Biosensor and chemo-enzymatic one-pot cascade applications to detect and transform PET-derived terephthalic acid in living cells

Summary Plastic waste imposes a serious problem to the environment and society. Hence, strategies for a circular plastic economy are demanded. One strategy is the engineering of polyester hydrolases toward higher activity for the biotechnological recycling of polyethylene terephthalate (PET). To provide tools for the rapid characterization of PET hydrolases and the detection of degradation products like terephthalic acid (TPA), we coupled a carboxylic acid reductase (CAR) and the luciferase LuxAB. CAR converted TPA into the corresponding aldehydes in Escherichia coli, which yielded bioluminescence that not only semiquantitatively reflected amounts of TPA in hydrolysis samples but is suitable as a high-throughput screening assay to assess PET hydrolase activity. Furthermore, the CAR-catalyzed synthesis of terephthalaldehyde was combined with a reductive amination cascade in a one-pot setup yielding the corresponding diamine, suggesting a new strategy for the transformation of TPA as a product obtained from PET biodegradation.


First bioreduction of terephthalic acid (TPA) by a carboxylic acid reductase in vivo
Real-time, highthroughput detection of TPA-derived aldehydes by luciferase LuxAB Bioluminescence reflects TPA amounts, assessing (engineered) PET hydrolase activity Transformation of TPA into the diamine through chemo-enzymatic one-pot cascade

INTRODUCTION
The global production of plastics is rapidly increasing. More than 8% of the global petrochemical production -4% as source for materials and 4% to cover energy demands -were consumed by plastic manufacturing industries (Hopewell et al., 2009). However, only a fraction of discarded plastic is recycled (Geyer et al., 2017). Consequently, efficient disposal and sustainable recycling strategies for plastic waste are urgently needed to reduce the risk of pollution imposed on ecosystems and human health (Eriksen et al., 2014;Rahimi and García, 2017;Vollmer et al., 2020;Wright and Kelly, 2017). Furthermore, to decrease both carbon dioxide (CO 2 ) emissions and the dependence on fossil fuel-based resources, a circular plastic economy is regarded as the central -and vital -approach (Raoul et al., 2021;Sarah and Gloria, 2021;Simon et al., 2021;Wei et al., 2020).
Particularly, the biocatalysis-based recycling of polyethylene terephthalate (PET), which is extensively used to manufacture food packaging and beverage containers, has become a vivid field of research with the discovery of microbial PET-degrading enzymatic activities (Kawai et al., 2019(Kawai et al., , 2020Tournier et al., 2020;Wei et al., 2020Wei et al., , 2022Yoshida et al., 2016). So far, PET hydrolases from actinomycetes including different Thermobifida strains (Herrero Acero et al., 2011;Mü ller et al., 2005;Wei et al., 2014), from the bacterium Ideonella sakaiensis (Yoshida et al., 2016), and a commercial cutinase from the fungi Thermomyces insolens, formerly known as Humicola insolens, have been employed (Ronkvist et al., 2009). Recently, a variant of the compost metagenome-derived and highly thermostable leaf-branch compost cutinase (LCC) (Sulaiman et al., 2012) was engineered toward increased PET-hydrolyzing activity, which pushed the enzymatic depolymerization of PET from laboratory scales to industrially relevant metrics by degrading amorphized (i.e., pretreated) postconsumer PET bottles in only 10 h reaction time . This and the fact that LCC as well as other PET hydrolases were found in public metagenome databases will certainly advance biotechnological plastic degradation and recycling in the near future (Bornscheuer, 2016;Danso et al., 2018;Wei et al., 2020Wei et al., , 2022. Despite the many achievements in the last two decades, the activity of PET hydrolases is still assessed by simply measuring the weight loss of the residual bulk PET polymer after depolymerization (Wei et al., 2019a(Wei et al., , 2019bYoshida et al., 2016) or the chromatographic analysis and quantification of degradation intermediates and/or products such as terephthalic acid (TPA) and its monoesters and diesters (Eberl et al., 2009;Herrero Acero et al., 2011;Palm et al., 2019). Recently, an isothermal titration calorimetry-based method has been established for directly assessing the enthalpy of ester hydrolysis, thus enabling a real-time monitoring of the enzymatic PET hydrolysis (Vogel et al., 2021). All these strategies suffer from laborious sample preparation and the only low to moderate sample throughput, impeding the characterization of novel biocatalysts -not only limited to polyester hydrolases -and the screening of large protein libraries (Markel et al., 2020;Wei et al., 2020Wei et al., , 2022Yi et al., 2021). This obstacle was addressed by a Fenton chemistry-mediated fluorometric detection assay for TPA in a 96-well microtiter plate format, suitable for high-throughput (HT) screening applications Wei et al., 2012). The assay is based on the formation of hydroxyl radicals mediated by an Fe(II)-ethylenediaminetetraacetic acid complex in the presence of molecular oxygen (O 2 ) (Saran and Summer, 1999;Wei et al., 2012;Welch et al., 2002); hydroxyl radicals and TPA then react to the fluorescent 2-hydroxyterephthalate (l excitation = 315 nm, l emission = 421 nm).
