Introduction

Top Abstract Introduction Results Discussion References

6-N-Formylamino-12,13-dihydro-1,11-dihydroxy-13-(-D-glucopyranosyl)5H-indolo[2,3-a]pyrrolo[3,4-c]carbazole-5,7(6H)-dione (compound 1; Fig. 1) is derived from a novel indolocarbazole antibiotic isolated from an Atreptoverticillium species (Ohkubo et al., 1996). Compound 1 demonstrated strong antitumor activity in vitro and in experimental animal tumor systems (Arakawa et al., 1995, 1996) and was developed as an anticancer agent (Sasaki et al., 1995). Compound 1 acts as a potent inhibitor of topoisomerase I by inducing the formation of stable topoisomerase I-DNA cleavage complexes and also inhibits the activity of DNA polymerase  and RNA polymerase II (Yoshinari et al., 1995; Fukasawa et al., 1998). Many derivatives of compound 1 have been synthesized in attempts to obtain a more effective and less toxic compound.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Chemical structure of the indolocarbazole compounds.

We previously reported that glucuronidation is one of the major pathways for compound 1 metabolism in mice, rats, and humans (Takenaga et al., 1999a,b). In the present study, however, we found that some structurally related compounds were not glucuronidated in human liver microsomes. We anticipate that the position of the phenolic hydroxy moiety would be important for substrate specificity of glucuronidation of these compounds. Furthermore, there might be other factors that are essential in order for the substrate to interact with UDP-glucuronosyltransferase (UGT¹). Glucuronidation is an important metabolic conjugation process for many endogenous compounds and xenobiotics catalyzed by numerous isoforms of UGT (Mackenzie et al., 1997). Some substrates for each isoform have previously been reported (Green et al., 1994, 1998; Green and Tephly, 1996; Burchell et al., 1995). The essential structural features required for the substrate to interact with UGT, however, remain unclear at present.

The present article describes the structure-activity relationship in the glucuronidation of indolocarbazole analogs in human liver microsomes. UGT activities toward indolocarbazole analogs were determined by measuring the radioactivity incorporated into the substrates using [¹⁴C]UDP-glucuronic acid (UDPGA). The substrate specificity of UGT was investigated by computer-modeling analysis. The relationship between the molecular size of substrates and UGT activities are reported here.

Experimental Procedures

Materials. The indolocarbazole analogs (Fig. 1) and glucuronide of compound 1 were synthesized at Banyu Tsukuba Research Institute (Tsukuba, Japan). Acetonitrile, ethanol, trifluoroacetic acid (TFA), N,N-dimethylformamide (all HPLC grade), HCl, Tris, and MgCl₂ were purchased from Wako Pure Chemicals (Osaka, Japan). Water was purified with a Milli-Q system (Millipore Corp., Tokyo, Japan). CHAPS and UDPGA were purchased from Sigma-Aldrich (Tokyo, Japan). Hionic-Fluor was purchased from Packard (Tokyo, Japan). [¹⁴C]UDPGA (338 mCi/mmol) was purchased from ICN Biomedicals (Costa Mesa, CA). Human pooled liver microsomes, a mixture from 15 subjects, were purchased from XenoTech, LLC (Kansas City, KS). The Sf9 insect cell microsomes expressing human UGT1A1, UGT1A6, UGT1A7, and UGT1A10 by the baculovirus expression system were purchased from PanVera Corp. (Madison, WI). The AHH-1 human B lymphoblastoid cell microsomes expressing human UGT1A4 and UGT1A9 using a pHEBo vector were purchased from GENTEST (Woburn, MA).

