Engineering mixed sugar metabolic channels in Pseudomonas putida to produce vanillic acid

Abstract

Hydrolysis of lignocellulosic biomass produces the mixed sugars of glucose (60–70 %), xylose (20–30 %), and arabinose (2–20 %), etc. Using mixed sugars instead of pure glucose for microbial biosynthesis will reduce the cost of carbon source and maximize utilization of biomass. However, carbon catabolite repression and poor adaptation of metabolic pathways are obstacles in the synergistic utilization of mixed sugars, thus resulting in low carbon utilization efficiency. Here, we engineered the mixed sugar metabolic channels in Pseudomonas putida to achieve their synergistic utilization for producing vanillic acid, a valuable aromatic compound with broad applications in the food, pharmaceuticals, cosmetics, and chemical industries. Expressing O-methyltransferase (OMT) and deleting vanillate-O-demethylase (vanAB) realized vanillic acid accumulation in P. putida from glucose. Introducing the xylose isomerase pathway enabled the strain to produce vanillic acid from xylose. Deleting glucose dehydrogenase (gcd) and transcriptional regulator (hexR), together expressing two critical pentose phosphate pathway enzymes (transketolase and transaldolase) effectively balanced glucose-xylose metabolic channels for vanillic acid production. Further assembling the arabinose oxidation pathway established the efficient metabolism of three sugars. The final engineered strain (VA12) produced 2.75 g/L vanillic acid in fed-batch fermentation with 20 g/L glucose, 10 g/L xylose and 10 g/L arabinose. This study effectively reduces carbon catabolite repression in the synergistic utilization of mixed sugars, and represents the first case utilizing glucose-xylose-arabinose for vanillic acid production, illustrating the capability of transforming lignocellulose hydrolyzate into valuable chemical products.

1. Introduction

Harnessing renewable resources like lignocellulosic biomass (LCB) to manufacture biofuels and biochemicals is pivotal in curbing fossil fuel dependency and fostering a greener society [1,2]. The cellulose and hemicellulose constitute 50–80 % of the dry weight of lignocellulosic biomass [3,4]. Their hydrolyzed sugars, including glucose, xylose, and arabinose, can subsequently be fermented to generate high-value biochemicals, however, the efficient utilization of the hydrolyzed sugars is essential for the economic viability of lignocellulosic biomass conversion [5]. Especially, improving the co-utilization of mixed sugars will significantly enhance the application level of lignocellulosic biomass [6,7]. As of now, there have been successful examples of using mixed sugar biosynthetic chemicals. Engineering Escherichia coli synthesized succinate using glucose-xylose from corn stover hydrolysate [8]. The acetone, butanol, and ethanol were synthesized by engineered Clostridium acetobutylicum through fermentation by co-utilizing glucose and xylose [9]. Nevertheless, these modifications were grounded in the strain's inherent capacity to metabolize sugar. The glucose-induced carbon catabolite repression (CCR) effect will inhibit the expression of genes associated with the transport and metabolism of pentoses such as xylose and arabinose, thus restricting the co-utilization of mixed sugars by microorganisms [10]. Because the CCR inhibitory effect involves the intricate transcriptional and translational regulatory mechanisms, achieving the co-utilization of mixed sugars in microorganisms that naturally metabolize these sugars by metabolic engineering operation remains a significant challenge [11]. Therefore, using microorganisms, such as P. putida, that inherently lack the ability to metabolize multiple sugars to design mixed sugar metabolic channels will be a promising choice.

P. putida, a gram-negative bacterium, lacks native xylose and arabinose metabolic pathways, which provides a clear opportunity for engineering the mixed sugar metabolic channels independently of CCR [12]. In addition, P. putida shows an impressive resistance to toxic cellular substances including furans, organic acids and phenolic compounds, demonstrating advantages over conventional microbial hosts like E. coli and Saccharomyces cerevisiae [[13], [14], [15]]. Previous studies have verified that the co-utilization of mixed sugars can be achieved by modifying Pseudomonas. For example, P. putida was engineered to produce muconate using glucose and xylose as carbon sources [16]. Through engineering glucose and cellobiose pathways, P. putida enhanced the pyruvate production [17].

