Multi-layered metabolic remodeling of Pseudomonas putida for efficient conversion of lignocellulosic sugars to the precursors of advanced aviation fuel

TL;DR

Scientists engineered Pseudomonas putida to efficiently convert plant sugars into isoprenyl acetate, a precursor for advanced aviation fuel.

Contribution

A multi-layered metabolic engineering strategy in Pseudomonas putida enables high-yield production of isoprenyl acetate from mixed sugars.

Findings

Deleting native esterases and introducing a heterologous enzyme reduced product degradation and enabled isoprenyl acetate production.

Engineering glucose and xylose co-utilization improved sugar consumption and ester yields.

Fed-batch bioreactor cultures achieved 1.9 g/L isoprenyl acetate from mixed lignocellulosic sugars.

Abstract

Isoprenyl acetate, a volatile ester derived from isoprenol, is a key biosynthetic intermediate for the advanced aviation fuel candidate, 1,4-dimethylcyclooctane. Here, we engineered Pseudomonas putida KT2440 for the production of isoprenyl acetate from mixed sugar substrates. We first generated isoprenyl acetate by introducing a heterologous alcohol acetyltransferase (ATF1) and deleting three promiscuous native esterases to reduce product degradation. Then, we engineered efficient glucose and xylose co-utilization by integrating a heterologous xylose isomerase pathway and deleting global regulators crc and hexR to alleviate catabolite repression. Additionally, intracellular acetyl-CoA flux was reinforced through the expression of auxiliary carbon-conserving routes, including non-oxidative glycolysis and acetate assimilation. Culture conditions were systematically optimized by adjusting medium composition, induction, and overlay solvent to maximize product yields and titers. These cumulative efforts achieved isoprenyl acetate titers of 1.5 g/L in shake flasks and 1.9 g/L in fed-batch bioreactor cultures from mixed sugars, corresponding to a yield of 0.067 g/g of total sugar consumed. Our work demonstrates the potential of P. putida as a robust microbial chassis for scalable biosynthesis of ester-based biofuels from lignocellulosic feedstocks.

•Isoprenyl acetate biosynthesis is demonstrated in Pseudomonas putida.•Product-degrading endogenous esterases are identified and deleted in P. putida.•Co-utilization of glucose and xylose is engineered by relieving catabolite repression.•Co-expression of Acs and Xfpk enhances acetyl-CoA flux and boosts ester yields.•Fed-batch cultivation led to 1.9 g/L isoprenyl acetate from lignocellulosic sugar mixture.

Isoprenyl acetate biosynthesis is demonstrated in Pseudomonas putida.

Product-degrading endogenous esterases are identified and deleted in P. putida.

Co-utilization of glucose and xylose is engineered by relieving catabolite repression.

Co-expression of Acs and Xfpk enhances acetyl-CoA flux and boosts ester yields.

Fed-batch cultivation led to 1.9 g/L isoprenyl acetate from lignocellulosic sugar mixture.

Figures

Overall scheme of strategies employed to modify *P. putida* metabolism for the production of isoprenyl acetate from glucose and xylose. Gene names are displayed only for targets modified during our engineering efforts. Gray: Seven gene knockouts and the IPP-bypass pathway previously established; Red: Alcohol acyltransferase for isoprenyl acetate production; Yellow: Knockouts of esterases involved in isoprenyl acetate hydrolysis; Green: Integration of the xylose utilization pathway and disruption of regulators; Blue: Overexpression of genes to increase the acetyl-CoA pool. Isoprenyl acetate is

Introduction of alcohol acetyltransferases (AATs) into *P. putida* and isoprenyl acetate production. (a) Pathway for isoprenyl acetate production from acetyl-CoA and the two plasmids used in this study. The isoprenol production plasmid, pIY670, was constructed in a previous study ([Banerjee et al., 2024](#bib6)). Five AAT variants were introduced into the pRSF1010 plasmid and expressed under the P_nagA_ promoter. The pathway diagram illustrates the flow of acetyl-CoA, redox equivalents, ATP, and CO_2_. (b) Isoprenyl acetate and isoprenol titers in *P. putida* denoted by organism of origin and

Identification and knockout of putative esterases in *P. putida*. (a) Degradation of isoprenyl acetate in *P. putida*. (Left) Reaction scheme showing the hydrolysis of isoprenyl acetate into isoprenol and acetate, potentially mediated by promiscuous esterase activity. (Right) Isoprenyl acetate degradation test. Isoprenyl acetate was externally added at concentrations ranging from 0 to 2 g/L. Residual isoprenyl acetate and its hydrolyzed byproduct, isoprenol, were measured after incubation in a cell-free control or with *P. putida* KT2440. Cultures were grown in 5 mL culture tubes and analyzed

Optimization of induction and culture parameters. (a) Optimization of salicylic acid (SA) concentration. (b) Correlation between isoprenyl acetate titer and ATF1 protein abundance. A regression line is shown in red. (c) Optimization of the inducer addition timing. Arabinose (Ara) and SA were added at 4, 8, or 12 h after inoculation, corresponding to OD_600_ values of 0.5 (early exponential), 2.5 (mid-exponential), and 5.6 (late exponential/early stationary), respectively. (d) Evaluation of overlay solvents. All experiments were conducted using the PIPA-E1.2 strain in 5 mL culture tubes, with s

Engineering of PIPA for integration of the xylose isomerase pathway. (a) Metabolic engineering strategy. Xylose is metabolized through the heterologous xylose isomerase pathway, where *xylA, xylB*, and *xylE* convert xylose into xylulose-5-phosphate. *tal* and *tkt* are overexpressed within the same operon. The introduced pathway is highlighted in green. The *gcd* knockout is indicated by an “X”. Additionally, *crc* and *hexR* were deleted, and enzymes under their regulatory control involved in glucose and xylose utilization are indicated with dotted light green lines. These include a subset o

Engineering of PIPAxyl through CCR-related regulatory modifications. (a) Gene regulation by *crc*, *hexR*, and other regulatory elements. The Crc-Hfq complex regulates gene expression at the mRNA level, indicated on each respective gene with a green (−) triangle. HexR, GltR-II, and GlnG function as repressors (−) or activators (+) as marked at the promoters of their respective operons. (b) Isoprenyl acetate production in the PIPAxyl-E3-K3 strain. Experiments were conducted in 50 mL flasks, with samples collected every 24 h. Arabinose and salicylic acid were added when the OD_600_ reached 0.6–0

Engineering strategies to enhance the acetyl-CoA pool. (a) Schematic representation of the engineered pathways. Genes selected for overexpression are highlighted in blue, and a summary of their functional descriptions is provided in Table S8. The offset diagram on the right illustrates key reactions catalyzed by Xfpk and enzymes in the pentose phosphate pathway. (b) Plasmid-based early screening of eight target genes (O1 to O8). Each gene was overexpressed (O) under the constitutive P_J23107_ promoter on pAAT1 and introduced into the PIPAxyl strain. A “-” indicates PIPAxyl-AAT1 control strain.

Fed-batch cultivation of PIPAxyl-E3-K3-O15 for isoprenyl acetate production. (a) Residual sugar concentrations and feeding rate over time. The culture was initiated with 13.3 g/L glucose and 6.7 g/L xylose and continuously fed with a solution containing 266.7 g/L glucose and 133.3 g/L xylose. (b) Time-course profile of isoprenyl acetate production. Titers represent the sum of isoprenyl acetate quantified from the aqueous phase, the Durasyn overlay in the bioreactor, and the Durasyn off-gas trap. Cultivations were performed with 1 L medium and 200 mL overlay with sampling approximately every 12

Full text

Introduction

1

Production of aviation fuels from biological resources is crucial for the advancement of the aviation sector. Many precursors of advanced aviation fuels are bio-derived chemicals that reduce reliance on fossil fuels (Harvey et al., 2014), with candidates including short-chain olefins such as isoprene, 1-butene, 1-pentene, and 1-hexene (Morris et al., 2019). Notably, the U.S. Navy developed a high energy density jet fuel, 1,4-dimethylcyclooctane (DMCO), chemically synthesized via the dimerization and hydrogenation of isoprene, which can be produced from isoprenol (3-methylbut-3-en-1-ol) (Baral et al., 2021; Kennedy et al., 2019). In this regard, isoprenol has emerged as an alternative biological DMCO precursor due to its superior stability and ease of microbial production. While robust isoprenol production from engineered strains of Escherichia coli (Kang et al., 2019), Saccharomyces cerevisiae (Kim et al., 2021), Corynebacterium glutamicum (Sasaki et al., 2019), and Pseudomonas putida (Banerjee et al., 2024; Wang et al., 2022) supports its potential as an advanced aviation fuel precursor over isoprene, isoprenol is toxic to the production hosts at high concentrations and remains challenging for downstream recovery from the production culture medium.

Isoprenyl acetate (3-methyl-3-buten-1-yl acetate), an acetate ester derived from isoprenol, exhibits favorable physicochemical properties that make it a more suitable biosynthetic precursor of DMCO (Table S1). Its lower aqueous solubility compared to isoprenol, attributed to the ester functional group, reduces product toxicity during fermentation, facilitates in situ recovery, and enhances downstream separation efficiency (Wilbanks and Trinh, 2017). Additionally, isoprenyl acetate itself possesses desirable fuel properties, including a higher derived research octane number (dRON) of 108.4, in contrast to 95 for isoprenol (Carruthers et al., 2023). Its synthesis is accomplished in a single enzymatic step catalyzed by an alcohol acyltransferase (AAT), which condenses the isoprenol with the high-energy thioester bonds in acyl-CoA intermediates. Among various AAT–substrate pairings, isoprenyl acetate production has been demonstrated in E. coli through the heterologous expression of ATF1, an AAT from S. cerevisiae (Carruthers et al., 2023). This system achieved a high titer of 28 g/L under fed-batch conditions in a 2 L bioreactor, showcasing its scalability and potential for industrial application.

P. putida presents practical advantages over E. coli as a production host for biofuels and bio-based compounds owing to its metabolic versatility (Martin-Pascual et al., 2021, Nikel and de Lorenzo, 2018). Together with the development of advanced synthetic biology tools such as genome-scale metabolic modeling (GSMM) (Banerjee et al., 2024), CRISPR-based approaches (Czajka et al., 2022), and Random Barcode Transposon Sequencing (RB-TnSeq) (Borchert et al., 2024), P. putida has become a strong candidate for large-scale bioprocesses (Weimer et al., 2020). It is particularly well-suited for valorizing lignocellulosic biomass, as it can thrive on recalcitrant and inhibitory compounds commonly found in lignocellulosic hydrolysates, including furfural and phenolic derivatives (Lechtenberg et al., 2024). In addition, its native ability to metabolize lignin-derived aromatics such as p-coumarate, ferulate, and benzoate (Banerjee et al., 2025; Johnson et al., 2019; Weiland et al., 2022) further supports its utility in lignocellulosic bioconversion.

