Latanoprost-induced pigmentation in human iridial melanocytes is fibroblast dependent.

The prostaglandin F2alpha derivative, latanoprost (LT), used in glaucoma treatment, can induce pigmentation in irises of patients with hazel or heterochromatic eye colour. The mechanism by which LT induces pigmentation in the iris is not yet established, although it does not appear to induce proliferation of iridial melanocytes. The purpose of this study was to develop an in vitro model in which to investigate this mechanism. The pigmentary responses to LT and prostaglandin F(2alpha) (PGF(2alpha)) were examined in human iridial melanocytes alone or in co-culture with epithelial cells (non-ocular human epidermal keratinocytes and iris pigment epithelial cells) or mesenchymal cells (non-ocular dermal fibroblasts or iridial fibroblasts). Melanogenesis was assessed after 4 days culture with prostanoids, using dopa oxidase activity. Prostaglandin FP expression on human iridial fibroblasts and melanocytes was investigated using an immunofluorescent technique employing antibody to PGF(2alpha) receptor and RT-PCR. Iridial melanocytes did not show a convincing increase in dopa oxidase when cultured alone but in the presence of fibroblasts (ocular or non-ocular) there was a significant increase (25-30%) in dopa oxidase activity in response to 10(-7)-10(-5)m LT and PGF(2alpha). Co-culture of melanocytes with epithelial cells, while leading to increased dopa oxidase activity, did not lead to any melanogenic response to LT or PGF(2alpha). FP receptor expression was detected on fibroblasts but not iridial melanocytes by immunocytochemistry and RT-PCR. The melanocyte/fibroblast co-culture model developed in this study also showed that LT and PGF(2alpha) increased dopa oxidase activity in melanocytes from donors with brown but not blue eyes. These results suggest that LT may be inducing pigmentation in the human iris indirectly through the FP receptor on adjacent fibroblasts.

Prostaglandin ethanolamides (prostamides): in vitro pharmacology and metabolism.

We investigated whether prostaglandin ethanolamides (prostamides) E(2), F(2alpha), and D(2) exert some of their effects by 1) activating prostanoid receptors either per se or after conversion into the corresponding prostaglandins; 2) interacting with proteins for the inactivation of the endocannabinoid N-arachidonoylethanolamide (AEA), for example fatty acid amide hydrolase (FAAH), thereby enhancing AEA endogenous levels; or 3) activating the vanilloid receptor type-1 (TRPV1). Prostamides potently stimulated cat iris contraction with potency approaching that of the corresponding prostaglandins. However, prostamides D(2), E(2), and F(2alpha) exhibited no meaningful interaction with the cat recombinant FP receptor, nor with human recombinant DP, EP(1-4), FP, IP, and TP prostanoid receptors. Prostamide F(2alpha) was also very weak or inactive in a panel of bioassays specific for the various prostanoid receptors. None of the prostamides inhibited AEA enzymatic hydrolysis by FAAH in cell homogenates, or AEA cellular uptake in intact cells. Furthermore, less than 3% of the compounds were hydrolyzed to the corresponding prostaglandins when incubated for 4 h with homogenates of rat brain, lung, or liver, and cat iris or ciliary body. Very little temperature-dependent uptake of prostamides was observed after incubation with rat brain synaptosomes or RBL-2H3 cells. We suggest that prostamides' most prominent pharmacological actions are not due to transformation into prostaglandins, activation of prostanoid receptors, enhancement of AEA levels, or gating of TRPV1 receptors, but possibly to interaction with novel receptors that seem to be functional in the cat iris.

Pharmacological characterization of a novel antiglaucoma agent, Bimatoprost (AGN 192024).

