Introduction

Top Abstract Introduction References

In 1949, Smith and Williams (1949) showed that the main urinary metabolite (more than 50% of the dose) of aniline in the rabbit was an acid-labile glucuronide and suggested that this metabolite was aniline N-glucuronide. Subsequently, many primary, secondary, and tertiary amines have been shown to be excreted as N-glucuronide metabolites. In general, glucuronidation is considered to be a significant means of detoxication and elimination from the body because of the conversion of hydrophobic xeno- and endobiotics to hydrophilic metabolites. Understanding N-glucuronidation and identifying the enzymes involved in this process and their regulation is especially important in humans for two reasons. First, N-glucuronidation of many primary and secondary aromatic amines has been suggested to have implications in the potential carcinogenicity of these compounds. In addition, metabolic elimination of many tertiary amine therapeutic agents used in humans (antipsychotic agents, antihistamines, and tricyclic antidepressants) is dependent on N-glucuronidation of the parent drugs to form quaternary ammonium-linked glucuronide metabolites.

Excretion of N-glucuronides of aromatic amines (e.g. - and -NA¹ 4-ABP, or benzidine) in urine accounts for only a low percentage of the administered dose of the compound. This is due to the competing metabolic pathways that exist for primary amines and to the lability of the aromatic amine N-glucuronides (Boyland and Manson, 1966; Kadlubar et al., 1977; Kadlubar et al., 1981; Hammons et al., 1985; Gorrod and Manson, 1986; Babu et al., 1992; Kadlubar et al., 1992). In general, primary aromatic amines are eliminated by three competing metabolic processes: N-acetylation, oxidation to form N- or ring-hydroxylated metabolites, and conjugation of the parent amine or of the N-hydroxylated metabolite. N-acetylation is thought to lead to detoxication of aromatic amines because it decreases the amount of amine that can be oxidized to N-hydroxylated compounds (Hein, 1988). Indeed, it has been shown that individuals expressing the slow acetylator phenotype had higher amounts of 4-ABP adduct formation, compared with high acetylators (Vineis et al., 1990; Vineis et al., 1994). In contrast to the inactivation of aromatic amines by N-acetylation, N-hydroxylation of aromatic amines is correlated with adduct formation from these compounds (Kadlubar et al., 1977). The presumptive reactive intermediates involved in this covalent binding are nitrenium ions formed from the N-hydroxylated aromatic amines (Kadlubar et al., 1977). Hydroxylation of the aromatic ring, on the other hand, results in the formation of metabolites that are substrates for sulfation and O-glucuronidation (Boyland and Manson, 1966; Gorrod and Manson, 1986). Sulfation or glucuronidation of the primary aromatic amine or its N-hydroxylated secondary amine metabolite are the two main conjugation reactions involved in the elimination of these compounds.

As shown by Smith and Williams (1949) and others (Babu et al., 1992; Babu et al., 1995), N-glucuronides of all primary amines and many secondary amines are very labile and are susceptible to cleavage to the parent amine and glucuronic acid under very mildly acidic conditions. For example, the half-time for stability of benzidine N-glucuronide has been shown to be approximately 3 min at pH 5.5 (Babu et al., 1992). At low pH, N-glucuronides of N-hydroxylated metabolites of -NA, -NA, and 4-ABP (secondary amines) have also been shown be converted to their corresponding N-hydroxylated derivatives, which bind DNA (Irving, 1977; Kadlubar et al., 1977; Gorrod and Manson, 1986). It has been proposed that N-glucuronides of N-hydroxy aromatic amines may act as the transport mechanism for delivering carcinogenic amines to the urinary bladder (Kadlubar et al., 1977; Poupko et al., 1979). Benzidine (4-4'-diaminobiphenyl) has two amine groups that can undergo N-acetylation, N-hydroxylation, and N-glucuronidation. Babu et al. (1995) suggested that hepatic N'-glucuronidation of N-acetylbenzidine can result in the delivery of the parent N-acetylated N'-amine to the bladder epithelium, whereas hepatic N'-glucuronidation of the N'-hydroxylated metabolites of N-acetylbenzidine leads to a product that has a higher acid stability. Thus the latter reaction might be considered a detoxification process.