UGT Activity Assay. The UGT activities toward compound 1 in human liver microsomes and UGT-expressing microsomes were determined by measuring the formation of glucuronide using the authentic standard. The reaction mixture (final volume, 500 µl) contained 0.1 M Tris-HCl buffer, pH 7.4, 10 mM MgCl₂, 4 mM UDPGA, 0.2 mg of microsomal protein/ml, and compound 1, unless otherwise indicated. Compound 1 was dissolved in N,N-dimethylformamide and added to the reaction mixture to make 1% of the final organic solution. For stimulating UGT activities, microsomes were preincubated with 0.5 mg of CHAPS/mg of microsomal protein at 4°C for 20 min. The reaction mixture was preincubated for 5 min at 37°C, and reactions were started by adding UDPGA. All enzymatic assays were conducted at 37°C under conditions that produced linear product formation with respect to time (20 min) and protein concentration (0.2 mg/ml). Reactions were terminated by adding equal volumes of ethanol. Proteins were removed by centrifugation at 1800g for 5 min, and aliquots of the supernatant were analyzed by HPLC.

HPLC Conditions. The formation of glucuronide from compound 1 was measured by HPLC (Hewlett Packard 1100 series; Yokogawa, Tokyo, Japan). The mobile phase consisted of 0.1% aqueous TFA and acetonitrile. A gradient scheme was used as follows: 0 to 20 min, linear gradient from 15 to 30% acetonitrile; and 20 to 25 min, 80% acetonitrile. The flow rate was 1 ml/min at 40°C. A Superiorex ODS (4.6 × 250 mm, 5 µm) column (Shiseido, Tokyo, Japan) was used for analysis, and a New Guard RP-18 (3.2 × 15 mm; Applied Biosystems, Foster City, CA) was used as a guard column. The column eluent was monitored by ultraviolet absorption at 305 nm.

Computer Modeling. Computer-modeling analyses were undertaken using WinMOPAC software (version 2.0; Fujitu, Tokyo, Japan). The molecular geometry of substrates was energy-minimized by the semiempirical molecular-orbitals calculation program (MOPAC97) using Austin model 1 Hamiltonians.

	Results

Top Abstract Introduction Results Discussion References

Glucuronidation of Indolocarbazole Analogs in Human Liver Microsomes. The UGT activities toward indolocarbazole analogs were determined using human liver microsomes (Table 1). The results obtained with compounds 1 to 10 showed that the position of the phenolic hydroxy moiety did not affect the glucuronidation activity, whereas results obtained with compounds 11 to 19 showed that the structure of the residue on the N-6 position had a strong impact on the glucuronidation activity.

Computer-Modeling Analysis. The three-dimensional structures of compounds, determined by energy minimization with semiempirical molecular orbital calculation, are shown in Fig. 2, and the values are described in Table 1. The molecular size in the direction of the x- and z-axes were not very different. The y-values, however, depended on the structure of the substitution at the N-6 position. The compounds with a bulky chain moiety at the N-6 position showed a larger diameter to the y-axis of the molecule when compared with compounds having a single chain moiety. With respect to the phenolic hydroxy moiety, the atom electron density and electrostatic potential charges calculated from the molecular orbital were not different among these compounds. The structure-activity correlation of UGT activity toward these indolocarbazole analogs is shown in Fig. 3. The compounds with a diameter larger than 14.5 Å on the y-axis were not glucuronidated.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Three-dimensional structure of compounds. The x-, y-, and z-axes represent the molecular dimensions.

View larger version (42K): [in this window] [in a new window]   Fig. 3.   Structure-Activity correlation in the glucuronidation of indolocarbazole compounds. The numbers next to the points represent the compound number. The x-, y-, and z-axes represent the molecular dimensions.

Glucuronidation of Compound 1 Catalyzed by Expressed Human UGTs. To determine which UGT isozymes contribute to the glucuronide formation of indolocarbazole compounds, compound 1 was incubated with Sf9 insect cell microsomes expressing human UGT1A1, UGT1A6, UGT1A7, and UGT1A10, or with AHH-1 human B lymphoblastoid cell microsomes expressing human UGT1A4 and UGT1A9 (Table 2). UGT1A9 and UGT1A10 catalyzed the glucuronidation of compound 1. 

