The differential response of the Pben promoter of Pseudomonas putida mt-2 to BenR and XylS prevents metabolic conflicts in m-xylene biodegradation



Pseudomonas putida mt-2 encompasses two alternative and potentially conflicting routes for benzoate metabolism, one meta pathway encoded by xyl genes of the pWW0 plasmid and mastered by the Pm promoter and XylS, and one chromosomally encoded ortho pathway initiated by Pben and the BenR protein. Any cross-activation of Pben promoter by XylS ought to cause a metabolic conflict during the degradation of m-xylene because 3-methylbenzoate (3MBz) generated as an intermediate can be channelled through the ortho pathway and produce toxic dead-end metabolites. The activation of Pben by XylS was revisited using both reporter technology and tiling arrays targeted to the sequences of interest around messenger RNA initiation of both Pben and Pm promoters. Analysis of supersensitive luxCDABE fusions, inspection of xylX versus benA transcripts and growth tests of benR mutants indicated that the natural expression ranges of XylS under various conditions are insufficient to cause a significant cross-regulation of Pben whether cells face endogenous or exogenous 3MBz. This seems to stem from the nature of the operators for binding either transcriptional factor, which in the case of the Pben promoter of P. putida mt-2 appear to have evolved for avoiding a strong interaction with XylS.


Biochemical pathways for metabolization of the vast array of chemical substrates found in nature often encode an evolutionary record of the problems that bacteria had to overcome before reaching a workable enzymatic and genetic solution (George and Hay, 2011). Examples include mutations in transcriptional activators, which alter effectors specificity expanding or narrowing the substrate range of the controlled pathway (Pavel et al., 1994; Jaspers et al., 2000; Choi et al., 2003), the modification of repressors preventing operator binding for allowing constitutive expression of required enzymes (Arai et al., 1999), or the creation of promoter elements by insertions/deletions resulting in increased transcription of catabolic genes (Kasak et al., 1997). A recurrent instance of adaptations required for avoiding metabolic conflicts is the inactivation of catabolic enzymes that produce toxic metabolites or result in misrouting of pathway intermediates. In these cases, the loss of enzyme activity is readily accomplished by point mutations or gene deletions (Gerischer and Ornston, 1995; Ju and Parales, 2010).

The TOL pathway, encoded by the archetypal plasmid pWW0, is probably the best characterized route for degradation of aromatics from a biochemical and genetic point of view (Ramos et al., 1997). This system, which has been described in the soil bacterium Pseudomonas putida mt-2, allows mineralization of toluene and m-xylene, and it is composed of two operons encoding an upper and a lower pathway. Through the upper pathway, the methyl group of toluene or m-xylene is sequentially oxidized to render benzoate (Bz) or 3-methylbenzoate (3MBz) respectively. Bz is then converted to catechol by either the lower route enzymes encoded by xylXYZL or by the products of the highly similar genes benABCD encoded in the chromosome (Fig. 1). Both gene clusters can be induced by Bz itself (Ramos et al., 1997; Cowles et al., 2000). Thereby, generated catechol can be channelled towards either a meta-cleavage pathway by catechol 2,3-dioxygenase encoded by the xylE gene or in an ortho-fission route initiated by the chromosomally encoded catechol 1,2-dioxygenase (Fig. 1). Instead, the transformation of 3MBz formed during m-xylene turnover into the corresponding 3-methylcatechol poses a metabolic conflict. This is because the action of catechol 1,2-dioxygenase over the methyl-substituted ring cleavage substrate generates 2-methyl-2-enelactone, a dead-end product (Schmidt et al., 1985). In this undesirable scenario, the activation of the unproductive chromosomally encoded ortho cleavage route should be avoided to ensure that the generated methylcatechol is channelled in its entirety through the meta fission pathway that is proficient for methyl-substituted substrates. Therefore, a finely tuned regulation able to sort out the expression of both cleavage pathways is a key requirement for a sound catabolic phenotype.

Figure 1.

Genetic, regulatory and biochemical connections of 3MBz catabolism in P. putida mt-2. 3MBz can be available either exogenously or generated intracellularly through biodegradation of m-xylene by the upper TOL pathway. The figure sketches both alternative metabolic itineraries involved in the further biodegradation of 3MBz, i.e. the xyl genes of the lower TOL operon (top), which are expressed from the Pm promoter activated by XylS, and the chromosomal ben genes (bottom) that are transcribed from promoter Pben upon activation by the XylS homologue BenR. Both the xyl and the ben routes make 3MBz to converge towards methylcatechol, but then diverge at that point: this compound can be subject to extradiol cleavage by XylE or intradiol fission by CatA and CatA2. This last reaction originates the dead-end intermediate 2-methyl-2-enelactone. Note that the action of the chromosomally encoded benzoate 1,2-dioxygenase on 3MBz may produce also a small fraction of 4-methylcatechol (Knackmuss et al., 1976), which becomes efficiently converted by CatA/CatA2 into the equally toxic compound 4-methyl-2-enelactone (not shown). The metabolic conflict is exacerbated if the TOL substrate were p-xylene, as it is converted to 4-methylbenzoate, then exclusively to 4-methylcatechol by the ben/xyl-encoded enzymes and further to the same toxic enelactone by CatA/CatA2.