Replacement of the carboxylic acid group of prostaglandin (PG) F(2alpha) with a nonacidic moiety, such as hydroxyl, methoxy, or amido, results in compounds with unique pharmacology. Bimatoprost (AGN 192024) is also a pharmacologically novel PGF(2alpha) analog, where the carboxylic acid is replaced by a neutral ethylamide substituent. Bimatoprost potently contracted the feline lung parenchymal preparation (EC(50) value of 35-55 nM) but exhibited no meaningful activity in a variety of PG-sensitive tissue and cell preparations. Its activity seemed unrelated to FP receptor stimulation according to the following evidence. 1) Bimatoprost exhibited no meaningful activity in tissues and cells containing functional FP receptors. 2) Bimatoprost activity in the cat lung parenchyma is not species-specific because its potent activity in this preparation could not be reproduced in cells stably expressing the feline FP receptor. 3) Radioligand binding studies using feline and human recombinant FP receptors exhibited minimal competition versus [(3)H]17-phenyl PGF(2a) for Bimatoprost. 4) Bimatoprost pretreatment did not attenuate PGF(2alpha)-induced Ca(2+) signals in Swiss 3T3 cells. 5) Regional differences were apparent for Bimatoprost but not FP agonist effects in the cat lung. Bimatoprost reduced intraocular pressure in ocular normotensive and hypertensive monkeys over a 0.001 to 0.1% dose range. A single-dose and multiple-dose ocular distribution/metabolism studies using [(3)H]Bimatoprost (0.1%) were performed. Within the globe, bimatoprost concentrations were 10- to 100-fold higher in anterior segment tissues compared with the aqueous humor. Bimatoprost was overwhelmingly the predominant molecular species identified at all time points in ocular tissues, indicating that the intact molecule reduces intraocular pressure.

The pharmacology of bimatoprost (Lumigan).

Bimatoprost (Lumigan) is a pharmacologically unique and highly efficacious ocular hypotensive agent. It appears to mimic the activity of a newly discovered family of fatty acid amides, termed prostamides. One biosynthetic route to the prostamides involves anandamide as the precursor. Bimatoprost pharmacology has been extensively characterized by binding and functional studies at more than 100 drug targets, which comprise a diverse variety of receptors, ion channels, and transporters. Bimatoprost exhibited no meaningful activity at receptors known to include antiglaucoma drug targets as follows: adenosine (A(1-3)), adrenergic (alpha(1), alpha(2), beta(1), beta(2)), cannabinoid (CB(1), CB(2)), dopamine (D(1-5)), muscarinic (M(1-5)), prostanoid (DP, EP(1-4), FP, IP, TP), and serotonin (5HT(1-7)). Bimatoprost does, however, exhibit potent inherent pharmacological activity in the feline iris sphincter preparation, which is prostamide-sensitive. Bimatoprost also resembles the prostamides in that it is a potent and highly efficacious ocular hypotensive agent. A single dose of bimatoprost markedly reduces intraocular pressure in dogs and laser-induced ocular hypertensive monkeys. Decreases in intraocular pressure are well maintained for at least 24 hr post-dose. Human studies have demonstrated that systemic exposure to bimatoprost is low and that accumulation does not occur. The sclera is the preferred route of accession to the eye. The high scleral permeability coefficient Papp is a likely contributing factor to the rapid onset and long-acting ocular hypotensive profile of bimatoprost.

Synthetic modification of prostaglandin f(2alpha) indicates different structural determinants for binding to the prostaglandin F receptor versus the prostaglandin transporter.

Several principles governing the binding of E series prostaglandins to EP receptors have emerged in recent years. The C-1 carboxyl group binds electrostatically to a conserved arginine residue in the seventh transmembrane segment of the receptor. Prostaglandin E analogs involving bioisosteric replacements of the carboxyl group, such as acylsulfonamide, are also active. In addition, structurally similar esters may also exhibit similar affinity, presumably by virtue of hydrogen bonding. Other regions of the substrate molecule appear to bind to other domains of EP receptors, either via hydrophobic interactions or by hydrogen bonding. Less information is available about the structural requirements for substrate binding to FP receptors. Prostanoids also bind to the prostaglandin transporter PGT. In this case, a conserved C-1 carboxyl group is critically important, since C-1 esters exhibit little affinity. Here we examined the binding of chemically diverse PGF(2alpha) structural analogs to the FP receptor and compared these with binding by the PG transporter. PGT recognized a wide range of anionic substituents. In contrast, the carboxylic acid group was essential for optimal binding to the FP receptor, since replacement by larger moieties with a similar pK(a), such as acylsulfonamide and tetrazole, substantially decreased binding affinity. Interestingly, insertion of cyclic substituents in the omega chain increased binding to the FP receptor but reduced affinity for PGT, and substitution for the 15-hydroxyl group produced only a modest reduction in FP receptor binding, but eliminated binding by PGT. Because extracellular PGF(2alpha) may compete for binding between FP receptors and PGT, these findings have implications for designing PGF(2alpha) analogs for treating disease states.

