Completing the acylation landscape in plants: GDSL enzymes join the BAHD and SCPL acyltransferase families

Abstract

Specialized metabolites are essential for plant adaptation to the environment. Their biosynthesis requires multiple chemical modifications of core structures, among which acylation is one of the most frequent. In plants, two families of acyltransferases are typically described. The BAHD superfamily, a family with many characterized members, uses mostly CoA esters as acyl donor, while the serine carboxypeptidase-like (SCPL) family uses mainly 1-O-β-D-glucose esters. Although these enzyme families catalyze most acyltransfer reactions in plants, enzymes from the Gly–Asp–Ser–Leu (GDSL) esterase/lipase family have recently been identified as acyltransferases in the biosynthesis of several specialized metabolites. These enzymes use various energy-rich donors, including CoA esters, chlorogenic acid, glycerol esters, and triacylglycerides, to acetylate a wide range of molecules. Interestingly, acyltransferases exhibit distinct subcellular localization: BAHD, SCPL, and GDSL are usually localized in the cytoplasm, vacuole, and apoplast, respectively. Together, these enzymes form a complex network of acyltransferases that modify specialized metabolites throughout plant cells. In this review, we aim to summarize the current knowledge on plant acyltransferases and their applications, particularly aiming to add GDSL transferases to BAHD and SCPL, which are usually considered for these types of reactions.

We provide an overview of the various acyltransferase families involved in specialized metabolic pathways, including the newly emerging GDSL acyltransferase family. Phylogenetic, biochemical, physiological, and biotechnological aspects are covered.

Introduction

Plants exhibit remarkable chemical diversity, enabling them to adapt to a wide range of selective pressures imposed by their environment (Weng et al., 2021). These molecules, specific to certain plant lineages and cell types or organs, are grouped together as specialized metabolites. They arise from a complex network of biosynthetic pathways that involve chemical modifications of a limited number of core structures derived from primary metabolism. Acylation constitutes, along with hydroxylation and methylation, some of the most common chemical modifications of these natural products (Wang et al., 2019). It notably affects stability, solubility, storage, and import (Wilson et al., 2016).

Acylation involves the transfer of an acyl group from an energy-rich donor molecule to the hydroxyl, amine, or thiol group of an acceptor molecule. This results in the formation of esters, amides, or thioesters, respectively. These reactions usually use a donor that has been activated through conjugation with CoA or 1-O-β-D-glucose (Wang et al., 2022). Although acyltransfer reactions are found in primary metabolism, notably in lipid biosynthesis, they are particularly important in the synthesis of specialized metabolites. In accordance, the expansion of gene families encoding these acyltransferases is closely associated with the chemical diversification during land plant evolution, highlighting their crucial role in adaptation to new environmental challenges (Kruse et al., 2022).

Two acyltransferase families have been canonically described in plants: the BAHD family and the serine carboxypeptidase-like (SCPL) family (Bontpart et al., 2015; Wang et al., 2022). The BAHD superfamily represents the largest group of plant acyltransferases. These enzymes mostly catalyze CoA-dependent acylation reactions and are involved in the biosynthesis of several compounds belonging to the major classes of specialized metabolites: phenylpropanoids, terpenoids, and nitrogen-containing compounds (Moghe et al., 2023). On the other hand, the SCPL family comprises acyltransferases that mostly use sugar-activated donors. Although fewer SCPLs have been characterized, they play critical roles in the biosynthesis of key specialized metabolites, such as sinapoyl esters in Brassicaceae and pigments (Mugford and Milkowski, 2012). Finally, a third family of acyltransferases has emerged in recent years: the GDSL acyltransferases, named after a conserved amino acid motif of Gly, Asp, Ser, and Leu. Traditionally known for their hydrolytic activity as GDSL esterase/lipase proteins (GELPs), some members have been shown to function as acyltransferases, catalyzing diverse reactions such as cutin biosynthesis, carotenoid esterification, and natural insecticide formation. Although not reviewed here, a fourth small family of acyltransferases, homologous to mammalian spermidine/spermine acetyltransferases (SSATs), catalyze amide synthesis in plants. In specialized metabolism, they have been identified in phenolamide synthesis, notably using tyramine and putrescine as acceptor molecules (Bassard et al., 2010; Roumani et al., 2025). Despite the first GDSL acyltransferases being identified >15 years ago and an increasing number of new members being discovered, no comprehensive review has yet covered this family. This review aims to fill this gap by providing an overview of the current knowledge on plant acyltransferases, with the objective to add GDSL acyltransferases, along with BAHDs and SCPLs, into the landscape of acylation mechanisms in plants.

The BAHD superfamily

BAHD acyltransferases represent one of the largest enzyme families in plants, exhibiting broad catalytic versatility and functional diversity. The name BAHD derives from the first four enzymes characterized in this family during the 1990s: benzoyl alcohol O-acetyl transferase (BEAT) (Dudareva et al., 1998), anthocyanin O-hydroxycinnamoyl transferase (AHCT) (Fujiwara et al., 1998), anthranilate N-hydroxycinnamoyl/benzoyl transferase (HCBT) (Yang et al., 1997), and deacetylvindoline 4-O-acetyl transferase (DAT) (St-Pierre et al., 1998). With the development of genomic, transcriptomic, and biochemical screening techniques, the discovery and characterization of new enzymes have rapidly accelerated, particularly in the last decade, resulting in the identification of hundreds of BAHD acyltransferases across various plant species. This family strictly uses CoA-activated molecules as acyl donor, although some of them can also use chlorogenic acid to form 3,5-dicaffeoylquinic acid (3,5-DiCQA) (Lallemand et al., 2012; Moglia et al., 2014; Miguel et al., 2020).

BAHDs have a conserved two-domain architecture and a solvent channel passing through the protein (Fig. 1). A His residue, located in the center of the solvent channel, acts as a base catalyzing the nucleophilic attack that allows the acyl transfer (Ma et al., 2005). First, both the acyl donor and acceptor enter the channel from opposite sides of the protein and bind independently near the active site (Unno et al., 2007) (Fig. 1). Then, the acyl acceptor, usually featuring a hydroxyl (-OH) or amine (-NH₂) group, is deprotonated by the catalytic His, allowing its nucleophilic attack on the electrophilic carbonyl carbon of the acyl-CoA donor (Wang et al., 2022). This generates a transient tetrahedral oxyanion intermediate between the two molecules, which quickly deprotonates to release the CoA-SH group (Xu et al., 2023). An ester or amide bond then forms. BAHD enzymes are defined by two conserved motifs that are critical for their catalytic activity (Ma et al., 2005). The first one, the HXXXD motif, is strictly conserved among BAHDs. The His residue is directly involved in the catalytic activity, while the Asp residue may stabilize the overall structure of the BAHD active site and the orientation of the His residue. The second motif, DFGWG, is located toward the C-terminus of the protein and may contribute to the CoA binding and overall conformation of the protein. Although the binding pocket architecture for the CoA thioesters is conserved among BAHDs of various classes, the acyl acceptor-binding site is highly variable, which is mainly responsible for the broad substrate range of these acyltransferases (Unno et al., 2007; C. Wang et al., 2021).

Fig. 1.

Crystal structure of Dm3MaT1 highlighting the conserved structure of BAHD enzymes. The crystal structure was retrieved from Unno et al. (2007). The two domains of BAHDs are displayed in blue and orange. The catalytic histidine residue (magenta) is located in the solvent channel formed by the two domains. The acyl donor pocket, with complexed malonyl-CoA, is shown in yellow on the front face. The acyl acceptor pocket is shown in cyan on the back face.

