Enzymatic galactosylation of phenolic compounds by permeabilized cells of Pantoea anthophila

Abstract

Abstract

The permeabilized cells of two Pantoea anthophila strains (BI 55.2 and BI 69.1) were evaluated for galactosylation of phenolic compounds with different molecular structure (gallic acid, caffeic acid, catechin, phlorizin, puerarin and mangiferin), the highest conversion was observed on puerarin (36.5 ± 8.4 and 53.3 ± 7.1), followed by phlorizin (33.5 ± 4.6 and 41.8 ± 0.4) and caffeic acid (21.4 ± 0.05 and 32.5 ± 0.5), respectively, no reaction was observed on mangiferin. They also exhibited a high catalytic promiscuity for most of the phenolic compounds (3 to 7 different galactosides). BI 69.1 was selected for a further kinetic characterization on phlorizin and puerarin as acceptors, elongation of the galactosyl change was observed, the mass spectrometry determined by UPLC-ESI-Qtof-MS showed the synthesis of digalactosides (759.1 and 739.08 m/z) from monogalactosides as starters (597.13 and 577.12 m/z) for phlorizin and puerarin, respectively. The major phenolic galactoside was purified and the molecular structure was elucidated by NMR, corresponding to a β-D-(1 → 6) puerarin monogalctoside.

Key points

• Permeabilized cells efficiently galactosylate diverse phenolic substrates

• High catalytic promiscuity observed with formation of multiple galactosides

• NMR confirmed a β-D-(1→6) monogalactoside structure derived from puerarin

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s00253-025-13688-8.

Introduction

Phenolic compounds are the most abundant secondary metabolites in plants, vegetables and fruits, characterized by having, at least, one phenol group (Lattanzio 2013). It has been reported that these compounds have bioactive properties such as antioxidant, cardio-vascular protective, neuro-protective, antiviral, anticancer, antidiabetic and anti-inflammatory activity (Coppin et al. 2013; Hamed et al. 2019; Lizarraga-Velázquez et al. 2020; Liu et al. 2022). However, their pour water solubility results in poor stability and bioavailability, limiting their use as nutraceutical and pharmaceutical for oral administration (Cao et al. 2020; Huang et al. 2017; Núñez-López et al. 2019).

It has been found that glycosylation of phenolic compounds increases water solubility, stability, and bioavailability as sugars are highly hydrophilic molecules capable to form hydrogen bonds between hydroxyl groups of sugars and water, which can be performed by chemical and enzymatic methods (Herrera-González et al. 2017; Hofer 2016; Xu et al. 2016). The former involves different steps of protection and deprotection of the acceptor for the desired glycosidic bonds coupled with the generation of toxic residues, the enzymatic synthesis is a friendly environmental method and normally has a high regioselectivity (De Winter et al. 2015).

Different Glycoside-hydrolase (GH) enzymes have been used for the glycosylation of phenolic compounds, some of them use activated sugars such as UDP-glucose (Leloir enzymes), where the breakdown of the phosphodiester linkage gives the energy for the formation of a new glycosidic bond with the acceptor, the principal drawback of these enzymes is the use of an expensive donor substrate such as UDP-glucose (Ohgami et al. 2014; Xu et al. 2016), on the contrary the glycoside hydrolase enzymes with glycosyltransferase activity, has the advantage of the use of natural, abundant and cheap substrates such as sucrose or lactose, these enzymes use the energy of the breakdown of the glycosidic bonds from the donor to form new glycosidic linkage with different acceptors (Xu et al. 2016).

In the case of sucrose and starch utilizing enzymes such as cyclodextrin glycosyltransferase, maltogenic amylase or glucansucrase, they exhibit the capacity to transfer a molecule of glucose to phenolic compounds as acceptors (Malbert et al. 2014). The glucosides obtained with these enzymes exhibit interesting properties, such as increased bioavailability (Yamada et al. 2006), increased stability (Woo et al. 2012), increased sensory quality of phenolics by reduction of astringency (Kim et al 2016) and depending of glucosylation position it can increase (Woo et al. 2012; Choung et al. 2017) or reduce (Kim et al. 2016) antioxidant activity, including new biological properties (Woo et al. 2012; Nishikawa et al. 2019; Huang et al. 2020; Tian et al. 2023). Other enzymes such as invertases and levansucrases from the GH 32 and GH 68 families respectively have successfully transferred a fructose to phenolic compounds (Herrera-González et al. 2017, 2021; Núñez-López et al. 2019).

