Introduction

Top Abstract Introduction Materials and Methods Results Discussion/Conclusions References

The uridine 5'-diphosphate glucuronosyl transferases (UDPGTs)¹ are members of the UDP glycosyltransferase superfamily. In humans, they are extensively involved in phase II metabolism, where they conjugate glucuronic acid to xenobiotics and endobiotics (Mackenzie et al., 1997). This is generally considered to be a detoxication reaction producing metabolites that are more readily eliminated. In some cases, however, it is possible to produce compounds with more potent activity than the parent analog, such as morphine-6-glucuronide (Meech and Mackenzie, 1997). Conjugates may also be immunoreactive, teratogenic, or carcinogenic; for example, ketoprofen acylglucuronide, all-trans-retinoyl--D-glucuronide, and the glucuronides of N-hydroxy arylamines, respectively (Babu et al., 1995; Nau et al., 1996; Terrier et al., 1999). Thus, studying the inherent capacity of an organ for glucuronidation and the glucuronidation characteristics of specific compounds is significant to many researchers in diverse fields.

Previously, UDPGT activity has been measured in microsomal fractions of tissues using radiolabeled substrates such as para-nitrophenol (Kyecombe and Franklin, 1973) or with HPLC (Ethell et al., 1998), both of which involve considerable time and cost. A fluorometric assay using the substrate 4-methylumbelliferone (4MU) has also been available since 1974 (Aitio, 1974a). Although a comprehensive examination of the isoform-specific metabolism of 4MU has not been carried out, it is considered to be suitable for a general screen because it has been shown to be metabolized by many isoforms. These include 1A1, 1A6, 2B7, 2B8 (now called 2B15) (Burchell et al., 1995); 1A7, 1A10 (Strassburg et al., 1998); and 1A8 (Cheng et al., 1998). However, the historical assay requires extraction and cleanup procedures and fixed incubation periods, making the assay cumbersome and time-consuming. It also suffers from a lack of sensitivity.

We have developed a modified microplate fluorometric method designed for high throughput formats, which allows the observation of changes in enzyme activity over time in a single preparation. Furthermore, the improved sensitivity of the procedure enables investigation of cell lines and tissues whose UDPGT expression was previously thought to be undetectable or absent.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion/Conclusions References

4MU was obtained from ICN Biomedicals (Aurora, Ohio). Bovine serum albumin (BSA; free fatty acid, fraction V) was obtained from Life Technologies Ltd. (Auckland, New Zealand). Uridine 5'-diphosphate glucuronic acid (UDPGA) was obtained from Roche Diagnostics (Auckland, New Zealand). Ninety-six-well "Spectraplate " microtiter plates were obtained from Whatman (London, UK). All other reagents were obtained from a variety of commercial sources and were analytical grade or better.

Preparation of Microsomes. Human liver was obtained from a healthy organ donor (female, 56 years old) under ethical approval from the New Zealand Health Funding Authority (Auckland) Ethical Committee. Microsomes were prepared according to the method of Gill et al. (1995), and protein content was determined using the bicinchoninic acid method with BSA as the protein standard (Smith et al., 1985).

Preparation of Cellular S9. Human placental trophoblast-derived cell lines (JEG-3, JAr, and BeWo) and amnion-derived cell lines (AV3 and WISH) were cultured in Hams/F12 media containing 10% heat-treated fetal calf serum, 2% kanamycin, and 1% penicillin/streptomycin/glutamine at 37°C in 10-cm Petri dishes. When confluent, cells were trypsinized (0.55% trypsin) and centrifuged at 1200 rpm for 7 min. The resulting pellet was resuspended to a density of 3 × 10⁶ cells/ml in 0.1 M Tris buffer with 2 mM phenylmethylsulfonyl fluoride, pH 7.8, and sonicated for 3 min. The homogenate was centrifuged for an additional 20 min at 10,000g, and the resulting supernatant was removed and assessed for protein content before assaying for UDPGT activity.

