Pseudomonas putida , the best-characterized member of the wide group of fluorescent Pseudomonas species, is a non-pathogenic soil bacterium with a remarkable and versatile metabolism that has evolved to withstand harsh environmental conditions. These properties have established it as one of the workhorses for bioremediation approaches and sustainable production of value-added chemicals3. Regarding the aromatic catabolic pathways, P. putida contains 12 out of the 14 identified central pathways4,5, with the ring-cleavage mechanism of β-ketoadipate and homogentisate pathway, mediated by dioxygenases, significantly involved in the reduction of most low-molecular-weight phenolic compounds found in OMW6, 7, 8,such as cinnamic, p-coumaric, caffeic, ferulic and vanillic acid ( Fig. 1).  Therefore, P. putida was selected as the primary bacterium, since it can effectively degrade a plethora of prominent monomeric phenolic compounds of OMW, exhibit phytotoxic9 and antimicrobial10 activities, as well as the other components that contribute to its high organic load (such as sugars, organic acids, and lipids) and utilize them as a source for the production of value-added products11. Those characteristics make P. putida fit in our design without the need of additional genetic modification for the detoxification aspect. Nevertheless, a series of metabolic engineering was designed to optimize the funneling of phenolics and other organic compounds into the production of biobased biodegradable polymers (Fig.1), known as polyhydroxyalkanoates, or PHAs (see “PHA production” section for more details).

 While some aerobic bacteria, such as P. putida, are effective against low-molecular-weight phenolic compounds, they are not as efficient with polymeric, lignin-like phenolics, such as tannins, lignans and catechol melaninic polymers12, mainly responsible for the dark coloration of OMW13. White-rot basidiomycetes, on the other hand, excel in breaking down these complex compounds14 due to a system consisting of extracellular ligninolytic enzymes, such as laccase, lignin peroxidase, and manganese-dependent peroxidase. Laccase, an oxidoreductase, catalyzes the non-specific oxidation of aromatic substrates while simultaneously reducing molecular oxygen to water15, offering an advantage over peroxidases that require costly hydrogen peroxide as a co-substrate.
 Therefore, in our design, E. coli plays a crucial supportive role in OMW detoxification by co-expressing and secreting a laccase derived from the white-rot fungus Trametes versicolor16, developing a laccase secretion system for the degradation of polymeric phenolic compounds found in OMW17. As this enzyme originates from an organism of a different kingdom, we had to implement modifications to ensure its successful integration. For this reason, the upstream coding sequence of the mature laccase was fused with the sequence of an optimal signal peptide and then the whole coding sequence was placed under the control of the well-known inducible system AraCThe AraC protein
serves a dual role for the regulation of the pBAD promoter .
It functions as a positive regulator, enhancing transcription when arabinose is present.
Conversely, it serves as a negative regulator when arabinose is absent, suppressing the initiation of transcription/PBAD 18(Fig.2.), since OMW contains significant concentrations of L-arabinose19, 20. Lastly, in order to overcome carbon catabolite repression (CCRCarbon catabolite
repression functions by suppressing the production of enzymes responsible
for breaking down carbon sources that are not the organism's preferred choice.
CCR is a significant component of the overall regulatory system in many bacteria and other microorganisms.) against our system’s induction, but also substrate competitions, in our theoretical design, we decided to include a knock out for the ptsG gene, encoding the major glucose phosphotransferase system, based on λ-red and I-SceI, as described by Lin Zhenquan21.

