressing effect, which is negligible when looking at the strong repression in all other cases. The repression of the operons encoding the funnelling reactions by elevated concentrations of succinate and acetate is meaningful in the context of energy preservation. The β-ketoadipate pathway and all the funnelling reactions do not lead to energy conservation. It thus seems advantageous to express these funnelling pathways only to the least necessary extent (an argument which also applies to the other regulatory phenomena, cross-regulation and vertical regulation, discussed below). Energy conservation occurs subsequently by oxidation of succinyl-CoA and acetyl-CoA. Lactate and pyruvate cause no or only a moderate repression of the operons (with the exception of lactate on catA). These two substrates may not be such abundant carbon sources for A. baylyi in its natural habitat and thus no regulatory mechanism may have evolved. The degradation of gluconate by A. baylyi is energy consuming, because gluconate has to be taken into the cell by active transport and then be converted to 6-phosphogluconate, which is also an energy-requiring step. Gluconate degradation then is accomplished by a modified Entner–Doudoroff pathway (Barbe et al., 2004; Young et al., 2005). Furthermore, the utilization of gluconate or glucose is unusual among species of Acinetobacter; thus it seems to fit that there is no strong repression of aromatic degradative pathways by gluconate as observed for succinate and acetate.
Crc is involved in the transcriptional expression control of numerous aromatic degradative pathways
The absence of Crc had a significant effect on the expression of the operons are, sal, ben, catB,C,I,J,F,D, dca, hca, vanA,B, vanK and sal, as shown here and earlier for pca-qui and pobA (Dal et al., 2002; Zimmermann et al., 2009). The inactivation of crc resulted in a significant transcriptional derepression of most operons analysed. Only catA, forming a separate regulatory unit in A. baylyi, did not show a significant dependence on Crc. While the results presented here focus on the transcriptional level, A. baylyi Crc also acts post-transcriptionally by (directly or indirectly) dramatically changing the mRNA half-life of the pca-qui transcript (Zimmermann et al., 2009). In P. putida it has been shown that Crc directly affects translation of the regulator AlkS (Yuste & Rojo, 2001). Similar observations have been made for P. putida BenR mRNA. In these bacteria, Crc blocks translation by binding the translation initiation site (Moreno & Rojo, 2008). In summary, Crc turns out to have complex effects at both the transcriptional and the post-transcriptional level. Furthermore, Crc affects all operons investigated so far, an observation supported by investigations of P. putida Crc, which is known to be a truly global regulator (Moreno et al., 2009).
Cross-regulation
Cross-regulation takes place between the two branches of the β-ketoadipate pathway, indicating that these branches are not regulated independently (Fig. 3⇓). Our results suggest that the consumption of benzoate (degraded via the catechol branch) is favoured over substrates of the PCA branch (e.g. vanillate or other hydroxycinnamates), as shown by the transcriptional repression of the respective operons in the presence of benzoate. Furthermore, the presence of increased benzoate concentrations led to a repression of the dca genes. The repression of the vanA,B and hca operons caused by benzoate is even stronger than the repression caused by the organic acids succinate and acetate in combination. It is important to note that all strains used for the analysis of cross-regulation are able to degrade benzoate to cis,cis-muconate via the catechol branch. Thus, cis,cis-muconate is available as an effector for BenM and CatM. These transcriptional regulators are in fact responsible for the regulated expression of the ben and cat genes in response to benzoate and cis,cis-muconate (Craven et al., 2008). However, while no interaction of the two regulators with the pobA promoter could be detected, both BenM and CatM are able to bind to a fragment upstream of pcaU, an interaction that is promoted by cis,cis-muconate (Brzostowicz et al., 2003). Subsequently, it was shown that pca-qui expression also undergoes cross-regulation by benzoate (Siehler et al., 2007). It is still unclear how these observations connect with the cross-regulation of pobA. It may be that POB transport is limited when the pca-qui operon (including pcaK) is repressed, but the involvement of the second POB transporter, VanK, needs to be characterized. Studies on VanK suggest an involvement in PCA uptake, since a double knockout of vanK and pcaK prevents growth on PCA, while a knockout of pcaK alone does not (D'Argenio et al., 1999). However, it has to be determined whether there is involvement of VanK in POB uptake. For all other operons under study, the presence of binding sites for BenM or CatM needs to be evaluated to determine if these proteins are centrally involved in the mechanism of cross-regulation.
Summary of the regulatory levels within the aromatic catabolic pathways of A. baylyi. All operons are affected negatively in expression by CCR triggered by succinate and acetate (not included in the scheme). Thin arrows indicate the direction of the metabolic pathway; thick arrows indicate transcriptional regulation effects. Repression (−) or activation (+) is symbolized by plus or minus signs next to the arrows.
Vertical regulation
We showed here that the addition of PCA to the medium containing the specific inducers (vanillate, p-coumarate) triggers a mechanism that represses the vanA,B and hca operons encoding funnelling reactions to PCA. This regulatory mechanism was termed vertical regulation (Fig. 3⇑). An analogous observation was made earlier for pobA. PobR had been excluded as a mediator of this effect (Siehler et al., 2007). Furthermore, PcaU, the transcriptional regulator of the pca genes, is naturally able to bind PCA but is not able to bind the pobA promoter (Siehler et al., 2007). The molecular basis of vertical regulation observed for the pobA promoter thus remains unknown, but the phenomenon seems to be common, as shown for the vanA,B and hca operons. The biological relevance appears to be that increased internal PCA concentrations prevent the expensive expression of any funnelling pathway leading to more PCA production. Only when the internal PCA level falls under a certain level will this mechanism cease to be active and more PCA can be produced from the respective precursor. Furthermore, external PCA in concentrations high enough to promote growth will also lead to the repression of funnelling pathways producing it, which also supports the organism in saving resources.
Conclusion
The first mechanisms of regulation found within the pathways for aromatic compound degradation in A. baylyi included induction brought about by specific regulators and their respective effectors. In the current study, additional levels of regulation were characterized more comprehensively, namely carbon catabolite repression, cross-regulation and vertical regulation. All of these consist of repressing effects. Besides efficient induction in response to the presence of the substrates, three different mechanisms decreasing expression under several conditions override specific induction. This indicates that the expression of genes for the degradation of aromatic compounds only occurs under well-defined conditions and in a well-defined order to best fit the needs of the bacterium.
Acknowledgments
We would like to thank Michael Vogl for his contributions. This work was supported by personal grants (LGFG) to F. S. B. and R. F. by the state of Baden-Württemberg, Germany.