Testosterone 1b-hydroxylation by human cytochrome P450 3A4 Joel A. Krauser 1 , Markus Voehler 2 , Li-Hong Tseng 3 , Alexandre B. Schefer 3 , Markus Godejohann 3 and F. Peter Guengerich 1 1 Department of Biochemistry and Center in Molecular Toxicology, Vanderbilt University School of Medicine and 2 Department of Chemistry and Center in Molecular Toxicology, Vanderbilt University, Nashville, TN, USA; 3 Bruker Bio-Spin GmbH, Rheinstetten, Germany Human c ytochrome P450 3A4 forms a series o f minor testosterone hydroxylation products in addition to 6b-hy- droxytestosterone, the major product. One of these, formed at the next highest rate after the 6 b-and2b-hydroxy prod- ucts, was identified as 1b-hydroxytestosterone. This p roduct was characterized from a mixture of testosterone oxidation products using an HPLC-solid phase extraction-cryoprobe NMR/time-of-flight mass spectrometry system, with an estimated total of  6 lg of t his product. Mass spectrometry established the formula as C 19 H 29 O 3 (MH + 305.2080). The 1-position of the added hydroxyl group was established by correlated s pectroscopy and heteronuclear spin quantum correlation experiments, and the b-stereochemistry of the added hydroxyl group was assigned with a nuclear Over- hauser correlated spectroscopy experiment (1a-H). Of several human P450s examined, only P450 3A4 formed this product. The p roduct was also formed in human liver microsomes. Keywords: cytochrome P450; NMR spectroscopy; HPLC- NMR combinations; testosterone. Cytochrome P450 (P450; also termed heme-thiolate P450 [1]) enzymes h ave long been of interest because of their roles in steroid metabolism [2,3]. These oxidations are most critical in steroidogenic tissues, and a set of  12 P450s are most important [4,5]. The hepatic P450s have also been studied extensively in the context of their abilities to hydroxylate steroids, even though few of the o xidations involve the generation of products with distinctive biological activities. Seminal in this area is the work o f C onney a nd his associates, who studied the hydroxylation of testosterone in rat liver systems a nd develope d the hypothesis t hat d ifferent hydroxylations are catalyzed by individual P450 enzymes [6]. Subsequently testosterone hydroxylation patterns have been utilized extensively as probes of the presence and function of individual rat liver P450s [7–11]. Testosterone hydroxylation has also been studied exten- sively with human liver microsomal P450s. Early work with liver microsomes resulted in reports of hydroxylation at the 2b,6a,6b,7a,15b,16a, and 17 positions (17-hydroxylation yields the ketone androstenedione) [12–17]. A human liver P450 was isolated that w as shown to b e t he maj or 6b-hydroxylase [18]; this P450 was originally termed nifedi- pine oxidase (P450 NF ) and subsequently named P450 3A4. Other work confirmed the role of P450 3A4 as the major enzyme involved in testosterone 6b-hydroxylation [19]. Other P450 3A subfamily enzymes (3A5, 3A7, 3A43) can also catalyze this reaction [20,21]. Testosterone hydroxylation is i n general use tod ay as one of the characteristic assays of P450 3A4, which was subsequently shown to be the most abundant P450 in human liver and small intestine [ 22] and involved in t he oxidation of approximately one-half t he drugs used today [23]. The major product is 6b-hydroxytestosterone [18,19] but several other hydroxylations occur, including those at the 2b and 1 5b positions [18,19] (Fig. 1). Not a ll of the products have been identified, however. In t he course of our investigations we noted that a peak (X) formed at a rate in the o rder 6b >2b >X 15b (hydroxylation) had not been characterized and did not correspond to any of our standards available in our set, including 2a-, 2 b-, 6a-, 6b-, 11b-, 15b-,16a-, or 16b-hydroxytestosterone or androsten- edione. We utilized an HPLC-solid phase extraction (SPE)- cryoprobe NMR/MS system and now provide a full spectral characterization o f t his p roduct f rom  6 lgof the p roduct i njected. The product is 1 b-hydroxytestosterone (Scheme 1) and is formed only by P450 3A4, of the set of human P450s examined. Experimental procedures Chemicals GC-grade acetonitrile (CH 3 CN) and HPLC-grade H 2 Ofor the separation (combined HPLC-MS-NMR) was from Merck (Darmstadt, Germany). CD 3 CN and CD 3 OD (both 99.8% deuterium-enriched) were from Deutero GmbH Correspondence to F. P. Guengerich, Department of Biochemistry and Center in Molecular T oxicology Vanderbilt Un iversity School of Medicine Nashville, Tennessee 37232–01 46, USA. Fax: +1 615 3223141, Tel.: +1 615 3222261, E-mail: f.guengerich@vanderbilt.edu Abbreviations: P450, cytochrome P450 (a lso termed h eme-thiolate P450, substrate, reduced flavoprotein: oxygen o xidoreductase); HSQC, heteronuclear spin quantum correlation; SPE, solid phase extraction. Enzymes: P450, substrate, reduced flavoprotein:oxygen oxidoreductase (EC 1.14.14.1). (Received 16 July 2004, accepted 19 August 2004) Eur. J. Biochem. 271, 3962–3969 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04339.x (Kastellaun, Germany). T estosterone (Sigma -Aldrich, St. Louis, MO, USA) was used without further purification. Hydroxytestosterone standards were purch ased from Stearoaloids (Newport, RI, USA). Enzymes Microsomes were prepared [24] from a h uman liver sample (denoted HL 97), which had been used in some previous investigations in this laboratory [25]. Recombinant P450 3A4 was used either in the form of Escherichia coli membranes in w hich both P450 3A4 and NADPH-P450 reductase were coexpressed [26] (termed Ôbicistronic mem- braneÕ system) o r m icrosomes prepared from insect cells infected with a baculovirus vector and e xpressing NADPH- P450 reductase in excess of P450 3A4 (PanVera, Madison, WI, U SA). Other human P450s were expressed together with NADPH-P450 reductase in the bicistronic membrane systems (E. coli membranes) for use [26]. Testosterone hydroxylation assays Incubations (0.5 mL total volume) were carried out with bacterial m embranes (from E. coli, bicistronic e xpression vectors, see above) containing P450 (100 pmol of P450 1A1, 1A2, 1B1, 2C9, or 2D6, or 40 pmol of P450 3A4) human liver microsomes containing 100 pmol total P450, or microsomes from baculovirus-infected insect cells contain- ing 4 pmol of P450 3A4. (Varying amounts of P450 were used because of differences in rates of the systems, in order to maintain linearity of product formation vs. time.) A typical system contained the NADPH-P450 reductase (see above), a n N ADPH-generating s ystem [ 24], a nd varying concentrations of testosterone, and was incubated for 8–10 min at 37 °C [27]. HPLC-UV HPLC-UV assays were used to quantify the rates of formation of i ndividual testosterone hyd roxylation prod- ucts. T he dichloromethane extract from each incubation was taken to dryness under a stream of N 2 . Aliquots were dissolved in 30 lL of methanol, injected into a 20-lL loop, and separated on a 4 .6 · 150 mm Phenomenex Prodigy ODS o ctadecylsilane HP LC column (C18, 3-lmparticle size, Phenomenex, Torrence, CA, U SA) with a gradient formed from solvent A (95% CH 3 CN, 5% H 2 O, v/v) and solvent B (H 2 O), u sing the schedule as follows: 0–5.5 min, 75% (v/v) solvent B; 5.5–12 min, 75% to 64% solvent B; 12–24 min, 64% (v/v) solvent B; 25–26 min, 64% to 75% (v/v) solvent B; and 26–30 min, hold at 7 5% solvent B. The pumping system was a Hitachi-L-7100 single pump ternary apparatus (Hitachi High Technologies America, San Jose, CA, USA). A 244 measurements were used, with a UV3000 rapid scanning detector (ThermoSeparations, Piscataway, NJ, U SA), and integration was done using t he software supplied by the manufacturer. HPLC-MS-NMR Sample preparation. A preparative incubation was done with the P450 3A4 bicistronic membrane preparation (200 pmol P450 3A4, total volume 5 mL) containing 500 l M testosterone a nd an NADPH-generating system [24] for 12 min at 37 °C. The reaction was extracted with dichloromethane (15 mL) an d the organic phase was washed with brine, dried over magnesium sulfate, filtered, and concentrated to dryness. The r esulting solid was dissolved in 300 lLofCD 3 OD and filtered prior to injection. HPLC (including UV) The H PLC s ystem c onsisted of an A gilent 1100 Syst em including a vacuum degasser, quaternary HPLC pump, an autosampler, and a diode array detector. Chromatographic separation