The logic layout of the TOL network of Pseudomonas putida pWW0 plasmid stems from a metabolic amplifier motif (MAM) that optimizes biodegradation of m-xylene

Abstract

Background: The genetic network of the TOL plasmid pWW0 of the soil bacterium Pseudomonas putida mt-2 for catabolism of m-xylene is an archetypal model for environmental biodegradation of aromatic pollutants. Although nearly every metabolic and transcriptional component of this regulatory system is known to an extraordinary molecular detail, the complexity of its architecture is still perplexing. To gain an insight into the inner layout of this network a logic model of the TOL system was implemented, simulated and experimentally validated. This analysis made sense of the specific regulatory topology out on the basis of an unprecedented network motif around which the entire genetic circuit for m-xylene catabolism gravitates.

Results: The most salient feature of the whole TOL regulatory network is the control exerted by two distinct but still intertwined regulators (XylR and XylS) on expression of two separated catabolic operons (upper and lower) for catabolism of m-xylene. Following model reduction, a minimal modular circuit composed by five basic variables appeared to suffice for fully describing the operation of the entire system. In silico simulation of the effect of various perturbations were compared with experimental data in which specific portions of the network were activated with selected inducers: m-xylene, o-xylene, 3-methylbenzylalcohol and 3-methylbenzoate. The results accredited the ability of the model to faithfully describe network dynamics. This analysis revealed that the entire regulatory structure of the TOL system enables the action an unprecedented metabolic amplifier motif (MAM). This motif synchronizes expression of the upper and lower portions of a very long metabolic system when cells face the head pathway substrate, m-xylene.

Conclusion: Logic modeling of the TOL circuit accounted for the intricate regulatory topology of this otherwise simple metabolic device. The found MAM appears to ensure a simultaneous expression of the upper and lower segments of the m-xylene catabolic route that would be difficult to bring about with a standard substrate-responsive single promoter. Furthermore, it is plausible that the MAM helps to avoid biochemical conflicts between competing plasmid-encoded and chromosomally-encoded pathways in this bacterium.

Figure 1

Overview of the TOL network. At the metabolic level, m-xylene is first converted to 3-methylbenzoate (3 MBz) through the action of the enzymes encoded by the upper operon, and this intermediate compound is further metabolized into the TCA cycle by the activities born by the meta operon. In the sketch, XylR and XylS are transcriptional regulators while Pu, Pm, Ps and Pr are promoters. At the regulatory level, the master regulatory gene xylR is encoded in a location adjacent to the end of the meta operon and expressed from the Pr promoter (not to scale). The corresponding TF is produced in the so-called inactive form (XylRi). This protein changes to an active form (XylRa) when bound to the inducer m-xylene or its first intermediate 3-methylbenzyl alcohol (3 MBA, not shown). XylRa then activates both Pu and Ps, which triggers expression of the upper pathway and stimulates production of XylS respectively [50]. In the absence of m-xylene, this second regulator XylS is produced at low levels, and it changes from the inactive form XylSi to the transcriptionally proficient XylSa by binding to 3 MBz [23]. This XylSa form is able to induce meta pathway expression by activating Pm. But, concomitantly, high levels of XylS triggered by XylRa-mediated Ps activation can also induce Pm activity. This activation loop is formalized as an alternative XylS form (XylSh, for hyper-expressed XylS [32])