Many chemicals that contain aliphatic tertiary amine moieties are clinically useful drugs, such as antihistamines, tricyclic antidepressants, and antipsychotic agents. The metabolic processes involved in the overall elimination of aliphatic tertiary amines include N-demethylation, hydroxylation reactions, and N-glucuronidation. Demethylation of tertiary amines generally results in the formation of compounds with reduced clinical efficacy. Oxidation of compounds with tertiary amine moieties also generally results in inactivation of the drugs and aids in their elimination by the formation of hydroxyl groups that can be conjugated to form O-glucuronides or sulfates. N-Glucuronidation of the tertiary amine group results in the formation of quaternary ammonium-linked glucuronides. Quaternary ammonium-linked glucuronide formation is highly species-dependent (Fischer et al., 1980; Dulik and Fenselau, 1987; Coughtrie and Sharp, 1991; Remmel and Sinz, 1991) and is an important elimination pathway for many tertiary amines in humans. For example, the quaternary ammonium-linked glucuronide of tripelennamine represents the major metabolic product of this compound found in the urine (Chaudhuri et al., 1976). In contrast to N-glucuronides of primary and secondary amines, which tend to be acid-labile, quaternary ammonium glucuronides tend to be susceptible to hydrolysis under mildly basic conditions. Dulik and Fenselau (1987) reported that at pH 8.5, about 40% of cyproheptadine quaternary-ammonium glucuronide is hydrolyzed over a 24-hr period and suggested that this is an explanation for the inability of investigators to detect the presence of cyproheptadine quaternary ammonium glucuronide in rabbit urine.

	UDP-Glucuronosyltransferases and N-Glucuronidation

Glucuronidation of endo- and xenobiotics is catalyzed by the UGT superfamily of enzymes. UGTs catalyze the transfer of glucuronic acid from UDP-glucuronic acid to compounds with amine, hydroxyl, and carboxylic acid moieties (Clarke and Burchell, 1994; Burchell et al., 1995). Individual UGT isoforms do not catalyze the conjugation of only a single type of chemical moiety; in other words, no single UGT catalyzes conjugation of only amine substratesUGTs generally demonstrate conjugation reactivity toward diverse chemical moieties. Despite this diversity of reactivity, each UGT isoform can be distinguished functionally from another based on its reactivity and/or nonreactivity with different classes of prototypical substrates. For example, expressed rat UGT2B1 protein catalyzes the glucuronidation of many simple phenolic compounds and coumarins like many other UGTs. However, it also catalyzes the glucuronidation of the 17-hydroxyl position of testosterone and related steroids, the hydroxyl group of chloramphenicol, the carboxylic acid moiety of profen nonsteroidal anti-inflammatory drugs and the phenolic 3-hydroxy group of morphine (Mackenzie, 1986; Mackenzie, 1987; Pritchard et al., 1994). The major pharmacologic importance of rat UGT2B1 is its ability to catalyze conjugation of morphine and other opioids. Recent studies in our laboratory suggest that rat UGT2B1 is the major enzyme in untreated rat liver that catalyzes the glucuronidation of morphine in this species (King et al., 1997). Another UGT isoform, rat UGT1A1, also catalyzes morphine glucuronidation, but at much lower rates compared to the UGT2B1 isoform (King et al., 1997). A discussion of the complex substrate specificities of UGTs is outside the scope of this review; however, recent reviews address this subject (Clarke and Burchell, 1994; Burchell et al., 1995).

Currently, UGTs are named in accordance with a nomenclature system based on evolutionary divergence of the UGT genes (Mackenzie et al., 1997), and this nomenclature will be used in this review. To date, three UGT families have been identified in humans: UGT1, UGT2, and UGT8. Of these three families, UGT1 and UGT2 proteins have been shown to catalyze glucuronidation of xenobiotics. The UGT1 and UGT2 genes appear to be structurally different in that UGT1 proteins result from alternate slicing of different first exons with five shared exons encoded by the UGT1 gene complex, while UGT2 proteins appear to be encoded by unique genes. In the human genome, at least 12 different first exons have been identified for the UGT1 gene (Ritter et al., 1992; Cho et al., 1995). A further consideration of the UGT1 gene complex and the UGT proteins encoded by it will be discussed later in this review. The UGT2 gene family is currently divided into three subfamilies: UGT2A, UGT2B, and UGT2C. Only one member of the UGT2A subfamily has been described, namely, UGT2A1. UGT2A1 was isolated from rat and bovine olfactory epithelium and is preferentially expressed in this tissue (Lazard et al., 1991). In contrast, over 18 members of the UGT2B subfamily have now been isolated (Mackenzie et al., 1997). In general, proteins encoded by the UGT2B subfamily have been characterized primarily for their ability to catalyze the glucuronidation of steroids. However, as noted above for rat UGT2B1, the pharmacologic and physiologic importance of the enzymes of the UGT2B gene subfamily may not relate entirely to their ability to catalyze steroid glucuronidation. Little is known about the substrate specificity of the UGT2C1 protein.