Despite its strengths, P. putida cannot natively metabolize xylose, the second most abundant sugar in lignocellulosic hydrolysates, typically present in a 1:2 ratio with glucose. To address this, three distinct xylose utilization pathways have been explored in P. putida: the Weimberg pathway, the Dahms pathway, and the xylose isomerase pathway (Bator et al., 2019). The Weimberg pathway oxidizes xylose to α-ketoglutaric semialdehyde for entry into the tricarboxylic acid (TCA) cycle (Stephens et al., 2007), and its efficiency has been improved through adaptive laboratory evolution (ALE) (Lim et al., 2021). The Dahms pathway diverges from the Weimberg intermediate to yield pyruvate and 2-phosphoglycerate for glycolysis (Dahms, 1974), but its application in P. putida remains limited due to its complexity. Notably, the xylose isomerase pathway converts xylose into xylulose-5-phosphate, which enters the pentose phosphate (PP) pathway. This pathway stands out for its simplicity, energy efficiency, and a carbon-efficient, non-oxidative entry of xylose into central metabolism, and it has been leveraged to produce compounds such as muconate (Ling et al., 2022), itaconate (Elmore et al., 2021), and medium-chain-length polyhydroxyalkanoate (mcl-PHA) (Le Meur et al., 2012) in P. putida. While these pathways enable xylose assimilation, carbon catabolite repression (CCR) remains a major barrier to efficient sugar co-utilization. Specifically, global regulators like Crc, which modulates carbon source preference at the translational level, and HexR, a repressor of glucose transport and the Entner-Doudoroff (ED) pathway, play key roles in sugar uptake regulation (Dvořák et al., 2024; Moreno et al., 2009). Alleviating these regulatory constraints could further improve xylose assimilation and overall metabolic efficiency.

A second significant challenge for isoprenyl acetate production in P. putida is a limited acetyl-CoA availability, which directly affects esterification efficiency and yield (Kozaeva et al., 2021; Ku et al., 2020). However, intracellular acetyl-CoA levels in P. putida are substantially lower than those in E. coli, with reported values under 0.05 μmol/gCDW (Kozaeva et al., 2021), whereas E. coli exhibits levels ranging from 0.25 to 1.5 μmol/gCDW during the exponential phase (Chohnan et al., 1998; Kudo et al., 2023; Ogata and Chohnan, 2015). Although carbon flux from pyruvate to acetyl-CoA is comparable between the two species (107–112 mmol/gCDW/h) (Kohlstedt and Wittmann, 2019; Maeda et al., 2016), the smaller pool in P. putida may arise from its evolutionary adaptation to diverse environments. Under carbon excess conditions, P. putida redirects acetyl-CoA to mcl-PHA accumulation as an energy reserve, which accelerates its turnover and lowers steady-state levels (Prieto et al., 2016). To expand the acetyl-CoA pool, several competitive pathways have been disrupted, including the glyoxylate shunt (aceA and glcB), fatty acid synthesis (accA and accC), and the TCA cycle branches (gltA and maeB) (Dvořák et al., 2024; Kozaeva et al., 2021; Thompson et al., 2020). Complementary strategies to boost acetyl-CoA generation have been explored through overexpression of rpe (ribulose-5-phosphate 3-epimerase, Rpe) and acs (acetyl-CoA synthetase, Acs), as well as introduction of xfpk (phosphoketolase, Xfpk) to improve carbon efficiency (Wirth et al., 2022; S. Yang et al., 2019). Together, these strategies help increase the acetyl-CoA pool and offer promising avenues to improve downstream biosynthetic efficiency.

In this study, we leverage these insights from previous studies to employ a multifaceted approach that enables high-titer production of isoprenyl acetate from glucose and xylose in P. putida (Fig. 1 and Table 1). We began by screening a panel of AAT candidates and identified the S. cerevisiae-derived ATF1 as the most effective AAT for isoprenyl acetate production. Next, we minimized product degradation by disrupting endogenous esterases (PP_1127, PP_3812, and PP_4218) involved in the hydrolysis of the ester product. For efficient utilization of mixed sugars, we introduced the xylose isomerase pathway and removed global regulatory repressors Crc and HexR to improve sugar uptake. Additionally, we expanded the intracellular acetyl-CoA pool by overexpressing acs and xfpk. Combined with optimization of cultivation parameters, these efforts led to the construction of the PIPAxyl-E3-K3-O15 strain, which produced 1.5 g/L of isoprenyl acetate in shake flask culture. This strain converted 28.0 g/L glucose and 12.5 g/L xylose into 1.9 g/L isoprenyl acetate in a 2 L bioreactor, corresponding to a yield of 0.067 g/g and 18.1% of the theoretical maximum. To our knowledge, this is the first demonstration of isoprenyl acetate production in P. putida, underscoring its potential as a robust microbial host for efficient biosynthesis of value-added esters.Fig. 1. Overall scheme of strategies employed to modify P. putida metabolism for the production of isoprenyl acetate from glucose and xylose. Gene names are displayed only for targets modified during our engineering efforts. Gray: Seven gene knockouts and the IPP-bypass pathway previously established; Red: Alcohol acyltransferase for isoprenyl acetate production; Yellow: Knockouts of esterases involved in isoprenyl acetate hydrolysis; Green: Integration of the xylose utilization pathway and disruption of regulators; Blue: Overexpression of genes to increase the acetyl-CoA pool. Isoprenyl acetate is converted to DMCO through chemical synthesis. Gene knockouts are indicated with "X" marks. Abbreviations: G6P, D-glucose-6-phosphate; 6PG, 6-phosphogluconate; 2KDPG, 2-dehydro-3-deoxy-6-phosphogluconate; G3P, D-glyceraldehyde 3-phosphate; FBP, fructose-1,6-bisphosphate; F6P, fructose-6-phosphate; PYR, pyruvate; Ru5P, D-ribulose-5-phosphate; R5P, ribose-5-phosphate; Xu5P, xylulose-5-phosphate; S7P, sedoheptulose-7-phosphate; E4P, erythrose-4-phosphate; CIT, citrate; ICIT, isocitrate; αKG, α-ketoglutarate; SUC, succinate; FUM, fumarate; MAL, malate; OAA, oxaloacetate; GLX, glyoxylate; CoA, coenzyme A; AAc-CoA, acetoacetyl-CoA; HMG-CoA, 3-hydroxy-3-methylglutaryl-CoA; MVA, mevalonic acid; MVAP, mevalonate-5-phosphate; IP, isopentenyl monophosphate; LAC, lactate; AcAc, acetoacetate; 3HB, 3-hydroxybutanoate; 3HACoA, 3-hydroxyacyl-CoA; PHA, polyhydroxyalkanoate; gcd, glucose dehydrogenase; xylE, D-xylose-H^+^ symporter; xylA, D-xylose isomerase; xylB, xylulokinase; tkt, transketolase; tal, transaldolase; xfpk, phosphoketolase; pta, phosphate acetyltransferase; acs, acetyl-CoA synthetase; MvaE, HMG-CoA reductase; MvaS, HMG-CoA synthase; MK, mevalonate kinase; PMD, phosphomevalonate decarboxylase; aphA, aminoglycoside 3′-phosphotransferase; ATF1, alcohol O-acetyltransferase; ldhA, D-lactate dehydrogenase; mvaB, hydroxymethylglutaryl-CoA lyase; hbdH, 3-hydroxybutyrate dehydrogenase; phaA, β-ketothiolase; phaB, acetoacetyl-CoA reductase; phaC, polyhydroxyalkanoate synthase; crc, catabolite repression control protein; hexR, transcriptional repressor; EMP, Embden-Meyerhof-Parnas; ED, Entner-Doudoroff; TCA, Tricarboxylic acid cycle; DMCO, 1,4-dimethylcyclooctaneFig. 1Table 1Key engineered strains and plasmids used in this study.Table 1. StrainsNameDescriptionKey features or notesTiter (mg/L); culture conditionsaPIPAP. putida KT2440 ΔphaABC (PP_5003-5005) ΔmvaB (PP_3540) ΔhbdH (PP_3073) ΔldhA (PP_1649) Δ86kb (4,536,184–4,627,926; PP_4023-PP_4092)

• •Competing routes deleted to limit byproducts
• •Strong supply of isoprenol precursors -PIPA-AAT1PIPA pIY670 pAAT1
• •Expression of ATF1 enabled isoprenyl acetate production 414; M9, Glc 20 g/L, Ara 2 g/L SA 125 μM at OD_600_ 0.6–0.8, Dod 20% (v/v)PIPA-E3PIPA ΔPP_1127 ΔPP_3812 ΔPP_4218 pIY670 pAAT1
• •Triple esterase knockout decreased hydrolysis
• •Improved stability of isoprenyl acetate 724; M9, Glc 20 g/L, Ara 2 g/L SA 125 μM at OD_600_ 0.6–0.8; Dod 20% (v/v)PIPAxylPIPA Δgcd (PP_1444):P_xylE_∗-xylE:P_tac_-xylAB:talB:tktA
• •Introduction of the xylose isomerase pathway
• •Enabled co-utilization of glucose and xylose -PIPAxyl-AAT1PIPAxyl pIY670 pAAT1
• •Xylose-utilizing strain expressing ATF1
• •Produced isoprenyl acetate from mixed sugars 181; M9, Glc/Xyl 10/10 g/L, (NH_4_)2_SO_4 40 mM, Ara 2 g/L SA 125 μM at OD_600_ 0.6–0.8; Dod 20% (v/v)PIPAxyl-E3PIPAxyl ΔPP_1127 ΔPP_3812 ΔPP_4218 pIY670 pAAT1
• •Xylose utilization plus esterase deletions
• •Reduced product degradation during mixed-sugar cultivation 383; M9, Glc/Xyl 10/10 g/L, (NH_4_)2_SO_4 40 mM, Ara 2 g/L SA 125 μM at OD_600_ 0.6–0.8; Dod 20% (v/v)PIPAxyl-E3-K3PIPAxyl ΔPP_1127 ΔPP_3812 ΔPP_4218Δcrc (PP_5292) ΔhexR (PP_1021) pIY670 pAAT1
• •Deletion of global repressors crc and hexR
• •Improved sugar uptake and increased ester titers 599; M9, Glc/Xyl 10/10 g/L + pulse, (NH_4_)2_SO_4 40 mM, Ara 2 g/L SA 125 μM at OD_600_ 0.6–0.8; Dod 20% (v/v)PIPAxyl-O2PIPAxyl pIY670 pO2
• •Overexpression of rpe
• •Improved isoprenyl acetate production -PIPAxyl-O6PIPAxyl pIY670 pO6
• •Overexpression of E. coli acs
• •Improved isoprenyl acetate production -PIPAxyl-O7PIPAxyl pIY670 pO7
• •Overexpression of Bifidobacterium xfpk
• •Improved isoprenyl acetate production -PIPAxyl-E3-O14PIPAxyl ΔPP_1127 ΔPP_3812 ΔPP_4218ΔampC (PP_2876):P_m_-acsEc-xfpkBa pIY670 pAAT1
• •Chromosomal integration of acsEc-xfpkBa operon
• •Improved isoprenyl acetate production -PIPAxyl-E3-O15PIPAxyl ΔPP_1127 ΔPP_3812 ΔPP_4218ΔampC (PP_2876):P_m_-acsEc-xfpk_Ba__RBS1 pIY670 pAAT1
• •RBS-optimized acsEc-xfpkBa operon integration
• •Highest flask-level titers among integration variants 819; M9-MOPS + YE, Glc/Xyl 10/10 g/L + pulse, (NH_4_)2_SO_4 40 mM, Ara 2 g/L at 0 h, SA 62.5 μM 3 MB 250 μM at OD_600_ 0.6–0.8; Dur 20% (v/v)PIPAxyl-E3-K3-O15PIPAxyl ΔPP_1127 ΔPP_3812 ΔPP_4218 Δcrc (PP_5292) ΔhexR (PP_1021) ΔampC (PP_2876):P_m_-acsEc-xfpk_Ba__RBS1 pIY670 pAAT1
• •Final consolidated strain
• •Robust and scalable isoprenyl acetate production 1500; M9-MOPS + YE, Glc/Xyl 13/7 g/L + pulse, (NH_4_)2_SO_4 40 mM, Ara 2 g/L at 0 h, SA 62.5 μM 3 MB 250 μM at OD_600_ 0.6–0.8; Dur 20% (v/v)PlasmidsNameDescriptionKey features or notesReferencespIY670pRK2-Kan-araC-P_BAD_-MvaS_ef_-MvaE_ef_-T_rpoH_-P_trc1-O_-MK_mm_-PMD_HKQ_-AphA
• •Supplies isoprenol through expression of a heterologous IPP-bypass MVA pathway Banerjee et al. (2024)pAAT1pRSF1010-Gm-NagR-P_nagAa_-ATF1-rrnB T1
• •Expresses ATF1 to catalyze esterification of isoprenol
• •Produces isoprenyl acetate from isoprenol and acetyl-CoA Carruthers et al. (2023)pO2pAAT1-P_J23107_-rpe-thrLABC
• •Introduces rpe
• •Improves PP pathway flux pO6pAAT1-P_J23107_-acsEc-thrLABC
• •Introduces E. coli acs
• •Enhances acetate assimilation pO7pAAT1-P_J23107_-xfpkBa-thrLABC
• •Expresses Bifidobacterium xfpk
• •Establishes a phosphoketolase shortcut to improve carbon yield toward acetyl-CoA aTiters represent the highest isoprenyl acetate concentrations obtained in flask cultures for each strain under the indicated conditions and correspond to the flask data points shown in Fig. 7c. Glc, glucose; Xyl, xylose; Ara, arabinose; SA, salicylic acid; 3 MB, 3-methylbenzoate; YE, yeast extract; Dod, dodecane; Dur, Durasyn 164. “Pulse” indicates that at 48 h an additional feed of glucose, xylose, and (NH_4_)2_SO_4 was supplied at the same concentrations as at 0 h to maintain the initial C:N ratio.

