Phlorizin

Enzymatic synthesis of phlorizin fructosides

Azucena Herrera-Gonza´lez a, b, Gema Nún˜ez-Lo´pez a, b, Nelson Nún˜ez-Dallos c,
Lorena Amaya-Delgado a, Georgina Sandoval a, Magali Remaud-Simeon b, Sandrine Morel b,
Javier Arrizon a, L´azaro Herna´ndez d,*
a Centro de Investigacio´n y Asistencia en Tecnología y Disen˜o del Estado de Jalisco, A.C.- Unidad Zapopan. Camino Arenero 1227, El Bajio del Arenal, 4519, Zapopan,
Jalisco, Mexico
b TBI, Universit´e de Toulouse, CNRS, INRAE, INSA, 135 Avenue de Rangueil, 31077, Toulouse, France
c Departamento de Química, Universidad de los Andes, Carrera 1 No. 18A-12, 111711, Bogota´, Colombia
d Centro de Ingeniería Gen´etica y Biotecnología, Ave. 31 e/ 158 y 190, Apartado, 6162, La Habana, Cuba

Abstract

Phlorizin is a low soluble dihydrochalcone with relevant pharmacological properties. In this study, enzymatic fructosylation was approached to enhance the water solubility of phlorizin, and consequently its bioavailability. Three enzymes were assayed for phlorizin fructosylation in aqueous reactions using sucrose as fructosyl donor. Levansucrase (EC 2.4.1.10) from Gluconacetobacter diazotrophicus (Gd_LsdA) was 6.5–fold more efficient than invertase (EC 3.2.1.26) from Rhodotorula mucilaginosa (Rh_Inv), while sucrose:sucrose 1-fructosyltransferase (EC 2.4.1.99) from Schedonorus arundinaceus (Sa_1-SST) failed to modify the non-sugar acceptor. Gd_LsdA synthesized series of phlorizin mono- di- and tri-fructosides with maximal conversion efficiency of 73 %. The three most abundant products were identified by ESI-MS and NMR analysis as β-D-fructofuranosyl-(2→6)-phlorizin (P1a), phlorizin-4’-O-β-D-fructofuranosyl-(2→6)-D-fructofuranoside (P2c) and phlorizin-4-O-monofructofuranoside (P1b), respectively. Purified P1a was 16 times (30.57 g L—1 at 25 ◦C) more soluble in water than natural phlorizin (1.93 g L—1 at 25 ◦C) and exhibited 44.56 % free radical scavenging activity. Gd_LsdA is an attractive candidate enzyme for the scaled synthesis of phlorizin fructosides in the absence of co-solvent.

1. Introduction

The dihydrochalcone phlorizin (phloretin-2’-β-D-glucopyranoside) is an intermediate of the flavonoid biosynthetic pathway in plants. This compound is the most abundant phenolic-glucoside in apple trees, being found in root bark, shoots, leaves, and fruit peel [1,2]. Phlorizin has interesting health-promoting properties, such as anti-diabetic, anti-py- retic, anti-cancer, anti-inflammatory, anti-microbial and anti-oXidant activities [1–4]. Phlorizin is poorly soluble in water, consequently its bioavailability and assimilation are rather low [5–7]. Tailored glyco- sylation is an attractive method to enhance the bioavailability of poly- phenolic compounds like flavonoids by increasing their aqueous solubility and protection from oXidation [6–10]. Glycosylation of poly- phenolic compounds can be performed via chemical or enzymatic syn- thesis [9,11]. The chemical process requires glycosyl activation and multiple steps of protection/deprotection of hydroXyl groups to control the regio-selectivity; it is hazardous and generates toXic wastes [8,9].

More attractively, the enzymatic transfer involves a one-step reaction under mild conditions, with the additional advantage of being regio- and stereo-selective [5,8,9]. Glycosyltransferases of Leloir and non-Leloir types have been suc- cessfully used for the glycosylation of dihydrochalcones [5,12]. Non-Leloir glycosyltransferases are of particular interest for biotechno- logical applications, as their donor substrate is the naturally abundant sucrose instead of expensive nucleotide-activated sugars. Pandey et al. [5] reported the in-vitro glycosylation of phloretin (phlorizin aglycone) with conversion efficiency of 95 % by using an UDP-glycosyltransferase from Bacillus licheniformis. Five different products were identified as phloretin 4’-O-glucoside, phloretin-2’-O-glucoside, phloretin 4’, 4-O-diglucoside, phloretin 4,6’-O-diglucoside, and phloretin 2’,4’, 4-O-triglucoside. Overwin et al. [12] reported the transfer of the glu- cosyl moiety of sucrose to phloretin by Escherichia coli cells expressing amylosucrase from Neisseria polysaccharea ATCC 43768. The biotransformation of phloretin by the recombinant non-Leloir glucosyltransferase resulted in the synthesis of phloretin-4’-O-glucoside, α-D-diglucosyl-(1→4)-phloretin and α-D-triglucosyl-(1→4)-phloretin at yield rates of 35 %, 32 % and 28 %, respectively [12]. To date, the enzymatic fructosylation of dihydrochalcones such as phlorizin remains unexplored.

Microbial and plant enzymes from the glycoside hydrolase (GH) families 32 and 68 (clan GH-J) initiate de novo fructan synthesis by transferring the fructosyl moiety from one sucrose molecule (acting as donor) to another sucrose molecule (acceptor). The fructosyl moiety can be also released to water resulting in substrate hydrolysis. The ratio of transferase versus hydrolase (T/H) activities, the regio-selectivity, and the polymerization capacity are intrinsic properties that vary among clan GH-J enzymes of different origins [13–15].
Sucrose hydrolysis is the predominant reaction of yeast invertases (GH32, EC 3.2.1.26), some of which also synthesize low levels of fruc- tooligosacharides (FOSs). Plant sucrose:sucrose 1-fructosyltransferases (GH32, EC 2.4.1.99) are distinguished for the preferential synthesis of 1-kestotriose with negligible fructose release. Bacterial levansucrases (GH68, EC 2.4.1.10) often yield a complex miXture of FOSs, including 1- kestotriose, 6-kestotriose and 6G-kestotriose, in addition to free fructose and the β-(2→6) linked polysaccharide levan. Levansucrases from diverse sources have shown the ability to fructosylate non-conventional acceptors such as other sugars, aromatic and aliphatic alcohols, and phenolic compounds [16–20].

In this study, three clan GH-J enzymes were assayed for the ability to fructosylate phlorizin in the absence of co-solvent using sucrose as donor substrate. Levansucrase from the Gram-negative bacterium Gluconace- tobacter diazotrophicus (Gd_LsdA) synthesized short-chain series of phlorizin fructosides including three positional mono-fructoside isomers. The water solubility and the anti-oXidant activity of the accumu- lated product β-D-fructofuranosyl-(2→6)-phlorizin were assessed and compared to those of phlorizin. To our knowledge, this is the first report of the enzymatic fructosylation of phlorizin.