Figure 2

Formalization the TOL network as a logic circuit. (a) Interplay between transcriptional factors. The scheme blows up the divergent Pr/Ps region that controls expression of XylR and XylS, respectively. In the absence of m-xylene, XylR represses weakly its own transcription from Pr, and an inactive form of XylS (XylSi) is expressed through a low-constitutive divergent promoter Ps2. The presence of m-xylene both increases XylR auto-repression and activates the σ⁵⁴-dependent Ps1 promoter, thereby strengthening XylS expression to the point of reaching a high concentration (XylSh) able to activate the Pm promoter of the lower operon (see text). (b) Basic logic gates (AND, OR and NOT) used for constructing the model presented in this work, along with the respective truth tables. (c) The minimal TOL logicome, which represents the core logic interactions taking place in the system. Expression of the upper pathway is represented by an AND gate having both XylR and m-xylene as inputs, the same being true for XylS production. 3 MBz synthesis is represented also as an AND gate with the upper pathway and m-xylene as the inputs. For expression of meta, the formation of XylSa (XylS plus 3 MBz) is presented as an AND gate where the output is connected to an OR gate, where the second input is overproduced XylS itself (XylSh). This is because Pm can be induced by either low level-XylS along with 3 MBz as an effector or high level, effector-free XylS (see text). Finally, degradation of 3 MBz into TCA metabolic intermediate is represented by an AND gate with the meta enzymes and the 3 MBz substrate as inputs. Note that the XylR auto-repression loop has been eliminated for the model, since the actual levels of this protein are known to change little in the presence/absence of m-xylene.

Figure 3

Simulation of the TOL logicome in the presence or absence of m-xylene. Piecewise-linear differential equations describing the regulatory and metabolic events of the network were implemented in GNA software and the behavior of the TOL simulated in response to m-xylene. (a) Non-inducing conditions. The state transition graph resulting from the simulation is shown to the left with the shortest path between defined states indicated in color. The plots to the right show the temporal sequence of qualitative states for the two regulators (XylR and XylS) and the two pathways (upper and meta) in the selected path of the transition graph. (b) Induced conditions. The transition graph is shown to the left while the temporal sequence of qualitative states is displayed to the right. As before, color states highlight the shortest path in the transition graph.

Figure 4

Analysis of of Pu and Ps activation dynamics by XylRa. (a) Proposed single-input module for XylR (SIM_XylR). In this motif, XylRa controls negatively its own expression and activates XylS and the upper pathway. While no other target is known for XylR, it cannot be excluded that this regulator controls additional genes (represented as X). (b) Simulations for upper and xylS expression under inducing condition show the synchrony of gene activation. (c) Genetic constructs used to analyze promoter kinetics. The architecture of Pu, Ps and Pm are sketched. The UAS (for upstream activator sequences) for XylR in Pu and Ps, and the XylS binding sites of Pm are shown, with an indication of the boxes for σ⁵⁴-RNAP (-12/-24) and σ⁷⁰-RNAP (-10/-35) recognition. Below, the main features of the broad host range lux reporter vector pSEVA226 were each of the promoters was cloned are indicated. (d) Light emission of reporter strains P. putida mt-2 (pSEVA226Pu), P. putida mt-2 (pSEVA226Ps) and P. putida mt-2 (pSEVA226Pm). Each of the strains was cultured in minimal medium with succinate and then added with 5 mM 3MBA as described in the Methods section. Light emission was recorded after 4 h and the figures of bioluminescence/OD₆₀₀ converted into arbitrary promoter activity units, A.U. (e) Induction kinetics of Pu and Ps assayed in minimal/succinate medium and 1 mM of 3MBA. Reporter strains P. putida mt-2 (pSEVA226Pu) and P. putida mt-2 (pSEVA226Ps) were treated as before but the aromatic inducer was present throughout the entire growth. (f) Relative induction kinetics of Pu, Ps and Pm in minimal medium with 5.0 mM 3MBA as the sole carbon source. Promoter activities were normalized in all cases in respect to their respective maximum values. Note the virtual identity between Pu and Ps promoters and the delay of Pm.

Figure 5

Modeling the effect of the XylSh loop in TOL system. (a) Signal transmission/conversion in the TOL system. The diagram sketches interactions between the active forms of the regulators and the metabolic intermediate 3-methylbenzoate (3 MBz). To the left, inactive XylR (XylRi) becomes activated by m-xylene to produce the transcriptionally competent form XylRa. This in turn, results in activation of the upper pathway and overproduction of XylS (XylSh), which can by itself activate meta pathway. Such a XylSh loop (marked in blue), which does not involve 3 MBz, links the meta pathway directly to m-xylene presence. To the right, XylS produced at low levels, insufficient for activating meta (XylSi) turns into an active form (XylSa) to the same end upon binding the 3 MBz produced by the action of upper on m-xylene. Finally, production of meta converts 3 MBz into Krebs' cycle intermediates. (b) Simulation conditions. Wild type considers the complete model where meta is concomitantly expressed through both XylSa-mediated and XylSh-mediated paths. In No XylS hyper-expressed conditions the effect of XylSh has been removed and meta is activated only by XylSa. In No XylS activation condition, the effect of XylSa has been deleted and meta is under the sole control of XylSh. (c) Temporal sequence of qualitative states for each of the three conditions. Each scenario was simulated until the system reached a steady state.