Researchers on microbial metabolism has shown that a range of impressive compounds can be produced through aromatic decomposition pathways utilizing sugars from lignocellulose [18]. Notable successful examples include engineered E. coli capable of producing 4-hydroxymandelic acid from a glucose-xylose mixture [19]. cis,cis-muconic acid and 4-hydroxybenzoic acid were synthesized by engineered E. coli using a sugar mixture obtained from lignocellulose (containing glucose and xylose) [20]. P. putida synthesized catechol from lignin-degraded ferulic acid and p-coumaric acid [21]. As an important aromatic phenolic acid, vanillic acid shows an array of beneficial biological functions like boosting the immune system, reducing inflammation, combating bacteria, serving as an antioxidant, and contributing to the prevention of heart disease and more, and even can be transformed into polyesters, vanillin and other performance-advantaged bioproducts [22,23]. It can also be synthesized by microorganisms such as P. putida, Sphingomonas sp. SYK-6, and Rhodococcus jostii RHA1 from carbohydrates, or aromatic compounds by the shikimate pathway [24,25]. This pathway initiates with the condensation of erythrose-4-phosphate (E4P) and phosphoenolpyruvate (PEP) to form 3-deoxy-d-arabinoheptulosonate 7-phosphate (DAHP). DAHP is then converted into 3-dehydroshikimic acid (3-DHS), a key intermediate in the metabolic pathway. Subsequently, 3-DHS is transformed into protocatechuic acid (PCA) by 3-dehydroshikimate dehydratase. Finally, PCA is transformed to vanillic acid by O-methyltransferase (OMT) [26] (Fig. 1). Presently, there are no reported instances of biosynthesis of vanillic acid through engineering the mixed sugar metabolic channels.

Fig. 1.

Schematic representation of metabolic engineering strategies for vanillic acid production via mixed sugar metabolic channels in Pseudomonas putida KT2440. The black arrows represent the original path of P. putida. xylA encodes xylose isomerase, xylB encodes xylulokinase, xylE encodes xylose-H+ symporte, araE₁ encodes arabinose-H+ symporter, araA₁ encodes Arabinose isomerase, araB₁ encodes ribulokinase, araD₁ encodes Lribulose-5-phosphate 4-epimerase, tal encodes transaldolase, tkt encodes transketolase, araA₂ encodes l-arabinose dehydrogenase, araB₂ encodes l-arabinolactonase, araC₂ encodes l-arabinoate dehydratase, araD₂ encodes 2-keto-3-deoxy-l-arabinoate dehydratase, araE₂ encodes 2-ketoglutarate semialdehyde dehydrogenase, gcd encodes glucose dehydrogenase, vanAB encodes vanillate-O-demethylase, OMT encodes O-methyltransferase. 2-KGn: 2-ketogluconate; 2K6PG: 2-ketogluconate-6-phosphate; KDPG: 2-dehydro-3-deoxyphosphogluconat; DHAP: dihydroxyacetone phosphate; G6P: Glucose-6-phosphate; 6PG: 6-phosphogluconate; G3P: d-glyceraldehyde 3-phosphate; F6P: fructose-6-phosphate; FBP: fructose-1,6-bisphosphate; R5P: ribose-5-phosphate; d-Ri5P: d-ribulose-5-phosphate; X5P: xylulose-5-phosphate; E4P: erythrose-4-phosphate; S7P: sedoheptulose-7-phosphate; 3 PG: 3-phosphoglycerate; PEP: phosphoenoylpyruvate; CIT: citrate; ICIT: isocitrate; 2-KGT: 2-ketoglutarate; SUCC: succinate; SUC-CoA: succinyl-coenzyme A; FUM: fumarate; MAL: malate; OAA: oxaloacetate; GLX: glyoxylate; DAHP: 3-deoxy-d-arabinoheptulosonate 7-phosphate; DHQ: 3-dehydroquinate; DHS: 3-dehydroshikimate; PCA: protocatechuic acid; Ri5P: l-ribulose-5-phosphat.

In this study, we aim to design a glucose-xylose-arabinose metabolic channel in P. putida and to produce vanillic acid through the co-utilization of mixed sugars. Introducing the O-methyltransferase (OMT) and knocking out the vanillin-O-demethylase (vanAB) gene in P. putida first produced a vanillic acid-producing strain from glucose. Then, based on the introduction of the xylose isomerase pathway, combining metabolic engineering including deleting glucose dehydrogenase and transcriptional regulator hexR, expressing two critical pentose phosphate pathway enzymes (transketolase and transaldolase) effectively mitigated the CCR effect. Finally, expressing the arabinose oxidation pathway achieved vanillic acid production through co-utilization of glucose, xylose and arabinose, with a titer of 2.75 g/L in batch-fed fermentation. Fig. 1 illustrates the metabolic engineering strategies for mixed sugar metabolic channels.

2. Materials and methods

2.1. Chemicals and reagents

Common chemicals and reagents were purchased from Solarbio (Beijing, China) or Thermo Fisher Scientific (Waltham, USA) unless otherwise noted. Molecular biology enzymes and related kits came from Takara Bio (Dalian, China), Vazyme (Nanjing, China), or New England Biolabs (Ipswich, UK). Genewiz (Suzhou, China) handled all primer synthesis and DNA sequencing services.

