nature communications 8 Article https://doi.org/10.1038/s41467-024-46812-9 Synthetically-primed adaptation of Pseudomonas putida to anon-native substrate D-xylose Received: 20 July 2023 Accepted: 11 March 2024 Published online: 26 March 2024 _*>j Check for updates Pavel Dvořák©1 , 9 , Barbora Burýšková©1 , 9 , Barbora Popelářova©1 , 9 , Birgitta E. Ebert©2 , 9 , Tibor Botka®1 , Dalimil Bujdoš©3 , 4 , Alberto Sánchez-Pascuala®5 , Hannah Schöttler©6 , Heiko Haven©6 , Victor de Lorenzo©7 , Lars M. Blank®8 & Martin Benešík1 To broaden the substrate scope of microbial cell factories towards renewable substrates, rational genetic interventions are often combined with adaptive laboratory evolution (ALE). However, comprehensive studies enabling a holistic understanding of adaptation processes primed by rational metabolic engineering remain scarce. The industrial workhorse Pseudomonasputida was engineered to utilize the non-native sugar D-xylose, but its assimilation into the bacterial biochemical network via the exogenous xylose isomerase pathway remained unresolved. Here, we elucidate the xylose metabolism and establish a foundation for further engineering followed by ALE. First, native glycolysis is derepressed by deleting the local transcriptional regulator gene hexR. We then enhance the pentose phosphate pathway by implanting exogenous transketolase and transaldolase into two lag-shortened strains and allow ALE tofinetunethe rewired metabolism. Subsequent multilevel analysis and reverse engineering provide detailed insights into the parallel paths of bacterial adaptation to the non-native carbon source, highlighting the enhanced expression of transaldolase and xylose isomerase along with derepressed glycolysis as key events during the process. Efficient utilization of the most abundant lignocellulose-derived sugars (D-glucose, D-xylose, L-arabinose) and aromatic chemicals (e.g., pcoumarate, ferulate) is a key prerequisite for the economic viability of biotechnological processes that leverage natural or recombinant microorganisms as biocatalysts1 4 . D-Xylose is a predominant monomeric hemicellulose component, which forms 1 0 - 5 0 % of the mass of lignocellulosic residues5 . However, many biotechnologically relevant microorganisms lack xylose metabolism { Zymomonas tnobilis, Corynebacterium glutamicum, Saccharomyces cerevisiae), while others can utilize it but prefer glucose or other available organic substrates as their primary carbon and energy source (e.g., Escherichia coli. Bacillus subtilisf. The desirable phenotype can be established using metabolic department of Experimental Biology, Faculty of Science, Masaryk University, Kamenice 753/5, 62500 Brno, Czech Republic. Australian I nstitute for Bioengineering and Nanotechnology, The University of Queensland, Cnr College Rd & Cooper Rd, St Lucia, QLD QLD 4072, Australia.3 APC Microbiome I reland, University College Cork, College Rd, Cork T12 YT20, Ireland.4 School of Microbiology, University College Cork, College Rd, Cork T12 Y337, Ireland, department of Biochemistry and Synthetic Metabolism, Max Planck Institute for Terrestrial Microbiology, Karl-von-Frisch-StraBe 10, 35043 Marburg, Germany, institute of Inorganic and Analytical Chemistry, University of Munster, CorrensstraBe 48, 48149 Munster, Germany. 7 Systems and Synthetic Biology Program, Centra Nacionál de Biotecnología CNB-CSIC, Cantoblanco, Darwin 3, 28049 Madrid, Spain, institute of Applied Microbiology, RWTH Aachen University, Worringer Weg 1, 52074 Aachen, Germany. 9 These authors contributed equally: Pavel Dvořák, Barbora Burýšková, Barbora Popelářova, Birgitta E. Ebert. e-mail: pdvorak@sci.muni.cz Nature Communications | (2024)15:2666 1 Article https://doi.org/10.1038/s41467-024-46812-9 engineering tools7 . The rational approach to metabolic engineering employs our vast yet incomplete knowledge of the cellular machinery to introduce non-native functions into microbial cell factories. However, the genetic instalment of new functions can cause imbalances in the metabolic network and is therefore often completed with natural selection8 1 1 . By harnessing the natural selection of the fittest, adaptive laboratory evolution (ALE) on the substrate(s) of choice can generate mutants with optimized functioning of the rewired metabolism. Gram-negative bacteria from the genus Pseudotnonas are sought after in biotechnology, particularly for their robustness, fast growth in inexpensive media, and versatile metabolism1 2 1 4 . P. putida KT2440 is one of the most studied pseudomonads, especially in the biodegradation and bioremediation field. Its capacity to degrade lignin and stream the resulting aromatic monomers into biomass and diverse bioproducts makes it a promising candidate for lignocellulose valorization1 5 1 7 . Still, a major drawback is P. putida KT2440's inability to utilize pentose sugars such as D-xylose or L-arabinose. To overcome this bottleneck, P. putida has previously been endowed with different exogenous xylose utilization pathways1 8 2 3 . The isomerase route is the most promising for biotechnological xylose valorization4 , 2 2 . It is composed of xylose isomerase XylA and xylulokinase XylB and its product xylulose 5-phosphate (X5P) enters directly into the EDEMP cycle, a pathway configuration specific to P. putida (Fig. I)2 4 . In our former study, we found that P. putida EM42, the Xylose (XYL) ^ Glucose (GLU) MEDIUM T • XYL XLN GLU ••••^ GLN 2KG PERIPLASM Gcd • Gcd Pyruvate — • TCA cycle Fig. 11 Pseudomonasputida EM42 PD310 used as a template strain in this study. Schematic illustration of (a) upper sugar metabolism of P. putida EM42 PD310 and (b) pSEVA2213_xy£4B£ plasmid construct. The PD310 strain18 capable of growth on D-xylose and its co-utilization with D-glucose bears low-copy-number plasmid pSEVA2213 with constitutive P E M 7 promoter and a synthetic operon that encodes exogenous xylose isomerase pathway (XylA xylose isomerase and XylB xylulokinase) and xylose/H* symporter XylE from Escherichia coli BL21(DE3). The xylA and xy/f genes are preceded by synthetic ribosome binding sites (RBS), while xylB was left with its native RBS. Note that the elements in the plasmid scheme are not drawn to scale. The PD310 strain was also deprived of the gcd gene, which encodes periplasms glucose dehydrogenase (PP_1444) to prevent the transformation of xylose to the dead-end product xylonate (XLN). The exogenous pathway converts xylose to xylulose 5-phosphate (X5P), which enters the EDEMP cycle formed by the reactions of the pentose phosphate pathway (brown arrows), the Embden-MeyerhofParnas pathway (orange arrows), and the Entner-Doudoroff pathway (blue arrows). Abbreviations: GLN, gluconate; G6P, glucose 6-phosphate; G3P, glyceraldehyde 3phosphate; Km, kanamycin; 2KG, 2-ketogluconate; 6PG, 6-phosphogluconate; RK2, a broad-host-range origin of replication; TO, transcriptional terminator. genome-reduced derivative of the strain KT24402 5 , requires the XylE xylose/H+ symporter from E. coli and the deletion of the gcd gene encoding periplasmic membrane-bound glucose dehydrogenase (PP 1444) to fully utilize xylose1 8 . This is consistent with the findings of Meijnen et al.2 0 , 2 6 . in P. putida S12 and Elmore and co-workers1 9 in P. putida KT2440. It was also demonstrated that recombinant P. putida can fully co-utilize xylose with glucose and other lignocellulosic sugars and aromatics4 '1 8 '1 9 '2 7 , 2 8 . P. putida, empowered with an efficient xylose metabolism based on the exogenous isomerase route represents a very attractive biocatalyst for the production of polyhydroxyalkanoates, ris,ris-muconate and other dicarboxylic acids, biosurfactants, some amino acids (e.g., threonine), and potentially many other compounds that can be synthesized from xylose alone or from complex lignocellulosic hydrolysates4 , 2 2 , 2 9 . However, the growth of our recombinant P. putida on xylose was still slow1 8 . Moreover, the xylose assimilation into the bacterium's biochemical network extended with the isomerase route was unresolved. This complicated further tailoring of the utilization and valorization of this sugar by P. putida. In the present study, we use an evolutionary approach primed by rational metabolic engineering and introduction of synthetic genetic constructs to (i) explore how the metabolism of P. putida adapts to the perturbations caused by carbon flux from a non-native sugar substrate, and (ii) accelerate xylose utilization in our previously reported P. putida EM42 PD310 strain (Fig. 1, Table 1). Carbon flux analyses and enzyme activity assays help us uncover xylose metabolism in P. putida, remove the repression of part of the EDEMP cycle, and identify potentially problematic nodes in the pentose phosphate pathway (PPP). ALE of strains with or without 6-phosphogluconate dehydrogenase and with implanted exogenous PPP genes doubles the growth rate and reduces the lag phase up to 5-fold. Genomic, proteomic, and enzyme activity analyses of mutant strains together with reverse engineering reveal that the bacterium's genomic and metabolic plasticity enabled it to find alternative solutions that led to equally improved phenotypes of selected mutants. This study elucidates xylose metabolism in P. putida and contributes to a better understanding of the adaptation of bacterial metabolism to a non-native substrate. Results and discussion Elucidation of xylose metabolism in recombinant P. putida Our previously constructed strain P. putida PD3101 8 (Fig. 1) grew fourfold slower on xylose (u = 0.11 h~') than o n glucose and exhibited a 5-times longer lag phase o n the pentose sugar (-10 h) (Table 2, Supplementary Fig. l a , b). Studies from other groups demonstrated faster growth (u -0.3 h- 1 ) of engineered and evolved P. putida gcd xylABE* on xylose1 9 , 2 0 , suggesting that P. putida has the capacity to utilize xylose more efficiently than observed for PD310. The