Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Human cytochrome P450 1B1 (CYP1B1) has been shown to be an important enzyme in the activation of diverse procarcinogens, such as arylarenes, nitroarenes, polycyclic aromatic hydrocarbons, and arylamines, to reactive metabolites that cause DNA damage (Shimada et al., 1996). CYP1B1 is constitutively expressed in steroidogenic tissues such as uterus, breast, ovary, testis, prostate, and the adrenal gland (Shimada et al., 1996) and is active in the metabolism of estradiol (Hayes et al., 1996), as well as of benzo[a]pyrene and 7,12-dimethylbenz[a]anthracene (7,12-DMBA), known procarcinogens. The enzyme is also present in many other extrahepatic tissues, including kidney, thymus, spleen, brain, heart, lung, colon, and intestine (Sutter et al., 1994; Shimada et al., 1996). A nuclear localization of the enzyme has been proposed based upon immunohistochemistry experiments (Muskhelishvili et al., 2001). CYP1B1 is expressed at a higher level in a wide range of human cancers, including cancers of the skin, brain, testis (Murray et al., 1997), and breast (McKay et al., 1995) compared with the nontransformed tissue.

The human CYP1B1 gene has been mapped to chromosome 2 (Sutter et al., 1994) and encompasses three exons with the coding region starting in exon 2 (Tang et al., 1996). The mRNA is 5.2 kilobases and encodes a protein of 543 amino acids (Sutter et al., 1994). Six different single nucleotide polymorphic sites (SNPs) have previously been reported in the human CYP1B1 gene (Stoilov et al., 1998), of which four cause amino acid substitutions. Three common CYP1B1 alleles have been identified so far. CYP1B1*2 has two linked polymorphisms in exon 2, 142CG (m1) and 355GT (m2), resulting in the Arg48Gly and Ala119Ser amino acid substitutions, respectively. CYP1B1*3, has 4326CG (m3), causing a Leu432Val exchange, and CYP1B1*4 has 4390AG (m4), leading to an Asn453Ser substitution (see http://www.imm.ki.se/CYPalleles/cyp1b1.htm).

Human CYP1B1 heterologously expressed in yeast shows higher specific activity toward 4-hydroxylation than 2-hydroxylation of 17-estradiol (Hayes et al., 1996) compared with CYP1A1 and CYP1A2, which preferentially oxidize the steroid at the 2-position (Yamazaki et al., 1998). The finding of a considerably higher rate of 4-hydroxylation of 17-estradiol in breast cancer microsomes, compared with a very low level of 4-hydroxylation in normal breast tissue (Liehr and Ricci, 1996), has received attention because 4-hydroxyestradiol has been suggested to be carcinogenic in both human myometrium (Liehr et al., 1995) and breast tissue (Liehr and Ricci, 1996). Increased CYP1B1 expression and elevated rates of 4-hydroxylation of estradiol in transformed tissue may be markers of a malignant phenotype and may also provide a basis for targeting of anticancer drugs that are metabolized by this enzyme.

The importance of CYP1B1 in chemical carcinogenesis has been illustrated using Cyp1b1-null mice. Embryonic fibroblast cells derived from Cyp1b1 null mice were found to be resistant to 7,12-DMBA mediated toxicity in contrast to cells from the wild-type animals (Buters et al., 1999). The Cyp1b1-null mice were also protected from 7,12-DMBA-induced malignant lymphomas (Buters et al., 1999), as well as bone marrow cytotoxicity and preleukemia (Heidel et al., 2000), compared with their wild-type counterparts.

Several studies have been devoted to the investigation of a potential relationship between SNPs present in the human CYP1B1 gene and incidence of various cancer forms (Bailey et al., 1998; Tang et al., 2000; Watanabe et al., 2000; Zheng et al., 2000). Some studies have revealed a significant relationship between the presence of the Ala119Ser variant and the incidence of breast cancer as well as squamous cell carcinoma in the lung (Watanabe et al., 2000). An increased risk for prostate cancer was also reported in subjects homozygous for the Leu432Val form (Tang et al., 2000). It is anticipated that these polymorphisms would cause an altered function of the enzyme to provide a role for the CYP1B1 polymorphism in determining interindividual differences in susceptibility for carcinogenesis.