In P. putida mt-2, the regulation of Bz and 3MBz degradation routes relies on transcriptional regulators belonging to AraC family, the so-called XylS and BenR activators. Members of this family present a characteristic modular structure. The conserved C-terminal domain includes two potential tetra-helical HTH DNA-binding motifs while the N-terminal domain is not well conserved and is involved in effector recognition and dimerization (Gallegos et al., 1997). The pWW0-encoded XylS regulator mediates transcriptional activation of the Pm promoter driving the expression of the meta pathway in response to 3MBz and Bz as inducers. Likewise, BenR is able to trigger the activity of Pben promoter by recognition of Bz as effector, allowing the expression of the benABCD-catA2 operon (Cowles et al., 2000). Both regulators share about 60% amino acid identity through their full sequence, and the similarity is even higher in the C-terminal DNA-binding domain suggesting that cross-activation of their target promoters could take place. In fact, the ability of BenR to activate the Pm promoter of TOL system in response to Bz has been recognized (Kessler et al., 1994). However, the role of XylS as an activator of Pben has not been entirely clarified. Earlier reports have suggested the occurrence of cross-activation of Pben by XylS in a different P. putida lineage, PRS2000 strain (Jeffrey et al., 1992; Cowles et al., 2000). A more recent publication has documented also this occurrence in P. putida mt-2 by using a plasmid harbouring a Pben-lacZ fusion (Dominguez-Cuevas et al., 2006). While the phenomenon may well happen, the question is how the corresponding parameters are set to avoid the predicted scenario of metabolic stress generated by the self-defeating activation of ortho-cleavage enzymes in the presence of 3MBz as described above.

In this work, we have employed a suite of alternative techniques to clarify (i) whether cross-activation of Pben by XylS actually occurs in wild-type P. putida mt-2 cells exposed to endogenously produced and exogenously added 3MBz, and (ii) if so, whether the parameters of the system under natural conditions lead to a metabolic conflict or instead the same parameters are poised to avoid it. To this end, we resorted to various configurations of supersensitive promoter fusions to luxCDABE, to transcriptomic analyses of benA and xylX under diverse conditions, and to growth tests of strains bearing different regulatory combinations. The results below reveal that while a degree of cross-activation of the ortho-cleavage pathway by XylS can be detected, its extent is kept low owing to changes in the Pben promoter that withhold the binding of the potential cross-regulator. This scenario provides a remarkable case of regulatory tinkering that facilitates recruitment of a catabolic pathway in a new host with a non-fully compatible metabolism.


Pben is only marginally activated by 3MBz in P. putida mt-2

In an attempt to examine the responses of the chromosomally encoded Pben promoter and pWW0-borne Pm promoter in entirely comparable and supersensitive conditions, we constructed reporter plasmids pSEVA226-Pb and pSEVA226-Pm. These are constructs based in the very low-copy replicon RK2 of pSEVA226 (Silva-Rocha et al., 2013) in which DNA fragments encompassing 500 bp (for Pben) and 80 bp (for Pm) regions of these promoters were cloned in front of a promoterless luxCDABE cassette. Note that the ribosomal binding sites (RBSs) are in both fusions those of the reporter operon (see Experimental procedures). Such segments bear the entire regulatory regions that bind, respectively, BenR and XylS (Kessler et al., 1994; Cowles et al., 2000), and thus are expected to report faithfully the expression of either the ortho or the meta pathway when introduced in strain P. putida mt-2. With these strains at hand, P. putida mt-2 (pSEVA226-Pb) and P. putida mt-2 (pSEVA226-Pm), we judged side by side the capacity and inducibility of each promoter in the presence of either Bz or 3MBz in a genetic background with wild-type levels of both transcriptional factors. To this end, each strain was pre-grown in M9 minimal medium with succinate as sole C source, diluted 1:20 in fresh media supplemented with 1 mM of either Bz or 3MBz, and incubated as explained in Experimental procedures. Light emission was quantified in each of the cultures 4 h after the addition, the results being shown in Fig. 2. Perusal of the data reveals at least three aspects that are relevant for the sake of this study. First, the basal level of Pben is visibly lower than that of Pm. Note that in the pSEVA226-derived constructs, there is a strong transcriptional terminator upstream of the cloned DNA, thereby ensuring that any luminescence of the reporter originates in the inserted promoter. Second, the Pm-luxCDABE fusion generates a level of luminescence with Bz that is markedly higher than that brought about by Pben under the same conditions. And, third, the data of Fig. 2 show that 3MBz induces Pm to a degree that is ≥ 11-fold higher than the same parameter in Pben. From a mere phenomenological point of view, these results put forward that Bz induces both the ortho and the meta pathway, but 3MBz seems to activate significantly only the pWW0-encoded route. Because these data somehow question the significance of previously reported experiments of cross-regulation between XylS/Pm and BenR/Pben (Cowles et al., 2000; Dominguez-Cuevas et al., 2006), we revisited the issue with three entirely different experimental approaches, which, as disclosed below, do not leave any reasonable doubt on the matter at stake.

