Page 189 - The Toxicology of Fishes

Biotransformation in Fishes 169

mammalian and teleost CYP3A genes were suggested using nearest-neighbor and maximum parsimony
methods (Celander and Stegeman, 1997). More recently, Bayesian analysis of 45 vertebrate CYP3A
deduced amino acid sequences suggest that teleost, diapsid, and mammalian CYP3A genes have under-
gone independent diversification and that an ancestral vertebrate genome contained a single CYP3A
gene (Hegelund and Celander, 2003; McArthur et al., 2003). Phylogenetic analyses suggest that the
divergence of CYP3A paralogs and additional subfamily members is likely due to successive gene-
duplication events. Whole genome duplications in teleosts have been suggested (Christoffels et al., 2004;
Furutani-Seiki and Wittbrodt, 2004), and multiple CYP3A paralogs have been identified in several
species, including medaka, rainbow trout, and killifish (Celander and Stegeman, 1997; Hegelund and
Celander, 2003; Kullman and Hinton, 2001; Kullman et al., 2000; Lee and Buhler, 2003; Lee et al.,
1998; Lemaire et al., 1996). Teleost CYP3A paralogs demonstrate high degrees of sequence similarity:
90% (CYP3A38 and CYP3A40), 94% (CYP3A27 and CYP3A45), and 98% (CYP3A30 and CYP3A56)
for medaka, trout, and killifish, respectively. Each sequence conforms to the specific structural features
associated with the cytochrome CYP gene superfamily and exhibits >40% sequence similarity to the
CYP3A subfamily. It has been suggested that the topologies of all CYP enzymes are similar, especially
regarding structurally conserved regions such as the heme-binding domain, oxygen-binding region, and
specific sites associated with redox interactions (Szklarz and Halpert, 1997). Differences in CYP catalytic
activities are suggested to be determined predominantly by amino acid composition in six substrate
recognition sites (SRS1 to SRS6) (Gotoh, 1992). Recently homology models for CYP3A genes have
been described (Harlow and Halpert, 1998; Yang et al., 1998). Key amino acids associated with CYP3A
substrate specificity, binding, and regio-specific catalysis have been suggested by using molecular
modeling and site-directed mutagenesis. Statistical comparisons using the DIVERGE program identified
regions in SRS1, SRS5, and SRS6 that appear to be associated with a general conserved CYP3A function,
whereas SRS2, SRS3, and SRS4 confer functional differences among different CYP3A enzymes
(McArthur et al., 2003). Alignments of medaka CYP3A38 and CYP3A40 demonstrate that 12 of 49
amino acid differences occur in SRS regions. These differences are predominately observed in SRS1,
SRS3, and SRS6. As noted below, it has been suggested that these amino acid substitutions are responsible
for the differing kinetic and catalytic properties of these two teleost paralogs.
Multiple CYP3A-like teleost proteins have additionally been observed using immunochemical detection
in numerous other species. Although gene sequences for these species have not been identified, cross-
reactivity with antibodies specific for either mammalian or teleost CYP3A proteins suggests that multiple
CYP3A-like proteins are present in the liver and intestine of several teleosts (Celander et al., 1996).
Function—Functionally, CYP3A enzymes are among the most versatile forms of CYPs as they have
unusually broad substrate specificities for both endogenous and exogenous substrates, including steroids,
bile acids, eicosanoids, retinoids, xenobiotics such as pharmaceuticals, and procarcinogens (Aoyama et
al., 1990; Gillam et al., 1993; Li et al., 1995; Smith et al., 1996; Waxman et al., 1998). CYP3A-like
proteins were initially purified from several teleost species, including scup, rainbow trout, and Atlantic
cod (Celander et al., 1989; Klotz et al., 1986; Miranda et al., 1989). Identification of these proteins as
CYP3A-like was based predominantly on steroid hydroxylase activity and cross-reactivity with CYP3A-
specific antibodies. In some instances, antibodies were additionally used as catalytic inhibitors. Purified
cytochrome P450A from scup and LMC5 from rainbow trout exhibited specific steroid hydroxylase
activity, similar to that observed with mammalian CYP3A enzymes. Each enzyme additionally demon-
strated minimal benzo(a)pyrene hydroxylase and ethoxycoumarin O-deethylase activities, suggesting a
functional difference from the previously identified and inducible CYP1A form. Further characterization
of purified teleost enzymes was undertaken by comparative reciprocal western blot analysis. Cross-
reactivity between teleost and mammalian antibodies further supported a close structural as well as
functional similarity between teleost and human CYP3A enzymes (Celander et al., 1996; Miranda et
al., 1991). Functional characterization of recombinant CYP3A enzymes has been determined for
CYP3A27, CYP3A45, CYP3A38, and CYP3A40 by heterologous expression in baculovirus systems.
Recombinant rainbow trout CYP3A27 exhibited a maximum CO-reduced spectrum at 450 nm and
comigrated with purified CYP3A27 (formerly denoted LMC5) on western blots. In reconstitution exper-
iments, recombinant protein exhibited catalytic activities for the 6β-, 2β-, and 16β-hydroxylation of