authors proposed complex metabolic and regulatory re-arrangements in an evolved recombinant strain of P. putida S12 (including altered glucose metabolism regulation) and emphasized the role of the upregulated XylE transporter in promoting the growth of evolved P. putida KT24401 9 , 2 0 . We expected sufficient expression of the xylE gene in PD310, as the gene is controlled by the strong constitutive Pem7 promoter and a strong synthetic ribosome binding site (RBS)3 0 . Expression of heterologous enzymes and transporters can place a burden on bacterial metabolism3 1 . However we hypothesized that this was not the major limitation of our strain's capacity to utilize xylose because it grew well on glucose (u = 0.46 h"1 ; Table 2). Therefore, we moved our attention to potential bottlenecks in the central carbon metabolism of P. putida and mapped the catabolism o f xylose in PD310 using 1 3 C-based metabolic flux analysis (MFA)3 2 . Distribution of carbon fluxes was previously determined for P. putida KT2440 or its derivatives cultured o n native substrates glucose, benzoate, glycerol, gluconate, or s u c c i n a t e 2 4 , 3 3 - 3 6 but never for xylose-assimilation via a heterologous isomerase pathway. The M F A performed with 1,2-1 3 C Nature Communications | (2024)15:2666 2 Article https://doi.org/10.1038/s41467-024-46812-9 Table 11 Pseudomonas putida strains used in this study Strain Characteristics Reference EM42 Genome-reduced derivative of strain P. putida KT2440: Aprophagesl, 2, 3, 4 ATn7 AendA7 AendA2 AhsdRMS Aflagellum ATn4652 25 EM42 Aged Strain EM42 with deletion of gcd gene (PPJ444) encoding periplasmic glucose dehydrogenase 18 EM42 Aged pSEVA2213_xy/AßE EM42 Aged freshly transformed with pSEVA2213_xyM6E plasmid bearing synthetic xytABE operon encoding XylA xylose isomerase, XylB xylulokinase, XylE xylose-H* symporter from E. co/j (EcoRI/HJndl II), see plasmid characteristics in Table S2 in Supplementary Information This study PD310 EM42 Aged with plasmid pSEVA2213_xy/A6E, -118 kbp multiplication (PP_2114 - PP_2219) in chromosome 18 PD505 EM42 Aged with additional deletion of gene edd (PPJ010) encoding phosphogluconate dehydratase, with pSEVA2213_xyM6E plasmid This study PD506 EM42 Aged with additional deletion of gene gnd (PP 4043) encoding 6-phosphogluconate dehydrogenase, with pSEVA2213_xyM6E plasmid This study PD507 EM42 Aged with additional deletions of genes pgj-l (PPJ808) and pgj-ll (PP_4701) encoding glucose 6-phosphate isomerase, with pSEVA2213_xy/A6E plasmid This study PD580 EM42 Aged with additional deletion of gene hexR (PPJ021) encoding DNA-binding transcriptional regulator This study PD580 rpoD* PD580 with Ser552Pro mutation in the rpoD gene encoding the RNA polymerase sigma factor a 7 0 This study PD584 EM42 Aged AhexR with pSEVA2213_xyM6E, -118 kbp multiplication in chromosome This study PD584 L3 Derivative of PD584 obtained after adaptive laboratory evolution (ALE) on xylose, -118 kbp multiplication in chromosome This study PD584 tt L3 Derivative of PD584 obtained after ALE on xylose, -118 kbp multiplication in chromosome This study PD689 EM42 Aged AhexR Agnd with pSEVA2213_x/M6E This study PD689 rpoD* PD689 with Ser552Pro mutation in the rpoD gene This study PD689 tt L1 Derivative of PD689 with chromosomally integrated expression cassette (PEM7-ta/6-t/ctA-SmR bearing genes for transaldolase and transketolase from E. co/j) obtained after ALE on xylose This study PD855 Reverse engineered strain: PD580 with pSEVA438_ta/ and mutated pSEVA2213_xyM6E plasmid isolated from PD584 L3 This study PEM7 constitutive promoter EM7, Sm, streptomycin. Table 2 | Growth parameters of Pseudomonas putida EM42 mutant strains in batch cultures with D-glucose or D-xylose as a sole carbon source Strain and u max (h1 )1 Yx/s (Qcdw q s (mmols Lag phase (h)' carbon 9s V 9cdw_ 1 Ivy source EM42Agcd 0.09 ±0.00 n.d. n.d. 41.94 ±0.72 xylABE* XYL PD310GLU 0.46 ±0.01 0.38 ±0.02 6.33 ± 0.42 1.40 ±0.14 PD310 XYL 0.11 ±0.00 0.31 ±0.03 1.98 ±0.32 10.04 ±0.53 PD506 XYL 0.07 ±0.00 n.d. n.d. 31.72 ±0.64 PD584 XYL 0.13 ±0.00 0.35 ±0.01 2.22 ±0.16 4.60 ±0.44 PD689 XYL 0.10 ±0.00 0.31 ± 0.02 1.14±0.06 18.25 ±1.28 PD584 L3 XYL 0.21 ±0.01 0.41 ± 0.07 3.07 ±0.46 3.40 ±0.60 PD689 tt 0.21 ±0.00 0.37 ±0.02 3.86±0.10 3.70 ± 0.35 L1 XYL PD855 0.20 ±0.00 0.35 ±0.03 3.46 ± 0.37 4.94 ±0.39 GLU D-glucose, XYL D-xylose, n.d. not determined. 'The maximal specific growth rate u max and growth lag were calculated from growth data obtained from cultures in 48-well plates using the deODorizer program80 . Presented values are means from six biological replicates ± standard deviation. 2 The biomass yield (Yx/s) and specific carbon uptake rate (qs) were calculated from data obtained from cultures in Erlenmeyer flasks. Shown values are means from three biological replicates ± standard deviation. Source data are provided as a Source Data file. D-xylose identified a partially cyclic upper xylose metabolism in PD310 (Fig. 2a). The majority (89%) of the carbon that initially entered the non-oxidative branch of the pentose phosphatepathway (PPP) via xylulose 5-phosphate (X5P) was converted into fructose 6-phosphate (F6P) and further to 6-phosphogluconate (6PG) through the reactions of glucose 6-phosphate isomerase (Pgi-I and Pgi-ll), glucose 6-phosphate 1-dehydrogenase (ZwfA, ZwfB, ZwfC), and 6-phosphogluconolactonase (Pgl). This is in contrast to the flux distribution in wild-type P. putida KT2440 grown on glucose, which is preferentially utilized via the periplasmic oxidative route and directly funneled into the EntnerDoudoroff (ED) pathway2 4 , 3 3 , 3 5 . At the 6PG node, the flux from xylose branches. Over 50% of the carbon enters the ED pathway while more than one-third is cycled back into the non-oxidative PPP via the 6-phosphogluconate dehydrogenase (Gnd) reaction to replenish the ribulose 5-phosphate (Ru5P) and ribose 5-phosphate (R5P) pools (Fig. 2a). This partial carbon cycling via Gnd was previously suggested also by Meijnen et al., for engineered, xylose-adapted P. putida S12 based on transcriptome data2 6 . Here, we hypothesize that the cycling compensates for the relatively weak flux through the ribulose-5phosphate 3-epimerase (Rpe) reaction. The bifurcation observed in our study reduced the ED pathway flux in xylose-grown PD310 (Fig. 2a) compared to wild-type P. putida cultured on g l u c o s e 2 4 3 3 3 5 . In line with Meijnen et al.2 6 , we determined an operational glyoxylate shunt (isocitrate lyase AceA and malate synthase GlcB) in PD310 (Fig. 2a). Glyoxylate shunt activity was reported in P. putida KT2440 grown on glycerol3 4 , benzoate3 7 , or a mixture of glucose and succinate3 8 but the shunt is typically inactive during growth on glucose as the sole carbon source2 4 , 3 3 , 3 5 , 3 9 . Meijnen et al. attributed glyoxylate shunt activation to the level of reducing cofactors, which was increased compared to cells cultured on glucose2 6 . Given the high fluxes measured for the four central dehydrogenases - Zwf, Gnd, pyruvate dehydrogenase complex Pdh, and malate dehydrogenase M d h (the middle two enzymes also have decarboxylating activity) - we hypothesize that the same holds true for strain PD310 (Fig. 2a). Overproduction of reducing cofactors in these reactions would also explain the negligible flux through the part of EMP pathway with fructose 1,6bisphosphatase (Fbp), fructose 1,6-bisphosphate aldolase (Fba), and triose phosphate isomerase (TpiA) (Fig. 2a). On glucose or glycerol, gluconeogenic operation of this section of the EDEMP cycle partially recycles triose phosphates back into hexose phosphates and secures the resistance of P. putida to oxidative stress through a supply of Nature Communications | (2024)15:2666 3 Article https://doi.org/10.1038/s41467-024-46812-9 Xylose (XYL) Glucose (GLU) C02 NADPH Embden-Meyerhof-Parnas pathway Pentose phosphate pathway Entner-Doudoroff pathway TCA cycle Anaplerotic reactions Xylose (XYL) Glucose (GLU) XYL ><••••• XLN GLU •><• GLL ••• GLN ••—• 2KG C 0 2 NADPH NADPH (generated by Zwf)2 4 , 3 3 3 5 , 4 0 . However, such gluconeogenic operation of Fbp, Fba, and TpiA would be redundant to PPP activity in xylose-grown PD310. MFA confirmed that malic enzyme MaeB activity significantly contributed to the pyruvate pool (Fig. 2a) similar to previous analyses on glucose2 4 , 3 3 , 3 5 . However, xylose metabolism differs in its low pyruvate carboxylase Pyc activity. To compute metabolic fluxes supporting optimal growth on xylose, we conducted flux balance analysis (FBA)4 1 using growth rate as objective function (Fig. 2b). Growth-optimised fluxes predicted by FBA showed several differences to the M F A data, including an almost evenly distributed flux from X5P into reactions of Rpe, Tkt, and Tal, and zero flux through Gnd, the glyoxylate shunt (AceA, GlcB), and the malic enzyme MaeB reaction in FBA. Enhanced fluxes from X5P to Ru5P Nature Communications | (2024)15:2666 4 Article https://doi.org/10.1038/s41467-024-46812-9 Fig. 21 Distribution of carbonfluxesin engineered strain Pseudomonas putida EM42 PD310 grown on xylose. Distribution of carbon fluxes in P. putida EM42 PD310 grown on xylose as determined by13 C metabolic flux analysis (MFA, a) and flux balance analysis (b). Fluxes are given as a molar percentage of the mean specific xylose uptake rate qs = 1.45 mmol gcDw1 h- 1 , which was set to 100%. Arrow thickness roughly corresponds with the given flux value. Flux values calculated in MFA (a) represent the mean ± standard deviation from two biological replicates (n = 2). Abbreviations (enzymes): AceA isocitrate lyase, Acn aconitate hydratase, Eda 2keto-3-deoxy-6-phosphogluconate aldolase, Edd, 6-phosphogluconate dehydratase, Eno phosphopyruvate hydratase, Fba fructose-l,6-bisphosphate aldolase, Fbpfructose-l,6-bisphosphatase,FumC fumarate hydratase. Gap glyceraldehyde-3phosphate dehydrogenase, Gcd glucose dehydrogenase, Gnd 6-phosphogluconate dehydrogenase, GlcB malate synthase, GltA citrate synthase, led isocitrate dehydrogenase, 2Kgd 2-ketoglutarate dehydrogenase, MaeB malic enzyme, Mdh malate dehydrogenase, Pdh pyruvate dehydrogenase, Pgi glucose-6-phosphate isomerase, Pgk phosphoglycerate kinase, Pgl 6-phosphogluconolactonase, Pgm and higher TCA cycle activity in the FBA simulation probably fully replaced the carbon cycling via Gnd observed in vivo. FBA also computed a reduced formation of hexoses F6P and G6P (by -33%) and, consequently, a lower flux through glucose 6-phosphate dehydrogenase Zwf. Importantly, FBA computed approximately 41% higher growth rate (u = 0.12 h_ 1 ) when constrained with the specific xylose uptake rate determined during the1 3 C labeling experiment (qs = 1.45 mmol gcDw1 h- 1 ) than experimentally observed with PD310 cells (u = 0 . 