Complementary, genetically encoded biosensor systems have been used for the detection of small molecules and included transcription factors (TFs), riboswitches, or enzyme-coupled sensor devices Dietrich et al., 2010;Lehtinen et al., 2017;Liu et al., 2015Liu et al., , 2017Yi et al., 2021). To date, only two biosensors have been reported to detect TPA in vivo. The first was assembled by Pardo et al. and comprised the TF TphR and its regulatory nucleotide sequences from Comamonas testosteroni and the superfolder green fluorescent protein (sfGFP) (Pardo et al., 2020). TphR is a transcriptional activator, which -upon binding of TPA -acts as the inducer of a gene cluster responsible for the conversion of TPA to protocatechuate in Comamonas strains (Kasai et al., 2010). Their TF-based biosensor system facilitated the screening of TPA transporter variants, in other words, the improved uptake of TPA from the environment in Acinetobacter baylyi ADP1 through fluorescence-activated cell sorting (Pardo et al., 2020). The second example featured sfGFP as the fluorescence reporter and the promiscuous TF XylS from Pseudomonas putida, which was engineered by Li and coworkers to bind TPA additionally to reported benzoic acid derivatives (Li et al., 2022). With the efficient detection of TPA in living cells, these sensing devices have yet to be tested for the directed evolution of PET hydrolases by the HT-assisted detection of TPA as PET degradation product.
Most recently, the luciferase LuxAB from Photorhabdus luminescens (P. luminescens) was introduced for the detection of structurally diverse aldehydes in Escherichia coli (E. coli) . In the present work, the carboxylic acid reductase from Mycobacterium marinum (CAR Mm ) was shown to transform TPA into the corresponding aldehydes 4-carboxybenzaldehyde (4-CBAL) and terephthalaldehyde (TAL) in vivo ( Figure 1). The coupling of the enzymatic reduction to the LuxAB biosensor device yielded bioluminescence that semiquantitatively reflected increasing amounts of TPA in PET hydrolysis samples obtained through various hydrolases. The system not only provides a biosensor-based HT assay for TPA but the first biocatalytic route toward highly reactive TPA-derived aldehydes such as TAL, avoiding hazardous chemical procedures (Barnicki, 2017;Snell and Weissberger, 1940). Following the transformation of TAL in the same reaction vessel, the corresponding diamine was yielded and will allow for potential industrial applications (Brindell et al., 1976;Rohan et al., 2015;Suematsu et al., 1983;Wang et al., 2021).

Optimization of whole-cell biotransformations and evaluation of HT assay conditions
In a previous study, the monooxygenase LuxAB from P. luminescens was expressed in E. coli K-12 MG1655 RARE (Kunjapur et al., 2014), herein referred to as E. coli RARE, and provided a reliable detection tool for aldehydes in living cells in a 96-well microplate format , importantly, beyond the previously reported long-chain aliphatic aldehydes (Colepicolo et al., 1989). Furthermore, LuxAB was suitable to sense aldehydes, including aromatic products such as benzaldehyde, cuminaldehyde, and 2-phenylacetaldehyde that were enzymatically produced from carboxylic acid substrates by the co-expression of CAR Mm in the same cell . Prompted by the structural relatedness of these aromatic aldehydes to TPA-derived aldehydes, the capabilities of (1) CAR Mm -to reduce one or both carboxylic acid functionalities of TPA to the aldehyde -and (2) LuxAB -to accept aldehyde products formed in situ, thereby yielding bioluminescence -were investigated.