The aim of the present study was to analyze for CYP1B1 missense mutations in an African population, develop methods for detection of the different possible haplotypes and evaluate their functional properties after expression of the enzyme variants in vitro. The results show altered kinetic properties for two of six CYP1B1 enzyme variants occurring in an Ethiopian population and emphasize the necessity of relating haplotypes rather than SNPs to disease when conducting molecular epidemiological studies.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Genomic Sequencing of the Human CYP1B1 Gene. The open reading frame of CYP1B1 and exon-intron boundaries were amplified by PCR in three different fragments (Stoilov et al., 1998) from five genomic DNA samples, previously genotyped for m1, m2, m3, and m4 (see below). The PCR products were purified using the Wizard PCR Preps DNA Purification kit (Promega, Madison, WI). Fragments one and two, which consist of overlapping regions in exon two, were sequenced in both directions using primers seq1F and seq1R and primers seq2F and seq2R, respectively. Fragment three, which comprises the whole coding region of exon three, was sequenced using primers seq3F and seq3R, respectively. Primer sequences are listed in Table 1. DNA sequencing was performed using the ABI PRISM BigDye terminator cycle sequencing ready reaction kit (Applied Biosystems, Foster City, CA) and analyzed on an ABI Prism 377 DNA sequencer.

Genotyping of the CYP1B1 Missense Mutations. The genomic DNA used was from 150 healthy unrelated Ethiopians who previously had participated in a study about genetic polymorphism in the CYP2D6 gene (Aklillu et al., 1996). The ethical committee at Karolinska Institutet approved the use of these samples for the current investigation.

View this table: [in this window] [in a new window]   TABLE 2 Primer combination and optimal PCR conditions used to genotype 142CG (m1), 355GT (m2), 4326CG (m3), 4390AG (m4), and 4390CG (m5), as well as linkage disequilibrium analysis (see Materials and Methods)

CYP1B1 Haplotype Analysis. Subjects heterozygous for m1 and m2 and for m3 and m4 were further analyzed for linkage disequilibrium of the polymorphisms as outlined in Fig. 1. Part of each exon covering the polymorphic sites amplified in PCR 1 (see above) was used as a template in PCR 2 after a 10-fold dilution in water. In PCR 2, the haplotypes were determined using two allele-specific primers in both forward and reverse directions in all four possible combinations. Reaction mixtures were as describe above, and primer combinations and optimal PCR conditions are summarized in Table 2.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   A schematic diagram of the PCR-based strategy used for analysis of SNPs and haplotypes in the human CYP1B1 gene. The positions of 142CG (m1), 355GT (m2), 4326CG (m3), 4390AG (m4), and 4360CG (m5) in the coding region of CYP1B1 gene are marked within the exon boxes. Horizontal arrows denote positions of primers used in the haplotype analysis namely, m1 versus m2, m3 versus m4, m3 versus m5, and m1 versus m3. Primers used in the first PCR (PCR 1) are indicated in bold.

View larger version (57K): [in this window] [in a new window]   Fig. 2.   Analysis of m3 and m4 linkage disequilibrium. Genomic DNA from subjects E1 (homozygous wt for both m3 and m4) and E65 (homozygous for the m3 SNP) were used as controls. ES11 was heterozygous for both SNPs. Part of exon 3 covering m3 and m4 was amplified with PCR (see Fig. 1) and used as a template in the second simultaneous allele-specific PCR for both m3 and m4. Combinations of primers used in PCR 2 are listed under each corresponding lane. In subject E1, PCR amplification was only obtained when both the wt-specific primers were used (lane 1) and in subject E65 PCR amplification was obtained only with the m3 mutant-specific and the m4 wt-specific primers (lane 3). Amplification of the region between m3 and m4 with all four possible combinations of primer give a PCR product in only two sets of primers in heterozygous subjects. Thus in subject ES11, PCR amplification was obtained only in lanes 2 and 3 using the m3wt/m4 mt and m3 mt/m4wt primer combinations, respectively, indicating that m3 and m4 are on separate alleles. 100 bp DNA Ladder (Invitrogen, Carlsbad, CA) was used as a marker.

