CYP1B1, a developmental gene with a potential role in glaucoma therapy.

The association of CYP1B1 gene alterations in primary congenital glaucoma individuals has been known for about a decade. Recent evidence has shown the involvement of CYP1B1 mutations in a number of forms of glaucoma and anterior segment disorders. This suggests a wide role for CYP1B1 in ocular physiology. Histochemical studies of eyes from individuals with primary congenital glaucoma revealed abnormalities in the anterior chamber angle, the region between the cornea and the iris, containing the trabecular meshwork. The cells of the trabecular meshwork serve as a filter to allow drainage of the aqueous humour, the fluid formed by the ciliary body that fills the anterior chamber. Mutations in CYP1B1 that affect its activity have frequently been shown to influence development of the trabecular meshwork, and it is thought that CYP1B1, a monooxygenase, acts to form or degrade some endobiotic compound that is necessary for proper development of the filtering structures. The rapidly developing area of stem cell research suggests a potential therapeutic approach for glaucomas resulting from deleterious mutations in CYP1B1, that is, the transfer of stem cells, differentiated to a specific lineage, containing wild-type CYP1B1 to specific regions of the eye, where they will develop into normal cells of that region and rectify the defect.

Physiological significance and expression of P450s in the developing eye.

Expression of 10 CYP orthologs (Families 1-3) in developing mouse conceptus is constitutive. These forms have specific temporal and spatial expression. Studies on CYP1B1 indicate its requirement for normal eye development, both in human and mouse. The distribution of the enzyme in the mouse eye is in three regions, which may reflect three different, perhaps equally important, functions in this organ. Its presence in the inner ciliary and lens epithelia appears to be necessary for normal development of the trabecular meshwork and its function in regulating intraocular pressure. Its expression in the retinal ganglion and inner nuclear layers may reflect a role in maintenance of the visual cycle. Its expression in the corneal epithelium may indicate a function in metabolism of environmental xenobiotics.

Comparative expression profiling of 40 mouse cytochrome P450 genes in embryonic and adult tissues.

This study is the first systematic investigation of the gestational age-dependent and adult tissue-specific expression patterns of each known mouse CYP family (40 genes) using normalized cDNA panels and uniform reverse transcriptase polymerase chain reaction-based assays. Twenty-seven of the P450s were constitutively expressed during development. The number gradually increased through the phases of gastrulation E7 (n=14), neural patterning and somitogenesis E11 (n=17), organogenesis E15 (n=20), and fetal period E17 (n=21). Cyp2s1, Cyp8a1, Cyp20, Cyp21a1, Cyp26a1, Cyp46, and Cyp51 were detected in each of the four stages studied. Members of family CYP1 demonstrated complex, nonoverlapping embryonic patterns of expression, indicating that Cyp1a1 and Cyp1a2 may not compensate for Cyp1b1 deficiency associated with abnormal eye development. Multiple Cyp forms were found to be constitutively expressed in each of the adult tissues studied: liver (n=31), kidney (n=30), testis (n=26), lung (n=24), and heart (n=13). The tissue-specific P450-expression profiles reported in this study provide a reference for more focused analysis of the tissue-specific and developmental functions of the cytochrome P450 monooxygenases.

Effect of two mutations of human CYP1B1, G61E and R469W, on stability and endogenous steroid substrate metabolism.

CYP1B1 is linked to normal eye development by the disease phenotype, primary congenital glaucoma (PCG). CYP1B1 mRNA was expressed in a number of human fetal tissue cDNA libraries, supporting the suggestion of its involvement in tissue development. Highest expression levels were found in thymus and kidney, followed by spleen. A considerably lower level was observed in lung, cardiac and skeletal muscle. No expression was detected in liver or brain. CYP1B1 is able to metabolize steroid hormones. Testosterone was a poor substrate and activity with progesterone was 6-fold higher, but estradiol was the preferred substrate, exhibiting a greater metabolite profile with CYP1B1 than with CYP1A2. Major metabolites were A-ring hydroxylations (75-80%). Others were 15alpha-, 6alpha-, 16alpha- and 6beta-hydroxy metabolites. Two CYP1B1 mutations found in families with the PCG phenotype in which incomplete penetrance is seen were expressed in Escherichia coli. G61E, a hinge region mutation, and R469W, a heme region mutation, were shown to code for holoenzymes. G61E had greatly diminished stability, while the R469W holoenzyme, if anything, was stabilized. Both mutants showed compromised catalytic activity. The extents of isomeric site activity diminution were not proportional, resulting in alterations in the metabolite profiles. The results suggest that if a metabolite of CYP1B1 or elimination of a metabolite by CYP1B1 is necessary for normal embryonic or fetal tissue development, the appearance of these two mutations could result in developmental abnormalities. The altered activities of the mutants and ability of CYP1B1 to respond to external challenge may be the basis for the observed incomplete penetrance.