2. Experimental methods
2.1. Chemical materials

Phlorizin was supplied by Carbosynth Limited (Compton, UK).Sugas carrier N2 and scan range m/z 100—1500. The data acquisition and processing were performed with the Chromeleon™ 7.2 Chromatography data system software.

2.3. Enzyme activity assays

The activity of the three enzymes in study was determined from the initial-rate release of reducing sugars (fructose and glucose) from 1 % (w/v) sucrose in 50 mM sodium phosphate buffer (pH 5.8) for Gd_LsdA or in 100 mM sodium acetate buffer (pH 5.5) for Sa_1-SST and Rm_Inv. All reactions were conducted at 42 ◦C. Reducing sugars were quantified by the dinitrosalicylic acid (DNS) method [23]. One unit (U) was defined as the amount of enzyme releasing 1 μmol of glucose per min under the reaction conditions described above.

Phlorizin fructosylation was assayed by reacting 1 U mL—1 enzyme with 146 mM sucrose and 25 mM phlorizin in 50 mM sodium phosphate buffer (pH 5.8) for Gd_LsdA or in 100 mM sodium acetate buffer (pH 5.5) for Sa_1-SST and Rm_Inv. After 24-h reaction at 42 ◦C, the enzyme was inactivated by heating at 95 ◦C for 10 min, the samples were diluted with DMSO, filtered through a 0.45 μm nylon syringe filter, and analyzed by HPLC-UV-MS. The percentage of phlorizin conversion was calculated as the difference between initial and final concentrations of
phlorizin, using the following equation: % Conversion = (([phlorizininitial]– [phlorizinfinal])/[phlorizininitial]) ×100.

2.4. HPLC analysis

Fructosylation miXtures were analyzed by HPLC-UV-MS using a Thermo Scientific Ultimate 3000 series chromatograph equipped with a Phenomenex PFP C18 column (Luna 5 μm, 100 Å, 250 4.6 mm, USA), a Dionex 340 UV/VIS detector and coupled with a simple quadruple mass spectrometer (MSQ Plus, Thermo Scientific). The following gradient composed of solvent A (water-formic acid 0.05 % (v/v)) and solvent B (acetonitrile-formic acid 0.05 % (v/v)) was used at 1 mL min—1 and 40 ◦C: from 10 to 50 % B over 20 min and from 50 to 95 % B over 5
min. Phlorizin and phlorizin fructosides were quantified by UV detection at 254 nm. The calibration curve was constructed using phlorizin as analytical standard in the range of 0.05–1 mM. Mass spectrometric analysis was performed using electrospray ionization (ESI) in the negative and positive ion mode with the following parameters: voltage crose, glucose, fructose, dimethyl sulfoXide (DMSO), formic acid, cone at 50, 80 and 110 V, temperature of the ESI ion source at 450 ◦C, acetonitrile (ACN) and n-butanol were supplied by Sigma Aldrich Inc. (MO, USA). All reagents were of high purity grade (≥ 98 %). Acetonitrile and n-butanol were HPLC-MS and HPLC grade, respectively. C18- reversed phase silica gel was purchased from Sigma Aldrich Inc. (MO, USA) and TLC silica gel 60 RP-18 F254s plates were from Merck KGaA, (Darmstadt, Germany).

2.2. Enzyme production

Levansucrase from Gluconacetobacter diazotrophicus (Gd_LsdA, E.C. 2.4.1.10) was recovered from the culture supernatant of the native strain SRT4, as previously reported [21]. Sucrose:sucrose 1-fructosyltransfer- ase from Schedonorus arundinaceus (Sa_1-SST, E.C. 2.4.1.99) was puri- fied from the culture supernatant of recombinant Pichia pastoris clone PGFT6X-308, as previously described [22]. The gene encoding the pre- cursor invertase from Rhodotorula mucilaginosa MB4 (Rm_Inv, E.C. 3.2.1.26) was cloned in pGAPZB vector (Invitrogen, Carlsbad, CA, USA) and expressed constitutively in Pichia pastoris X-33. The recombinant yeast was grown in 2-L flasks containing 200 mL of YPGS medium [1 % (w/v) yeast extract, 2 % (w/v) peptone, 2 % (v/v) glycerol and 5 % (w/v) sucrose] for 3 days at 30 ◦C in an orbital shaker with constant agitation (200 rpm). Cells were removed by centrifugation (6000 rpm, 10 min), the culture supernatant was concentrated by ultrafiltration and used as a crude invertase preparation.

The quantification of sucrose remaining along the phlorizin fructo- sylation reaction was carried out using HPLC-RI on a Thermo Scientific
DIONEX Ultimate 3000 chromatograph equipped with a Biorad Carbo- hydrate Analysis column (Aminex HPX-87 K, 300 × 7.8 mm, USA) and a Shodex refractometer. The mobile phase was ultrapure water (18 MΩ cm), the flow rate was 0.65 mL min—1 and the temperature of the oven was 65 ◦C. Sucrose was used as standard and the samples were diluted in ultrapure water. Data acquisition and processing were performed with the Chromeleon™ 6.8 Chromatography data system software.

2.5. Effects of sucrose, phlorizin and enzyme concentrations on the synthesis of phlorizin fructosides by Gd_LsdA

Sucrose at initial concentrations of 0.146, 0.5, 1 and 1.5 M was miXed with phlorizin (25 mM) and reacted with Gd_LsdA (1 U mL—1). The influence of the acceptor phlorizin (25, 50, 75 and 100 mM) was assayed at fiXed sucrose concentration (1 M). Gd_LsdA at varied dosages (1, 2.5, 5, 10 and 15 U mL-1) was incubated with 1.5 M sucrose and 25 mM phlorizin. The time-course reactions were performed in 50 mM
sodium phosphate buffer (pH 5.8) at 42 ◦C with vigorous agitation (800 rpm) using an IKA® RCT basic safety control (Staufen, Germany). The reaction volume was 2 mL. Aliquots (50 μL) were taken at predefined time intervals and boiled at 95 ◦C for 10 min to stop the reaction.
Subsequently, the samples were diluted with DMSO, filtered through a 0.45 μm nylon syringe filter, and analyzed by HPLC-UV-MS and HPLC- RI. All the experiments were performed in triplicate.