View larger version (48K): [in this window] [in a new window]   Fig. 3.   PCR-restriction fragment length polymorphism-based analysis of linkage between m3 and m5. The two alleles were separated with respect to the presence of m3 using genomic DNA from three representative subjects. Subject E5 was homozygous mutated for m3 and heterozygous for m5, E33 was heterozygous for both m3 and m5. E112, wt for both m3 and m5, was used as control. The two forward primers, Fm3wt or Fm3 mt, and a common reverse primer, ex3R (indicated under each lane), were used to separate the two alleles with respect to m3 genotype resulting in a 146-bp PCR fragment. This fragment was subsequently subjected to digestion with BstNI as the 4360CG mutation (m5) introduces a BstNI restriction site. Complete digestion is only seen in PCR fragments obtained with the m3 mt primer (Fm3 mt), indicating that m5 is genetically linked with m3. DNA fragment sizes are indicated with arrows on the right side. A 100-bp DNA Ladder (Invitrogen) was used as a marker.

View larger version (80K): [in this window] [in a new window]   Fig. 4.   Linkage analysis of m1 and m3. A, alleles were separated with respect to the presence of m1 by allele-specific long PCR with two forward primers, m1Fwt or m1Fmt, and a common reverse primer, 1B13R. The 1-kilobase-plus DNA ladder (Invitrogen) was used as a marker. Subjects E1, homozygous for m1wt, and E65, homozygous for m1 and m3 mutated, were used as control. E97 and E113 were heterozygous for both m1 and m3. B, analysis of PCR fragments from A, for the presence of m3 (PCR 2). Fragments from PCR 1 (A) were subjected to PCR amplification using a common exon 3 forward primer (ex3F) and reverse m3Rwt/m3Rmt primers (indicated under each lane). As expected, E1 was only amplified with the m3Rwt primer and E65 with the m3Rmt primers. E97, positive for both m1wt and m1 mt PCR fragments in PCR 1, amplified in the second PCR with both m3Rmt and m3Rwt primers, respectively, indicating the absence of linkage between m1 and m3. Individual E113, heterozygous for m1, fragment obtained with m1wt in PCR 1 was amplified with the m3Rwt in PCR 2 and the fragment obtained with m1 mt primer in PCR 1 was positive with m3Rmt in PCR 2, indicating the presence of m1 and m3 on the same allele and the other allele being wt for both SNPs.

Evaluation of the Haplotype Method. A 4949-bp fragment (from 3676-8624 in GenBank accession number U56438), containing the whole coding regions of exon 2 and exon 3 as well as the entire intron 2 were amplified by long PCR using primers clone 1B1-F (5'-TGC AAA GCT TTC AGA GAG TCA GCT CCG-3') and clone 1B1-R (5'-CTA GCC GCG GTA TGG AGC ACA CCT CAC CTG-3') from seven subjects assigned as having different CYP1B1 haplotypes by the above described haplotyping method. The HindIII and SacII restriction sites in primer clone 1B1-F and 1B1-R, respectively, are underlined. The conditions for long PCR were as described above except using 1.0 mM MgO(Ac)₂ and 35 reaction cycles. The long PCR fragments were purified using the Wizard PCR Preps DNA Purification kit (Promega) and cloned into pBluescript SK after digestion with HindIII and SacII (New England Biolabs, Beverly, MA). The whole coding region and exon-intron boundaries were sequenced in three fragments as described above using the ABI PRISM BigDye Terminator cycle sequencing Ready reaction kit and analyzed on ABI PRISM 377 DNA sequencer.