P. putida produces medium-chain-length polyhydroxyalkanoate (mcl-PHA) copolymers (monomers ranging from C6-C14) via two primary and complementary pathways, one using PHA-related carbon sources such as free fatty acids, as they are structurally similar to the key monomers that make up PHA, and a second pathway utilizing PHA-nonrelated carbon sources such as sugars, aromatic compounds, and organic acids24. In the presence of PHA-related carbon sources, the biosynthesis of PHAs is orchestrated by β-oxidation, since it provides the peripheral PHA biosynthetic pathway with the intermediates of fatty acid degradation. When PHA-nonrelated substrates are fed to P. putida, carbon fluxes within central catabolic pathways are directed to the key central intermediate acetyl-CoA and channeled toward PHA formation via the de novo fatty acid synthesis (Fig. 3).  Although most of the previous strategies for improved PHA production are limited to enhancing only one of the involved pathways, we considered that such a scenario would limit the extent to which such a rich substrate as OMW could be valorized. So, we came up with two different regulatory circuits, aiming for the activation of the producing devices directly from the contents of the OMW, both from the related and nonrelated carbon sources.  Starting with the PHA production from nonrelated substrates (e.g. glucose, phenolic acids and acetic acid), we have pinpointed two key enzymes, acyl-CoA synthase (alkK) and hydroxyacyl-ACP acyl-transferase (phaG), which play a pivotal role in connecting the de novo fatty acid synthesis with the PHA biosynthetic pathway25. To control the expression of the two genes, we designed a regulatory system based on the transcription factor PcaU, an activator found in Acinetobacter baylyi ADP126 (Fig. 3b). PcaU has the unique property of binding to specific operator regions whether the effector molecule is present or absent27. In our design, this system would be triggered by the presence of the intermediate compound protocatechuic acid (PCA), formed during the catabolism of OMW phenolic compounds through the protocatechuate branch of the β-ketoadipate pathway (see "OMW detoxification"). Hence, this regulatory mechanism would enhance the funneling of nonrelated substrates towards the production of PHA .  Regarding the PHA production from related substrates (e.g. free fatty acids), the phaJ gene was chosen to be incorporated in a regulatory system since it encodes the trans-enoyl-CoA hydratase that catalyzes the linking step between β-oxidation and the PHA machinery by converting the β-oxidation intermediate 2-trans-enoyl-CoA to (R)-3-hydroxyacyl-CoA28 (Fig. 3a). For the regulation of its expression, a free fatty acid (FFA)-sensitive system (Fig. 4) was constructed, based on the E. coli fatty acid metabolism repressor FadR and a synthetic promoter with FadR recognition/binding sites29. So, when fatty acids accumulate, they would be first activated to the corresponding acyl-CoA by an acyl-CoA synthetase (fadD)30. Acyl-CoAs in turn would bind to FadR, releasing in that way the promoter region and simultaneously activating the metabolic hub sitting between β-oxidation and PHA synthesis.

Fig. 4. Schematic representation of the PfadBA-FadR system. In absence of free fatty acids (FFAs), the consistently produced FadR protein inhibits the function of the PfadBA promoter, suppressing the phaJ expression. When FFAs are present, they are activated to acyl-CoAs by acyl-CoA synthase (encoded by fadD). Acyl-CoAs bind to the FadR protein, allowing the PfadBA promoter to function properly, leading to the transcription of the phaJ gene.

 Besides the incorporation of the phaJ gene, weakening β-oxidation is considered another rational approach for boosting mcl-PHA production from related substrates31. Therefore, two well-identified genes encoding 3-hydroxyacyl-CoA dehydrogenase (fadB) and 3-ketoacyl-CoA thiolase (fadA) were chosen to be knocked out in P. putida, leading the conversion of most fatty acids into 3-hydroxyacyl-CoA32 (Fig. 3.d).  Both main pathways converge in the creation of (R)-3-hydroxyacyl-CoA units, which are then transformed by two mcl-PHA synthases33 (or polymerases) into a polyester chain by releasing one CoA molecule per catalytic cycle. To enhance the polymerization of PHA precursors the phaC1 and phaC2 genes, encoding the PHA polymerases, were designed to be overexpressed using the Anderson promoter J23100, a choice that was supported again by Dr. Tsampika Manoli Maria (Fig. 3.c).  However, we must keep in mind that PHA production is a dynamic process, alternating between synthesis and consumption of the polymers depending on the environmental and feeding conditions. Therefore, in order to prevent the degradation of the final product into free (R)-HA monomers, we have included the deletion of the phaZ gene, encoding the PHA depolymerase34 (Fig. 3.d).