2.6. Purification of phlorizin fructosides

The preparative synthesis of phlorizin fructosides was carried out using 1.5 M sucrose, 25 mM phlorizin and 1 U mL—1 Gd_LsdA in 50 mM sodium phosphate buffer (pH 5.8) at 42 ◦C with agitation (800 rpm). The reaction was monitored by HPLC-UV-MS during 50 h and then stopped by inactivating the enzyme at 95 ◦C for 10 min. Subsequently, the re- action miXture (20 mL) was diluted with ultrapure water (50 mL) and
subjected to a liquid-liquid extraction using n-butanol in volume pro- portion 1:1. The aqueous phase containing the sugars was discarded. The n-butanol phase containing phlorizin and phlorizin fructosides was concentrated with a rotary evaporator, and separated on a C18 silica gel column (30 2 cm) using as mobile phase a miXture of water/aceto- nitrile (65:35). Three different fractions of phlorizin fructosides were recovered and lyophilized. The first fraction contained 15 mg of β-D- fructofuranosyl-(2→6)-phlorizin (P1a). The second fraction contained 7 mg of phlorizin-4’-O-β-D-fructofuranosyl-(2→6)-D-fructofuranoside (P2c) and the third fraction (7.5 mg) corresponded to a miXture con- sisting of 45 % phlorizin-4-O-monofructofuranoside (P1b) and 35 % P2c, as identified later on by 1H, 13C NMR and ESI-MS analysis. The purification process was monitored by thin layer chromatog- raphy (TLC) on silica gel 60 RP-18 F254s plates using a water:acetonitrile miXture (65:35) as the mobile phase. The TLC plates were exposed to UV light (254 nm) to visualize phlorizin fructosides.

2.7. NMR spectroscopy

NMR spectroscopic characterization of phlorizin and the product β-D-fructofuranosyl-(2→6)-phlorizin (P1a) was conducted on an Advance 500 MHz spectrometer (Bruker) operating at 500 MHz for 1H and 125 MHz for 13C. The data were processed using the Topspin 3 software. 1H and 13C NMR chemical shifts are reported in parts per million (ppm) and coupling constants (J) in Hz. The samples were dissolved in methanol-d4, and the residual solvent peak was used for referencing (1H NMR δ: 3.31 ppm and 13C NMR δ: 49.0 ppm). Hetero-nuclear single quantum coherence (HSQC) and heteronuclear multiple bond correlation (HMBC) spectra were acquired using the standard Bruker software.

NMR data of phlorizin, phlorizin-4’-O-β-D-fructofuranosyl-(2→6)-D- fructofuranoside (P2c) and the miXture (45:35) of phlorizin-4-O-mono- fructofuranoside (P1b) and P2c were recorded in DMSO-d6 on a Bruker Avance 400 spectrometer (400.13 MHz for 1H; 100.61 MHz for 13C) with
the residual solvent peak used as an internal reference (1H NMR δ: 2.50 ppm and 13C NMR δ: 39.5 ppm). All measurements were performed at 298 K.

2.8. Water solubility and anti-oxidant activity of β-D-fructofuranosyl- (2→6)-phlorizin (P1a)

Phlorizin and β-D-fructofuranosyl-(2→6)-phlorizin (P1a) were solu- bilized until saturation in 300 μL of ultrapure water. The samples were incubated at 25 ◦C during 12 h at 500 rpm followed by centrifugation at 13,000 g during 10 min. The supernatant was diluted in DMSO and
analyzed by HPLC to determine the aqueous solubility.

The anti-oXidant activity of phlorizin and P1a was estimated by the 2,2-diphenyl-1-picrylhydrazyl (DPPH) assay as described by Lee et al. [24]. MiXtures of 0.15 mM DPPH (150 μL) and 6.69 mM phlorizin or P1a in 80 % (v/v) methanol (150 μL) were incubated at 37 ◦C for 30 min in
the dark, and the absorbance was measured at 517 nm using a Thermo Scientific™ Multiskan™ GO microplate spectrophotometer. The assay was carried out in triplicate. The percentage of DPPH inhibition was calculated using the following equation. % DPPH inhibition = (1-(Asample / Acontrol)) × 100 Asample is the absorbance of phlorizin or P1a and Acontrol is the absorbance of DPPH solution. TroloX™ (6-hydroXy-2,5,7,8-tetrame- thylchroman-2-carboXylic acid) was used as positive control.

3. Result and discussion
3.1. Enzyme screening for phlorizin fructosylation

Three clan GH-J enzymes differing in their origin and catalytic properties were assayed for transferring the fructosyl residue of sucrose (146 mM) to the non-conventional acceptor phlorizin (25 mM). Re- actions were deliberately performed in water to avoid any addition of inhibitory co-solvents. The synthesis of phlorizin fructosides by sucrose: sucrose 1-fructosyltransferase from the plant Schedonorus arundinaceus (Sa_1-SST, E.C. 2.4.1.99, GH32), invertase from the yeast Rhodotorula mucilaginosa (Rm_Inv E.C. 3.2.1.26, GH32) and levansucrase from the bacterium Gluconacetobacter diazotrophicus (Gd_LsdA, E.C. 2.4.1.10, GH68) was analyzed by HPLC-UV-MS (Fig. 1). In parallel, the course of sucrose transformation into mono- and oligosaccharides was monitored by HPLC-RI.

Sa_1-SST yielded short-chain β-(2→1) linked FOSs (data not shown) but failed to modify phlorizin at the end of the reaction (24 h) (Fig. 1a). Sa_1-SST is a regio-selective enzyme as it converts sucrose (Fru2-1Glc) into 1-kestotriose (Fru2-1Fru2-1Glc) very efficiently, but it does not synthesize 6-kestotriose (Fru2-6Fru2-1Glc) or 6G-kestotriose (Fru2- 6Glc1-2Fru) [22]. The enzyme inability to form glycosyl β-(2→6) link- ages may explain the failure of coupling a fructosyl unit to the hydroXyl group at position C6’’ of the only glucosyl moiety of phlorizin. It remains unknown if a sucrose-transforming fructosyltransferase of plant origin can use a non-sugar compound as an alternative fructosyl acceptor.

Rm_Inv synthesized phlorizin mono-fructoside, but the conversion efficiency was only 6 % 2.3. An almost equimolar release of fructose and glucose was observed by HPLC-RI revealing that sucrose hydrolysis was the predominant reaction of Rm_Inv at the assayed substrate con- centration (146 mM) (data not shown). We decided not to evaluate if the yield of phlorizin fructoside could be enhanced by increasing the initial sucrose concentration to above 1 M, a reaction condition that tends to improve the fructosylation/hydrolysis rate of microbial invertases [25, 26].

The HPLC-UV chromatogram of the Gd_LsdA reaction shows at least nine visible peaks of different intensities that eluted before phlorizin (retention time 14.6 min) (Fig. 1a). The electrospray ionization mass spectrometry (ESI-MS) analysis in negative ion mode revealed that
products with retention times 12.6, 11.4 and 11.04 min display the same ion peak at m/z 597.77 [C27H34O15-H]— equivalent to phlorizin mono- fructosides (Fig. 1b). The products recorded at 11.7, 10.1 and 9.8 min showed a common ion peak at m/z 759.73 [C33H44O20-H]— corre- sponding to phlorizin di-fructosides, while the less intense products (retention times 9.6, 9.3 and 9.1 min) with the ion peak at m/z = 921.68 [C39H54O25-H]— are phlorizin tri-fructosides (Fig. 1b). Gd_LsdA converted 39 % 1.5 of the initial phlorizin content under the assayed reaction conditions, and thus it was selected as the target enzyme for further optimization studies.