Construction of pYeDP60-CYP1B1 Expression Plasmids. The various CYP1B1 cDNAs were generated by site-directed mutagenesis and cloned into the pYeDP60 expression plasmid (Urban et al., 1990). The CYP1B1*1 and CYP1B1*2 expression plasmids are as described in (McLellan et al., 2000). The CYP1B1*1 plasmid was used as a template to generate the m3 SNP using the forward primer 5'-GGT CTG TGA ATC ATG ACC CAG TGA AGT GGC CTA ACC CGG-3', the reverse primer 5'-CCG GGT TAG GCC ACT TCA CTG GGT CAT GAT TCA CAG ACC- 3' and the QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. The m1+m2+m3 variant, CYP1B1*6, was made using the NotI-SacI fragment from the CYP1B1*3 plasmid to replace the corresponding part in the CYP1B1*2 plasmid. The other expression plasmids were generated using the same principle as described for the CYP1B1*3 plasmid. The m4 SNP was introduced into the CYP1B1*1 plasmid using the forward primer 5'-GGA CGG CCT CAT CAG CAA GGA CCT GAC CAG C-3' and the reverse primer 5'-GCT GGT CAG GTC CTT GCT GAT GAG GCC GTC C-3'. The m1+m2+m3+m5 variant, CYP1B1*7, was generated by introducing the m5 polymorphism into the CYP1B1*6 plasmid using the forward primer 5'-GGA GAA CTT TGA TCC AGG TCG ATT CTT GGA CAA GG-3' and the reverse primer 5'-CCT TGT CCA AGA ATC GAC CTG GAT CAA AGT TCT CC-3'. The m3+m5 variant was generated by introducing the m5 polymorphism into the CYP1B1*3 plasmid. The forward primer was 5'-GGA GAA CTT TGA TCC AGG TCG ATT CTT GGA CAA GG-3' and the reverse primer 5'-CCT TGT CCA AGA ATC GAC CTG GAT CAA AGT TCT CC-3'. For production of the m3+m4 variant, the m4 polymorphism was introduced into the CYP1B1*3 plasmid. The forward primer was 5'- GGA CGG CCT CAT CAG CAA GGA CCT GAC CAG C-3' and the reverse primer 5'-GCT GGT CAG GTC CTT GCT GAT GAG GCC GTC C-3'. The SNP, intended to be introduced in the corresponding cDNA, is shown in bold. All plasmids were sequenced using the ABI Prism Big Dye terminator cycle sequencing kit and analyzed with ABI Prism 377 DNA sequencer to ensure the correct constructs and to exclude any potential PCR artifacts.

Expression of CYP1B1 Variants in Yeast. The yeast strain INVSc1-HR MAT  his31 leu2 trp1-289 ura3-52 (pFL-35 human reductase) expressing human cytochrome P450 reductase was a gift from the LINK project (a program of the University of Dundee/Biotechnology and Biology Research Council/Department of Trade and Industry/Pharmaceutical Industry) (Masimirembwa et al., 1999). Yeast cells were transfected with the pYeDP60-CYP1B1 expression plasmids described above and were subsequently inoculated into selective medium and grown as described in (McLellan et al., 2000), where also the method for isolation of microsomes is described.