Figure 2.

Activity of Pben and Pm promoters as measured by transcriptional luxCDABE fusions located in low-copy plasmids harboured by P. putida mt-2. Strains bearing reporter plasmids pSEVA226-Pb and pSEVA226-Pm were pre-grown overnight in M9 medium, diluted 1:20 in the same fresh medium, added with 1 mM of each of the aromatic inducers indicated or kept as untreated control (Nil) and light emission recorded after 4 h. Promoter activities in each condition are expressed as relative units (RU: luminiscence/OD600). Error bars represent standard deviation of a minimum of three experiments. Nil, untreated condition.

Quantification of benA and xylX transcripts in response to aromatic inducers

In order to dissect the effect of Bz and 3MBz on expression of chromosomally encoded benA (driven by Pben) and the plasmid-borne xylX (transcribed by Pm), we inspected the corresponding transcripts by means of a customized tiling array of overlapping 50-mers that covered the entirety of the target genes (see Experimental procedures). Employing succinate-grown P. putida mt-2 cells as the reference, we exposed the cultures not only to exogenously added Bz and 3MBz, but also to m-xylene and o-xylene. In the first case, 3MBz is produced endogenously through the conversion of m-xylene to the cognate carboxylic acid through the action of the upper TOL enzymes (Fig. 1). But note also that m-xylene suffices to overexpress xylS, which in turn activates Pm even before 3MBz is formed (Ramos et al., 1997). To sort this out, we run a control with o-xylene, a non-metabolizable but otherwise efficient inducer of the upper pathway and the xylS gene (Abril et al., 1989). This allowed assessing how much cross-activation of Pben by XylS can be attributed to the complex effector/regulator and how much to the mere overproduction of the protein. To follow the evolution of xylS expression under such conditions, the tiling array included also sequences covering this gene. The results of the corresponding experiments are shown in Fig. 3. While the data match fairly well those of Fig. 2 generated with luxCDABE reporter fusions to Pben and Pm, the tiling array revealed details of the cross-talk that were not noticeable before. First, 3MBz endogenously produced from m-xylene is expectedly a good inducer of xylX but it only triggers a quite limited expression of benA (note vertical axes scales of the panels). Such minor activation of Pben could be due to either cross-regulation by XylS or the consequence of a residual recognition of 3MBz as an effector of BenR (Silva-Rocha and de Lorenzo, 2012). Note that BenR levels remained low and invariant in all tested conditions (data not shown). A second detail exposed by the tiling array experiment was that overproduction of XylS by o-xylene altogether failed to cause any activation of Pben (Fig. 3). This is in contrast with Pm, which is indeed triggered by such increased levels of XylS even in the absence of any inducer. Finally, as anticipated by the luxCDABE fusions of Fig. 2, exogenously added Bz brings xylX and benA to high levels while 3MBz does so efficiently only in the first case (again, note different scales of vertical axes). While the origin of the residual activation of Pben by 3MBz and m-xylene is addressed below, the phenomenological picture that emerges from the data of Fig. 3 is that Bz elicits the ortho and meta routes to approximately the same expression range. In contrast, 3MBz seems to predominantly bring forth the operation of the meta pathway. This state of affairs makes sense for avoiding the metabolic conflicts mentioned above, but say little on the mechanism(s) that allow such differential response of Pben and Pm to either unsubstituted or substituted Bz.

Figure 3.

Transcriptomic profile of (A) xylS and (B) the Pm (xylX) and (C) Pben (benA) regions of P. putida mt-2. The panels show the signals from a high-density tiling array hybridized with RNA extracted from P. putida exposed to m-xylene, o-xylene, Bz, 3MBz and succinate as indicated. The bars express relative level of each messenger RNA segment in respect to the reference condition (i.e. terminal growth of P. putida mt-2 on succinate, see Experimental procedures). The horizontal axes indicate the extent of the genomic region of interest covered by the subset of oligonucleotides of the tiling array. (A) Entire xylS gene. (B) 5′ untranslated region and leading 50 bp of the xylX coding region. (C) 5′ untranslated region and leading 35 bp of the benA coding region. The vertical axes correspond to the log2 of such a ratio. Note different scales of the Y values in each case.