was carried out on a Phenomenex Prodigy ODS3 (5-lm particle size, 4.6 · 250 mm, Phenomenex, Torrence, CA, USA). The chromatographic conditions were as follows: solvent A, CH 3 CN; solvent B, H 2 O; initial conditions 5% A/95% B (v/v), followed by a linear gradient to 95% A/5% B (v/v) over 30 min; 10-min linear gradient to 1 00% A and held for 5 min; back t o initial conditions in 0.1 min; re-equilibration for 10 min at a flow rate of 0.8 mLÆmin )1 . The peaks were detected at a wavelength o f 244 nm using a diode array detector. MS (time-of-flight) An aliquot (5%, v/v) of the eluent from the HPLC column was s plit to the mass spectrometer u sing a splitte r from LCPackings (Amsterdam, the Netherlands). The split ratio was guided to a MicroTOF mass spectrometer (Bruker Daltonic, Bremen, Germany) equipped with an orthogonal electrospray ion source. Measurements were carried out in the positive mode with a scan range from 20 to 800 mass- to-charge ratio (m/z). The capillary was set to 4500 V w ith an end-plate offset of )400 V. The nebulizer was operated Scheme 1. 1b-H ydroxytestosterone. Fig. 1. HPLC of testosterone ox idatio n product s. Ó FEBS 2004 1b-Hydroxytestosterone (Eur. J. Biochem. 271) 3963 at 1.3 bar and the dry gas was set to 4.3 LÆh )1 at a temperature of 200 °C. The capillary e xit was to 120 V with a skimmer voltage of 40 V. The h exapole RF was set to 50 Vpp (volts peak to peak) to enable the detection of smaller masses. Solid phase extraction of peaks After detection of peaks with the diode array detector, H 2 O was added using a Knauer K120 pump operated at a flow rate of 1.6 mLÆmin )1 . The flow was guided to a modified Prospekt 2 solid phase extraction unit from Bruker/Spark (Bruker Biospin, Rheinstetten, Germany/Spark Holland, Emmen, the Netherlands). Peaks were automatically trapped o n 2 · 10 mm SPE cartridges filled with Hysphere GP, a cross-linked polystyrene-divinylbenzyl copolymer (Spark Holland). After the trapping step the cartridges were automatically dried for 30 min under a stream of N 2 gas and eluted into the NMR flow cell with CD 3 OD. Cryo-NMR An Avance spectrometer equipped with a Dual Inverse 1 H/ 13 C30-lL Cryofit Probe operated at 600.13 MHz f rom Bruker BioSpin (Rheinstetten, Germany) was used for NMR investigation. The data was obtained after threefold trapping of the peak on G P cartridges a nd elution with subsequent on-line NMR analysis. The analyses were performed with three 20-lL injections, each containing 420 lg o f t otal m aterial (substrate plus other products). The chromatographic separation is shown below (Fig. 1). The spectra of testosterone and 6b-hydroxytestosterone were recorded but are not presented here. Sp ectra of the previously unidentified o xidation product were recorded eluting at 12.1 min, which was subsequently identified as 1b-hydroxytestosterone using the LC-NMR data discussed below. The trapped product was eluted in a mixture of d 6 -methanol and D 2 O with small amounts of residual CH 3 CN present. The temperature was controlled at 25 ± 0.1 °C. Chemical shifts were referenced to the water resonance a t 4 .88 p.p.m. at 25 °C . The 1D s pectrum u tilized double presaturation to minimize any residual water and methanol signals. A total of 65 536 complex data points were recorded with a sweep width of 5531 Hz and 32 scans. The data was processed with a line broadening of 0.3 Hz. Two-dimensional techniques ( 1 H- 1 HCOSY, 1 H- 1 H NOESY, and 1 H- 13 C HSQC) were also used for the structure elucidation of the trapped compounds. The parameters for the phase sen sitive (States-TPPI mode) 1 H- 1 H COSY spectra with water suppression were: spectral width, 5531 Hz, 4096 complex data points, relaxation delay 2 s, and eight scans for each of the 512 increments. The same parameters were used for phase sensitive 1 H- 1 H NOESY with H 2 O suppression on the water signal except for the number of scans (32), the number of data points (2 k) and number of increments ( 256). T he mix ing t ime was 500 ms. The 1 H- 13 C HSQC experiment was acquired in the phase sensitive mode with sensitivity enhancement, echo/anti-echo-TPPI gradient selection and adiabatic car- bon decoupling during evolution