Classification of vanillin according to origin

Bacterial vanillin synthesis

As a model organism, Escherichia coli has been genetically engineered to produce a vast number of compounds (22–24), including vanillin (9). In 2005, Yoon et al. (25) introduced a feruloyl-CoA synthetase (fcs) and enoyl-CoA hydratase/aldolase (ech) from Amycolatopsis sp. into E. coli via a plasmid (Fig. 1 and Table 1 (25-39)). Using ferulic acid as a carbon source, the engineered strain produced 1.1 g/L vanillin (25). Enhancing the expression of the citrate synthase gene (gltA) of the tricarboxylic acid (TCA) cycle resulted in 1.98 g/L vanillin from 3 g/L ferulic acid (26). To further increase the yield, the consumption of acetyl-CoA was optimised by activating the glyoxylate bypass through the deletion of icdA gene. Amplification of the gltA gene and deletion of the icdA gene, together with the use of a polystyrenic resin XAD-2 to reduce vanillin toxicity, yielded 5.14 g/L vanillin with a molar conversion rate of 86.6 % (26).

Fig. 1

‘Natural’ vanillin production routes from the precursors ferulic acid, (iso)eugenol and glucose. IEM=isoeugenol monooxygenase, EhyAB=eugenol hydroxylase, CalA=coniferyl alcohol dehydrogenase, CalB=coniferyl aldehyde dehydrogenase, FCS=feruloyl-CoA synthetase, ECH=enoyl-CoA hydratase/aldolase, 3DSD=3-dehydroshikimate dehydratase, COMT=catechol-O-methyltransferase, CAR=carboxylic acid reductase, PROGO=4-propylguaiacol oxidase, NOV1=isoeugenol dioxygenase

Due to its low price and availability, glucose is considered a more cost-effective carbon source than ferulic acid (40). In 1998, Li and Frost (41) developed a microbial biosynthetic pathway for vanillin production from glucose. In this pathway, E. coli was engineered to catalyse the dehydration of 3-dehydroshikimic acid and the regioselective methylation of the resulting protocatechuic acid to produce vanillic acid. Vanillic acid was then reduced to vanillin by aryl-aldehyde dehydrogenase, purified from Neurospora crassa lysate. While effective, the in vitro reduction step depended on expensive cofactors (adenosine triphosphate (ATP) and nicotinamide adenine dinucleotide phosphate (NADPH)), making the process economically unfeasible for large-scale production. In 2024, Wu et al. (27) optimised the protocatechuic acid pathway in E. coli K12 MG1655 by screening for highly active carboxylic acid reductases and catechol O-methyltransferases. Using this vanillin production route can lead to the formation of vanillyl alcohol as a major side product. Therefore, the authors deleted three endogenous aldo–keto reductases and three alcohol dehydrogenases, while also regulating competitive metabolic pathways. Vanillin was produced from glucose at a yield of 481.2 mg/L (27).

Using isoeugenol monooxygenase (IEM), vanillin can be produced in one step from isoeugenol, a compound found in several essential oils, e.g. derived from clove, nutmeg or cinnamon (42). A recombinant plasmid carrying the IEM gene from Pseudomonas putida IE27 under the control of the T7 promoter was introduced into E. coli BL21(DE3), yielding 28.3 g/L vanillin from 230 mM isoeugenol. This represents a molar conversion yield of 81 % after 6 h of reaction and no accumulation of side products such as vanillic acid or acetaldehyde was reported (28). In 2021, the IEM from Pseudomonas nitroreducens Jin1 was cloned into E. coli, resulting in a recombinant strain that produced 38.34 g/L vanillin with >99 % purity (29). Marić et al. (30) presented a strategy to convert the lignin-derived monomer 4-n-propylguaiacol (4PG) into vanillin using engineered E. coli strains expressing 4-propylguaiacol oxidase (PROGO) and isoeugenol dioxygenase (NOV1). To overcome substrate inhibition, i.e. 4PG competitively inhibiting the second-step enzyme, and to accommodate different temperature optima (37 °C for PROGO and 25 °C for NOV1), a stepwise one-pot strategy was established. Using this method, a vanillin yield of 66 % from raw spruce-derived lignin oil was achieved on a small laboratory scale (250 µL reactions).