3.2. Time-course synthesis of phlorizin fructosides by Gd_LsdA

The simultaneous analysis of sucrose transformation and phlorizin fructosylation in time-course reactions is shown in Fig. 2. Two main phlorizin mono-fructosides (peaks P1a and P1b) were synthesized at a similar proportion during the first 6 h (Fig. 2a). Then, P1b started decreasing with a sharp decline at the interval 10 24 h coinciding with sucrose depletion (Fig. 2b), while peak P1a kept its progressive increase over time. At the end of the reaction (24 h), Gd_LsdA synthesized series of products containing one, two and three fructosyl moieties attached to the acceptor flavonoid. Mono-fructoside P1a and a di-fructoside (peak P2c) represented the first and second major phlorizin fructo-conjugates, respectively. Structural characterization of products P1a, P1b and P2c revealed that neither P1a nor P1b is the precursor of P2c (see section 3.4).

Fig. 1. Enzyme screening for phlorizin fructosylation. a) HPLC-UV-MS chromatograms of the reactions catalyzed by Gd_LsdA, Sa_1-SST, and Rm_Inv. b) ESI-MS spectra in negative ion mode of mono-fructosides (peaks P*), di-fructosides (peaks P**) and tri-fructosides (peaks P***) synthesized by Gd_LsdA.

3.3. Optimized reaction conditions for phlorizin conversion by Gd_LsdA

The influence of sucrose concentration (0.146–1.5 M) on phlorizin fructosylation was investigated at the enzyme dose of 1 U/mL using 25 mM phlorizin, pH 5.8 and 42 ◦C in time-course reactions monitored during 24 h (Fig. 3). A low sucrose concentration (146 mM) accelerated
the synthesis of phlorizin fructosides in the first time interval (0–2 h) but at the end of the reaction only 39 % of the initial phlorizin content was transformed into products (Fig. 3a) due to a rapid depletion of the fructosyl donor (Fig. 3b). The efficiency of phlorizin conversion increased to 51, 52 and 48 % in the reactions at 0.5, 1 and 1.5 M sucrose, respectively (Fig. 3c). Under these conditions, the donor substrate remained available along the entire incubation time (Fig. 3b), which avoided the hydrolysis of phlorizin mono-fructoside P1b (Fig. 3c). The fructosylation of phlorizin occurred more slowly at high sucrose con- centrations. This behavior can be explained by the fact that sucrose not only provides the fructosyl unit to be transferred onto phlorizin but it is also the main competitor in the acceptor reaction. The lower accessi- bility to water molecules in the treatment with 1.5 M sucrose favored the prevalence of transferase activity over hydrolytic reactions. On the other hand, the raise of the initial sucrose concentration from 146 mM to 1.5 M increased the solubility of phlorizin at 42 ◦C from 5.01 to 5.71 g L—1, respectively, thus favoring its accessibility to the enzyme active site. The increase of phlorizin concentration from 25 to 100 mM in reactions at fiXed 1.5 M sucrose had a negative effect on the conversion efficiency, probably influenced by a decreased solubility of the acceptor in the aqueous medium (Fig. 4a). The activity of Gd_LsdA was partially inhibited at high phlorizin concentrations, as deduced from the differences in remaining sucrose observed between the treatments after 24-h of reaction (Fig. 4b).

Phlorizin fructosylation was accelerated as the Gd_LsdA dosage was raised from 1 to 15 U mL—1 at fiXed 25 mM phlorizin and 1.5 M sucrose
(Fig. 5a). At the end of the incubation period (24 h), similar conversion rates (63–73 %) were observed in all the treatments but the time-course curves showed different behaviors. During the last time interval (8 24 h), a slight reconversion to unmodified phlorizin occurred as sucrose was depleted in the treatments containing the highest enzyme amounts (Fig. 5b). In the condition for the slowest reaction (1 U mL-1 enzyme), phlorizin conversion increased continuously over time reaching a final value of 63 %. When the reaction was stopped at 24 h, sucrose remained notably high (0.83 M). At this point, 22.5 mM fructose was linked to phlorizin forming phlorizin mono-, di- and tri-fructosides in the pro- portion 65.7, 25.9 and 8.5 %, respectively. The amount of donor sub- strate converted into phlorizin fructosides was 3.4 %. The rest of transformed sucrose (647.5 mM) was consumed via hydrolysis or in the synthesis of prebiotic fructans, a reaction which initiates with two su- crose molecules acting as the fructosyl- donor and acceptor, respectively.

A preparative Gd_LsdA reaction (20 mL) was performed under the optimized conditions (1 U mL—1 enzyme, 1.5 M sucrose, 25 mM phlorizin, pH 5.8 and 42 ◦C). At the end of a prolonged incubation (50 h), at least three series of phlorizin mono-, di- and tri-fructosides were syn- thesized; the phlorizin conversion reached the maximal value of 73 0.4 % and the remaining sucrose concentration decreased to 0.25 M (Fig. 6). The estimated percentage composition of phlorizin fructosides in the reaction miXture was P1a (35.2 %), P2c (29.8 %), P1b (7.4 %) and others (27.6 %).

In all the assayed conditions, the synthesis of phlorizin fructosides occurred as a secondary reaction. The fructosyl unit of sucrose was preferentially transferred to the natural acceptors sucrose (fructan syn- thesis), fructan (elongation reaction) or even water (sucrose hydrolysis). The addition of a co-solvent may increase the rate of phlorizin fructo- sylation by increasing the solubility of the non-sugar acceptor. Gd_LsdA is known to be active in the presence of 20 % (v/v) dimethyl sulfoXide (DMSO) or acetone [18,19]. However, from the biotechnological standpoint it is desirable conducting the enzymatic fructosylation of flavonoids in the absence of organic solvents, unless good conversion rates cannot be achieved in an aqueous medium. In the optimized aqueous reaction phlorizin conversion reached 73 %.

Fig. 2. Time-course fructosylation of phlorizin by Gd_LsdA. a) HPLC-UV-MS chromatograms of phlorizin fructo-conjugates. b) Phlorizin conversion and sucrose consumption along the reaction. The enzyme (1 U mL—1) was reacted with a miXture of sucrose (146 mM) and phlorizin (25 mM) in 50 mM sodium phosphate buffer (pH 5.8) at 42 ◦C with vigorous agitation (800 rpm). The time-course reaction was monitored during 24 h.