BAHDs are usually found in the cytosol, are monomeric and soluble, with no signal peptide (Bontpart et al., 2015). However, one BAHD involved in alkaloid synthesis has recently been localized to the mitochondria (Zeng et al., 2024), and another has dual targets in the cytoplasm and the vacuole (Moglia et al., 2014). Although they have been extensively studied in plants, BAHD homologs are also found in other lineages. In fungi, such as Saccharomyces cerevisiae and Fusarium species, a few members have been characterized and are involved in the biosynthesis of mycotoxins (Kimura et al., 1998; Alexander et al., 2002). Algal genomes typically contain 1–5 BAHD copies, though only one member, the hydroxycinnamoyl-CoA:quinate hydroxycinnamoyl transferase-like (CbHQT-like) from Chara braunii, has been assessed and shown to acetylate quinate (Kruse et al., 2022). Other acyl acceptors were not tested. Such activity is common with HQTs involved in the formation of chlorogenic acid in higher plants. While chlorogenic acid is absent in algae, CbHQT-like might play a role in the synthesis of pre-lignin phenolic polymers using other aromatic alcohols. The hydroxycinnamoyl-CoA:shikimate hydroxycinnamoyl transferase (HCT) from the moss Physcomitrium patens, which diverged early after plant terrestrialization, supports this hypothesis (Kriegshauser et al., 2021). This enzyme can form shikimate esters, as found in higher plants. Although PpHCT also forms threonate esters in vitro, which are present at much higher concentrations than shikimate esters in P. patens, functional genomics revealed that PpHCT is only involved in moss cuticle formation through formation of shikimate esters, revising the previously hypothesized role of threonate esters (Renault et al., 2017). Similarly, AaHCT6 from the hornwort Anthoceros agrestis can already synthesize hydroxycinnamic acid esters with shikimate and 2,3-dihydroxybenzoic acid, but also amides with 3-hydroxyanthranilic acid (Ernst et al., 2022). However, it is the drastic expansion of BAHDs throughout land plant evolution that makes it a superfamily. The number of BAHD copies increased from a few in algae to up to 200 in angiosperm genomes (Moghe et al., 2023). This diversification is characteristic of enzyme families associated with specialized metabolism, where weaker selective pressure favors rapid evolution, substrate promiscuity, and specific expression patterns (Roberts and Josephs, 2023). Consequently, the functional prediction of BAHD enzymes based only on sequence homology or in vitro assays should be approached with caution. Although many of the enzymes discussed in this review have not been functionally assessed, in planta validation in a native context constitutes the most reliable form of characterization.

In 2006, D’Auri first ranked BAHDs in six clades (D’Auria, 2006). Recently, this classification was revised due to the increasing number of characterized BAHDs and now includes eight clades (Kruse et al., 2022). Although there is some discrepancy between the phylogenetic classification of these enzymes in clades and the substrate class used, common biochemical characteristics can be attributed to each clade (Kruse et al., 2022; Moghe et al., 2023). As hundreds of BAHDs have been characterized to date and as several excellent reviews already cover many of these enzymes individually, only major activities of BAHDs in plant will be discussed (D’Auria, 2006; Bontpart et al., 2015; Petersen, 2016; Moghe et al., 2023; Xu et al., 2023).

BAHD acyltransferases play a significant role in phenylpropanoid metabolism. These enzymes catalyze the transfer of hydroxycinnamoyl or malonyl groups to a wide range of flavonoids, including anthocyanins and isoflavonoids (Luo et al., 2007; Suzuki et al., 2007; Taguchi et al., 2010; Rinaldo et al., 2015; Irmisch et al., 2018). In the lignin pathway, BAHD hydroxycinnamoyl transferases are essential for the formation of caffeoyl-shikimate, a key intermediate (Hoffmann et al., 2004; Wagner et al., 2007). In addition, they contribute to the acylation of coumaroyl and feruloyl moieties into cell wall polymers, such as hemicellulose and lignin (Wilkerson et al., 2014; Karlen et al., 2016; de Souza et al., 2018; Yang et al., 2024). BAHD acyltransferases are also responsible for the formation of several caffeic acid esters including chlorogenic acid (caffeoyl-quinate), phaselic acid (caffeoyl-malate), caftaric acid (caffeoyl-tartrate), and rosmarinic acid (caffeoyl-3,4-dihydroxyphenyl lactate) (Niggeweg et al., 2004; Berger et al., 2006; Sullivan, 2009; Sullivan and Zarnowski, 2011; Fu et al., 2021). Beyond cell wall polymers, these enzymes are involved in defense by catalyzing the mono- or polyacylation of hydroxycinnamic acids with polyamines and aromatic amines, such as putrescine, spermidine, spermine, tyramine, and serotonin, leading to the formation of phenolamides (Grienenberger et al., 2009; Onkokesung et al., 2012; Peng et al., 2016; C. Wang et al., 2021; Bernard et al., 2022; Cao et al., 2024). BAHDs further contribute to the production of volatile phenylpropanoids and benzenoids, as well as volatile terpenoids, through the conjugation of various alcohols or acetylation of coniferyl-alcohol in the biosynthetic pathway of eugenol and isoeugenol (Dudareva et al., 1998; Nam et al., 1999; Aharoni et al., 2000; D’Auria et al., 2002; Shalit et al., 2003; Dexter et al., 2007; Dhar et al., 2020). Members of this family are also involved in the esterification of alkaloids, including tropane and monoterpene indole alkaloids (Power et al., 1990; Grothe et al., 2001; Schmidt et al., 2015; Zeng et al., 2024), as well as in the formation of acyl sugars in the trichomes and roots of Solanum species (Nadakuduti et al., 2017; Kerwin et al., 2024). Additionally, BAHD enzymes catalyze key steps in the biosynthesis of the anticancer compound paclitaxel in Taxus species such as acetylation and benzoylation of the taxane core to form Baccatin III, and its further acylation with phenylalanyl-CoA and benzoylation (T. Wang et al., 2021; Liang et al., 2025; McClune et al., 2025) (Fig. 2). Finally, they participate in cuticle formation by catalyzing the elongation of very-long-chain fatty acids and the acylation of hydroxycinnamic esters, required for the biosynthesis of cuticular waxes and cutin formatio n (Xia et al., 1996; Haslam et al., 2012; Rautengarten et al., 2012; Wang et al., 2017).

Fig. 2.

The multiple roles of BAHD acyltransferases in taxol biosynthesis. BAHD enzymes are shown in red, and the colors of the catalyzed reactions are based on the type of acyl donor used. Dotted double arrows indicate that other enzymes besides BAHD are involved in the displayed biosynthetic steps. The linear biosynthetic pathway proposed by McClune et al. (2025) is shown, but taxol synthesis is rather a complex array of crossed enzymatic steps. TAT, taxadien-5α-ol-O-acetyltransferase; DBAT, 10-deacetylbaccatin III-10-O-acetyl transferase; TBT, taxane 2α-O-benzoyltransferase; T7AT, taxane C-7β-O-acyltransferase; BAPT, baccatin III:3-amino-3-phenylpropanoyl transferase; T3'NBT, taxoid-3'-N-benzoyltransferase.

Overall, the expansion and specialization of the BAHD superfamily in land plants enabled the diversification of plant specialized metabolism through new acylation mechanisms. These acyltransferases contribute significantly to the metabolic plasticity of plants in their changing environments by participating, alongside other enzymes, in the synthesis of volatile compounds for intra- and interspecies communication, protective compounds against UV exposure, defense compounds against biotic stresses, and rigid and hydrophobic cell wall polymers.