2.2. Plasmids, strains, and culture conditions

Table 1 lists the strains and plasmids used in this research. E. coli DH5α served in gene cloning and plasmid amplification. P. putida KT2440 facilitated protein expression and target compound synthesis. The genes, being of heterogeneous origin, were chemically synthesized by Genewiz (Suzhou, China) after undergoing codon optimization. Table S1 catalogs the oligonucleotide sequences employed in plasmid engineering, gene knockout, and verification procedures.

Table 1.

Strains and plasmids used in this study.

Name	Characteristics	Reference
Plasmids
pBBR1MCS-5	pBBR1 oriV, pBBR1 Rep, GmR, PT7:mcs-lacZa	Novagen
pK18mobsacB	pUC origin, nptII, sacB, Plac:mcs-lacZa	Novagen
pBBR-OMT	pBBR1MCS-5 harboring PJ23119-OMT cassette	This study
pBBR-xylE-xylA-xylB	pBBR1MCS-5 harboring PJ23105-xylE:PJ231119--xylAB cassette	This study
pK18-vanAB	pK18mobsacB harboring ΔvanAB cassette	This study
pK18-vanAB::OMT	pK18mobsacB harboringΔvanAB::PJ23119-OMT cassette	This study
pK18-gcd	pK18mobsacB harboring Δgcd cassette	This study
pK18-hexR	pK18mobsacB harboring ΔhexR cassette	This study
pK18- hexR::talB-tktA	pK18mobsacB harboring ΔhexR::PEM7-talB-tktA cassette	This study
pK18-gcd::xylE-xylA-xylB	pK18mobsacB harboring Δgcd::PJ23105-xylE:PJ231119--xylAB cassette	This study
pK18-endA::araE1-araB1-araA1-araD1	pK18mobsacB harboring ΔendA::PJ23105-araE1:PJ231119- araB1-araA1:Prpsh- araD1 cassette	This study
pK18-endA::araE1-araB2-araC2-araD2araA2-araE2	pK18mobsacB harboring ΔendA::PJ23105-araE1:Plpp2.0- araB2-araC2:PJ231119-araD2araA2-araE2 cassette	This study
Strains
VA01	P. putida KT2440 Wide-type harboring pBBR1MCS-5	This study
VA02	P. putida KT2440 Wide-type harboring pBBR-OMT	This study
VA03	P. putida KT2440 Wide-type ΔvanAB harboring pBBR1MCS-5	This study
VA04	P. putida KT2440 Wide-type ΔvanAB harboring pBBR-OMT	This study
VA05	P. putida KT2440 Wide-type ΔvanAB::PJ23119-OMT	This study
VA06	VA05 harboring pBBR-xylE-xylA-xylB	This study
VA07	VA05 Δgcd harboring pBBR-xylE-xylA-xylB	This study
VA08	VA05 Δgcd::PJ23105-xylE:PJ231119--xylAB	This study
VA09	VA08 ΔhexR	This study
VA10	VA09 ΔhexR::PEM7-talB-tktA	This study
VA11	VA10 ΔendA::PJ23105-araE1:PJ231119- araB1-araA1:Prpsh- araD1	This study
VA12	VA11 ΔendA::PJ23105-araE1:Plpp2.0- araB2-araC2:PJ231119-araD2araA2-araE2	This study

E. coli DH5α and P. putida KT2440 were cultivated in Luria-Bertani (LB) medium. For fermentation, a modified M9 medium was employed, consisting of 1 × M9 minimal salts (6.78 g/L Na₂PO₄, 3 g/L KH₂PO₄, 0.5 g/L NaCl, 1.0 g/L NH₄Cl), supplemented with 2 mM MgSO₄, 0.1 mM CaCl₂, and 20 % glucose. If required, gentamycin concentration in the medium was adjusted to 50 μg/mL.

2.3. Plasmid construction

Gibson assembly [27] was used to create plasmids. The pBBR1MCS-5 plasmid was used to construct expression plasmids. pBBR-xylE-xylA-xylB exemplified a concise illustration. Initially, the linearized pBBR1MCS-5 plasmid was prepared by reverse PCR amplification with primers pBBR-F and pBBR-R. Subsequently, the insert xylE, xylA, xylB genes were amplified with the corresponding primers xylE-F/R, xylA-F/R and xylB-F/R, respectively. Finally, ClonExpress Ultra (Vazyme Biotech, Nanjing, China) facilitated ligation of the linearized plasmid and insert. Transformed E. coli DH5α, containing ligation products, were plated and incubated at 37 °C overnight to identify successful clones.