and acquisition [28–30]. Further parameters were: spectral width of 5531 Hz, 4096 data points in the 1 H dimension, 25 000 Hz with 256 data points in the 13 C dimension and a relaxation delay of 2 s. The data was processed using Bruker XWINNMR software on an SGI workstation (Silicon Graphics, Mountain View, CA, USA). The d ata was zero-filled in the acquisition dimension and linear prediction was applied in the indirect dimension. Results HPLC of testosterone oxidation products The chromatogram acquired at 244 nm (Fig. 1) showed a number of UV absorbing peaks eluting at shorter retention times when compared with the major peak (t R ¼ 23 min), which can be easily assigned to the substrate testosterone. This indicates the presence of more polar components at much lower concentrations. Several of these were known because of their coelution with standards in this and previous work. However, the peak eluted at 12.1 min did not correspond to any o f the available s tandards in our collection (2a-, 2b-, 6a-, 6b-, 11b-,15b-,16a-or16b- hydroxytestosterone or androstenedione), and the sample was submitted for HPLC-solid phase extraction-cryoprobe NMR/time-of-flight MS analysis. Mass spectrometry Preliminary HPLC-electrospray MS experiments indicated an [M + H] + ion at m/z 305, corresponding to a mono- hydroxylated testosterone product. The r esult was con- firmed in the HPLC-MS-NMR w ork with the MicroTOF instrument, yielding MH + at m/z 305.2080 (theoretical m/z for C 19 H 29 O 3 305.2111) (Fig. 2). NMR The total amount of the product estimated to have been collected for the analysis is  6 lg. The 1D 1 Hspectrumwas devoid of impurities (Fig. 3). The carbinol peak of interest was noted at d 3.95 p.p.m., observed as a multiplet. The COSY spectrum (Fig. 4) was very informative. The carbinol proton of interest (d 3.95 p.p.m.) was coupled to two protons in the d 2.5 p.p.m. region, indicating that the hydroxylation w as at either C-1 or C -7, i.e. the carbinol is coupled to either an H-2 or H-6 proton. The lack o f coupling t o the H-8 proton ( d 1.69 p.p.m.) indicates that the proton can only be at C-1. The HSQC spectrum (Fig. 5) allowed complete assign- ment of all proton-attached 13 C s ignals, confirming the basic structure. The resulting information i s presented in Table 1. The NOESY spectrum (Fig. 6) clearly indicates that the added hydroxyl g roup at C-1 m ust be b.TheH-1 carbinol proton clearly shows correlation peaks with protons established as C-2 (d 2.53, 2.46 p.p.m.), C-9 (d 1.15 p.p.m.) and the equatorial positioned proton C-11 (d 2.07 p .p.m.) (but not with the C-19 methyl group). Thus, the carbinol proton must be in the a-p osition. If the proton were b, it would b e e xpected to show a strong interaction with the C-19 methyl, as indicatedinFigs5and6. 3964 J. A. Krauser et al. (Eur. J. Biochem. 271) Ó FEBS 2004 One synt hesis of 1b-hydroxytesto sterone was found in the literature (seven steps from dihydrotestosterone benzoate) [31]. The chemical shifts presented in that paper (1, 2, 4, 17, 18 and 19 protons assigned) are simila r to ours. However, the J 1a,2a and J 1a,2b values differ. Our assignments are also consistent with those reported for 1a-hydroxytestosterone, except for the differences at and near C-1 (http://www. unibas.ch/mdpi/ecsoc-4/a0099/a0099.htm). Formation of 1b-hydroxytestosterone by recombinant P450 3A4 systems The 1b-hyroxylation of testosterone was observed in both bacterial- and baculovirus-based P450 3A4 expression sys- tems (Table 2). Rates of formation of the products were similar. The 1b-hydroxy product accounts for  5% of all testosterone products formed in both systems. The formation of 1b-hydroxytestosterone was also observed in human liver microsomes. The ratio of 1b-to 6b-hydroxylation was less than that measured with the recombinant P 450 3A4 systems due to contribution of s ome other P450s to 6b-hydroxylation (Table 3; see below also). The liver sample used (HL 97) had previously been shown to have a concentration of P450 3A4 intermediate between that of high and low individuals [25]. Testosterone hydroxylation by other human P450s Several human P450s were examined for the ability to form the individual hydroxylated testosterone products, at a Fig. 2. MS of previously unidentified tes to- sterone oxidation product. (A) Experimental spectrum. (B) Theoretical. The molecular ion (MH + 305.2080) corresponds to the formula C 19 H 29 O 3 (theoretical 305.2111). Fig. 3. COSY ( 1 H) NMR spectrum of testo- sterone oxidation product. Eight scans, 4096 · 512 acquisition matrix, 2-s relaxation delay, with water s uppression. See Table 1 for assignments. Ó FEBS 2004 1b-Hydroxytestosterone (Eur. J. Biochem. 