3.4. Structural characterization of three major phlorizin fructosides synthesized by Gd_LsdA

Three main peaks (P1a, P2c and P1b) are distinguishable in the HPLC-UV-MS chromatograms of phlorizin fructo-conjugates synthesized by Gd_LsdA along the time (Figs. 2 and 3). The ESI-MS spectra in negative ion mode revealed that products P1a and P1b are monofructoside isomers (ion peak at m/z = 597.77 [C27H34O15-H]—), while product P2c is a di-fructoside (ion peak at m/z 759.73 [C33H44O20- H]—) (Fig. 1b). The chemical structure of the three compounds was identified by ESI-MS, 1H-NMR, 13C-NMR and 2D NMR analysis.

Table 1 shows the chemical shifts from 13C and 1H NMR data of phlorizin and mono-fructoside P1a in CD3OD. The 13C spectrum of P1a
revealed 25 signals, two of them being overlapped. Ten signals were in the range 60 90 ppm and were assigned to the two mono-saccharide units (fructose and glucose). Compared with the 13C NMR spectrum of phlorizin, siX additional signals are identified in the spectrum of P1a (Fig. S2). Notably, the signal of the carbon atom at 105.72 ppm is characteristic of the C2’’’ of the fructosyl moiety. The HMBC spectrum of P1a was used to deduce the position of fructosylation (Fig. 7). The long-range correlations between C2’’’ of fructose (δC 105.72 ppm) and H6’’ (δH 3.81–4.07 ppm) of the glucosyl unit revealed that P1a corre- sponds to β-D-mono-fructofuranosyl-(2’’’→6”)-phlorizin.

Fig. 3. Effect of sucrose concentration on the synthesis of phlorizin fructosides by Gd_LsdA. a) Time-course conversion of the acceptor phlorizin. b) Time-course transformation of the donor sucrose. c) HPLC-UV-MS chromatograms of samples reacted for 24 h.

Product P2c was identified as phlorizin-4’-O-β-D-fructofuranosyl- (2→6)-D-fructofuranoside as revealed by ESI-MS and 1D and 2D NMR
analysis (Figs. S9–S16). The ESI-MS data confirmed the attachment of two fructose units to the acceptor phlorizin. The 13C NMR spectrum
revealed 30 signals (two overlapped at 165.3 ppm). Ten signals in the range 70-90 ppm were assigned to the methine carbons (CH) of two fructose units and one glucose unit, while five signals in the range 60-70 ppm were assigned to the methylene carbons (CH2) of the three mono- saccharides. Compared with the 13C NMR spectrum of phlorizin, 12 additional signals are identified in the spectrum of P2c, as expected for the proposed structure (Figs. S10 and S11). Notably, the signals of the quaternary carbons at 104.28 and 104.21 ppm are characteristic of the C2’’’ and C2’’’’ of the two fructosyl moieties, respectively. The HMBC spectrum of P2c showed long-range correlations between C2’’’’ of the second fructosyl unit (δC 104.21 ppm) and H6’’’ (δH 3.41 ppm) of the first fructosyl unit, demonstrating the interconnection between the two fructosyl moieties (Fig. S14). The HMBC spectrum of P2c shows no long- range correlations between C2’’’ of fructose and H6’’ of glucose unit (Fig. S14), suggesting that the first fructose unit is not linked to the only glucose moiety of phlorizin. This assumption is confirmed by the fact that the signal at 4.59 ppm corresponding to the 6’’-OH of the glucose unit is present in the 1H NMR spectra of phlorizin and P2c in DMSO-d6 (Fig. S16). Moreover, the signal of the phenolic hydroXyl at 10.58 ppm (bs, 1H, 4’-OH) is missing in the 1H NMR spectrum of P2c, indicating that the fructose unit is connected to the aglycone at the C4’ position. The assignation of the signals for phenolic hydroXyl protons was sup- ported by HMBC spectrum of phlorizin (Fig. S8), and the spectra data agreed with those described in the literature for phlorizin [27]. In the phlorizin structure, the signal at 10.58 ppm for phenolic hydroXyl at C4’ position is broad and the HMBC long-range correlations are not observed probably due to the phenolic proton exchange, as a result of the increased acidity of that proton in para position to a carbonyl group.

Fig. 4. Effect of phlorizin concentration on the synthesis of phlorizin fructosides by Gd_LsdA. a) Time-course conversion of the acceptor phlorizin. b) Time-course transformation of the donor sucrose.

Fig. 5. Effect of enzyme loading on the synthesis of phlorizin fructosides by Gd_LsdA. a) Time-course conversion of the acceptor phlorizin. b) Time-course trans- formation of the donor sucrose.

Fig. 6. Time-course synthesis of phlorizin fructosides under optimized reaction conditions. Gd_LsdA (1 U/mL) was reacted with a miXture of sucrose (1.5 M) and phlorizin (25 mM) in 50 mM sodium phosphate buffer (pH 5.8) at 42 ◦C
with vigorous agitation (800 rpm). Phlorizin conversion on the content of remaining sucrose were determined at time intervals during 50 h.

We failed to isolate a pure fraction of P1b. Instead, a miXture of P1b and P2c (45:35) in DMSO-d6 was analyzed by 1D and 2D NMR spec- troscopy. The in situ NMR analysis revealed that P1b corresponds to phlorizin-4-O-monofructofuranoside (Figs. S17–S20). In the 1H-NMR
spectrum of the miXture, the signal at 2.84 ppm as triplet (t, J 7.0 Hz, 2 H) corresponds to the aliphatic hydrogens (H7) of P1b (Fig. S17). The aromatic signals at 6.13 (H3’) and 5.92 (H5’) ppm in the tetra- substituted ring were observed as singlets. Furthermore, one single ar- omatic signal at 7.09 ppm that corresponds to the hydrogens in the di- substituted ring (in para position) was detected (Fig. S18). The shift of this aromatic signal to low field as well as the observed roof effect suggest that the fructose unit was linked to the aglycone of phlorizin at C4 position (Fig. S20).

3.5. The phlorizin fructosylation reaction catalyzed by Gd_LsdA

One, two and three fructosyl moieties were attached to phlorizin, a natural dihydrochalcone consisting of a glucose moiety and two phenolic aromatic rings joined by an alkyl spacer. At least three regioisomers of phlorizin mono-fructoside were formed. The primary hydroXyl group positioned at the C6’’ of the glucose unit and three phenolic hydroXyls (C4, C4’ and C6’) of the aglycone phloretin are for the steric hindrance of the adjacent carbonyl group. Moreover, there is a stable intramolecular hydrogen bond between the hydroXyl group at C6’ and the carbonyl group in the phlorizin structure (Fig. 8a) [28,29]. Gd_LsdA made no distinction for coupling the fructosyl unit to the C4 or C4’ positions of phlorizin, probably because these two aromatic hy- droXyls present similar pKa values (approX. 10.8) [16,30]. The synthesis of different mono-fructoside isomers reflects the low regio-selectivity of the phlorizin fructosylation reaction catalyzed by Gd_LsdA. On the other hand, the formation of glycosyl β-(2→6) linkages matches with the strict stereo-selectivity of the β-retaining GH families 32 and 68 [14].