Roles of cytochrome p450 in development.

Cytochrome P450 (CYP) forms are ubiquitous in nature, appearing in almost all phyla, with many forms appearing in any organism. About 50 different forms have been identified in man, and some of these are found in the embryo, some showing temporal dependence. Many of the forms of cytochrome P450 present in one species have homologues in other species. For example, CYP1A2 is present in many species, including man, rabbits, rodents, fish and fowl. The amino acid sequence identity of these homologues is often in excess of 70%. CYP26, too, has more than 61% identity in amino acid sequence between fish, fowl and mammals. In view of the high degree of conservation of sequence as well as of enzymatic activities, it is only reasonable to assume that such strong conservation of sequence also reflects a conservation of function. Since the 'xenobiotic metabolizing' enzymes predate the production of the many xenobiotics they are known to metabolize, perhaps it is reasonable to consider endobiotics as natural substrates for their metabolism. Of the identified forms of cytochrome P450 that are present in embryonic tissue, we consider the possibility that they serve the organism in support of morphogenesis of the embryonic tissue. These forms may either function to generate morphogenic molecules or to keep regions free of them, thereby creating temporal and spatial regions of morphogen action and supporting region-specific changes in cells. One known morphogen, retinoic acid, has the enzymes retinal dehydrogenase (RALDH) and CYP26 maintaining its actions, the former responsible for its generation and the latter for its elimination. Another form of cytochrome P450, CYP1B1 appears also to be involved in differentiation of tissue, with its absence resulting in primary congenital glaucoma. However, the nature of the morphogen it may maintain still remains to be elucidated.

Charge-dependent sidedness of cytochrome P450 forms studied by quartz crystal microbalance and atomic force microscopy.

Quartz crystal microbalance (QCM) resonance measurements were used to examine the surface charge characteristics of cytochrome P450 forms and the influence of charge on the docking of redox partners like cytochrome b5. The distal surface of cytochrome P450 (CYP)101 (pI = 4.5), relative to the heme, is fairly anionic, as is the proximal surface. The latter, however, also has two cationic clusters. A considerably greater extent of CYP101 binding was seen to the cationic, polyethylene-surfaced resonators. CYP2B4 (pI = 8.5) preferentially bound to the polyanionic, polystyrene sulfonate-surfaced resonators. Cytochrome b5 is an acidic protein that had a preferential binding to the poly(ethyleneimine (PEI)-surfaced resonators. When binding to CYP2B4-surfaced films, cytochrome b5 preferentially bound to those cytochrome P450 molecules that were adsorbed to cationic (PEI) films. It is suggested that adsorption of CYP2B4 to an anionic poly(styrenesulfonate) (PSS) surface is with cationic clusters that include the cytochrome b5 docking domain. This diminishes the extent of docking of the cytochrome b5. In contrast, when CYP2B4 is adsorbed to a cationic film the proximal surface with the cytochrome b5-docking site is available for cytochrome b5 binding. A film of the polycation PEI was adsorbed to the silver QCM surface. It formed polymer islands when viewed with atomic force microscopy. Polyanionic PSS was adsorbed intermittently with the PEI. By the third and fourth layer of polyions the polymer islands were essentially merged and protein adsorption as a fourth or fifth layer formed a nearly continuous film. CYP101 was seen to adsorb as globules with a molecular diameter of about 10 nm. CYP2B4 adsorbed to the polyionic films had a slightly elliptical globular shape, also with a molecular diameter of about 10 nm.

Enhanced expression of CYP1B1 in Escherichia coli.