Analysis of Estradiol Metabolism. Rates of CYP1B1-dependent metabolism of estradiol in yeast microsomes were analyzed as described previously (McLellan et al., 2000) employing reversed-phase high-performance liquid chromatography with a C18 column (Merck, 250 × 4 mm, 5 µm) and acetic acid/methanol/acetonitrile/water [1:1:38:60 (v/v)] as mobile phase. In brief, incubation mixtures contained microsomes corresponding to 7 pmol of the CYP1B1 variants, varying concentration (0-12.8 µM) of ¹⁴C-labeled 17-Estradiol (PerkinElmer Life Sciences), and was buffered to pH 7.4 with 100 mM sodium phosphate in a final volume of 0.5 ml. After 2-min preincubation at 37°C, the reaction was initiated by adding NADPH to a final concentration of 0.5 mM. In all incubations, equivalent protein amounts (2 mg/incubation) were used by addition of microsomes from yeast transfected with the empty pYeDP60 plasmid. After 5 min, the incubations were terminated by addition of 2 ml of dichloromethane. The organic phase extracts were collected after centrifugation for 10 min at 3500 rpm and evaporated under a nitrogen stream. The residue was dissolved in 160 µl of mobile phase and 100 µl was injected into the high-performance liquid chromatography.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of CYP1B1 Polymorphisms in the Ethiopian Population. To detect new polymorphic sites in the CYP1B1 gene, we sequenced the whole open reading frame and exon-intron boundaries of CYP1B1 using genomic DNA from five different subjects previously genotyped for the CYP1B1 missense mutations m1-m4. Sequencing was carried out using genomic DNA from 1) subject E1, homozygous wt for all SNPs; 2) subject E15, homozygous for both m1 and m2; 3) subject E6, homozygous for m3; 4) subject E65, homozygous for m1, m2, and m3; and 5) subject ES11, heterozygous for m1, m2, m3, and m4. The results confirmed the presence of the SNPs determined by the allele-specific PCR analysis and also detected three new SNPs in subjects E65 and ES11. Two silent mutations, 729GC and 4110TC, were identified in exon 2 and exon 3, respectively. Furthermore, a missense mutation 4360CG (m5), which causes an Ala443Gly substitution, was detected in exon three in both E65 and ES11.

Genotyping for m1, m2, m3, m4, and m5. The five missense mutations of CYP1B1, m1, m2, m3, m4, and m5, were analyzed in genomic DNA from 150 healthy unrelated Ethiopian persons using allele-specific PCR methods as detailed under Materials and Methods. The frequencies of the SNPs were 51, 50, 53, 2, and 7%, respectively. None of the subjects was homozygous for m4 or m5.

SNP Mapping and Determination of the CYP1B1 Haplotypes. We considered it important to determine the haplotypes of the various CYP1B1 genes in the population because of the high frequency of the missense mutations and incomplete knowledge of their linkage disequilibrium. All subjects who were heterozygous for two or more SNP sites were further analyzed by three different PCR methods. Linkage of m1 in exon 2 with m3, m4, or m5 in exon 3 was achieved by separation of the two chromosomes by allele-specific long PCR using m1 wt or mutant specific forward primers and a common reverse primer in the 3'-end of exon 3 as outlined in Fig. 1. The m1 separated alleles were further genotyped for the m3, m4, or m5 using allele-specific PCR.

CYP1B1 Haplotypes in Ethiopians. With five missense mutations in the CYP1B1 gene, 2⁵ = 32 possible different CYP1B1 haplotypes could theoretically be present, but only seven were found. These (CYP1B1*1, *2, *3, *4, *5, *6, and *7) were present at allele frequencies of 8, 37, 39, 2, 0.7, 6, and 7%, respectively, in the Ethiopian population examined (Table 3). Eighteen different combinations of these haplotypes were present among the subjects examined, of which the CYP1B1*2/*3 combination was found to be the most frequent one (Table 4). One subject was found to be homozygous for CYP1B1*6, whereas no subject was homozygous for CYP1B1*4, CYP1B1*5, or CYP1B1*7. In fact, only one subject was homozygous for the CYP1B1*1 allele. The frequency distribution of the observed haplotype combination did not significantly deviate from the ones expected by the Hardy-Weinberg law using the ² test by performing a 2 × 2 cross table for each genotype group against the rest using the computer program Statistica (version 5.5; StatSoft, Tulsa, OK). Fisher's exact test was used for genotype groups where number of expected frequencies was less than 5. 

Evaluation of Catalytic Properties of the Variant Forms of CYP1B1. The human CYP1B1 cDNAs corresponding to the different alleles were cloned into the pYeDP60 expression vector and expressed in Saccharomyces cerevisiae strain INVSc1-HR. This strain is genetically modified to over express the human P450 reductase gene (see above). The kinetic properties of the six most common variants CYP1B1.1, CYP1B1.2, CYP1B1.3, CYP1B1.4, CYP1B1.6, and CYP1B1.7, as well as constructs with the m3+m4 and m3+m5 in combination, were analyzed using 17-estradiol as a substrate and microsomes from yeast transfected with the respective cDNA variant. All CYP1B1 variants expressed well in yeast and folded properly to give a typical P450 spectrum in the CO-reduced form with a peak at 450 nm. The expression levels were similar, although not identical for all forms (Table 5). These results indicate that neither of the missense mutations alone or in combination influences the gross structure of the enzyme or the folding efficiency.