Regarding enzymatic galactosylation of phenolic compounds, it is interesting as galactose in glycoconjugates is involved in different biological recognition processes (Lu et al. 2015a; Darsandhari et al. 2018) and improves solubility, stability and bioavailability, and some phenolic galactosides exhibited better biological activities versus their phenolic glucoside counterparts (Lu et al. 2015a; Darsandhari et al. 2018; Xing et al. 2024), including modulation of the intestinal mycrobiota such as prebiotics (Sun et al. 2025). Phenolic galactosylation is favored by catalytic promiscuity of galactosyltransferase enzymes (Mohideen and Kawan 2023; He et al. 2021; Xing et al. 2024; Potocká et al. 2015). The synthesis of phenolic galactosides has been successfully obtained by plant UDP-galactose using glycosyltransferases (Mohideen and Kawan 2023; He et al. 2021; Wang et al. 2019; Li et al. 2025); however, UDP-galactose is an expensive donor compared with lactose as galactosyl donor, which is an abundant and cheaper substrate used by several microbial enzymes (Lima de Albuquerque et al. 2021; Porciuncula-González et al. 2025; Lu et al. 2015a, b).

The enzymatic glycosylation of phenolic compounds using lactose as a donor has been performed with different microbial β-galactosidases, Carevic et al. (2015) synthesized salicin galactoside by a β-galactosidase from Aspergillus oryzae, Xu et al. (2020) synthesized myricetin 3–0-galactoside by whole cells of a recombinant E. coli, Darsandhari et al. (2018) synthesized galactosides from kaempferol, emodin, rutin, curcumin, mangostin, biochanin, formononetin and resveratrol with a β(1,4) galactosyltransferase from Neisseria meningitidis, Lu et al. (2015a, b) synthesized 4`-O-β-caffeic acid galactoside with an engineered β-galactosidase from Lactobacillus bulgaricus, with the same enzyme they synthesized galactosides with phenol, hydroquinone, catechol, and pyrogallol as acceptors (Lu et al. 2015a), Xing et al. (2024) produced tyrosol-galactoside with a β-galactosidase from Vulcanisaeta thermophila, and recently, Kossuthova et al. (2025), carried out the synthesis of tyrosol-galactoside in order to produce and analogue of salidroside (β-tyrosol glucoside), a compound with neuroprotective properties.

Therefore, the enzymatic galactosylation of phenolic compounds is a promising area for the development of new bioactive enzymes for different applications. The Pantoea anthophila permeabilized cells (PAPC) showed high transglycosylation activity for galactooligosaccharides (GOS) production with a high conversion (Yañez-Ñeco et al. 2017), as it contains the β-galactosidase from GH42 family, which transfer galactose from lactose to synthesize GOS with β (1 → 3) and β (1 → 6) galactosyl linkages (Yañez-Ñeco et al. 2021); this is the first time that permeabilized cells of P. anthophila is evaluated for the galactosylation of six phenolic compounds with different molecular structure, one phenolic ring (gallic acid and caffeic acid), two phenolic rings (catechin) and glycosylated compounds with two phenolic rings (phlorizin, puerarin and mangiferin) for the discovery of new bioactive molecules.