Materials and methods

2

Strains and plasmid construction

2.1

The key strains and plasmids are listed in Table 1. Details of all strains and plasmids used in this study are provided in Table S2. These strains and plasmids, along with relevant information, have been deposited in the public domain through the JBEI Registry (https://public-registry.jbei.org) and are available upon request.

Plasmid construction was performed using the NEBuilder HiFi DNA Assembly Master Mix (New England Biolabs (NEB), Ipswich, MA, USA) following the manufacturer's protocol. Overlapping regions of 30 bp were designed for efficient assembly. All PCR amplifications were performed using Q5 High-Fidelity DNA Polymerase (NEB) according to the supplier's instructions. Synthesized oligos (Integrated DNA Technologies, Coralville, IA) are listed in Table S3. E. coli XL1-Blue was used as the cloning host. Plasmid DNA extraction was performed using the QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany). For genotyping, colony PCR was performed using OneTaq™ 2X Master Mix (NEB), and sequencing was performed by GENEWIZ (South San Francisco, CA, USA). Genomic integration was performed through homologous recombination of kanamycin-sucrose counterselection (Marx, 2008). Gene knockouts were generated using a catalytically inactive Cpf1 endonuclease (Cpf1-D917N) from Francisella novicida U112 along with a corresponding sgRNA expression cassette (Czajka et al., 2022). Genomic DNA was extracted using the PureLink™ Genomic DNA Mini Kit (Thermo Fisher Scientific, Waltham, MA).

S. cerevisiae-derived ATF1 [NCBI: NP_015022.3] and ATF2 [NCBI: NP_011693.1] were amplified from JBEI-136483 and JBEI-136484, respectively (Carruthers et al., 2023). The SAAT [GenBank: AAG13130.1] from Fragaria × ananassa, codon-optimized for expression in E. coli, was obtained from JBEI-231873 (Carruthers et al., 2023). The R. hybrida-derived AAT [GenBank: AY850287.1] and E. coli-derived CAT [UniProtKB: P00484], carrying the Y20F mutation, were synthesized with P. putida-optimized codons by Twist Biosciences (Seo et al., 2021; Wang et al., 2022). Additionally, codon-optimized xfpk [UniProtKB: A1A185], derived from Bifidobacterium adolescentis, was obtained from the JBEI registry.

Reagents and growth medium

2.2

Reagents for cell culture were obtained from BD Biosciences (Sparks, MD, USA). All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA) unless specified otherwise.

P. putida strains were cultured in either LB medium, M9, modified M9, or M9-MOPS. M9 medium was prepared with the following components: 2 g/L (NH_4_)2_SO_4, 6.8 g/L Na_2_HPO_4_, 3 g/L KH_2_PO_4_, 0.5 g/L NaCl, 1 mL/L trace element solution (Teknova, Hollister, CA), 0.1 mM CaCl_2_, and 2 mM MgSO_4_. For experiments requiring modified nitrogen levels, the concentration of (NH_4_)2_SO_4 was increased to 40 mM and is referred to as modified M9. M9-MOPS was prepared with the following components: M9 salts (6.78 g/L Na_2_HPO_4_, 3 g/L KH_2_PO_4_, 1 g/L NH_4_Cl, and 0.5 g/L NaCl), 75 mM 3-morpholinopropane-1-sulfonic acid (MOPS), 1 mg/L thiamine, 10 nM FeSO_4_, micronutrients (3∗10^−8^ M (NH_4_)6_Mo_7_O_24, 4∗10^−6^ M boric acid, 3∗10^−7^ M CoCl_2_, 1.5∗10^−7^ M CuSO_4_, 8∗10^−7^ M MnCl_2_, and 1∗10^−7^ M ZnSO_4_), 2 mM MgSO_4_, and 0.1 mM CaCl_2_. If stated, 1 g/L or 5 g/L of yeast extract was added. Unless otherwise noted, cultures contained 20 g/L total sugar (either glucose alone or a 2:1 glucose:xylose mixture); conditions with 40 g/L total sugar or a 1:1 glucose:xylose ratio are specified where used. Where applicable, kanamycin sulfate (50 μg/mL) and gentamicin sulfate (30 μg/mL) were added to the growth media for antibiotic selection. Where indicated, 100 g/L L-arabinose (Ara), 0.1 M salicylic acid (SA), or 0.5 M m-toluic acid (3-methylbenzoate, 3 MB) stocks were added to reach the final concentrations specified for each experiment.

Isoprenyl acetate hydrolysis test

2.3

Isoprenyl acetate hydrolysis was evaluated in two stages: first, basal degradation was assessed using wild-type P. putida KT2440 to determine background hydrolysis; second, candidate esterases were screened in the engineered PIPA strain, a metabolically streamlined P. putida variant optimized for isoprenol production, to identify esterases driving hydrolysis (Banerjee et al., 2024).

First, initial degradation tests were performed using wild-type P. putida KT2440. A single colony was inoculated into 5 mL of LB medium in 50 mL culture tubes and grown overnight. For adaptation to minimal media, cultures were diluted 1:50 into 5 mL of M9 minimal medium in the same 50 mL culture tubes and cultured overnight. This adaptation process was repeated once more. The adapted cells were used in a hydrolysis assay by inoculating them into 5 mL of fresh M9 medium supplemented with varying concentrations of isoprenyl acetate (0–2 g/L) in 50 mL culture tubes. After 24 h, samples were collected for the analysis of isoprenyl acetate and isoprenol levels. M9 medium without isoprenyl acetate was used as a blank control.

Next, for esterase identification, the PIPA strain was transformed with plasmids pD1–pD17 and pU1–pU17, yielding PIPA-D1 to PIPA-D17 and PIPA-U1 to PIPA-U17 strains, respectively. These strains underwent the same two-step adaptation process in M9 minimal medium in 50 mL culture tubes with a 5 mL working volume as described above. Following the second adaptation, cells were transferred into 5 mL of M9 medium in 50 mL culture tubes and incubated overnight to accumulate sufficient biomass for the hydrolysis assays. The resulting cultures were then resuspended in 5 mL of fresh M9 medium containing 1 g/L isoprenyl acetate, incubated in 50 mL conical tubes, and tightly sealed to minimize evaporation. After 24 h, samples were collected to measure OD_600_, isoprenyl acetate, isoprenol, and proteomic profiles. All assays were performed at 30 °C with shaking at 200 rpm.

Isoprenyl acetate-supplemented M9 medium was prepared by adding the ester directly to 50 mL of M9 medium in conical tubes. Tubes were immediately sealed to prevent evaporation, shaken vigorously, wrapped in aluminum foil to prevent light exposure, and left undisturbed overnight at room temperature before use.

Isoprenyl acetate production test

2.4

All P. putida strains were inoculated from freshly transformed colonies. In each case, a single colony was inoculated into 5 mL of LB medium with appropriate antibiotics in 50 mL culture tubes and grown overnight at 30 °C with shaking at 200 rpm. For adaptation to minimal medium conditions, the overnight culture was transferred to 5 mL of M9 minimal medium at an initial OD_600_ of 0.1 in 50 mL culture tubes and incubated overnight. This step was repeated once more under identical conditions for a second adaptation. The final culture was then initiated at an OD_600_ of 0.2 for experimental testing. Tube cultures were conducted in 5 mL of medium in 50 mL culture tubes, with samples collected at 48 h, while flask cultures were performed in 50 mL of medium in 250 mL unbaffled shake flasks, with sampling every 24 h. At an OD_600_ of 0.6–0.8, cultures were induced, and a 20% (v/v) dodecane overlay was added.

Gas chromatography analysis for isoprenol and isoprenyl acetate

2.5

To analyze isoprenyl acetate and isoprenol concentrations, samples were processed using gas chromatography-flame ionization detection (GC-FID) (Thermo Focus GC, Thermo Scientific) equipped with a DB-WAX column (15 m length, 0.32 mm inner diameter, 0.25 μm film thickness, Agilent, USA).

For overlay phase quantification, 400 μL of organic phase was separated by centrifugation at 13,000 g for 5 min. Prior to GC analysis, samples were diluted 1:100 in ethyl acetate containing 1-butanol (30 mg/L) as an internal standard to reach a final volume of 1 mL, then transferred to an amber vial.