SCPL acyltransferases

Although BAHDs, which utilize CoA esters as acyl donors, are widely found in plant metabolism, they are not the unique protagonists of acyl transfer reactions. The involvement of 1-O-β-D-glucose esters as donors was first described over half a century ago, particularly in the biosynthesis of specialized metabolites (Kosuge and Conn, 1961; Kojima and Uritani, 1972). However, despite the purification of several enzymes catalyzing such reactions (Bokern et al., 1992; Gräwe et al., 1992; Szerszen et al., 1994), the first cloning and sequencing of a cDNA coding for one of these enzymes was not carried out until 2000 in Solanum pennellii (Li and Steffens, 2000). This enzyme, which forms 1,2-di-O-acyl-β-glucose from two molecules of 1-O-acyl-β-glucose, displays sequence homology with hydrolases belonging to the S10 family of peptidases named serine carboxypeptidases (SCPs).

This enzyme family has therefore been classified as SCP-like proteins (SCPLs). While SCPLs share the α/β fold and the conserved catalytic triad of SCPs (Ser, Hhis, and Asn), they primarily function as acyltransferases rather than hydrolases (Table 1) (Mugford and Milkowski, 2012; Bontpart et al., 2015; Ciarkowska et al., 2019). In Arabidopsis thaliana, a bioinformatic analysis has classified SCPs into two major clades based on sequence homology (Fraser et al., 2005). Clade I is further subdivided into Clades IA and IB. Clade IA comprises all the SCPLs that have been identified to date with acyltransferase activity. Using energy-rich 1-O-β-D-glucose esters as acyl donors, SCPL proteins participate in a wide range of cellular processes including specialized metabolism, hormone conjugation and storage, and glucose polyester synthesis. Besides the conserved motifs, SCPLs feature a predicted vacuolar targeting signal peptide at their N-terminus, and their localization in this organelle has been validated in several studies using fluorescence microscopy or immunogold labeling (Hause et al., 2002; Mugford et al., 2013; Nishizaki et al., 2013; Fu et al., 2021). Finally, to be active, some SCPLs undergo post-translational modifications such as glycosylation, cleavage of an internal peptide releasing two subunits which act as heterodimers, and cleavage of both the N- and C-terminus (Mugford et al., 2009; Yao et al., 2022; Wang et al., 2024). As a consequence, several of them have been shown to be inactive when expressed in prokaryotic systems and impossible to purify using N- or C-terminal tags (Shirley and Chapple, 2003; Weier et al., 2007; Stehle et al., 2008; Wang et al., 2024).

Table 1.

Characterized SCPL acyltransferases

Enzymes	Substrates		Products	Species	Localization	Roles	Class	References
	Acyl donors	Acyl acceptors
AT	1-O-acyl-β-D-glucose	1-O-acyl-β-glucose	1,2-di-O-acyl-β-glucose	Solanum pennellii	Glandular trichomes	Insect resistance	Acylsugars	Li and Steffens (2000)
AtSMT	Sinapoyl-glucose	Malic acid	Sinapoyl-malate	Arabidopsis thaliana	Leaves	UV protection	Sinapic acid esters	Lehfeldt et al. (2000)
AtSCT/BnSCT	Sinapoyl-glucose	Choline	Sinapoyl-choline	Arabidopsis thaliana Brassica napus	Seeds	—	Sinapic acid esters	Shirley and Chapple (2003); Baumert et al. (2005)
AtSST	Sinapoyl-glucose	Sinapoyl-glucose	Disinapoyl-glucose	Arabidopsis thaliana	Seeds	—	Sinapic acid esters	Fraser et al. (2007)
AtSAT	Sinapoyl-glucose	Anthocyanins	Sinapoylated anthocyanins	Arabidopsis thaliana	Senescing leaves	Oxidative stress response	Anthocyanins	Fraser et al. (2007)
DcSAT1	Sinapoyl-glucose	Cyanidin 3-xylosyl/glucosyl/galactoside	Cyanidin 3-xylosyl-sinapoylglucosyl-galactoside	Daucus carrota	Roots	Oxidative stress response	Anthocyanins	Xu et al. (2019); Curaba et al. (2020)
AtFPT2	Sinapoyl-glucose Caffeoyl-glucose p-Coumaroyl-glucose	Glycosylated flavonols	Saiginols	Arabidopsis thaliana	Flowers	UV protection	Flavanols	Tohge et al. (2016)
DgSCPL2	p-Hydroxybenzoyl-glucose	Delphinide derivatives	Viodelphin/Cyanodelphin	Delphinium grandiflorum	Flowers (vacuole)	Pollinator attraction	Anthocyanins	Nishizaki et al. (2013)
ECGT/CsSCPL4/CsSCPL5/DkSCPL1/DkSCPL2/VvGAT1/VvGAT2	1-O-galloyl-β-D-glucose	Catechin	Galloylated catechins	Camellia sinensis L. Diospyros kaki Vitis vinifera	Young leaves/buds	Pathogen defense	Flavanols	Liu et al. (2012); Chiu et al. (2016); Yao et al. (2022)
FvSCPL3/FaSCPL3	1-O-β-glucogallin	Glucose	Gallotanins	Fragaria vesca Fragaria ananassa	Fruits	Herbivore repelling	Hydrolyzable tannins	Zhang et al. (2025)
EpCAS/LsSCPL2/LsSCPL4	Chlorogenic acid	Caftaric acid	Chicoric acid	Echinacea purpurea Lactuca sativa	Whole plant (vacuole) Hairy roots (vacuole)	UV protection	Caffeoyltartaric acid esters	Fu et al. (2021); Xiang et al. (2025)
EpCCE1/EpCCE2	Chlorogenic acid	Chlorogenic acid	Isochlorogenic acid	Echinacea purpurea	Seeds (vacuole)	Germination control	Caffeoylquinic acid esters	Huang et al. (2024)
LsSCPL1/LsSCPL3	Chlorogenic acid	L-Tartaric acid	Caftaric acid	Lactuca sativa	Hairy roots (vacuole)	UV protection	Caffeoyltartaric acid esters	Xiang et al. (2025)
SCPL1	N-methyl anthraniloyl-O-glucose	Des-acyl avenacin A	Avenacin A-1	Avena strigosa	Roots (vacuole)	Pathogen defense	Triterpenes	Mugford et al. (2009, 2013)
LS	Phenyllactyl-glucose	Tropine	Littorine	Atropa belladonna	Roots	Pathogen defense	Tropane alkaloids	Qiu et al. (2020)
LjSCPL10/LjSCPL12	Syringyl-glucose	4-guanidinobutanol	Leonurine	Leonurus japonicus	Leaves (vacuole)	—	Alkaloids	Li et al. (2024)
IAInos synthase/OsIAA-At	1-O-indole-3-acetyl-β-D-glucose	Myo-inositol	IA–myo-inositol	Zea mays Oryza sativa	Seeds	Auxin homeostasis	Acylated hormones	Kowalczyk et al. (2003); Ciarkowska et al. (2018)
1-O-IAGlc-SugAc	1-O-(indole-3-acetyl)-β-D-glucose	Monosaccharides Disaccharides Trisaccharides	IA–saccharide conjugates	Zea mays	Seeds	Auxin homeostasis	Acylated hormones	Starzyńska and Kowalczyk (2012)
SCPL17	Benzoyl-glucose Sinapoyl-glucose	Glucosinolates	Benzoyl-glucosinolates Sinapoyl-glucosinolates	Arabidopsis thaliana	Seeds	Pathogen defense Auxin metabolism	Glucosinolates	Lee et al. (2012)