0 8 h _ 1 ± 0 . 0 2 ; note that the growth rate and xylose uptake rate in the M F A experiment differ from the values reported in Table 2 due to different experimental conditions, while the biomass yield on carbon was the same in both setups). In addition, the comparison of the sum of fluxes through oxidoreductase reactions that generate the reducing equivalents N A D P H and N A D H revealed 1.6-fold (NADPH) and 1.5-fold (NADH) higher values in the M F A model (Supplementary Data 1). The net rate of C 0 2 production determined in M F A was also higher (1.6-fold) than the net rate calculated by FBA (Supplementary Data 1). As mentioned above, the surplus of NAD(P) H and loss of carbon in the f o r m of C 0 2 is caused by high fluxes through Zwf and G n d reactions in PD310. The high flux through decarboxylating enzymes such as G n d also contributes to the lower carbon efficiency of PD310 strain grown on xylose compared to glucose (see YX /s biomass yields in Table 2)4 2 . FBA with the genomescale metabolic model has some limitations; it cannot model enzyme capacity, burden of protein synthesis, and some other cellular constraints. Nevertheless, a comparison of its results to the MFA showed that the flux distribution of PD310 grown on xylose was wasteful. Next, we verified the importance of EDEMP cycle enzymes that were identified by the flux analyses as key for the growth of PD310 on xylose. Genes encoding Pgi-I (PPJ808), Pgi-ll (PP_4701), Gnd (pp 4043), and Edd (PPJ010) were knocked out in P. putida EM42 Aged and the growth of the resulting mutants with inserted pSEVA2213_xylABE plasmid (named PD505, PD506, PD507, Table 1) was tested on solid and in liquid medium (Fig. 3). These experiments demonstrated that none of the deletions was detrimental for growth on citrate, representing a gluconeogenic growth regime (Fig. 3b) and only the edd deletion disabled growth on glucose. In contrast, growth on xylose was affected by all three deletions. The Aged Apgi-\ Apgi-\\ (PD507) and Aged Aedd (PD505) mutants did not show any growth in a liquid medium within three days, confirming the essentiality of Pgi and Edd for xylose catabolism. Interestingly, the Aged Agnd mutant (PD506) grew on xylose but with a reduced growth rate and substantially prolonged lag phase when compared to PD310 (Table 2, Fig. 3c). This result showed that the carbon flux into the non-oxidative branch of PPP through G n d is not essential for xylose utilization by P. putida, in line with the FBA data. phosphoglycerate mutase, Ppc phosphoenolpyruvate carboxylase, Pyc pyruvate carboxylase, Pyk pyruvate kinase, Rpe ribulose-5-phosphate 3-epimerase, RpiA ribose-5-phosphate isomerase, Sdh succinate dehydrogenase, Tal transaldolase, Tkt transketolase, Tpi triosephosphate isomerase, Zwf glucose-6-phosphate dehydrogenase. Abbreviations (metabolites): DHPA dihydroxyacetone phosphate, E4P erythrose 4-phosphate, FBP fructose 1,6-bisphosphate, F6P fructose 6-phosphate, GLL glucono-8-lactone, GLN gluconate, G3P glyceraldehyde 3-phosphate, G6P glucose 6-phosphate, KDPG 2-keto-3-deoxy-6-phosphogluconate, 2KG 2ketogluconate, 2KG-6P 2-ketogluconate 6-phosphate, 2KT a-ketoglutarate, MAL malate, NADH reduced nicotinamide adenine dinucleotide, NADPH reduced nicotinamide adenine dinucleotide phosphate, OAA oxaloacetate, PEP phosphoenolpyruvate, 3PG 3-phosphoglycerate, 6PG 6-phosphogluconate, R5P ribose 5phosphate, Ru5P ribulose 5-phosphate, S7P sedoheptulose 7-phosphate, XLN xylonate, X5P xylulose 5-phosphate. Note that F6P and G3P are products of both Tal and Tkt reactions. Source data are provided as a Source Data file. To deepen our insight into the operation of the EDEMP cycle in PD310 grown on xylose, we screened the activities of eight enzymes, which contribute to the cycle together with the activities of the exogenous XylA and XylB (Fig. 3d, Supplementary Method 1). The specific activity of both XylA and XylB was significantly higher (p<0.05) on xylose than on glucose substrate. This difference may reflect growth condition dependent epigenetic regulation of the respective genes4 3 . Activities of XylB, G n d , Zwf, Pgi, and Edd-Eda in cells cultivated on either sugar were higher than the activities in cells grown in LB medium. In the case of Zwf and Edd-Eda, the difference can be attributed to the (de)repression of these genes in cells consuming glucose or xylose. Their genes are placed in operons controlled by the HexR transcriptional repressor, which binds 2-keto-3-deoxy-6-phosphogluconate (KDPG), an intermediate of the ED pathway formed during catabolism of glucose and xylose (Fig. 2)4 4 . This interaction causes HexR dissociation, which leads to transcriptional activation of these operons. The de-repression seemed to be more efficient on glucose than on the non-native substrate xylose (Fig. 3d). The low activities of Zwf and the ED pathway enzymes in cells grown on xylose (compared to glucose) could pose a bottleneck for pentose catabolism in this part of the EDEMP cycle. In contrast, the activities of transketolase, transaldolase, and Tpi were comparable across the three tested conditions, which indicates that the levels of these enzymes in P. putida cells are relatively constant irrespective of the substrate. The metabolic flux analyses, gene deletion experiments, and enzyme activity measurements elucidated carbon distribution in the EDEMP cycle of xylose-grown PD310. Interestingly, the distribution of fluxes determined by M F A - especially in the PPP segment of the EDEMP cycle - resembled the situation in glucose-grown P. putida under oxidative stress4 0 . Given that the complex biochemical network of P. putida is evolved primarily towards the utilization of organic acids, aromatic compounds, or glucose4 5 , it is plausible that the introduction of an exogenous xylose isomerase pathway led to metabolic imbalances and, consequently, to slow growth differing significantly from rates attainable on native substrates. Synthetically-primed enhancement of xylose metabolism in P. putida Del Castillo et al.4 6 and Bentley et al.4 7 previously showed that the deletion of hexR gene de-represses zwf-pgl-eda and edd-glk operons and improves growth on glucose (Fig. 4a). Down-regulation of hexR was identified also in the evolved P. putida S12 and linked to its boosted xylose-utilization phenotype2 6 . Therefore, we argued that the removal of the repressor could have a positive effect on xylose metabolism in PD310. The hexR gene (PP_1021) was deleted in P. putida EM42 Aged and the resulting strain P. putida EM42 Aged AhexR pSEVA2213_xylABE, named here PD584, showed improved growth compared to PD310 (Table 2, Fig. 4 d , Supplementary Fig. lc). Notably, the lag phase on Nature Communications | (2024)15:2666 5 Article https://doi.org/10.1038/s41467-024-46812-9 Citrate Glucose Xylose Ctrl Ctrl + - — 0 20 40 60 Time (h) 0.08-1 X y l A „ 0.201 X y l B 0.061 T k t 0.15-1 T a l 0.06 "1 G n d ill ll. Ill ill ll.0.00 O . O O - l - ^ - T r - ^ - T - ^ - i 0 . 0 0 - I - ^ - t - ^ - t - ^ - i 0 . 0 0 4 - ^ - t - ^ - t - ^ - i 0.00 I ^ i ^ i ^ i 0.15 GLU XYL LB GLU XYL LB GLU XYL LB GLU XYL LB 0.20-. ^ Z w f 0 25-, 3.0-1 0.15-1 ^ d - E d a tlllx ElLU llllllllii GLU XYL LB GLU XYL LB GLU XYL LB GLU XYL LB GLU XYL LB Fig. 3 | Growth of three Pseudomonas putida deletion mutants on selected carbon sources and activities of XylA, XylB, and selected EDEMP cycle enzymes in P. putida PD310 cells, a Scheme of the EDEMP cycle with the incorporated xylose isomerase pathway (abbreviations and color coding are the same as in Figs. 1,2) and highlighted reactions that were eliminated by knocking out the respective genes (red crosses), b Growth of PD505 (Aged Aedd, green arrow), PD506 (Aged Agnd, brown arrow), and PD507 (Aged ApgH ApgHI, red arrow) mutants on solid M9 agar medium with 2g L1 citrate, D-glucose, or D-xylose used as the sole carbon and energy source. Ctrl- stands for negative control (P. putida EM42 Aged with empty pSEVA2213) and Ctrl+ stands for positive control (PD310). Cells pre-cultured in LB medium were washed with M9 medium and 10 pL of cell suspension of OD 2.0 was dropped on the agar and incubated for 24 h (agar with glucose or citrate) or 48 h (agar with xylose) at 30 °C. c Growth of the three mutants and control PD310 in M9 minimal medium with 2g L1 D-xylose in a 48-well microplate. Data points are shown as mean ± standard deviation from three (n = 3) biological replicates, d Specific activities of 10 selected enzymes (activity of Eda and Edd was measured in a combined assay) were determined in cell-free extracts (CFE) prepared from PD310 cells cultured till mid-exponential phase in rich LB medium (LB), in M9 medium with 2 g L- 1 glucose (GLU), or in M9 medium with 2 g L- 1 xylose (XYL) as detailed in Supplementary Method 1 in Supplementary Information. Data are shown as mean ± standard deviation from three (n = 3) biological replicates. Asterisks (**) denote statistically significant difference between two means at p < 0.01 calculated using two-tailed Student f test (p values = 4.07 x 10 2 for Zwf and 3.73 x 104 for Edd-Eda). Source data are provided as a Source Data file. xylose was reduced by more than 2-fold, from 10.0 to 4.6 h. The significantly increased activities of Zwf and the ED pathway in PD584 on xylose (Fig. 4b) underpin that this improvement can be attributed to the de-repression of native glycolysis enzymes. The reduced lag phase was also observed for PD584 grown on hexose substrates glucose and fructose (Supplementary Fig. 2). The high growth rate of PD584 on fructose (umax = 0.30 ± 0 . 