For aqueous phase quantification, 400 μL of culture supernatant was extracted by mixing it in a 1:1 ratio with ethyl acetate containing the same internal standard. The mixture was vortexed at 3000 rpm for 10 min and centrifuged at 13,000 g for 5 min to separate the organic and aqueous phases. A 200 μL of the ethyl acetate fraction was carefully transferred to an amber vial with a fused insert. All GC-FID runs were completed by injection of 1 μL analyte.

The GC oven temperature program was as follows: initial hold at 50 °C for 1 min, ramp at 15 °C/min to 100 °C, then increased at 30 °C/min to 230 °C, with a final hold at 230 °C for 1 min. The inlet temperature was maintained at 200 °C. Standard curves for quantification were established using serial dilutions of isoprenol and isoprenyl acetate.

Liquid chromatography analysis for glucose and xylose

2.6

Glucose and xylose concentrations were measured using high-performance liquid chromatography (HPLC) (Agilent 1200 Series, Agilent Technologies, USA) equipped with a refractive index detector (RID) and an Aminex HPX-87H column (Bio-Rad, USA). Prior to analysis, culture samples were filtered through a 0.22 μm syringe filter to remove particulates. A 4 mM sulfuric acid solution was used as the mobile phase in isocratic mode, with a flow rate of 0.6 mL/min. The column temperature was maintained at 65 °C, and the RID detector was set to 45 °C. Serial dilutions of glucose and xylose standards were used to determine the concentration of sugars in the samples. Data acquisition and analysis were performed using ChemStation software (Agilent Technologies).

Proteomics analysis

2.7

Protein was extracted from cell pellets and tryptic peptides were prepared by following established proteomic sample preparation protocol (Gin et al., 2023). Briefly, cell pellets were resuspended in Qiagen P2 Lysis Buffer (Qiagen, Germany) to promote cell lysis. Proteins were precipitated with addition of 1 mM NaCl and 4 × vol acetone, followed by two additional washes with 80% acetone in water. The recovered protein pellet was homogenized by pipetting mixing with 100 mM ammonium bicarbonate in 20% methanol. Protein concentration was determined by the DC protein assay (BioRad, USA). Protein reduction was accomplished using 5 mM tris2-(carboxyethyl)phosphine (TCEP) for 30 min at room temperature, and alkylation was performed with 10 mM iodoacetamide (IAM; final concentration) for 30 min at room temperature in the dark. Overnight digestion with trypsin was accomplished with a 1:50 trypsin:total protein ratio. The resulting peptide samples were analyzed on an Agilent 1290 UHPLC system coupled to a Thermo Scientific Orbitrap Exploris 480 mass spectrometer for discovery proteomics (Chen et al., 2022). Briefly, peptide samples were loaded onto an Ascentis® ES-C18 Column (Sigma–Aldrich, USA) and were eluted from the column by using a 10 min gradient from 98% solvent A (0.1 % FA in H2O) and 2% solvent B (0.1% FA in ACN) to 65% solvent A and 35% solvent B. Eluting peptides were introduced to the mass spectrometer operating in positive-ion mode and were measured in data-independent acquisition (DIA) mode with a duty cycle of 3 survey scans from m/z 380 to m/z 985 and 45 Tandem mass spectrometry (MS2) scans with precursor isolation width of 13.5 m/z to cover the mass range. DIA raw data files were analyzed by an integrated software suite DIA-NN (Demichev et al., 2020). The database used in the DIA-NN search (library-free mode) is P. putida KT2440 latest Uniprot proteome FASTA sequence plus the protein sequences of the heterologous proteins and common proteomic contaminants. DIA-NN determines mass tolerances automatically based on first pass analysis of the samples with automated determination of optimal mass accuracies. The retention time extraction window was determined individually for all MS runs analyzed via the automated optimization procedure implemented in DIA-NN. Protein inference was enabled, and the quantification strategy was set to Robust LC = High Accuracy. Output main DIA-NN reports were filtered with a global false discovery rate set at 0.01 (FDR ≤0.01) on both the precursor level and protein group level. The Top3 method, which is the average MS signal response of the three most intense tryptic peptides of each identified protein, was used to plot the quantity of the targeted proteins in the samples (Ahrné et al., 2013; Silva et al., 2006).

The generated mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD067010 (Perez-Riverol et al., 2022). DIA-NN is freely available for download from https://github.com/vdemichev/DiaNN.

Fed-batch cultivation

2.8

Fed-batch cultivation was conducted in a 2 L DASGIP parallel bioreactor system (Eppendorf) with a working volume of 1.2 L. Seed cultures were prepared through a three-step adaptation. A single colony was inoculated into 5 mL of LB medium in 50 mL culture tubes and grown overnight at 30 °C with shaking at 200 rpm. The culture was then transferred to 5 mL of modified M9-MOPS minimal medium in 50 mL culture tubes containing 20 g/L glucose at an initial OD_600_ of 0.1 and incubated for 24 h. This adaptation step was repeated once using modified M9-MOPS containing 13.3 g/L glucose and 6.7 g/L xylose, followed by a third adaptation in 100 mL of the same medium in 1 L unbaffled shake flasks for 8 h. The adapted culture was used to inoculate the bioreactor at an initial OD_600_ of 0.3.

The bioreactor was maintained at 30 °C with dissolved oxygen (DO) controlled at 15%. Initially, agitation and aeration were held at 400 rpm and 6 sL/min, respectively, corresponding to approximately 5 vvm for the 1.2 L working volume. DO was then maintained with a control cascade, first adjusting agitation from 400 to 800 rpm, then adjusting aeration from 6 sL/min to 40 sL/min, corresponding to approximately 5–33 vvm. The pH was held at 6.5 using 15% ammonium hydroxide. Antifoam B (Sigma-Aldrich) was added as needed to suppress foaming. Induction was performed at OD_600_ 0.6 by addition of 2 g/L Ara, 62.5 μM SA, and 250 μM 3 MB. To mitigate product loss, 20% (v/v) Durasyn 164 (Univar Solutions) was used as an overlay. An additional 200 mL of Durasyn was added to an off-gas trap connected to the vent line.

The feeding solution consisted of 266.7 g/L glucose and 133.3 g/L xylose and was prepared using the same basal composition as the modified M9-MOPS culture medium with 1 to 5 g/L yeast extract. Continuous feeding was implemented to maintain residual sugar concentrations below 5 g/L throughout the cultivation. Samples of 5 mL to 10 mL were collected approximately every 12 h for analysis of OD_600_, residual sugars by HPLC, and isoprenyl acetate levels by GC-FID. In addition, 1 mL samples were withdrawn from the off-gas trap every 24 h and analyzed by GC-FID.

Results

3

Production of isoprenyl acetate in P. putida

3.1

The P. putida strain used for isoprenol production was adapted from a previous study (Banerjee et al., 2024). Our base strain, PIPA, incorporated the GSMM-guided deletions of six genes (phaABC, mvaB, hbdH, and ldhA) from that study to redirect the metabolic flux towards isoprenol synthesis along with a deletion of the leucine degradation pathway to reduce native isoprenol degradation, which proceeds via initial oxidation and CoA activation before entering this pathway (Fig. 2a and Table 1) (Thompson et al., 2020). Additionally, five genes of the isopentenyl diphosphate (IPP)-bypass pathway (mvaS, mvaE, MKMm, PMDHKQ, and aphA) were expressed from a medium-copy plasmid (Kang et al., 2019). This bypass leverages the promiscuous activity of diphosphomevalonate decarboxylase (PMD) to convert mevalonate-5-phosphate directly to isopentenyl monophosphate (IP), avoiding ATP-dependent synthesis of IPP, which is a toxic intermediate to the production host (Kang et al., 2016). The resulting IP is then dephosphorylated to isoprenol by the promiscuous monophosphatase AphA from E. coli, thereby supplying the C5 precursor for isoprenyl acetate, while the native MEP pathway is retained for essential isoprenoid biosynthesis. In the previously reported engineered P. putida KT2440 background, this strain supported isoprenol titers of up to 762 mg/L in shake flasks and 3.5 g/L in fed-batch cultures (Banerjee et al., 2024).Fig. 2. Introduction of alcohol acetyltransferases (AATs) into P. putida and isoprenyl acetate production. (a) Pathway for isoprenyl acetate production from acetyl-CoA and the two plasmids used in this study. The isoprenol production plasmid, pIY670, was constructed in a previous study (Banerjee et al., 2024). Five AAT variants were introduced into the pRSF1010 plasmid and expressed under the P_nagA_ promoter. The pathway diagram illustrates the flow of acetyl-CoA, redox equivalents, ATP, and CO_2_. (b) Isoprenyl acetate and isoprenol titers in P. putida denoted by organism of origin and the name of each AAT enzyme. Tests were conducted in 5 mL culture tubes, and samples were analyzed after 48 h. Arabinose and salicylic acid were added at final concentrations of 2 g/L and 125 μM, respectively, when culture OD_600_ reached 0.6–0.8. Dodecane was used as an overlay at 20% (v/v). Error bars represent the standard deviation of biological triplicates.Fig. 2

We evaluated a panel of AAT enzymes to identify their compatibility with isoprenol. AATs are known for their broad substrate specificity toward diverse acyl-CoAs (e.g., acetyl-CoA, propionyl-CoA, and butyryl-CoA) and alcohols (e.g., ethanol, butanol, and benzyl alcohol) (Liu et al., 2023), and prior studies have screened natural variants and engineered their substrate scopes (Lin et al., 2016; Seo et al., 2021). Based on their reported activity toward short- to medium-chain alcohols (C2–C10), we selected five candidates: ATF1 and ATF2 from S. cerevisiae (Tai et al., 2015; Yuan et al., 2016), SAAT from Fragaria × ananassa (strawberry) (Lee and Trinh, 2022), AAT from Rosa hybrida (Wang et al., 2022), and CAT from E. coli (Seo et al., 2022) (Table S4). The CAT variant included a Y20F mutation, an engineered chloramphenicol acetyltransferase that shows enhanced activity toward a broad range of alcohol substrates (Seo et al., 2021). Each AAT was expressed from a high copy pRSF1010 plasmid under the salicylic acid (SA) inducible P_nagA_ promoter to ensure high expression without interfering with the pIY670 pathway plasmid (Fig. 2a).

Among the AAT variants tested, the PIPA-AAT1 strain (i.e., PIPA harboring pIY670 and pAAT1) expressing ATF1 from S. cerevisiae achieved the highest isoprenyl acetate production with a titer of 405 mg/L from 20 g/L glucose (Fig. 2b). This aligns with prior findings in E. coli (Carruthers et al., 2023). ATF1 exhibits notable activity toward branched-chain alcohols such as isobutanol and isoamyl alcohol (Tai et al., 2015; Yuan et al., 2016), and in this study, enabled over 50% conversion of isoprenol to isoprenyl acetate, underscoring its effectiveness for ester biosynthesis in P. putida.