View this table: [in this window] [in a new window]   TABLE 5 Kinetic analysis of the 4- and 2-hydroxylation of 17-estradiol by human CYP1B1 variants expressed in S. cerevisiae Kinetic parameters are presented as means ± S.D. from three independent analyses using two separate microsomal preparations for each enzyme variant.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Based on the assumption that allele-specific PCR analysis for the known SNPs in the human CYP1B1 gene might not completely reveal the true haplotypes, we sequenced the complete open reading frame of the CYP1B1 genes from five different subjects and identified three novel SNPs, one causing an amino acid change. Methods for haplotype analysis of the four previously known missense mutations and for the newly identified missense mutation (m5) were developed and genomic DNA from 150 unrelated healthy Ethiopians were analyzed, revealing the presence of seven different CYP1B1 haplotypes in the population. In total, we detected 18 different combinations of haplotypes in the population (Table 4), illustrating the complexity by which the various SNPs are distributed and the manner by which they would influence the phenotype with respect to CYP1B1 enzyme function. Only two of the haplotypes, however, having a combination of three or four different amino acid substitutions, respectively, exhibited altered kinetic properties for estradiol metabolism compared with the wild-type enzyme. The amino acid changes in common of the enzyme variants with altered properties were Arg48Gly, Ala119Ser, and Leu432Val.

Previous work by Stoilov and collaborators has identified 17 different inactivating mutations in the human CYP1B1 gene, cosegregating with primary congenital glaucoma (Stoilov et al., 1998). In addition, six SNPs in the CYP1B1 gene that are not associated with the disease phenotype have been described previously (Stoilov et al., 1998; Bejjani et al., 2000). Thus, together with the three novel SNPs identified in this study, there are at present at least nine polymorphic sites in the CYP1B1 gene, of which five cause amino acid substitutions.

The frequency of m1, m2, and m3 among Ethiopians, the first African population to be studied with respect to the CYP1B1 genotype, was higher than that described for whites (Stoilov et al., 1997) and Japanese (Inoue et al., 2000). The frequency of m3 in the Ethiopian population (54%) was somewhat lower than among African Americans (75%) but higher than among whites (43%) and Chinese (17%) (Tang et al., 2000). The m4 polymorphism, was rare in the Ethiopian population examined and has previously been shown to be more frequent in whites compared with African-Americans (Bailey et al., 1998), and not at all detected in Japanese population (Inoue et al., 2000). The variant having m4 alone (Asn453Ser) has been shown to be present in MCF-7 cells (Spink et al., 2000). These findings indicate the existence of pronounced interethnic variations in the distribution of different CYP1B1 alleles.

We considered it important to study the functional properties of the haplotypes of CYP1B1 found in the population. For this purpose, we employed the heterologous yeast system with a strain over expressing human cytochrome P450 reductase. All CYP1B1 enzyme variants displayed typical P450 spectra indicating no influence on major structural aspects of the enzyme. The kinetics was, however, significantly different in enzymes having a combination of m1, m2, and m3 (i.e., for CYP1B1.6 and CYP1B1.7) in which the V_max was lower and K_m was higher for estradiol hydroxylation (Table 5). The influence of the new missense mutation m5 (Ala443Gly) identified in the present study was small, but the affinity for estradiol was further reduced in CYP1B1.7 that had this SNP compared with CYP1B1.6. The Ala443 amino acid is conserved both in the mouse, rat and human CYP1B1 sequences. Because glycine residues are often involved in protein bending, it is reasonable to speculate that this amino acid substitution might be of some structural importance.