0 0 IT1 ), which is similarly to xylose metabolized through Pgi, Zwf, Pgl, and the ED pathway4 8 , indicated that these reactions are not limiting xylose catabolism. However, the growth rate of PD584 on xylose increased only modestly (by -18%) compared to PD310 (Table 2) and we thus sought additional targets to eliminate further bottlenecks in the metabolism. The most evident differences between the distribution of fluxes in the EDEMP cycle in FBA and M F A were within the PPP. During growth on native sugar substrates such as glucose or fructose, the role of PPP in P. putida is rather complementary and predominantly anabolic4 0 . Periplasmic glucose oxidation is the preferred route for glucose uptake and therefore only a smaller fraction of carbon (-20-50%) flows through Zwf and Pgl and even much less (-1-10%) is directed to the non-oxidative branch of the PPP to provide metabolic precursors for nucleotides and some amino acids3 3 , 3 5 , 4 9 . This changes during the growth of recombinant P. putida on xylose for which PPP is the metabolic entry point. The situation is reminiscent of the exposure of P. putida to oxidative stress in which case activities of Zwf and Gnd Nature Communications | (2024)15:2666 6 Article https://doi.org/10.1038/s41467-024-46812-9 Zwf ftexR eda J n . E M 4 2 Aged PD310 Aged xy//4B£+ PD584 Aged AftexR xylABE+ PD506 Aged Agnd xylABE+ PD584 L3 PD689 Aged Agnd AftexR xylABE + PD689 ttL1 fltM to/S + 0.15 0.125 0.100 =i 0.075co 0.050 & 0.025 0.00 0.000 584 584 310 GLU XYL XYL 584 584 310 GLU XYL XYL • PD310 •PD584 #PD584L3 OPD855 PD689 • PD689ttL1 < SZ S O 0.2 Fig. 4 | Synthetically-primed adaptation of Pseudomonas putida to D-xylose. a Genetic organization of relevant genes in operons regulated by HexR transcriptional regulator. The elements in this scheme are not drawn to scale. Abbreviations used are the same as in Fig. 1. b Specific activity of Zwf and Edd-Eda measured in cell-free extracts from PD310 or PD584 cells grown in M9 medium with 2g L"1 glucose (GLU), or in M9 medium with 2g L- 1 xylose (XYL). Data are shown as mean ± standard deviation from six (n = 6) biological replicates. Asterisks (**) denote statistically significant difference between two means at p < 0.01 calculated using two-tailed Student f test (p values = 2.25 x 10 5 for Zwf and 6.81 x 10"6 for EddEda). c Pedigree of P. putida mutant strains used in this study. Abbreviations are the same as in the previous figures. The graphical scheme shows genes deleted and introduced rationally in these strains and highlights two lineages of mutants with (PD310, PD584, PD584 L3) and without (PD506, PD689, PD689 tt LI) gnd gene. d Growth of PD310, PD506, PD584, PD584 L3, PD689, PD689 tt LI, and reverse engineered PD855 in M9 medium with 2g L 1 D-xylose in a 48-well microplate. Data are shown as mean from six (n = 6) biological replicates. Error bars are omitted for clarity. Source data are provided as a Source Data file. increase multiple times to generate reducing equivalents for the elimination of reactive oxygen species (ROS)4 0 . This adaptation has no significant effect on the growth rate and indicates that the bacterium has the capacity to adjust its PPP in favor of non-native pentose metabolism. However, during the growth on xylose, all carbon enters the PPP at the point of X5P, not G6P, as is the case during P. putida's physiological response to ROS. Hence, the suboptimal activity of the non-oxidative branch of PPP might still be limiting xylose metabolism. Elmore and co-workers (2020) accelerated the growth of KT2440 xylABF on xylose by enhancing it with additional transketolase (tktA) and transaldolase (talB) genes from E. coll, which grows well on pentoses1 9 . A similar approach has been successfully employed for other bacteria, e.g., Zymomonas mobllls'0 . Guided by these studies and our flux analysis results, we decided to modulate the PPP in strain PD584 to further improve xylose utilization. To mimic the FBA scenario with zero flux through the Gnd reaction and to test the carbon-saving potential of this setup, we prepared a P. putida strain designated PD689 with deletions of gcd, hexR, and gnd and harboring the pSEVA2213_xylABE plasmid (Table 1, Fig. 4c, Supplementary Fig. 3). PD689 demonstrated a lower growth rate and 4-fold longer lag phase compared to the PD584 reference (Table 2, Fig. 4d), which confirmed that the null mutation of gnd decelerates growth but is not fatal for xylose utilization by P. putida (Fig. 3). We then integrated an expression cassette bearing either talB-tktA or talB-tktA-rpe-rplA synthetic operon assembled from F. coll genes into the genome of PD584 and PD698. We presumed that the activity of Rpe epimerase and RpiA isomerase (Fig. 2), whose genes were included in the second variant of the operon, could provide a higher carbon pull, support the conversion of X5P to R5P, and thus replenish metabolites depending on the Gnd reaction. The expression cassettes were inserted randomly into the host's chromosome using mini-Tn5 delivery plasmid p B A M D l - 4 5 1 . This allowed for a chromosome position effect and selection of transformants with optimal gene expression. The cassettes were complemented with the constitutive P E M 7 promoter ensuring transcription at various integration sites2 8 . Following the transformation with p B A M D l - 4 constructs and a 4-5 day selection period, strains PD584 and PD689 were subjected to ALE on xylose for two weeks (Fig. 4c, Fig. 5, Methods)8 , 9 1 9 2 2 , 5 2 . From an array of candidates isolated during ALE (Supplementary Fig. 4a, b), we selected three clones that demonstrated the fastest growth on xylose in shake flasks and reached O D 6 0 o > 3.5 within 24 h. The three clones, designated PD584 L3, PD584 tt L3, and PD689 tt LI, were isolates from the end of the evolutionary experiment ( - 6 0 - 7 0 generations). Interestingly, mutant PD584 L3 was isolated from the control culture (PD584 transformed with pure water instead of any of the p B A M D l - 4 constructs), and strains PD584 tt L3 and PD689 tt LI from cultures of transformants with the inserted p B A M D l - 4 talB-tktA construct. No isolates with the integrated talB-tktA-rpe-rplA operon outperformed the other strains in shake flasks (Supplementary Fig. 4c). Nature Communications | (2024)15:2666 7 Article https://doi.org/10.1038/s41467-024-46812-9 PD584 Aged AftexR xylABE+ PD689 Aged AftexR Agnd xylABE+ Control o ACD "* < 5 6 8 10 12 Time (days) 20 0 6 8 10 12 14 16 18 20 Time (days) PD584 PD689 -f 1 1 1~ 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 Time (days) Time (days) L.T500 * „ f * K K ^ K ™ P D 5 8 4 T^-'fflfcMwylHMWEp- -y P D 6 8 9 6 8 10 12 14 16 18 20 0 Time (days) 6 8 10 12 14 16 18 20 Time (days) Fig. 5 | Adaptive laboratory evolution (ALE) on xylose of Pseudomonas putida PD584 and PD689 strains with and without integrated synthetic operons bearing pentose phosphate pathway (PPP) genes from Escherichia coli. Cells were cultured in 20 mL of M9 medium with 5g L 1 D-xylose and kanamycin and passaged in time intervals indicated in the graphs (upper two graphs depict ALE of PD584 and PD689 controls without implanted PPP genes and the lower four graphs depict ALE of PD584 and PD689 transformants after insertion of pBAMDl-4 plasmid constructs with PPP genes). The initial period where PD584 and PD689 transformants were cultured in M9 medium with xylose and two antibiotics (kanamycin, streptomycin) to select for the integration of tktA-talB and tktA-talB-rpe-rpiA cassettes is indicated by pink shading. ALE of cultures in M9 medium with xylose and a single antibiotic kanamycin followed with regular culture transfers. Days in which samples of the cultures were withdrawn to prepare glycerol stocks and select individual clones are highlighted by orange shading. The first withdrawal occurred when the turbidity of a given culture (OD6oo) first reached the value of > 3.5 within 24 h of growth (the only exception was the control culture of PD689 which did not pass this threshold during the evolutionary experiment). The elements in the operon schemes are not drawn to scale. EM7 constitutive promoter, T500 transcriptional terminator. The talB, tktA, rpe, and rpiA denote genes that encode f. coli transaldolase B, transketolase A, ribulose-phosphate 3-epimerase, and ribose-5phosphate isomerase, respectively. Source data are provided as a Source Data file. It is plausible that the E. coli rpe and rpiA genes were not needed for the improved xylose utilization in P. putida. The ALE and subsequent screening experiments showed that the evolution of the PD584 strain can provide variants with substantially accelerated growth on xylose even without supplementation of any additional exogenous genes. That is particularly intriguing considering other recent studies that endowed P. putida with E. coli genes for better growth on xylose1 9 , 2 1 . It demonstrates that a similar outcome can be achieved with fewer engineering steps. A 3-week control ALE of P. putida PD310 did not result in enhanced growth on xylose (Supplementary Fig. 5). This suggests that the hexR deletion in PD584 and PD689 is an important factor for the enhanced carbon passage through the native glycolysis reactions (namely, the Zwf-Pgl-Edd-Eda part of the EDEMP cycle) and fast evolution of these strains1 2 . Growth assays confirmed that all three evolved strains utilize xylose more efficiently than their ancestors (Table 