Materials and methods

Bacterial strains and materials

In this work, two P. anthophila strains, isolated from tejuino, BI-55.2 and BI-69.1, respectively, are belonging to the CIATEJ microorganisms’ collection. Lactose and sodium chloride were supplied by golden bell (California, USA). Dimethyl sulfoxide (DMSO), potassium and sodium phosphates, O-nitrophenol, o-nitrophenyl-β-D-galactopyranoside (oNPG), formic acid, caffeic acid, phlorizin, puerarin, gallic acid, catechin, mangiferin, were purchased from Sigma-Aldrich (Missouri, USA), all phenolic compounds and reactive were high purity grade (≥ 98%). Yeast extract, peptone from casein and bacteriological agar were acquired from Bioxon (Toluca, Mexico). And acetonitrile (ACN) HPLC grade was supplied by J.T. Baker (Madrid, España). And TLC silica gel 60 F254s plates were from Merck KGaA, (Darmstadt, Germany), Flash pure eco flex C18 column (20 g, 50 µm) from Büchi (Uster, Switzerland).

Permeabilized cells production

A single colony of each P. anthophila strain was taken and resuspended in inducer culture medium (20 gL⁻¹ lactose, 10 gL⁻¹ bactopeptone, 10 gL⁻¹ yeast extract, 5 gL⁻¹ NaCl), the culture was incubated at 30 °C, 250 rpm and stopped at the exponential growth phase when OD (600 nm) was higher than 1.0, the cells were harvested by centrifugation (6000 rpm, 10 min, 4 °C) and washed with phosphate buffer (50 mM, pH 6.5). For permeabilization, cells were resuspended in ethanol 50% (v/v) solution and stirred for 15 min, then washed with phosphate buffer (Rodriguez-Colinas et al. 2011). This cells solution was used as biocatalyst for the galactosylation reactions.

Phenolic galactosylation screening by P. anthophila permeabilized cells

Β-galactosidase activity of permeabilized cells was determined with o-nitrophenyl-β-D-galactopyranoside (oNPG) as substrate at 30 °C, 20 μL of permeabilized cells was added to 480 μL of 22 mM oNPG in sodium phosphate buffer (50 mM, pH 6.5), after 10 min, the reaction was stopped by addition of 750 μL of Na₂CO₃ (0.4 M). Then, the release of O-nitrophenol was measured by absorbance at 420 nm (Yañez-Ñeco et al. 2017). The enzymatic galactosylation of phenolic compounds was carried out with a 400 gL⁻¹ lactose solution (50 mM sodium phosphate buffer, pH 7.0), followed by the addition of each phenolic compound to a final concentration of 10 mM (caffeic acid, gallic acid, catechin, phlorizin, mangiferin and puerarin). The reaction was conducted at 50 °C for 24 h and then quenched in a boiling water bath. All samples were analyzed by HPLC–UV, and the conversion rate was determined based on the difference between the initial and final acceptor concentrations, following the methodology described by Herrera-González et al. (2021).

$$\%conversion=({[Acceptor]}_{\mathit{initial}}-{[Acceptor]}_{\mathit{final}})/{[Acceptor]}_{\mathit{initial}}\times 100$$

HPLC–UV analysis

Samples were analyzed by an Agilent Technologies_HPLC, (Infinitely Better 1260), coupled to a UV detector, a Phenomenex Luna 5u C18 (2) 100 Å (3.90 mm × 150 mm) column was used. The mobile phase was applied at 0.8 mL/min and 40 °°C. Two solvents were used, (A) composed of water-formic acid (0.05% v/v) and (B) composed of acetonitrile-formic acid 0.05% v/v. A gradient was applied as follows: 90% A-10% B (min 0), 50% A – 50% B (1–20 min), 5% A – 95% B (20–25 min), 90% A – 10% B (25–30 min), phenolic compounds were measured at 280 nm (Núñez-López et al. 2019).

UPLC-ESI-QTof-MS analysis

UPLC-QTof analysis was performed using a Waters Acquity UPLC H-Class system (Milford, MA, USA). The separation was carried out in a Waters Acquity UPLC BEH C18 (2.1 × 100 mm, 1.7 μm) column at 40 °C, using the solvents a gradient of solvent A (water-formic acid 0.05% v/v) and solvent B (acetonitrile-formic acid 0.05% v/v), the gradient was as 0–1 min (95% A), 1–11 min (95–30% A), 11–12 min (30–95% A), 12–15 min (95% A), the flow was 0.3 mL/min and the injection volume was 10 μL). Mass spectrometry was performed in negative mode in a Waters Xwvo G2-XS QTof spectrometer with electrospray ionization (ESI), the parameters were: cone voltage of 40 kV, a capillarity voltage of 2.5 kV, a font temperature of 100 °C, desolvation temperature of 250 °C and collision energy of 6 eV. The mass range was from 50 to 1000 m/z.