Early studies on SCPL acyltransferases focused on the formation of sinapate esters in Brassicaceae (Fig. 3). This pathway enables the synthesis of various phenylpropanoid derivatives in response to biotic and abiotic stresses, notably UV-B exposure. Even prior to the identification of SCPLs, the use of sinapoyl-glucose as an acyl donor was demonstrated in numerous steps of this pathway. Sinapoyl-glucose is produced by glucosyltransferases, such as UGT84A2 in A. thaliana which has been characterized in vitro (Lim et al., 2001). The first SCPL gene identified in this pathway, SNG1, encodes the sinapoyl-glucose:malate sinapoyltransferase (AtSMT), which catalyzes the formation of sinapoyl-malate from sinapoyl-glucose and malic acid (Lehfeldt et al., 2000). Subsequently, a close homolog, SNG2, was found in seeds of Arabidopsis. It encodes the sinapoyl-glucose:choline sinapoyltransferase (AtSCT). AtSCT catalyzes the acylation of choline to form sinapoyl-choline using sinapoyl-glucose as acyl donor (Shirley et al., 2001; Shirley and Chapple, 2003). In Brassica napus seeds, silencing of the gene responsible for sinapoyl-glucose synthesis significantly reduced the accumulation of disinapoyl-glucose, suggesting the requirement of this substrate for its synthesis (Baumert et al., 2005). In accordance, the sinapoyl-glucose:sinapoyl-glucose sinapoyltransferase (AtSST) was then identified in A. thaliana as a new SCPL catalyzing the condensation of two sinapoyl-glucose molecules to generate disinapoyl-glucose (Fraser et al., 2007). While other SCPLs, such as AtSMT, also exhibit this activity, it does not represent their primary function (Fraser et al., 2007). Additionally in the same study, the sinapoyl-glucose:anthocyanin sinapoyltransferase (AtSAT) was identified as an enzyme involved in flavonoid metabolism, where it catalyzes sinapoylation of anthocyanins. This acylation is expected to increase their stability and to modify their biochemical properties (Luo et al., 2022).

Fig. 3.

Biosynthesis of sinapic acid esters catalyzed by SCPL acyltransferases in the Brassicaeae. Sinapoyl glucose is imported into the vacuole, where SCPLs use it to acetylate several derivatives. SAT, sinapoyl-glucose:anthocyanin sinapoyltransferase; SCT, sinapoyl-glucose:choline sinapoyltransferase; SMT, sinapoyl-glucose:malate sinapoyltransferase; SST, sinapoyl-glucose:sinapoyl-glucose sinapoyltransferase; UGTs, uridine diphosphate-dependent glycosyltransferases.

The involvement of SCPL proteins in the acylation of anthocyanins has been observed in other species. In Delphinium grandiflorum, DgSCPL2 has been shown to participate in the synthesis of anthocyanins (Nishizaki et al., 2013). Using p-hydroxybenzoyl-glucose (pHBG) as an acyl donor, DgSCPL2 catalyzes polyacylation of delphinidine derivatives. These subsequent modifications lead to the formation of violdelphin and cyanodelphin, two anthocyanins that are responsible for the blue coloration of delphinium flowers. In purple carrots, a gene putatively involved in anthocyanin acylation was also identified but the enzyme was not characterized (Xu et al., 2019; Curaba et al., 2020). Still in polyphenol metabolism, the involvement of SCPLs was first suspected in the synthesis of tannins, notably proanthocyanidins in Vitis vinifera and Diospyros kaki (Ikegami et al., 2009; Carrier et al., 2013; Cheynier et al., 2013). In the tea plant, Camellia sinensis, the formation of galloylated catechin, a flavan-3-ol involved in defense against pathogens, was shown to be catalyzed by a purified epicatechin:1-O-galloyl-D-glucose O-galloyltransferase (ECGT). This enzyme, mapped to a SCPL gene, mediates catechin acylation using 1-O-galloyl-β-D-glucose as an acyl donor (Liu et al., 2012; Chiu et al., 2016). In 2022, a new study confirmed the involvement of CsSCPLs in catechin galloylation. Two genes, CsSCPL4 and CsSCPL5, were identified and evidenced to interact to promote the accumulation of galloylated catechins (Yao et al., 2022). While CsSCPL5 does not play a direct role in galloylation, it acts as a non-catalytic companion necessary for the post-translational modifications and activation of CsSCPL4 (Chen et al., 2025). Homologs of these proteins were characterized simultaneously in grapes, persimmons, Camellia oleifera (Yao et al., 2022), and later on in strawberries (Zhang et al., 2025). An additional SCPL, SCPL3, involved in galloylation was further identified in strawberries and C. oleifera. In both plants, SCPL3 catalyzes consecutive galloylation of β-glucogallin, resulting in the formation of 1,2,3,4,6-O-pentagalloylglucose (PGG) (Wang et al., 2024; Zhang et al., 2025). Finally, flavonol-phenylacyltransferase 2 (AtFPT2) has been identified as an SCPL involved in the formation of phenylacylated-flavonols (saiginols) in A. thaliana, which constitute a chemical signature of UV-B resistance in flowers (Tohge et al., 2016). In Echinacea purpurea, three enzymes from the SCPL family were recently shown to catalyze the formation of caffeic acid esters. One was shown to be involved in the last step of the synthesis of chicoric acid (dicaffeoyl-tartaric acid), a caffeic acid ester displaying promising health-promoting properties (Fu et al., 2021). The characterized enzyme, chicoric acid synthase (EpCAS), uses chlorogenic acid as an acyl donor to acetylate caftaric acid, suggesting that SCPL enzymes might be able to use donors other than glucose esters. The two additional SCPLs are involved in the formation of 3,5-DiCQA during seed germination (Huang et al., 2024). These two enzymes, named chlorogenic acid condensing enzyme 1 (EpCCE1) and 2 (EpCCE2), form this product using chlorogenic acid as both the acyl donor and acceptor. This activity has been suspected to be involved in the regulation of chlorogenic acid levels during germination as this metabolite has an inhibitory effect at this developmental stage. The reaction catalyzed by EpCCEs to give rise to 3,5-DiCQA is identical to that catalyzed by isochlorogenic synthase (IbICS), a GDSL protein, which will be discussed below (Miguel et al., 2020). Very recently, in Lactuca sativa, four enzymes belonging to the SCPL family that also used chlorogenic acid as acyl donor have been identified. LsSCPL1 and LsSCPL3 catalyze the synthesis of caftaric acid (monocaffeoyl-tartaric acid) using L-tartaric acid as the acyl acceptor, whereas LsSCPL2 and LsSCPL4 use caftaric acid as the acceptor to form chicoric acid (Xiang et al., 2025). Synthesis of caftaric acid using either a BAHD enzyme as discussed above or SCPL enzymes in closely related species highlights a convergent evolution which demonstrates the great adaptability of plant specialized metabolism.