Concomitantly, to address plastic waste pollution, a novel approach to vanillin production has been explored. An E. coli MG1655 RARE strain was engineered to carry a novel synthetic pathway that can produce vanillin from polyethylene terephthalate (PET). Achieving a 79 % conversion of terephthalic acid to vanillin (119 mg/L), the study proposed a 157-fold improvement in vanillin yield through process optimisation, i.e. reaction temperature, medium optimisation and in situ product removal. This work documents the first biological upcycling of post-consumer PET into a single value-added small molecule. The findings suggest that PET waste may serve as a carbon source for producing high-value chemicals, contributing to a more circular economy and reducing plastic pollution (31).

Beyond E. coli, various other bacteria have also been used for vanillin production (Table 1). Pseudomonas putida KT2440 was optimised to convert ferulic acid to vanillin by enhanced chromosomal expression of the genes fcs and ech. To decrease further vanillin metabolism, Graf and Altenbuchner (32) deleted the vanillin dehydrogenase gene (vdh); however, this did not fully prevent vanillin turnover. High initial conversion rates and molar vanillin yields of up to 86 % were achieved within just 3 h, with little by-product formation. The highest vanillin titre reached in this approach was 1.31 g/L (32). Amycolatopsis sp. ATCC 39116 is attractive as a vanillin production host due to its high vanillin tolerance. After the introduction of fcs and ech through a plasmid, deletion of vdh and an optimisation of ferulic acid feeding strategies, a vanillin yield of 22.3 g/L was achieved. The deletion of vdh led to a more than 90 % decrease in vanillin turnover (33). A similar engineering strategy was used for converting ferulic acid to vanillin with Pseudomonas fluorescens, yielding 1.28 g/L (34).

Lactic acid bacterium Pediococcus acidilactici BD16 produced 0.48 g/L vanillin from 0.16 g/L [sic!] ferulic acid per milligram of recombinant cell biomass within 20 min of biotransformation. Heterologous genes fcs and ech were introduced via plasmids, and the process was statistically optimised and scaled up (35). The same recombinant P. acidilactici strain was used to explore vanillin production from ferulic acid in rice bran, targeting a more economical vanillin production. The strain yielded 4.01 g/L vanillin within 24 h of incubation with rice bran medium (36).

To address the problem of vanillin toxicity and product inhibition, the ability of resins to adsorb vanillin in situ was tested during the bioconversion of ferulic acid to vanillin using Streptomyces sp. strain V-1, which was isolated from soil samples and characterized. Using resin DM11, the highest amount of vanillin and the lowest amount of ferulic acid were adsorbed, resulting in a yield of 19.2 g/L vanillin within 55 h (37).

Several enzymes involved in reducing aromatic aldehydes to their corresponding alcohols were identified in Corynebacterium glutamicum. After the deletion of pcaHG, vanAB and NCgl0324 and the introduction of a carboxylic acid reductase and mutated catechol O-methyltransferase, the strain produced 0.31 g/L vanillin from glucose (38). Using eugenol as a carbon source, Ashengroph et al. (39) screened for eugenol-degrading bacteria and found Pseudomonas resinovorans. Without further optimisation, resting cells of P. resinovorans SPR1 produced 0.24 g/L vanillin, i.e. a molar yield of 10 %.