Purification of puerarin galactoside

Purification of the major puerarin galactoside was performed in a Büchi flash chromatograph (Uster, Switzerland), equipped with a Pump Manager C-615, two Pump Module C-605 and a Fraction Collector C-660, the separation was carried out in a Flash pure eco flex C18 column (20 g, 50 µm, Uster, Switzerland) at room temperature. Water (A) and acetonitrile (B) were used solvents for the mobile phase at a flux of 3 mL/min. 10 mL of sample was injected, then 100% A (15 min) was applied followed by a gradient from 100 to 80% A and 0 to 20% B (30 min), followed by a second gradient from 80 to 70% A and 20 to 30% B (30 min). Fractions were monitored by TLC silica gel 60 F254s plates with a mix of acetonitrile and water (9:1 v/v) as mobile phase. The separation of products was analyzed by UV lamp with a wavelength of 280 and 354 nm.

NMR analysis

For Nuclear Magnetic Resonance (NMR) analysis, samples were dissolved in deuterated dimethylsulfoxide (DMSO-d₆). 1D and 2D NMR spectra of purified puerarin galactoside and puerarin were recorded at 25 °C on a Jeol ECA 600 spectrometer (Tokio, Japan) at 600 MHz for ¹H and 150 MHz for ¹³C. The data were processed using Delta 6.3.0 software. Chemical shifts were given in ppm. All the samples were dissolved in DMSO-d₆ and the residual solvent peak was used as reference (¹H δ: 2.49 ppm and ¹³C δ: 39.5 ppm). Assignments were supported by homonuclear and heteronuclear correlated 1D and 2D techniques, including: distortionless enhancement by polarization transfer (DEPT 90), correlation spectroscopy (COSY), heteronuclear multiple quantum coherence (HMQC) and heteronuclear multiple bond correlation (HMBC) experiments.

Results

Screening of enzymatic galactosylation of phenolic compounds

Table 1 shows the results of the enzymatic galactosylation of phenolic compounds. The Panthoea anthophila permeabilized cells (PAPC) BI-69.1 showed a higher conversion than PAPC BI-55.1 for all the compounds (Table 1), regarding the acceptor, the highest conversion was observed on puerarin (53.3 and 36.5%), followed by phlorizin (41.8 and 33.5%), caffeic acid (32.5 and 21.4%), gallic acid (18.3 and 17.4%) and catechin (1.37%) for BI-69.1 and BI-55.2 respectively, while no reaction was observed on mangiferin with both biocatalysts (Table 1).

Table 1.

Screening of the enzymatic galactosylation on different phenolic compounds by Pantoea anthophila permeabilized cells (PAPC) BI-55.2 and BI-69.1

Galactoside synthesis profile on different phenolic compounds

Regarding the galactoside synthesis profile from the different phenolic compounds, a HPLC analysis was carried out at 0 and 24 h of reaction for both PAPC. In general, several galactosides were observed for most of the phenolic compounds with both biocatalysts (Figs. S1–S5). The two highest galactosylation conversions were observed on puerarin and phlorizin by PAPC 69.1 (Fig. 1). On puerarin five galactosides were observed with puerarin-galacoside 4 (G-Pu4) as the major product (Fig. 1a), for phlorizin seven galactosides were observed, interestingly phlorizin-galactoside 7 (G-Ph7) was increased until 8 h of reaction, then it was decreased (24 h), on the contrary phlorizin-galactoside 4 (G-Ph 4) was increased all the time during reaction, thus it could be possible that G-Ph7 is the precursor of G-Ph4, probably with galactosyl elongation (Fig. 1b). In the case of caffeic acid, five galactoside products were observed with different retention times; however, there were some differences in the profile of products between the two biocatalysts, for the products observed at 6.5 and 8.5 min, the production was higher for PAPC BI-69.1 than BI-55.2, while for the galactoside observed at 4.5 min, the production was higher for PAPC BI-55.2 than BI-69.1 (Fig. S3). For gallic acid and catechin, four and five different galactosides were observed (Figs. S4 and S5, respectively).