SCPLs also catalyze reactions in terpenoid and alkaloid metabolism. In Avena spp., SCPL1 has been shown to catalyze the transfer of N-methylanthranilic acid or benzoic acid from their glucose esters to avenacin, a triterpenoid (Mugford et al., 2009). As sad7 mutants lacking SCPL1 are more susceptible to pathogen infection, the acylation of avenacin is expected to play a key role in defense against soil pathogens, particularly in oat roots (Mugford et al., 2009, 2013). In alkaloid metabolism, one SCPL protein has been identified in the synthesis of littorine, a tropane alkaloid found in Atropa belladonna (Qiu et al., 2020). This littorine synthase (LS) catalyzes the esterification of tropine and phenyllactate, using activated phenyllactyl-glucose as the acyl donor. Littorine then serves as an intermediate in the synthesis of alkaloids of interest in human health such as hyoscyamine and scopolamine. In Leonurus japonicus, SCPL proteins have recently been identified through genomic, metabolic, and biochemical studies in the biosynthesis of leonurine (Li et al., 2024), an alkaloid with interesting activities against cardiovascular disorders. These vacuolar enzymes, LjSCPL10 and LjSCPL12, displayed leonurine synthase activity in vitro using activated syringic acid glucose ester and 4-guanidinobutanol as substrates. Both genes were highly expressed in leaves, in accordance with the accumulation pattern of leonurine, and reduced levels of this metabolite were observed following the silencing of LjSCPL10 and LjSCPL12. Interestingly, both genes seem to arise from recent gene duplication and neofunctionalization in the Leonurus genus. LS, SCPL1, and LjSCPLs are clear examples of the wide array of reactions catalyzed by SCPL enzymes, which are then involved in the synthesis of metabolites from the three main families of specialized metabolites: polyphenols, terpenes, and alkaloids. SCPL activity has also been proposed in the acylation of glucosinolates in A. thaliana seeds, although the corresponding enzymes have not been fully characterized (Lee et al., 2012).

Lastly, enzymes from the SCPL family have been characterized in auxin homeostasis. In maize, two proteins from the SCPL family, 1-O-(indole-3-acetyl)-β-D-glucose:myo-inositol indoleacetyl transferase (IAInos synthase) and 1-O-(indole-3-acetyl)-β-D-glucose:sugar indoleacetyl transferase (1-O-IAGlc-SugAc), were shown to efficiently catalyze the conjugation of indole-3-acetic acid (IAA) with either inositol or saccharides, respectively (Kowalczyk et al., 2003; Starzyńska and Kowalczyk, 2012). IAA esters are considered as storage forms of IAA. Therefore, both enzymes could be involved in the homeostasis of free IAA. While both enzymes use 1-O-IAGlc as the acyl donor, IAInos synthase forms IA–myo-inositol using myo-inositol, whereas 1-O-IAGlc-SugAc can use various monosaccharides such as mannose, glucose, and galactose, or polysaccharides such as melibiose, gentiobiose, and raffinose as acyl acceptors, resulting in the formation of IA–saccharide conjugates. It is important to note here that IAInos synthase has also been shown to exhibit activity using monosaccharides (Ciarkowska et al., 2018). A homolog of IAInos synthase was later identified in rice (Ciarkowska et al., 2018).

All known SCPL acyltransferases belong to Clade IA (Fig. 4). Clade IA itself can be divided into two distinct subclades. Subclade IA.2 comprises enzymes that utilize sinapoyl-glucose as an acyl donor. On the other hand, subclade IA.1 consists of two main groups: one includes enzymes with galloyltransferase activity, while the other contains enzymes that use chlorogenic acid as an energy-rich acyl donor. Additionally, SCPL enzymes involved in the acylation of terpenes and alkaloids are scattered throughout Clade IA with no clear evolutionary links. SCPL1 is found in a clade separate from other SCPL acyltransferases, probably corresponding to monocot SCPLs as already suggested (Mugford et al., 2009). A phylogenetic comparison of the SCP family in higher plants and algae revealed that Clade IA, which contains SCPLs with transferase activities, is specific to land plants. In contrast, the algal and moss genomes only contain members of Clades IB and II (Mugford and Osbourn, 2010). The diversification of this family in land plants has given rise to a new acyltransferase function using glucose esters, notably linked to the synthesis of specialized metabolism, which promotes, together with BAHD, their metabolic diversity.

Fig. 4.

Phylogenetic relationships between enzymes of the SCPL family. The multiple alignment of selected sequences was performed using the Muscle algorithm in Mega 7.0 (Edgar, 2004; Kumar et al., 2016). The phylogenetic tree was constructed using the maximum-likelihood method with the same software (Felsenstein, 1973). The length of the lines indicates the relative distance between nodes, using the scale bar provided. Bootstrap values were calculated with 1000 replicates. Values higher than 75 are shown. Characterized SCPLs were aligned together with uncharacterized SCPLs from A. thaliana to replace previously described clades (Fraser et al., 2005). * marks enzymes that have not been biochemically characterized. AT (AAF64227); AtFPT2 (Q8S8P0); AtSAT (O64810); AtSCT (Q8VZU3); AtSMT (Q8RUW5); AtSST (Q8H780); BnSCT (AAQ91191); CoSCPL4 (MZ462936); CoSCPL5 (MZ462937); CsSCPL4 (MZ462934); CsSCPL5 (MZ462935); DcSAT1 (MK572824); DgSCPL2 (BAO04183); DkSCPL1 (AB195285); DkSCPL2 (AB473186); EpCAS (QRI59129); EpCCE1 (PP341507); EpCCE2 (PP341508); LjSCPL10; LjSCPL12; LS (CAJ40778); OsIAA-At (EEE56946); SCPL1 (ACT21078); SCPL17 (AAS99709); VvGAT1 (XM_002272080); VvGAT2 (XM_002269403).

Regarding their reaction mechanism, hydrolases generally use a double displacement (ping–pong) mechanism. Based on their homology with SCP proteins, SCPLs were first proposed to use the same mechanism to perform acylation (Shirley and Chapple, 2003), although some biochemical data suggest that they use a random bi–bi catalytic mechanism (Stehle et al., 2009). Unlike the ping–pong mechanism, this second hypothesis suggests that both the acyl donor and acceptor are fixed to the enzyme prior to the nucleophilic attack, thereby preventing water access to the active site. Analysis of kinetic parameters has not yet led to a clear conclusion regarding the type of reaction used (Shirley and Chapple, 2003; Stehle et al., 2009; Wilson et al., 2016). However, the presence of minor hydrolytic activity towards acyl donors in several SCPLs is in favor of a ping–pong mechanism (Kowalczyk et al., 2003; Fu et al., 2021). Several SCPLs have been shown to localize to the vacuole, while the glucosyl transferases that form their acyl donors are cytosolic enzymes. In order to be used by SCPLs, glucose esters may therefore be imported into the vacuole. This process is thought to be mediated by MATE or ABC-type transporters, as these have been shown to mediate the import of flavonoid glycosides into the vacuole. However, none of these transporters has yet been functionally characterized (Goodman et al., 2004; Zhao and Dixon, 2009; Zhao et al., 2011).

GDSLs: a new acyltransferase family

GDSL esterases/lipases represent an additional class of α/β hydrolases characterized by the presence of a conserved Gly–Asp–Ser–Leu (GDSL) motif at their N-terminus. This feature sets them apart from classical lipolytic enzymes, which typically exhibit a GXSXG motif located toward the center of the protein sequence (Ding et al., 2019). Like other lipases, GDSLs contain a catalytic serine residue in the active site that performs the nucleophilic attack on the substrate (Akoh et al., 2004). These esterases participate in the hydrolysis of a broad range of substrates and play a role in several biological processes, including morphological development, abiotic stress response, pathogen defense, and specialized metabolism (Lai et al., 2017). Apart from the GDSL motif, GDLSs contain five conserved blocks with four invariable residues: Ser, Gly, Asn, and His. Due to the presence of these four amino acids, these proteins are often referred to as SGNH hydrolases (Shen et al., 2022), although other enzymes also share this classification.