These results indicate that the true haplotype is important in determining the overall function of the specific form of CYP1B1, indicating that combinations of SNPs would influence the structure of the active site after protein folding. Such an example has previously been registered (e.g., with the CYP2D6.17 variant having the amino acid substitutions Thr107Ile, Arg296Cys, and Ser486Thr). Incorporation of these polymorphisms separately into wt CYP2D6 cDNA did not influence the catalytic function when the enzyme was expressed in yeast and in mammalian COS-1 cells. However, constructs having both the Thr107Ile and Arg296Cys substitutions displayed a reduced affinity for the CYP2D6 substrates bufuralol and codeine compared with the wt enzyme (Oscarson et al., 1997). These findings are in agreement with phenotype analysis in vivo, where the activity of CYP2D6.17 has been found to be severely impaired when debrisoquine was given as a probe drug (Masimirembwa and Hasler, 1997). It can be speculated that the folding of the enzyme is affected by the combination of substitutions resulting in enzyme products with a different active site conformation. A similar mechanism for the specific effects seen for CYP1B1.6 and CYP1B1.7 can therefore be hypothesized. Thus, altered active site structure is only achieved when the substitutions Ala48Gly, Ala119Ser, and Leu432Val exist together as in CYP1B1.6 and CYP1B1.7 (see Table 4).

It should be noted that the folding is partially dependent on the expression system and that a similar relationship might not be evident when the constructs are expressed in, for example, bacteria. Thus, variable K_m values for estradiol metabolism catalyzed by different CYP1B1 variants has been achieved in different expression systems, but variations have also been registered between laboratories (Shimada et al., 1999; Hanna et al., 2000; Li et al., 2000; McLellan et al., 2000; Spink et al., 2000). Although all these discrepancies cannot easily be explained, it is considered important to have similar conditions and expression systems when comparing allelic variants with different missense mutations. Moreover, a similar haplotype activity relationship as observed in the yeast expression system may not hold in mammalian cells or in humans in vivo. For instance, higher levels of testosterone hydroxylation by CYP1B1 are observed when expressed in human lymphoblastoid cells (Crespi et al., 1997) compared with S. cerevisiae (Shimada et al., 1997). Furthermore, as seen with CYP2D6.17 (Oscarson et al., 1997), the haplotypes would be expected to influence the efficiency of metabolism in a substrate-specific manner. Use of in vitro metabolic conditions to assess the activity profile of the various CYP1B1 variants is justified because, unlike several CYP enzymes, such as CYP2D6 and CYP2C19, currently there is no CYP1B1-specific probe drug available for use in vivo. Further studies, if possible employing CYP1B1-specific probes, can perhaps give an answer to the in vivo relevance of these alleles. In addition, it is of importance to study the effect of the CYP1B1.6 and CYP1B1.7 on the metabolism of other CYP1B1 substrates such as benzo[a]pyrene and 7,12 dimetylbenza[a]anthracine, known procarcinogens and substrates of CYP1B1 in vitro.

CYP1B1 is known to metabolize a variety of putative human carcinogens as well as estrogens In addition, CYP1B1 shows elevated expression in a wide range of human tumors including mammary and ovarian tumors. Thus, CYP1B1 might have a potential role in tumor development and progression. Polymorphisms in this enzyme may thus influence the metabolism of environmental estrogens and hormone-like substances and affect the interindividual risk of developing cancer. Thus, it would be of potential importance to investigate for any association of the CYP1B1 haplotypes here identified yielding enzyme products with altered function and incidence of, for example, breast or ovarian cancer.

In conclusion, our study shows the extreme complexity by which five different SNPs in the human CYP1B1 gene are distributed into a human population, resulting in seven different haplotypes and CYP1B1 enzyme variants, of which two exhibit altered kinetic properties for estradiol hydroxylation. Consequently, the CYP1B1 genetic polymorphism may contribute to interindividual and interethnic variation in susceptibility toward the development of cancer upon exposure of procarcinogens or estrogen over a long period of time, but a true investigation regarding the genetic influence of the polymorphic CYP1B1 on the susceptibility of different forms of cancer requires determination of the CYP1B1 haplotypes.