Formation of Phosphoglycosides in Caenorhabditis elegans: A Novel Biotransformation Pathway

Background Caenorhabditis elegans (C. elegans) has become a widely used model to explore the effect of food constituents on health as well as on life-span extension. The results imply that besides essential nutrients several flavonoids are able to impact the aging process. What is less investigated is the bioavailability and biotransformation of these compounds in C. elegans. In the present study, we focused on the soy isoflavone genistein and its metabolism in the nematode as a basis for assessing whether this model system mimics the mammalian condition. Principal Findings C. elegans was exposed to 100 µM genistein for 48 hours. The worm homogenate was extracted and analyzed by liquid chromatography (LC). 11 metabolites of genistein were detected and characterized using LC electrospray ionization mass spectrometry. All genistein metabolites formed by C. elegans were found to be sugar conjugates, primarily genistein-O-glucosides. The dominant metabolite was identified as genistein-7-O-phosphoglucoside. Further interesting metabolites include two genistein-di-O-glycosides, a genistein-O-disaccharide as well as a genistein-O-phosphodisaccharide. Conclusions/Significance Our study provides evidence for a novel biotransformation pathway in C. elegans leading to conjugative metabolites which are not known for mammals. The metabolism of genistein in mammals and in C. elegans differs widely which may greatly impact the bioactivity. These differences need to be appropriately taken into consideration when C. elegans is used as a model to assess possible health or aging effects.


Introduction
Metabolic transformation plays a key role in mammalian and non-mammalian defense against xenobiotics, which might exert beneficial or harmful effects on biological systems. Biotransformation generally includes phase I reactions in which functional groups are either introduced to the xenobiotic or exposed, followed by conjugation reactions with water-soluble moieties. These chemical modifications change the physicochemical properties of the xenobiotic which in turn alters also its biological and physiological effects. Consequently, detailed information on the metabolism of a foreign compound is crucial for a reliable assessment of potential benefits but also risks associated with its exposure. When using an animal model to investigate the biological activity of a substance it is important to keep in mind that its response, especially its metabolism, may differ between the model system and that of humans. Non-mammalian model organisms such as the nematode Caenorhabditis elegans (C. elegans) have become an important tool in systems biology research. The genome of C. elegans is fully sequenced and many physiological functions found in mammals have been identified in the nematode as well. Due to the relatively high degree of complexity and the small total number of 959 somatic cells C. elegans provides a controllable experimental system employed in developmental research, cell physiology and aging [1].
The aging trajectory is controlled by genetic mechanisms involving signal transduction pathways that are evolutionarily conserved in yeasts, worms, flies and mammals including humans [2]. In nutrition research, C. elegans has become a widely used model to explore the effect of food constituents on health as well as on life-span extension. The results imply that essential nutrients like tocopherols but also plant secondary compounds from fruit, vegetables and herbs, especially polyphenols, are able to impact the aging process. Such effects on life span of C. elegans have recently been described for example for the proanthocyanidin fraction derived from blueberries, epigallocatechin gallate from green tea, or quercetin derivatives from onions [3][4] [5]. However, what is frequently not assessed in those studies is the extent by which the test compounds are bioavailable and moreover, to which extent and into which metabolites the foodderived compounds are converted. Knowledge about metabolic conjugation reactions in C. elegans is thus very limited and so far, only a few reports have addressed this question [6] [7].
We recently assessed for four structurally related flavonoids the apparent bioavailability in C. elegans and observed indications for a substantial metabolism [8]. In the present study we characterized the biotransformation of genistein, a prominent isoflavone found in soy and soy products by C. elegans. Very recently Fischer et al. reported that genistein affect the immunity of the worm by acting antiestrogenic via a reduced expression of the estrogen-responsive vitellogenin-genes [9].
In mammalian systems the metabolism of genistein is well described and reveals an impressive complexity [10]. Our aim was to identify and quantify the metabolites to which genistein is converted in C. elegans as a basis for assessing whether this model system mimics the mammalian condition.
Caenorhabditis elegans strain and culture conditions C. elegans (strain Bristol N2) and Escherichia coli (E. coli) strain OP50 were obtained from the Caenorhabditis Genetics Center (CGC), University of Minnesota, USA. The nematodes were grown at 20uC on nematode growth medium (NGM) agar plates as previously described [12] [13]. We used heat killed E. coli OP50 as food source in order to prevent any bacterial transformation of genistein. Bacteria were grown overnight, concentrated 5-fold by centrifugation and heat killed at 65uC for 30 min according to [14]. Heat killed OP50 feeding solution and genistein stock solution were added to the NGM plates. The final concentration of genistein was 100 mM.