Members from the GDSL family are ubiquitously found in plants, bacteria, and animals (Akoh et al., 2004). GDSLs are present throughout the plant kingdom, from charophytes and chlorophytes to land plants (Vujaklija et al., 2016). However, the number of GDSL copies varies considerably during plant evolution. Embryophyta possess ∼100 copies in their genome (Chepyshko et al., 2012; Lai et al., 2017; Cao et al., 2018; Ni et al., 2020), whereas green algae contain ∼10 times fewer copies (Vujaklija et al., 2016). This significant increase in the number of GDSL copies is closely linked to plant terrestrialization as P. patens and Selaginella moellendorffii, which diverged early after the appearance of land plant, already possess ∼50 and 100 GDSL copies, respectively (Vujaklija et al., 2016). These observations suggest that the functional diversification of GDSL proteins following gene duplication may have contributed to the adaptation of plants to the constraints of terrestrial environments, such as UV exposure, water loss, and new pathogen infection (Renault et al., 2017). This hypothesis gains further support from the characterization of several GDSLs directly involved in the metabolism of structural polymers such as cutin (Takahashi et al., 2010; Girard et al., 2012; Yeats et al., 2012), hemicellulose (Zhang et al., 2017, 2019), and suberin (Ursache et al., 2021).

Although the majority of described GDSL enzymes exhibit hydrolytic activity, numerous protein members with acyltransferase functions have been recently discovered (Table 2). These results have led us to add this third family, GDSL acyltransferases, to the two families of acyltransferases that are canonically described in plants: BAHD and SCPL, as already suggested in Petersen (2016).

Table 2.

Characterized GDSL acyltransferases

Enzymes	Substrates		Products	Species	Localization	Roles	Class	References
	Acyl donors	Acyl acceptors
SlCGT	Chlorogenic acid	Glucarate Galactarate	Caffeoyl-glucarate Caffeoyl-galactarate	Solanum lycopersicum	Leaves (apoplast)	—	Caffeoyl esters	Teutschbein et al. (2010)
TcGLIP	Chrysanthemoyl-CoA Pyrethroyl-CoA	Pyrethrolone Jasmololone Cinerolone	Pyrethrin Jasmolin Cinerin	Tanacetum cinerariifolium	Bud/leaves (apoplast)	Insect defense	Pyrethrins	Kikuta et al. (2012)
CD1/GDSL1	2-mono(10,16-dihydroxyhexa- decanoyl)glycerol (2-MHG)	Growing cuticle	2-MHG polymer (cuticle)	Solanum lycopersicum	Fruits (apoplast)	Cuticle formation	Cuticular waxes	Yeats et al. (2012); Girard et al. (2012)
XAT	Trilinolein (triacylglyceride) and other lipids	Xanthophylls	Xanthophyll esters with fatty acids	Triticum aestivum	Seeds (apoplast)	Oxidative stress response	Carotenoids	Watkins et al. (2019)
IbICS	Chlorogenic acid	Chlorogenic acid	Isochlorogenic acid	Ipomoea batatas	Roots	—	Caffeoylquinic esters	Miguel et al. (2020)

The first GDSL enzyme with transferase activity was reported in tomato (Solanum lycopersicum) leaves in 2010 (Teutschbein et al., 2010). This apoplast-localized GDSL, named chlorogenate:glucarate caffeoyltransferase (SlCGT), transfers the caffeoyl group from chlorogenic acid to glucarate or galactarate to yield caffeoyl-glucarate and caffeoyl-galactarate, respectively. Generally, acyltransferases utilize CoA-activated molecules (BAHD) or 1-O-β-glucose esters (SCPL). Therefore, SlCGT initially stood out as an atypical acyltransferase able to use chlorogenic acid (5-CQA). Since its discovery, additional BAHD, SCPL, and GDSL enzymes using 5-CQA as a substrate have been characterized (Lallemand et al., 2012; Moglia et al., 2014; Miguel et al., 2020; Fu et al., 2021; Huang et al., 2024). Interestingly, this enzyme displays regiospecificity toward 5-CQA and is unable to metabolize the chlorogenate isomers 3-CQA and 4-CQA. SlCGT was first thought to use the same Ser, Asp, His catalytic triad as other GDSLs, but mutagenesis experiments performed on these amino acids suggested that they were not fully responsible for the acyltransferase activity. In fact, enzyme activity was reduced by no more than 40% when these residues were mutated. Furthermore, SlCGT exhibits a complete loss of hydrolytic activity, indicating an evolutionary shift toward acyltransferase activity in this GDSL.

Subsequently, in the Asteraceae species Tanacetum cinerariifolium, a second GDSL with transferase activity was identified (Kikuta et al., 2012). In this species, TcGLIP catalyzes the efficient synthesis of pyrethrins by transferring acyl groups from CoA thioesters to pyrethrolone. Interestingly, this enzyme also exhibits hydrolytic activity toward pyrethrin I in vitro and catalyzes the reverse acylation reaction in the presence of CoA. Both processes result in the release of pyrethrolone, although with a much slower reaction rate. TcGLIP uses chrysanthemoyl-CoA preferentially as an acyl donor, with jasmololone being the preferred acceptor. Subcellular localization following fusion of TcGLIP with enhanced green fluorescent protein (eGFP) showed an apoplastic expression pattern. Unlike the previously identified SlCGT, mutagenesis of the catalytic Ser in TcGLIP resulted in a complete shutdown of both transferase and hydrolytic activities. This enzyme then appears to use the same nucleophilic Ser as hydrolytic GDSLs, implying a distinct evolution of catalytic determinants between TcGLIP and SlCGT. In T. cinerariifolium, pyrethrins play an important role in defense against insects and have been used for a long time as natural insecticides. Thus, understanding their synthesis could facilitate the production of natural alternatives to synthetic pyrethroids.

Cutin is an apolar polymer that emerged early in the process of plant terrestrialization (Renault et al., 2017). As the primary constituent of the cuticle, it serves as a hydrophobic barrier that is crucial for preventing water loss and infection by pathogens (Nawrath, 2006). In tomato, a GDSL protein has been found to be involved in cutin synthesis (Yeats et al., 2012). This enzyme was identified during the screening of fruit cuticle-deficient mutants. One of these mutants, cutin deficient 1 (cd1), exhibited a drastic reduction in fruit cuticle with ∼5–10% of the wild-type content. Genetic mapping revealed that the linked mutation was in a GDSL-coding gene. Localized in the apoplast near the developing cuticle, the CD1 protein possesses transferase activity and catalyzes transesterification between 2-mono(10,16-dihydroxyhexadecanoyl)glycerol (2-MHG) units, leading to cuticle polymerization. Mutation of the catalytic Ser in CD1 resulted in a complete loss of enzyme activity, as observed in TcGLIP. Here, CD1 represents the first example of a GDSL enzyme with acyltransferase activity involved in the metabolism of structural polymers. The identification of this enzyme in the synthesis of these components is important for understanding the evolutionary mechanisms that allowed plant development on land and increased their resistance to various biotic and abiotic stresses. At around the same time, another GDSL, SlGDSL1, was identified as a second acyltransferase catalyzing cutin synthesis and deposition in tomato fruit (Girard et al., 2012).