2, Fig. 4 d , Supplementary Fig. 6). Remarkable was the 5-fold reduction of the lag phase of PD698 tt LI compared to PD689 (Table 2, Fig. 4d). PD689 tt LI utilized xylose as efficiently as PD584 L3, which, however, originated from the much better-growing ancestor PD584 (Table 2, Fig. 4d). Additional cultivation experiments with glucose (Supplementary Fig. 7a, Supplementary Note 1) revealed a partially negative effect of metabolic adjustments on the growth rate and lag phase of PD584 tt L3 and PD689 tt LI on this substrate. However, both PD689 tt LI and PD584 L3 maintained the previously reported ability of the ancestral strain PD310 to co-utilize glucose and xylose (Supplementary Fig. 7b)1 8 . Unveiling the causes of the improved xylose utilization We combined proteomic analysis with whole-genome sequencing to correlate proteome changes with genomic alterations in P. putida Nature Communications | (2024)15:2666 8 Article https://doi.org/10.1038/s41467-024-46812-9 PD310, PD584, PD584 L3, and PD689 tt LI grown on xylose. Strain PD584 tt L3 was excluded from these characterizations because its growth on glucose was impaired (Supplementary Fig. 7a) and sequencing revealed that, in contrast to PD689 tt LI, the talB-tktA cassette was not incorporated into its genome, and therefore its genotype was similar to strain PD584 L3 (Supplementary Note 2 and Supplementary Table 1). Altogether, 3981 proteins (92 duplicities removed) were uniquely detected in the four strains (Supplementary Data 2). The proteomic analysis showed relatively subtle differences between PD584 and PD310 (118 downregulated and 87 upregulated proteins, log2 fold change >1.0, p < 0.05, Supplementary Fig. 8) and between the evolved strain PD584 L3 and its template PD584 (83 downregulated and 58 upregulated proteins. Supplementary Fig. 8), while the proteomes of the two evolved strains PD689 tt LI and PD584 L3 varied in 667 proteins (376 downregulated and 292 upregulated. Supplementary Fig. 8). Visualizing the changed protein abundances in the P. putida central carbon metabolism map verified the effects of our targeted engineering interventions and disclosed additional mutations selected in the ALE experiment (Fig. 6). Sequence analysis revealed various polymorphisms affecting the sequence of the encoded proteins in parental strains PD584 and PD689 (Supplementary Data 3,4). In the evolved strains, changes represented by de novo mutations were significantly more prevalent than fixed polymorphisms. Due to the detected genomic changes, the chromosomes of PD584 L3 and PD689 tt LI showed 99.8% and 99.9% pairwise identity, respectively, to their parental strains (using whole-genome alignment by Mauve Plugin in Geneious Prime, Supplementary Data 3, 4). Given the known variability of pseudomonad genomes, such changes are not surprising5 3 . Numerous mutations accumulated in PP_0168 (locus tag PED37_RS00910 in PD584) encoding the large adhesive protein LapA responsible for biofilm formation5 4 . Mutations in biofilm genes can occur in response to stress or selection of planktonic cells during multiple transfers in ALE experiments, as was observed previously4 7 , 5 5 . However, we did not observe a significant difference in biofilm forming ability between PD584 L3, PD689 tt LI, and PD584 used as a control (Supplementary Fig. 9 and Supplementary Method 2). Importantly, we identified an intriguing alteration in the genomes of PD584 and PD584 L 3 - a multiplication of a large region (118,079 bp), discovered by an approximately 6-fold higher sequencing coverage from locus PP_2114 to PP_2219 (Fig. 7a, Supplementary Table 2, Supplementary Data 5). Further investigation revealed that this multiplication was already present in the previously published strain PD3101 8 , but was not present in its EM42 Aged template (Supplementary Table 2). Freshly prepared EM42 Aged pSEVA2213_xy/-4BF exhibited a similar growth rate as the original PD310 but the fresh clones passed through a very long lag phase (-40 h of slow linear growth) before growing exponentially (Table 2, Fig. 7b). This experiment demonstrated a positive effect of the multiplication on xylose utilization. The multiplication was maintained in the whole lineage of PD310, PD584, and PD584 L3 strains (Fig. 4c) but was absent in PD689 tt LI and its ancestor PD689. As its two border open reading frames (ORFs) encode transposases, we presume that the multiplication occurred via repeated transposition events. The multiplied region includes 108 CDS of which 9 encode hypothetical proteins. The remaining CDS encode transporters or their subunit (3 genes), transcriptional regulators (7 genes), or enzymes (53 genes) (Supplementary Data 5). Notably, the latter set includes the gene of transaldolase Tal (PP_2168). We cloned the tal gene into the expression plasmid pSEVA438 (Supplementary Method 3, 4, Supplementary Data 6) and inserted the resulting construct into the strain EM42 Aged pSEVA2213_xy£4B£. Growth comparison of this strain on xylose with control (EM42 Aged pSEVA2213_xy£4B£ +pSEVA438) demonstrated that overexpression of the tal gene alone ensured an improved phenotype, which in the case of PD310 was made possible by amplification of the entire 118 kb segment including tal (Fig. 7c). We hypothesized that the transposition and segment multiplication in strains PD310 and PD584 occurred during their restreaking on agar plates with xylose, which we performed to check the desired phenotype before glycerol stock preparation. To verify this, we streaked several clones of the freshly prepared EM42 Aged pSEVA2213_xy£4B£strain without multiplication on a xylose agar plate. When the clones were re-streaked on a fresh plate and then cultivated in liquid microplate cultures, we indeed observed faster growth, comparable to the PD310 control (Supplementary Fig. 10). This growth acceleration can thus be attributed to the replication of the large genomic segment with the tal gene, which was identified in the chromosomes of all four clones (Supplementary Table 2). Proteome comparison of PD584 and PD310 found that the hexR deletion in PD584 manifested in increased abundances of the enzymes ZwfA, Pgl, Edd, Eda, and Gap (Fig. 6a) as expected. A potential redox imbalance caused by higher activity of de-repressed dehydrogenases Zwf and Gap could explain the apparent upregulation of glyoxylate shunt enzymes isocitrate lyase AceA (PP 4116) and malate synthase GlcB (PP 0356), as discussed in the first part of the Results and discussion section. The modestly increased abundance of Tal transaldolase (PP_2168) (Fig. 6a, Supplementary Data 5) can be attributed to a higher copy number of the multiplied PP 2114-PP 2219 segment in the chromosome of PD584 (six copies) compared to PD310 (four copies, Supplementary Table 2). Inspection of downregulated and upregulated proteins in the central carbon metabolism of PD584 L3 compared to PD584 revealed only minute changes including a small decrease in the quantity of ZwfA in the EDEMP cycle, AceA in the glyoxylate shunt, and malate dehydrogenase M d h (PP 0654) in the TCA cycle (Fig. 6b). The multiplication could not be responsible for the improved phenotype of the evolved mutant PD584 L3 as the same number of copies was detected in both strains (Supplementary Table 2), and other genomic changes that we identified and were able to interpret did not clearly explain the improved phenotype either. Finally, sequencing of the pSEVA2213_xylABE plasmid revealed a perfect 32 bp duplication upstream of the xylA gene in PD584 L3 (Fig. 7f). The duplication encompassed the synthetic Shine-Dalgarno sequence (RBS), the ATG start codon and the following eight nucleotides of the xylA gene. Analysis of the resulting mRNA sequence with RBS Calculator3 0 revealed that the RBS in the duplication with a predicted translation initiation rate of 310 a.u. became the strongest RBS upstream of thexy£4 gene while the strength of the original xylA RBS was reduced 10-fold from 292 to only 29 a.u. It is plausible that the effect of the duplication lies in the emergence of a mechanism similar to a translational coupler, that is, a region downstream of the promoter that encodes a short leading peptide stabilizing translation of the downstream gene5 6 . However, no such peptide was identified in the PD584 L3 proteome, so the specific molecular effect of the duplication remains to be elucidated (Methods). Since an increase in exogenous XylA abundance was confirmed on the protein level (Fig. 6b), it was probable that the duplication caused higher xylA expression. To verify the hypothesized effect of the duplication on xylose utilization by P. putida, PD584 L3 was deprived of the plasmid by sub-culturing in rich LB medium without antibiotics and then transformed either with the same mutated plasmid or with the original pSEVA2213_xy//4Bf. The strain with the mutated plasmid grew equally well on xylose compared to PD584 L3 (n = 0.21 ± 0 . 0 0 h_ 1 ), while the strain transformed with the original plasmid grew 33% slower (n = 0.14 ± 0 . 