Conditions for the optimal expression of the human CYP1B1 hemoprotein in Escherichia coli have been investigated. CYP1B1 cDNA was prepared from a retinal cDNA template and used to generate cDNA fragments with modified 5'-sequences reported to allow enhanced expression in E. coli DH5alpha. Plasmids were constructed, using the pCWori+ expression vector and were used to examine necessity for thiamine, delta-aminolevulinic acid (ALA), and IPTG. The optimal shaking speed in an orbital incubator was 150 rpm at 30 degrees C. Higher speeds resulted in increased cell death and lower speeds resulted in lower expression of cytochrome P450. IPTG was necessary for this expression system, which makes use of the lac repressor, but levels above 0.5 mM were without additional benefit. We were able to show thiamine to be unnecessary in this expression system, although included by others expressing CYP1B1. ALA has been reported to enhance expression of several different forms of cytochrome P450. We examined the dependence of CYP1B1 expression on ALA. The expression proved to be highly dependent upon this heme precursor, with levels of CYP1B1 increasing approximately 20-fold, to 920 nmol/l in the presence of up to 2.5 mM ALA. The question of whether heme synthesis and apoprotein synthesis were coupled was then investigated. It could be shown that although heme synthesis was not limiting (CYP101 holoenzyme expression in the absence of ALA was four times higher than the ALA-supported CYP1B1 holoenzyme expression), it was necessary for optimal expression of CYP1B1. CYP1B1 protein synthesis appears to be coupled to heme precursor availability, as seen by SDS-PAGE, because in the absence of heme precursor apocytochrome P450 1B1 does not accumulate.

Applications of polyion films containing biomolecules to sensing toxicity.

This paper describes several applications of polyion-biomolecule films on electrodes related to future development of in vitro chemical toxicity sensors. In the first example, composite films of DNA and ionomers cast onto pyrolytic graphite (PG) electrodes are shown to be useful for detecting DNA damage during incubation with the carcinogen styrene oxide at pH 5.5. Single electrodes can be used to estimate relative damage rates by derivative square wave voltammetry. Films containing the ionomer Nafion gave better reproducibility than another ionomer, Eastman AQ38. In the second example, films containing redox proteins myoglobin (Mb) and cytochrome (cyt) P450cam were constructed in alternate layers with polyions including DNA on rough PG electrodes. Films with reversible protein FeIII/FeII electrochemistry with up to 7 electroactive layers were made. Amounts of electroactive protein on rough PG that were 7 to 17-fold larger than in similar films on smooth gold were achieved because many more layers were electroactive. Films of Mb/DNA also showed oxidation peaks after short incubations with styrene oxide that may be attributable to DNA damage. Results are relevant to the future design of enzyme-DNA films which convert pollutants and drugs to reactive metabolites, followed by electrochemical detection of the resulting DNA damage.

Substrate influence on interaction between cytochrome P450 and cytochrome b5 in microsomes.

We have been able to demonstrate that cytochrome b4 interacts closely with cytochrome P450 in the microsomal membrane, and that substrates can serve to order the interaction: Using the water-soluble carbodiimide EDC we could crosslink cytochrome b5 and CYP2B4 in microsomes from phenobarbital-treated rabbits and cytochrome b5 and CYP1A2 in microsomes from beta-naphthoflavone-treated animals. The substrate benzphetamine increased the specific interaction between cytochrome b5 and CYP2B4, decreasing the formation of higher molecular weight oligomeric complexes with cytochrome b5. In contrast, no ordering of the interactions were obtained on addition of 7-ethoxycoumarin, a substrate of CYP1A2, or of benzphetamine to microsomes of beta-naphthoflavone-treated animals in the presence of EDC. Of interest, although evidence could be shown for complementary charge-pairing between cytochrome b5 and a number of other microsomal proteins in the membranes, and while the extent of CYP1A2 and CYP2B4 interaction with cytochrome b5 each exceeded 30% in the presence of substrate, no significant complexation of the P450s was obtained with any other microsomal proteins.

The isolation and comparison of multiple forms of CYP2B from untreated and phenobarbital-treated rabbit liver microsomes.