Figure 6

Experimental strategy for quantification of XylSa-dependent and XylSh-dependent Pm/meta activation. (a) Default scenario, i.e. Pm is activated by both XylSa and XylSh. The inducer employed in this case is m-xylene (or its proxy 3MBA), which both activates XylR (and thus triggers the XylSh loop and is metabolized by upper to produce 3 MBz, necessary for XylSa formation. (b) XylSa alone i.e. no XylSh. The added inducer is 3 MBz, which is specific for XylS. (c) XylSh alone i.e., no XylSa. The inducer employed is ortho-xylene (o-xylene), which fully activates XylR (thus generating high levels of XylS = XylSh) but cannot be converted into 3 MBz and therefore XylSa cannot be formed. (d) Pu activation by m-xylene and o-xylene. Reporter strain P. putida mt-2 (pSEVA226Pu) was patched on the surface of minimal-succinate agar plates, grown overnight and then exposed to saturating vapors of either inducer as indicated. Bioluminescence was captured along time and the figures in arbitrary units represented with a color code according to the signal intensity (bar on the right represents the scale). Nil: Control with no inducer. (e) Promoter activities on the basis of the densitometry of the images of panel (d). Values were normalized in respect to maximum activity as above.

Figure 7

Pm regulation through alternative control loops. (a) Pm activation in response to vapors of m-xylene or o-xylene assayed in solid media. Patches of the reporter strain P. putida mt-2 (pSEVA226Pm) were grown on the surface of M9-succinate agar and then exposed to saturating vapors of m-xylene (which triggers the appearance of both XylSh and XylSa), or o-xylene (which makes cells to produce only XylSh). (b) Pm promoter activity deduced from images of panel (a) processed identically as in Fig. 6. (c) Pm activation kinetics in liquid media added with 1 mM 3MBA (m-xylene proxy, leading to both XylSh and XylSa) or 1 mM of 3 MBz (appearance of XylSa only). Promoter activities of reporter strain P. putida mt-2 (pSEVA226Pm) are shown in respect to the maximal value reached with 3MBA induction. (d) Contribution of each regulatory device to Pm activity. The bar diagram compares standardized promoter activities brought about by the XylSa-dependent loop (3 MBz), by the XylS hyper-expression loop (o-xylene induced) and both (m-xylene or 3MBA induced). Promoter activity is represented as the maximal value obtained in every experimental condition relative to the highest m-xylene (or 3MBA) induction value.

Figure 8

The inner logic of the TOL regulatory network. (a) Layout of a canonical type I coherent Feed Forward Loop (FFL). In such a motif, a master regulator × activates expression of a target Z both directly and indirectly. Indirect regulation takes through activation of a second transcriptional factor Y which in turn has Z as a target as well. S_X and S_Y are the signals which trigger the activity of × and Y, respectively. The ara operon is shown as an example of this type of FFL, as its expression depends on the interplay between the CRP and AraC regulators, cAMP and arabinose being the S_X and S_Y inducers respectively. (b) Metabolic Amplifier Motif (MAM) found in the TOL network. Compared to the type-I FFL motif, the indirect regulation of Z through the X→Y node remains, but the direct interaction X→Y makes a detour that involves a metabolic (rather than regulatory) action. Specifically, the master regulator × now activates the production of an enzyme (or a metabolic pathway) W, which converts the signal S_X into S_Y. In the TOL system, × and Y are represented by XylR and XylS, while m-xylene (S_X) is converted to 3 MBz (S_Y) by the action of the upper pathway.