 While unable to synthesize PHAs natively, E. coli has many characteristics that make it suitable for the supporting role in our system. E. coli was selected and designed to preferentially utilize L-arabinose, D-xylose and D-galactose, sugars found in OMW which P. putida cannot metabolize35, by knocking out the ptsG gene (see “OMW detoxification”), and at the same time produce free fatty acids (FFAs), an ideal substrate for the synthesis of PHAs by Pseudomonas putida, according to "Maria Tsampika" , a Postdoctoral researcher at Biological Research Center Margarita Salas. However, P. putida contributes also to the consortium by utilizing potentially toxic intermediate metabolites of E. coli, such as acetic acid36, as carbon source for the production of PHAs (see “PHA production”).  Under normal conditions, the biosynthesis of fatty acid is strictly regulated37. Unlike most oleaginous microorganisms or plants, E. coli is unable to accumulate significant quantities of free fatty acids naturally38. Even though most engineering approaches to improve yields and productivities include overexpressing bottleneck enzymes and bypassing native regulations of the fatty acid metabolism, these imbalanced conditions can prove destructive for the host39. For this reason we propose a malonyl-CoA-based negative feedback system in E. coli, based on Liu et al40, for improved fatty acid titer and productivity.  Malonyl-CoA is synthesized from acetyl-CoA by acetyl-CoA carboxylase encoded by accABCD (acc). The acetyl-CoA carboxylase consists of four subunits: a biotin carboxyl carrier protein, a biotin carboxylase, and two carboxyltransferase subunits. This is the first and rate-limiting step of fatty acid biosynthesis41. To increase the malonyl-CoA supply while preventing the harmful effects of overexpressing the acc, we developed a system that can sense malonyl-CoA levels and regulate acc expression accordingly (Fig. 6). We built this system around a malonyl-CoA-responsive element called FapR, a transcription factor which binds to a specific 17-bp DNA sequence, thus negatively regulating fatty acid and phospholipid metabolism in the Gram-positive bacteria Bacillus subtilis42.  When malonyl-CoA binds to FapR, it causes a conformational change in the FapR protein, leading to the dissociation of the FapR-DNA complex. We leveraged this mechanism to create a feedback loop for the negative regulation of acc expression. In this system, the acc is controlled by a T7 promoter that is regulated by the LacI, a repressor protein. LacI expression is under the control of a synthetic promoter called PFR1 , which is regulated by FapR, since it contains FapR-binding sites. Additionally, FapR is influenced by malonyl-CoA levels.  When there is an excess of malonyl-CoA, the biosensor activates the expression of lacI. LacI then, through the inhibition of the PT7 , down-regulates the expression of acc, which helps prevent the toxic effects associated with acetyl-CoA carboxylase overproduction, such as the depletion of free CoA pool by malonyl-CoA accumulation43. With malonyl-CoA levels dropping again, FapR in its free state, can bind to the PFR1 , thus inhibiting the transcription of lacI. PT7 is then active to transcribe the acc. This negative feedback circuit ensures a balanced and tightly controlled acc expression and at the same time an enhanced malonyl-CoA supply for efficient fatty acid biosynthesis.

Fig. 6. Schematic representation of the malonyl-CoA-based negative feedback system for the regulation of acc expression. In the presence of low malonyl-CoA levels, the FapR protein, in its unbound state, is able to bind to the PFR1 promoter causing its inhibition. Thus, the lacI is not expressed and the PT7 is active transcribing the acc. The expression of acc is equivalent to the rise of Malonyl CoA levels, which, above a concentration threshold, bind to and restrict the FapR protein, leaving the PFR1 active for lacI transcription. LacI inhibits the PT7 , switching the expression of the acc off and consequently, the malonyl CoA levels drop.