Therefore, chemically competent E. coli BL21(DE3) cells were transformed with pACYCDuet-1/car Mm :ppt Ni to co-express CAR Mm and a phosphopantetheinyl transferase from Nocardia iowensis (PPT Ni ) . The PPT is required to posttranslationally modify apoCARs to yield the functional holo-CAR enzymes (Akhtar et al., 2013;Finnigan et al., 2017;Horvat and Winkler, 2020). Whereas TPA was not converted in resting cells (RCs) of untransformed E. coli, the detection of 4-(hydroxymethyl) benzaldehyde (4-HMBAL) and 1,4-benzenedimethanol (1,4-BDM; 32.7 G 3.5% combined yields) by gas chromatography equipped  Table 1) (Bayer et al., 2017;Kunjapur et al., 2014;Kunjapur and Prather, 2015). However, biotransformation mixtures contained up to 75% unreacted TPA besides the over-reduced products after 24 h ( Figure S1A). Although a similar conversion of TPA was achieved with RCs of E. coli RARE ( Figure S1B), the utilization of E. coli BL21(DE3) Dlpp enhanced the bioreduction of TPA significantly ( Figure 2A). RCs of the engineered strain harboring pACYCDuet-1/ car Mm :ppt Ni were prepared and biotransformations were carried out as outlined below. The resulting suspension contained a mixture of TPA (31.1 G 5.9%), 4-CBAL (36.8 G 9.9%), 4-HMBAL (6.5 G 2.2%), and 1,4-BDM (13.7 G 5.7%) according to GC/FID ( Figure 2A); the highly reactive TAL could only be detected in traces. The nonessential lpp gene encodes one of the most abundant cellular proteins in terms of copy number (Li et al., 2014) and controls the (mechanical) properties of the inner and outer membrane (Asmar et al., 2017;Mathelié -Guinlet et al., 2020). Not only was its deletion suggested to affect the permeability of the cellular envelope for small molecules (Ni et al., 2007); it increased expression levels of CAR Mm according to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis ( Figure S2). This may be explained by the reallocation of cellular resources (Li et al., 2014) and might provide a general approach to improve heterologous protein production.
Subsequently, RCs of E. coli BL21(DE3) Dlpp as well as E. coli RARE were prepared either expressing only LuxAB or the luciferase together with CAR Mm /PPT Ni . E. coli RARE exhibits reduced aromatic aldehydereducing activity (Kunjapur et al., 2014) and has been employed by various groups to increase the persistence of aldehydes for both their production in vivo (Bayer et al., 2017;Horvat and Winkler, 2020;Kunjapur et al., 2016) and their efficient detection Ressmann et al., 2019).
Satisfyingly, the previously established HT assay conditions yielded bioluminescence in the presence of TPA-derived aldehydes in both E. coli strains expressing LuxAB . At 1 mM final concentration, the highest fold-increase in bioluminescence was observed in the presence of 4-CBAL and TAL, both elevating bioluminescence about 8-fold above background in RCs of E. coli BL21(DE3) Dlpp after 15 min, followed by 4-HMBAL (4-fold) ( Figure 2B). As expected, TPA did not increase bioluminescence in RCs only expressing the biosensor, but signals increased more than 4-fold when co-expressing LuxAB and CAR Mm /PPT Ni in the same cell ( Figure 2C). Similar results were obtained with RCs of E. coli RARE upon the addition of TPA ( Figure   (2) TPA can be reduced to the corresponding dialdehydes and monoaldehydes by CAR Mm (accessory PPT Ni not shown). These aldehydes are sensed by LuxAB, thereby emitting bioluminescence. Endogenous enzymes further reduce aldehydes to the corresponding primary alcohols. (3) The reactive TAL can be captured as aldoxime (not shown) and further converted to the diamine by reductive amination and basic work-up in a one-pot cascade, interconverting polymer precursors as future upcycling option after further optimization.