Fig. 1.

HPLC–UV analysis of time-course of enzymatic galactosylation by permeabilized cells of the Pantoea anthophila BI-69.1 strain on puerarin (a) and phlorizin (b)

Mass spectrometry by UPLC-ESI-Qtof-MS

A mass spectrometry analysis by UPLC-ESI-QTof was carried out to determine the degree of galactosylation by PAPC BI-69.1 with puerarin and phlorizin as acceptors (Figs. 2 and 3, respectively). As it can be seen, a high catalytic promiscuity is observed, thus probably the galactosylation occurs at different hydroxyl groups of these phenolic compounds, resulting in three monogalactosides for puerarin (Pu*) with retention times of 4.47, 4.7, and 5.09 min (Fig. 2) and a m/z = 577.12 (Fig. 2b). Three monogalactosides for phlorizin (Ph*) with retention times of 6.12, 6.36, and 6.79 min (Fig. 3a) and a m/z = 597.13 (Fig. 3b). The biocatalyst has the capacity to elongate the galactosyl chain from puerarin producing a digalactoside with a retention time of 3.81 min (Fig. 2a) and a m/z = 739.08 (Fig. 2b), as well as from phlorizin resulting in two different digalactosides with retention times of 5.8 and 6.63 min and a m/z = 759.1 (Fig. 3b).

Fig. 2.

a UPLC-ESI-Qtof-MS chromatogram of puerarin galactosylation, b mass spectra in negative ion mode of puerarin (Pu), mono-galactoside (Pu*) and di-galactoside (Pu**)

Fig. 3.

a UPLC-ESI-QTof-MS chromatogram of phlorizin galactosylation, b mass spectra in negative ion mode of phlorizin (Ph), mono-galactoside (Ph*) and di-galactoside (Ph**)

Molecular structure elucidation by NMR of puerarin monogalactoside

The major galactoside produced by PAPC 69.1 with puerarin as acceptor (G-Pu4, Fig. 1a) was purified by Flash chromatography and the molecular structure elucidated by NMR spectroscopy. The ¹³C spectra (Fig S8) shows the presence of 12 carbons corresponding to the glucose linked to puerarin (Red text, Table 2) and the galactose linked to glucose (Blue text, Table 2), comprising the range from 60 to 105 ppm. One of them (103.8 ppm) corresponds to the 1´´´ carbon of galactose, which in a standard spectrum of galactose appears at 92.6 ppm. Thus, it is probable that galactose was linked to glucose in that position. On the other hand, the 6´´ carbon of puerarin is located at 61.4 ppm and in the glycoconjugate it´s located at 69.2 (Table 2). This change in the signal shift indicates that the 6´´ carbon of puerarin forms the glycosidic bond.

Table 2.

¹³C NMR data of puerarin and puerarin galactoside. Chemical shifts in ppm

^aRed text, 12 carbons corresponding to the glucose linked to puerarin

^bBlue text, the galactose linked to glucose

HMQC analysis (Fig. S11) of the glycoconjugate indicates that the 1‴ carbon (103.8 ppm) of galactose corresponds to a proton at 4.07 ppm, the same signal was observed in the ¹H NMR (Fig. S10). Additionally, the 6″ carbon of puerarin (69.2) corresponds to a proton also at 4.07 ppm (Fig. S6). HMBC shows proton interactions at 4.07 ppm with carbon 6″ (Fig. S7). Since this proton signal belongs to both carbon 6″ and carbon 1‴ and in the HMBC experiment the interaction of a carbon with its own protons cannot be observed, it is concluded that carbons 6″ and 1‴ interact through an glycosyl-1‴−6″-puerarin bond which confirms what was reported in the analysis of the ¹³C displacements shown in Table 2 corresponding to β-D-(1 → 6) puerarin monogalactoside (Fig. 4).