At low levels of phenobarbital induction two forms of isoenzyme 2 (LM2; CYP2B4) were obtained during purification of cytochrome P450 from rabbit liver microsomes. At high levels of induction only one form (LM2A) was present. Although the two purified forms (LM2A and LM2B) were very similar they differed in: (a) peak elution on CM-Sepharose, (b) wavelength maximum of the reduced P450-CO spectrum, and (c) metabolism of several substrates, where the activities of LM2B ranged from 0.6 to 2.65 times that of LM2A. A third LM2 fraction (2C) was isolated from untreated rabbit liver and, although homogenous on sodium dodecyl sulfate-polyacrylamide gel electrophoresis, appeared to be a mixture of LM2B and a form of P450 LM2 other than LM2A. LM2A was not found in the untreated rabbit liver microsomes. On CM-Sepharose the elution of fraction 2C overlapped that of LM2B. The apparent molecular weight and immunoresponse to anti-LM2A IgG were the same for fraction 2C as for LM2A and LM2B. Peptide mapping using trypsin showed no difference between LM2A and LM2B, but consistently revealed at least one extra band with fraction 2C. After CNBr cleavage and high-pressure liquid chromatography separation of the LM2A and LM2B fragments the peptide beginning with Pro(347) of LM2A (peak 4A) eluted 1/2 min later than that of LM2B (peak 4B) indicating a difference in the fragments, although partial NH2-terminal amino acid sequences and molecular masses were the same. The corresponding CNBr fragment of fraction 2C splits into two peaks (4C:1 and 4C:2) with retention times corresponding to 4B and 4A, respectively. The mass of 4C:1 was the same as that of 4B, while the mass of 4C:2 markedly differed from that of 4A and 4B. Both fragments had the same partial NH2-terminal amino acid sequence as 4A and 4B. After comparing the physiochemical properties as well as catalytic activities of these isolated and purified LM2 forms with the cDNA-expressed forms 2B-B0, 2B-B1, 2B-B2, and 2B-Bx [see R. Ryan et al. (1993) Arch. Biochem. Biophys. 304, 454-463], the data suggest that LM2A is the form designated as 2B-B0 (LM2), LM2B is 2B-Bx, and fraction 2C is a mixture containing 2B-B1 and 2B-Bx. This is the first isolation and identification of the three isozymic LM2 proteins from rabbit liver.

Influence of ionic strength on the P450 monooxygenase reaction and role of cytochrome b5 in the process.

A stimulatory effect of increased salt content on the metabolism of benzphetamine, 7-ethoxycoumarin, and coumarin by rabbit liver microsomes, CYP2B4 and rabbit CYP1A2, was seen, indicating that the effect was not specific for either substrate or form of cytochrome P450. The stimulation was not due to an action on the cytochrome P450 itself as increased salt concentration minimally affected the substrate turnover when cumene hydroperoxide was used as the source of active oxygen. The elevation of ionic strength increased the coupling efficiency of the monooxygenase reaction with benzphetamine as substrate. Cytochrome b5 also can increase the monooxygenase coupling efficiency. At low ionic strength cytochrome b5 did not much influence the reduction of P450, but the rate constant of the cytochrome b5 reduction was increased about 15-fold by its binding to cytochrome P450. A stimulatory effect of cytochrome b5 on benzphetamine oxidation was seen at low ionic strength, but it was lost at elevated ionic strength as the binding of cytochrome b5 to cytochrome P450 was weakened. At the higher ionic strength cytochrome b5 competes with cytochrome P450 for the reductase, an action that slows cytochrome P450 reduction. Based upon these observations, plus those in the literature, a scheme is suggested that proposes the stimulatory effect of cytochrome b5 on the cytochrome P450-mediated monooxygenation reaction is due to an increase in the efficiency of the electron transfer reaction: With cytochrome b5 bound to cytochrome P450, two electrons can be provided from the reductase to the P450-b5 complex in a single interaction, obviating the need for a second interaction with the reductase.

The NH2-terminal region of rabbit CYP2E1 is not essential for interaction with NADPH-cytochrome P450 reductase.

Recently we reported that electrostatic forces interfere with the formation of the NADPH-cytochrome P450 reductase/cytochrome P450 electron transfer complex and suggested that this complex is formed by the attraction of the complementary hydrophobic patches (Voznesensky, A., and Schenkman, J. (1994) 269 J. Biol. Chem, 15724-1573). In this report we evaluate the role of the NH2-terminal hydrophobic region of CYP2E1 in the interaction with the reductase by comparing ionic strength dependence of the reduction of the full-length and truncated delta 3-29 CYP2E1. Increasing ionic strength stimulates reduction of both full-length and truncated CYP2E1. The neutralization of electrostatic interactions by increasing ionic strength revealed no impairment of the delta 3-29 CYP2E1 reduction compared to that of the full-length CYP2E1, indicating that the NH2-terminal region is not essential for the interaction of the cytochrome with the reductase.

Quantitative analyses of electrostatic interactions between NADPH-cytochrome P450 reductase and cytochrome P450 enzymes.