OPEN ACCESS
iScience 25, 104326, May 20, 2022 3 iScience Article established reduction of 2-phenylacetic acid to 2-phenylacetaldehyde by CAR Mm was included as positive control for the HT assay because the latter is accepted by LuxAB ( Figure 2B-2C). DMSO slightly increased background luminescence over time, which had also been shown for other cosolvents like ethanol and acetonitrile . Supporting the results of the HT assay ( Figure 2C), the activity of the CAR enzyme toward 4-CBAL and 4-HMBA could be confirmed by GC/FID analysis of extracts from biotransformations employing CAR Mm /PPT Ni (Figures S1C-S1D).
Motivated by the functional CAR/luciferase biosensor couple for the detection of TPA, the assay was tested with hydrolysate samples obtained after the enzymatic degradation of PET.

Assaying TPA in PET hydrolysis samples under HT conditions
For the preparation of PET hydrolysates, the codon-optimized genes of LCC, the engineered variant LCC-ICCG , and the polyester hydrolase-1 (PES-H1) (Zimmermann et al., 2019) were expressed from pET26b vectors in E. coli BL21(DE3) cultivated in auto-induction medium (AIM) supplemented with kanamycin and finally purified as described in this study.
The enzymatic degradation of amorphous PET film (Gf-PET, purchased from Goodfellow Ltd.) by LCC, LCC-ICCG, and PES-H1 was adapted from Tournier et al. as outlined below . Hydrolysates were processed as described in this study and analyzed by the CAR Mm /LuxAB biosensor system under HT conditions ( Figure 3) as well as calibrated high-performance liquid chromatography (HPLC; Table S2).
In the presence of 1 mM TPA, the bioluminescence increased about 4-fold and 5-fold in RCs of E. coli BL21(DE3) Dlpp and E. coli RARE, respectively, after 1 h. While the fold-increase in bioluminescence plateaued in E. coli BL21(DE3) Dlpp for 4 h ( Figure S4), it increased more than 17-fold in E. coli RARE cells during the same reaction time (Figure 3). This difference can be explained by the distinct metabolic backgrounds of the two strains as highlighted earlier. The knockout of several alcohol dehydrogenases and aldo-keto reductases in E. coli RARE increases the persistence of (aromatic) aldehydes in vivo, including TPA-derived aldehydes (Kunjapur et al., 2014). In contrast, the activity of these endogenous enzymes in E. coli BL21(DE3) Dlpp continuously reduces reactive aldehydes to the corresponding primary alcohols such as 1,4-BDM ( Figure S1), which is not a substrate for LuxAB ( Figure 2B).
Whereas bioluminescence signals were elevated >3-fold with PET hydrolysates obtained by the wild-type enzymes PES-H1 and LCC, bioluminescence increased >7-fold in LCC-ICCG samples after 4 h (Figure 3). This may be attributed to higher concentrations of potassium terephthalate salts in PET hydrolysates obtained by the LCC-ICCG variant compared to LCC, for example. Based on the fold-increases, TPA concentrations in the supernatants of the three hydrolysates were calculated and suggested 38.3 G 3.8 mM, 39.3 G 1.3 mM, and 95.5 G 10.7 mM for PES-H1, LCC, and LCC-ICCG, respectively. Similar TPA yields in the same concentration range (56 mM, 47 mM, and 111 mM, respectively) were determined by HPLC (Table S2). The retention times for benzoic acid and 2-phenylacetic acid and their corresponding aldehydes and primary alcohols as well as GC/FID-based quantification were reported previously . Relative response factors (RFFs) were used as mean values of independently prepared standard solutions (n R 3) analyzed by GC/FID.

OPEN ACCESS
Given that the biosensor system is operating in living cells, the cytotoxicity of both carboxylates and the corresponding aldehydes (Bayer et al., 2017Kunjapur and Prather, 2015), as well as the transient nature of bioluminescence signals (Fleiss and Sarkisyan, 2019) may interfere with the quantitative determination, allowing for marginal deviations from HPLC data. Nonetheless, the analysis of PET hydrolysates under HT conditions employing RCs of E. coli RARE yielded a reproducible fold-increase in bioluminescence based on the enzymatic transformation of TPA into the corresponding aldehydes and their detection by LuxAB, ultimately, reflecting TPA concentrations in PET hydrolysate samples semiquantitatively.