Fungal vanillin synthesis

In 1996, Lesage-Meessen et al. (43) described a two-step process using different filamentous fungi, Aspergillus niger and Pycnoporus cinnabarinus, to convert ferulic acid to vanillin (Table 2 (17, 18, 43, 44, 47-51)). First, A. niger transformed ferulic acid to vanillic acid, then P. cinnabarinus reduced vanillic acid to vanillin. Vanillic acid was produced at a molar yield of 88 %; however, the reduction to vanillin only yielded 22 %. Low vanillin yields resulted from P. cinnabarinus predominantly producing methoxyhydroquinone from vanillic acid. After the addition of phenolic precursors, the final vanillin yield was 237 mg/L (43). Following optimisation of medium components (i.e. carbon and nitrogen), environmental factors (such as pH), and a one-step biotransformation process under statistically optimal conditions, the molar yield of P. cinnabarinus increased to 54 %, producing 126 mg/L vanillin (43). Screening for isoeugenol-tolerant yeasts from soil samples led to the identification of Trichosporon asahii. This strain effectively converted isoeugenol to vanillin in a resting cell biotransformation without any genetic modifications. T. asahii produced 2.4 g/L vanillin from 5 g/L isoeugenol (44). Van den Heuvel et al. (45) demonstrated the potential of the flavoprotein vanillyl alcohol oxidase (VAO) from Penicillium simplicissimum for the ‘natural‘ synthesis of vanillin from creosol and capsaicin. The study identified two primary enzymatic pathways: (i) a two-step oxidative hydroxylation of creosol and (ii) the deamination of vanillyl amine (derived from capsaicin hydrolysis). While the capsaicin route was highly efficient, achieving nearly 100 % molar yields, the creosol route was limited by competitive inhibition and the formation of non-productive covalent flavin-substrate complexes at neutral pH.

To address the problems in the creosol pathway, a later study (46) used directed evolution to optimise VAO for vanillin production from creosol. While the wild-type enzyme was hindered by the formation of a stable, non-productive covalent flavin adenine dinucleotide (FAD) N-5-creosol adduct, a single round of error-prone polymerase chain reaction (PCR) followed by high-throughput screening identified seven mutants with enhanced activity. Among them, a variant with four amino acid exchanges (I238T, F454Y, E502G and T505S) showed up to a 40-fold increase in catalytic efficiency (k_cat/K_m) at pH 10. The improvement resulted from destabilization of an abortive adduct, which shifted the reaction towards product formation.

Hansen et al. (17) achieved the first successful microbial vanillin biosynthesis from glucose by integrating the complete vanillin biosynthesis pathway into a single microorganism. This was demonstrated in the yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe, with vanillin production reaching 45 and 65 mg/L, respectively. Key enzymes included 3-dehydroshikimate dehydratase from Podospora pauciseta, an aromatic carboxylic acid reductase from Nocardia, and an O-methyltransferase from Homo sapiens. These findings suggested that the engineered yeasts represent a sustainable alternative to petrochemically derived vanillin (17). More recently, researchers used a MARE (minimal aromatic aldehyde reduction) yeast platform to improve vanillin yield from glucose by minimising vanillin reduction to vanillyl alcohol, optimising cofactor supply and reconfiguring the yeast's central metabolism. Overall, a total of 24 genetic modifications were introduced into S. cerevisiae, yielding a final vanillin titre of 365.55 mg/L (47). Building on these advances in metabolic engineering, Guo et al. (48) reported the first successful de novo synthesis of vanillin in the yeast Komagataella phaffii, reaching a titre of 731.3 mg/L using glucose as the sole carbon source. To prevent vanillin conversion to vanillic acid and vanillyl alcohol by endogenous enzymes, the researchers performed a systematic combinatorial knockout of 14 oxidoreductase genes. Notably, deletion of PAS14, PAS15 and PAS21 led to an 11.1-fold increase in vanillin production. Further improvements were achieved by rewiring metabolic pathways to increase the intracellular supply of l-tyrosine, NADPH and S-adenosylmethionine (SAM). To enhance the activity of the rate-limiting enzyme in the pathway, caffeic acid O-methyltransferase (NtCOMT), Guo et al. (48) combined molecular docking with saturation mutagenesis to generate the NtCOMT^N312A/H315N variant. This variant exhibited a widened substrate channel and a 49.7 % increase in catalytic activity. Together, these strategies establish K. phaffii as a robust industrial chassis for the production of vanillin.