Natural XylS levels of P. putida mt-2 fail to cross-activate Pben

The data above leave the question open on what regulator activates what promoter in the presence of what inducer in the naturally occurring P. putida mt-2. Available literature indicates that (i) XylS recognizes both Bz and 3MBz as bona fide effectors (Ramos et al., 1997), (ii) BenR is responsive to Bz but only marginally to 3MBz (Silva-Rocha and de Lorenzo, 2012) and (iii) BenR can activate Pm (Jeffrey et al., 1992; Kessler et al., 1994). Alas, the condition that completes the whole picture of interactions between the two regulatory systems (i.e. testing whether XylS can activate Pben in the absence of BenR) has not been unambiguously tested. In an effort to address this question with a gene dosage and stoichiometry that faithfully reproduces the natural conditions, we generated transcriptional fusions of Pben and Pm to luxCDABE in a Tn7 mini-transposon vector (Choi et al., 2005) that placed the whole reporter device in the same neutral site of the chromosome of a P. putida KT2440 (i.e. the pWW0-cured derivative strain) variant lacking the benR gene (see Experimental procedures). The resulting reporter strains were then used as recipients of either the wild-type pWW0 or plasmids pJB653 (Blatny et al., 1997), pSEVA438 (Silva-Rocha et al., 2013) and pVLT43 (de Lorenzo et al., 1993), each of them capable of expressing XylS at increasingly higher levels (pJB653 < pSEVA438 < pVLT43). Note that high XylS concentrations lead to activation of Pm without any need for inducer view (Ramos et al., 1997). The reporter output was then recorded for each of the strains added with Bz or 3MBz as before. In the case of strains bearing pVLT43, the cultures were amended or not with IPTG in order to bring about still higher intracellular levels of XylS. The data of Fig. 4A revealed that in the absence of BenR, the ordinary levels of XylS stemming from the pWW0 plasmid were unable to trigger any significant induction of Pben with either Bz or 3MBz. The same held true when XylS was supplied by a low-copy number plasmid (pJB653). A minimal induction was detected with the XylS levels produced by a medium-copy number plasmid (pSEVA438), but luminescence took off the background only when xylS expression was strongly stimulated by IPTG in the overexpressing plasmid pVLT43 (Fig. 4B; note the scales of the vertical axes). In contrast, the Pm promoter was always induced – albeit to very different levels by all combinations of XylS and inducers (Fig. 4C), including overproduction of the regulator in the absence of aromatics (Fig. 4D). Taken together, these data indicate that physiological concentrations of XylS in P. putida mt-2 do not grant any significant cross-regulation of the ortho-cleavage route-encoding genes. It is necessary to artificially increase such XylS levels (e.g. by using multicopy vectors or overexpressing plasmids) for causing a still moderate induction of Pben by this regulator. But even under such unlikely conditions, the responses of Pben to XylS are 10–15 times weaker than Pm under the same dosage of the regulator. In contrast, the presence of BenR affords a strong induction of Pben by Bz (Fig. 2).

Figure 4.

Activity of (A, B) Pben and (C, D) Pm promoters measured by transcriptional luxCDABE fusions located in chromosome of a P. putida KT2440 benR mutant and entered with different xylS-encoding plasmids. Strains were grown overnight in M9 medium with 0.3% citrate, diluted 1:20 in the same medium and either added with 1 mM of each of the inducers or kept untreated (Nil). Light emission was recorded after 3–4 h, on the basis of which promoter activities were determined for each condition, calculated as relative luminiscence units (RU). Error bars indicate standard deviations of a minimum of three experiments. In A and C, each xylS-encoding plasmid provides a different level of XylS: wild-type copy number (pWW0), low-copy number (pJB653) and medium-copy number (pSEVA438). In B and D, the pVLT43 plasmid carries the xylS gene with an improved translation initiation region and under the control of the IPTG-inducible Ptac promoter, thereby providing high intracellular levels of XylS upon induction.

The results on Pben above deal with exogenously added 3MBz, but still they did not clarify the state of affairs when this effector is generated intracellularly as a consequence of the action of upper TOL enzymes on m-xylene. In order to check if increased levels of XylS naturally reached during m-xylene metabolism (leading to 3MBz production) are enough to cross-activate Pben, we re-examined the activity of this promoter (and Pm as a positive control) in the P. putida KT2440 benR strains bearing the pWW0 plasmid and engineered with the same luxCDABE fusions as before. Because m-xylene might cause a degree of physiological stress that could interfere with reporter readout, we took the extra precaution of testing also 3-methylbenzyl alcohol (3MBA), a less-toxic substrate of the upper TOL pathway that is also a good XylR effector and yields 3MBz as a metabolic intermediate as well (Ramos et al., 1997). The results shown in Fig. 5 revealed that neither m-xylene nor 3MBA were able to induce high enough XylS levels to trigger the Pben promoter in the strain that lacks benR gene. On the contrary, the physiological overexpression of XylS by either upper pathway inducer triggered induction of Pm. Note that these results clarify also the origin of the residual Pben activation by m-xylene in the wild-type P. putida mt-2 strain during tiling array experiments (Fig. 3, see above). That the Pben promoter remains silent in cells of P. putida KT2440 (pWW0) exposed to upper pathway inducers when BenR is lacking implies that in the benR+ wild-type context, the response of Pben to m-xylene necessarily stems from a residual recognition of the endogenously generated 3MBz as an effector of BenR.

Figure 5.

