Abstract
β-glucuronidase cleavage of 4-methylumbelliferyl β-d-glucuronide generates the highly fluorescent compound, 4-methylumbelliferone. We show that other β-d-glucuronide compounds act as competitors in this assay. The 4-methylumbelliferyl β-d-glucuronide cleavage assay can easily be adapted to high throughput formats to detect the presence of β-D glucuronides generated using recombinant glycosyl transferase preparations.
Thousands of potentially important therapeutic agents are being generated through the use of combinatorial chemistry. Many of the candidate compounds chosen for development fail in clinical trials due to poor pharmacokinetic properties. Obtaining relevant in vitro drug metabolism data at an early stage is considered to help improve the successful selection of candidate compounds. This includes, as a first step, identifying the enzymes responsible for metabolism of a compound and then determining kinetic parameters, which, through appropriate mathematical manipulations, show utility in predicting pharmacokinetic parameters in humans (Obach et al., 1997).
UDP glycosyltransferases (UGTs)1 are a class of enzymes involved in the conjugation of glucuronic acid to a variety of drugs, xenobiotics, and endogenous steroid compounds (Burchell et al., 1995). In humans, at least 15 isoenzymes are encoded by two gene subfamilies (Mackenzie et al., 1997). Glucuronidation is known to be an important detoxification pathway that can lead not only to changes in the pharmacokinetic and pharmacodynamic properties of therapeutic agents, but also to the production of reactive metabolites (Parkinson, 1996). Traditional analytical procedures for measuring glycosyltransferase activities and metabolites, using radiolabeled substrates or cofactors with resolution by thin-layer chromatography or HPLC (Bansal and Gessner, 1980; Ethell et al., 1998), are not suited for high throughput formats. A more rapid, nonradioactive assay that can identify glucuronide formation or presence would be of great utility.
Materials and Methods
Methods are included in the text and figure legends. β-d-glucuronidase (Part G-7396), α-napthyl β-d-glucuronide, β-trifluoromethylumbelliferyl β-d-glucuronide, β-estradiol 3-(β-d-glucuronide), p-acetominophenyl β-d-glucuronide, 5β-androstane-3α,17α-diol-11-one-17-carboxilic acid-3-(β-d-glucuronide), uridine 5′-diphosphoglucuronic acid (UDPGA), and 4-methylumbelliferone β-d-glucuronide were obtained from Sigma Chemical Co. (St. Louis, MO). Tetrahydrocortisone 3-β-d-glucuronide was obtained from Molecular Probes (Eugene, OR). Recombinant control and UGT1A6 membrane preparations were generated at PanVera Corporation and are commercially available from PanVera (Madison, WI). All other reagents were analytical grade or better and purchased from a variety of commercial sources.
Results and Discussion
We sought to develop an assay that can be used to demonstrate the presence of β-d-glucuronides that might be generated from different classes of UGT enzymes. An assay based on β-glucuronidase activity provides a high throughput screening method to identify structurally different β-d-glucuronides. Cleavage of 4-methylumbelliferyl β-d-glucuronide yielded the highly fluorescent compound 4-methylumbelliferone. Under linear conditions of protein concentration and incubation time, the apparentKm value for cleavage of 4-methylumbelliferyl-β-d-glucuronide was approximately 56 μM (Fig. 1).
To show the feasibility of developing a high throughput inhibition assay, we examined the potential of a variety of structurally dissimilar commercially available β-d-glucuronides to act as inhibitors for the cleavage of 4-methylumbelliferyl-β-d-glucuronide. These β-d-glucuronides represent compounds that might potentially be formed by UGT activity, e.g., phenol and steroid glucuronides. Figure 2 shows that all of the β-glucuronides tested inhibited the production of 4-methylumbelliferone by β-glucuronidase. The potential to inhibit 4-methylumbelliferyl-β-d-glucuronide cleavage appeared to be dependent on the chemical nature of the substituted aglycone, with β-estradiol-3-(β-d-glucuronide) and α-naphthyl β-d-glucuronide showing the strongest and weakest inhibition, respectively. UDPGA, an essential cofactor for UGTs, was not an effective competitor over the range tested, probably due to the α configuration of the sugar bond to UDP (Parkinson, 1996).