In bread wheat, the GDSL xanthophyll acyltransferase (XAT) allows the esterification of lutein and other xanthophylls with fatty acids (Watkins et al., 2019). Esterification of these carotenoids increases their stability to UV light and heat treatments, both in vitro and in vivo (Watkins et al., 2019). Given their recognized status as efficient antioxidants (Pérez-Gálvez et al., 2020), XAT indirectly contributes to plant resistance against oxidative stress. Comprehensive analyses involving mutant studies, and in vitro experiments confirmed that XAT catalyzes the efficient synthesis of diverse xanthophyll esters, using several lipid substrates as the acyl donor. Triacylglycerides, such as trilinolein, were identified as the preferred acyl donors, with lutein as the favored acyl acceptor. This enzyme is predominantly expressed in seeds, with high expression in the endosperm. At the level of the plant cell, XAT exhibits an apoplastic accumulation pattern, in accordance with the localization of other GDSLs. As previously described, XAT exhibits both transferase and hydrolytic activities, with the Ser of the catalytic triad playing a key role in enzyme activity.

Finally, our team recently characterized a GDSL in Ipomoea batatas (Miguel et al., 2020). This enzyme, isochlorogenic acid synthase (IbICS), catalyzes a transferase reaction between two molecules of chlorogenic acid, enabling the synthesis of 3,5-DiCQA, a promising molecule for human health (Kurata et al., 2007). As the synthesis of 3,5-DiCQA was first thought to be performed by BAHDs as shown in tomato and coffee (Lallemand et al., 2012; Moglia et al., 2014), IbICS represents an original pathway leading to the synthesis of this compound. IbICS displays higher efficiency than BAHD for 3,5-DiCQA synthesis. This enzyme was expressed in Pichia pastoris and successfully led to an efficient platform for the industrial production of 3,5-DiCQA. IbICS was stable in aqueous solutions and showed production of 3,5-DiCQA with a 61.3% conversion rate. The synthesis of 3,5-DiCQA is an additional example of convergent evolution as in E. purpura its synthesis relies on the activity of SCPL members as described above.

The characterization of these several GDSLs with transferase activities has revealed common characteristics among these enzymes. Firstly, all GDSL transferases feature a signal peptide at their N-terminus, which is predicted to direct the protein into the secretory pathway. Indeed, subcellular localization analyses, performed for all characterized enzymes except IbICS, revealed an apoplastic signal, which is consistent with their role in the synthesis and catabolism of parietal compounds, as well as their involvement in defense against pathogens. All of these GDSLs possess the conserved blocks of the GDSL family, including the catalytic triad: Ser, Asp, His. With the exception of SlCGT, mutation of one of these residues strongly decreases enzyme activity, indicating their crucial role in transferase activity. When produced in prokaryotic heterologous systems, many GDSLs with transferase or hydrolytic activities remain inactive. This was notably the case for SlCGT and IbICS which were not active when produced in Escherichia coli. This lack of activity may be attributed to poor protein solubility or to the absence of some of the post-translational modifications in these organisms (Miguel et al., 2020). Therefore, the most suitable system for their expression seems to be Nicotiana benthamiana. This species provides a fast and efficient platform for the transient production of catalytically active GDSLs in a cellular environment close to that of other plants (Bally et al., 2018). Despite the diversity of substrates employed by GDSL transferases, many display hydrolytic activity toward their products. Enzymes such as esterases and lipases that exhibit both hydrolytic and acyltransferase activity have been well documented (Müller et al., 2021). GDSL proteins appear to operate via a ping–pong mechanism for both types of reaction. Initially, the enzyme interacts with the first substrate, which is typically a CoA-activated donor or an ester, and this results in the formation of an enzyme–substrate complex. The enzyme then interacts with the second substrate and catalyzes a nucleophilic attack between the fixed donor and the acceptor molecule (Akoh et al., 2004). The acceptor can be a metabolite, in the case of an ester-forming reaction, or water in the case of hydrolytic activity. Hydrolytic and transferase activities are then closely related, as esterase activity can be considered as a transfer of an acyl group from an ester to a water molecule as the acceptor (Müller et al., 2021), resulting in the breakdown of the ester bond. To favor transferase over hydrolytic activity, GDSLs with transferase activity may employ mechanisms to minimize access of free water to the active site. This can be achieved either by creating a hydrophobic environment or by sequestrating water through interactions with amino acids surrounding the active site (Müller et al., 2021). GDSL proteins have been shown to undergo conformational changes following docking of the substrate (Akoh et al., 2004). For instance, a flexible loop located at the entry point of the SGNH hydrolase-binding pocket helps transition from a distorted to a structured form. This stabilizes the oxyanion hole and obstructs solvent access to the active site (Wang et al., 2023). This mechanism creates a hydrophobic environment, which reduces hydrolytic activity. GDSLs that exhibit transferases activity may possess hydrophobic active sites and a high affinity for the acceptor in the transferase reaction, to prevent this side activity. These characteristics could arise from several changes in the protein sequence, explaining why specific amino acids cannot be strictly attributed to transferase over hydrolytic activity (Yamashiro et al., 2022). Consequently, GDSL transferases do not cluster in phylogenetic analyses, which slows their identification based on sequence homology (Fig. 5). As an example, chlorogenate hydrolases were recently identified in Cichorium intybus (CiCQEs) and Helianthus annuus (HaCAE) while searching for homologs of IbICS, which acts as a transferase in 3,5-DiCQA synthesis (Guo et al., 2025; Mallavergne et al., 2025) (Fig. 5).

Fig. 5.

Phylogenetic classification of characterized GDSLs. The phylogenetic tree was constructed as described in Fig. 4. GDSLs with transferase activity are marked with orange squares. Green asterisks highlight homologs of IbICS with hydrolase activity. AtCDEF1 (OAO99887); AtCUS2 (NP_198322); AtDAD1 (Q948R1); AtEXL4 (Q9LY09); AtEXL6 (NP_177721); AtGELP12 (Q3E7I6); AtGELP22 (ABF58999); AtGELP38 (AAO24551); AtGELP49 (O64468); AtGELP51 (O80470); AtGELP55 (O22927); ATGELP73 (Q9M2R9); ATGELP77 (NP_001329284); AtGELP81 (Q8VY93); AtGELP96 (Q9FHQ1); AtGER1 (Q9FLN0); AtGLIP1 (NP_198915); AtPLIP2 (F4HXL0); AtPLIP3 (Q940L4); AtSFAR4 (Q9STM6); BnSCE3 (AAX58135); CaGLIP1 (AAX20033); CD1 (JF968592); CiCQE1 (PX093951); CiCQE3 (PX093953); GhGLIP (AWD40208); HaCAE (XP_022035318); IbICS (MT291823); JmJNP1 (ABY59947); OsBS1 (XP_015626582); OsDARX1 (XP_015637555); OSP1 (NP_178536); SlCGT (FR667689); SlGDSL1 (XP_004233002); TcGLIP (AFJ04755); XAT (MN233788).

Recent advances have made the GDSL family an attractive target for identifying new acyltransferases in plants. However, this family is still poorly characterized, especially considering the large number of GDSL copies in the genomes of land plants. Given the wide array of substrates used by these proteins, it is reasonable to assert that numerous GDSL transferases remain to be identified that play roles in various biological processes. Some of them are anticipated to have acyltransferase activities, but they require further biochemical characterization, such as the GDSLs involved in suberin polymerization in A. thaliana roots (Ursache et al., 2021), CER-ZV, which is likely to be involved in cutin deposition in barley leaves (Li et al., 2017), and cutin synthase 2 (AtCUS2), which is required for cutin formation in A. thaliana sepals (Hong et al., 2017). The GDSL family is often overlooked as a source of acyltransferases in plants compared with the BAHD and SCPL families due to the limited number of members characterized so far. Thus, this review brings them back to center stage as an efficient catalyst of acyltransferase reactions in plants.