Prevention of product degradation

3.2

P. putida is known for its robust aerobic metabolism that enables degradation of a wide variety of organic and inorganic compounds (Loh and Cao, 2008). Given this capability, we investigated the potential degradation of isoprenyl acetate by culturing P. putida for 24 h in a medium supplemented with 0–2 g/L isoprenyl acetate. Beyond the severe volatilization in the abiotic control, P. putida cultures showed an additional loss of 34 mg/L in isoprenyl acetate, accompanied by appreciable isoprenol (Fig. 3a). This hydrolytic activity represents one of the key bottlenecks in bioproduction of the ester.Fig. 3. Identification and knockout of putative esterases in P. putida. (a) Degradation of isoprenyl acetate in P. putida. (Left) Reaction scheme showing the hydrolysis of isoprenyl acetate into isoprenol and acetate, potentially mediated by promiscuous esterase activity. (Right) Isoprenyl acetate degradation test. Isoprenyl acetate was externally added at concentrations ranging from 0 to 2 g/L. Residual isoprenyl acetate and its hydrolyzed byproduct, isoprenol, were measured after incubation in a cell-free control or with P. putida KT2440. Cultures were grown in 5 mL culture tubes and analyzed after 24 h. (b) Identification of esterase activity using the CRISPR/dCas9 knockdown system and (c) plasmid-based overexpression. Gene names denote either the target of sgRNAs or the overexpressed gene on plasmid pRSF1010. (b, c) “Blank” indicates cell-free medium, and “Control” represents the PIPA-C harboring empty pRSF1010. Assays were conducted with 1 g/L isoprenyl acetate in sealed 50 mL conical tubes with 5 mL culture. Samples were analyzed after 24 h. (d) Production levels of isoprenyl acetate from engineered strains with knockout combinations after 48 h in 5 mL culture tubes. Each deleted esterase (E) is denoted with a number. (e) Time-course of isoprenyl acetate production by the control strain (PIPA-ATF1) and the triple esterase knockout strain (PIPA-E3). Experiments were conducted in 50 mL flasks containing M9 minimal medium and samples were collected every 24 h. (d, e) Arabinose and salicylic acid were added when the OD_600_ reached 0.6–0.8, at final concentrations of 2 g/L and 125 μM, respectively. Dodecane was used as an overlay at 20% (v/v). Error bars indicate the standard deviation of biological triplicates.Fig. 3

A total of 33 putative esterases were initially identified from the Biocyc database for candidate screening (Table S5) (Karp et al., 2019). From this list, phosphodiesterases and thioesterases were excluded due to substrate incompatibility, while lipases and other α/β-hydrolase fold enzymes were prioritized based on their known activity toward carboxylic esters (Ali et al., 2012). Following selection criteria outlined in a previous study (Lu et al., 2021), the list was refined to 16 candidate esterases for further investigation (Table S5).

First, we employed a CRISPR/dCas9 (or CRISPRi) system to assess the impact of individual esterase downregulation (Czajka et al., 2022; Yunus et al., 2024). The CRISPRi system is a useful tool for metabolic studies and pathway optimization as it allows for reversible, tunable, and targeted repression without permanent genetic modification. In our experimental setup, each esterase-targeting sgRNA and a non-targeting control sgRNA for gfp were individually cloned into a high copy plasmid pRSF1010 also harboring dCas9 to generate plasmids pD1–pD17 (Table S2).

We added 1 g/L of isoprenyl acetate to the medium, then cultured strains in tightly sealed conical tubes to minimize evaporation. After 24 h, most strains harboring sgRNA-dCas9 plasmids showed similar degradation levels to the PIPA-C, a control strain harboring empty pRSF1010 plasmid (Fig. 3b). Notably, the strain harboring an sgRNA targeting PP_3812 retained 83% of the initial isoprenyl acetate while reducing isoprenol accumulation, thereby demonstrating an effective suppression of ester hydrolysis. This result is consistent with previous studies showing that deletion of Est12 (PP_3812) reduced the degradation of hexyl acetate (Lu et al., 2021) and fatty alcohol acetate esters (Lu et al., 2023), supporting its broad substrate specificity and promiscuous esterase activity. To probe protein-level responses to CRISPR/dCas9-mediated repression, we performed label-free quantitative proteomics on the same culture samples used for the ester hydrolysis assay (Fig. 3b), as described in the Materials and methods section. Among the 16 targeted esterases, eight were detected, of which five showed successful knockdown (PP_0364, PP_1127, PP_4218, PP_4286, and PP_4337), while the remaining three were unaffected (PP_1296, PP_1343, and PP_4916) (Fig. S1). Even among the knocked-down targets, however, the efficiency of gene repression varied considerably. In particular, downregulation of PP_4854 significantly affected many protein levels compared to the control (Fig. S1a). These responses may result from dCas9-associated stress responses, downstream regulatory effects of target inhibition, or off-target activity of the sgRNA, although the exact cause remains unclear (Table S6). Despite these unresolved effects, the final OD_600_ values after 24 h were comparable to that of the PIPA-C control for all sgRNA-dCas9 strains, indicating that the CRISPR/dCas9 perturbations did not cause major growth defects under the ester hydrolysis assay conditions (Fig. S1e).

We further implemented a plasmid-based overexpression system to identify additional esterases potentially missed by the CRISPR/dCas9 approach. Each esterase gene was cloned into the pRSF1010 plasmid under the constitutive promoter P_J23100_ to ensure robust expression, generating plasmids pU1–pU17 (Table S2). In addition to PP_3812, we found that overexpression of PP_1127 and PP_4218 also resulted in complete hydrolysis of isoprenyl acetate, implicating these three enzymes as key contributors to ester degradation (Fig. 3c).

Based on these findings, PIPA derivatives with combinatorial knockouts of PP_1127, PP_3812, and PP_4218 were constructed to evaluate their effect on isoprenyl acetate stability under production conditions containing 20 g/L glucose. Initial 48 h tube assays in culture tubes showed that individual knockouts of each esterase increased isoprenyl acetate titers compared to the control strain PIPA-AAT1, with improvements of 1.3-, 1.9-, and 1.6-fold, respectively, and that both the double-knockout strain PIPA-E2.1 (PIPA ΔPP_1127 ΔPP_3812 harboring pIY670 and pAAT1) and the triple-knockout strain PIPA-E3 (PIPA ΔPP_1127 ΔPP_3812 ΔPP_4218 harboring pIY670 and pAAT1) achieved the highest endpoint titers (Fig. 3d). Because this experiment captured only a single 48 h snapshot, we next compared the strains in longer flask time-course cultivations. Under these conditions, the triple knockout strain PIPA-E3 demonstrated the most promising performance in M9 medium containing 20 g/L glucose, with titers increasing from 414 mg/L to 724 mg/L, a 1.7-fold increase (Fig. 3e, Table 1, and Fig. S2). These results confirm that the deletion of these three esterases improved product stability without compromising cell growth and motivated the choice of PIPA-E3 for subsequent characterization.

Tuning pathway expression and product recovery

3.3

In our IPP-bypass MVA pathway design, isoprenyl acetate production places a high demand on acetyl-CoA as three acetyl-CoA molecules are used for isoprenol synthesis and an additional molecule for the AAT-catalyzed esterification step (Fig. 2a). Regarding this, we focused on two key factors to optimize pathway performance: (1) inducer concentration and (2) induction timing. Additionally, we assessed extraction efficiency across various overlays to improve product recovery.

The pIY670 plasmid harbors the IPP-bypass pathway, with the upper module (mvaE and mvaS) under the P_BAD_, an Ara-inducible promoter and the lower module (MKMm, PMDHKQ, and aphA) expressed constitutively under the P_trc_ promoter (Banerjee et al., 2024) (Fig. 2a). In this system, we strongly induced the isoprenol biosynthetic pathway by adding 2 g/L Ara to ensure maximal precursor availability, whereas the ATF1 expression was modulated via SA. Under these conditions in M9 medium containing 20 g/L glucose, the PIPA-E1.2 strain produced the highest isoprenyl acetate titer of 583 mg/L with 250 μM SA (Fig. 4a). Proteomic analysis further revealed that the ATF1 abundance strongly correlated with production (R^2^ = 0.8323) (Fig. 4b). However, SA levels above 500 μM led to decreased titers, coinciding with downregulation of key pathway enzymes, including MvaE, MvaS, PMD_HKQ_, and AphA (Fig. S3). This reduction is likely due to the metabolic burden imposed by overexpression of the ATF1, a membrane-associated enzyme known to disrupt cellular homeostasis when expressed at high levels (Lin and Wheeldon, 2014), highlighting the need to fine-tune inducer concentration.Fig. 4. Optimization of induction and culture parameters. (a) Optimization of salicylic acid (SA) concentration. (b) Correlation between isoprenyl acetate titer and ATF1 protein abundance. A regression line is shown in red. (c) Optimization of the inducer addition timing. Arabinose (Ara) and SA were added at 4, 8, or 12 h after inoculation, corresponding to OD_600_ values of 0.5 (early exponential), 2.5 (mid-exponential), and 5.6 (late exponential/early stationary), respectively. (d) Evaluation of overlay solvents. All experiments were conducted using the PIPA-E1.2 strain in 5 mL culture tubes, with samples analyzed after 48 h. Overlays were used at 20% (v/v). Unless otherwise specified, Ara and SA were added when the OD_600_ reached 0.6–0.8, at final concentrations of 2 g/L and 125 μM, respectively, and dodecane was used as an overlay. Error bars indicate the standard deviation of biological triplicates.Fig. 4

We next optimized the timing of the SA addition. The optimal condition was achieved by adding Ara at inoculation and SA during the early exponential phase, allowing isoprenol to accumulate before ATF1 expression in M9 medium containing 20 g/L glucose (Fig. 4c). This minimized promiscuous activity of ATF1 on native P. putida alcohols (Lin et al., 2016; Nancolas et al., 2017). In contrast, induction during mid- or late exponential phase led to lower titers, consistent with previous findings that acetyl-CoA availability declines under nutrient-limited and stress conditions (Kozaeva et al., 2021).

The choice of overlay solvent can affect the extraction efficiency of isoprenyl acetate (due to the difference in partitioning of the products) and the oxygen-transfer characteristics during cultivation and collectively influence overall titers. We evaluated four solvents: dodecane (Peralta-Yahya et al., 2011), hexadecane (Layton and Trinh, 2016), oleyl alcohol (Carruthers et al., 2023), and Durasyn 164 polyalphaolefin (Table S7) (Adamczyk et al., 2025). Durasyn 164 polyalphaolefin (or Durasyn) is a synthetic hydrocarbon known for its chemical stability, low volatility, and low cost (Bagheri et al., 2013). Among these, dodecane yielded the highest titer at the culture tube scale in M9 medium containing 20 g/L glucose (Fig. 4d), while both dodecane and Durasyn supported high titers in flask-scale experiments (Fig. S4). For the flask cultures, cells were grown in phosphate-buffered M9 medium containing 20 g/L glucose, and a single pulse of 20 g/L glucose was added at 48 h to avoid carbon limitation during the extended cultivation and to compensate for periplasmic oxidation of glucose to gluconate; under these conditions, the PIPA-E1.2 strain produced 1.4 g/L isoprenyl acetate at 96 h when using Durasyn. Durasyn and hexadecane selectively capture isoprenyl acetate without co-extracting isoprenol, which is beneficial for downstream purity. In contrast, oleyl alcohol led to abnormal turbidity (OD_600_ > 50) or excessive emulsification, making it unsuitable for further use. Given its scalability and cost advantages, Durasyn is a promising solvent for industrial-scale applications, whereas dodecane remains the preferred choice for small-scale setups due to its ease of handling.