Fig. 4.

Structure of β-D-(1 → 6) puerarin monogalactoside elucidated by NMR. The carbons that form the glycosidic bond between galactose and puerarin are highlighted in red

Discussion

This is the first time that β-galactosidases of Pantoea anthophila permeabilized cells (PAPC) were used in galactosylation reactions with caffeic acid, gallic acid, catechin, phlorizin, puerarin and mangiferin as acceptors. As expected, conversion varied in function of the molecular structure of the acceptor (Table 1), in the case of molecules with one phenolic ring like gallic acid, conversions were 17.4 and 18.3% for PAPC of BI 55.2 and BI 69.1 respectively (Table 1). For caffeic acid, an increase in conversions was observed with 21.4 and 32.5% for PAPC of BI 55.2 and BI 69.1, respectively (Table 1). Lu et al. (2015a, b) reported a conversion of 7% with an engineered β-galactosidase from Lactobacillus bulgaricus, the same authors (Lu et al. 2015a) galactosylated with the same enzyme catechol, phenol, pyrogallol and hydroquinone with conversions of 7.6, 13.8, 32.3 and 53.1%, respectively. Potocka et al. (2015) synthesized tyrosol galactoside (10.9%) with a β-galactosidase from Aspergillus orizae, Qi et al. (2017) reported the galactosylation of tyrosol (36.3%) by a recombinant β-galactosidase from Enterobacter cloacae B5, Xing et al. (2024) galactosylated tyrosol by β-galactosidases from Vulcanisaeta thermophila, wild type (19.2%) and mutans, respectively (28.9–34.2%), and recently tyrosol-galactoside synthesis reported by Kossutova et al. (2025) with a β-galactosidase from Aspergillus orizae with a 11% of conversion. By the other hand Carevic et al. (2015) carried out the galactosylation of salicin, a molecule with a glucosylated phenolic ring has been carried out by a β-galactosidase from A. orizae with a conversion of 28%. Therefore, PAPC compete with other β-galactosidades for galactosylation of one phenolic ring molecules such as gallic acid and caffeic acid. Regarding galactosylation of catechin, a molecule with two phenolic rings connected by a heterocyclic ring, the conversion was very low 1.37 for both BI 55.2 and BI 69.1 PAPC (Table 1). On the contrary, higher conversions were observed on phlorizin (33.5 and 41.8%) and puerarin (36.5 and 53.3%) for BI 55.2 and BI 69.1 PAPC respectively (Table 1), glucosylated compounds with two phenolic rings. Other flavonoids such as myricetin, quercetin and kaempferol, molecules with two phenolic rings with a similar structure but different number of hydroxyl groups in the phenolic ring B, were successfully galactosylated in the position 3 of the heterocyclic ring, producing 3-O-galactosides by a flavonol 3-O-galactosyltransferase from a plant cloned in E. coli, requiring activated sugars such as UDP-galactose as a donor for galactosylation (Xu et al. 2020). Mohideen and Kawan (2023) reported also the galactosylation of quercetin, myricetin and kaempferol with mutants of a UDP-glucose: flavonoid 3-O-glycosyltransferase (VvGT1) from red grape, which was cloned in E. coli, these mutants were able to use UDP-galactose as a donor. Myricetin was also galactosylated by a mutant of plant UDP-galactose anthocyanidin 3-O-galactosyltransferase (UA3GalT) cloned in E. coli using UDP-galactose as donor (He et al. 2021). Floretin, the aglycon of phlorizin, was also galactosylated by a plant C-glycosyltransferase from Fortunella crassifolia using UDP-galactose as donor (Li et al. 2025). The galactosylation conversion was not reported on these four recombinant plant UDP-glycosyltransferases. A microbial β(1,4)-galactosyltransferase from Neisseria meningitidis, which used also UDP-galactose as donor, showed a highly promiscuous catalysis for the galactosylation of different phenolic glycosides such as kaempferol 3-O-β-D-glucoside, biochanin A 7-O-β-D-glucoside, formononetin 7-O-β-D-glucoside, resveratrol 3-O-β-D-glucoside, curcumin 4′-O-β-D-glucoside, emodin 3-O-β-D-glucoside, aloe emodin 3-O-β-D-glucoside, rutin, diosmin, and α-mangostein 3-O-β−2-deoxy-D-glucoside, in all these compounds, galactose was linked by a β-galactosyl linkage to the glucose producing the respective phenolic lactosides (Darsandhari et al. 2018). The enzymatic synthesis of phenolic galactosides has been carried out by