0 0 h"1 . Supplementary Fig. 11a). Similarly, when the mutant plasmid from PD584 L3 was inserted into PD584 strain devoid of its own pSEVA2213_xy//4B£ construct, the resulting strain showed faster growth on xylose than Nature Communications | (2024)15:2666 9 Article https://doi.org/10.1038/s41467-024-46812-9 PD584 vs. PD310 Xylose (XYL) 3PG PD689 tt L1 vs. PD584 L3 XylE (1.00) Catalase (KatE, PP_0115) f (9.53) XYL Glutathione peroxidase (PP_0777) 02} Fig. 61 Changes in protein abundances in the upper xylose pathway and central carbon metabolism in selected Pseudomonasputida strains. Changes in protein abundances for (a) strain PD584 compared to PD310, (b) PD584 L3 compared to PD584, and (c) PD689 tt LI compared to PD584 L3. Normalized and imputed protein intensities were used for differential expression using LIMMA statistical test. The figures show log2(fold change, FC) values for significantly differentially expressed proteins (adj.p< 0.05). Thep values adjustment on multiple hypothesis CO2 NADPH f f s-0.5 <-1 <-2 log2FC >2 >1 >0.5 i i testing was done using Benjamini & Hochberg method. The metabolic map and abbreviations used are the same as in Fig. 2. Please note that Tal and Tkt represent native Pseudomonas putida transaldolase (PP_2168) and transketolase (PP_4965), respectively, while TalB (KEGG ID:JW0007) and TktA (KEGG ID:JW5478) stand for the respective enzymes from Escherichia coli. Note that proteins with less significant changes in abundance (log2FC > 0.5 or < -0.5) were also visualized. Source data are provided as a Source Data file. Nature Communications | (2024)15:2666 10 Article https://doi.org/10.1038/s41467-024-46812-9 1,255 t PD584 162 2,372,035 2,490,113 1,506 453 PD689 tt L1 2,369,985 2,488,063 (>(>«iai 3.5 within 24 h of growth, and secondly at the end of the experiment (after - 6 0 - 7 0 generations. Fig. 5). Cells were plated on M 9 agar plates with 2 g L"1 xylose and K m and 2 - 3 fastest growing clones from each cultivation were picked for further characterization. Growth of the selected clones on xylose was first tested in a 48-well plate format and nine fastest-growing candidates were verified in 24 h-long shake-flask cultures (Supplementary Fig. 4). The number of generations in the evolution experiment was calculated for each of the evolved strains using the Eq. (I)4 : Number of generations = l n ( O D 6 0 0 f i n a | /OD6 0 0 i n i t i a l )/ln(2) (1) Calculations of dry cell weight and growth parameters Dry cell weight (DCW), maximal specific growth rate (umax), lag phase (in h), biomass yield (YX /S ), and biomass-specific substrate uptake rate (as ) of characterized P. putida strains were determined as described in the Supplementary Method 7. 13 C labeling experiments and analysis of metabolic fluxes An initial pre-culture in LB medium was inoculated from a cryogenic glycerol stock of strain PD310 (Table 1) and incubated in a rotary shaker at 30 °C, 3 0 0 rpm and 5 c m amplitude. The second preculture was performed in 100 mL shake flask using 10 mL M 9 medium with 5 g L"1 xylose. The medium was inoculated with the first preculture to an O D 6 0 o of 0.05 and incubated overnight under otherwise identical conditions as the LB preculture. Three 250 mL Erlenmeyer flasks containing 25 mL M 9 medium were inoculated with the second overnight culture to an O D 6 0 o of0.025. The media contained 5 g L"1 1,2-1 3 C xylose, i.e., xylose labeled with 1 3 C isotopes at positions CI and C2 (SigmaAldrich, 99% purity). Cultures were grown under agitation (300 rpm, amplitude: 5 cm) at 30 °C. Substrate uptake was monitored by HPLCUV/RI analysis of the fermentation broth. Biomass concentrations were monitored using the cell growth quantifier (Scientific Bioprocessing) and determined manually by measuring the optical density at 6 0 0 nm with an Ultrospec 10 Cell Density Meter (GE Healthcare). The conversion factor of OD6 oo to cell dry weight (CDW) in g/L was gravimetrically determined to 0.39. During the exponential phase, OD and xylose concentrations were determined in 1 h intervals to allow accurate determination of growth and substrate uptake rates. At mid-exponential growth ( O D 6 0 o of 1.0-1.3), samples were taken to quantify the 1 3 C isotope incorporation Nature Communications | (2024)15:2666 14 Article https://doi.org/10.1038/s41467-024-46812-9 into proteinogenic amino acids and free intracellular metabolites. A fast filtration method7 3 was applied for intracellular metabolites to rapidly sample the biomass and quench the metabolism. Briefly, samples corresponding to 10 mg biomass were taken and the biomass was collected by vacuum filtration (Durapore, PVDF, 0.45 urn, 47 m m Sigma-Aldrich). After one washing step with 0.9% saline, the filter was placed (upside down) into a small Petri dish (5 cm diameter) filled with methanol (pre-cooled at - 8 0 °C), incubated at - 8 0 °C for l h . Filters were scraped and rinsed to ensure all cell material was in the liquid solvent, the solvent and filter were transferred to microfuge tubes and vigorously vortexed at - 2 0 °C. The filter was removed, the extract centrifuged at 17,000 g at 4 °C for 5 min, and the supernatant removed and stored at - 8 0 °C. Filter and solvent were transferred into a tube and vortexed at - 2 0 °C. Extracts were dried in a lyophilizer and analyzed using capillary IC-MS analysis (Supplementary Method 8 and Supplementary Table 4, 5). The metabolites from upper glycolysis were important for a better resolution of fluxes of this part of the metabolism, especially cyclic fluxes. Capillary IC-MS data were corrected for the natural abundance of heavy isotopes using IsoCorr7 4 . Determination of 13 C-labeling patterns in proteinogenic amino acids A standard protocol was used as described in Schmitz et al.7 5 . In brief, samples corresponding to 0.3 m g CDW were taken and centrifuged for 10 min at 4 °C at 17,000 g. The pellets were washed and resuspended in 5 N HCI, transferred to GC vials, and the samples were hydrolyzed at 105 °C for 6 h . Derivatization of the dried hydrolysate was done in a mix of 30 uLacetonitrileand 30 uL N-methyl-N-tert-butyldimethylsilyltrifluoroacetamide (CS-Chromatographie Service GmbH) at 85 °C for 1 h. The derivatized samples were analyzed on a single-quadruple gas chromatography-mass spectrometry (GC-MS) system. Gas chromatography separation was performed using TRACE™ G C Ultra (Thermo Fisher Scientific, Waltham, MA, USA) equipped with an AS 3 0 0 0 autosampler. The column used consisted of TraceGOLD TG-5SilMS fused silica (length, 30 m; inner diameter, 0.25 m m ; film thickness, 0.25 um). Helium was used as carrier gas at a constant gas flow rate of 1 mL min 1 and a split ratio of 1:15. The injector temperature was set to 270 °C, and the column oven was heated according to a ramped program. The initial temperature of 140 °C was held for 1 min, the temperature was then increased at a rate of 10 °C/min to a final value of 310 °C, which was held time for l m i n . Mass spectrometry analysis utilized a Thermo Scientific ISQ single quadrupole mass spectrometer (Waltham, MA, USA). The transfer line and ion source temperatures were maintained at 280 °C, and ionization was achieved via electron impact (El) ionization at 70 eV. Subsequent analysis of GC-MS raw data was conducted using Xcalibur software. Data were corrected for unlabeled biomass (introduced with the inoculum) and natural abundance of heavy isotopes using the software iMS2FLUX7 6 . Metabolic flux analysis was performed using the Matlab-based tool INCA7 7 . The model was constrained with mass isotopomer distribution data of intracellular metabolites and proteinogenic amino acids and the specific growth and xylose uptake rates. Confidence intervals were determined by parameter continuation using INCA's in-build function. Genome-scale metabolic model simulations The most recent iJN1463 genome-wide metabolic reconstruction of P. putida KT24407 8 was downloaded from the BIGG database (http://bigg. ucsd.edu/). Two metabolites (xylose and xylulose) and four reactions xylose transport to the periplasm and to cytosol (XYLtex, XYLt2pp respectively), xylose isomerase (XYLI1) and xylulokinase (XYLK) were added to the model. The glucose dehydrogenase reaction (GCD) was deleted from the model to prevent the oxidation of glucose to gluconate and 2-ketogluconate. To predict optimal flux distributions by flux balance analysis (FBA)4 1 either COBRApy library or MATLAB COBRA toolbox7 9 was used. Xylose uptake rate was fixed at experimentally determined 1.45 mmol gcDw1 h"1 . Model reactions of central carbon metabolism, TCA cycle, and CO2 production were then constrained with upper and lower bounds as determined by metabolic flux analysis (Fig. 2). To get an upper and lower bound from MFA data, firstly, the standard error was calculated from two independent measurements and then the lower and upper bound on the fluxes were calculated as ± 1.96 x standard error (1.96 corresponds to the 97.5t h percentile of a standard normal distribution). iJN1463 contains two malate dehydrogenase reactions (MDH and MDH2). The combined flux through these two reactions was set to the upper and lower bound calculated from MFA. The two models - one constrained with MFA data (MFA model) and another one constrained only with xylose uptake (FBA model) - were then compared (both modified models are available from GitHub). The glpk solver was used for all simulations. Maximizing the biomass formation rate was the objective function in all simulations. Whole-genome sequencing and proteomic analyses Details on whole-genome sequencing of selected P. putida strains are provided in Supplementary Method 9. Details on proteome characterization of selected P. putida strains are provided in Supplementary Method 10 and Supplementary Table 6. Analytical methods The optical density in cell cultures was recorded at 6 0 0 nm using UV/VIS spectrophotometer Genesys 5 (Spectronic). Analytes from cultures were collected by withdrawing 0.5 ml of culture medium. The sample was then centrifuged (20,000 g, 10 min). The supernatant was filtered through 4 m m / 0.45 urn LUT Syringe Filters (Labstore) and stored at - 2 0 °C. Prior to the HPLC analysis, 50 m M H 2 S 0 4 in degassed miliQ water was added to the samples in a 1:1 ratio to stop any hydrolytic activity and to dilute the samples. Highperformance liquid chromatography (HPLC) was used to quantify xylose and glucose. HPLC analysis was carried out using Agilent 1100 Series system (Agilent Technologies) equipped with a refractive index detector and Hi-Plex H, 7 . 