Sample preparation
The nematodes were grown on NGM plates containing 100 mM of genistein for 48 hours. After exposure, worms were washed five

LC-DAD and LC-MS analyses
The LC-DAD analyses were performed on a Shimadzu LC system equipped with a controller (CBM-20A), a degasser (DGU-20A3), two pumps (LC-20AD), an autosampler (SIL-20AC HT), a column oven (CTO-20AC) and a DAD (SPD-M20A). The LC system was controlled by the software LCsolution 1.24. LC separation was carried out on a YMC Pack Hydrosphere C18 column (15063.0 mm, particle size 3 mm) equipped with a Phenomenex SecurityGuard (C18, 4.063.0 mm). Eluent A was ammonium acetate (25 mM, pH 7.0) and eluent B was acetonitrile. A linear gradient was used with a flow rate of 0.8 ml/min and the following elution profile: 0-5 min isocratic with 10% B, 5-16 min from 10% to 20% B, 16-30 min from 20% to 42% B, 30-31 min from 42% to 95% B, 31-36 min isocratic with 95% B, 36-37 min from 95% to 10% B and 37-47 min isocratic with initial conditions. The column oven and the DAD flow cell were adjusted to 40uC. The injection volume was 20 ml. The chromatograms were recorded from 200 to 400 nm and the 260 nm trace was used to monitor the analytes.
The LC-MS analyses were performed on two systems. The first system was an ABSciex QTrap 5500 mass spectrometer equipped with a Shimadzu LC system, which consisted of a controller (CBM-20A), a degasser (DGU-20A5), two pumps (LC-30AD), an autosampler (SIL-30AC) and a column oven (CTO-20AC). The LC-MS system was controlled by the software Analyst 1.5.2. The LC conditions were the same as described above. The electrospray ionization (ESI) source was operated in the negative mode using the following parameters: Curtain gas (CUR) 20 psi, ion spray voltage (IS) 23500 V, ion source gas-1 (GS 1) 60 psi, ion source gas-2 (GS 2) 40 psi, ion source gas-2 temperature (TEM) 600uC. The MS parameters were adjusted as follows: scan rate 20,000 Da/sec, declustering potential (DP) 2250 V, entrance potential (EP) 210 V. The MS full scans (enhanced MS mode) were performed with a scan range from 100 to 755 m/z and a collision energy voltage (CE) of 210 V. The MS/MS measurements (enhanced product ion mode) were executed with a collision energy voltage (CE) of 240 V and collision energy spread of 30 V. Nitrogen was used as collision gas.
The second LC-MS system was an Agilent 6540 QToF mass spectrometer equipped with an Agilent 1290 Infinity LC system, which consisted of a controller, a degasser, a binary pump, an autosampler, a column oven and a DAD. The system was controlled by the software MassHunter B.03.01 Build 3.1.346.0. The LC conditions were the same as described above. The jet stream ESI source was operated in the negative mode and the MS parameters were adjusted as follows: Nebulising pressure 50 psi, gas temperature 350uC, gas flow 10 l/min, capillary voltage (VCap) 3500 V, skimmer voltage 65 V, fragmentor voltage 175 V, Oct 1 RF Vpp voltage 750 V, scan range from 100 to 1100 m/z, scan rate 1 spectrum/sec. The MS/MS experiments were performed with a CID voltage of 30 V. Nitrogen was used as collision gas.

Results
C. elegans converted the isoflavone genistein to at least 11 metabolites which were detected and characterized by LC-DAD and LC-MS analysis. Figure 1 shows the LC-UV chromatogram of the organic extract derived from C. elegans exposed to 100 mM genistein for 48 h. The peaks (denoted as M1 to M11) were numbered according to their order of elution. All peaks that were not present in control incubations without genistein were assumed to be metabolites and characterized by interpretation of their MS, MS/MS and MS-TOF data. In sum, all metabolites detected were identified as genistein sugar conjugates. In the following, the mass spectrometric data of each metabolite will be discussed and explained. Because of the similarity of their mass spectra to those of M10 and M11, we assume that these metabolites are genistein-7-and genistein-49-O-hexoside stereoisomers, carrying a different hexose sugar moiety, e.g. galactose instead of glucose. The presence of a genistein-5-O-glucoside seems less likely because the 5-OH group of the A-ring is involved in intramolecular hydrogen bond formation with the 4-carbonyl group of the isoflavone C-ring [16]. However, this alternative cannot be completely ruled out. Figure 2A   hexosides, the MS/MS spectrum of peak 7 exhibited, if at all, a relatively low abundance of the fragment ion at m/z 431. Furthermore an intense genistein aglycone ion at m/z 269 was detected, indicating the neutral loss of 324 Da (m/z 593R269), which corresponds to the characteristic fragment mass of an anhydro-disaccharide unit, is preferred ( Figure 3A). We observed the same behaviour when using p-nitrophenyl-b-D-maltoside to study the fragmentation of O-disaccharide-substituted phenols. Therefore, genistein-7-O-b-D-maltoside and genistein-49-O-b-Dmaltoside were prepared enzymatically from genistein-7-O-b-Dglucoside and genistein-49-O-b-D-glucoside, respectively. The maltosides had exactly the same mass spectra as peak 7, but eluted later. In conclusion, M7 is a genistein-O-disaccharide, conceivably a genistein-glucosyl-glucoside since glucose conjugates represented the vast majority of genistein metabolites formed by C.

Detection of genistein-O-monohexosides
elegans. Because we can exclude the (1R4)-linked maltoside, we propose that M7 most likely is a genistein-O-b-glucosyl-(1R6)glucoside (genistein-gentiobioside) or a genistein-O-b-glucosyl-(1R2)-glucoside as these two linkages are the most common ones found in nature. The fragmentation pattern of genistein-di-Ohexosides compared to genistein-O-disaccharides is depicted in Figure 3.

Detection of genistein-O-phospoglucosides
The main metabolite detected was M5 with a molecular [M-H] 2 ion at m/z 511, which was confirmed by precursor ion experiments scanning for m/z 269, the genistein anion. In addition, the corresponding sodium [M-2H+Na] 2 and potassium [M-2H+K] 2 adduct ions at m/z 533 and m/z 549, respectively, were observed ( Figure 4A). The accurate mass of the molecular  Table 1. Predicted formula, determined and theoretical mass/charge ratios as well as the mass error of fragment ions observed in the product ion mass spectrum of m/z 511.  ion was determined to be 511.0652 by (ESI)-TOF experiments ( Table 1). The EPI scan at m/z 511 produced, as illustrated in Figure 4B, fragment ions at m/z 269, 241, 97, and 79.     Figure 4C). Overall the MS data indicates that the main metabolite of genistein in C. elegans is a genistein-O-monophosphohexoside. To confirm these results, peak 5 was isolated and incubated with acid phosphatase, which led to the release of genistein-7-O-b-D-glucoside (data not shown). We therefore conclude that the main metabolite formed is a genistein-7-O-(60-O-phospho)-b-D-glucoside. The proposed fragmentation pathway is depicted in Figure 5.  Figure 6B). Thus we assume that the metabolite M4 is a genistein-O-phosphodissacharide. Table 2 summarizes the experimental data as well as the proposed structure of each metabolite.

Discussion
Our studies on the metabolism of the isoflavone genistein in C. elegans revealed a novel biotransformation pathway. All metabolites identified are products formed by the worm since the animals were provided with heat-killed bacteria. Moreover, vital E. coli were found as unable to produce any of the genistein metabolites detected (data not shown).
All genistein metabolites formed by C. elegans were identified as sugar conjugates, primarily genistein-O-glucosides. Similar glucosylation reactions of flavonoids in non-mammalian systems are primarily reported for some bacteria species, especially aerobic or facultative anaerobic rod-shaped bacteria such as Bacillus subtilis, Bacillus cereus, Xanthomonas campestris and Lactobacillus delbrueckii [17] [18]. More recently Laing et al. [6] were the first to describe an analogous conjugation reaction in the nematode C. elegans where a glucose moiety was introduced into the anthelmintic drug albendazole as a xenobiotic response and a variety of quercetin glycosides including sulfate-glycosyl derivatives were recently found by Surco-Laos et al. [7] as metabolites of the flavonoid quercetin by C. elegans.
In our study, the dominant genistein metabolite in N2 C. elegans was reliably identified as a 7-O-phosphoglucoside. To our knowledge, this is the first time that phosphoglucosides as biotransformation reaction products are identified in C. elegans. By the high concentrations found, it seems that this is the predominant conjugation pathway in the worms. Since we observed both, the genistein-7-O-glucoside and the corresponding phosphoglucoside, we speculate that the formation may occur in a two-step reaction. C. elegans expresses a variety of UDP-glycosyltransferases which could catalyze the generation of an O-glucoside, followed by phosphorylation of the glucose moiety by the action of a hexokinase (Figure 7). 265 glycosyltransferase genes have been identified for C. elegans to date of which 29% belong to the UDPglycosyltransferase family [19]. Preliminary experiments in our lab using RNA interference in C. elegans with the goal to identify UDPglycosyltransferases that could catalyze the formation of the Oglucosides failed as in most cases silencing was lethal (data not shown). To identify the relevant enzymes involved in this pathway will be challenging.
Genistein-O-disaccharides and phosphorylated genistein-O-disaccharides were only formed in minor amounts. However, disaccharide conjugates are a new metabolite class as well and have never been reported as products of the C. elegans xenobiotic defense system. The pathway of their formation is equally unclear. We have not identified any hydroxylated genistein metabolites in worms known to be formed by cytochrome P450 enzymes in mammalian species (data not shown). However, even in mammals this reaction is only a side-pathway since genistein already possesses three hydroxyl groups. Figure 8 compares the conjugative metabolism of genistein in C. elegans and humans. In contrast to the formation of phosphoglucosides and glucosides described here for C. elegans, mammalian systems appear to metabolize genistein to glucuronides, sulfates and mixed sulfoglucuronides exclusively. [20] [21]. In addition, in humans, the microbial metabolism and degradation of genistein results in products like dihydrogenistein, 69-hydroxy-O-desmethylangolensin, and 2-(4-hydroxyphenyl)-propionic acid which influence the bioactivity dramatically. In C. elegans none of these microbial metabolites were detected.
Very recently, Fischer et al. made the interesting observation that daidzein exhibits an estrogenic effect in C. elegans which is proposed to be responsible for lifespan prolongation of the nematode in the presence of pathogenic bacteria, whereas genistein acts in an antiestrogenic manner, diminishes the resistance against the pathogen and reduces lifespan under the same conditions [9]. Based on our results, we suggest that the effect observed might not be caused by the parent compound itself but by the phophoglucosides identified as the main metabolites that were formed. It is described that the nature of the moiety attached can affect the bioactivity of a metabolite considerably. Glucuronidation of daidzein at the 49-and 7-position resulted, for example, in a significant reduction in estrogenicity, although the activity was not completely eliminated [22]. Sulfation of daidzein in the 49-position only modestly influenced the estrogen agonist activity and even more, the 7-O-sulfate of daidzein exerted a much higher activity than daidzein itself [23]. In the case of genistein, the 7-O-sulfation acts in an opposite way and reduced the estrogenic activity substantially. However, not only estrogenicity has been shown to be affected by the type of conjugation. Daidzein sulfoconjugates are also described as potent inhibitors of sterol sulfatase while daidzein does not affect this enzyme [24].
In conclusion, our study provides evidence for a novel biotransformation pathway in C. elegans leading to the formation of glucoside-and phosphoglucosides derivatives which so far have not been identified in a mammalian system. The metabolism of genistein in mammals and C. elegans therefore shows huge differences which may greatly impact the bioactivity. These differences need to be appropriately taken into consideration when C. elegans is used as a model to assess possible health or aging effects.