Plant vanillin synthesis

Vanillin is produced in the pods of V. planifolia (4). In the plant, the enzyme vanillin synthase (VpVAN) appears to convert ferulic acid to vanillin, which is glycosylated to protect the plant from the toxicity of the compound (5). In Capsicum frutescens (hot chilli pepper), ferulic acid and vanillin were found to be the intermediates in its phenylpropanoid biosynthetic pathway (49). Using biolistics, C. frutescens cells were transformed with a codon-optimised VpVAN (Table 2). For the biotransformation, immobilized cell cultures were used and the vanillin content of transformed calli was 0.057 % compared to 0.0003 % of untransformed calli (49). To influence the phenylpropanoid pathway and phenolic compound accumulation in Ocimum sanctum, the VpVAN gene was overexpressed via Agrobacterium-mediated transformation. LC–MS/MS analysis showed increased vanillin production in transgenic lines compared to O. sanctum wild type, with the highest vanillin content being (1.98±0.01) mg/g extract (50). A similar approach was used with rice calli derived from embryonic rice cells, engineered to carry a codon-optimised VpVAN gene. The vanillin yield in the fresh callus of rice cell culture was 544.72 μg/g (51). Compared to bacterial and fungal systems, plant vanillin synthesis shows lower yields and is also affected by growth environment and processing costs. At present, this makes recombinant plant hosts unsuitable to meet the vanillin market demand (2).

Microalgae are another versatile host for vanillin production, due to their rapid growth, high biomass productivity and metabolic versatility. They can produce several important molecules such as fatty acids, feedstocks for food, fuel and phenolic compounds (52, 53). Rico et al. (54) showed that Phaeodactylum tricornutum growing in natural seawater under iron and copper stress could produce several precursors of vanillin biosynthesis pathways. P. tricornutum was able to produce protocatechuic, vanillic, caffeic, coumaric and ferulic acids (54). Cell cultures of Haematococcus pluvialis were studied for their biotransformation of ferulic acid, p-coumaric acid and coniferyl aldehyde. When immobilised, H. pluvialis produced 10.6 mg/L vanillin, 5.4 mg/L vanillic acid, 3.3 mg/L vanillyl alcohol, 1.6 mg/L protocatechuic acid, 1.1 mg/L p-coumaric acid and 1.4 mg/L p-hydroxybenzoic acid from these substrates (55). Following these results, Tazon et al. (40) hypothesised the presence of putative enzymes that could be used for vanillin biosynthesis. Based on the sequences of already characterised enzymes such as VpVAN of V. planifolia, FCS and ECH of Streptomyces sp. V-1 and IEM of Pseudomonas nitroreducens Jin1, potential microalgal homologues were identified using BlastP in National center for biotechnology information (NCBI) (56). In four species of microalgae (Chlamydomonas reinhardtii, P. tricornutum, H. pluvialis and Chlorella vulgaris), homologues to the already characterised enzymes were identified. While microalgae are not yet ready for large-scale vanillin production, they represent an interesting alternative to the already existing hosts and may see increased use in the future.

Mechanisms of vanillin toxicity in microorganisms

In general, vanillin is considered a membrane-active compound capable of forming pores in membranes and thus, destroying the membrane integrity in E. coli (59, 60). However, the underlying molecular mechanism of vanillin toxicity and its targets in the membrane remain unclear. A recent study suggested that adding exogenous vanillin disturbed key metabolic pathways in E. coli, leading to the accumulation of intracellular reactive oxygen species (ROS) and activation of stress response pathways (58). In S. cerevisiae, vanillin: (i) induced oxidative stress responses, possibly causing mitochondrial fragmentation (61), and (ii) repressed translation (62). However, the effects of endogenously produced vanillin may be different from those of externally added vanillin.

Biotechnological strategies to mitigate vanillin toxicity