271) 3965 single substrate concentration of 100 l M (Table 3). All of the P450s examined produced some products, but only P450 3A4 formed 1b-hydroxytestosterone. Discussion The use of a combined HPLC-MS-NMR system facilitated the c haracterization of one of the minor hydroxylation products of testosterone, with an estimated total amount of  6 lg. Spectroscopy alone yielded an unequivocal assign- ment of the product. Traces of a product designated 1a,b-hydroxytestosterone had been reported previously in rat and mouse liver systems but only on the basis of the expected t R [11,32,33]. The 1(b)-hydroxylation of androgens has been reported previously. Dodson et al. [34] r eported that microbial ppm 20 30 40 50 60 70 80 4.0 3.5 3.0 2.5 1.5 1.0 ppm ppm ppm 125 130 6.0 5.8 2.0 Fig. 4. HSQC NMR spectrum of testo- sterone oxidation product. Sixty-four scans, 4096 · 256 acquisition matrix, 2-s relaxation delay, with PFG coherence selection. The inset shows a cross-peak out of the range of the rest of the scale. Fig. 5. NOESY ( 1 H)NMR spectrum of tes- tosterone oxidation product. Thirty-two scans, 2048 · 256 acquisition matrix, 2-s relaxation delay, with water suppression. H-1 a cross- peaks are boxed. 3966 J. A. Krauser et al. (Eur. J. Biochem. 271) Ó FEBS 2004 (Xylaria sp.) oxidation of androstenedione yield ed a product identified as the 1b-hydroxy derivative. The assignment was based largely on chemical conversion to D 1,2 -dehydrotesto- serone and the optical rotation [34]. This compound was reduced to 1b-hydroxytestosterone, which has been used as a standard o r substrate in m ost subsequent work, e ither directly or by indirect comparisons. Gustafsson’s group reported 1b-hydroxylation of testosterone by human fetal liver micorosomes, using the Xylaria-derived product as a standard [35,36]. On the basis of our own s tudy, it may be speculated t hat the e nzyme re spo nsible i s P450 3A7, an enzyme closely r elated to P450 3A4 and fetal-specific (P450 3A4 is not expressed until after birth) [37]. Other work with the 1b-hydroxytestosterone derived from Xyleria oxidations [34] has yielded reports that it is a weak inhibitor of human placental aromatase (P450 19A1), the enzyme that oxidizes testosterone to 17b-estradiol, with an IC 50 value of  1m M [38]. Another report i ndicated that human placental microsomes used 1b-hydroxytestosterone  30% as efficiently as testosterone or antrostenedione [35], but apparently has not been confirmed. Very recent work on possible functions of 1b-hydroxy- testosterone has appeared in a paper published after our own work was submitted [39]. Porcine gonadal P450 19A1 (aromatase) converted testosterone to significant amounts of 1b-hydroxytestosterone, as well as 19-hydroxy- and 19-oxotestosterone and 17b-estradiol [39]. The assignment of the structure was based on (a) comparison of an MS fragmentation pattern w ith an earlier literature s pectrum [35] (going back to the original Xylaria product [34]), and (b) labilization of 3 Hfrom[1b- 3 H]-testosterone [39]. Corbin Table 1. NMR shifts (see Figs 3–5). N/A, no protons attached; Ô–Õ indicates that the sh ift was not ide ntified. 1b-OH, 1b-hydroxyt estosteron e; 2b-OH, 2b-hydroxytestosterone. Atom 1b-OH 2b-OH Testosterone 1 H d (p.p.m.) Multiplicity and coupling, J (Hz) 13 C d (p.p.m.) 1 H d (p.p.m.) 13 C d (p.p.m.) 1 H d (p.p.m.) 13 C d (p.p.m.) 1a 3.95 dd 1H, J ¼ 10.2 Hz, H 2a , 4.6 Hz, H 2b 74.3 2.31 40.46 1.64 36.22 1b – – – 1.48 40.46 2.02 36.22 2a 2.46 dd 1H, J ¼ 4.6 Hz, H 1 , 16.1 Hz, H 2b 44.05 4.11 69.06 2.18 34.33 2b 2.53 dd 1H, J ¼ 10.2 Hz, H 1 , 16.1 Hz, H 2b 44.1 N/A N/A 2.41 34.33 3 N/A N/A – N/A – N/A – 4 5.75 – 123.55 5.70 119.33 5.64 123.84 5 N/A N/A – N/A – N/A – 6a 2.31 m 1H 34.3 2.20 33.18 2.23 33.12 6b 2.54 m 1H 34.3 2.51 33.18 2.40 33.12 7a 1.00 m 1H 33.7 1.09 37.08 1.01 37.17 7b 1.90 m 1H 33.7 1.77 37.08 1.77 37.17 8 1.69 m 1H 37.35 1.90 30.55 1.59 36.19 9 1.15 ddd 1H, J ¼ 4.2 Hz, 10.3Hz, 16.2 Hz 55.65 1.39 50.90 0.90 54.76 10 N/A N/A – N/A – N/A – 11a 2.07 m 1H 24.20 1.71 22.89 1.56 21.30 11b 1.55 m 1H 24.20 1.15 22.89 1.43 21.30 12 1.09 dd 1H, J ¼ 3.9 Hz, H 11 , 13.1 Hz, H 12 37.8 0.94 34.96 0.94 32.32 12b 1.84 m 1H 37.8 1.92 34.96 1.81 32.32 13 N/A N/A – N/A – N/A – 14 0.98 m 1H 51.65 0.96 51.03 0.93 51.09 15a 1.60 m 1H 24.1 1.51 23.57 1.56 23.78 15b 1.31 m 1H 24.1 1.27 23.57 1.26 23.78 16a 1.98 m 1H 30.2 1.88 30.60 1.91 30.48 16b 1.47 m 1H 30.2 1.36 30.60 1.38 30.48 17 3.56 dd 1H, J ¼ 8.5 Hz 82.2 3.49 81.45 3.50 81.54 18 0.78 s 3H 11.4 0.71 11.36 0.72 11.39 19 1.24 s 3H 13.25 1.15 22.88 1.18 17.48 Fig. 6. Space-filling model of 1b-hydroxytestosterone. The model was produced with the program CHEM 3 DPRO v.5, CambridgeSoft Corp. (Cambridge, MA, USA). Black denotes oxygen, medium gray denotes carbon, and light gray denotes hydrogen atoms. Ó FEBS 2004 1b-Hydroxytestosterone (Eur. J. Biochem. 271) 3967 et al. [39] postulated physiological activity of 1b-hydroxy- testosterone and showed activation of the androgen recep- tor in tw o different cell lines. 1b-Hydroxytestosterone was not (enzymatically) reduced to the D 4,5 derivative. Interest- ingly, the 1b-hydroxylation reaction was not catalyzed by human P450 19A1 or by any other (tissue-specific) form of porcine P450 19A1. Although some biological activity has been demonstrated, the relevance o f 1b-hydroxytestosterone to human physiology is not clear at this point. The biological properties of 1b-hydroxytestosterone, although speculated (see above), a re currently unknown. It is of interest to note that almost all of the P450 3A4- catalyzed hydroxylations of testosterone are on the b face. This information is of interest in considerations of the steroselectivity of P450 3A4 and general considerations about the juxtaposition of the substrate in the active site, particularly in predicting sites and rates of P450 3A4 reactions deals with models based on chemical reactivity. The concept has often been proposed that P450 3A4 has a relatively open a ctive site and that reactions are influenced largely by the chemical lability of C-H bonds [40,41]. However, the striking stereochemical selectivity at each of the several hydroxylation positions would appear to argue against this and in favor of a relatively large but organized active site. Acknowledgements The authors t hank M. V. Martin and C . G. T u rvy for preparing bacterial membranes. This work was supported i n part by United States Public Health Service grants R01 CA90426, P30 ES00267, and T32 ES07028. References 1. Pa lmer, G. & R e edijk, J. (1992) Nomenclature of electron-transfer proteins: recommendation s 1989. J. Biol. Chem. 267, 665–677. 2. Dorfman, R., Cook, J.W. & H amilton, J.B. (1939) Conversion by the human of the testis hormone, testosteron e, into the urinary androgen, androsterone. J. Biol. Chem. 130, 285–295. 3. Ryan, K.J. (1958) Conversion of androstenedione to estrone by placental microsomes. Biochim. Biophys. Acta 27, 658–662. 4. Kagawa, N. & Waterman, M.R. (1995) Regulation of steroido- genic a nd related P 450s. In Cytochrome P4 50: Structure, Mechanism, and Biochemistry, 2nd edn. (Ortiz de Montellano, P.R., ed.), pp. 419–442. Plenum Press, New York. 5. Guengerich, F.P. (2003) Cytochrome P450s, drugs, and diseases. Mol. Interventions 3, 8–18. 6. Conney,A.H.,Levin,W.,Jacobsson,M.&Kuntzman,R.(1969) Specificity in the regulation of the 6b,7a,and16a-hydroxylation of testosterone by rat live r microsomes. In Microsomes and Drug Oxidations (Gi llette, J.R., ed.), pp. 279–301. Academic Press, New York. 7. Wood, A.W., Ryan, D.E., Thomas, P.E. & Levin, W. (1983) Regio- an d stereoselective met abolism o f two C19 steroids by five highly purified and reconstituted rat hepatic cytochrome P-450 isozymes. J. Biol. Chem. 258, 8839–8847. 8. van der Hoeven, T. (1984) Assay of hepatic microsomal testo- sterone hydroxylases by high-performance liquid chromatogra- phy. Anal. Biochem. 138, 57–65. 9. Dutton, D.R., McMillen, S.K., Sonderfan, A.J., Thomas, P .E. & Parkinson, A. (1987) Studies on the rate-determining factor in testosterone hydroxylation by rat liver microsomal cytochrome P-450: evidence against cytochrome P-450 isozyme: isozyme interaction. Arch. Biochem. Biophys. 255, 316–328. 10. Sond erfan, A.J., A rlotto, M.P. & Parkinson, A. (1989) Identifi- cation of the cytochrome P -450 isozyme s respon sible for te sto- sterone oxidation in rat lung, kidney, and testis: evidence that cytochrome P-450a (P450IIA1) is the physiologically important testosterone 7a-hydroxylase in rat testis. Endocrinology 125, 857–866. 11. Purdon, M.P. & Lehman-McKeeman, L.D. (1997) Im proved high-performance liquid chromatographic procedure for the separation and quantification o f hydroxytestosterone metabolites. J. Pharmacol. Toxicol. Methods 37, 67–73. 12. Gustafsson, J.A. & Lisboa, B.P. (1968) Biosynthesis of 6-beta- hydroxytestosterone from testosterone by human fetal liver microsomes. Steroids 11, 555–563. 13. Lisboa, B.P. & Gustafsson, J.A. (1969) S tudies on the metabolism of ster oids in th e foetus: bi osynthesis of 6a-hydroxytestosterone in the human foetal liver. Biochem. J. 115, 5 83–586. 14. Yaffe,S.J.,Rane,A.,Sjoqvist,F.,Boreus,L.O.&Orrenius,S. (1970) The presence of a monooxygenase system in human fetal liver microsomes. Life Sci. II, 1189–1200. Table 2. Hydroxylation of testosterone by rec ombinant human P450 3A4 systems and hu man liver microsomes. The range of substrate concen tratio ns used in most cases was 25–400 l M . Product E. coli membranes Baculovirus microsomes Human liver microsomes k cat (min )1 ) K m (l M ) k cat /K m k cat (min )1 ) K m (l M ) k cat /K m k cat (min )1 ) a K m (l M ) k cat /K m 1b-OH 4.1 ± 0.1 10 ± 2 0.40 7.1 ± 0.3 17 ± 2 0.41 1.9 ± 0.1 55 ± 9 0.035 2b-OH 11 ± 1 49 ± 6 0.23 14 ± 4 44 ± 4 0.30 12 ± 1 170 ± 40 0.072 6b-OH 78 ± 2 26 ± 3 3.0 78 ± 3 23 ± 2 3.4 88 ± 5 90 ± 10 0.98 15b-OH 3.0 ± 0.2 41 ± 12 0.072 7.1 ± 0.2 32 ± 3 0.22 8.4 ± 0.8 81 ± 20 0.10 a Based on total P450. Table 3. Testosterone hydroxylation by various human P450 enzymes. Assays were done (in triplicate) with bacterial membranes (Ôbicistro- nicÕ) containing P450 a nd NADPH-P450 reductase [26] using a single testosterone concentration of 100 l M . Ô–Õ Indic ates rate < 0.1 min )1 . P450 Rate (min )1 ) 1b-OH 2b-OH 6b-OH 15b-OH 1A1 – – 1.1 ± 0.1 – 1A2 – – 6.7 ± 0.3 – 1B1 – – 7.7 ± 0.6 – 2C9 – – 0.93 ± 0.03 – 2D6 – 0.31 ± 0.01 0.91 ± 0.01 – 3A4 4.8 ± 0.1 11 ± 0.1 83 ± 1 5.0 ± 0.1 3968 J. A. Krauser et al. (Eur. J. Biochem. 271) Ó FEBS 2004 15. Krem ers, P., Beaune, P., C resteil, T., de Graeve, J., Co lumelli, S., Leroux,J.P.&Gielen,J.E.(1981)CytochromeP-450mono- oxygenase activities in human and rat liver microsomes. Eur. J. Biochem. 118, 599–606. 16. Cresteil,T.,Beaune,P.,Kremers,P.,Flinois,J.P.&Leroux,J.P. (1982) Drug-metabolizing enzymes in human foetal liver: partial resolution of multiple cytochromes P450. Pediatr. Pharma col. (New York) 2, 199–207. 17. Distlerath, L.M. & Guengerich, F.P. (1987) Enzymology of human liver cytochrome s P-450. In Mammalian C ytochromes P-450, Vol. 1 (Guengerich, F.P., ed.), pp. 133–198. CRC Press, Boca Raton, Florida. 18. Guengerich, F .P., Martin, M.V., Beaune, P.H., Kremers, P., Wolff,T.&Waxman,D.J.(1986)Characterizationofratand human liver microsomal cytochrome P-450 forms involved in nifedipine oxidation, a prototype for genetic polymorphism in oxidative drug metabolism. J. Biol. Chem. 261, 5051–5060. 19. Waxman, D.J., At tisano, C., Guengerich, F.P. & La penson, D.P. (1988) Cytochrome P-450 steroid hormone metabolism catalyzed by human liver microsomes. Arch. Biochem. Biophys. 263,424– 436. 20. Kitada, M., Kamataki, T., Itahashi, K., Rikihisa, T. & Kanakubo, Y. (1987) Significance of cytochrome P-450 (P-450 HFLa) of human fetal livers in the steroid and drug oxidations. Biochem. Pharmacol. 36, 453–456. 21. Wrighton, S.A., Brian, W.R., Sari, M.A., Iwasaki, M., Guenge- rich, F.P., Raucy, J.L., Molowa, D.T. & VandenBranden, M. (1990) Studies on the expression and metabolic capabilities of human liver cytochrome P450 IIIA5 ( HLp3). Mol. Pharmacol. 38, 207–213. 