Flavonol 3-O-galactosides (F3Gals) are important flavonoid glycosides, in which a galactose is attached to the C3-OH of flavonol aglycones. Common F3Gals include trifolin (kaempferol 3-O-galactopyranoside, 1c), hyperoside (quercetin 3-O-galactopyranoside, 2c), and myricetin 3-O-galactopyranoside (3c) (Fig. 1A). F3Gals exhibit a variety of activities, including anticancer activity [1], anti-inflammatory property, antinociceptive action [2], insecticidal effect [3], antifungal activity [4], hepatoprotective potential [5], and antioxidant action [6]. In addition, F3Gals are rich in many edible plants, such as the leaves of Zanthoxylum bungeanum [6], flowers of Citrus [7], and fruits of Prunus genus [8]. Hence, F3Gals have the promising potential to be developed as food supplements and innovative drugs.

Additionally, F3Gals exist widely in the form of esters. The acylation of most known F3Gals occurs at the primary 6′′-OH of the galactose moiety, forming various 6′′-O-esters such as acetates [9,10], malonates [11], p-hydroxybenzoates [12], gallates [13], coumarates [14], and caffeates [15] (Fig. S1). These F3Gal 6′′-O-esters display diverse activities, including anti-complement activity [16], antioxidant action [17], and inhibitory effect on aldose reductase [14]. The evidence collectively indicated that acylation at the primary 6′′-OH of the galactose moiety was conducive to the beneficial properties of F3Gals.

Of these 6′′-O-esters, F3Gal 6′′-O-acetates (F3Gal-6′′As), such as kaempferol-3-O-(6′′-O-acetyl)-galactoside (trifolin 6′′-O-acetate, T6′′A, 1d) [9,10], quercetin-3-O-(6′′-O-acetyl)-galactoside (hyperoside 6′′-O-acetate, H6′′A, 2d) [11,18,19], and myricetin-3-O-(6′′-O-acetyl)-galactoside (MGal6′′A, 3d), are the common esters (Fig. 1A). These F3Gal-6′′As display diverse pharmaceutical activities. For example, T6′′A and H6′′A exhibit antioxidant activity [20,21]. Moreover, T6′′A is the major component of Sambucus chinensis Lindl, the traditional Chinese medicine exhibiting bacteriostatic, anti-inflammatory, hepatoprotective and antioxidant activities [22]. These data suggested that the F3Gal-6′′As had the positive effects on the human health. More F3Gal-6′′As should be obtained for further investigation. However, the content of F3Gal-6′′As in plants is low. For example, the content of T6′′A in Trifolium repens L, Pinus densiflora and Apocynum venetum is 0.002222 % [21], 0.00047 % [20], and 0.000116 % [23], respectively. The low content limited the large-scale preparation of F3Gal-6′′As, which was not conducive to their further investigation. Many attempts have been made for chemical acetylation of F3Gal-6′′As [[24], [25], [26], [27], [28]]. However, no monoacetylated derivatives of F3Gals have been obtained because of the poor regioselectivity of chemical approaches. Chemical synthesis typically involves the usage of the toxic agents. Consequently, a promising and eco-friendly alternative capable of producing F3Gal-6′′As in large-scale should be pursued.

The great strides made in the synthetic biology and metabolic engineering is now opening a new avenue for the bioproduction of F3Gal-6′′As through the cell factories, in which the biosynthetic pathways of F3Gal-6′′As are introduced. The biosynthesis of F3Gal-6′′As from flavanones undergoes four tailoring steps. First, flavanones are converted into flavanonols by a flavanone 3-hydroxylase (F3H). Subsequently, flavanonols are converted to flavonols by a flavonol synthase (FLS). Alternatively, flavanones can be converted into flavonols, catalyzed by a bifunctional FLS with both F3H and FLS activities. Next, the flavonols are galactosylated at their C3-OH groups by a flavonol 3-O-galactosyltransferase (F3GalT) to yield F3Gals. At last, F3Gals can be acetylated to obtain F3Gal-6′′As, catalyzed by an acetyltransferase (AT) (Fig. 1A). As such, F3Gal-6′′As are synthesized from their respective flavanone substrates via the four-step pathway consisting of four enzymes: F3H, FLS, F3GalT, and AT (Fig. 1A).

Among these four enzymes, the members belonging to F3H, FLS, and F3GalT are usually promiscuous [[29], [30], [31], [32], [33], [34]]. For example, two F3Hs, CsF3Ha and CsF3Hb, could convert naringenin (1) and eriodictyol (2) into dihydrokaempferol (1a) and dihydroquercetin (2a), respectively [29]. The flavonol synthases AcFLS-HRB [32] and FtFLS1 [31] recognize both dihydrokaempferol (1a) and dihydroquercetin (2a). In addition, the bifunctional flavonol synthase OcFLS2 is a promiscuous enzyme that recognizes multiple flavanones, such as naringenin (1), eriodictyol (2), pinocembrin, and homoeriodictyol [33]. Members of F3GalT, PhUGT [34], and DkFGT [30], were capable of acting on kaempferol (1b), quercetin (2b), and myricetin (3b), and displayed flavonol substrates promiscuity. Substrate flexibility makes the promiscuous enzymes the promising candidates for constructing a universal pathway, which can yield various metabolites depending on different substrates. The four enzymes F3H, FLS, F3GalT, and AT, however, had never been patched together to synthesize F3Gal-6′′As.

The second hurdle in the construction of the F3Gal-6′′A biosynthetic pathway is the uncertainty of the flavonol 3-O-galactoside 6′′-O-acetyltransferase. To date, limited progress has been made in the functional identification of promiscuous acetyltransferases that are capable of recognizing multiple F3Gals. These disadvantages collectively pose significant challenges to the construction of a platform cell factory capable of yielding diverse F3Gal-6′′As.

Herein, a platform cell factory consisting of four promiscuous enzymes, CsF3Ha, AcFLS-HRB, PhUGT and GAT-D17W, was developed to synthesize diverse F3Gal-6′′As (T6′′A, H6′′A and MGal6′′A). The enzymatically synthesized T6′′A, H6′′A and MGal6′′A displayed an improved liposolubility. In addition, T6′′A and MGal6′′A showed hepatoprotective activities. This study not only provided a platform cell factory for the bioproduction of active F3Gal-6′′As, but also formed a new insight for the application of the promiscuous enzymes.