Replacement of the carboxylic acid group of prostaglandin f(2alpha) with a hydroxyl or methoxy substituent provides biologically unique compounds.

Replacement of the carboxylic acid group of PGF(2alpha) with the non-acidic substituents hydroxyl (-OH) or methoxy (-OCH(3)) resulted in an unexpected activity profile. Although PGF(2alpha) 1-OH and PGF(2alpha) 1-OCH(3) exhibited potent contractile effects similar to 17-phenyl PGF(2alpha) in the cat lung parenchymal preparation, they were approximately 1000 times less potent than 17-phenyl PGF(2alpha) in stimulating recombinant feline and human FP receptors. In human dermal fibroblasts and Swiss 3T3 cells PGF(2alpha) 1-OH and PGF(2alpha) 1-OCH(3) produced no Ca(2+) signal until a 1 microM concentration was exceeded. Pretreatment of Swiss 3T3 cells with either 1 microM PGF(2alpha) 1-OH or PGF(2alpha) 1-OCH(3) did not attenuate Ca(2+) signal responses produced by PGF(2alpha) or fluprostenol. In the rat uterus, PGF(2alpha) 1-OH was about two orders of magnitude less potent than 17-phenyl PGF(2alpha) whereas PGF(2alpha) 1-OCH(3) produced only a minimal effect. Radioligand binding studies on cat lung parenchymal plasma membrane preparations suggested that the cat lung parenchyma does not contain a homogeneous population of receptors that equally respond to PGF(2alpha)1-OH, PGF(2alpha)1-OCH(3), and classical FP receptor agonists. Studies on smooth muscle preparations and cells containing DP, EP(1), EP(2), EP(3), EP(4), IP, and TP receptors indicated that the activity of PGF(2alpha) 1-OH and PGF(2alpha) 1-OCH(3) could not be ascribed to interaction with these receptors. The potent effects of PGF(2alpha) 1-OH and PGF(2alpha) 1-OCH(3) on the cat lung parenchyma are difficult to describe in terms of interaction with the FP or any other known prostanoid receptor.

A single amino-acid substitution in the EP2 prostaglandin receptor confers responsiveness to prostacyclin analogs.

A high degree of homology between the four Gs-coupled prostaglandin (PG) receptors [EP2, EP4, prostacyclin (IP), PGD2 (DP)] and the four Gq/Gi-coupled receptors [EP1, EP3, PGF2alpha (FP), thromboxane A2 (TP)] suggests that prostaglandin receptors evolved functionally from an ancestral EP receptor before the development of distinct binding epitopes. If so, ligand selectivity should be determined by a limited number of amino acids. EP2 receptor transmembrane domain residues that are similar to those in the EP4 receptor but differ from those in the IP receptor were mutated to the corresponding IP receptor residue. Activation of the mutant receptors by PGE2 (EP2 ligand), iloprost (stable prostacyclin analog), and PGE1 (EP2/IP ligand) was determined using a cAMP-dependent reporter gene assay. A Leu304-to-tyrosine substitution in the seventh transmembrane domain enhanced iloprost potency approximately 100-fold. A glycine substitution at Ser120 in the third transmembrane domain had no effect on drug potency but improved the response of the Tyr304 mutant. The potency of the natural prostaglandins PGF2alpha and PGD2 was not enhanced by the mutations. In contrast, the potency of all prostaglandins was reduced 10- to 100-fold when arginine 302, which is thought to be a counterion for the prostaglandin carboxylic acid, was mutated. Thus, a single amino acid change resulted in a selective gain of function for iloprost, which is consistent with the proposed phylogeny of the prostaglandin receptors.

Elucidation of amino acid residues critical for unique activities of rabbit cytochrome P450 2B5 using hybrid enzymes and reciprocal site-directed mutagenesis with rabbit cytochrome P450 2B4.