Activity of Pben and Pm promoters measured by transcriptional luxCDABE fusions located in chromosome of a P. putida KT2440 benR mutant bearing the pWW0 plasmid. Strains were grown overnight in M9 medium with 0.3% citrate, diluted 1:20 in the same medium and added with 1 mM of 3MBA or exposed to vapours of m-xylene released by a 50% dilution in dibutyl phthalate. The untreated condition is also shown (Nil). Light emission was recorded after 3–4 h, on the basis of which promoter activities were determined for each condition as relative units (RU). Error bars indicate standard deviations of a minimum of three experiments.

Taken together, the data above indicated that while Pben can be activated by XylS under somewhat unusual circumstances, such activation is not significant under physiological conditions. Prediction is, therefore, that low-level XylS cannot replace the ordinary function of BenR. Yet, such a substitution could occur when XylS levels are artificially increased. The experiments below shed light on this conjecture.

Physiological levels of XylS do not restore growth on Bz of P. putida mt-2 lacking BenR

The experiments presented thus far indicate that Pben can be shown responsive to XylS in vivo by manipulating the intracellular pool of the regulator. The question is whether any possible cross-activation of this sort has metabolic consequences or the phenomenon is just the result of the engineered experimental set-up. The simple experiment of Fig. 6 was made to answer this question. In this test, the growth on Bz of a control wild-type P. putida KT2440 without XylS was compared with that of a benR mutant bearing a plasmid encoding XylS in various configurations. The positive control (wild-type strain) grew rapidly in the medium with Bz, while the negative control (isogenic benR strain), as expected, failed to grow entirely. Then a xylX::bla variant of pWW0 was transferred to the benR strain. This plasmid bears an inactivated allele of the first gene of the lower TOL pathway that encodes a subunit of the toluate 1,2-dioxygenase that starts degradation of Bz/3MBz through that route. pWW0-xylX::bla thus provides physiological levels of XylS but does not contribute to any possible metabolization of Bz, which has to go necessarily through the chromosomally encoded ortho cleavage pathway driven by Pben promoter. As shown in Fig. 6, such strain failed to grow at all on 5 mM Bz. The same happened when P. putida benR was transformed with low-copy number xylS+ pJB653 that produces low levels of XylS. The situation changed, however, when the concentration of the regulator was increased in vivo by either supplying the medium-copy number xylS+ pSEVA438 or the xylS expression plasmid pVLT43. In these cases, the benR strain could eventually resume growth on Bz after some prolonged lag phase, signifying that a functional replacement of benR by xylS had taken place under such conditions. These results are fully consistent with those reported above with the Pben → luxCDABE and Pm→ luxCDABE fusions. All the data thus coherently accredit that any possible cross-regulation of Pben by XylS lacks physiological significance in the original genetic architecture of strain P. putida mt-2.

Figure 6.

Growth curves of a P. putida KT2440 benR mutant supplemented with different xylS-encoding plasmids when cultured on Bz (5 mM) as the sole carbon source. Cultures of each strain were pre-grown on succinate 0.3% and diluted 100× in Bz-containing medium. The mean values of the OD600 of duplicate cultures are shown.


Strain P. putida mt-2 provides a good example of the evolutionary roadmap that a bacterium had to go through during expansion of its metabolic repertoire, in this case from Bz to 3MBz. Although both molecules are chemically similar, their main sources in the environment are different. Bz is more abundant as a component of plant root exudates (Uren, 2007), while methylated Bzs are typically produced as side-biodegradation products of aromatic hydrocarbons like those often found in the volatile fraction of petroleum (BTEX compounds; Vandecasteele and Monot, 2008). In the case of P. putida mt-2, it is likely that a pre-existing Bz-degrader and plasmid-less bacterium received the xyl genes for metabolism of m-xylene through horizontal transfer of a conjugative plasmid. This scenario is supported by the fact that the ortho pathway at stake is found in every sequenced P. putida strain, suggesting that such metabolic activity is a core trait in this species. In contrast, the TOL pathway is an exceptional attribute found in just a few strains that suggest a more recent evolutionary origin. While non-substituted Bz can be channelled productively through both ortho and meta catabolic pathways, the routing of methylated aromatics into the chromosomally encoded intradiol route leads to methyl-substituted 4-carboxymethylbut-2-en-4-olides (trivially called methyl-2-enelactones) as dead-end metabolites (Fig. 1). Because the two catechol 1,2-dioxygenases encoded in P. putida mt-2 genome have activity on both catechol, 3-methylcatechol and 4-methylcatechol (Jimenez et al., 2014), it is plausible that a significant fraction of m- and p-xylene finish as dead-end metabolites. How has P. putida solved this problem? The results presented above expose the solution: under physiological conditions, Bz activates expression of both the ortho and the meta pathways genes, whereas methylated substrates trigger only the meta route, leaving silent transcription of the chromosomal ben genes. Such a simple way out conflicts, however, with some of the existing literature on the regulatory interplay between the meta and the ortho routes for degradation of Bz in P. putida. In particular, by using multicopy Pben-lacZ transcriptional fusions Cowles and colleagues (2000) and Dominguez-Cuevas and colleagues (2006) reported that XylS can induce the promoter of benABCD-catA2 operon in response to Bz and 3MBz in two different strains of P. putida. If such a cross-induction had a physiological value, then growth of P. putida mt-2 in m-xylene or 3MBz would be hardly possible (Fig. 1). Our data above show, instead, that while XylS can indeed regulate Pben under some engineered conditions, the extent of the cross-talk is not significant for the wild-type strain in physiologically significant settings.