Over the last several years, many UGTs, from a number of different species, have been shown to catalyze the glucuronidation of amines. The purpose of this review is to present current information about the enzymes that catalyze N-glucuronidation. N-Glucuronidation of primary amines has been demonstrated for a number of purified and cloned and expressed UGT proteins. More recently, human and rabbit UGT cDNAs have been shown to encode for proteins that catalyze the glucuronidation of primary, secondary, and tertiary amines. In addition, evidence will be presented which suggests that there is a correlation between species that are able to conjugate sapogenins with glucuronic acid and those that can form quaternary ammonium-linked glucuronides. Finally, a genetic basis for the ability, or inability, of a species to glucuronidate some classes of tertiary amines will be postulated.

N-Glucuronidation of Primary and Secondary Amines Catalyzed by Purified and Cloned and Expressed Rat UGTs. N-Glucuronidation of primary amines occurs in many species, but this process has been studied most extensively in the rat. The most commonly investigated substrates have been - and -NA, and 4-ABP, - and -NA are planar amines, whereas 4-ABP is considered to be more "bulky." Three purified rat liver UGTs have been shown to catalyze N-glucuronidation of primary amines (table 1). Purified rat liver 3-hydroxysteroid UGT (UGT2B2), 17-hydroxysteroid UGT (UGT2B3), and a 3-methylcholanthrene-inducible p-nitrophenol UGT (UGT1A6) catalyze the glucuronidation of - and -NA, whereas glucuronidation of 4-ABP is only catalyzed by UGT2B2 (Green and Tephly, 1987). In addition, purified rat UGT2B2 has been shown to catalyze the glucuronidation of aniline (Roy Chowdhury et al., 1986). For each purified enzyme, -NA was the best substrate, in that its catalytic efficiency (V_max/K_m) was at least tenfold higher than either -NA or 4-ABP (Green and Tephly, 1987). Primary amines are not substrates for purified rat liver digitoxigenin mono-digitoxoside UGT, morphine UGT, and 4-hydroxybiphenyl UGT (von Meyerinck et al., 1985; Puig and Tephly, 1986; Styczynski et al., 1991). Glucuronidation of secondary amines by purified rat liver UGTs has received less attention. Only purified rat liver UGT1A6 has been tested for reactivity toward a secondary amine substrate (N-OH -NA) (Bock et al., 1979).

N-Glucuronidation of Primary and Secondary Amines Catalyzed by Human and Rabbit UGTs. Rabbits have been a useful species for the characterization of oxidative drug-metabolizing enzymes' genes and have proven useful as an animal model to investigate the molecular events associated with the expression of some of these genes. In addition, based upon the speciation period of rodents, legamorphs, and humans, legamorphs appear to be more closely related in evolution to humans than are rodents. For example, the predicted primary amino acid sequence of rabbit UGT1A6 is most related to the human (81%) UGT1A6, compared with the rat (78%) and mouse (78%) enzymes (Lamb et al., 1994). Therefore, the identification and characterization of UGT genes and gene products from rabbits could serve to broaden our understanding of how these genes have evolved, while providing an additional mechanism to study how the UGT genes may be regulated in humans.

Glucuronidation of Tertiary Amines Catalyzed by Expressed Human and Rabbit UGT Proteins. While expressed human UGT1A9, 1A4, 1A3, and 1A6 react with some of the same primary and secondary amine substrates, their reactivity with tertiary amine substrates varies considerably. In humans, tertiary amines are an important class of substrates because many clinically important therapeutic agents have this chemical moiety and are extensively conjugated to form quaternary ammonium-linked glucuronides. Expressed human UGT1A3 and 1A4 catalyze the glucuronidation of a wide variety of aliphatic tertiary amines (table 3).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Chemical structures for various amine-containing drugs. Prototypical polycyclic amines that are substrates for expressed human UGT1A4 are depicted in row A. Other tertiary amine substrates for expressed human UGT1A4 are shown in row B. Amines that are not substrates for the expressed enzyme are shown in row C.