The molecular basis for the unique activities of rabbit cytochrome P450 2B5, compared with the highly related rabbit 2B4, was investigated using hybrid enzymes and site-directed mutagenesis. Alterations in androstenedione hydroxylase profiles observed with 2B4-2B5 hybrids expressed in COS cells showed that key amino acids are present in both the N-terminal ApaI fragment (codons 1-122) and an internal SstI fragment (codons 220-393). Based on these results, data obtained with other cytochromes P450 2B, and correlation to the six substrate recognition sites proposed by Gotoh (1992, J. Biol. Chem. 267, 83-90), reciprocal 2B4-2B5 mutants were constructed at positions 114, 294, 363, and 367. Wild-type and mutant enzymes were expressed in Escherichia coli, and the oxidation of a number of substrates was analyzed. All residues studied were found to be important for regio- and stereospecificity of androstenedione hydroxylation. Mutations at these positions also caused alterations in the oxidation of progesterone, benzyloxyresorufin, pentoxyresorufin, ethoxycoumarin, and benzphetamine, with the magnitude and direction of the changes dependent upon the enzyme, residue, and substrate. Major changes in activity were consistently observed upon mutation of residues 114 and 294 in both enzymes, and some of these alterations were interpreted with the help of a 3-D model of P450 2B4. For example, in the 2B4 Ile-114--> Phe mutant, Phe prevents androstenedione from assuming a 16 beta-binding orientation and also hinders binding of benzyloxyresorufin, leading to a loss of activity. Conversely, the presence of Phe-114 stabilizes a 16 alpha-binding orientation of androstenedione, resulting in an increase in this activity.

Molecular characterization and ocular hypotensive properties of the prostanoid EP2 receptor.

The cloning of the genes that encode for prostaglandin (PG) receptors has resolved much of the complexity and controversy in this area by confirming the classification proposed by Coleman, et al. Two issues that remained unresolved were (1) the inability of the EP2 agonist butaprost to interact with the cloned putative EP2 receptor and (2) molecular biological confirmation of a fourth PGE2-sensitive receptor, which was pharmacologically designated EP4. In order to provide clarification, we attempted to clone further PGE2-sensitive receptors. By using a cDNA probe that encodes for the human EP3A receptor, a cDNA clone that encoded for a novel PGE2-sensitive receptor was obtained by screening a human placenta library. This cDNA clone was transfected into COS-7 cells for pharmacological studies. The cDNA clone obtained from human placenta had only about 30% amino acid identity with cDNAs for other PG receptors, including those that encode for the previously proposed murine and human EP2 receptors. Radioligand binding studies on the novel EP receptor expressed in COS-7 cells revealed that selective EP2 agonists such as butaprost, AH 13205, AY 23626 and 19(R)-OH PGE2 all competed with 3H-PGE2 for its binding sites, whereas selective agonists for other PG receptor subtypes had minimal or no effect. This receptor was coupled to adenylate cyclase and EP2 agonists caused dose-related increases in cAMP. It appears that the cDNA described herein encodes for the pharmacologically defined EP2 receptor. Ocular studies revealed that AH 13205 decreased intraocular pressure in normal and ocular hypertensive monkeys by a mechanism that does not appear to involve inhibition of aqueous humor secretion.

Cloning of a novel human prostaglandin receptor with characteristics of the pharmacologically defined EP2 subtype.