Historical Extraction UDPGT Assay Protocol. The method used is based on the protocol of Aitio (1974a). Briefly, 50 µl of microsomal protein (10 mg/ml stock) and 400 µl of 4MU (1-1000 µM final concentration) in 0.1 M Tris-HCl buffer, containing 5 mM MgCl₂ and 0.05% BSA, pH 7.4, were added to microtubes and preincubated for 2 min. UDPGA, (50 µl, 2 mM final concentration) was added, and the tubes were vortexed and incubated for 20 min at 37°C. The final volume was 0.5 ml. The reaction was stopped with 10% trichloroacetic acid (0.5 ml) and centrifuged for 5 min at 3000 rpm. Supernatant (0.5 ml) was transferred into a clean glass test tube, 1 ml of water-saturated chloroform was added and vortexed, and the organic phase was discarded. The chloroform wash was repeated, and 0.25 ml of the aqueous phase was transferred into a clean tube. Two milliliters of 2 M glycine, pH 10.3, was added, and fluorescence was measured in a Hitachi F-2000 Fluorometer (355-nm excitation and 460-nm emission). Results were transformed to nmol/min/mg of protein using a standard curve generated with 4MU (r² = 0.98).

Microplate UDPGT Assay Protocol. A 96-well microtiter plate containing 30 µl of microsomal protein or cellular S9 (2 and 1 mg/ml stock, respectively), 105 µl of 4MU (0-1000 µM, final concentration) in 0.1 M Tris-HCl buffer containing 5 mM MgCl₂ and 0.05% BSA, pH 7.4, was placed in a Victor Wallac 1420 Multilabel Counter set to read fluorescence at 355-nm ex and 460-nm em (15-nm bandwidth). The cofactor UDPGA (15 µl, 2 mM final concentration) was added to initiate the reaction. Final volume was 150 µl. Fluorescence was measured every 2 min for 10 min. Results were transformed to nmol/min/mg of protein using a standard curve generated with 4MU (r² = 0.99).

Kinetic Evaluation and Modeling. Maximum reaction velocities (V_max) and Michaelis-Menten constants (K_m) were determined by fitting the results to a one binding site model with least-squares nonlinear regression using GraphPad Prism 3.0 Software (San Diego, CA). Best-fit one-site binding curves for the extraction method are constrained by strict criteria for convergence, wherein the fit for five consecutive iterations reduces the absolute sum of squares by less than 0.000001%. Student's t test, means, standard deviations, and coefficients of variation were calculated using Excel 7.0 (Microsoft, Seattle, WA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion/Conclusions References

Initial experiments with the microplate method using 15, 30, 60, 120, and 150 µg of microsomal protein at 0.01 to 10 mM 4MU established 60 µg as the optimal amount of protein per well. Concentrations below this gave a less linear reaction, whereas quenching of fluorescent signal was observed above 60 µg. The initial rate of reaction was investigated using 4MU (0.01-10 mM) for up to 4 h. Subsequent experiments were measured over 10 min because this time was in the linear range of reaction for all substrate concentrations (average r² coefficient at 10 min 0.995, n = 3) with a minimum acceptable number of points (six).

Four surfactants, Brij 58 (used in the extraction method), Tween 20, Triton X-100, and BSA (Trapnell et al., 1998), which act as activators for microsomal UDPGT, were investigated for ability to increase microsomal activity. BSA was chosen as the best activator (Table 1).

The addition of the -glucuronidase inhibitor D-saccharic 1,4-lactone (sacchrolactone, 3 mM) did not influence the initial rate of reaction. Furthermore, addition of sacchrolactone quenched fluorescence nonsignificantly with 200 µM 4MU, but significantly with 2 mM 4MU (P < .00001, n = 3); thus, sacchrolactone was not included in further assays.