microbial β-galactosidases with lactose as a donor and galactosyltransferases that require UDP-galactose as a donor, the former has the advantage to use an abundant and cheaper substrate than UDP-galactose. Therefore, PAPC has potential for the synthesis of new galactosides from phenolics with different molecular structures. The highest galactosylation of PAPC was observed in glucosylated phenolics such as puerarin and phlorizin. It has been found for other enzymatic glycosylations, that the presence of glucose in phenolic compounds allowed the enzymatic glucosylation with GH 13 and GH 70 enzymes (Tian et al. 2023), in the same way the presence of glucose in different phenolics favors galactosylation by the β-(1,4) galactosyltransferase from N. meningitidis (Darsandhari et al. 2018), and higher conversions for the enzymatic fructosylation of puerarin and phlorizin (≤ 80%) were also observed by a levansucrase from Gluconacetobacter diazatrophicus (Núñez-López et al. 2019; Herrera-Gonzalez et al. 2021). It could be possible that the presence of glucose in phlorizin and puerarin facilitates enzymatic galactosylation with a high conversion by PAPC, as it is closer to lactose, its natural acceptor. It could be interesting in the future to perform a docking analysis to predict which hydroxyls could be galactosylated by β-galactosidases from P. anthophila on the different phenolics evaluated Mangiferin was not galactosylated by PAPC (Table 1), in another previous research (Núñez-López et al. 2019), it was difficult to enzymatically fructosylate this molecule because of its very low solubility in aqueous conditions, additionally mangiferin is a rigid and big molecule composed of two aromatic rings connected by heterocyclic ring linked to glucose. Emodin, a molecule of two phenolic rings connected by a heterocyclic ring, was hardly galactosylated by mutants of a plant UDP-glycosyltransferase (Mohideen and Kawan 2023). It could be possible that steric hindrance difficulted the galactosylation of these molecules. According to the profile of galactosides observed by HPLC, PAPC carried out galactosylation in different positions, the same behavior has been observed for the glucosylation of flavonoids by some GH70 enzymes (Tian et al. 2023) but it is not common in the galactosylation of phenolic compounds by β-galactosidases. In other reports, the enzymatic galactosylation of phenolic compounds produce only one monogalactoside as the principal product, for example Xu et al. (2020) reported only one galactoside as a product from myricetin, kaempferol and quercetin as acceptors, by whole cell biocatalyst with an engineered E. coli strain with an UDP-galactose depending galactosyl-transferase. Carevic et al. (2015) reported also only one monogalactoside from salicin as acceptor by a β-galactosidase from A. orizae, Potocka et al. (2015) reported only one galactoside from tyrosol with the same A. orizae β-galactosidase, and Xing et al. (2024) reported one galactoside from tyrosol with a β-galactosidase from V. thermiphila, while Lu et al. (2015a) reported one galactoside from phenol, hydroquinone, and catechol and two galactosides from pyrogallol, and the same research group reported the synthesis of two caffeic acid galactosides with a β-galactosidase from L. bulgaricus, however both of them are monogalactosides (Lu et al. 2015a, b; Lu et al 2015a). Therefore, the β-galactosidases present PAPC showed a high catalytic promiscuity resulting in a high diversity of products with aglycones and phenolic glycosides used as acceptors, some of them possibly elongated by polymerization. To investigate the degree of galactosylation of the products obtained from puerarin and phlorizin as acceptors with PAPC 69.1, a mass spectrometry analysis was performed by UPLC-ESI-QTof-MS. As the number of monogalactosides are lower than digalactosides for phlorizin and puerarin as acceptors respectively, probably some of these monogalactosides are starters for elongation to obtain the difference of the digalactosides. The presence of them proves that a polymerization reaction was carried out by transgalactosylation. Therefore, the β-galactosidases from PAPC are less specific than other β-galactosidases reported (Carevic et al. 2015; Potocka et al. 2015; Lu