A cDNA that when expressed has the binding and functional characteristics of the pharmacologically defined EP2 prostaglandin (PG) receptor [Cardiovasc. Drug Rev. 11:165-179 (1993)] has been cloned from a human placenta library. This clone, known as Hup-4, encodes a protein of 358 amino acids that has only approximately 30% overall identity with other PG receptors, including mouse and human clones that have been designated as EP2 receptors [J. Biol. Chem. 268:7759-7762 (1993); Biochem. Biophys. Res. Commun. 197:263-270 (1993)]. In COS-7 cells transfected with Hup-4, PGE2 stimulated the formation of cAMP with an EC50 of approximately 50 nM. The EP2-selective agonists AH13205 and butaprost were also active, with EC50 values in the range of 2-6 microM. The order of potency of PGs for competition with binding of [3H]PGE2 to membranes prepared from COS-7 cells transfected with Hup-4 was PGE2 > or = PGE1 > 16,16-dimethyl-PGE2 > or = 11-deoxy-PGE1 > butaprost > AH13205 > 19(R)-OH-PGE2. Natural PGs and analogues that are selective for the FP (PGF2a), DP (PGD2), EP1 (sulprostone), EP3 (MB 28767), and EP4 (1-OH-PGE1) receptors were inactive or competed poorly with the binding of [3H]PGE2 (< 50% displacement of specific binding at 10 microM). Northern blot analysis showed the presence of a Hup-4 message of approximately 3.1 kilobases in mRNA from human lung and placenta. Reverse transcription-polymerase chain reaction studies also indicated that Hup-4 is probably expressed in human uterus and in HL-60 (human promyelocytic leukemia) cells. Our findings suggest that Hup-4 encodes the pharmacologically defined EP2 receptor, whereas the mouse and human cDNAs previously classified as EP2 may represent another EP receptor subtype or the recently defined EP4 subtype [Prostaglandins 47:151-168 (1994)].

Site-directed mutagenesis of putative substrate recognition sites in cytochrome P450 2B11: importance of amino acid residues 114, 290, and 363 for substrate specificity.

Eleven amino acid residues unique to dog cytochrome P450 (P450) 2B11, compared with rat 2B1 and 2B2, rabbit 2B4 and 2B5, and mouse 2B10, in the putative substrate recognition sites [J. Biol. Chem. 267:83-90 (1992)] were mutated to the residues found in 2B1 or 2B5. The mutants were expressed initially in COS cells and screened for activity toward androstenedione and 2,2',4,4',5,5'-hexachlorobiphenyl (245-HCB). P450 2B11 mutants V107I, M199L-N200E-V204R, V234I, A292L, Q473R, and I475S showed no differences from wild-type P450 2B11 in metabolite profiles with either substrate. Mutants V114I, D290I, and L363V exhibited altered androstenedione metabolite profiles and were expressed in Escherichia coli for further study with androstenedione, testosterone, 7-ethoxycoumarin, (R)- and (S)-warfarin, and 245-HCB. With V114I, hydroxylation of steroids and warfarin and 2-hydroxylation of 245-HCB were decreased, whereas 7-ethoxycoumarin O-dealkylation and 3-hydroxylation of 245-HCB were unaltered. For D290I, activities toward all substrates were decreased, except for 16 beta-hydroxylation of testosterone. The activity of L363V was increased 5-6-fold for 16 alpha-hydroxylation of androstenedione and testosterone but was decreased to 40-50% of wild-type activity with 7-ethoxycoumarin and warfarin and to 6-8% of control for 2-hydroxylation of 245-HCB. Alignment of P450 2B11 with P450 101 and super-imposition of the 11 mutated 2B11 residues on a P450 101 three-dimensional model suggest that only residues 114, 290, and 363 represent substrate contact residues, in excellent agreement with the experimental results. The data indicate the importance of the three residues 114, 290, and 363 in substrate specificity and regio- and stereoselectivity of P450 2B11 and also demonstrate that the effects of the mutations vary considerably with different substrates.

Molecular cloning and expression of human EP3 receptors: evidence of three variants with differing carboxyl termini.

1. The polymerase chain reaction (PCR) was used in combination with plaque hybridization analysis to clone four variants of the EP3 prostaglandin receptor from a human small intestine cDNA library. 2. Three of these variants, i.e. the EP3A, EP3E and EP3D, share the same primary amino acid sequence except for their carboxyl termini, which diverge from one another at the same point, approximately 10 amino acids away from the end of the seventh membrane spanning domain of the receptor. The fourth variant (EP3A1) has a nucleotide coding sequence identical to EP3A but has a completely different 3' untranslated sequence. 3. The carboxyl termini of the three isoforms differ most obviously in length with the EP3A being the longest (41 amino acids) and the EP3E being the shortest (16 amino acids). They also differ in content with the EP3A containing 9 serine and threonines in its carboxyl terminus and the EP3E none. 4. Transient expression in eukaryotic cells showed that the human EP3 receptor variants had similar but not identical radioligand binding properties and differed in their functional coupling to second messenger pathways. Up to 3 pmol mg-1 protein of [3H]-prostaglandin E2 binding could be obtained with more than 95% specific binding. Using a reporter gene assay, as a measure of intracellular cyclic AMP levels, the EP3A coupled more efficiently to the inhibition of adenylyl cyclase than did the EP3E. 5. PCR was used to confirm the presence of mRNAs encoding the four human EP3 receptor variants in tissues of the human small intestine, heart and pancreas. These findings indicate that the EP3 receptor variants identified here are likely to be expressed in tissues. The differences in the carboxyl termini at the protein level, and in the 3' untranslated regions at the mRNA level, could be profound in terms of the regulation and functional coupling of these receptor isoforms.