The effect of successive freeze-thaw cycles on the UDPGT activity of human liver microsomes was investigated. For each freeze-thaw cycle, two samples in triplicate were assayed. There was no significant difference between samples or replicates within freeze-thaw cycles; however, there was a trend toward decreasing activity after the third freeze/thaw cycle. Analysis of kinetic data, performed in replicates of six on 6 days, suggested a one binding site, Michaelis-Menten curve was most appropriate. Typical kinetic parameters for the extraction method are presented (Fig. 1A); mean V_max is 19.9 ± 3.2 nmol/min/mg of protein, and mean K_m is 652.5 ± 157.4 µM. Coefficients of variation (CV) were 16 and 24%, respectively. For the microplate method (Fig. 1B), mean V_max was 36.2 ± 1.3 nmol/min/mg of protein (CV = 3.7%), significantly (P < .0001) higher than for the extraction method. The mean K_m, 175.4 ± 24.2 µM (CV = 14.5%), was significantly lower (P < .0001) than the extraction method.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Michaelis-Menten kinetic curves for the determination of UDPGT activity with the substrate 4MU using the extraction method (A) and the microplate method with six replicate curves within one assay (B).

In the placental trophoblast-derived cell lines, JEG-3, JAr, and BeWo activity (5.5, 4.1, and 2.6 nmol/min/mg of protein, respectively) was detected, whereas in the amnion-derived AV3 and WISH cell lines it was not (Fig. 2). The rate of glucuronidation at 200 µM 4MU observed in JEG-3 cells was significantly higher than that noted in BeWo cells, (P < .05) but not significantly different from that in JAr cells. Activity in JAr cells did not differ significantly from that in BeWo cells. This suggests JEG-3 is the most appropriate cell line for investigating glucuronidation in the placenta.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Mean UDPGT activity of placental cell lines. Bars are mean ± S.D., n = 3. The assays were performed using the microplate method with 200 µM 4MU concentration as detailed in the text. *P < .05 versus BeWo (t test). ND, not detectable.

	Discussion/Conclusions

Top Abstract Introduction Materials and Methods Results Discussion/Conclusions References

We have developed a simple, high throughput assay with good reproducibility and reliability. When assessed in replicates of six over 6 different days, average intra- and interassay coefficients of variation for V_max and K_m were 9 and 22%, and 8 and 35%, respectively.

The microplate assay has the advantage of observing successive kinetics rather than a fixed time point for evaluation, allowing the researcher to identify the most linear (initial) rate of reaction. In contrast, for the extraction method, a linearity study is initially performed, and all samples are then incubated for a fixed, predefined period. This may explain the observed difference in K_m and V_max between the two methods. Specifically, the lower V_max and higher K_m values observed in the extraction method may be caused by the incubation period spanning more than the initial, linear rate of reaction. Consequently this may lessen the slope of the observed reaction [when a linear relation between time (zero) and time (final) is constructed] to yield an apparently larger K_m and smaller V_max. Thus the ability to observe only the initial rate makes this protocol not only faster, but considerably more accurate than the extraction method.

Time is additionally conserved due to the high throughput format, which uses a 96-well microtiter plate and is capable of processing many samples in a single preparation, compared with the extraction method, which accommodates one replicate per cuvette. Furthermore, the absence of extraction and cleanup procedures in the microplate method also saves time, money, and reagents.

Of particular interest to researchers with limited sample stocks, this assay is able to detect activity with approximately 10% of the sample needed in the extraction method (0.06 mg of microsomal protein for the microplate method as opposed to 0.5 mg of microsomal protein used in the extraction method).

Moreover, the direct microplate assay has been shown to be more sensitive than the extraction method; it could detect UDPGT activity in human placental cell lines where the extraction method could not. This result suggests that the enzyme UDPGT may be present and active in the placenta, which is in contrast to previous studies that have variously reported no UDPGT activity (Chakraborty et al., 1972; Kyecombe and Franklin, 1973) or limited activity (no significant UDPGT activity in 24/40 placentae sampled) (Aitio, 1974b) in subcellular fractions of placental trophoblast.

In conclusion, the microplate method for quantifying UDPGT activity is faster, more reliable, simpler, cheaper, and more sensitive than the historical method. It presents an easy, effective opportunity for researchers to investigate glucuronidation in tissues and does not require chromatographic or radioactive resolution. In addition, it has potential, due to the high throughput format and rapid resolution, to be used as a general screening assay in pharmacological studies.