7 x 3 0 0 m m , 8 urn HPLC column (Agilent Technologies). Analyses were performed using the following conditions: mobile phase 5 m M H 2 S 0 4 , mobile phase flow 0.5 mL m i n 1 , injection volume 2 0 uL, column temperature 65 °C, Rl detector temperature 55 °C. Xylose and glucose standards (SigmaAldrich) were used for the preparation of calibration curves. Xylose concentrations in labeling experiments were determined using a Beckman System Gold 126 Solvent Module equipped with a System Gold 166 UV-detector (Beckman Coulter) and a Smartline Rl detector 2300 (Knauer). Analytes were separated on the organic resin column Metab A A C (Isera) eluted with 5 m M H 2 S 0 4 at an isocratic flow of 0.6 mL min"1 at 4 0 °C for 4 0 min. Glucose and xylose concentrations in culture supernatants were alternatively determined also by Glucose (GO) Assay Kit (SigmaAldrich, USA) and Xylose Assay Kit (Megazyme, Ireland), following the manufacturer's instructions. Product concentrations were measured spectrophotometrically using Infinite M Plex reader (Tecan). Data and statistical analyses The number of repeated experiments or biological replicates is specified in figure and table legends. The mean values and corresponding standard deviations are presented. When appropriate, data were treated with a two-tailed Student's f-test in Microsoft Office Excel 2013 (Microsoft) and confidence intervals were calculated for the given parameters to test a statistically significant difference in means between two experimental datasets. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Nature Communications | (2024)15:2666 15 Article https://doi.org/10.1038/s41467-024-46812-9 Data availability All sequencing data and assembled whole-genome sequences were deposited under NCBI BioProject PRJNA914626. The detailed information can be found in Supplementary Table 1. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRI DE partner repository with the dataset identifier PXD047537. The raw gas chromatography-mass spectrometry and ion chromatography-mass spectrometry data used for the metabolic flux analysis in this study have been deposited in the Zenodo repository [https://zenodo.org/records/10732391]. Source data are provided with this paper. Code availability Two modified genome-scale metabolic models of P. putida used in this study are available from GitHub [https://github.com/DalimilBujdos/P_ putida_xylose_metabolism]. References 1. Narisetty, V. et al. Valorisation of xylose to renewable fuels and chemicals, an essential step in augmenting the commercial viability of lignocellulosic biorefineries. Sust ain. Energy Fuels 6, 2 9 - 6 5 (2021). 2. Baral, N. R. et al. Approaches for more efficient biological conversion of lignocellulosic feedstocks to biofuels and bioproducts. ACS Sustain. Chem. Eng. 7, 9062-9079 (2019). 3. Sun, L. et al. Complete and efficient conversion of plant cell wall hemicellulose into high-value bioproducts by engineered yeast. Nat. Commun. 12, 4975 (2021). 4. Ling, C. et al. Muconic acid production from glucose and xylose in Pseudomonas put ida via evolution and metabolic engineering. Nat . Commun. 13, 4925 (2022). 5. Taha, M. et al. Commercial feasibility of lignocellulose biodegradation: possibilities and challenges. Curr. Opin. Biot echnol. 38, 190-197(2016). 6. Görke, B. & Stülke, J. Carbon catabolite repression in bacteria: many ways to make the most out of nutrients. Nat . Rev. Microbiol. 6, 613-624 (2008). 7. Volk, M. J. et al. Metabolic engineering: methodologies and applications. Chem. Rev. 123, 5521-5570 (2022). 8. Dragosits, M. & Mattanovich, D. Adaptive laboratory evolution principles and applications for biotechnology. Microb. Cell Factories 12, 64 (2013). 9. Sandberg, T. E., Salazar, M. J., Weng, L. L, Palsson, B. O. & Feist, A. M. The emergence of adaptive laboratory evolution as an efficient tool for biological discovery and industrial biotechnology. Met ab. Eng. 56,1-16 (2019). 10. Porcar, M., Danchin, A. & de Lorenzo, V. Confidence, tolerance, and allowance in biological engineering: the nuts and bolts of living things. BioEssays 37, 95-102 (2015). 11. Mavrommati, M., Daskalaki, A., Papanikolaou, S. & Aggelis, G. Adaptive laboratory evolution principles and applications in industrial biotechnology. Biot echnol. Adv. 54,107795 (2022). 12. Nikel, P. I ., Martinez-Garcia, E. & de Lorenzo, V. Biotechnological domestication of pseudomonads using synthetic biology. Nat . Rev. Microbiol. 12, 368-379 (2014). 13. Sullivan, K. P. et al. Mixed plastics waste valorization through tandem chemical oxidation and biological funneling. Science 378, 207-211 (2022). 14. Dvorak, P., Nikel, P. I ., Damborsky, J. & de Lorenzo, V. Bioremediation 3.0: Engineering pollutant-removing bacteria in the times of systemic biology. Biot echnol. Adv. 35, 845-866 (2017). 15. Salvachüa, D. et al. Outer membrane vesicles catabolize ligninderived aromatic compounds in Pseudomonas put ida KT2440. Proc. Nat l Acad. Sei. 117, 9302-9310 (2020). 16. Liu, Z.-H. et al. Transforming biorefinery designs with 'Plug-I n Processes of Lignin' to enable economic waste valorization. Nat . Commun. 12, 3912 (2021). 17. Kohlstedt, M. et al. Biobased PET from lignin using an engineered cis, c/'s-muconate-producing Pseudomonas put ida strain with superior robustness, energy and redox properties. Met ab. Eng. 72, 337-352 (2022). 18. Dvořák, P. & de Lorenzo, V. Refactoring the upper sugar metabolism of Pseudomonas put ida for co-utilization of cellobiose, xylose, and glucose. Met ab. Eng. 48, 94-108 (2018). 19. Elmore, J. R. et al. Engineered Pseudomonas put ida simultaneously catabolizes five major components of corn stover lignocellulose: Glucose, xylose, arabinose, p-coumaric acid, and acetic acid. Metab. Eng. 62, 62-71 (2020). 20. Meijnen, J.-P., de Winde, J. H. & Ruijssenaars, H. J. Engineering Pseudomonas put ida S12 for efficient utilization of D-xylose and L-arabinose. Appl. Environ. Microbiol. 74, 5031-5037 (2008). 21. Lim, H. G. et al. Generation of Pseudomonas put ida KT2440 strains with efficient utilization of xylose and galactose via adaptive laboratory evolution. ACS Sust ain. Chem. Eng. 9, 11512-11523 (2021). 22. Bator, I., Wittgens, A., Rosenau, F., Tiso, T. & Blank, L. M. Comparison of three xylose pathways in Pseudomonas put ida KT2440 for the synthesis of valuable products. Front . Bioeng. Biot echnol. 7, 480 (2020). 23. Wang, Y. et al. Growth of engineered Pseudomonas put ida KT2440 on glucose, xylose, and arabinose: Hemicellulose hydrolysates and their major sugars as sustainable carbon sources. GCB Bioenergy 11, 249-259 (2019). 24. Nikel, P. I ., Chavarría, M., Fuhrer, T., Sauer, U. & de Lorenzo, V. Pseudomonas put ida KT2440 strain metabolizes glucose through a cycle formed by enzymes of the entner-doudoroff, embdenmeyerhof-parnas, and pentose phosphate pathways. J. Biol. Chem. 290, 25920-25932 (2015). 25. Martínez-García, E., Nikel, P. I ., Aparicio, T. & de Lorenzo, V. Pseudomonas 2.0: genetic upgrading of P. putida KT2440 as an enhanced host for heterologous gene expression. Microb. Cell Factories 13,159 (2014). 26. Meijnen, J.-P., Winde, J. Hde & Ruijssenaars, H. J. Metabolic and regulatory rearrangements underlying efficient d-xylose utilization in engineered Pseudomonas put ida S12 *. J. Biol. Chem. 287, 14606-14614 (2012). 27. Dvořák, P., Kováč, J. & de Lorenzo, V. Biotransformation of D-xylose to D-xylonate coupled to medium-chain-length polyhydroxyalkanoate production in cellobiose-grown Pseudomonas put ida EM42. Microb. Biot echnol. 13,1273-1283 (2020). 28. Bujdoš, D. et al. Engineering of Pseudomonas put ida for accelerated co-utilization of glucose and cellobiose yields aerobic overproduction of pyruvate explained by an upgraded metabolic model. Met ab. Eng. 75, 2 9 - 4 6 (2023). 29. Weimer, A., Kohlstedt, M., Volke, D. C , Nikel, P. I . & Wittmann, C. Industrial biotechnology of Pseudomonas put ida-. advances and prospects. Appl. Microbiol. Biot echnol. 104, 7745-7766 (2020). 30. Salis, H. M., Mirsky, E. A. & Voigt, C. A. Automated design of synthetic ribosome binding sites to control protein expression. Nat . Biotechnol. 27, 9 4 6 - 9 5 0 (2009). 31. Ceroni, F., Algar, R., Stan, G.-B. & Ellis, T. Quantifying cellular capacity identifies gene expression designs with reduced burden. Nat. Met hods 12, 415-418 (2015). 32. Wiechert, W. 1 3 C metabolic flux analysis. Met ab. Eng. 3, 195-206 (2001). 33. Kohlstedt, M. & Wittmann, C. GC-MS-based 1 3 C metabolic flux analysis resolves the parallel and cyclic glucose metabolism of Pseudomonas put ida KT2440 and Pseudomonas aeruginosa PA01. Metab. Eng. 54, 35-53 (2019). Nature Communications | (2024)15:2666 16 Article https://doi.org/10.1038/s41467-024-46812-9 34. Beckers, V., Poblete-Castro, I., Tomasch, J. & Wittmann, C. Integrated analysis of gene expression and metabolic fluxes in PHAproducing Pseudomonas putida grown on glycerol. Microb. Cell Factories 15, 73(2016). 35. Kukurugya, M. A. et al. Multi-omics analysis unravels a segregated metabolic flux network that tunes co-utilization of sugar and aromatic carbons in Pseudomonas putida. J. Biol. Chem. 294, 8464-8479 (2019). 36. Wilkes, R. A., Waldbauer, J. & Aristilde, L. Analogous Metabolic Decoupling in Pseudomonas putida and Comamonas testosteroni implies energetic bypass to facilitate gluconeogenic growth. mBio 12, e03259-21 (2021). 37. Sudarsan, S. et al. Dynamics of benzoate metabolism in Pseudomonas putida KT2440. Metab. Eng. Commun. 3, 97-110 (2016). 38. La Rosa, R., Nogales, J. & Rojo, F. The Crc/CrcZ-CrcY global regulatory system helps the integration of gluconeogenic and glycolytic metabolism in Pseudomonas putida. Environ. Microbiol. 17, 3362-3378 (2015). 39. Sasnow, S. S., Wei, H. & Aristilde, L. Bypasses in intracellular glucose metabolism in iron-limited Pseudomonas putida. Microbiol. Open 5, 3-20(2016). 40. Nikel, P. I. et al. Reconfiguration of metabolic fluxes in Pseudomonas putida as a response to sub-lethal oxidative stress. ISMEJ. 15, 1751-1766 (2021). 41. Orth, J. D., Thiele, I. & Palsson, B. 0 . What is flux balance analysis? Nat. Biotechnol. 28, 245-248 (2010). 42. Cheng, K.