Cloning, sequencing, and functional studies of phenobarbital-inducible forms of cytochrome P450 2B and 4B expressed in rabbit kidney.

Expression of several forms of cytochrome P450 2B and of P450 4B1 in rabbit kidney was investigated by cloning from cDNA libraries constructed with renal mRNA from animals treated with phenobarbital. Isolation and sequencing of several cDNAs demonstrated that: (i) cytochrome P450 2B-B0 can be found in rabbit kidney along with a newly discovered form of P450 2B termed "P450 2B-Bx." P450 2B-Bx differs from P450 2B-B0 at 25 nucleotide positions and at four positions in the derived sequence of 491 amino acids. Two previously identified forms of cytochrome P450 2B, 2B-B1 and 2B-B2, are not detected in rabbit kidney. cDNA encoding cytochrome P450 4B1 was also cloned from the kidney library and found to be identical in sequence to cDNAs cloned from rabbit hepatic and pulmonary libraries. Analysis of renal mRNA indicates that forms of cytochrome P450 2B and P450 4B1 are expressed in a number of species but induced by phenobarbital in rabbit only (4B1) or rabbit and hamster (2B). Relatively high levels of mRNA related to P450 4B1 were detected in samples from untreated and phenobarbital-treated mice. Analysis of protein by immunoblotting was less sensitive but produced results consistent with those obtained by analysis of mRNA; protein related to cytochrome P450 2B was detected in renal microsomal samples from rabbit and hamster (phenobarbital > untreated), and protein related to P450 4B1 in samples from rabbits (phenobarbital > untreated) and mice (phenobarbital = untreated). The four forms of cytochrome P450 2B were expressed in COS-7 cells, and their activities were evaluated with androstenedione, testosterone, and 7-ethoxycoumarin as substrates. Three of the P450 2B forms, B0, B1, and Bx, metabolize these substrates in a manner characterized by preference for 16 beta-hydroxylation of androstenedione, low testosterone 16-hydroxylation, and high ethoxycoumarin O-deethylation. The fourth form, P450 2B-B2, is catalytically distinct from the others, with activities characterized by high androstenedione 16 alpha- and 15 alpha-hydroxylation and high testosterone 16-hydroxylation. Since P450 2B-B2 is catalytically distinct from the other forms, the metabolic profiles of phenotypes that include P450 2B-B2 might differ significantly from those of phenotypes that lack P450 2B-B2.

Hybrid enzymes for structure-function analysis of cytochrome P-450 2B11.

Previous work has shown that P-450 2B11 is responsible for the unique ability of dogs to metabolize and eliminate certain highly-chlorinated biphenyls such as 2,2',4,4',5,5'-hexachlorobiphenyl (245-HCB), whereas the related P-450 2B forms in rat and rabbit are unable to metabolize the compound to any significant degree. To determine the structural basis for this functional diversity, hybrid enzymes were generated. Success with this approach required a careful choice of second enzyme and common substrate with which to assess the functional integrity of the hybrid proteins. The choices of P-450 2B5 from rabbit as the second enzyme and androstenedione as the substrate were based in part on the finding that P-450 2B11 and P-450 2B5 hydroxylate androstenedione with similar overall activities but distinct profiles. Enzymatic studies with eight hybrid enzymes provided evidence for two regions of P-450 2B11 and 2B5, between residues 95-239 and 240-370, that appear to be involved in defining substrate specificity for androstenedione, and three regions of P-450 2B11, between residues 95-239, 240-370, and 371-494, that contain amino acids necessary for metabolism of 245-HCB. This deliberate approach to the creation of hybrid cytochromes P-450 has generated a series of enzymes that will be central to further structure-function studies of the cytochromes P-450 2B.