et al. 2015a, b; Lu et al. 2015a; Xu et al. 2020, Xing et al. 2024). This is the first time that polymerization or galactosyl chain elongation from phenolic galactosides as acceptors is reported. The characterization of the molecular structure of one major galactoside produced by PAPC 69.1 was determined by NMR spectroscopy. According with the signals of Table 2 and Fig. 4, the major galactoside produced from puerarin corresponded to β-D-(1 → 6) puerarin monogalactoside. Furthermore, the ¹³C spectra showed the presence of 12 carbon signals correlated with a sugar region. One of them with 103.8 ppm corresponds to the 1‴ carbon of galactose, normally in a standard spectrum of galactose, it appears at 92.6 ppm. Thus, galactose was linked to glucose in that position. Lu et al. (2015a, b) found a similar movement of the shift in the carbon signal of galactose (101.1 ppm). On the other hand, the 6″ carbon of puerarin is located at 61.4 ppm and in the glycoconjugate it’s located at 69.2 (Table 2). This change in the signal shift indicates that the 6″ carbon of puerarin forms the glycosidic bond. Previously, Li et al. (2004) performed the synthesis of puerarin glycoconjugates forming an α-D-gluycosyl-1–6-puerarin bond in which the 6″ carbon shift of puerarin moves from 61.4 ppm to 66.2 ppm which is similar to the present work. When galactosylation was performed with a β-(1,4) galactosyltransferase from N. meningitidis on different phenolic glucosides galactose was linked to the molecule of galactose, Carevic et al. (2015) also reported a link of galactose to the molecule of glucose of salicin, and in other phenolic glycosylation reports, puerarin has been fructosylated in the same position by a levansucrase from Gluconacetobacter diazatrophicus (Nuñez-López et al. 2019) as well as phlorizin (Herrera-González et al. 2021), in both cases the fructose was linked to the carbon 6 of the glucose linked to the aglycone of puerarin and phlorizin, as the hydroxyl group of the carbon 6 of glucose is the most reactive in transglycosylation reactions with glucose as acceptor (Núñez-López et al. 2019; Herrera-Gonzalez et al., 2021). Therefore, probably the presence of glucose in puerarin favors glycosylation for levansucrases and β-galactosidases, as it is closer to their natural acceptors sucrose and lactose, respectively, both with a glucosyl group, this could also explain the highest conversion percentage from the phenolic compounds evaluated in the present work. It could be interesting in the future to perform a docking analysis to predict which hydroxyls could be galactosylated by β-galactosidases from P. anthophila on the different phenolics evaluated.

Conclusion

This is the first time that permeabilized cells of Pantoea anthophila from two different strains were evaluated for the galactosylation of different phenolic compounds (gallic acid, caffeic acid, catechin, puerarin, phlorizin, and mangiferin). It was observed that the highest galactosylation conversions were favored by the presence of glucose naturally linked to the aglycone of the phenolic compounds (puerarin and phlorizin) with 53.3 and 41.8% of conversion, respectively, with the exception of mangiferin, in fact the NRM spectroscopic analysis of the major galactoside purified revealed that galactose was linked by a β (1 → 6) linkage to the glucose of puerarin. The permeabilized cells of P. anthophila as biocatalysts also exhibited transgalactosylation capacity from monogalactosides as starters, which is reported for the first time in enzymatic galactosylation reactions with phenolic compounds as acceptors.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

JRB conducted experiments, JGTT analyzed data, MGA and LAD provided financial support, RLR provided reactive and material, AHG methodology, conceived and designed research, editing, JA conceived and designed research, supervision, funding acquisition, writing.

Funding

This research was supported by Consejo Estatal de Ciencia y Tecnología de Jalisco (FODECIJAL 10233–2022 project).

Data availability

The data of this research can be supplied by a suitable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

All authors have consented for publication.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.