To evaluate whether the fluorescent assay could be used to detect β-glucuronides generated from a biochemical assay, two approaches were taken. First, α-naphthyl β-d-glucuronide was generated in an incubation with recombinant human UGT1A6 and α-naphthol. Following extraction, 4-methylumbelliferyl-β-d-glucuronide assay reagents were added directly to the extracted residues. In a separate experiment, it was shown that aglycones are extracted with ethyl acetate and hence transferred to the β-glucuronidase reporter assay. Figure 3A shows that residues from the UGT1A6 assay but not the control microsome assay inhibited 4-methylumbelliferyl-β-d-glucuronide cleavage. Residues from the zero time point extractions do not appear to inhibit the 4-methylumbelliferone cleavage assay, which indicated that the aglycone α-naphthol does not interfere nonspecifically with this assay under these conditions. Inhibition of 4-methylumbelliferyl-β-d-glucuronide cleavage was dependent on the incubation time of the UGT1A6 assay, presumably reflecting increased α-naphthyl glucuronide accumulation. (Accumulation of α-naphthyl glucuronide in the UGT1A6 assay was independently verified in a parallel experiment using [14C]UDPGA; data not shown.) In the second approach, 4-methylumbelliferyl-β-d-glucuronide assay reagents were added directly to an α-naphthyl glucuronidation assay that was performed in a 96-well plate. Figure 3B shows that 4-methylumbelliferyl β-d-glucuronide cleavage was 60% slower when added to the UGT1A6 assay compared with the control. The homogeneous format circumvented the need for extraction but required longer incubation periods with the recombinant enzyme. These results demonstrate the following: 1) The 4-methylumbelliferyl-β-d-glucuronide assay can be used to detect β-d-glucuronides generated in biological preparations; 2) The choice of α-naphthol β-d-glucuronide, which is a weak competitive inhibitor (Fig. 2), adds credence to this assay. More potent inhibitors such as estradiol β-d-glucuronide (Fig. 2) are likely to be more effective; 3) The assay is amenable to a high throughput format as a two-step procedure; and 4) Heterologous expression systems are useful for the generation of high concentrations of β-d-glucuronides, such as α-naphthyl glucuronide, which was a weak competitor (Fig. 2), in the β-d-glucuronide cleavage assay.
Some caveats to performing this type of assay should be mentioned. The rate of cleavage is important and compared with O-linked glucuronides, N-linked glucuronides appear to be cleaved more slowly by β-d-glucuronidase (Parkinson, 1996). Also, carboxylic acid β-d-glucuronides form reactive acyl-glucuronides that could form covalent linkages to assay components (Parkinson, 1996). The use of human biological tissue preparations to generate β-glucuronides might be complicated by several factors including the presence of naturally occurring fluorescent compounds and the common use of saccharolactone to inhibit levels of endogenous β-glucuronidase. The type of β-glucuronidase and the reaction conditions (such as pH) are important (Hawes, 1998). Cleavage by β-glucuronidase may result in compounds with fluorescent properties similar to that of 4-methylumbelliferone. This problem could be solved by either adding an aglycone control or performing a second β-glucuronidase assay, in parallel with the 4-methylumbelliferyl β-d-glucuronide assay, which used a substrate that had different fluorescent emission properties, e.g., resorufin β-d-glucuronide. In addition, organic extraction was used to obtain glucuronides from the UGT assay in the experiment described in Fig. 3A. Extraction can be avoided, as shown in Fig 3B. However, using a solid-phase extraction method in a 96-well format or centrifugation to remove microsomal membranes would also be more consistent for high throughput assays. Also, replacing 4-methylumbelliferone β-d-glucuronide with a lower affinity substrate in the β-glucuronidase assay may improve the potential of the assay for a high throughput format.
In summary, we demonstrated the feasibility of using a fluorescent assay that can be adapted to a high throughput format to measure the presence of β-d-glucuronides. The assay described here uses relatively inexpensive reagents and does not require chromatographic resolution (thin-layer chromatography or HPLC) or radioactivity. The assay can be performed in a homogeneous format and could be easily automated. The assay will not eliminate the usefulness of other analytical procedures needed to determine the kinetic parameters of glucuronidation, but shows potential as a rapid preliminary screening method similar to that developed for cytochrome P-450 (Crespi et al., 1997). In addition, the principles and procedures described here could be applied to other important classes of phase II conjugating enzymes. For example, sulfate conjugates from sulfonation assays would be expected to compete in a similarly designed assay that used a sulfatase to generate a fluorescent product.
Footnotes
-
Send reprint requests to: Peter M. Shaw, Bristol-Myers Squibb, New Jersey, P.O. Box 5400, Princeton, NJ 08543-5300. E- mail:shawp{at}bms.com; and Olga V. Trubetskoy, PanVera Corporation, 545 Science Dr., Madison, WI 53711. E-mail: olgat{at}panvera.com
- Abbreviations used are::
- UDPGA
- uridine 5′-diphosphoglucuronic acid, UGT
- UDP-glycosyltransferase.
- Received September 3, 1998.
- Accepted January 29, 1999.
- The American Society for Pharmacology and Experimental Therapeutics