Role of residue 478 as a determinant of the substrate specificity of cytochrome P450 2B1.

Two allelic variants and eight site-directed mutants of cytochrome P450 2B1 differing at residue 478 have been expressed in COS cells and assayed for androstenedione hydroxylase activities. The 478Gly and 478Ala variants and five mutants (Ser, Thr, Val, Ile, and Leu) exhibited 16 beta-OH:16 alpha-OH ratios ranging from 0.7 to 9.3, whereas the Pro, Glu, and Arg mutants were expressed but inactive. The seven samples active toward androstenedione also exhibited testosterone 16 beta-OH:16 alpha-OH ratios ranging from 0.4 to 2.3. With both steroids, the Gly variant had the highest 16 beta-hydroxylase activity, and the 16 beta-OH:16 alpha-OH ratio increased with the size of aliphatic size chains (Ala, Val, and Ile/Leu). The highest ratio of androgen 15 alpha:16-hydroxylation was observed with the Ser mutant. On the basis of previous work indicating decreased susceptibility of the 478Ala variant in liver microsomal and reconstituted systems to inactivation by chloramphenicol analogs, methodology was refined for monitoring enzyme inactivation in COS cell microsomes. The Gly and Ala variants were inactivated by chloramphenicol with similar rate constants, whereas the Ser and Val mutants were inactivated more slowly, and the Leu mutant was refractory. Only the Gly variant was inactivated by the chloramphenicol analog N-(2-p-nitrophenethyl)chlorofluoroacetamide. Thus, the side chain of residue 478 appears to be a major determinant of enzyme inactivation as well as of androgen hydroxylation.(ABSTRACT TRUNCATED AT 250 WORDS)

Functional expression of mammalian cytochromes P450IIB in the yeast Saccharomyces cerevisiae.

Three mammalian cytochromes P450 from the IIB subfamily, P450IIB11 from canine and P450IIB4 and P450IIB5 from rabbit, have been expressed in the yeast Saccharomyces cerevisiae by use of an autonomously replicating vector containing the galactose-inducible gal10 promoter. Cytochromes P450IIB4 and P450IIB5 are closely related proteins, with only 11 amino acid substitutions between them. P450IIB11 is a homologous protein, likely orthologous with IIB4 or IIB5, with 102 amino acid substitutions compared with the P450IIB4 protein and 106 compared with the P450IIB5 protein. The expressed proteins are functional in yeast microsomes, exhibiting activity toward androstenedione, 7-ethoxycoumarin, and, in some cases, progesterone. Expressed cytochromes P450IIB4 and P450IIB11 hydroxylate androstenedione with regio- and stereoselectivity characteristic of the purified, reconstituted proteins. A striking difference in the androstenedione metabolite profiles of IIB4 and IIB5 was observed, with IIB4 producing almost exclusively the 16 beta-hydroxy metabolite and IIB5 producing the 16 alpha-hydroxy and 15 alpha-hydroxy products. This is the first time that 15 alpha-hydroxylase activity has been associated with IIB4/IIB5. This activity has also been detected in liver microsomes from some, but not all, individual phenobarbital-induced rabbits tested and is largely inhibited by anti-rabbit P450IIB immunoglobulin G. These studies illustrate the utility of the yeast expression system for defining catalytic activities of individual mammalian cytochromes P450 and identifying new marker activities that can be utilized in liver microsomes.

Molecular basis for a functionally unique cytochrome P450IIB1 variant.

Liver microsomes from phenobarbital-treated rats of four inbred strains expressing distinct allelic variants of cytochrome P450IIB1 were analyzed. The Wistar Munich (WM) strain exhibited 5- to 10-fold lower androstenedione 16 beta-hydroxylase activity (a specific P450IIB1 marker) than the Lewis, Wistar Kyoto, and Wistar Furth strains. The androstenedione 16 beta-hydroxylase in the WM liver microsomes was refractory to inactivation by N-(2-p-nitrophenethyl)chlorofluoroacetamide, a selective P450IIB1 inactivator in the other three strains. Purified P450IIB1-WM was insensitive to the inactivator and exhibited 5-fold lower androstenedione 16 beta-hydroxylase, testosterone 16-hydroxylase, and 7-ethoxycoumarin deethylase activities but the same benzphetamine demethylase activity and slightly higher androstenedione 16 alpha-hydroxylase activity than a P450IIB1 purified from outbred Sprague-Dawley rats, which appears to correspond to the form in Lewis rats. The stereoselectivity of androstenedione 16-hydroxylation catalyzed by P450IIB1-WM (16 beta-OH:16 alpha-OH = 1.4) is thus distinct from that (16 beta-OH:16 alpha-OH = 12-15) of other P450IIB1 preparations described. A cDNA encoding P450IIB1-WM was cloned and sequenced, revealing a single amino acid substitution (Gly-478----Ala) compared with the published sequence (Fujii-Kuriyama, Y., Mizukami, Y., Kawajiri, K., Sogawa, K., and Muramatsu, M. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, 2793-2797). Heterologous expression of P450IIB1 and P450IIB1-WM confirmed the striking difference in androstenedione metabolite profiles, strongly implicating the involvement of Ala-478 in defining the distinctive catalytic properties of P450IIB1-WM.

Characterization and comparative structural features of the gene for human interstitial retinol-binding protein.

We have cloned the gene for human interstitial retinol-binding protein (IRBP) and compared its nucleotide sequence with that of the corresponding cloned cDNA. The human IRBP gene is approximately 9.5 kilobase pairs (kbp) in length and consists of four exons separated by three introns. The introns are 1.6-1.9 kbp long. The gene is transcribed by photoreceptor and retinoblastoma cells into an approximately 4.3-kilobase mRNA that is translated and processed into a glycosylated protein of 135,000 Da. The amino acid sequence of human IRBP can be divided into four contiguous homology domains with 33-38% identity, suggesting a series of gene duplication events. In the gene, the boundaries of these domains are not defined by exon-intron junctions, as might have been expected. The first three homology domains and part of the fourth are all encoded by the first large exon, which is 3,180 base pairs long. The remainder of the fourth domain is encoded in the last three exons, which are 191, 143, and approximately 740 base pairs long, respectively. This unusual structure is shared with the bovine IRBP gene. A large (1.7 kbp) fragment appears to have been lost from the 3'-noncoding region of the last human exon. We conclude that the human and bovine genes have similar evolutionary histories.

Comparison of the aspartate transcarbamoylases from Serratia marcescens and Escherichia coli.

The aspartate transcarbamoylases (ATCase, EC 2.1.3.2) of Escherichia coli and Serratia marcescens have similar dodecameric enzyme structures (2(c3):3(r2] but differ in both regulatory and catalytic characteristics. The catalytic cistrons (pyrB) of the ATCases from E. coli and S. marcescens encode polypeptides of 311 and 306 amino acids, respectively; there is a 76% identity between the DNA sequences and an overall amino acid homology of 88% (38 differences). The regulatory cistrons (pyrI) of these ATCases encode polypeptides of 153 and 154 amino acids, respectively, and there is a 75% identity between the DNA sequences and an overall amino acid homology of 77% (36 differences). In both species, the two genes are arranged as a bicistronic operon, with pyrB promoter proximal. A comparison of the deduced amino acid sequences reveals that the active site and the allosteric binding sites, as well as most of the intrasubunit interactions and intersubunit associations, are conserved in the E. coli and the S. marcescens enzymes; however, there are specific differences which undoubtedly contribute to the catalytic and regulatory differences between the enzymes of the two species. These differences include residues that have been implicated in the T-R transition, c1:r1 interface interactions, and the CTP binding site. A hybrid ATCase assembled in vivo with catalytic subunits from E. coli and regulatory subunits from S. marcescens has a 6 mM requirement for aspartate at half-maximal saturation, similar to the 5.5 mM aspartate requirement of the native E. coli holoenzyme at half-maximal saturation. However, the heterotropic response of this hybrid enzyme is characteristic of the heterotropic response of the native S. marcescens holoenzyme: ATP activation and CTP activation. Activation by both allosteric effectors indicates that the heterotropic response of this hybrid holoenzyme (Cec:Rsm) is determined by the associated S. marcescens regulatory subunits.