A decline in the ionic interactions in the medium with increasing ionic strength (decrease in the ionic activity coefficients) was accompanied by an increase in the fast phase rate constants of CYP2B4 and CYP1A2 reduction. The stimulations were observed both in reconstituted P450 systems and in microsomes. An increase in the ionic strength from 10 to 100 mM sodium phosphate resulted in a 7-fold decrease in the Km of CYP1A2 for NADPH-cytochrome P450 reductase, while the Vmax was unchanged. When ionic interactions were neutralized without changing the ionic strength by addition of charged oligopeptides (polylysine and polyglutamic acid), stimulations of CYP1A2 and CYP2B4 reduction were observed. Increase in the ionic strength also enhanced the rate of cytochrome P450 reduction in control and phenobarbital-induced rat liver microsomes and in reconstituted systems containing purified rat liver CYP2C6, CYP2C12, CYP2C13, and CYP2E1, and rat reductase. A method was devised for the quantification of the number of charges involved in protein-protein interactions based on the estimation of the ionic activity coefficients. Different numbers of charged residues are involved in the repulsion between different P450 forms and the reductase. The product of the number of charges involved in the interaction between rabbit reductase and CYP2B4 is 10.84 compared with the value of 6.64 for the reductase-CYP1A2 interaction.

Inhibition of cytochrome-P450 reductase by polyols has an electrostatic nature.

The present study was undertaken to examine the nature of the inhibitory action of glycerol on the liver microsomal monooxygenase system. In agreement with earlier observations, glycerol inhibited benzphetamine N-demethylation by liver microsomes of the phenobarbital-treated rabbit. The presence of glycerol in the medium did not affect binding of the substrate to cytochrome P450. Another polyol, ethylene glycol, was equally efficient in inhibiting benzphetamine N-demethylation. Both also inhibited reduction of rabbit cytochrome P450 LM2, cytochrome c and potassium ferricyanide by NADPH-cytochrome-P450 reductase in microsomes. Recently, we showed that the stimulation of electron transfer by increased ionic strength is due to neutralization of electrostatic interaction between NADPH-cytochrome-P450 reductase and its charged redox partners [Voznesensky, A. I. & Schenkman, J. B. (1992) J. Biol. Chem. 267, 14669-14676]. Polyols have an opposite effect to that of salt on ionic properties of a solution. They decrease the dielectric constant, thereby promoting electrostatic interactions between proteins. Addition of polyols decreased the conductivity of the medium. When rates of electron transfer to charged acceptors, cytochrome P450, cytochrome c and potassium ferricyanide, at various salt and polyol concentrations, relative to activities in 200 mM sodium phosphate, were plotted as a function of the conductivity the data for each acceptor fit on the same line. In contrast, neither alteration of ionic strength nor polyol addition affected the rate of electron transfer from NADPH-cytochrome-P450 reductase to an uncharged acceptor 1,4-benzoquinone. The data obtained is consistent with our earlier suggestion that charge repulsion limits redox interactions between rabbit cytochrome P450 LM2 and its reductase at low ionic strength, and suggest that the observed action of polyols is the result of enhancement of electrostatic interactions that inhibits electron transfer between NADPH-cytochrome-P450 reductase and its charged redox partners. In congruence with the hypothesis, the Km of rabbit cytochrome P450 LM2 for NADPH-cytochrome-P450 reductase was increased almost one order of magnitude by elevating the glycerol content from 5% to 25% (by vol.) without a change in Vmax.

The cytochrome P450 2B4-NADPH cytochrome P450 reductase electron transfer complex is not formed by charge-pairing.

Attempts to covalently link NADPH-cytochrome P450 reductase to cytochrome P450 2B4 using a water-soluble carbodiimide, 1-ethyl-3-(3-dimethylisopropyl)carbodiimide, were unsuccessful, despite the fact that under the same conditions about 30% of P450 2B4 could be covalently linked with cytochrome b5 in a functionally active complex (Tamburini, P. P., and Schenkman, J. B. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 11-15). This suggested that the functional electron transfer complex between P450 2B4 and reductase is not stabilized by electrostatic forces. Raising the ionic strength of the medium is disruptive to salt bridges and was used to further test whether P450 2B4 and the reductase form charge-pairing complexes. Instead of inhibiting electron transfer, high ionic strength increased the apparent fast phase rate constant and the fraction of P450 2B4 reduced in the fast phase. The possibility that electron transfer between NADPH-cytochrome P450 reductase and P450 2B4 is diminished by charge repulsion was examined. Consistent with this hypothesis, the Km of P450 2B4 for reductase was decreased 26-fold by increasing the ionic strength from 10 to 100 mM sodium phosphate without affecting the Vmax. The rate of benzphetamine N-demethylation also was increased by elevation of the ionic strength. Electron transfer from the reductase to other charged redox acceptors, e.g. cytochrome c and ferricyanide, was also stimulated by increased ionic strength. However, no similar stimulation was observed with the uncharged acceptor 1,4-benzoquinone. Polylysine, a polypeptide that binds to anionic sites, enhanced electron transfer from NADPH to ferricyanide and the apparent fast phase of reduction of cytochrome P450. The results are consistent with the hypothesis that charges on NADPH-cytochrome P450 reductase and cytochrome P450 decrease the stability of the electron transfer complex.