Cross-activation between AraC-type regulators besides BenR/XylS has been reported (Ledger et al., 2002; Carl and Fetzner, 2005), indicating a degree of promiscuity in respect to operator binding specificities in this family of transcriptional factors. Such a cross-binding of activators to non-discriminating regulatory sequences surely plays a role in the evolution of novel catabolic routes, as it allows primal regulation of genes acquired by horizontal transfer (de Lorenzo and Perez-Martin, 1996). Once a regulatory device has been recruited, it has to be further finely tuned into the indigenous transcriptional and metabolic network for avoiding detrimental effects on the host physiology (Cases and de Lorenzo, 2001). This evolutionary step implies additional genetic modifications in regulators and/or regulatory sequences to achieve more specific interactions. Our results suggest that such an adaptation has occurred during the acquisition of the TOL pathway by P. putida. This notion is strengthened upon examination of the regulatory sequences involved. In fact, inspection of the DNA upstream of the Pben and Pm promoters (Fig. 7) provides an explanation on how the respective BenR and XylS transcriptional factors ended up controlling differentially each of their cognate gene clusters. The two DNA sequences at the top of the alignment of Fig. 7 show the organization of the regulatory regions of the Pm and the Pben promoters (Kessler et al., 1994). Note that Pm has two operator sequences for XylS, one partially overlapping the −35 promoter motif (proximal site or Om-p) and a second upstream site (distal or Om-d). Each operator includes two conserved boxes (A and B), the modification of which eliminates responsiveness to XylS (Kessler et al., 1993). Perusal of the corresponding Pben sequences in P. putida KT2440 reveals the presence in this promoter of three out of the four boxes that are required for physiological XylS binding. In particular, the A box of Om-d has lost any similarity to the TGCA sequence. The reconstitution of a full TGCA box in the Om-d operator of Pben enhances BenR inducer sensitivity particularly in response to 3MBz (Silva-Rocha and de Lorenzo, 2012), revealing the key role of the A box in productive operator recognition by this transcriptional regulator. It is plausible that the lack of XylS affinity for Pben caused by the missing A box in the Om-d operator can be compensated by overproducing the regulator, as repeatedly shown in this work, but leaves the promoter inactive under physiological levels of the factor. As shown also in Fig. 7, the modification of the distal A box that plausibly makes the Pben promoter unresponsive to XylS is shared by other P. putida strains, e.g. those named Idaho (Tao et al., 2011) and B6-2 (Tang et al., 2011). These strains use methyl-substituted aromatics such as m-xylene and p-xylene (Cruden et al., 1992) or (methylated) carbazoles, (methylated) dibenzothiophenes and (methylated) benzothiophenes (Tang et al., 2011) as carbon source. These methylaromatics are most likely metabolized through methylcatechols and therefore would run into the same metabolic conflict as with P. putida mt-2. In contrast, other strains not known to deal with substituted aromatics keep the regulatory region of their Pben promoters very similar to that of Pm. These alignments suggest that Pben promoters impaired to respond to XylS prevail in strains evolved to degrade methylaromatics. The adaptation that has permitted the coexistence of somewhat incompatible pathways for degradation of Bz and 3MBz seems therefore to rely on a discrete change in the regulatory region of Pben that makes it to remain responsive to BenR but impervious to XylS. In contrast, both regulators can activate Pm, although with very different kinetics and dynamic parameters (Kessler et al., 1994; Silva-Rocha and de Lorenzo, 2012). Such a lose-to-win scenario exemplifies how intricate metabolic problems can be solved through an unsophisticated regulatory solution that alternates the appearance of divergent pathways rather than merging the two routes or inactivating trouble-making enzymes. Also, such a simple genetic tinkering suggests strategies for orthogonalization of new pathways implanted in a pre-existing metabolic chassis by engineering the regulation in a fashion that prevents misrouting or unproductive catalysis (Mampel et al., 2013).

Figure 7.

Comparative analysis of the Pben promoter region in different P. putida strains. The structure of the Pm promoter of the pWW0 plasmid is shown on top as a reference. The predicted distal (Om-d) and proximal (Om-p) operators for the cognate transcription factors are each formed by two conserved boxes (labelled as A and B) as indicated. Additional features such as the −10/–35 boxes and the predicted transcriptional start sites (+1) are shown.