The synthesis of P1a (β-D-fructofuranosyl-(2→6)-phlorizin) and P1b (phlorizin-4-O-monofructofuranoside) at a similar rate during the first time intervals of the time-course reactions (Figs. 2 and 3) reflects that Gd_LsdA had no preference for coupling the fructosyl unit to either the glucose moiety (P1a) or to an aromatic ring (P1b) of phlorizin. P1a and P1b were hardly elongated to di- or tri-fructosides. P1a was not hydro- lyzed after sucrose depletion and thus accumulated to represent 35.2 % of all the fructose-conjugates at the end of the optimized 50-h reaction peak at prolonged incubation times (10 24 h) in the reactions per- formed at a low sucrose concentration (146 mM) (Figs. 2 and 3). The hydrolysis of P1b releases free fructose and phlorizin, and thus it has a negative effect on the final percentage of phlorizin conversion.

Gd_LsdA synthesized at least a third mono-fructoside isomer (P1c), which is the immediate precursor of the structurally characterized di- fructoside P2c (4’-O-di-β-(2→6)-fructosyl phlorizin) (Fig. 8b). The low abundance of P1c (phlorizin-4’-O-monofructofuranoside) even at early reaction times suggests its rapid elongation to P2c by creating a β-(2→6) fructosylation resulted in miXture of different mono-fructosylated linkage between the two fructose residues. The progressive increased concentration of P2c along the reaction and the low intensity of the peaks corresponding to phlorizin tri-fructosides (Figs. 2 and 3) indicate that Gd_LsdA does not further fructosylate P2c, at least in an effective way. P2c was not hydrolyzed when sucrose was depleted and thus remained as the second most abundant phlorizin fructo-conjugate (29.8 %) after 50-h reaction (Fig. 6). Two decades ago, Hern´andez et al. [34] demonstrated that Gd_LsdA hardly use the self-synthesized FOSs as alternative fructosyl donors. In the absence of the natural substrate su- crose, the enzyme does cleave the terminal fructose unit (exohydrolysis) of the β-(2→6) linked polysaccharide levan.

Fig. 7. HMBC spectrum of β-D-fructofuranosyl-(2→6)-phlorizin (P1a) in CD3OD. The observed correlations H→C which demonstrate the interconnection between phlorizin and the fructosyl moiety are indicated with curved arrows.

Fig. 8. Enzymatic synthesis of phlorizin fructosides by Gd_LsdA. a) Potential fructosylation sites in the 3D structure of phlorizin taken from crystal structure [28,29]. b) Schematic synthesis of phlorizin mono-fructosides P1a, P1b and di-fructoside P2c.

The acceptor promiscuity of levansucrases offers the possibility of transferring the fructosyl moiety of sucrose to different phenolic com- pounds. The flavonoids puerarin (daidzein-8-C-glucoside) and phlorizin (phloretin-2’-β-D-glucopyranoside) were fructosylated by Gd_LsdA with high conversion (73-93 %) but different degrees of regio-selectivity [19,20, this work]. The enzyme synthesized β-D-fructofuranosyl- (2→6)-puerarin as the sole mono-fructoside, while phlorizin products. Three other levansucrases (Bs_SacB, Lm_LevS and Zm_LevU) fructosylated puerarin with lower efficiencies (7–24 %) and formed mainly puerarin-4’-O-β-D-fructofuranoside [20]. These findings reveal that the efficiency and selectivity of flavonoid fructosylation by levan- sucrases are clearly dependent on both the intrinsic properties of the enzyme and the chemical structure of the acceptor. The number of phenolic rings, presence of sugar substituents and reactivity of the hy- droXyl groups are key aspects that determine the recognition and appropriate positioning of the acceptor at the enzyme active site enabling a productive fructosyl transfer event. Increasing the specificity of the binding interactions between Gd_LsdA and phlorizin via protein engineering may help improve the regio-selectivity of the acceptor reaction.

3.6. Water solubility and anti-oxidant activity of β-D-fructofuranosyl- (2→6)-phlorizin (P1a)

The product β-D-fructofuranosyl-(2→6)-phlorizin (P1a) accumulated as the major phlorizin fructo-conjugate in the Gd_LsdA reaction miXture. Purified P1a was almost 16 times more soluble in water (30.57 g L—1 at 25 ◦C) than the natural compound phlorizin (1.93 g L—1 at 25 ◦C)
polyfructosylation of puerarin through β-(2→6) linkages improved in 3 orders the aqueous solubility of the flavone [20]. Our finding provides further evidence that the solubility, and by consequence the bioavail- ability, of plant flavonoids can be greatly enhanced via enzymatic fructosylation.

The free radical scavenging capacity dropped from 62 % in phlorizin to 44 % in P1a (Table 2). Even thus, the anti-oXidant activity of P1a remains higher than the values reported for puerarin (33 %) and β-D- fructofuranosyl-(2→6)-puerarin (26 %) [19]. It is notable that the fructosyl unit linked to the C6’’ of the only glucosyl residue of either phlorizin or puerarin masked the free radical scavenging capacity of the two phenolic compounds. Contrasting, puerarin-7-O-fructoside dis- played a 5 % increase in its anti-oXidant activity compared to untrans- formed puerarin [35]. In the present study, Gd_LsdA synthesized at least three regioisomers of phlorizin mono-fructoside. Unlike P1a, phlor- izin-4-O-monofructofuranoside (P1b) and phlorizin-4’-O-mono- fructofuranoside (P1c) were not accumulated in the reaction miXture, which hampered their purification and functional characterization. Thus, it remains unknown if the anti-oXidant activity can be influenced by the position of the incorporated fructosyl unit in phlorizin mono-conjugates.

4. Conclusion

Enzymatic fructosylation was approached as a method to enhance the aqueous solubility of the natural dihydrochalcone phlorizin and consequently its bioavailability. Levansucrase from Gluconacetobacter diazotrophicus (Gd_LsdA) distinguished among three enzymes assayed for catalyzing fructosyl transfer reactions from sucrose to the non- conventional acceptor. Gd_LsdA synthesized series of phlorizin mono- di- and tri-fructosides with the conversion efficiency reaching 73 % in an optimized aqueous reaction. The three major products were character- ized as β-D-fructofuranosyl-(2→6)-phlorizin (P1a), phlorizin-4’-O-β-D- fructofuranosyl-(2→6)-D-fructofuranoside (P2c), and phlorizin-4-O- monofructofuranoside (P1b). Purified P1a was 16-fold more soluble in water than phlorizin and retained considerable anti-oXidant activity. Gd_LsdA is an attractive candidate enzyme for the scaled one-step syn- thesis of phlorizin fructosides in the absence of co-solvent. These new fructoconjugates may combine the anti-oXidant activity of phlorizin and the prebiotic property of fructans.