Introduction of a xylose utilization pathway

3.4

Biomass hydrolysates are typically composed of glucose and xylose as the major sugars, derived from various feedstocks. To replicate this composition, glucose and xylose are commonly added in a 2:1 ratio (Elmore et al., 2020; Ling et al., 2022) or in equimolar amounts (Dvořák and de Lorenzo, 2018; Elmore et al., 2020; Wang et al., 2019). Minor byproducts such as L-arabinose and acetate are included in trace amounts, generally less than 4% (Elmore et al., 2020). Among these components, we prioritized engineering a strain for efficient co-utilization of glucose and xylose.

Of the xylose assimilation pathways studied in P. putida, we selected the xylose isomerase pathway based on its distinct advantages: (1) minimal carbon loss during entry into central metabolism, (2) independence from redox regulation (Shen et al., 2020), and (3) a concise three-enzyme route that channels xylose into the PP pathway, potentially boosting acetyl-CoA pool (Fig. 5a). We introduced codon-optimized E. coli genes xylE, xylA, and xylB, encoding a xylose transporter, isomerase, and kinase, respectively. Additionally, codon-optimized E. coli tal and tkt genes, which encode a transaldolase and transketolase, respectively, were introduced to support flux through the PP pathway. To decouple the transport from the catabolism, xylE remained under its native promoter but with an ALE-identified single adenosine insertion at −10 region to boost the uptake (Elmore et al., 2020). Finally, the complete expression cassette (P_xylE∗-xylE:P_tac-xylAB:talB:tktA) was integrated into the gcd locus, disrupting glucose dehydrogenase to prevent conversion of xylose to xylonate, thereby yielding strain PIPAxyl from the parental strain PIPA (Table 1).Fig. 5. Engineering of PIPA for integration of the xylose isomerase pathway. (a) Metabolic engineering strategy. Xylose is metabolized through the heterologous xylose isomerase pathway, where xylA, xylB, and xylE convert xylose into xylulose-5-phosphate. tal and tkt are overexpressed within the same operon. The introduced pathway is highlighted in green. The gcd knockout is indicated by an “X”. Additionally, crc and hexR were deleted, and enzymes under their regulatory control involved in glucose and xylose utilization are indicated with dotted light green lines. These include a subset of over 134 enzymes, several of which are involved in glucose and xylose utilization. (b) Optimization of adaptation media and C:N ratio. The PIPAxyl-AAT1 strain was cultured in 5 mL tubes, and samples were analyzed after 48 h. (c) Isoprenyl acetate production in glucose-xylose mixtures. The left panel shows results from PIPAxyl-AAT1, and the right panel from PIPAxyl-E3. Here, experiments were conducted in 50 mL flasks, with samples collected every 24 h. Conditions for adaptation media and C:N ratio were based on prior optimization, using a glucose–xylose mixture and 40 mM (NH_4_)2_SO_4. (b, c) Arabinose and salicylic acid were added when the OD_600_ reached 0.6–0.8, at final concentrations of 2 g/L and 125 μM, respectively. Dodecane (20%, v/v) was used as an overlay. Error bars indicate the standard deviation of biological triplicates.Abbreviations: oprB, carbohydrate-selective porin; PP_1015, mannose/glucose ABC transporter substrate-binding protein; PP_1016, ABC transporter permease; PP_1017, mannose/glucose ABC transporter permease; PP_1018, mannose/glucose ABC transporter ATP binding protein; glk, glucokinase; zwf, glucose 6-phosphate-1-dehydrogenase; pgl, 6-phosphogluconolactonase; edd, phosphogluconate dehydratase; eda, KHG/KDPG aldolase; gapA, glyceraldehyde-3-phosphate dehydrogenase.Fig. 5

Isoprenyl acetate production by the PIPAxyl-AAT1 strain (PIPAxyl harboring pIY670 and pAAT1) was assessed in M9 medium containing 10 g/L glucose and 10 g/L xylose. However, the resulting titer was only 25 mg/L (Fig. 5b), significantly lower than the 405 mg/L produced by PIPA-AAT1 with glucose alone (Fig. 2b). To address this limitation, we modified both the preculture medium and the C:N ratio to enhance sugar adaptation and support amino acid and nucleotide biosynthesis in P. putida (Dvořák and de Lorenzo, 2018; Schmidt et al., 2022). Notably, the highest titer of 198 mg/L was achieved when both sugars were included during adaptation and 40 mM (NH_4_)2_SO_4 was supplied, corresponding to a C:N molar ratio of 8.33:1 (Fig. 5b) (Elmore et al., 2020; Kang et al., 2019). Under these optimized conditions, PIPAxyl-AAT1 simultaneously consumed 10 g/L glucose and 10 g/L xylose in flask cultures, with a modest tendency for xylose to be consumed more rapidly than glucose (Fig. 5c). This accelerated consumption likely resulted from combination of deleting gcd, which removes the major periplasmic glucose oxidation route, and strongly expressing the heterologous xylE–xylAB–talB–tktA cassette, collectively favoring xylose uptake and assimilation under our conditions (Nikel et al., 2015). As in the PIPA parent strain, significant isoprenyl acetate degradation was again observed due to esterase activity. To mitigate this, we constructed a triple esterase knockout strain, PIPAxyl-E3 (PIPAxyl ΔPP_1127 ΔPP_3812 ΔPP_4218 harboring pIY670 and pAAT1). This engineered strain achieved a titer of 383 mg/L, a 2.1-fold increase over PIPAxyl-AAT1, with minimal isoprenol accumulation (Fig. 5c and Table 1).

In P. putida, glucose uptake primarily occurs via ABC transporters rather than the phosphotransferase system (PTS), unlike in E. coli or Bacillus subtilis (Pause et al., 2024). Consequently, P. putida employs a unique CCR system governed by the Crc-Hfq complex and small RNAs such as CrcZ and CrcY (Rojo, 2010). Other factors, including cytochrome o ubiquinol oxidase and PTS proteins, may also be involved, though their roles remain unclear. Among known CCR-related global regulators, crc and hexR have been widely targeted for knockout to mitigate potential CCR and improve mixed utilization of glucose, xylose, and aromatic acids (Elmore et al., 2020; Johnson et al., 2017; Shrestha et al., 2023) (Fig. 6a). Crc, a global translational regulator functioning with Hfq, controls at least 134 genes to optimize metabolic balance across different growth stages (Moreno et al., 2009). Another key regulator, HexR, represses genes involved in the Entner–Doudoroff (ED) pathway and glucose transport (Dvořák et al., 2024; Ling et al., 2022).Fig. 6. Engineering of PIPAxyl through CCR-related regulatory modifications. (a) Gene regulation by crc, hexR, and other regulatory elements. The Crc-Hfq complex regulates gene expression at the mRNA level, indicated on each respective gene with a green (−) triangle. HexR, GltR-II, and GlnG function as repressors (−) or activators (+) as marked at the promoters of their respective operons. (b) Isoprenyl acetate production in the PIPAxyl-E3-K3 strain. Experiments were conducted in 50 mL flasks, with samples collected every 24 h. Arabinose and salicylic acid were added when the OD_600_ reached 0.6–0.8, at final concentrations of 2 g/L and 125 μM, respectively. Dodecane (20%, v/v) was used as an overlay. Error bars indicate the standard deviation of biological triplicates. (c) Proteomic comparison of proteins regulated by crc and hexR, as annotated in Fig. 6a. Bars represent protein abundance in PIPAxyl-E3 and PIPAxyl-E3-K3 at representative time points for each protein. Numerical fold-change values are labeled above each bar. (d) Time-course proteomic analysis of IPP-bypass pathway enzymes. Each bar graph is labeled with the corresponding enzyme name.Fig. 6

Given their central roles in carbon regulation, we constructed single and double knockout strains targeting crc and hexR. Isoprenyl acetate production and sugar consumption were evaluated in M9 medium containing 10 g/L glucose and 10 g/L xylose in shake flasks, with a single pulse at 48 h consisting of 10 g/L glucose, 10 g/L xylose, and 40 mM (NH_4_)2_SO_4. The single crc knockout strain, PIPAxyl-E3-K1, showed markedly faster glucose consumption, with complete depletion by 96 h (Fig. S5). This aligns with previous findings that crc deletion accelerates glucose metabolism by upregulating glucose transport-related porins, ABC transporters, and key ED pathway genes (edd, eda, pgi, and gapA) (Moreno et al., 2009). The hexR knockout strain PIPAxyl-E3-K2 also showed an increased glucose consumption rate immediately after the 48 h pulse, together with more simultaneous co-consumption of xylose, although residual carbon remained at the end of cultivation and the improvement in isoprenyl acetate titer was modest compared with the crc/hexR double knockout (Fig. S5). Notably, the double knockout strain PIPAxyl-E3-K3 (PIPAxyl ΔPP_1127 ΔPP_3812 ΔPP_4218 Δcrc ΔhexR harboring pIY670 and pAAT1) achieved the highest isoprenyl acetate titer of 599 mg/L at 144 h (Fig. 6b and Table 1), representing a 1.3-fold increase over PIPAxyl-E3 (Fig. S5).

Proteomics revealed that the deletion of crc and hexR in PIPAxyl-E3-K3 broadly upregulated key proteins in central carbon metabolism between 1.6 and 4.3-fold compared to PIPAxyl-E3 (Fig. 6c). Importantly, pathway protein levels in the PIPAxyl-E3-K3 strain remained stable during the stationary and death phases, supporting continued production and contributing to the highest isoprenyl acetate titers observed (Fig. 6d). While these outcomes highlight the synergistic effect of crc and hexR deletion, it is worth noting that carbon metabolism in P. putida is inherently complex and mapping this metabolic landscape will be key to guiding further optimization of sugar co-utilization (Lim et al., 2022).

Rewiring native and synthetic routes to enhance acetyl-CoA flux

3.5

Acetyl-CoA is a key metabolite connecting glycolysis with the TCA cycle (Fig. 7a) and crucial for the bioproduction of IPP-bypass derived isoprenyl acetate. Diverse metabolic engineering strategies including amplification of native pathways, elimination of competing routes, and enhancement of coenzyme A availability have been developed to increase the biosynthesis of acetyl-CoA-derived compounds such as lipids, amino acids, and isoprenoids (Krivoruchko et al., 2015). Based on these prior engineering efforts in P. putida, we focused on overexpressing both native and heterologous genes to boost acetyl-CoA biosynthesis (Table S8). We overexpressed the native talB and tktA, which were already integrated into the PIPAxyl strain, as well as additional genes in the PP pathway, including gnd, rpe, rpiA, and glpX (Dvořák et al., 2024; Ling et al., 2022; Wirth et al., 2022) (Fig. 7a). Here, gnd encodes 6-phosphogluconate dehydrogenase, rpe and rpiA encode non-oxidative PP enzymes that interconvert pentose phosphates, and glpX encodes fructose-1,6-bisphosphatase 2, which regenerates fructose-6-phosphate from lower glycolytic intermediates to reinforce flux through the PP/upper glycolytic node from both glucose and xylose. Then, we introduced a heterologous acetyl-CoA synthetase (acs) from E. coli to recycle any acetate that might be transiently produced back into acetyl-CoA and a phosphoketolase (xfpk) from Bifidobacterium adolescentis to directly convert xylulose-5-phosphate or fructose-6-phosphate into acetyl-phosphate, thereby minimizing carbon loss via a non-oxidative route (Bogorad et al., 2013; J. Yang et al., 2019). Co-expression of pta from E. coli with xfpk was also explored to enable efficient conversion of acetyl-phosphate to acetyl-CoA (Bruinsma et al., 2023; Wirth et al., 2022).Fig. 7. Engineering strategies to enhance the acetyl-CoA pool. (a) Schematic representation of the engineered pathways. Genes selected for overexpression are highlighted in blue, and a summary of their functional descriptions is provided in Table S8. The offset diagram on the right illustrates key reactions catalyzed by Xfpk and enzymes in the pentose phosphate pathway. (b) Plasmid-based early screening of eight target genes (O1 to O8). Each gene was overexpressed (O) under the constitutive P_J23107_ promoter on pAAT1 and introduced into the PIPAxyl strain. A “-” indicates PIPAxyl-AAT1 control strain. Gene origins are indicated as subscripts: Ec for E. coli and Ba for B. adolescentis. Each additionally overexpressed (O) gene is denoted with a number. Experiments were conducted in 5 mL culture tubes and analyzed after 48 h. Arabinose (Ara) and salicylic acid (SA) were added at an OD_600_ of 0.6–0.8 to final concentrations of 2 g/L and 125 μM, respectively. Dodecane (20%, v/v) was used as an overlay. (c) Schematic overview of strain engineering steps derived from the PIPA strain. Maximum isoprenyl acetate titers achieved by each strain under its respective optimized culture conditions are shown in red. Values without parentheses were obtained in 5 mL tube cultures, and values in parentheses were obtained in shake flask cultures. The corresponding culture formats for each strain are summarized in Table 1. AAT, E, K, and O represent variant numbers for each engineering module. The final consolidated strain, PIPAxyl-E3-K3-O15, integrates all key modifications. (d) Cultivation profiles of the PIPAxyl-E3-O15 and PIPAxyl-E3-K3-O15 strains under optimized culture conditions. Cultures were grown in M9-MOPS medium supplemented with 1 g/L yeast extract. Induction was achieved by adding 2 g/L Ara at inoculation, followed by 62.5 μM SA and 250 μM m-toluic acid at OD_600_ of 0.6–0.8. Durasyn 164 (20%, v/v) was used as an overlay. Experiments were conducted in 50 mL flasks with sampling every 24 h. Error bars represent the standard deviation of biological triplicates.Abbreviations: acka, acetate kinase; pta, phosphate acetyltransferase; gnd, 6-phosphogluconate dehydrogenase; rpe, ribulose-5-phosphate 3-epimerase; rpiA, ribose-5-phosphate isomerase A; glpX, fructose-1,6-bisphosphatase 2.Fig. 7

For preliminary evaluation, overexpression candidates were individually cloned into the ATF1-carrying plasmid pAAT1 under the constitutive promoter P_J23107_, yielding constructs pO1 through pO8 (Table 1). Initial screening identified that overexpression of the native rpe as well as the heterologous acs from E. coli and xfpk from B. adolescentis (PIPAxyl-O2, O6, and O7) improved isoprenyl acetate production, attaining titers of 317 ± 24, 337 ± 11, and 292 ± 36 mg/L, respectively, up to 1.7-fold higher than that of the PIPAxyl-AAT1 control in M9 medium containing 10 g/L glucose and 10 g/L xylose (Fig. 7b).

To achieve stable and tunable expression, each overexpression candidate was chromosomally integrated at the ampC locus, which encodes a β-lactamase that provides a neutral landing pad for stable single-copy expression, into the esterase-deficient strain PIPAxyl-E3 (ΔPP_1127, ΔPP_3812, and ΔPP_4218). Expression was controlled by the tightly regulated, 3 MB-inducible P_m_ promoter, widely validated in P. putida strains (Martin-Pascual et al., 2021). All engineered strains outperformed the PIPAxyl-E3 control in M9 medium containing 10 g/L glucose and 10 g/L xylose (Fig. S6), with the most pronounced improvement in PIPAxyl-E3-O14, which co-expressed acs and xfpk in an operon and reached 500 mg/L of isoprenyl acetate. This marked increase reflects the synergistic contribution of two distinct acetyl-CoA-generating pathways that together increase the precursor availability (Wirth et al., 2022). To explore the expression variability, we employed five modular RBSs with experimentally validated strengths, each paired with standardized linkers from the iGEM Parts Registry (Fig. S7) (Storch et al., 2015; Yunus et al., 2024). Notably, the PIPAxyl-E3-O15 strain (PIPAxyl ΔPP_1127 ΔPP_3812 ΔPP_4218 ΔampC::P_m_-acs-xfpk_RBS1 harboring pIY670 and pAAT1), which harbors the weakest RBS, achieved the highest isoprenyl acetate titer of 660 mg/L in M9 medium containing 10 g/L glucose and 10 g/L xylose. This corresponds to a 2.2-fold increase over PIPAxyl-E3, consistent with an improved carbon allocation to product rather than diversion to competing sinks. Curiously, titers decreased with increasing RBS strength, except for a modest rebound at the highest level, highlighting the importance of precise translational tuning to balance pathway flux within polycistronic operons.

A summary of the key strain modifications implemented in this study is provided in Fig. 7c and Table 1. These modifications included (1) AAT introduction: the base strain PIPA was co-transformed with the isoprenol-producing plasmid (pIY670) and the isoprenyl acetate-producing plasmid (pAAT1) (PIPA-AAT1, Fig. 2b); (2) esterase knockouts: three native esterases (PP_1127, PP_3812, and PP_4218) were deleted to reduce product degradation (PIPA-E3, Fig. 3e); (3) xylose pathway engineering: a heterologous xylose isomerase pathway was integrated, along with knockouts of global sugar regulators (PIPAxyl-E3-K3, Fig. 6b); (4) acetyl-CoA flux enhancement: xfpk and acs were genomically integrated to redirect flux toward ester biosynthesis (PIPAxyl-E3-O15, Fig. S7).

Building on this foundation, we further optimized the production by testing different medium compositions (Fig. S8) and fine-tuning the concentrations of SA and 3 MB in the induction system (Fig. S9). Under these refined conditions along with previously established inducer timing and overlay compatibility (Fig. 4c and Fig. S4), the PIPAxyl-E3-O15 strain produced 819 mg/L of isoprenyl acetate in M9-MOPS medium with 10 g/L glucose, 10 g/L xylose, and 1 g/L yeast extract in shake flasks, with a single pulse at 48 h consisting of glucose, xylose, and (NH_4_)2_SO_4 added at the same concentrations as at 0 h (Fig. 7d). However, as observed previously, sugar consumption stalled at 96 h due to the absence of crc and hexR deletions (Fig. S5). Incorporation of both regulatory knockouts yielded the final strain, PIPAxyl-E3-K3-O15, which sustained mixed-sugar utilization and achieved the highest titer of 1.5 g/L at a 2:1 glucose-to-xylose ratio (Fig. 7d and Table 1). This improvement represents an 8.3-fold increase over the initial unoptimized PIPAxyl strain (Fig. 5c), highlighting the cumulative impact of the multi-layered metabolic pathway engineering and the process engineering to effectively maximize isoprenyl acetate production (Fig. 7c).

Isoprenyl acetate production in fed-batch cultivation

3.6

High titer ester production prompted evaluation in a 2 L bioreactor under fed-batch conditions to assess the production scalability. The initial sugar concentrations were set at 13.3 g/L glucose and 6.7 g/L xylose to mimic the composition of lignocellulosic hydrolysates previously reported (Elmore et al., 2020; Ling et al., 2022). To minimize evaporation losses, a 200 mL Durasyn overlay was added to the bioreactor in addition to connecting the vent line to an off-gas bottle containing an additional 200 mL of Durasyn for product capture. DO, pH, and temperature were maintained at 15%, 6.5, and 30 °C, respectively. A concentrated feed solution containing 266.7 g/L glucose and 133.3 g/L xylose was supplied at a rate of 0.5–1.0 mL/h and adjusted every 12 h based on sugar consumption profiles (Fig. 8a). Specifically, feeding was carefully controlled to keep residual sugar concentrations below 5 g/L. All other parameters followed previously optimized flask-scale conditions, with detailed reactor settings described in the Materials and Methods section.Fig. 8. Fed-batch cultivation of PIPAxyl-E3-K3-O15 for isoprenyl acetate production. (a) Residual sugar concentrations and feeding rate over time. The culture was initiated with 13.3 g/L glucose and 6.7 g/L xylose and continuously fed with a solution containing 266.7 g/L glucose and 133.3 g/L xylose. (b) Time-course profile of isoprenyl acetate production. Titers represent the sum of isoprenyl acetate quantified from the aqueous phase, the Durasyn overlay in the bioreactor, and the Durasyn off-gas trap. Cultivations were performed with 1 L medium and 200 mL overlay with sampling approximately every 12 h. Arabinose, salicylic acid, and m-toluic acid were added at OD_600_ 0.6–0.8 to final concentrations of 2 g/L, 62.5 μM, and 250 μM, respectively.Fig. 8

Under these fed-batch conditions, the PIPAxyl-E3-K3-O15 strain reached a final titer of 1.9 g/L isoprenyl acetate from a total feed of 28.0 g/L glucose and 12.5 g/L xylose, corresponding to a yield of 0.067 g/g and 18.1% of the theoretical maximum (Fig. 8b). The maximum OD_600_ was 13.2, lower than the 16.8 observed in flasks (Fig. 7d), due to tighter control of sugar availability and prevention of overflow metabolism. Glucose and xylose were co-consumed throughout cultivation without apparent CCR, consistent with the behavior of the crc/hexR knockout strain PIPAxyl-E3-K3 (Fig. 6b). However, xylose consumption was slower, presenting an opportunity for further optimization. In contrast to E. coli, which typically accumulates acetate as a byproduct (Carruthers et al., 2023), P. putida produced no detectable acetate. This reflects its inherent metabolic traits, notably the absence of acetate kinase (ackA), key enzyme in acetate overflow metabolism in facultative anaerobes (Weimer et al., 2024). Instead, P. putida utilizes acetyl-CoA synthetase (acs) to assimilate any available acetate. Although alternative pathways involving acetaldehyde dehydrogenase (aldB) and succinyl-CoA:acetate CoA-transferase (scpC) can generate minor amounts of acetate under carbon excess or stress (Henríquez et al., 2023; Turlin et al., 2023), these routes were likely inactive under the controlled cultivation conditions used in this study.

Mass balance analysis indicated that 1.35 mol of carbon input yielded 0.104 mol of isoprenyl acetate, 0.17 mol of biomass, and an estimated 0.178 mol of CO_2_. A substantial portion of the carbon remained unaccounted for, likely due to conversion into unknown metabolic byproducts or volatilization losses. Notably, 52% of the total isoprenyl acetate was recovered from the off-gas trap at the end of the 168-h cultivation, despite the inclusion of a 20% Durasyn overlay in the culture (Table S9). This observation underscores the high volatility of isoprenyl acetate under aerobic cultivation and points to the need for more effective in situ capture strategies to improve recovery and quantification. Nonetheless, this volatility will prove beneficial at scale for cost-effective separation during cultivation. Taken together, these findings represent the first demonstration of P. putida producing isoprenyl acetate from a mixed glucose-xylose substrate.

Discussion

4

Isoprenyl acetate is an acetate ester of isoprenol and also a direct biosynthetic precursor to DMCO, a high-energy-density synthetic aviation fuel (SAF) candidate. While the ester offers an advantage over isoprenol for downstream separation and recovery, microbial biosynthesis of esters remains technically challenging due to acetyltransferase promiscuity, native degradation pathways, precursor limitations, and the requirement for precise metabolic regulation. Beyond pathway construction, the use of lignocellulosic sugars introduces further complexity in achieving economically viable and scalable bio-based fuel production. Here, we present the first demonstration of isoprenyl acetate production from glucose and xylose in P. putida, reaching a titer of 1.9 g/L in fed-batch cultivation. This result marks a significant step toward the consolidated bioprocessing of isoprenoid esters from renewable feedstocks.

Several industrially relevant microbial hosts are known to degrade ester products due to promiscuous esterase activity. Previous studies have used E. coli, which does not show any detectable hydrolysis of target esters, as a neutral screening chassis and expressed genes from organisms such as P. putida and Clostridium thermocellum to identify their esterases (Fig. S10) (Lu et al., 2021, 2023; Seo et al., 2020). Here, rather than relying on conventional gene knockouts or heterologous gene expression screening, we used a CRISPRi system to directly and systematically repress native esterase genes in P. putida (Fig. 3b). CRISPRi enables conditional and tunable gene downregulation, which is particularly useful for studying essential genes (Fenster et al., 2022). Furthermore, its modularity supports multiplexed repression, enabling the high-throughput targeting of redundant enzyme families (Batianis et al., 2020). Despite its strengths, the CRISPRi approach also presented challenges (Fig. S1). Variable repression efficiency and inaccuracies in target prediction limited the consistent downregulation of intended genes (Calvo-Villamañán et al., 2020; Huang et al., 2021). These limitations complicate data interpretation and highlight the need for improved sgRNA design, optimized dCas9 expression systems, and engineered dCas9 variants with stronger transcriptional repression (Yu et al., 2024).

Despite extensive metabolic rewiring, complete conversion of isoprenol to isoprenyl acetate was not achieved throughout this study. Residual isoprenol consistently accumulated in the fermentation broth at concentrations of 50–100 mg/L, suggesting a bottleneck at the final esterification step. Several factors may underlie this incomplete conversion. First, residual activities of uncharacterized esterases may partially hydrolyze the product. In line with our screening results, minor contributors such as PP_0418 likely play a role and represent additional targets for future strain optimization. Definitive confirmation of this will require purified-enzyme assays, but it would be beyond the scope of this work, and we designate it as a future work. Second, the IPP-bypass MVA pathway used here imposes a high acetyl-CoA demand; three acetyl-CoA molecules are required to generate the isoprenol precursor, and a fourth is consumed as the acyl donor for isoprenyl acetate. As a result, limited intracellular acetyl-CoA availability could restrict flux toward ester formation, which may be addressed by eliminating competing native pathways or overexpressing type III pantothenate kinase to increase intracellular CoA availability (Kudo et al., 2023; Ogata and Chohnan, 2015). For instance, gltA has been downregulated or deleted, with essential citrate synthase function in some cases partially compensated by the substrate promiscuity of methyl citrate synthase PrpC (Kozaeva et al., 2021; Favoino et al., 2024; Dolan et al., 2022). GSMM could further guide identification of relevant targets (Banerjee et al., 2024). Third, the catalytic efficiency of ATF1 toward isoprenol may be suboptimal. Structure-guided engineering, directed evolution, or mining of novel AAT orthologs could help improve substrate specificity and kinetics (Seo et al., 2021).

Co-expression of xfpk and acs yielded the highest isoprenyl acetate titers among the acetyl-CoA–boosting modules tested (Fig. S6). In the phosphoketolase background, we hypothesize that Acs improves carbon recovery by re-assimilating low levels of acetate derived from acetyl-phosphate or overflow metabolism back into the acetyl-CoA pool. Although the mock hydrolysate in this study contained only glucose and xylose, future work could examine acetate co-feeding at levels typical of lignocellulosic hydrolysates to better exploit the Acs pathway and assess its effect on acetyl-CoA supply and isoprenyl acetate production. At the same time, expression of xfpk and acs resulted in a non-monotonic trend in isoprenyl acetate production, likely due to a stoichiometric imbalance arising from differences in their optimal expression levels. At intermediate expression, flux may be constrained by a rate-limiting enzyme, whereas high expression of both genes may alleviate this bottleneck. Independent tuning of gene expression within an operon is therefore essential for pathway optimization. RBS library screening offers one strategy, though operon-level factors such as gene order (Gerngross et al., 2022) and intergenic sequences affecting translational coupling (Tian and Salis, 2015) must be carefully considered. Genome engineering platforms such as serine recombinase-assisted genome engineering (SAGE), which support stable, markerless integration of multiple expression cassettes, are well-suited for high-throughput evaluation of such expression designs (Elmore et al., 2023). Alternatively, dual-level CRISPR-Cas systems enabling simultaneous control of transcription and translation provide a versatile approach to modulate gene expression within polycistronic operons (Cardiff et al., 2023). Quantification of acetyl-CoA pools and ^13^C-metabolic flux analysis would provide direct guidance for dissecting the xfpk–acs synergy and for further pathway optimization, but it will be probably out of scope of the current work.

Regarding sugar utilization, PIPAxyl-AAT1 displayed slightly faster depletion of xylose than glucose during flask co-fermentations (Fig. 5c). We regard this as a strain- and condition-specific behavior rather than a general shift in substrate preference and note that a more detailed kinetic and mechanistic analysis will be required in future work to fully explain this phenotype. In addition, incomplete consumption of both glucose and xylose was observed during pulsed co-fermentation (Fig. S5), likely reflecting nutrient limitations during prolonged cultivation together with altered carbon routing in the Δgcd background, where gcd was deleted to prevent periplasmic oxidation of xylose to xylonate (Fig. 5a). In wild-type P. putida, approximately 90% of periplasmic glucose is oxidized by Gcd to gluconate, most of which is subsequently metabolized through the ED pathway as 6-phosphogluconate (Nikel et al., 2015). The remaining ∼10% of glucose is taken up via the ATP-dependent ABC transporter and enters the ED pathway as glucose-6-phosphate (Dvořák and de Lorenzo, 2018; Weimer et al., 2020). Without Gcd, carbon flux is rerouted through the minor oxidative branch catalyzed by Zwf, potentially creating a metabolic bottleneck and elevating redox stress. Gcd variants with an altered substrate specificity for selective glucose oxidation, developed for biomedical glucose sensing, may be repurposed here following careful assessment of solubility, membrane localization, and cofactor compatibility (Haque et al., 2019; Yamada et al., 2003). In parallel, improving xylose utilization may require optimization of both the substrate specificity and expression level of the xylE transporter. Native XylE is subject to competitive inhibition by glucose, but this limitation can be alleviated using engineered variants such as Q175I and L297F (Bazzone et al., 2022; Madej et al., 2014). Furthermore, enhancing xylE expression could increase transport efficiency; however, overexpression must be carefully balanced to prevent potential membrane stress (Elmore et al., 2020). These limitations indicate opportunities for further improvement in sugar utilization; nevertheless, the system enabled robust growth and isoprenyl acetate production with minimal nutrient supplementation, demonstrating practical viability of the current strain design.

In addition to these pathway-level bottlenecks, the high volatility of isoprenyl acetate posed an additional challenge during aerobic cultivation in the bioreactor. Although off-gas trapping recovered some evaporated ester, gas-phase losses constrained titer gains compared to the isoprenol case (Table S9). Addressing this issue will require more effective in situ capture strategies or sealed cultivation systems.

Together with the metabolic, regulatory, and process-level optimizations described above, our multi-layered engineering approach establishes a strong foundation for improving ester yields and process efficiency, particularly from mixed sugar feedstocks such as glucose and xylose. Notably, these efforts enabled an over 10-fold increase in isoprenyl acetate production, from 181 mg/L in the unoptimized PIPAxyl strain to 1.5 g/L in shake flasks and 1.9 g/L under fed-batch cultivation with the consolidated PIPAxyl-E3-K3-O15 strain. This positions P. putida as the near-term chassis for hydrolysate-based industrial production from mixed sugars, despite currently lower absolute titers than E. coli.

Conclusions

5

The development of SAFs is a pressing priority for the advancement of the aviation sector. Here, we report the biosynthesis of isoprenyl acetate, a promising SAF precursor, in P. putida using glucose-xylose mixtures as carbon sources. By implementing a multi-faceted metabolic engineering strategy, we achieved gram-per-liter scale production in fed-batch cultivation.

Our work highlights the metabolic robustness of P. putida as a chassis for ester biosynthesis for lignocellulosic biomass valorization. Native challenges such as product degradation and limited precursor availability were addressed through strategic pathway engineering, which included deletion of endogenous esterases and introduction of heterologous pathways designed to channel carbon flux toward acetyl-CoA biosynthesis. These modifications, combined with disruption of carbon catabolite control, enabled efficient conversion of mixed sugars to the target volatile ester. Although product losses due to volatility remain, these may be mitigated through improved downstream in situ capture. Overall, our findings lay a foundation for future optimization of P. putida as a production platform for advanced esters and other next-generation biofuels derived from renewable feedstocks.

CRediT authorship contribution statement

Chae Won Kang: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. David N. Carruthers: Writing – review & editing, Methodology, Investigation, Data curation, Conceptualization. Joshua McCauley: Methodology, Investigation, Formal analysis. Yan Chen: Methodology, Formal analysis, Data curation. Jennifer W. Gin: Methodology, Formal analysis. Christopher J. Petzold: Writing – review & editing, Methodology, Investigation, Data curation. Blake A. Simmons: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization. Taek Soon Lee: Writing – review & editing, Writing – original draft, Supervision, Project administration, Investigation, Funding acquisition, Data curation, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Blake Simmons reports a relationship with Erg Bio that includes board membership. Taek Soon Lee, Chae Won Kang, David Carruthers have patent GENETICALLY MODIFIED PSEUDOMONAS HOST CELLS AND METHODS USEFUL FOR PRODUCING ISOPRENYL ACETATE FROM BIOMASS HYDROLYSATE pending to THE REGENTS OF THE UNIVERSITY OF CALIFORNIA. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.