Experimental procedures

Bacterial strains, culture conditions and general DNA methods

All P. putida strains were derived from P. putida mt-2 (ATCC 33015) or P. putida KT2440 (Bayley et al., 1977; Franklin et al., 1981). Escherichia coli strains were routinely grown in Luria–Bertani (LB) medium at 37°C. Pseudomonas putida strains were grown at 30°C either in LB or in M9 minimal medium (Sambrook and Russell, 2001) supplemented with citrate (0.3%), succinate (0.3%) or Bz (5 mM) as carbon source. Aromatic compounds used as inducers (Bz, 3MBz, 3MBA, m-xylene and o-xylene) were all purchased from Sigma-Aldrich at the highest purity grade available. Where appropriate, antibiotics were added to the medium at the following concentrations (μg ml–1): kanamycin, 50; gentamycin, 10; ampicillin, 150; tetracycline, 20; and streptomycin, 50. Plasmids were transferred to P. putida derivatives by electroporation using a previously described protocol (Choi et al., 2006). Mutant strain P. putida mt-2 (pWW0-xylX::bla) was generated by disruptive homologous recombination as follows. An internal region of xylX gene was amplified by PCR using Go Taq DNA polymerase (Promega) and oligonucleotides xylX-int-FW (5′CAGATTCCCGAGAAGAACGA3′) and xylX-int-RV (5′GTTCTCGACCTGCACTTTCC3′). This fragment was cloned using the pCR2.1-TOPO system (Invitrogen Life Technologies) to generate pTOPO-ΔxylX plasmid that can only replicate in E. coli. When transformed into P. putida mt-2 by electroporation, this plasmid was integrated in pWW0 by homologous recombination that disrupted the target sequence. Mutant derivatives were selected on LB agar containing ampicillin (750 μg ml–1), and correct recombinational insertion was confirmed by PCR. The pWW0-xylX::bla plasmid was transferred by bipartite conjugation on membrane filters (MFTM, 0.45 μm, Millipore) to a P. putida KT2440 derivative having a benR gene inactivated by mini-Tn5 insertion (KmR) obtained from a genome-wide mutant library of P. putida (Duque et al., 2007; Molina-Henares et al., 2010). After 8 h of incubation at 30°C on LB agar, the conjugation mixture was placed on minimal media plates containing kanamycin and ampicillin to select transconjugants of benR mutant harbouring the pWW0-xylX::bla plasmid.

Construction of transcriptional fusions

The assayed promoters were cloned in two different luxCDABE reporter vectors. For transcriptional fusions to be located in chromosome, we used the pUC18-mini-Tn7T-Gm-lux plasmid (GmR) designed as a single-copy integration system (Choi et al., 2005). The Pben and Pm promoters were PCR amplified with Pfu DNA polymerase (Promega) using the primers pairs SacI-Pben-FW (5′ATGGAGCTCACCTGGTAGCTGCAAAAGGA3′)/BamHI-Pben-RV (5′GTCGGATCCGCCAGGGTCTCCCTTGTTAT3′) and SacI-Pm-FW (5′ATGGAGCTCGGCGACGTTCAAGAAGTAT3′)/BamHI-Pm-RV (5′GTCGGATCCTTGTTTCTGTTGCATAAAGCCTA3′), and genomic DNA from P. putida mt-2 strain as template. These primers introduced SacI and BamHI restriction sites (underlined) at the 5′ and 3′ ends respectively. The PCR products were gel purified, digested with SacI/BamHI and ligated to the pUC18-mini-Tn7T-Gm-lux vector previously cut with the same restriction enzymes. The resulting plasmids were sequenced to check integrity and transferred to P. putida by co-electroporation with the helper plasmid pUX-BF13 as described by Choi and colleagues (2005). Primers Pput-glmSUP (5′AGTCAGAGTTACGGAATTGTAGG3′)/PTn7L (5′ATTAGCTTACGACGCTACACCC3′) and Pput-glmSDN (5′TTACGTGGCCGTGCTAAAGGG3′)/PTn7R (5′CACAGCATAACTGGACTGATTTC3′) were used to check GmR colonies for integration of the reporter construct into the natural Tn7 insertion site of P. putida (Choi et al., 2005). For constructing transcriptional fusions in the low-copy and broad-host range pSEVA226 vector (RK2 replication origin, KmR; Silva-Rocha et al., 2013), the Pben promoter was PCR amplified with Pfu DNA polymerase (Promega) using the primers PBF (5′TGGATGAATTCGACAGTACCCTCC3′) and PBR (5′GCGCGGATCCGGCCAGGGTCTCCCTTG3′). This procedure introduced EcoRI and BamHI restriction sites (underlined) at the 5′ and 3′ ends, respectively, of the amplified promoter. The PCR product was gel purified, digested with EcoRI/BamHI and ligated to pSEVA226 digested with the same enzymes. The resulting plasmid was named pSEVA226-Pb and its DNA re-sequenced for ensuring integrity. The equivalent reporter plasmid for the Pm promoter (pSEVA226-Pm) is described elsewhere (Silva-Rocha et al., 2011). Both reporter plasmids were transferred independently to P. putida mt-2 by tri-parental mating with a helper E. coli strain (de Lorenzo and Timmis, 1994).