CRediT authorship contribution statement

Azucena Herrera-Gonza´lez: Investigation, Visualization, Writing original draft. Gema Nún˜ez-Lo´pez: Investigation, Formal analysis, Visualization. Nelson Nún˜ez-Dallos: Formal analysis, Investigation, Data curation, Writing – review & editing. Lorena Amaya-Delgado: Resources, Funding acquisition. Georgina Sandoval: Resources, Fund- ing acquisition. Magali Remaud-Simeon: Project administration, Su- pervision, Funding acquisition in France. Sandrine Morel: Conceptualization, Methodology, Investigation, Visualization, Writing- review & editing. Javier Arrizon: Project administration, Supervision, Visualization, Funding acquisition in Mexico. La´zaro Herna´ndez: Conceptualization, Methodology, Supervision, Visualization, Writing- review & editing a Free radical scavenging activity of 2,2-diphenyl-1-picrylhydrazyl (DPPH) was considered as 100 %.

Author agreement statement

We declare that this manuscript is original, has not been published before and is not currently being considered for publication elsewhere. We confirm that the manuscript has been read and approved by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the manuscript has been approved by all of us.

Compliance with ethical standards

This article does not contain any studies conducted by any of the authors involving human participants or animals.

Declaration of Competing Interest

The authors report no declarations of interest.

Acknowledgements

We acknowledge the bilateral project CONACYT-ANUIES/ECOS- NORDM14A01, the CONACYT project CB-2012-01/000000000181766
for financial support and the CONACYT Ph.D. fellowship 237785. We are grateful to Metasys, the Metabolomics & FluXomics Center at the Laboratory for Engineering of Biological Systems & Processes (Toulouse, France), for NMR experiments. We thank the ICEO facility dedicated to enzyme screening and discovery, and Part of the Integrated Screening Platform of Toulouse (PICT, IBiSA) for providing access to HPLC equipment and protein purification system.

Appendix A. Supplementary data

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.enzmictec.2021.10 9783.

References

[1] J.R.L. Ehrenkranz, N.G. Lewis, C.R. Kahn, J. Roth, Phlorizin: a review, Diabetes Metab. Res. Rev. 21 (2005) 31–38, https://doi.org/10.1002/dmrr.532.
[2] B.M. Rezk, G.R.M.M. Haenen, W.J.F. Van der Vijgh, A. Bast, The antioXidant activity of phloretin: the disclosure of a new antioXidant pharmacophore in flavonoids, Biochem. Biophys. Res. Commun. 295 (2002) 9–13, https://doi.org/ 10.1016/S0006-291X(02)00618-6.
[3] Z. Xiao, Y. Zhang, X. Chen, Y. Wang, W. Chen, Q. Xu, P. Li, F. Ma, EXtraction, identification, and antioXidant and anticancer tests of seven dihydrochalcones from Malus ‘Red Splendor’ fruit, Food Chem. 231 (2017) 324–331, https://doi.org/ 10.1016/j.foodchem.2017.03.111.
[4] X. Qin, Y. Xing, Z. Zhou, Y. Yao, Dihydrochalcone compounds isolated from crabapple leaves showed anticancer effects on human cancer cell lines, Molecules 20 (2015) 21193–21203, https://doi.org/10.3390/molecules201219754.
[5] R.P. Pandey, F. Li, E. Kim, T. Yamaguchi, I. Park, S. Kim, K. Sohng, Enzymatic synthesis of novel phloretin glucosides, Appl. Environ. Microbiol. 79 (2013) 3516–3521, https://doi.org/10.1128/AEM.00409-13.
[6] L. Xu, T. Qi, L. Xu, L. Lu, M. Xiao, Recent progress in the enzymatic glycosylation of phenolic compounds, J. Carbohydr. Chem. 35 (2016) 1–23, https://doi.org/ 10.1080/07328303.2015.1137580.
[7] B. Hofer, Recent developments in the enzymatic O-glycosylation of flavonoids, Appl. Microbiol. Biotechnol. 100 (2016) 4269–4281, https://doi.org/10.1007/ s00253-016-7465-0.
[8] V. Kˇren, Glycoside vs. aglycon: the role of glycosidic residue in biological activity, in: J. Fraser-Reid, Tatsuta B.O, Thiem K (Eds.), Glycoscience, second edi, Springer, Berlin Heidelberg, Berlin, Heidelberg, 2008, pp. 2589–2644, https://doi.org/ 10.1007/978-3-540-30429-6_67.
[9] T. Desmet, W. Soetaert, P. Bojarova, V. Kaˇren, L. Dijkhuizen, V. Eastwick-Field,
A. Schiller, Enzymatic glycosylation of small molecules: challenging substrates require tailored catalysts, Chem. – Eur. J. 18 (2012) 10786–10801, https://doi.org/ 10.1002/chem.201103069.
[10] A. Herrera-Gonza´lez, G. Nún˜ez-Lo´pez, S. Morel, L. Amaya-Delgado, G. Sandoval,
A. Gschaedler, M. Remaud-Simeon, J. Arrizon, Functionalization of natural compounds by enzymatic fructosylation, Appl. Microbiol. Biotechnol. 101 (2017) 5223–5234, https://doi.org/10.1007/s00253-017-8359-5.
[11] N.H. Thuan, J.K. Sohng, Recent biotechnological progress in enzymatic synthesis of glycosides, J. Ind. Microbiol. Biotechnol. 40 (2013) 1329–1356, https://doi.org/ 10.1007/s10295-013-1332-0.
[12] H. Overwin, V. Wray, B. Hofer, Biotransformation of phloretin by amylosucrase yields three novel dihydrochalcone glucosides, J. Biotechnol. 211 (2015) 103–106, https://doi.org/10.1016/j.jbiotec.2015.07.002.
[13] B.L. Cantarel, P.M. Coutinho, C. Rancurel, T. Bernard, V. Lombard, B. Henrissat, The Carbohydrate-Active EnZymes database (CAZy): an expert resource for Glycogenomics, Nucleic Acids Res. 37 (2009) D233–D238, https://doi.org/ 10.1093/nar/gkn663.
[14] W. Lammens, K. Le Roy, L. Schroeven, A. Van Laere, A. Rabijns, W. Van Den Ende, Structural insights into glycoside hydrolase family 32 and 68 enzymes: functional implications, J. EXp. Bot. 60 (2009) 727–740, https://doi.org/10.1093/jXb/ ern333.
[15] E.T. O¨ ner, L. Hernandez, J. Combie, Review of levan polysaccharide: from a
century of past experiences to future prospects, Biotechnol. Adv. 34 (2016) 827–844, https://doi.org/10.1016/j.biotechadv.2016.05.002.
[16] A. Mena-Arizmendi, J. Alderete, S. A´guila, A. Marty, A. Miranda-Molina, A. Lo´pez-
Munguía, E. Castillo, Enzymatic fructosylation of aromatic and aliphatic alcohols by Bacillus subtilis levansucrase: reactivity of acceptors, J. Mol. Catal. B Enzym. 70 (2011) 41–48, https://doi.org/10.1016/j.molcatb.2011.02.002.
[17] O. Kirtel, C. Men´endez, M. Versluys, W. Van den Ende, L. Hern´andez, E. Toksoy O¨ ner, Levansucrase from Halomonas smyrnensis AAD6T: first halophilic GH-J clan
enzyme recombinantly expressed, purified, and characterized, Appl. Microbiol. Biotechnol. 102 (2018) 9207–9220, https://doi.org/10.1007/s00253-018-9311-z.
[18] L.E. Trujillo, R. Go´mez, A. Banguela, M. Soto, J.G. Arrieta, L. Hern´andez, Catalytical properties of N-glycosylated Gluconacetobacter diazotrophicus levansucrase produced in yeast, Electron. J. Biotechnol. 7 (2004) 115–123, https:// doi.org/10.2225/vol7-issue2-fulltext-4.
[19] G. Nún˜ez-Lo´pez, A. Herrera-Gonz´alez, L. Hern´andez, L. Amaya-Delgado,
G. Sandoval, A. Gschaedler, J. Arrizon, M. Remaud-Simeon, S. Morel, Fructosylation of phenolic compounds by levansucrase from Gluconacetobacter diazotrophicus, Enzyme Microb. Technol. 122 (2019) 19–25, https://doi.org/ 10.1016/j.enzmictec.2018.12.004.
[20] G. Nún˜ez-Lo´pez, S. Morel, L. Hern´andez, A. Musacchio, L. Amaya-Delgado,
A. Gschaedler, M. Remaud-Simeon, J. Arrizon, One-pot bi-enzymatic cascade synthesis of puerarin polyfructosides, Carbohydr. Polym. 247 (2020), 116710, https://doi.org/10.1016/j.carbpol.2020.116710.
[21] J. Arrieta, L. Herna´ndez, A. Coego, V. Suarez, E. Balmori, C. Men´endez, M.-F. Petit- Glatron, R. Chambert, G. Selman-Housein, Molecular characterization of the levansucrase gene from the endophytic sugarcane bacterium Acetobacter diazotrophicus SRT4, Microbiology 142 (1996) 1077–1085, https://doi.org/ 10.1099/13500872-142-5-1077.
[22] L. Herna´ndez, C. Men´endez, E.R. P´erez, D. Martínez, D. Alfonso, L.E. Trujillo,
R. Ramírez, A. Sobrino, Y. Mazola, A. Musacchio, E. Pimentel, Fructooligosaccharides production by Schedonorus arundinaceus sucrose:sucrose 1- fructosyltransferase constitutively expressed to high levels in Pichia pastoris,
J. Biotechnol. 266 (2018) 59–71, https://doi.org/10.1016/j.jbiotec.2017.12.008.
[23] G.L. Miller, Use of dinitrosalicylic acid reagent for determination of reducing sugar, Anal. Chem. 31 (1959) 426–428, https://doi.org/10.1021/ac60147a030.
[24] C.H. Lee, L. Yang, J.Z. Xu, S.Y.V. Yeung, Y. Huang, Z.-Y. Chen, Relative antioXidant activity of soybean isoflavones and their glycosides, Food Chem. 90 (2005) 735–741, https://doi.org/10.1016/j.foodchem.2004.04.034.
[25] P. Guti´errez-Alonso, L. Ferna´ndez-Arrojo, F.J. Plou, M. Ferna´ndez-Lobato, Biochemical characterization of a β-fructofuranosidase from Rhodotorula dairenensis with transfructosylating activity, FEMS Yeast Res. 9 (2009) 768–773, https://doi.org/10.1111/j.1567-1364.2009.00526.X.
[26] C. Men´endez, D. Martínez, E.R. P´erez, A. Musacchio, R. Ramírez, A. Lo´pez- Munguía, L. Hern´andez, Engineered thermostable β–fructosidase from Thermotoga maritima with enhanced fructooligosaccharides synthesis, Enzyme Microb. Technol. 125 (2019) 53–62, https://doi.org/10.1016/j.enzmictec.2019.02.002.
[27] K. Xü, R. Deng, H.T. Lü, Preparative isolation and purification of phlorizin from apple tree bark by high-speed counter-current chromatography, Asian J. Chem. 23 (2011) 3807–3810.
[28] M. Auf’mkolk, J. Koehrle, R.-D. Hesch, S.H. Ingbar, V. Cody, Crystal structure of phlorizin and the iodothyronine deiodinase inhibitory activity of phloretin analogues, Biochem. Pharmacol. 35 (1986) 2221–2227, https://doi.org/10.1016/ 0006-2952(86)90595-2.
[29] W.-C. Shin, The crystal and molecular structure of phlorizin dihydrate, Bull. Korean Chem. Soc. 6 (1985) 7–11.
[30] A. Gutmann, C. Krump, L. Bungaruang, B. Nidetzky, A two-step O- to C-glycosidic bond rearrangement using complementary glycosyltransferase activities, Chem. Commun. 50 (2014) 5465–5468, https://doi.org/10.1039/C4CC00536H.
[31] A. Homann, R. Biedendieck, S. Gotze, D. Jahn, J. Seibel, Insights into polymer versus oligosaccharide synthesis: mutagenesis and mechanistic studies of a novel levansucrase from Bacillus megaterium, Biochem. J. 407 (2007), 289-198.
[32] A. Miranda-Molina, E. Castillo, A. Lopez, A novel two-step enzymatic synthesis of blastose, a β-D-fructofuranosyl-(2→6)-D-glucopyranose sucrose analogue, Food Chem. 227 (2017) 202–210.
[33] L. Herna´ndez, Y. Quintero, A. Musacchio, Fructan Biosynthesis in Bacteria, in: Microbial EXopolysaccharides: Current Research and Developments, 2019,
pp. 133–164, https://doi.org/10.21775/9781912530267.06.
[34] L. Hernandez, J. Arrieta, C. Menendez, R. Vazquez, A. Coego, V. Suarez, G. Selman,
M.F. Petit-Glatron, R. Chambert, Isolation and enzymic properties of levansucrase secreted by Acetobacter diazotrophicus SRT4, a bacterium associated with sugar cane, Biochem. J. 309 (Pt 1) (1995) 113–118, https://doi.org/10.1042/ bj3090113.
[35] G. Liu, L. Sun, S. Wang, C. Chen, T. Guo, Y. Ji, X. Li, G. Huang, H. Wei, Y. Dai,
S. Yuan, HydroXylation modification and free radical scavenging activity of puerarin-7-O-fructoside, Folia Microbiol. Praha 56 (2011) 305–311, https://doi. org/10.1007/s12223-011-0052-y.