-K. et al. Global metabolic network reorganization by adaptive mutations allows fast growth of Escherichia coli on glycerol. Nat. Commun. 5, 3233 (2014). 43. Breckell, G. L. & Silander, O. K. Growth condition-dependent differences in methylation imply transiently differentiated DNA methylation states in Escherichia coli. G3 Genes Genomes Genet. 13, jkac310 (2022). 44. Udaondo, Z., Ramos, J., Segura, A., Krell, T. & Daddaoua, A. Regulation of carbohydrate degradation pathways in Pseudomonas involves a versatile set of transcriptional regulators. Microb. Biotechnol. 11, 442-454 (2018). 45. Rojo, F. Carbon catabolite repression in Pseudomonas-, optimizing metabolic versatility and interactions with the environment. FEMS Microbiol. Rev. 34, 658-684 (2010). 46. del Castillo, T., Duque, E. & Ramos, J. L. A set of activators and repressors control peripheral glucose pathways in Pseudomonas putida to yield a common central intermediate. J. Bacteriol. 190, 2331-2339 (2008). 47. Bentley, G. J. et al. Engineering glucose metabolism for enhanced muconic acid production in Pseudomonas putida KT2440. Metab. Eng. 59, 64-75 (2020). 48. Volke, D. C , Olavarria, K. & Nikel, P. I. Cofactor specificity of Glucose-6-Phosphate dehydrogenase isozymes in Pseudomonas putida reveals a general principle underlying glycolytic strategies in bacteria. mSystems 6, e00014-e00021 (2021). 49. Noor, E., Eden, E., Milo, R. & Alon, U. Centralcarbon metabolismasa minimal biochemical walk between precursors for biomass and energy. Mol. Cell 39, 8 0 9 - 8 2 0 (2010). 50. Zhang, M., Eddy, C , Deanda, K., Finkelstein, M. & Picataggio, S. Metabolic engineering of a pentose metabolism pathway in ethanologenic Zymomonas mobilis. Science 267, 240-243 (1995). 51. Martinez-Garcia, E., Aparicio, T., de Lorenzo, V. & Nikel, P.I. New transposon tools tailored for metabolic engineering of gramnegative microbial cell factories. Front. Bioeng. Biotechnol. 1, 46 (2014). 52. Espeso, D. R., Dvorak, P., Aparicio, T. & de Lorenzo, V. An automated DIY framework for experimental evolution of Pseudomonas putida. Microb. Biotechnol. 14, 2679-2685 (2020). 53. Silby, M. W., Winstanley, C , Godfrey, S. A. C , Levy, S. B. & Jackson, R. W. Pseudomonas genomes: diverse and adaptable. FEMS Microbiol. Rev. 35, 6 5 2 - 6 8 0 (2011). 54. El-Kirat-Chatel, S., Beaussart, A., Boyd, C. D., OToole, G. A. & Dufrene, Y. F. Single-cell and single-molecule analysis deciphers the localization, adhesion, and mechanics of the biofilm adhesin LapA. ACS Chem. Biol. 9, 485-494 (2014). 55. Chu, E. K., Kilic, O., Cho, H., Groisman, A. & Levchenko, A. Selfinduced mechanical stress can trigger biofilm formation in uropathogenic Escherichia coli. Nat. Commun. 9, 4087 (2018). 56. Mutalik, V. K.etal. Precise and reliable gene expression via standard transcription and translation initiation elements. Nat. Methods 10, 3 5 4 - 3 6 0 (2013). 57. Paget, M. S. Bacterial sigma factors and anti-sigma factors: structure, function and distribution. Biomolecules 5,1245-1265 (2015). 58. Conrad, T. M. et al. RNA polymerase mutants found through adaptive evolution reprogram Escherichia coli for optimal growth in minimal media. Proc. Natl Acad. Sei. Usa. 107, 20500-20505(2010). 59. Tenaillon, O. et al. The molecular diversity of adaptive convergence. Science 335, 457-461 (2012). 60. Choe, D. et al. Adaptive laboratory evolution of a genome-reduced Escherichia coli. Nat. Commun. 10, 935 (2019). 61. Tomko, T. A. & Dunlop, M. J. Expression of heterologous sigma factor expands the searchable space for biofuel tolerance mechanisms. ACS Synth. Biol. 6,1343-1350 (2017). 62. Tomatis, P. E., Schütz, M., Umudumov, E. & Plückthun, A. Mutations in sigma 70 transcription factor improves expression of functional eukaryotic membrane proteins in Escherichia coli. Sei. Rep. 9, 2483 (2019). 63. Pang, C. et al. Engineering sigma factors and chaperones for enhanced heterologous lipoxygenase production in Escherichia coli. Biotechnol. Biofuels Bioprod. 15,105 (2022). 64. Wagner, S., Bader, M. L, Drew, D. & de Gier, J.-W. Rationalizing membrane protein overexpression. Trends Biotechnol. 24, 364-371 (2006). 65. Schmitz, S., Nies, S., Wierckx, N., Blank, L. M. & Rosenbaum, M. A. Engineering mediator-based electroactivity in the obligate aerobic bacterium Pseudomonas putida KT2440. Front. Microbiol. 6, 284 (2015). 66. Hassan, H. M. & Fridovich, I. Mechanism of the antibiotic action pyocyanine. J. Bacteriol. 141,156-163 (1980). 67. DeBritto, S. et al. Isolation and characterization of nutrient dependent pyocyanin from Pseudomonas aeruginosa and its dye and agrochemical properties. Sei. Rep. 10,1542 (2020). 68. Tan, Z. L. et al. In vivo continuous evolution of metabolic pathways for chemical production. Microb. Cell Factories 18, 82 (2019). 69. Wirth, N. T. et al. A synthetic C2 auxotroph of Pseudomonas putida for evolutionary engineering of alternative sugar catabolic routes. Metab. Eng. 74, 83-97 (2022). 70. Abril, M. A., Michan, C , Timmis, K. N. & Ramos, J. L. Regulator and enzyme specificities of the TOL plasmid-encoded upper pathway for degradation of aromatic hydrocarbons and expansion of the substrate range of the pathway. J. Bacteriol. 171, 6782-6790 (1989). 71. Martinez-Garcia, E. & de Lorenzo, V. Transposon-based and plasmid-based genetic tools for editing genomes of gram-negative bacteria. Methods Mol. BioL Clifton NJ 813, 267-283 (2012). 72. Horton, R. M., Cai, Z., Ho, S. N. & Pease, L. R. Gene splicing by overlap extension: Tailor-made genes using the polymerase chain reaction. BioTechniques 8, 528-535 (1990). 73. Bennette, N. B., Eng, J. F. & Dismukes, G. C. An LC-MS-Based chemical and analytical method for targeted metabolite quantification in the model cyanobacterium Synechococcus sp. PCC 7002. Anal. Chem. 83, 3808-3816 (2011). Nature Communications | (2024)15:2666 17 Article https://doi.org/10.1038/s41467-024-46812-9 74. Millard, P., Letisse, F., Sokol, S. & Portais, J.-C. IsoCor: correcting MS data in isotope labeling experiments. Bioinforma. Oxf. Engl. 28, 1294-1296 (2012). 75. Schmitz, A., Ebert, B. E. & Blank, L. M. GC-MS-Based Determination of Mass Isotopomer Distributions for1 3 C-Based Metabolic Flux Analysis, in Hydrocarbon and Lipid Microbiology Protocols: Genetic, Genomic and System Analyses of Pure Cultures (eds. McGenity, T. J., Timmis, K. N. & Nogales, B.) 223-243 (Springer, 2017). https://doi. org/10.1007/8623_2015_78. 76. Poskar, C. H. et al. iMS2Flux- a high-throughput processing tool for stable isotope labeled mass spectrometric data used for metabolic flux analysis. BMC Bioinforma. 13, 295 (2012). 77. Young, J. D. INCA: a computational platform for isotopically nonstationary metabolic flux analysis. Bioinforma. Oxf. Engl. 30, 1333-1335 (2014). 78. Noga les, J. et a I. H ig h-q ua lity genome-sca le meta bolic modelli ng of Pseudomonas putida highlights its broad metabolic capabilities. Environ. Microbiol. 11, 255-269 (2020). 79. Heirendt, L. et al. Creation and analysis of biochemical constraintbased models using the COBRA Toolbox v.3.0. Nat. Protoc. 14, 639-702 (2019). 80. Swain, P. S. et al. Inferring time derivatives including cell growth rates using Gaussian processes. Nat. Commun. 7,13766 (2016). Acknowledgements We thank Dr. Adam Feist and Dr. Hyungyu Lim for valuable discussions on genomic data of engineered and evolved P. putida strains, and Dr. Ludmilla Aristilde for valuable discussion on flux analyses. This work was funded by Czech Science Foundation Project 22-12505 S and Grant Agency of Masaryk University GAMU Project MASH Junior 2022 (MUNI/J/ 0003/2021) granted to P.D. and Brno Ph.D. Talent granted to B.B. CIISB. This work was also supported by the project National Institute of Virology and Bacteriology (Programme EXCELES, ID Project No. LX22NPO5103), funded by the European Union - Next Generation EU. Instruct-CZ Centre of Instruct-ERIC EU consortium, funded by MEYS CR infrastructure project LM2023042, is gratefully acknowledged for the financial support of the measurements at the CEITEC Proteomics Core Facility. Computational resources were provided by the e-INFRA CZ project (ID:90254), supported by MEYS CR. We thank Dr Hendrik Ballersedt for support with G C - M S analysis and Tim Langhorst for esatablishing sampling procedures for intracellular metabolites. Author contributions P.D., B.B., B.P., and B.E.E. contributed equally, they conceptualized the work, performed experiments, cured data, and drafted the manuscript. P.D., in addition, conceived the work, secured resources, and supervised B.B., B.P., D.B., and M.B. T.B. designed and performed whole genome sequencing and cured data. D.B. designed and performed genomescale metabolic modeling and cured data. A. S.-P. contributed to the constructions of plasmids and strains. H.S. performed analyses (capillary IC-MS analysis), cured data. H.H. performed analyses (capillary ICMS analysis), supervised H.S. and cured data. V. de L. contributed to the work conceptualization and supervised A.S.-P. L.M. B. contributed to the work conceptualization, read and approved the manuscript. M.B. contributed to the work conceptualization, plasmid and strain constructions, and cell cultures, and cured data. All authors read, edited and approved the final manuscript. Competing interests The authors declare no competing interests. Additional information Supplementary information The online version contains Supplementary Material available at https://doi.org/10.1038/s41467-024-46812-9. Correspondence and requests for materials should be addressed to Pavel Dvorak. Peer review information Nature Communications thanks Magnus Carlquist and the other, anonymous, reviewers) for their contribution to the peer review of this work. A peer review file is available. Reprints and permissions information is available at http://www.nature.com/reprints Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. © The Author(s) 2024 Nature Communications | (2024)15:2666 18