 The integration of the negative feedback system will provide an increased flux of malonyl-CoA to the fatty acid biosynthesis (fatty acid elongation cycle), which, in contrast to β-oxidation, operates on acyl-ACP thioesters intermediates. The release of free fatty acids requires the function of the acyl-ACP thioesterase, which terminates fatty acyl group extension by hydrolyzing the acyl moiety from the acyl-ACP. So, to further optimize the whole process we chose to knocked out the fadD gene, encoding the long-chain acyl-CoA synthetase44 and introduced the acyl-ACP thioesterase from the castor plant (Ricinus communis), an enzyme that can produce significant quantities of three major straight chain free fatty acids (C14, C16 and C16:1) when heterologously expressed in E.coli as confirmed by Zhang et al45. Hence, E. coli will provide P.putida with a preferred carbon source for PHA synthesis, contributing drastically to the efficiency and the overall cost of the process.

P.Putida

 Bacteria accumulate PHAs as insoluble and highly hydrophobic intracellular inclusion bodies, composed of the polymer proper and an organized layer of granule-associated proteins47 (GAPs). Since traditional methods, involving hydrolytic enzymes, sonication, high temperatures, and environmentally-unfriendly solvents/detergents reagents, present many limitations48 (e.g. high energy and chemical demands) there is a great need for a novel efficient recovery process. Therefore, we decided to create a controllable system that, upon induction, would cause the breakdown of the cellular structure, resulting in the release of the PHA granules. As the central element of this system, we selected lysozyme, an enzyme that hydrolyses the 1,4-beta-linkages between N-acetylmuramic acid and N-acetyl-D-glucosamine residues in the peptidoglycan layer, a major constituent that provides integrity to the cell wall49. To ensure the translocation of the recombinant lysozyme into the periplasmic space, the lysozyme C precursor sequence of Gallus gallus was fused at the N-terminus with the sequence of a signal peptide described for a naturally secreted protein in Pseudomonas stutzeri 50.  All that remained was to choose the appropriate inducible expression system for the regulation of the genetic SP-lysozyme entity. This selection was based on some key criteria such as the tight system regulation, in order to avoid a premature cell lysis, high expression levels in the presence of the inducer for high final biopolymer titer and the low cost of the inducer for large-scale extraction. Ultimately, the MekR/PmekA system, a methyl ethyl ketone (MEK)-inducible system, derived from the strain Pseudomonas veronii MEK70051, emerged as the ideal choice that met our requirements. This system, which is positively regulated by MekR, a member of the AraC/XylS family of regulators (Fig. 7), presents extremely tight regulation accompanied with three magnitudes of fold increase of gene expression after treatment with MEK, which is considered an ideal inducer since it can act also as a promising ‘green’ solvent for mcl-PHA extraction52.

Co-dependency
 To ensure thebiocontainment of our genetically engineered bacteria, we had to make them dependent on specific conditions inside the bioreactor. Through our search, we concluded that the best choice would be the establishment of a cross-feeding community of E. coli and P. putida, based on amino acid exchange, thus introducing a layer of co-dependency between the two bacteria.

Double auxotrophy
 The selection of the amino acids for the creation of cross-feeder strains is considered particularly important. According to the literature and the thorough study of the Wageningen UR iGEM 2021 team, it was determined that the most promising amino acids are Tyrosine (Tyr) and Tryptophan (Trp). Specifically, E. coli was designed to be auxotrophic for Trp and overproducing Tyr, which require the deletion of trpE and tyrR genes respectively53. On the other hand, P. putida was chosen to be auxotrophic for Tyr, which requires a two-step deletion of tyrA and pheA genes54 (Fig. 8), and overproducing Trp, relying on an overexpression vector harboring trpES40F, a feedback-resistant allele of trpE55. Overall, this strategy, along with a lysozyme-based lysis system for PHAs recovery (see“PHA recovery” ), ensures the safety of “oPHAelia”(Fig. 9 ).  It is worth noting that the implementation of this strategy provides not only a layer of safety but can also play a critical role in the maintenance of the two bacteria relative ratio during co-culture56, a parameter which can affect the stability of the microbial consortium.

Fig. 8 Schematic representation of aromatic amino acids (phenylalanine, tyrosine and tryptophan) biosynthetic pathways in P.putida KT2440