22. Guengerich, F.P. (1995) Human cytochrome P450 enzymes. In Cytochrome P450: S tructure, Mechanism, a nd Bioc hemsitry (Orti z de Montell ano, P.R., ed.), pp. 473–535. Plenum Pre ss, New York. 23. Evans, W.E. & Relling, M.V. (1999) Pharmacogenomics: trans- lating function genomics into rational therapeutics. Science 286, 487–491. 24. Guengerich, F.P. (2001) Analysis and characterization o f enzymes and nucleic acid s. In Principles and M ethods of Toxicology (Hayes, A.W., ed.), pp. 1625–1687. Taylor & Francis, Philadelphia. 25. Guengerich, F.P. (1988) Oxidation o f 17a-ethynylestradiol by human liver cytochrome P-450. Mol. Pharmacol. 33, 500–508. 26. Parikh, A ., Gillam, E.M.J. & Guengerich, F.P. (1997) Drug metabolism by Escherichia coli expressing human cytochromes P450. Nat. Biotechnol. 15, 784–788. 27. Yamaza ki, H., Nakano, M., Imai, Y., Ueng, Y F., Guengerich, F.P. & Shimada, T. (1996) Roles of cytochrome b 5 in the oxidation of testosterone and n ifedipine by recombinant cytochrome P450 3A4 and by human liver microsomes. Arch. Biochem. Biophys. 325, 174–182. 28. Palmer, A.G., 3rd, Cavanagh, J., Wright, P.E. & Rance, M. (1991) Sensitivity improvement in proton-detected two-dimensional heteronuclear correlation NMR spectroscopy. J. Mag. Res. 93, 151–170. 29. Kay, L.E., Keifer, P. & S aarinen, T. (1992) P ure absorption gradient enhanced heteronuclear single quantum correlation spectroscopy with improved sensitivity. J. Am. Chem. Soc. 114, 10663–10665. 30. Sc hleuch er, J., Schwendinger, M., Sattler, M., Schmidt, P., Schedletzky, O., Glaser, S.J., Sorensen, O.W. & Griesinger, C. (1994) A general enhancement scheme in heteronuc lear multi- dimensional NMR e mploying pulsed field gradients. J. Biomol. NMR 4, 301–306. 31. Sharma, P.K. & Akhila, A. (1991) A facile synthesis of 1b-hy- droxytestosteron e. Indian J. Chem., Sect. B 30B, 554–556. 32. Ford, H.C., Wheeler, R. & Engel, L.L. (1975) Hydroxylation of testosteroneatcarbons1,2,6,7,15and16bythehepatic microsomal fraction from adult female C57BL/6J mice. Eur. J. Biochem. 57, 9–14. 33. Halvorson, M., Greenway, D., Eberhart, D., Fitzgerald, K. & Parkinson, A. (1990) Reconstitution of testosterone oxidation by purified rat cytochrome P450p (IIIA1). Arch. Biochem. Biophys. 277, 166–180. 34. Dodson,R.M.,Kraychy,S.,Nicholson,R.T.&Mizuba,S.(1962) Microbiological transformations. IX. The 1b-hydroxylation of androstenedione. J. Org. Chem. 27, 3159–3164. 35. Lisboa, B.P. & Gustafsson, J.A. (1968) Biosynthesis of two new steroids in the human foetal liver, 1b-and2b-hydroxytestosterone. Eur. J. Biochem. 6, 419–424. 36. Ingelman-Sundberg, M., Rane, A. & Gustafasson, J.A. (1975) Properties of hydroxylase systems in the human fetal liver active on free and sulfoconjugated steroids. Biochemistry 14, 429–437. 37. Komori, M., Nishio, K., Kitada, M., Shiramatsu, K., Muroya, K., Soma,M.,Nagashima,K.&Kamataki,T.(1990)Fetus-specific expression of a form of cytoc hrome P -450 in human livers. Biochemistry 29, 4430–4433. 38. Schwarzel, W.C., Kruggel, W.G. & Brodie, H .J. (1973) Studies on the mechanism of estrogen biosynthesis.8.Thedevelopmentof inhibitors of the enzym e system in human placenta. Endo crinology 92, 866–880. 39. Corbin, C.J., Mapes, S.M., Marcos, J., Shackleton, C.H., Mor- row, D., Safe, S., Wise, T., Ford, J.J. & Conley, A.J. (2004) Paralogues of porcine aromatase cytochrome P450: a novel hydroxylase a ctivity is associated with the survival of a duplicated gene. Endocrinology 145, 2157–2164. 40. Smith, D.A. & Jones, B.C. (1992) Speculations on the substrate structure–activity relationship (SSAR) of cytochrome P450 enzymes. Biochem. Pharmacol. 44, 2089–2098. 41. Singh, S.B., Shen, L.Q., Walker, M.J. & Sheridan, R.P. (2003) A model for predicting likely sites of CYP3A4-mediated metabolism on drug-like molecules. J. Med. Chem. 46, 1330–1336. Ó FEBS 2004 1b-Hydroxytestosterone (Eur. J. Biochem. 271) 3969 . concentration of P450 3A4 intermediate between that of high and low individuals [25]. Testosterone hydroxylation by other human P450s Several human P450s were. of human liver cytochrome P450 IIIA5 ( HLp3). Mol. Pharmacol. 38, 207–213. 22. Guengerich, F.P. (1995) Human cytochrome P450 enzymes. In Cytochrome P450: