Glucuronate, the precursor of vitamin C, is directly formed from UDP-glucuronate in liver Carole L. Linster and Emile Van Schaftingen Laboratory of Physiological Chemistry, Universite ´ Catholique de Louvain and the Christian de Duve Institute of Cellular Pathology, Brussels, Belgium Formation of free glucuronate from UDP-glucuronate can be considered as the first step in the synthesis of vitamin C (Fig. 1), a pathway that occurs in most ver- tebrates, although not in guinea pigs and primates, including humans [1]. Free glucuronate can also be converted to pentose phosphate intermediates via the ‘pentose pathway’ [2]. The latter is inter- rupted in subjects with pentosuria, who have a deficiency in l-xylulose reductase and excrete abnormal amounts of l-xylulose [3]. We recently reinvestigated Keywords glucuronate; glucuronate 1-phosphate; UDP-glucuronosyltransferases; vitamin C; xenobiotics Correspondence E. Van Schaftingen, Laboratory of Physiological Chemistry, UCL-ICP, Avenue Hippocrate 75, B-1200 Brussels, Belgium Fax: +32 27 647 598 Tel: +32 27 647 564 E-mail: vanschaftingen@bchm.ucl.ac.be (Received 12 January 2006, revised 2 February 2006, accepted 10 February 2006) doi:10.1111/j.1742-4658.2006.05172.x The conversion of UDP-glucuronate to glucuronate, usually thought to proceed by way of glucuronate 1-phosphate, is a site for short-term regula- tion of vitamin C synthesis by metyrapone and other xenobiotics in isola- ted rat hepatocytes [Linster CL & Van Schaftingen E (2003) J Biol Chem 278, 36328–36333]. Our purpose was to explore the mechanism of this effect in cell-free systems. Metyrapone and other xenobiotics stimulated, by approximately threefold, the formation of glucuronate from UDP-glucuro- nate in liver extracts enriched with ATP-Mg, but did not affect the forma- tion of glucuronate 1-phosphate from UDP-glucuronate or the conversion of glucuronate 1-phosphate to glucuronate. This and other data indicated that glucuronate 1-phosphate is not an intermediate in glucuronate forma- tion from UDP-glucuronate, suggesting that this reaction is catalysed by a ‘UDP-glucuronidase’. UDP-glucuronidase was present mainly in the micro- somal fraction, where its activity was stimulated by UDP-N-acetylglucosa- mine, known to stimulate UDP-glucuronosyltransferases by enhancing the transport of UDP-glucuronate across the endoplasmic reticulum mem- brane. UDP-glucuronidase and UDP-glucuronosyltransferases displayed similar sensitivities to various detergents, which stimulated at low concen- trations and generally inhibited at higher concentrations. Substrates of glucuronidation inhibited UDP-glucuronidase activity, suggesting that the latter is contributed by UDP-glucuronosyltransferase(s). Inhibitors of b-glucuronidase and esterases did not affect the formation of glucuronate, arguing against the involvement of a glucuronidation–deglucuronidation cycle. The sensitivity of UDP-glucuronidase to metyrapone and other stim- ulatory xenobiotics was lost in washed microsomes, even in the presence of ATP-Mg, but it could be restored by adding a heated liver high-speed supernatant or CoASH. In conclusion, glucuronate formation in liver is catalysed by a UDP-glucuronidase which is closely related to UDP- glucuronosyltransferases. Metyrapone and other xenobiotics stimulate UDP-glucuronidase by antagonizing the inhibition exerted, presumably indirectly, by a combination of ATP-Mg and CoASH. Abbreviations ER, endoplasmic reticulum; 4-Np-UGT, 4-nitrophenylglucuronosyltransferase; UDPGlcNAc, UDP-N-acetylglucosamine. 1516 FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS the mechanism by which some xenobiotics stimulate the formation of vitamin C in animals and enhance the excretion of l-xylulose in humans with pentosuria and have shown that aminopyrine, metyrapone and other xenobiotics cause an almost instantaneous increase in the conversion of UDP-glucuronate to glucuronate in isolated rat hepatocytes [4]. The precise mechanism by which free glucuronate is formed remains unclear. It is usually stated that glucuronate formation from UDP- glucuronate is the result of two successive reactions comprising the hydrolysis of UDP-glucuronate to glu- curonate 1-phosphate and UMP by a pyrophospha- tase, followed by dephosphorylation of glucuronate 1-phosphate [5,6]. However, neither the pyrophospha- tase nor the phosphatase implicated in these reactions has been identified. Furthermore, other mechanisms, in which glucuronate is directly formed by hydrolysis of UDP-glucuronate or indirectly through the transfer of glucuronide to an endogenous (unknown) acceptor by a UDP-glucuronosyltransferase, followed by the hydrolysis of the glucuronidated acceptor, need to be considered [4,7,8]. The purpose of this study was to check if the effect of aminopyrine, metyrapone and chloretone to stimu- late the formation of glucuronate from UDP-glucuro- nate could be reproduced in cell-free systems and to progress in the identification of the enzyme(s) implica- ted in this conversion. Results Glucuronate and glucuronate 1-phosphate formation in crude liver extracts Our first attempts were aimed at identifying conditions under which aminopyrine, metyrapone and chloretone stimulated the formation of glucuronate from UDP- glucuronate in crude liver extracts. These experiments UDP-D-glucuronate D-glucuronate-1-P + UMP (-) ATP UDP- D-GlcNAc UDP- D-glucose Plasma membrane UDP-D-glucuronate D-glucuronate + UDP UDP- D-GlcNAc (+) D-glucuronate L-gulonate L-gulono-1,4-lactone L-gulono-1,4-lactone L-ascorbate 3-dehydro-L-gulonate L-xylulose ATP-Mg + CoASH (-) Aglycones (-) Metyrapone Aminopyrine Chloretone (-) ER cytosol (8) (1) Sorbinil (-) (2) (5) (6) (3) (4) xylitol (7) Pentosuria Fig. 1. Pathways of vitamin C, L-xylulose and glucuronate 1-phosphate formation. 1, UDP-glucuronidase; 2, glucuronate reductase; 3, aldono- lactonase; 4, L-gulono-1,4-lactone oxidase; 5, L-gulonate 3-dehydrogenase; 6, 3-dehydro-L-gulonate decarboxylase; 7, L-xylulose reductase; 8, nucleotide pyrophosphatase. As shown in this study (see Discussion), glucuronate appears to be formed directly from UDP-glucuronate by a membrane-bound enzyme in the endoplasmic reticulum (ER). Metyrapone, aminopyrine and chloretone stimulate this formation by antagon- izing the inhibitory effect exerted, presumably indirectly, by a combination of ATP-Mg and CoASH. C. L. Linster and E. Van Schaftingen Glucuronate formation in liver cell-free systems FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS 1517 were performed in the presence of sorbinil, an inhibitor of aldose reductase and aldehyde reductase [9], to block the conversion of glucuronate to l-gulonate, and in the presence of UDP-N-acetylglucosamine (UDPGlcNAc), which stimulates glucuronate forma- tion (see below). As shown in Fig. 2, xenobiotics had no effect on the formations of glucuronate and glucur- onate 1-phosphate in extracts that were not supple- mented with ATP-Mg. ATP-Mg inhibited the formation of free glucuronate and, more powerfully, that of glucuronate 1-phosphate, but the first effect was counteracted by xenobiotics, whereas the second was not, suggesting that glucuronate formation was independent of glucuronate 1-phosphate formation. In the presence of ATP-Mg, the rate of hydrolysis of 0.5 mm glucuronate 1-phosphate amounted to 0.04 nmolÆmin )1 Æmg )1 protein irrespective of the pres- ence or absence of xenobiotics (not shown), therefore being much lower than the rate of glucuronate formation from UDP-glucuronate in the presence of xenobiotics (0.2 nmolÆmin )1 Æmg )1 protein). Even lower activities were observed at concentrations of glucuro- nate 1-phosphate < 0.5 mm, indicating that the glu- curonate 1-phosphate phosphatase activity was not underestimated because of substrate inhibition. These results further argued against glucuronate 1-phosphate being an intermediate in the formation of glucuronate from UDP-glucuronate (see Discussion). Localization of the enzyme forming glucuronate in microsomes Liver extract fractionation showed that the enzyme responsible for glucuronate formation from UDP-glu- curonate (henceforth called ‘UDP-glucuronidase’) was mainly present in the microsomal fraction (Table 1), as were UDP-glucuronosyltransferase and UDP-glucuro- nate pyrophosphatase. Interestingly, metyrapone sti- mulated UDP-glucuronidase activity in the microsomal fraction by only 20%, despite the presence of ATP- Mg. It is shown below that this is due to loss of the inhibitory effect of ATP-Mg, consequent to the removal of a heat-stable cofactor present in the high- speed supernatant. Accordingly, the total recovery of UDP-glucuronidase activity in the mitochondrial and microsomal fractions was much higher than 100% if metyrapone was omitted (first column of Table 1), but was close to 100% if the assays were performed in the presence of this xenobiotic. The microsomal fraction contained only minimal glucuronate 1-phosphatase activity (0.09 nmolÆmin )1 Æmg )1 protein, i.e.  10% of the UDP-glucuronidase activity in the same fraction). This activity was not modified in the presence of 0.1% Triton X-100. UDP-glucuronate is used in the lumen of the endo- plasmic reticulum (ER) by UDP-glucuronosyltrans- ferase [10] and its transport into this organelle appears to be stimulated by UDPGlcNAc, explaining the sti- mulation that this nucleotide exerts on glucuronidation [11]. To test whether the enzymes catalysing the forma- tion of glucuronate and glucuronate 1-phosphate were present in the lumen of the ER (or had their catalytic site oriented towards the lumen of this organelle), we checked the effect of UDPGlcNAc on their activity. As shown in Fig. 3, UDPGlcNAc exerted a marked stimulatory effect on UDP-glucuronidase, similar to that observed for UDP-glucuronosyltransferase, but did not stimulate the formation of glucuronate 1-phos- phate. This indicated that UDP-glucuronate must cross the ER membrane to reach the catalytic site of UDP-glucuronidase, but not of UDP-glucuronate pyrophosphatase. As a matter of fact, UDPGlcNAc Fig. 2. Effect of metyrapone, aminopyrine and chloretone on the formation of free glucuronate and glucuronate 1-phosphate in crude liver extracts incubated in the absence (A, B) or presence (C, D) of ATP-Mg. Crude liver extracts were incubated with 1 m M UDP-glu- curonate, 1 m M UDPGlcNAc, 0.5 mM sorbinil, without or with 10 m M ATP-Mg and ⁄ or 1 mM of the indicated xenobiotic (open diamonds, no xenobiotic added; filled triangles, aminopyrine; filled circles, chloretone; filled squares, metyrapone). A control incubation containing 0.5% dimethylsulfoxide (solvent for chloretone) was also performed (open circles). When incubations were run without liver extract, no glucuronate, but 6.7 ± 0.6 l M (mean ± SEM, n ¼ 12) glucuronate 1-phosphate, resulting from acid hydrolysis of UDP- glucuronate, was measured. This value was subtracted from those found in the presence of liver extract. Note that the scale of the ordinate in Fig. 2B differs from the other panels by sixfold. Glucuronate formation in liver cell-free systems C. L. Linster and E. Van Schaftingen 1518 FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS and ATP-Mg inhibited UDP-glucuronate pyrophos- phatase, 50% inhibition being reached at  4 and 0.5 mm, respectively (Fig. 3 and not shown). By con- trast, ATP-Mg did not affect UDP-glucuronidase activity in the microsomal fraction, although, as shown above, it did inhibit this activity in crude extracts. Implication of UDP-glucuronosyltransferases in the formation of free glucuronate Because they are located in the same subcellular com- partment and use the same nucleotide substrate, it was of interest to compare the properties of UDP-glucu- ronidase and UDP-glucuronosyltransferases. The latter are sensitive to several detergents [12,13], because they are integral membrane proteins [14,15]. We therefore compared the effect of various detergents on free glu- curonate formation and glucuronidation of 4-nitrophe- nol. All incubations were performed in the presence of UDPGlcNAc to stimulate the entry of UDP-glucuro- nate into undisrupted microsomes. The four tested detergents had similar effects on both activities: stimu- lation was observed with low concentrations and, except for polyoxyethylene ether W-1, inhibition was observed at higher concentrations with the following order of potency: deoxycholate > b-octylglucoside > Triton X-100 (Fig. 4). By contrast, 0.5% deoxycholate and 1.8% b-octylglucoside, which both completely inhibited UDP-glucuronidase and UDP-glucuronosyl- transferase activities, only slightly affected the activity of glucose-6-phosphatase (10 and 20% inhibition, respectively), another integral membrane protein of the ER [16]. To determine whether UDP-glucuronosyltransferas- es are directly implicated in the formation of free Table 1. Subcellular distributions of UDP-glucuronidase, 4-Np-UGT and UDP-glucuronate pyrophosphatase. UDP-glucuronidase was assayed at 37 °C in the presence of 0.5 m M sorbinil without or with 1 mM metyrapone. ATP-Mg was omitted from the UDP-glucuronate pyrophos- phatase assay, also performed at 37 °C. 4-Np-UGT was assayed with 0.2 m M 4-nitrophenol and the assay was started by addition of the enzyme preparation. Results are means ± SEM for three experiments or individual values obtained in two independent experiments. UDP-glucuronidase 4-Np-UGT UDP-glucuronate pyrophosphataseNo metyrapone 1 m M metyrapone Specific activity (nmolÆmin )1 Æmg )1 protein) Heavy mitochondrial fraction 0.05 ± 0.01 0.08 ± 0.01 0.27, 0.25 2.76, 2.70 Light mitochondrial fraction 0.18 ± 0.01 0.23 ± 0.01 0.90, 0.78 4.93, 4.96 Microsomal fraction 0.87 ± 0.02 1.02 ± 0.01 3.07, 2.88 12.2, 11.8 Final supernatant 0.01 ± 0.00 0.01 ± 0.00 0.00, 0.04 0.71, 0.77 Total activity (nmolÆmin )1 Æg )1 liver) Post-nuclear supernatant 14.3 ± 0.8 31.3 ± 0.3 74.3, 73.8 354.9, 364.2 Sum of fractions 25.4 ± 1.0 30.4 ± 0.8 89.2, 86.8 426.3, 420.9 Yield (%) 179 ± 4 97 ± 2 120, 118 120, 116 Fig. 3. Stimulation of glucuronate and b-glucuronide formation (A) and inhibition of glucuronate 1-phosphate formation (B) by UDPGlc- NAc in microsomes. Microsomes were incubated at 30 °C with 1m M UDP-glucuronate, the indicated concentrations of UDPGlcNAc and without (open symbols) or with (filled symbols) 10 m M ATP- Mg. For the assay of 4-nitrophenylglucuronoslytransferase (4-Np-UGT), the medium additionally contained 0.2 m M 4-nitrophe- nol and 1 m M saccharo-1,4-lactone. The reactions were initiated by the addition of microsomes. Perchloric acid extracts were prepared after 8 min to measure b-glucuronide (triangles) and after 20 min to measure glucuronate (squares) and glucuronate 1-phosphate (dia- monds). Glucuronate 1-phosphate formation from UDP-glucuronate was corrected for acid hydrolysis as in Fig. 2. UDPGAse, UDP- glucuronidase. C. L. Linster and E. Van Schaftingen Glucuronate formation in liver cell-free systems FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS 1519 glucuronate, we tested the effect of glucuronidation substrates on UDP-glucuronidase activity. These experiments were performed in the presence of ATP- Mg, to inhibit UDP-glucuronate breakdown by the pyrophosphatase, saccharo-1,4-lactone, an inhibitor of b-glucuronidase [17], to block hydrolysis of the b-glu- curonides and 0.1% Triton X-100, to prevent any limi- tation in UDP-glucuronate supply due to saturation of a transport mechanism. As shown in Fig. 5, 4-methyl- umbelliferone and valproate both dose-dependently inhibited the formation of free glucuronate. Remark- ably, the effect of 4-methylumbelliferone disappeared after it had been completely glucuronidated (Fig. 6), indicating that inhibition was truly due to the presence of this substrate of glucuronidation. No such decrease in the inhibition was observed with time in the case of valproate, which was more slowly metabolized. Inhibi- tion of glucuronate formation was also observed with other substrates of glucuronidation including resorci- nol, 4-nitrophenol and chloramphenicol (not shown). A potential explanation for the involvement of UDP-glucuronosyltransferases in the formation of free glucuronate could be a glucuronidation–deglucuroni- dation cycle involving an unknown glucuronidated intermediate. The latter would be hydrolysed by b- glucuronidase or possibly by esterases, in which case it would be an acylglucuronide. However, saccharo-1,4- lactone (3 mm) did not affect glucuronate formation from UDP-glucuronate in microsomes, whereas it powerfully inhibited b-glucuronidase in this subcellular fraction. Fifty per cent inhibition was observed at pH 7.1 with 10–15 lm saccharo-1,4-lactone when 0.5 mm 4-nitrophenylglucuronide or 0.5 mm 4-methyl- umbelliferylglucuronide were used as substrates (not shown). Similarly, preincubation of microsomes with 1mm bis-p-nitrophenylphosphate, an esterase inhibitor [18], for 30 min at 37 °C did not affect their UDP- glucuronidase activity, whereas it suppressed their capacity to hydrolyse 3 mm o-nitrophenylacetate (not shown). Fig. 4. Effect of various detergents on glucuronate (A, C) and b-glucuronide (B, D) formation. Microsomes were incubated at 30 °Casdes- cribed in Experimental procedures, but without ATP-Mg. UDP-glucuronate and UDPGlcNAc, as well as the indicated concentrations of the various detergents (squares, Triton X-100; circles, b-octylglucoside; diamonds, polyoxyethylene ether W-1; triangles, deoxycholate) were included in the assays. UDP-glucuronidase (UDPGAse) was measured in the presence of 1 m M metyrapone and 4-Np-UGT in the presence of 0.2 m M 4-nitrophenol and 1 mM saccharo-1,4-lactone. The reactions were initiated by addition of microsomes. Perchloric acid extracts were prepared after 8 and 20 min to measure b-glucuronide and glucuronate, respectively. PE W-1, polyoxyethylene ether W-1. Glucuronate formation in liver cell-free systems C. L. Linster and E. Van Schaftingen 1520 FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS Role of a heat-stable cofactor in the sensitivity of UDP-glucuronidase to metyrapone and other xenobiotics The data obtained with purified microsomes suggested that a cofactor required for inhibition of UDP-glucu- ronidase by ATP-Mg had been lost during the prepar- ation of this subcellular fraction. Accordingly, addition of a liver high-speed supernatant inhibited microsomal UDP-glucuronidase in the presence of ATP-Mg (Fig. 7A). This inhibition was much less important in the presence of metyrapone. Similar results were obtained with a high-speed supernatant that had been heated for 5 min at 95 °C, indicating that the cofactor was heat stable. This heat-stable cofactor was dependent on ATP-Mg for its action and the inhibition that it exer- ted together with ATP-Mg was antagonized by metyra- pone, aminopyrine and chloretone (Fig. 7B,C). Further characterization of the cofactor indicated that it was retained on charcoal (Fig. 7A) and on the anion-exchan- ger Q-Sepharose (not shown). No inhibitor was appar- ently eluted from the column by applying a salt gradient. However, incubation of the eluted fractions for 90 min with 5 mm dithiothreitol at 25 °C allowed us to recover  15% of the initial inhibitory activity in the fraction eluted with 500 mm NaCl. As this inhibitory fraction contained CoASH, we tested the effect of this nucleotide on glucuronate formation. Like the heat-sta- ble cofactor, CoASH inhibited free glucuronate forma- tion in an ATP-dependent manner and its inhibitory effect was antagonized by metyrapone, aminopyrine and chloretone (Fig. 8). The effect of CoASH was half- maximal at 30 lm. Discussion Lack of involvement of glucuronate 1-phosphate in glucuronate formation Previous results obtained with isolated hepatocytes have indicated that free glucuronate formation is Fig. 5. Effect of 4-methylumbelliferone (4-MU) and valproate on the formation of free glucuronate (A) and the rate of their glucuronid- ation (B). Microsomes were incubated at 30 °C with 3 m M UDP-glu- curonate, 0.1% Triton X-100, 10 m M ATP-Mg, 1 mM saccharo-1, 4-lactone, 1 m M metyrapone and the indicated concentrations of 4-methylumbelliferone (squares) or valproate (triangles). The react- ions were initiated by addition of UDP-glucuronate after 10 min pre- incubation. Perchloric acid extracts were prepared 10 min later to measure glucuronate and b-glucuronides. Fig. 6. Transience of the inhibitory effect of 4-methylumbelliferone (4-MU) but not of valproate on the formation of free glucuronate. Microsomes were incubated in the same conditions as for Fig. 5 but without (open triangles) or with a fixed concentration of valpro- ate (1 m M; closed triangles) or 4-methylumbelliferone (0.5 mM; closed squares). Perchloric acid extracts were prepared at various times after the addition of UDP-glucuronate to determine glucuro- nate (A) and b-glucuronide (B) concentrations. A control incubation containing 1% dimethylsulfoxide (solvent for 4-methylumbelliferone) was also performed (open squares). The dashed line represents an extrapolation of the initial rate of glucuronate formation in the pres- ence of 4-methylumbelliferone over the whole incubation period. C. L. Linster and E. Van Schaftingen Glucuronate formation in liver cell-free systems FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS 1521 rapidly stimulated by aminopyrine, metyrapone and other xenobiotics, and that this formation takes place at the expense of UDP-glucuronate [4]. We were able to reproduce this effect in liver extracts enriched with ATP-Mg (see below). However, although metyrapone and other xenobiotics stimulated the formation of glu- curonate in these preparations, they did not affect the formation of glucuronate 1-phosphate. This indicates that glucuronate 1-phosphate is not an intermediate in the formation of glucuronate. If it were, its concentra- tion would either increase or decrease, depending on whether the stimulation by xenobiotics was exerted on its formation or its hydrolysis. Other observations further argue against glucuro- nate 1-phosphate being an intermediate in glucuronate formation. First, is the finding that the rate of glucuro- nate 1-phosphate hydrolysis is several-fold slower than the rate of glucuronate formation from UDP-glucuro- nate under various conditions (e.g. in liver extracts incubated in the presence of ATP-Mg and xenobiotics; in microsomes). Second, is the finding that glucuronate A B C Fig. 7. Requirement of a heat-stable cofactor for the effect of metyrapone (MP) and other xenobiotics on microsomal UDP-glucu- ronidase. Microsomes (ms, 2.5 mg proteinÆmL )1 )and⁄ or a high- speed supernatant (HSS, 12.2 mg proteinÆmL )1 ) were incubated in the same conditions as the crude liver extracts in Fig. 2. The effect of a high-speed supernatant (untreated, heated for 5 min at 95 °C or heated and subsequently treated with 2% charcoal) on micro- somal glucuronate formation was tested in the presence of 10 m M ATP-Mg and in the absence (black bars) or presence (grey bars) of 1m M metyrapone (A). The effect of the heated high-speed super- natant was further analysed in the absence (B) or presence (C) of 10 m M ATP-Mg and in the absence (black bars) or presence of 1m M metyrapone (light grey bars), aminopyrine (AP, white bars) or chloretone (CL, dark grey bars). Perchloric acid extracts were pre- pared 0 and 20 min after initiation of the reaction by addition of UDP-glucuronate and UDPGlcNAc to measure glucuronate. The dif- ference between the concentrations determined at 0 and 20 min of incubation is shown. A B Fig. 8. ATP-dependent inhibition of free glucuronate formation by CoASH. Microsomes were incubated in the same conditions as the crude liver extracts in Fig. 2 except that sorbinil was omitted from the incubation medium. Glucuronate formation was measured with- out (light grey bars) or with (dark grey bars) 100 l M CoASH and in the absence (A) or presence (B) of 10 m M ATP-Mg. The effect of 1m M metyrapone (MP), aminopyrine (AP) or chloretone (CL) on glucuronate formation in the presence of CoASH was also tested. Perchloric acid extracts were prepared 20 min after initiation of the reaction by addition of UDP-glucuronate and UDPGlcNAc to meas- ure glucuronate. Glucuronate formation in liver cell-free systems C. L. Linster and E. Van Schaftingen 1522 FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS 1-phosphate formation in microsomes is profoundly inhibited by ATP-Mg, whereas glucuronate formation, under the same conditions, is unaffected by this nuc- leotide (Fig. 3). Furthermore, low concentrations of UDPGlcNAc stimulate the formation of glucuronate although not that of glucuronate 1-phosphate in microsomes. These data indicate, therefore, that UDP-glucuro- nate hydrolysis to glucuronate 1-phosphate is unre- lated to free glucuronate formation. The enzyme that forms glucuronate 1-phosphate from UDP-glucuronate most likely corresponds to nucleotide pyrophosphatase (Fig. 1). This enzyme, which is mainly present on the outer face of the plasma membrane, hydrolyses a series of nucleotide diphosphate sugars, as well as triphos- phonucleotides [19–23]. The finding that ATP-Mg and UDPGlcNAc (Fig. 3), as well as UDP-glucose (not shown), inhibit the formation of glucuronate 1-phos- phate supports this interpretation. It is therefore likely that nucleotide pyrophosphatase does not serve physi- ologically to hydrolyse UDP-glucuronate, because it is not present in the same compartment as this potential substrate. Similarly, the low glucuronate 1-phosphate phosphatase activity detected in liver extracts and microsomes most likely corresponds to a nonspecific phosphatase. Lack of involvement of a glucuronidated intermediate The enzyme forming free glucuronate from UDP-glu- curonate shares several properties with UDP-glucuron- osyltransferases (see below). Because liver microsomes contain b-glucuronidase [24–26], the formation of free glucuronate from UDP-glucuronate observed in this preparation could be the result of a glucuronidation– deglucuronidation cycle, with a hypothetical acceptor present in the microsomal fraction. Against this is the finding that saccharo-1,4-lactone did not affect the for- mation of glucuronate despite completely blocking hydrolysis of 4-nitroph enyl- and 4-methylumbelliferyl- glucuronide. As esterases are also present in micro- somes [27], and some UDP-glucuronosyltransferases use carboxylic acids as acceptors [28], we had to consi- der the possibility that an acylglucuronide could form as an intermediate. The finding that bis-p-nitrophenyl- phosphate, although blocking esterase activity, did not affect the formation of glucuronate from UDP-glu- curonate allowed us to discard this second possibility. Although we may not formally exclude that glucuro- nate formation involves the hydrolysis of a hypothet- ical glucuronidated intermediate by an unknown enzyme that would not be affected by these inhibitors, our observations indicate that UDP-glucuronate is directly hydrolysed to glucuronate and UDP, i.e. that glucuronate formation is catalysed by a UDP-glucuro- nate glucuronyl hydrolase, which we designate UDP- glucuronidase for the sake of simplicity. Probable identity of UDP-glucuronidase and UDP-glucuronosyltransferases To date, very few enzymes have been described that hydrolyse a nucleotide diphosphate sugar to a free sugar and a nucleotide diphosphate. A well-characterized example of this type of reaction is the one catalysed by GDP-mannose hydrolase, an enzyme that was initially characterized in Escherichia coli [29], but whose physio- logical function is not known. Like other members of the Nudix family, GDP-mannose hydrolase is a soluble protein, and therefore very different from liver UDP- glucuronidase, a membrane-bound enzyme. UDP-glucuronidase shares several properties with UDP-glucuronosyltransferases. Both enzymes are pre- sent in liver microsomes and are stimulated by UD- PGlcNAc. The stimulatory effect of UDPGlcNAc on UDP-glucuronosyltransferases depends on the integrity of the microsomal membrane [30] and has been attrib- uted to the ability of this nucleotide to stimulate UDP- glucuronate influx into microsomes [11]. This involves conversion of UDP-glucuronate to UMP in the lumen of the vesicles and exchange of the latter with cytosolic UDPGlcNAc. Once inside microsomes, the latter can, in turn, be exchanged with cytosolic UDP-glucuronate thanks to a UDP-glucuronate–UDPGlcNAc antiport. Stimulation of UDP-glucuronidase by UDPGlcNAc indicates that this enzyme is present in the same type of vesicles as UDP-glucuronosyltransferases. Further analogy between the two types of enzymes is found in the similarity of the effect of detergents. All tested detergents stimulated both enzymatic activit- ies at low concentrations, consistent with the idea that both types of enzymes have their catalytic site oriented towards the lumen of the vesicles and that disruption of the vesicular membrane increases accessibility to UDP-glucuronate. Some of the detergents exerted inhi- bition of the enzymatic activity at higher concentra- tions and it is striking that the same order of potency (deoxycholate > b-octylglucoside > Triton X-100) was observed for UDP-glucuronosyltransferases and for UDP-glucuronidase. This indicates that their activity has the same type of requirement in terms of phos- pholipidic environment. That the UDP-glucuronidase activity may actually be a side activity of UDP-glucuronosyltransferases themselves is suggested by the fact that glucuronidable C. L. Linster and E. Van Schaftingen Glucuronate formation in liver cell-free systems FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS 1523 substrates (4-methylumbelliferone, valproate) inhibited formation of free glucuronate. 4-Methylumbelliferone was more potent than valproate as an inhibitor of glu- curonate formation consistent with the former being a substrate for many UDP-glucuronosyltransferase iso- forms [31], which is not the case for valproate [32]. Taken together, these findings indicate that UDP- glucuronosyltransferase (or at least some UDP-glucu- ronosyltransferase isoforms) may actually catalyse not only the transfer of a glucuronosyl group to an accep- tor, but also the hydrolysis of the glycosidic linkage in UDP-glucuronate. From the data shown in Fig. 5 this reaction would be substantial, amounting to  7% of the rate of glucuronidation of 4-methylumbelliferone, one of the best substrates for glucuronidation. The involvement of UDP-glucuronosyltransferases in glu- curonate formation is consistent with the finding that 3-methylcholanthrene (an inducer of UDP-glucurono- syltransferases of the UGT1A family) stimulates vita- min C formation in normal rats, although not in Gunn rats [8], in which all UGT1A isoforms are defici- ent [33,34]. However, Gunn rats produce vitamin C, which, if our hypothesis is correct, would mean that UGT2 family isozymes may also be involved in the formation of glucuronate. Interestingly, vitamin C for- mation is induced in Gunn rats by phenobarbital [8], an inducer of UGT2s, which is indirect evidence for the involvement of members of the UGT2 family. To the best of our knowledge, very few studies on purified UDP-glucuronosyltransferases have investi- gated the capacity of these enzymes to hydrolyse UDP-glucuronate to UDP and glucuronate. A UDP- glucuronosyltransferase purified from pig liver (GT 2P ) was shown to hydrolyse UDP-glucuronate to free glu- curonate and UDP at a rate corresponding to  0.001% of its activity as a transferase [35]. This ‘a-glucuronidase’ activity was enhanced by the pres- ence of phenylethers and lysophosphatidylcholines up to  0.03% of its transferase activity. This value is much lower than that observed in this study for a non- purified enzyme, indicating that if indeed free glucuro- nate production is due to an a-glucuronidase activity of UDP-glucuronosyltransferases, the hydrolytic activ- ity must be stimulated by phospholipids or other com- pounds present in the microsomal membrane. Another possibility is that the a-glucuronidase activity may be more substantial in the case of some UDP-glucurono- syltransferases than others, or that one or several members of the UDP-glucuronosyltransferase family only act as hydrolases. Our conclusions on the involvement of UDP-glucu- ronosyltransferases in the formation of glucuronate are only tentative at this stage. Purification attempts involving solubilization of UDP-glucuronidase with detergents followed by chromatography (CL Linster & E Van Schaftingen, unpublished results) failed because the UDP-glucuronidase activity was inhibited by the detergents or because the detergents were unable to solubilize the enzyme properly. Ongoing experiments with overexpressed UGT1A6 in HEK cells (CL Lin- ster, CP Strassburg & E Van Schaftingen, unpublished results) indicate that this enzyme has modest UDP-glu- curonidase activity that is stimulated by menadione (a stimulator of glucuronate formation in isolated hepato- cytes) [4] and lysophosphatidylcholine (reported to be a stimulator of the UDP-glucuronidase activity of ‘GT 2P ’ [35]). Under the ‘best’ conditions, the UDP- glucuronidase activity amounted to  0.4% of the UDP-glucuronosyltransferase activity, which is  10- fold higher than the highest ratios observed by Hochman & Zakim, but which is still far from the  7% UDP-glucuronidase ⁄ UDP-glucuronosyltrans- ferase activity described for intact liver microsomes (this study). Further work is needed to identify the UGT isozymes and potential cofactor(s) involved in free glucuronate formation. Conditions required to observe the effect of xenobiotics in cell-free systems The stimulation exerted by several xenobiotics on vitamin C formation has recently been attributed to a rapid effect of these agents to stimulate glucuronate formation in intact liver cells [4]. We were able to reproduce the stimulation of glucuronate formation in liver extracts and microsomes. With the first type of preparation we noticed that ATP-Mg behaved as an inhibitor of the UDP-glucuronidase activity, and that metyrapone, aminopyrine and chloretone could then show a ‘stimulatory effect’ (a deinhibitory effect in fact) of about the same order of magnitude as in intact hepatocytes. This inhibitory effect of ATP-Mg was no longer present in washed liver microsomes, but could be restored in this last preparation by add- ing a heated liver high-speed supernatant or low (physiological) concentrations of CoASH. Both the heated liver extract and CoASH also restored (in the presence of ATP-Mg) the sensitivity to metyrapone and other xenobiotics and it is likely that the effect of the heated liver extract can be entirely ascribed to CoASH or CoA derivatives. This identification may, for instance, account for the loss of inhibitor upon anion-exchange chromatography of the heated high- speed supernatant and its partial recovery upon treatment of the fractions with dithiothreitol, as CoASH was found to be largely oxidized during this Glucuronate formation in liver cell-free systems C. L. Linster and E. Van Schaftingen 1524 FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS purification procedure. The finding that the effect of CoASH depends on the presence of ATP (although not of other NTPs such as GTP and UTP; not shown) suggests that it is indirectly mediated via the formation of acyl-CoAs from fatty acids present in the microsomal preparation by microsomal acyl-CoA synthetase. Interestingly, acyl-CoAs are known to inhibit UDP-glucuronosyltransferases [36]. Our conclusions on glucuronate formation and its regulation are summarized in Fig. 1. We have provided evidence for the fact that glucuronate formation in liver appears to proceed through direct hydrolysis of UDP-glucuronate rather than via an intermediate, and that UDP-glucuronosyltransferase or a closely related enzyme seems to be involved in this conversion. How- ever, the enzyme responsible for the synthesis of glu- curonate 1-phosphate from UDP-glucuronate remains a pending problem that needs further research. The identification of conditions that allow one to observe the stimulation of glucuronate formation by xenobiot- ics in cell-free systems is an important step towards the identification of the detailed mechanisms by which these compounds act and of the enzyme implicated in glucuronate formation. Experimental procedures Materials Glucuronate 1-phosphate was prepared by incubating 80 mm UDP-glucuronate in the presence of 3.3% perchloric acid (w ⁄ v) at 50 °C for 1 h in a total volume of 0.6 mL. After neutralization with K 2 CO 3 and elimination of the resulting salt precipitate by centrifugation, the preparation was treated twice with 5% charcoal in the presence of 25 mm Hepes, pH 7.1 to eliminate nucleosides and nucleo- tides, and centrifuged to remove charcoal. The resulting supernatant was chromatographed on a Dowex 1 · 8 resin (1 mL), from which glucuronate 1-phosphate was eluted with a NaCl gradient. Four fractions of 0.5 mL, eluted with 250–350 mm NaCl and containing between 4 and 6 mm glu- curonate 1-phosphate, were obtained in this way. These fractions did not contain any free glucuronate. E. coli b-glucuronidase, alkaline phosphatase and ATP (disodium salt) were purchased from Roche Applied Sci- ence (Mannheim, Germany). Dimethylsulfoxide, MgCl 2 , 4-nitrophenol, sodium deoxycholate and sodium phosphate were from Merck (Darmstadt, Germany). Aminopyrine, metyrapone, charcoal, saccharo-1,4-lactone, polyoxyethyl- ene ether W-1, b-octylglucoside, Triton X-100 and the sodium salts of CoASH (from yeast), UDP-glucuronic acid and UDPGlcNAc were from Sigma-Aldrich (St Louis, MO). Chloretone and Dowex 1 · 8 were from Acros Organics (Geel, Belgium). 4-Methylumbelliferone was from Koch-Light (Colnbrook, UK) and sodium valproate from Labaz-Sanofi (Brussels, Belgium). Sorbinil was a kind gift of Pfizer. All other reagents, whenever possible, were of analytical grade. Preparation of crude liver extracts, microsomes and other subcellular fractions All steps of the described procedures were carried out at 4 °C. Liver extracts were prepared from overnight-fasted male Wistar rats. Livers were homogenized in a Potter-Elv- ehjem apparatus with 3 vol. (v ⁄ w) of a buffer containing 25 mm Hepes, pH 7.1, 25 mm KCl, 0.25 m sucrose, 5 lgÆmL )1 antipain and 5 lgÆmL )1 leupeptin. The homogen- ate was centrifuged for 20 min at 18 000 g. The resulting supernatant (crude liver extract) was centrifuged for another 45 min at 100 000 g to obtain a high-speed supernatant and a microsomal pellet. The latter was washed twice in the homogenization buffer and resuspended in the same buffer to get a microsomal preparation containing  40 mg pro- teinÆmL )1 . For subcellular fractionation (Table 1), livers from two overnight fasted male Wistar rats were homo- genized as described but in a buffer containing 10 mm Hepes, pH 7.1, 0.25 m sucrose, 2.5 lgÆmL )1 antipain and 2.5 lgÆmL )1 leupeptin. The homogenate was submitted to differential centrifugation [24]. The extracts and subcellular fractions were stored at )80 °C. Protein was measured according to Lowry et al. [37], with bovine serum albumin as a standard. Assay of enzymatic activities UDP-glucuronidase and UDP-glucuronate pyrophosphatase were assayed at 30 or 37 °C through the conversion of UDP-glucuronate to glucuronate and glucuronate 1-phos- phate, respectively. Unless otherwise stated, the assay med- ium contained 20 mm sodium phosphate, pH 7.1, 2 mm MgCl 2 ,10mm ATP-Mg, 1 mm UDPGlcNAc, 1 mm UDP- glucuronate and  3 (microsomes) or 15 (crude extracts) mg proteinÆmL )1 . In most experiments, the enzyme prepar- ation was preincubated for 10 min with all assay compo- nents except UDP-glucuronate and UDPGlcNAc, and the assay was initiated by addition of these two nucleotides. Where indicated, the assay was initiated by the addition of the enzyme preparation to an otherwise complete assay mixture. The reaction was stopped after 0–30 min by mix- ing a portion of the incubation medium with 0.5 vol. of cold 10% (w ⁄ v) perchloric acid. Glucuronate-1-phosphatase was measured through the formation of glucuronate under similar conditions, except that UDP-glucuronate was replaced by 0.5 mm glucuronate 1-phosphate. UDP-glucu- ronosyltransferase was also similarly assayed, at 30 °C, through the formation of b-glucuronides in an incubation mixture containing 20 mm sodium phosphate, pH 7.1, 2mm MgCl 2 ,10mm ATP-Mg, 1 mm saccharo-1,4-lactone, C. L. Linster and E. Van Schaftingen Glucuronate formation in liver cell-free systems FEBS Journal 273 (2006) 1516–1527 ª 2006 The Authors Journal compilation ª 2006 FEBS 1525 [...]... respectively The incubation conditions for hydrolysis by these two enzymes were the same as those described previously [38] except that, because of the presence of saccharo1,4-lactone, an inhibitor of b-glucuronidase, in the samples where b-glucuronides had to be determined, the amount of b-glucuronidase and the incubation time with the latter enzyme were doubled Unless otherwise indicated, the results shown in. .. stimulation of free glucuronate formation by non-glucuronidable xenobiotics in isolated rat hepatocytes J Biol Chem 278, 36328–36333 5 Ginsburg V, Weissbach A & Maxwell ES (1958) Formation of glucuronic acid from uridinediphosphate glucuronic acid Biochim Biophys Acta 28, 649–650 6 Puhakainen E & Hanninen O (1976) Pyrophosphatase ¨ and glucuronosyltransferase in microsomal UDPglucuronic-acid metabolism in the. .. Biochem J 52, 464–472 18 Heymann E, Mentlein R, Schmalz R, Schwabe C & Wagenmann F (1979) A method for the estimation of esterase synthesis and degradation and its application to evaluate the in uence of insulin and glucagon Eur J Biochem 102, 509–519 19 Touster O, Aronson NN Jr, Dulaney JT & Hendrickson H (1970) Isolation of rat liver plasma membranes Use of nucleotide pyrophosphatase and phosphodiesterase... (1955) Tissue fractionation studies 6 Intracellular distribution patterns of enzymes in rat -liver tissue Biochem J 60, 604–617 25 Himeno M, Nishimura Y, Tsuji H & Kato K (1976) Purification and characterization of microsomal and lysosomal b-glucuronidase from rat liver by use of immunoaffinity chromatography Eur J Biochem 70, 349–359 26 Owens JW & Stahl P (1976) Purification and characterization of rat liver. .. UDPglucuronic-acid metabolism in the rat liver Eur J Biochem 61, 165–169 1526 7 Pogell BM & Leloir LF (1961) Nucleotide activation of liver microsomal glucuronidation J Biol Chem 236, 293–298 8 Horio F, Shibata T, Makino S, Machino S, Hayashi Y, Hattori T & Yoshida A (1993) UDP glucuronosyltransferase gene expression is involved in the stimulation of ascorbic acid biosynthesis by xenobiotics in rats J Nutr 123, 2075–2084... the results shown in this study are means ± SEM for three experiments using different enzymatic preparations Acknowledgements This work was supported by the Concerted Research ´ Action Program of the Communaute Francaise de Bel¸ gique; the Interuniversity Attraction Poles Program, Belgian Science Policy; and the Fonds de la Recherche ´ Scientifique Medicale CLL is a fellow of the Fonds National de la...Glucuronate formation in liver cell-free systems C L Linster and E Van Schaftingen 1 mm UDPGlcNAc, 1 mm UDP-glucuronate and 0.2 mm 4-nitrophenol, 0.5 mm 4-methylumbelliferone or 1 mm valproate In all cases, perchloric acid extracts were centrifuged at 4 °C and the supernatants neutralized by the addition of K2CO3 These perchloric acid extracts were treated with 2% charcoal in experiments in which 4-methylumbelliferone... References 1 Smirnoff N (2001) l-Ascorbic acid biosynthesis Vitam Horm 61, 241–266 2 Hiatt HH (2001) Pentosuria In The Metabolic and Molecular Bases of Inherited Disease (Scriver CR, Beaudet AL, Sly WS & Valle D, eds), 8th edn, Vol I, pp 1589–1599 McGraw-Hill, New York 3 Wang YM & Van Eys J (1970) The enzymatic defect in essential pentosuria N Engl J Med 282, 892–896 4 Linster CL & Van Schaftingen E (2003)... S, Das B, Ansari NH & Srivastava SK (1990) Inhibition kinetics of human kidney aldose and aldehyde reductases by aldose reductase inhibitors Biochem Pharmacol 39, 1115–1124 ´ 10 Fulceri R, Banhegyi G, Gamberucci A, Giunti R, Mandl J & Benedetti A (1994) Evidence for the intraluminal positioning of p-nitrophenol UDP-glucuronosyltransferase activity in rat liver microsomal vesicles Arch Biochem Biophys... Dietrich LS (1972) The purification and properties of detergent-solubilized rat liver nucleotide pyrophosphatase J Biol Chem 247, 5071–5078 23 Bischoff E, Tran-Thi TA & Decker KFA (1975) Nucleotide pyrophosphatase of rat liver A comparative study on the enzymes solubilized and purified from plasma membrane and endoplasmic reticulum Eur J Biochem 51, 353–361 24 De Duve C, Pressman BC, Gianetto R, Wattiaux . Glucuronate, the precursor of vitamin C, is directly formed from UDP-glucuronate in liver Carole L. Linster and Emile Van Schaftingen Laboratory of. vitamin C for- mation is induced in Gunn rats by phenobarbital [8], an inducer of UGT2s, which is indirect evidence for the involvement of members of the