	Discussion

Top Abstract Introduction Results Discussion References

In this study, we showed the possible reason for substrate specificity of UGT activity toward indolocarbazole analogs in human liver microsomes. To find the structural features required for the substrate to interact with UGT in human liver microsomes, several indolocarbazole analogs were chosen for substrates. We anticipated that the position of the phenolic hydroxy moiety would be important for causing glucuronidation. The compounds with an NHCHO residue on the N-6 position with a phenolic -OH moiety on the 10 or 11 position were glucuronidated; however, none of the compounds having an NHCH(CH₂OH)₂ residue on the N-6 position with a phenolic -OH moiety on the 9, 10, or 11 position were glucuronidated. The results showed, unexpectedly, that there was no association between the glucuronidation activity and the position of the hydroxy moiety. To find the essential features of the substrate structure for UGT, compounds that differed at the N-6 residue were synthesized and analyzed for a correlation between glucuronidation activities and the molecular orbits of these compounds. The atom electron density and electrostatic potential charges at the hydroxy position indicated that the chemical reactivities of the hydroxy moiety were not different in these substrates. The molecular dimensions, however, represent an important factor for glucuronidation. For glucuronidation to occur, it was essential to have a structure less than 14.5 Å in diameter in the direction perpendicular to the phenolic O-position.

The values of glucuronidation activities of the compounds, which could be glucuronidated in human liver microsomes, did not correlate with the size or atom electron density and electrostatic potential charges at the hydroxy position. This might be due to other factors responsible for the catalytic activities of UGT (e.g., stereo-hindrance between amino acid residues lining the catalytic site of UGT and substrate or other UGTs involved in the metabolism of remaining 18 molecules not assessed as expressed UGTs). To further understand the quantitative structure-activity correlation in O-glucuronidation, to identify the UGT isoforms for each indolocarbazole analogs and to evaluate the glucuronidation activity in expressed human UGTs would be helpful.

We previously reported that the glucuronidation of compound 1 by human liver microsomes was a monophasic reaction, and the apparent K_m value was 75 µM (Takenaga et al., 1999a). The UGTs represent a superfamily of enzymes, and the nomenclature has already been established (Mackenzie et al., 1997). The UGT1 family consists of phenol and bilirubin UGTs that result from alternative splicing of at least 13 divergent first exons, with exons 2 to 5 in common (Gong et al., 2001). The UGT2A subfamily comprises at least two olfactory-specific genes (Lazard et al., 1991; Tukey and Strassburg, 2001); one (UGT2A1) has been demonstrated to be functionally active (Jedlitschky et al., 1999). The UGT2B subfamily contains phenobarbital-inducible genes and numerous genes that are involved in the glucuronidation of endogenous steroids and xenobiotics (Mackenzie, 1986; Burchell et al., 1995). To determine which isoforms catalyze the indolocarbazole analogs, some commercially available microsomes that express human UGT were tested. The UGT activity toward compound 1 was observed in microsomes expressing UGT1A9 and UGT1A10. These results indicate that at least UGT1A9 and UGT1A10 contribute to the glucuronidation of compound 1. It was reported that UGT1A9 is expressed in the liver and intestine (Strassburg et al., 1997). UGT1A10, however, is expressed in the biliary and gastric tissue but is not observed in the liver (Strassburg et al., 1997). The UGT activity toward compound 1 observed in human liver microsomes would reflect the UGT1A9 activity. The possibility still remains that glucuronidation of compound 1 is catalyzed by other UGT1A and UGT2B subfamilies.

In conclusion, structure-activity correlation analysis indicated that the substitution position of the hydroxy moiety did not affect glucuronidation in the tested compounds. The molecules measuring over 14.5 Å in diameter in the direction perpendicular to the phenolic O-position were not glucuronidated. These results suggest that a molecular size less than 14.5 Å might be required for a substrate to interact with catalytic site of UGT.