Transformation of TAL by a chemo-enzymatic cascade in one pot
The chemical synthesis of (aromatic) aldehydes can be troublesome because of the high reactivity of the carbonyl group (Ferguson, 1946;Kunjapur and Prather, 2015). A promising alternative to specifically synthesize aldehydes are the well-established enzymatic reductions of carboxylates by CARs (Bayer  (Figures 2A and S1). To the best of our knowledge, the CAR-catalyzed reduction of TPA is the first reported biocatalytic route forming TPA-derived aldehydes such as TAL, substituting hazardous chemical procedures (Snell and Weissberger, 1940). Depending on the purity and downstream application of plastic monomers from biocatalytic degradations, not all TPA is suitable for the resynthesis of virgin PET. Therefore, (bio)chemical transformation strategies for the re-use (i.e., upcycling) of plastic precursors is of interest (Tiso et al., 2021). Recently, Sadler and Wallace synthesized vanillin from hydrolyzed waste PET by combining TPA-transforming enzymes from Comamonas sp. to yield intermediate catechol that was converted to the product by the activities of a CAR and an engineered catechol O-methyltransferase in E. coli RARE (Kunjapur and Prather, 2019;Sadler and Wallace, 2021).
In the following proof-of-concept example, benzaldehyde and TAL were produced from benzoic acid and TPA, respectively, by CAR Mm /PPT Ni in E. coli BL21(DE3) or E. coli RARE RCs. The aldehydes were quenched in the presence of an excess of hydroxylamine hydrochloride (NH 2 OH $ HCl) to form the corresponding aldoximes. Subsequently, reductive amination was performed in one pot by the addition of zinc powder and acidification . After extraction under basic conditions, the expected primary aminesbenzylamine (BAM; 35.3 G 0.7%) and 1,4-bis-(aminomethyl) benzene (1,4-bis-AMB; 15.0 G 5.0%) -could be detected by GC/FID ( Figure 4); benzyl alcohol and 1,4-BDM, respectively, were the major byproducts. Structurally related diamines find applications in synthesis of polyurethanes and polyamides, for example (Wang et al., 2021). In addition, although not further investigated in this study, the formation of imines might contribute to the low yield and the poor recovery of material in reactions starting from TPA

DISCUSSION
The expanding number of new PET hydrolases from natural resources including metagenomes as well as protein engineering endeavors calls for tools for their rapid characterization (Wei et al., 2022;Wiltschi et al., 2020). Furthermore, the functional assessment of these enzymes depends -with very few exceptions (Pfaff  (Markel et al., 2020;Wei et al., 2020;Yi et al., 2021). To address this issue, this work coupled the activity of CAR Mm to reduce TPA in different E. coli strains to the corresponding aldehydes (4-CBAL, TAL, and 4-HMBAL) with the genetically encoded biosensor LuxAB from P. luminescens. The latter emits detectable bioluminescence in the presence of TPA-derived aldehydes (Figures 1, 2 and S3). As TPA is a building block of PET, the CAR/LuxAB couple was employed to detect terephthalates in hydrolysate samples obtained from PET degradation, catalyzed by the wild-type polyester hydrolases PES-H1 and LCC and the engineered variant LCC-ICCG. Not only was TPA reliably detected by reproducible fold-increase in bioluminescence values in independently carried out assay set-ups under HT conditions ( Figures 2C and  3); samples containing terephthalate from PET hydrolysis by PES-H1, LCC, and LCC-ICCG exhibited steady fold-increases over 4 h under HT assay conditions in E. coli RARE, which was in agreement with HPLC data. This sufficed to distinguish between wild-type enzymes and a variant with increased PET degradation activity and offers a semiquantitative screening tool for PET hydrolase libraries in the future.
Lastly, with the biocatalytic production of TAL from TPA, we accessed a highly reactive aldehyde intermediate that could be transformed into the corresponding primary diamine, for example, in aqueous reaction media ( Figure 4A). The chemo-enzymatic three-step cascade also yielded >30% BAM from benzoic acid via the aldehyde and aldoxime intermediates ( Figure 4B).