	Possible Genetic Basis for Species Differences in Quaternary Ammonium Glucuronide Formation

The rat and human UGT1 gene complexes are structurally similar (fig. 2). The human UGT1 gene complex appears to encode for at least 12 unique gene products, each of which is formed by alternate splicing of multiple first exons with common exons 2-5 (Ritter et al., 1992; Cho et al., 1995). Similarly, at least nine unique first exons have been described for the rat UGT1 gene complex (Emi et al., 1995). Alternate splicing results in the production of mRNAs that code for proteins with unique amino acid sequences in the amine terminal portion of the protein, whereas the carboxyl portion of the protein is identical within the UGT1 gene family. The first five exon 1s (human A1-F1 and rat B1-B5) have been described to code for proteins that are "bilirubin-like," while the rest of the first exons code for UGTs that are "phenol-like" (Ritter et al., 1992; Emi et al., 1995). Human exon A1 and rat exon B1 both code for the amino terminal portion of UGT1A1. Expressed UGT1A1 proteins catalyze the glucuronidation of bilirubin, opioids, and other compounds, and we have recently shown that the rat and human UGT1A1 proteins are functionally similar and are likely to be orthologous enzymes (King et al., 1996). Likewise, human, rabbit, mouse, and rat UGT1A6 have been shown to be functionally similar and have been suggested to be orthologous (Lamb et al., 1994; Emi et al., 1996). Finally, the recent results of Green et al. (1998) and Bruck et al. (1997) show that expressed human and rabbit UGT1A4 proteins catalyze glucuronidation of tertiary amines to quaternary ammonium-linked glucuronides. These data suggest that the UGT1 gene structure and the function of the different proteins coded by this gene are highly conserved between different species.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Structures of the human and rat UGT1 gene complex. The multiple first exons of the human [1A-1L(P)] and rat [B1-B5 and A1-A4(P)] UGT1 gene complexes are indicated. The drawing is not represented to scale and is not meant to imply accurate distances between the first exons. The common exons 2-5 are depicted as a block without the intervening introns. The human UGT1 gene complex is compiled from Ritter et al. (1992) and Cho et al. (1995), while the rat data is from Emi et al. (1995, 1996). A (P) indicates that the exon codes for a pseudo gene. In the rat gene complex, A1* indicates an extension of the first exon that codes for a 5' nontranslated portion of the UGT1A6 mRNA that has been shown to have upstream transcriptional activation factors associated with it.

In the rat genome, UGT1 exon B2 (UGT1A2) codes the amine terminal portion of a second protein that catalyzes the glucuronidation of bilirubin (Sato et al., 1990; Sato et al., 1994). However, this exon encodes a premature stop codon and is, therefore, a pseudo-gene in the human genome (Ritter et al., 1992). Thus rats express two bilirubin UGTs, whereas humans express only one. These observations are consistent with the fact that patients with Crigler-Najjar type I syndrome, who have genetic defects only in UGT1 exon A1, are unable to glucuronidate bilirubin. These data also indicate that it is unlikely that the other "bilirubin-like" first exons code for proteins that catalyze the glucuronidation of bilirubin in humans. The amine terminal portion of human UGT1A4 is encoded by UGT1 exon D1. The corresponding rat enzyme is encoded by rat UGT1 exon B4. However, in the rat, exon B4 is a pseudo-gene and does not code for a full-length UGT protein. Therefore, it seems reasonable to suggest that the reason that rats do not glucuronidate many tertiary amines is that they cannot produce a functional UGT1A4 due to a genetic defect in first exon B4. The recent observation (Bruck et al., 1997) that expressed rabbit UGT1A4 also catalyzes glucuronidation of tertiary amines further supports this hypothesis.

At the present time, it is not known whether UGTs other than human UGT1A3 and UGT1A4 catalyze quaternary ammonium-linked glucuronide formation. A significant contribution of human UGT1A3 to the overall metabolism of primary and tertiary amines is somewhat doubtful, given that these compounds tend to exhibit higher apparent K_m values, compared with UGT1A4 (Green et al., 1998) and that the expression of UGT1A3 is very low in human liver (Mojarrabi et al., 1996). Human UGT1A5 is >90% identical in primary amino acid sequence to human UGT1A3 and 1A4, and it is possible that this protein may also catalyze glucuronidation of amine substrates. However, to date, expression of UGT1A5 has not been demonstrated in a number of human tissues that have been examined (Strassburg et al., 1997). If human UGT1A7 has the ability to catalyze the formation of quaternary ammonium-linked glucuronides, as does the rabbit protein (Bruck et al., 1997), its contribution to the overall metabolism of tertiary amines is also probably insignificant because of the possible higher apparent K_m values of the amine substrates and because human UGT1A7 is not expressed in liver (Strassburg et al., 1997). The observation that UGT1A7 is expressed in gastric tissue but is not expressed in human liver suggests that tissue-specific expression of UGT isoforms may have significance for the local metabolism of xenobiotics and endobiotics. Future studies are necessary to further characterize and identify the proteins responsible for amine glucuronidation in human tissues.