Promoter activity assays

For analyses of promoter activity, the P. putida derivatives harbouring the different Pben- and Pm-luxCDABE transcriptional fusions were inoculated into 2 ml of M9 minimal medium supplemented with citrate (0.3%) as carbon source and allowed to grow for 16 h (overnight). The cells were then diluted 20 times in the same fresh minimal media added with 1 mM Bz or 3MBz. Two hundred microlitres aliquots of thereby diluted cells were placed in 96-well microplates (OptiluxTM, BD Falcon) and analysed in a WallacVíctor II 1420 Multilabel Counter (Perkin Elmer). The optical density at 600 nm (OD600) and the bioluminescence were recorded at time intervals of 15 min. Promoter activities were determined as the ratio between absolute bioluminescence and cell density, and are given in relative units. The values reported represent the maximal activity reached after at least 3–4 h of incubation.

RNA isolation and tiling microarray experiments

For RNA extraction, P. putida mt-2 strain was precultured overnight in M9 medium with succinate, diluted 100-fold in the same medium and grown to OD600 = 0.3. Samples were then exposed for 2 h to vapours of m-xylene or o-xylene released by a 50% dilution of these aromatics in dibutyl phthalate (a non-effector) in flask for 2 h. Alternatively, soluble substrates Bz (5 mM) or 3MBz (5 mM) were added to the cultures once the growth had reached OD600 ∼ 0.3. As a reference, the same cells were grown for 96 h at 30°C in M9 medium. Once grown, samples were transferred to 1/10 volume of ice-cold ethanol/phenol solution (5% phenol in ethanol) for preventing RNA degradation and harvested by centrifugation (3.800 r.p.m. for 15 min at 4°C). After aspiration of supernatant, the pellets were frozen in liquid nitrogen and stored at −80°C, until required. Total RNA was extracted by using miRNeasy Mini kit (Qiagen) with some modifications. The collected pellets were re-suspended into 0.3 ml Tris-HCl (pH 7.5) containing 2 mg ml–1 lysozyme and incubated for 10 min at 37°C. Then 0.1 ml of the lysate was used in order to extract total RNA according to the manufacturer's instructions. RNase-free DNase (Qiagen) treatment was performed during the isolation procedure to eliminate the residual DNA. The quality of the RNA was evaluated by using a model 2100 bioanalyzer (Agilent Technologies). For transcriptomic analysis, a customized tiling microarray chip was used consisting of 60-mers oligonucleotide probes designed to be spaced every 10 bp and overlapping 50 nucleotides, covering the wholeness of target genes (Kim et al., in preparation). Total RNA (20 μg) of each sample was retrotranscribed and aminoallyl-labelled using SuperScript Indirect cDNA Labeling System (Invitrogen) and 5-(3-aminoallyl)-2′deoxyuridine-5′-triphosphate (aa-dUTP, Ambion) following the manufacturer's instructions. To avoid antisense artefacts of second-strand complementary DNA (cDNA) during reverse transcription, actinomycin D was added after the denaturing step at 70°C to a final concentration of 6 μg ml–1, as described (Perocchi et al., 2007). For each sample, aminoallyl-labelled cDNA was re-suspended in 0.1 M Na2CO3 (pH 9.0) and conjugated with either Cy3 or Hyper 5 Mono NHS Ester (CyTMDye Post-labelling Reactive Dye Pack, Amersham), following a dye-swap strategy. The samples were purified with MegaclearTM (Ambion) following the manufacturer's instructions. Cy3 and Hyper 5 incorporation was measured in a Nanodrop spectrophotometer (Nanodrop Technologies). Preparation of probes and hybridizations were performed as described (Two-Color Microarray-Based Prokaryote Analysis, Agilent Technologies). The samples were placed on ice and quickly loaded onto arrays, hybridized at 65°C for 17 h and then washed once in GE wash buffer 1 at room temperature (1 min) and once in GE wash buffer 2 at 37°C (1 min). Arrays were drained by centrifugation at 2000 r.p.m. for 2 min. Images from Cy3 and Hyper5 channels were equilibrated and captured with a GenePix 4000B scanner (Axon) and spots quantified using GenePix software (Axon). After scanning, both datasets were analysed using the same methods: raw intensities were background-corrected by normexp method with an offset of 50 (Smyth and Speed, 2003). Background-corrected intensities were converted to log2 scale and normalized by adjusting the quartiles of all replicates as described in Bolstad and colleagues (2003). After normalization, differential expression for every probe was calculated as log2Ratios = log2Intensity (experimental condition) – log2Intensity (reference condition). Integrated Genome Browser (Nicol et al., 2009) was used to represent the log2Ratios of all probes along the particular regions.


This study was supported by the BIO programme of the Spanish Ministry of Science and Innovation, the ST-FLOW, ARISYS and EVOPROG Contracts of the EU, the ERANET-IB Program and the PROMPT Project of the Autonomous Community of Madrid. D. P. P. is the holder of a Marie Curie Actions Program grant of the EC for visiting scholars PIIF-GA-2009-253825. J. K. is a beneficiary of the JAE Program of the CSIC. Authors declare no conflict of interest.