In conclusion, the presented work featured a complementary biosensor tool for the HT detection of TPA in living cells and suggested new routes for the bio-based interconversion of polymer building blocks, supporting efforts toward a circular plastic economy, the reduction of CO 2 emissions, and the stewardship of resources.

Limitations of the study
Although the utilization of E. coli BL21(DE3) Dlpp significantly increased the CAR-catalyzed conversion of TPA in vivo, it was not superior to the established E. coli RARE strain for the LuxAB-based detection of aldehydes over longer reaction times because of their different metabolic backgrounds. However, the reproducible detection of TPA by the CAR Mm /LuxAB-coupled biosensor under HT assay conditions in E. coli RARE enabled the semiquantitative assessment of terephthalate salts in the supernatants obtained from the biocatalytic degradation by various PET hydrolases. Even though calculated yields were in the same concentration range according to calibrated HPLC, discrepancies arise from operating the biosensor system in whole-cells of E. coli because of the cytotoxicity of TPA and the corresponding aldehydes, for example. Accordingly, the bioluminescence yielded by the LuxAB-catalyzed reaction is transient and influenced by the metabolic background, the viability and physiological state of cells including aeration; because LuxAB is a monooxygenase, the generation of bioluminescence depends on the aldehyde substrate and O 2 . In addition, expression levels of enzymes, intracellular cofactor availability, and the background luminescence in living cells can add to variations but are easily addressed by appropriate (negative) controls, and the normalization of bioluminescence signals as discussed in detail previously .
The interconversion of TPA into 1,4-bis-AMB through a three-step chemo-enzymatic cascade operating in one-pot only yielded only 15.0 G 5.0% of the diamine and could not be improved by employing E. coli RARE, for example. The poor recovery of material (<50%) can be explained by the low solubility of TPA in aqueous solutions and the volatility of reaction intermediates. Furthermore, the formation of imines from aldehyde and amine precursors in aqueous solutions has been reported (Godoy-Alcá ntar et al., 2005;Simion et al., 2001) and will be investigated as a contributing factor in the future. Nonetheless, the reductive amination could be achieved in an aqueous buffer system, which advances the original protocol  and puts it in the context of transforming PET-derived TPA.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:   . Chemo-enzymatic one-pot cascades Carboxylates are reduced by CAR Mm in RCs of E. coli BL21(DE3) to the corresponding aldehydes; PPT Ni is omitted for clarity. In the presence of NH 2 OH $ HCl, the oximes are formed (not shown), which are reduced to the primary amines (shades of blue) after the addition of Zn/HCl to the same reaction vessel. (A) The TAL intermediate yields the desired 1,4-bis-AMB, besides 1,4-BDM as the major byproduct. Recoveries were reduced due to low solubility of TPA in RCM containing 5% (n/n) DMSO as organic co-solvent, the volatility of reaction compounds, and the formation of yet to be identified byproducts such as imines (Godoy-Alcá ntar et al., 2005;Simion et al., 2001). (B) Benzoic acid in the presence of 5% (n/n) ethanol was reduced to benzaldehyde, yielding the desired BAM after reductive amination and benzyl alcohol as the sole byproduct. Experiments were performed in RCs (OD 600 z 10.0) co-expressing enzymes from pACYCDuet-1/car Mm :ppt Ni  . Sampling: (1) after the addition of NH 2 OH $ HCl (2.2 and 1.1 equiv for TPA and benzoic acid, respectively) and carboxylic acid and mixing; (2) after performing the reductive amination in one-pot. GC yields are presented as mean values + SD [mM] of biological replicates (n = 3). Performance was similar with RCs of E. coli RARE producing 27.2 G 6.6% BAM and 13.1 G 8.0% 1,4-bis-AMB (n = 2). Herrero Acero, E., Ribitsch, D., Steinkellner, G., Gruber, K., Greimel, K., Eiteljoerg, I., Trotscha, E., Wei, R., Zimmermann, W., Zinn, M., et al. (2011 Table S1 This paper (Thermo Scientificä) N/A lpp-up_R, primer for