All the gene knock-outs and knock-ins were constructed by the suicide plasmid pK18mobsacB. The knockout of the gcd gene was used as an example. The plasmid pK18mobsacB was first cut with primers pK18-F and pK18-R to produce a linear fragment. This fragment was subsequently ligated with the flanking regions of the gcd gene, each approximately 1000 bp in length, resulting in the plasmid pK18-gcd. Plasmid pK18-gcd was electroporated into P. putida KT2440 and chromosomally integrated [28]. The initial recombinant colonies resulting from the first crossover event were selected on LB agar with 50 μg/mL kanamycin (LB_Kan). Subsequent recombinants produced through the second crossover were identified by plating on LB medium with 20 g/L sucrose. Positive clones were then confirmed through PCR amplification and DNA sequence analysis. To confirm whether the strain had lost the plasmid, it was streaked onto two types of LB agar plates—one containing kanamycin (50 μg/mL) and another without any antibiotics—then incubated at 30 °C. The strain that showed growth exclusively on the antibiotic-free plate was designated as P. putida Δgcd. Additional P. putida KT2440 mutants were generated using an identical protocol, with the non-essential endA gene serving as the insertion site. All engineered strains underwent verification via PCR and sequencing. The DNA sequences for three synthetic gene cassettes (xylEAB, araE₁B₁A₁D₁, and araE₁B₂C₂D₂A₂E₂) are provided in Table S2.

2.4. Shake-flask fermentation

Engineered P. putida strains were first cultivated as single colonies in 5 mL of LB broth, incubated overnight at 30 °C with vigorous shaking at 220 rpm. Following this initial growth phase, a 3 mL aliquot from the preliminary culture was sterilely introduced into a 250 mL flask holding 50 mL of 2 × M9 minimal salts medium. And the medium was further enriched using 2 mM MgSO₄ and 0.1 mM CaCl₂. Gentamycin at a concentration of 50 μg/mL was introduced into the medium as required. To synthesize the desired compounds, precise quantities of glucose, xylose, and arabinose were introduced into the culture medium. The cells were then incubated at 30 °C under constant agitation at 220 rpm. At predetermined intervals, samples were taken and analyzed via HPLC to monitor production progress.

2.5. Fed-batch fermentation

Vanillic acid was produced through batch-feeding fermentation using the Minifors Cell desktop bioreactor (INFORS, Switzerland). To prepare the seed culture, individual colonies were introduced into a 250 mL flask holding 50 mL of LB broth and then incubated at 30 °C with constant agitation (220 rpm) for 12 h. After this initial growth phase, the seed culture was inoculated into a 6 L bioreactor filled with 4 L of fermentation medium, which consisted of 1 × M9 minimal salts, 5 g/L yeast extract, 20 g/L glucose, 10 g/L xylose, 10 g/L arabinose, 2 mM MgSO₄, and 0.1 mM CaCl₂, at an inoculation rate of 5 %. A batch of concentrated glucose (50 %, w/v), xylose (50 %, w/v), and arabinose (50 %, w/v) solutions was introduced into the fermenter at calculated rates, tailored to the residual sugar levels, to maximize product yields. This feeding regimen began 24 h after the initial inoculation and continued twice daily until the 96-h mark. To keep the fermentation broth's pH at a steady 7.0 ± 0.5, NH₄OH (25 %, v/v) was periodically added. The airflow was regulated to a steady 1 L/min, and the dissolved oxygen (DO) content was meticulously managed at 30 % by adjusting the stirrer speed from 200 to 900 rpm. Samples were taken at regular checkpoints to track cellular proliferation and to perform high-performance liquid chromatography (HPLC) analyses.

2.6. Chemical analyses and quantification

Vanillic acid and related intermediates were analyzed via HPLC (Shimadzu, Kyoto, Japan) using an Agilent C18 column (4.6 × 150 mm, 2.6 μm, Santa Clara, CA, USA). The fermentation extract was sonicated in a 50 % ethanol solution for half an hour, then spun at 12,000 rpm for 3 min to separate the supernatant. This clear liquid was subsequently filtered through a 0.22 μm nylon syringe filter and transferred into HPLC vials for storage at 4 °C until further analysis. The HPLC methodology followed the protocols outlined in prior research [29].

Glucose, xylose, arabinose and gluconate were determined on an HPLP system (Shimadzu, Kyoto, Japan) equipped with a refractive index detector (RID-20A) and an Amine HPX-87 H column (300 × 8.0 mm, 6 μm, Bio-Rad, Hercules, USA). Prior to analysis, the samples were centrifuged at 12,000 revolutions per minute (rpm) for 3 min. The resulting supernatants were then filtered through a 0.22 μm PES syringe filter into HPLC vials and stored at 4 °C. A 25 μL aliquot of each sample was injected for analysis. The fractionation occurred at 60 °C, using 5 mM sulfuric acid as the mobile phase at a rate of 0.6 mL/min. Each experiment was conducted in triplicate, and the results are presented as the mean ± standard deviation, with error bars indicating variability in the figures.