Bioengineering using acyltransferases from plants

The identification and characterization of plant acyltransferases from various biosynthetic pathways have created new opportunities for the microbial production of complex natural products using engineered hosts, mainly the yeast S. cerevisiae. Enzymes from the three previously described families have been used to reconstruct both partial and complete pathways for synthesizing health-promoting compounds. This bypasses the need to extract these compounds from plants, where they are often poorly accumulated and/or lineage specific, and avoids costly organic synthesis processes. The insertion of acyltransferases, either alone or alongside a set of other genes in yeast, allows the production of the desired compounds by providing specific substrates. Notably, this has been achieved with hydroxycinnamoyl transferases from the BAHD family, which are co-expressed with 4-coumarate:CoA ligase (4CL) to generate CoA-activated donors. This enables the production of chlorogenic acid, shikimic acid esters, rosmarinic acid, hydroxycinnamoyl-glycerol conjugates, clovamide, and several trihydroxycinnamoyl-spermidines in yeast (Kim et al., 2013; Eudes et al., 2016; Bouchez et al., 2019; Perrin et al., 2021). In certain cases, this strategy could also promote the synthesis of new-to-nature molecules that may have original biological properties (Perrin et al., 2021). Another example is the use of a single GDSL enzyme, IbICS from sweet potato, which led to the production of 3,5-DiCQA from chlorogenic acid as the unique substrate in yeast (Miguel et al., 2020). Although substrate promiscuity in acyltransferases allows the formation of novel or rare compounds, it also diverts metabolic flux away from the desired products. This presents both opportunities and challenges for metabolic engineering. The final goal of pathway identification remains the complete de novo reconstruction of the biosynthesis of valuable compounds in yeast, independent of specific substrate supplementation. Several examples of full pathway reconstruction exist to date. For instance, the expression of the BAHD salutaridinol 7-O-acetyltransferase (PsSalAT) from Papaver somniferum, along with >20 other genes, enabled the biosynthesis of thebaine and hydrocodone in yeast, two alkaloids analogous to morphine and codeine (Galanie et al., 2015). In the biosynthesis of tropane alkaloids, cinnamoyl-tropine was formed in yeast following 15 genetic modifications, including the insertion of Erythroxylum coca cocaine synthase (EcCS), a BAHD catalyzing the last biosynthetic step (Srinivasan and Smolke, 2019). The synthesis of two other tropane alkaloids, scopolamine and hyoscyamine, was achieved by fully reconstructing the biosynthetic pathway in yeast, including the LS acyltransferase, from the SCPL family, which allowed the formation of littorine, an intermediate in their biosynthesis (Srinivasan and Smolke, 2020). For taxol production, its full pathway reconstruction has not been achieved, although several BAHD enzymes involved in intermediate steps have been expressed in yeast as part of partially reconstructed pathways (DeJong et al., 2006; Engels et al., 2008; Liu et al., 2024; Liang et al., 2025). Vinblastine is an important anti-cancer compound from Catharanthus roseus. Its complete pathway was inserted into yeast through the integration of >30 genes and host metabolism engineering (Zhang et al., 2022). This system allows the production of vindoline and catharanthine, followed by in vitro chemical coupling to produce vinblastine. An inserted BAHD enzyme, the deacetylvindoline-O-acetyltransferase (CrDAT), catalyzes the final step in vindoline biosynthesis. Lastly, expressing the BAHD acyltransferase HpBAHD3 in a yeast background that produces betalain enabled the formation of phyllocactin, a malonylated betalain pigment (Babaei et al., 2023; Glitz et al., 2025). These examples demonstrate how individual acyltransferases, when placed within a broader context of pathway reconstruction in engineered yeast, unlock the production of a wide array of specialized metabolites with high commercial and therapeutic value.

In addition to their roles in specialized metabolism, several acyltransferases, particularly from the BAHD family, are involved in the biosynthesis of cell wall polymers. These enzymes have drawn significant interest due to their industrial relevance in cellulose processing for biofuels and papermaking. Removing lignin and other parietal polymers is a costly process in these industries, which drives research towards altering plant cell wall composition to improve digestibility. Early approaches have focused on reducing total lignin content by silencing genes in the monolignol biosynthetic pathway, including HCTs. This strategy has been applied to species such as Brachypodium distachyon and alfalfa, resulting in reduced lignin levels (Chen and Dixon, 2007; Pattathil et al., 2012; Tong et al., 2015; Serrani-Yarce et al., 2021). However, these reductions often lead to severe growth defects and reduced biomass yields, which limit their utility on an industrial scale (Shadle et al., 2007; Li et al., 2010). Later studies have then focused on modifying the composition of lignin, and other parietal compounds, using BAHDs. These enzymes impact the proportion of ferulate and coumarate derivatives that are incorporated, without modifying total lignin content (Bartley et al., 2013; Cheng et al., 2013; Petrik et al., 2014; Wilkerson et al., 2014; Smith et al., 2015; Sibout et al., 2016; de Souza et al., 2018). These modifications led to improved saccharification efficiency without affecting plant growth or biomass. Regarding caffeic acid esters, such as chlorogenic acid, they are better substrates for polyphenol oxidases (PPOs) than caffeic acid. As oxidation of these polyphenols leads to formation of reactive quinones that induce tissue browning, silencing the genes responsible for their synthesis, such as HQT, could result in vegetables with reduced browning for applications in the food industry (Amaki et al., 2011). Finally, acyltransferases are also involved in anthocyanin synthesis. Introducing these enzymes into other species could alter their anthocyanin profiles and floral pigmentation, as demonstrated with the Vitis vinifera anthocyanin 3-O-glucoside-6''-O-acyltransferase (Vv3AT) in tobacco (Rinaldo et al., 2015). Modifying this profile can alter the color of flowers for ornamental purposes or change how plants interact with their pollinators.

Conclusion

Since the initial characterization of BAHD acyltransferases in the late 1990s, a diverse array of plant acyltransferases has been identified, highlighting their central roles in plant development, specialized metabolism, and responses to environmental challenges. The evolutionary expansion of these enzyme families, which closely follows plant terrestrialization, suggests a complex functional diversification. However, many members of these families, particularly GDSL and SCPL enzymes, remain largely uncharacterized in higher plants. Recent advances in genomics, transcriptomics, and high-throughput biochemical screening methods have accelerated the pace of acyltransferase discovery, leading to the identification of a growing number of non-BAHD enzymes with unique catalytic properties and functions. Therefore, recent findings on GDSL acyltransferases in both specialized metabolism pathways and cell wall formation suggest that they are key contributors to the acylation landscape in plants, together with SCPLs and BAHDs (Fig. 6).

Fig. 6.

Multiple localizations and functions of acyltransferases in plant cells. Simplified scheme of a plant cell showing the subcellular localization of the three acyltransferase families: SCPL, GDSL, and BAHD. These enzymes use different activated donors to acylate their acceptor molecules. On the other hand, chlorogenic acid serves as an acyl donor for enzymes belonging to each family. Depending on the type of metabolites synthesized and their localization, acyltransferases play important roles in several biological processes, which are shown in blue. Yellow stars represent oxidative stresses. CW, cell wall; ER, endoplasmic reticulum; GA, Golgi apparatus; IAGlc, 1-O-indole-3-acetyl-β-D-glucose; IAInos, indole-3-acetyl-myo-inositol; N, nucleus; PM, plasma membrane.

Funding

AM was supported by a doctoral fellowship from University of Lille, the Region Hauts-de-France, and Florimond-Desprez. This review is part of the project of the Chicory for One Health (CHIC41H) joint team between Florimond-Desprez and the University of Lille, funded by the European Union, and the European Regional Development Fund of the Region Hauts-de-France.