strain engineering, see Table S1 This paper (Thermo Scientificä) N/A lpp-down_F, primer for strain engineering, see Table S1 This paper (Thermo Scientificä) N/A lpp-down_R, primer for strain engineering, see Table S1 This paper (Thermo Scientificä) N/A pTarget_F, primer for strain engineering, see Table S1 This paper (Thermo Scientificä) N/A pTarget_R, primer for strain engineering, see Table S1 This paper (Thermo Scientificä) N/A Dlpp-gRNA_F, primer for strain engineering, see Table S1 This paper (Thermo Scientificä) N/A Dlpp-gRNA_R, primer for strain engineering, see Table S1 This paper (Thermo Scientificä) N/A For the assembly of the pTarget-Dlpp plasmid in this study, the templates pCas (#62225) and pTarget (#62226) were purchased from Addgene (Watertown, USA). Subsequently, pTarget-Dlpp and pCas were used to knock-out the lpp gene from the genome of E. coli BL21(DE3). The genes encoding the leaf-branch compost cutinase (LCC) and the LCC-ICCG variant  and the polyester hydrolase-1 (PES-H1) (Zimmermann et al., 2019) were codon-optimized for the expression in E. coli, synthesized, and cloned in frame with the C-terminal 6xHis tag present in pET26b by the BioCat GmbH (Heidelberg, Germany). Accession numbers of proteins are provided in the Key resources table.
There are restrictions to the availability of the previously constructed pCDFduo/luxAB, herein referred to as pLuxAB, and pACYCDuet-1/car Mm :ppt Ni plasmids  due to material transfer agreements (MTAs). Further information is available from the lead contact upon request.
Otherwise, this study did not generate new unique reagents.
Data and code availability d The genome of E. coli BL21(DE3) and associated metadata were retrieved from the National Center for Biotechnology Information (NCBI; GenBank: CP001509.3). The accession numbers of protein sequences are listed in the Key resources table.
d This paper does not report original code.
d Any additional information required to re-analyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS
E. coli BL21(DE3), E. coli BL21(DE3) Dlpp, E. coli DH5a, and E. coli RARE were propagated in 4-5 mL lysogeny broth (LB) medium (25 g L À1 ; Sigma-Aldrich, Buchs, Switzerland) in Infors HT Multitron incubator shakers (Bottmingen, Switzerland) at 37 C with shaking (150-180 rpm) for 12-16 h. If not stated otherwise, chemically competent E. coli cells were produced by using 0.1 M CaCl 2 and transformed with plasmid DNA (25-100 ng) by heat-shock at 42 C for 45 s as previously described . For the efficient transformation of E. coli RARE, plasmids were passed through E. coli DH5a . E. coli transformants harboring pLuxAB and pACYCDuet-1/car Mm :ppt Ni were propagated in LB medium supplemented with streptomycin (25 mg$mL -1 ) and chloramphenicol (34 mg$mL -1 ), respectively. Only half the concentration of antibiotics was used for the selection and subsequent propagation of strains harboring both plasmids.
For the selection and propagation on plates, LB containing 1.5% (u/n) agar (Carl Roth, Karlsruhe, Germany) and supplemented with antibiotics -if applicable -was used.

Strain engineering
The non-essential lpp gene encodes a cellular 'bulk' protein (Li et al., 2014), which controls the (mechanical) properties of the inner and outer membrane and the width of the periplasmic space (Asmar et al., 2017;Mathelié -Guinlet et al., 2020). The deletion of the lpp gene from the E. coli genome has been suggested to affect the permeability of the cellular envelope for small molecules (Ni et al., 2007) that might also influence the uptake of TPA and derivatives. Furthermore, expression levels of CAR Mm /PPT Ni were increased in E. coli BL21(DE3) Dlpp according to SDS-PAGE analysis ( Figure S2). This may be due to the re-allocation of cellular resources (Li et al., 2014).
E. coli BL21(DE3) Dlpp was constructed by using a previously developed two-plasmid-based CRISPR/Cas9 system . The two key plasmids, pCas (#62225) and pTarget (#62226), were purchased from Addgene (Watertown, USA). The pTarget-Dlpp plasmid was constructed by first engineering the flanking sequence of the lpp gene by the assembly of three DNA fragments using a sequence-and ll OPEN ACCESS iScience Article iScience Article