Molecular cloning of a cDNA for rat diabetes-inducible cytochrome P450RLM6: hormonal regulation and similarity to the cytochrome P4502E1 gene.

1. A polyclonal, monospecific antibody to a constitutive, diabetes-inducible and insulin-reversible cytochrome P-450 isozyme (RLM6) was used to screen a male rat liver cDNA library in lambda gt 11. Six clones harbouring the RLM6 cDNA insert were isolated initially from the expression library and three of these were further plaque-purified and sub-cloned. A 1.1 Kb cDNA insert, representing approximately 65% of the expected full length cDNA was characterized by restriction endonuclease mapping and sequenced by the dideoxy chain-termination method. Comparison of the nucleotide sequence of RLM6 cDNA to that of ethanol-inducible P4502E1 rat cDNA showed the two cDNAs to be identical, the RLM6 cDNA corresponding to nucleotides 310-1402 of the P4502E1 sequence. 2. RLM6 cDNA probe was used in Northern blot and RNA dot blot hybridization analysis to demonstrate that both streptozotocin-induced diabetes and fasting significantly elevated the steady-state level of RLM6 mRNA in male rat liver. Increased RLM6 mRNA level in the diabetic rat resulted in increased RLM6 apoprotein synthesis when polysomal RNA was used in a cell-free, protein-synthesizing system, indicating that the elevated RLM6 level observed in diabetic rats was correlated directly with the increased RLM6 mRNA concentration. 3. Daily insulin treatment of diabetic rats reversed the diabetes-dependent increase in RLM6 mRNA in a time-dependent manner, returning to control values after approximately 2 weeks of continuous insulin treatment. This insulin-dependent decrease of the RLM6 mRNA level was paralleled by a similar time-dependent decrease in serum acetone concentration. 4. Treatment of the male diabetic rat with testosterone also resulted in a decrease in both RLM6 mRNA and in vitro translated apoprotein. 5. Modulation of RLM6 mRNA level in the diabetic rat by insulin and testosterone, and the nucleotide sequence similarity with that of P4502E1 confirms that diabetes-inducible P450RLM6 and ethanol-inducible P4502E1 are coded for by the same gene.

Cholate solubilization of liver microsomal membrane components which promote NADPH-supported lipid peroxidation.

NADPH-supported lipid peroxidation monitored by malondialdehyde (MDA) production in the presence of ferric pyrophosphate in liver microsomes was inactivated by heat treatment or by trypsin and the activity was not restored by the addition of purified NADPH-cytochrome P450 reductase (FPT). The activity was differentially solubilized by sodium cholate from microsomes, and the fraction solubilized between 0.4 and 1.2% sodium cholate was applied to a Sephadex G-150 column and subfractionated into three pools, A, B, and C. MDA production was reconstituted by the addition of microsomal lipids and FPT to specific fractions from the column, in the presence of ferric pyrophosphate and NADPH. Pool B, after removal of endogenous FPT, was highly active in catalyzing MDA production and the disappearance of arachidonate and docosahexaenoate, and this activity was abolished by heat treatment and trypsin digestion, but not by carbon monoxide. The rate of NADPH-supported lipid peroxidation in the reconstituted system containing fractions pooled from Sephadex G-150 columns was not related to the content of cytochrome P450. p-Bromophenylacylbromide, a phospholipase A2 inhibitor, inhibited NADPH-supported lipid peroxidation in both liver microsomes and the reconstituted system, but did not block the peroxidation of microsomal lipid promoted by iron-ascorbate or ABAP systems. Another phospholipase A2 inhibitor, mepacrine, poorly inhibited both microsomal and pool-B'-promoted lipid peroxidation, but did block both iron-ascorbate-driven and ABAP-promoted lipid peroxidation. The phospholipase A2 inhibitor chlorpromazine, which can serve as a free radical quencher, blocked lipid peroxidation in all systems. The data presented are consistent with the existence of a heat-labile protein-containing factor in liver microsomes which promotes lipid peroxidation and is not FPT, cytochrome P450, or phospholipase A2.