3. Results and discussion

3.1. Construction of vanillic acid biosynthetic pathway in P. putida

P. putida KT2440 can synthesize protocatechuic acid (PCA), the vanillic acid precursor, through the naturally occurring shikimic acid pathway. To establish the vanillic acid production from PCA, the O-methyltransferase (OMT) from Homo sapiens was expressed under the control of a constitutive promoter PJ23119 in the plasmid pBBR-MCS5 with a middle-copy BBR1 origin. pBBR-MCS5 (control) and pBBR-OMT (expression vector) were electroporated into wild-type P. putida KT2440, respectively, generating strains VA01 and VA02. After 48 h of cultivation, strain VA02 produced 32.3 mg/L of vanillic acid from glucose. However, the titer of vanillic acid decreased from 32.3 mg/L to 16.9 mg/L between 48 h and 72 h (Fig. 2A).

Fig. 2.

De novo biosynthesis of vanillic acid (VA). Production of VA by expressing OMT (A); The effect of knocking out vanAB on protocatechuic acid (PCA) (B) and VA synthesis (D); Genomic integration of OMT to increase VA production (C).

To explain the decline of vanillic acid in the late fermentation period, we detected other products and found 16.5 mg/L of protocatechuic acid at 72 h in strain VA02 (Fig. 2B). We speculated that vanillate-O-demethylase (VanAB), encoded by vanAB (PP_3736 and PP_3737) in P. putida, probably demethylated vanillic acid to protocatechuic acid [30]. Thus, the vanAB genes were deleted from the strains VA01 and VA02, generating strains VA03 and VA04, respectively. This modification increased vanillic acid to 113 mg/L (strain VA04), 3.5-fold higher than that of VA02, simultaneously preventing the accumulation of protocatechuic acid (Fig. 2C and D). Finally, to produce vanillic acid without the addition of inducers or antibiotics, OMT was integrated into the vanAB locus of P. putida KT2440, resulting in the strain VA05. The vanillic acid titer reached 316 mg/L, 2.8-fold higher than VA04 (Fig. 2C), which may due to the low plasmid stability (Figure_S1) and growth (Figure_S2).

3.2. Introducing the xylose isomerase pathway into P. putida for vanillic acid production

P. putida is considered a promising chassis for lignocellulosic bioconversion and aromatic compound production, however, it does not possess the inherent capability to metabolize xylose [31]. Thus, the introduction of an exogenous metabolic pathway for xylose utilization is essential to obtain a mixed-sugar-utilizing strain. Phosphorylation-based regulation of sugar metabolism is considered an effective strategy for channeling xylose into the shikimic acid pathway. The xylose isomerase pathway represents a commonly utilized route for xylose metabolism, converting it to xylulose-5-phosphate (X5P) via isomerase and xylulokinase (XK) [32]. Thus, we first introduced the related genes xylA, xylB, and xylE from E. coli into the strain VA05, resulting in the strain VA06 (Figure_S3). Compared with strain VA05, the strain VA06 was capable of utilizing xylose as the sole carbon source for growth and production of vanillic acid (Fig. 3A and B), and no significant differences when cultured with 20 g/L glucose (Fig. 3C and D). However, the vanillic acid produced by the strain VA06 using xylose is only about 80 % that produced with glucose, which may cause by the lower xylose consumption rate (0.15 g/L/h) than glucose (0.29 g/L/h) and lower synthesis efficiency of xylose isomerase pathway than glucose pathway (Fig. 3B and D).

Fig. 3.

The xylose isomerase pathway for the synthesis of vanillic acid (VA). Shake-flask profiles of strains VA05 and VA06 in M9 medium with 20 g/L xylose (A) and (B), 20 g/L glucose (C) and (D), and 10 g/L glucose + 10 g/L xylose (E) and (F).

When strains were cultured using 10 g/L glucose and 10 g/L xylose, the VA06 strain produced 267 mg/L vanillic acid, high 28 % than the VA05 strain (Fig. 3E and F). However, compared to using glucose or xylose as the sole carbon source, the consumption rates of glucose (0.13 g/L/h) and xylose (0.06 g/L/h) during mixed sugar fermentation of strain VA06 reduced 55 % and 60 %, respectively (Fig. 3B–D and F). The CCR phenomenon results in preferential glucose metabolism, which delays xylose uptake and affects vanillic acid production in P. putida. Conversely, xylose also affected the glucose metabolism through competitive binding or allosteric regulation hampering the glucose uptake [33]. In addition, the simultaneous metabolism of two sugars could break the energy and redox balance, systemically down-regulating the consumption rates of glucose and xylose [34].

3.3. Optimization of glucose-xylose dual metabolic channels for vanillic acid synthesis

To alleviate the inhibitory impact on carbon metabolism of the glucose-xylose dual metabolic channel, balancing the metabolic flux distribution between glucose and xylose is a strategy. P. putida naturally possesses glucose dehydrogenase (encoded by gcd, PP_1444), a key enzyme involved in glucose metabolism that oxidizes glucose to gluconate [35]. Glucose dehydrogenase can also oxidize xylose to xylonate [36]. While this activity has minimal impact on the oxidative Weimberg pathway for xylose catabolism, it is detrimental to the xylose utilization of the isomerase pathway [37]. Thus, we first deleted glucose dehydrogenase (gcd) from the strain VA06, generating the strain VA07. Compared with the strain VA06, the strain VA07 increased growth rate by 1.5-fold when cultured with 10 g/L glucose and 10 g/L xylose (Fig. 4A and B), but exhibited about 24-h delay when cultured with 10 g/L glucose (Fig. 4C). This phenomenon may be attributed to the temporary limitation of intracellular glucose phosphorylation flux caused by the inhibition of glucose oxidation pathway [38]. Nikel et al. reported that up to 90 % of glucose was oxidized into gluconate in P. putida, thus resulting in carbon loss [39]. In this study, modification of glucose oxidation pathway also completely prevented the gluconate accumulation (Figure_S4). The deletion of gcd effectively minimized carbon loss and redirected a greater carbon flux toward the central metabolic pathway [40]. In addition, the reallocation of carbon fluxes resulted in a 1.8-fold increase in xylose consumption by strain VA07 compared to strain VA06 (Fig. 4D), suggesting that the gcd deletion accelerated the co-utilization rate of glucose and xylose [41]. Altogether, this modification increased the production of vanillic acid by 19 % in the strain VA06 (Fig. 4E). This improvement can be attributed to the reduction in carbon loss and the enhancement of central metabolic flux. Knocking out the gcd to inhibit the glucose oxidation pathway also enhanced the availability of key precursors such as acetyl-CoA, thereby promoting the synthesis of products like polyhydroxyalkanoate [42]. The integration of the xylose metabolic pathway into the gcd locus of strain VA05 resulted in strain VA08, showing the same results as strain VA07, further validating the universality of this strategy for aromatic acid production (Fig. 4E).

Fig. 4.

Balancing flux through sugar catabolic pathways. (A) Growth rates, (B) growth curves, (C) glucose consumption, (D) xylose consumption, and (E) vanillic acid (VA) production of VA06, VA07, VA08, VA09, and VA10 on M9 medium with 10 g/L glucose + 10 g/L xylose. μ represents absolute growth rate.

Then, we deleted hexR, a transcriptional regulator influencing zwf, pgi, and edd-eda gene expression, which are critical for utilizing glucose or xylose [43]. Previous studies have confirmed that knocking out the hexR gene can alleviate the growth inhibition of glucose or xylose in the strains [44]. The knockout of hexR from the strain VA08 produced the strain VA09. When cultured with 10 g/L glucose and 10 g/L xylose, the growth rate of strain VA09 was enhanced by 1.2-fold compared to VA08 (Fig. 4A and B), supporting the hypothesis that the deletion of hexR promotes growth by increasing the metabolic flux through glucose metabolism pathways [45]. The strain VA09 also showed a 47 % increase in glucose consumption within 24 h and no significant alteration in xylose metabolism (Fig. 4C and D). This intervention relieved the inhibition of glucose metabolism more effectively than that of xylose metabolism, consistent with findings reported in previous studies [46]. This result thus suggests that xylose, as a non-native carbon source, its metabolism may be influenced by other regulatory mechanisms, such as transport competition or inhibitors independent of hexR [47,48]. For vanillin acid production, the fermentation time of VA09 has been shortened by 24 h to reach 300 mg/L compared to that of VA08 (Fig. 4E), which is largely from the enhanced carbon flux of the shikimic acid pathway caused by the knockout of hexR, increasing the expression of 3-dehydroquinate synthase (aroB) [44].

However, through knocking out the gcd and hexR genes, the growth rate of strain VA09 increases to 0.062 h⁻¹ when cultivated with 10 g/L glucose and 10 g/L xylose, but this remains lower than the growth rate (0.076 h⁻¹) when cultivated with 20 g/L glucose. To further enhance the growth rate by promoting xylose utilization, the enzymes transketolase (tktA) and transaldolase (talB), which play pivotal roles in the E. coli pentose phosphate pathway, were integrated into the hexR locus of strain VA09, resulting in the strain VA10. Becasuse the provision of NADPH and pentose phosphate was improved by upregulating the transketolase and transaldolase [[49], [50], [51]], the growth rate of the strain VA10 cultured with 10 g/L glucose and 10 g/L xylose increased to 0.078 h⁻¹, which was comparable to the level cultured with 20 g/L glucose (Fig. 4A). And also, the time required for the VA10 strain to consume 10 g/L of glucose and xylose reduced to 24 h compared to the VA09 strain (Fig. 4C and D). This improvement is attributed to the overexpression of tktA and talB, which effectively optimized the EDEMP cycle, pentose phosphate pathway, Entner-Doudoroff pathway, and gluconeogenesis reactions in P. putida [52]. The improved mixed sugar utilization also promoted the production of vanillic acid to 342 mg/L, increased by 18 % compared only with glucose culture (Fig. 4E).

The conventional perspective suggests that P. putida exhibits strict substrate preference in mixed sugar environments, exemplified by its suppression of xylose utilization when glucose is present [53]. Here, through optimizing dual-channel synchronous metabolism of glucose and xylose, we have circumvented carbon catabolite repression and enhanced growth rate and sugar consumption, and finally promoted the production of vanillic acid.

3.4. Establishing a glucose-xylose-arabinose metabolic channel for vanillic acid synthesis

Arabinose typically constitutes 2 %–20 % of the total soluble sugars present in lignocellulosic hydrolysate [[54], [55], [56], [57]]. To broaden the range of carbon sources available to the strain, based on the strain VA10 capable of utilizing glucose and xylose, we integrated the arabinose isomerase pathway containing a codon-optimized araA, araB1, araD, and araE from E. coli into the endA locus (PP_3375), resulting in the strain VA11 (Figure_S5A). As depicted in Fig. 5A, B, and C, strain VA11 was capable of utilizing 10 g/L arabinose as the sole carbon source for growth and produced 146 mg/L of vanillic acid. However, when cultured in the medium containing 10 g/L of glucose, xylose and arabinose each, the growth rate and sugar consumption rate of strain VA11 decreased by 21 % and 12 % respectively compared to strain VA10 (Fig. 5D, E, J, H, I), resulting in only 260 mg/L vanillic acid, about 76 % of strain VA10 (Fig. 5F). The reduced utilization for xylose and arabinose indicated that their metabolisms were more severely rate-limited. We guess that this is possibly attributable to metabolic competition between xylose and arabinose because both converge into the pentose phosphate pathway at the same metabolic node, xylose-5-phosphate (Fig. 1) [58]. Thus, we attempted to use the non-phosphorylated route-arabinose oxidation pathway to replace the arabinose isomerase pathway.

Fig. 5.

Arabinose pathways for the synthesis of vanillic acid (VA). (A) Growth curves, (B) arabinose consumption, and (C) vanillic acid production of VA10, VA11 and VA12 on M9 medium with 10 g/L arabinose, respectively; (D) Growth curves, (E) arabinose consumption, (F) vanillic acid production, (G) glucose consumption, (H) xylose consumption, and (I) growth rates of VA10, VA11 and VA12 on M9 medium with 10 g/L glucose + 10 g/L xylose + 10 g/L arabinose, respectively. μ represents absolute growth rate.

Assembling the arabinose oxidation pathway, including the codon-optimized araE1 gene from E. coli, and araA2, araC2, araB2, araD2, araE2 genes from Burkholderia multivorans into the endA locus of strain VA10 generated strain VA12 (Figure_S5B). The obtained strain VA12 was capable of growing on 10 g/L arabinose (Fig. 5A and B), but no vanillic acid was detected (Fig. 5C). In a mixed culture with 10 g/L of glucose, xylose, and arabinose each, strain VA12 exhibited improved growth and sugar consumption efficiency over VA11 (Fig. 5D, E, G, H, I). The growth rate of strain VA12 reached 0.086 h⁻¹, 1.4-fold of strain VA11 and 2.4-fold of P. putida utilizing glucose and xylose for synthesis of muconic acid [16]. The μ_max of strain VA12 (0.541 h⁻¹) and sugar consumption rate (0.29 g/L/h) are close to those of recently published studies (Table S3).

The vanillic acid produced by VA12 reached 393 mg/L, representing a 42 % and 15 % increase compared with VA11 and VA10, respectively (Fig. 5F), indicating the effectiveness of the arabinose oxidation pathway strategy. The efficient metabolism of three sugars can be attributed to their entry into central metabolic pathways at distinct points [59]. The arabinose oxidation pathway encompasses the conversion of arabinose to 2-ketoglutarate via a series of enzymatic reactions mediated by araE1, araB2, araC2, araD2, araA2 and araE2, which subsequently enter the tricarboxylic acid (TCA) cycle [60,61]. 2-ketoglutarate can be metabolized into pyruvate via the TCA cycle, consequently influencing the synthesis of phosphoenolpyruvate (PEP). As the primary substrate in the shikimate pathway, PEP provides the essential precursor necessary for vanillic acid biosynthesis [62]. In addition, ATP and NADPH produced during the TCA cycle can also enhance glycolysis and the pentose phosphate pathway, thereby indirectly modulating the supply balance between erythrose-4-phosphate (E4P) and PEP, and providing the necessary energy for vanillic acid synthesis [63,64]. Dynamic regulation of competitive nodes in the TCA cycle, such as pyk, can redirect a greater carbon flux toward the shikimate pathway [65,66]. This also demonstrates a strong correlation between TCA cycle flux and shikimate pathway. However, compared with the carbon yield (3.0 %) of strain VA05 fermented with glucose, the carbon yield (1.7 %) of strain VA12 fermented with mixed sugar decreased by 43.3 %. The longer and more energy-intensive metabolic pathways of xylose and arabinose usually lead to the lower carbon yield than that of glucose, commonly observed in engineered strains [67]. The additional carbon source was used to support microbial growth or sustain metabolic activities rather than directly channeled into the synthesis of target compound [[68], [69], [70]]. Future research should further elucidate the dynamic interaction network between sugar metabolism, TCA cycle and shikimate pathway, as this understanding will facilitate the production of more valuable metabolites for multi-carbon source utilization.

3.5. Vanillic acid production through fermentation optimization

To optimize the fermentation process, we first examined the impact of mixed sugar ratio on vanillic acid production (Fig. 6A). The strain VA12 produced 338 mg/L of vanillic acid at a 1:1 glucose-to-xylose ratio (10 g/L each). A higher glucose ratio (2:1) enhanced vanillic acid production by 23 %. Based on this sugar ratio, supplementing arabinose to a ratio of 2:1:1 (20 g/L glucose, 10 g/L xylose and 10 g/L arabinose) resulted in 495 mg/L of vanillic acid production. However, further increasing the xylose and arabinose to 2:2:2 (each at 20 g/L alongside glucose) did not promote vanillic acid production. These findings suggest that glucose utilization has the greatest impact on vanillic acid production.

Fig. 6.

Optimization experiment for vanillic acid production through fermentation. (A) The vanillic acid (VA) production by strain VA12 with different ratios of sugars in the shake-flask fermentation. The ratios of glucose, xylose and arabinose were adjusted from 1:1 (10 g/L glucose and 10 g/L xylose) to 2:2:2 (20 g/L glucose, 20 g/L xylose and 20 g/L arabinose). Additionally, the residual sugar levels are illustrated; Bioreactor profiles from strain VA12 in fed-batch mode: (B) time curves of growth and vanillic acid production, (C) the concentrations of residual sugars, (D) the utilization of glucose, xylose and arabinose.

Then, the strain VA12 was cultured in a 6-L fermenter to assess the yield of vanillic acid. The glucose, xylose and arabinose were maintained at about 20 g/L, 10 g/L and 10 g/L, respectively. As illustrated in Fig. 6B, the strain exhibits robust growth and remains unaffected by the presence of three sugars. As observed in the shake flask experiment, the three sugars were consumed simultaneously (Fig. 6C). Consistent with the design, glucose serving as the primary carbon source for vanillic acid synthesis exhibited the highest utilization. Arabinose was predominantly utilized for energy provision during cell growth and demonstrated a higher utilization rate compared to xylose (Fig. 6D). After 120 h fermentation, the strain VA12 consumed 63.3 g/L of glucose, 34.2 g/L of xylose, and 42.9 g/L of arabinose to produce 2.75 g/L vanillic acid (Fig. 6B and D), a 5.6-fold increase in productivity of the shake flask experiment, which is also the best yield reported for the utilization of the three sugars in P. putida.

4. Conclusions

In this study, we demonstrated the efficient co-utilization of glucose, xylose, and arabinose, the primary carbohydrates in lignocellulose, for the production of vanillic acid. Expressing OMT in P. putida efficiently catalyzed the conversion of protocatechuic acid to vanillic acid. Knocking out the vanAB and gcd genes significantly reduced by-product accumulation and enhanced vanillin production. Introducing the xylose isomerase pathway and arabinose oxidation pathway enabled the strain to produce vanillic acid using three sugars. In addition, carbon catabolite repression in the synergistic utilization of mixed sugars was mitigated by knocking out of hexR gene and over-expressing the tktA and talB genes. Finally, the modified strain ultimately produced 2.75 g/L of vanillic acid through combined sugar metabolism. The findings demonstrated the feasibility of synthesizing vanillic acid from mixed sugar sources, highlighting a promising and sustainable approach. However, the vanillic acid titer still limits scaling to industrial levels. Optimization of metabolic flux within the shikimate pathway (e.g., overexpressing aroG, aroQ and aroB to increase the supply of precursors) can enhance product yield, and strengthening product export (e.g., overexpressing transport proteins to reduce intracellular toxicity) can also contribute to this improvement.

CRediT authorship contribution statement

Peihan Wu: Writing – review & editing, Writing – original draft, Methodology, Formal analysis, Data curation, Conceptualization. Haifeng Ding: Methodology, Formal analysis, Data curation. Zhiqing Xu: Formal analysis, Data curation. Yan Zhang: Writing – review & editing, Supervision. Yanyan Dang: Writing – review & editing, Supervision. Bo Gao: Writing – review & editing, Supervision. Genlin Zhang: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix. ASupplementary data

The following is the Supplementary data to this article: