Association of Multiple Developmental Defects and Embryonic Lethality with Loss of Microsomal NADPH-Cytochrome P450 Oxidoreductase*

The microsomal flavoprotein NADPH-cytochrome P450 oxidoreductase (CYPOR) is believed to function as the primary, if not sole, electron donor for the microsomal cytochrome P450 mixed-function oxidase system. Development of the mammalian embryo is dependent upon temporally and spatially regulated expression of signaling factors, many of which are synthesized and/or degraded via the cytochromes P450 and other pathways involving NADPH-cytochrome P450 oxidoreductase as the electron donor. Expression of CYPOR as early as the two-cell stage of embryonic development (The Institute for Genomic Research Mouse Gene Index, version 5.0, www.tigr.org/tdb/mgi) suggests that CYPOR is essential for normal cellular functions and/or early embryogenesis. Targeted deletion of the translation start site and membrane-binding domain of CYPOR abolished microsomal CYPOR expression and led to production of a truncated, 66-kDa protein localized to the cytoplasm. Although early embryogenesis was not affected, a variety of embryonic defects was observable by day 10.5 of gestation, leading to lethality by day 13.5. Furthermore, a deficiency of heterozygotes was observed in 2-week-old mice as well as late gestational age embryos, suggesting that loss of one CYPOR allele produced some embryonic lethality. CYPOR -/- embryos displayed a marked friability, consistent with defects in cell adhesion. Ninety percent of CYPOR -/- embryos isolated at days 10.5 or 11.5 of gestation could be classified as either Type I, characterized by grossly normal somite formation but having neural tube, cardiac, eye, and limb abnormalities, or Type II, characterized by a generalized retardation of development after approximately day 8.5 of gestation. No CYPOR -/- embryos were observed after day 13.5 of gestation. These studies demonstrate that loss of microsomal CYPOR does not block early embryonic development but is essential for progression past mid-gestation.


INTRODUCTION
NADPH-cytochrome P450 oxidoreductase (CYPOR) is an essential component of the microsomal P450 mixed-function oxidase system (1), mediating electron transfer from NADPH to the cytochromes P450, and a variety of other acceptors, including heme oxygenase (2), fatty acid elongase (3), cytochrome b 5 (4) and squalene monoxygenase (5). Electron transfer to the cytochromes P450 as well as membrane anchoring requires the hydrophobic, N-terminal membrane-binding domain of this microsomal flavoprotein (6,7). No other physiological electron donor to the cytochromes P450 has been identified in vertebrates, although plants contain multiple CYPOR genes (8) and electron transfer from the cytochrome b 5 /cytochrome b 5 system to P450 has been reported in yeast lacking CYPOR (9,10).
Development of the mammalian embryo requires temporally-and spatially regulated biosynthesis and degradation of signalling factors, many of which, such as retinoic acid, sterols, prostaglandins, and steroids, are dependent upon cytochrome P450 or other CYPOR-dependent pathways. For example, squalene monooxygenase and CYP51 are necessary for biosynthesis of cholesterol (5,11), while catabolism of retinoic acid proceeds via CYP26A1 (10,11). The heme oxygenase pathway regulates heme homeostasis (14), and steroidogenesis is dependent upon several of the cytochromes P450 (15). CYPOR is expressed as early as the 2-cell stage of embryonic development (16)(17)(18), suggesting that one or more CYPOR-dependent processes are important in early embryonic development.
To determine whether CYPOR is an essential enzyme in the mouse and the sole electron donor to the cytochromes P450, we have examined the effects of CYPOR gene disruption by removal of the natural translation initiation site and deletion of the membrane-binding domain of 7 Chimeric males were mated with C57/BL6 females to yield heterozygous F1 progeny, which were then mated with each other to produce CYPOR +/+, +/-and -/-offspring. Mouse tail DNA was isolated using the Qiagen DNeasy kit (www.qiagen.com)and analyzed using two separate PCR reactions (Fig. 1B). The first employed the primers Intron 1C and Common 1 (5'-TCTGAGGGCACCACTCACGTTGT-3'), shown in Fig. 1A, which yielded bands of 4.7 Kbp and 4.0 Kb, respectively, for the wild-type and disrupted alleles. A second, confirmatory, reaction, yielding 210 bp and 260 bp fragments, respectively, for the wild-type and disrupted alleles, employed 3 primers: Common 1, which spans the Exon 2/Intron 2 junction and is found in both the wild-type and disrupted alleles, WT1 (5'-GTGTCACCAACATGGGGGACTCT-3'), located in the deleted segment of Intron 1 and found only in the wild-type allele, and Neo4 (5'-CTTCCATTTGTCACGTCCTGCAC-3'), located in the Neo gene of the disrupted allele (Fig.   1A).

Analysis of Embryos.
Detection of vaginal plugs in the morning was taken as gestation day 0.5 (E0.5). Embryos were genotyped by removal of either the tail or the hindlimb bud prior to fixation. For histology, embryos were fixed for at least 4 hr in 4% paraformaldehyde in phosphate-buffered saline, dehydrated, and embedded in paraffin. Ten-micron sections were stained with hematoxylin and eosin (23). For scanning electron microscopy, embryos were fixed in Karnovsky's buffer, postfixed in osmium tetroxide, dehydrated, critical point dried, coated with gold-palladium, and viewed on a Hitachi S-570 scanning electron microscope.
Analysis of Subcellular Fractions. For preparation of embryonic S1 and microsomal fractions, whole embryos were homogenized in 50 mM Tris, pH 7.7, 0.15 M KCl, 1 mM EDTA, containing 100 µM PMSF and 10 µg/ml aprotinin, and centrifuged at 14,000 x g for 30 minutes.
This low-speed supernatant fraction, designated S1, was centrifuged at 100,000 x g for 60 min glycerol. Protein was determined using the Pierce BCA Protein Assay Kit (http://www.piercenet.com). Western blotting was carried out using a polyclonal anti-rat CYPOR antibody, with ECL detection using the BM Chemiluminescence Western blotting kit from Roche (http://biochem.roche.com). EROD and cytochrome c reductase activities were measured as described (24,25). Cytochrome c reductase assays contained 23-50 (liver) or 10-30

RESULTS
Homologous recombination between the targeting construct and the mouse CYPOR gene deleted 2.5 Kbp of the 3'-end of Intron 1 and all but the last 6 bases of Exon 2 (Fig. 1A), thus removing both the translation start site and the membrane-binding domain necessary for electron transfer to cytochrome P450 (20). Two separate clones, designated G1 and H11, yielded chimeras that subsequently transmitted the disrupted allele to their progeny. Mice analyzed in this study were obtained from five founder males, four derived from the G1 clone and one from H11, and were maintained on a mixed 129SvJ/C57/BL6 background. No differences were noted among the progeny of the five founders.
Heterozygote crosses produced an average litter size of 6 ± 2, with a normal male: female ratio (0.52) and no apparent perinatal mortality. Genotyping of 2-week old offspring of heterozygote crosses yielded no homozygous -/-offspring (Table 1). Heterozygotes were obtained at a frequency of 56% (Table 1), significantly less (P< 0.05) than the 67% expected for a 1:2 wild-type: heterozygote ratio, suggesting that loss of one CYPOR allele reduced survival.
Genotyping of E8.5 to E10.5 embryos from heterozygote crosses produced the expected 1:2:1 Mendelian ratios; however, the frequency of homozygotes declined thereafter and no homozygous embryos were observed after E13.5. Genotyping of late gestational age embryos also revealed a deficiency of heterozygous embryos, suggesting an increased mortality of +/embryos.
Fore and hind limb buds were present and well-formed; however, neural tube abnormalities were evident. Fig Other gross abnormalities include a fluid-filled dilated pericardial cavity ( Fig. 2B and C), alteration in the size and positioning of the branchial arches relative to the fronto-nasal region ( Fig. 2A vs. 2B), as well as truncation of the tail (Fig. 2B). Petechial hemorrhaging was noted in the CYPOR -/-embryos and was particularly prominent at the edges of the hindbrain neural folds (Fig. 2B). Failure of the caudal neuropore to close was also observed (not shown).
Examples of a more severe, generalized retardation of development, the Type II phenotype, are presented in Fig. 2C. Although the majority of E10.5 Type II -/-embryos were turned, they were, in comparison with wild-type littermates, poorly-developed and small (Fig.   2C). These embryos displayed an observable heartbeat in spite of obvious pericardial edema.
Fore-and hindlimb formation, as well as cranial development, were delayed. Finally, -/embryos with an apparently normal yolk sac, but no recognizable embryonic structures were also isolated (Fig. 2C).
Scanning electron microscopy of the Type I CYPOR -/-embryos provided a detailed view of the neural tube defects (Fig. 3). By E10.5, the neural tube was completely closed in the wild-type embryo (Fig. 3A). In the hindbrain, the neural folds remained in a convex position in spite of continued proliferation of the neuroepithelium. Alternate patterns were also observed; Fig. 3C shows a Type I embryo where the cranial portion of the neural tube was closed, but the hindbrain neural folds failed to elevate and fuse. Malformation of the branchial arches and eye abnormalities were apparent in these embryos, as well as a rough surface, consistent with the observed increased friability, and suggesting defects in cell adhesion properties and/or membrane structure (Fig. 3, B and C). No Type II embryos remained intact through the fixation process.
Histological examination of the homozygous null embryos also demonstrates the convex position and lack of elevation of the neural folds in the Type I CYPOR -/-embryo (compare Fig.   4A and 4B). As was observed in the intact embryo, red blood cells were present in the region of the neural folds (Fig. 4, B and C). Numerous cells containing pyknotic nuclei were also detected in the CYPOR -/-embryo adjacent to the fronto-nasal midline (Fig. 4D), consistent with apoptosis. Apoptosis was also noted in the apical epidermal ridges of the limb bud (not shown), although the fore-and hindlimb buds were well-formed overall. Significantly, no evidence of necrosis was observed in the CYPOR -/-embryos. embryos also displayed histological evidence of defective heart development. Western blot analysis of both S1 and microsome fractions from CYPOR +/+ embryos demonstrated, as expected, a single immunoreactive 78 kDa protein, which was absent in E10.5 CYPOR -/-Type I embryos, consistent with deletion of the wild-type translation start site (Fig.   6). However, the -/-embryos contained a new immunoreactive protein with an apparent molecular weight of 66 kDa, found in the S1 fraction but not in microsomes. While both proteins were found in the S1 fraction of CYPOR +/-embryos, only the 78 kDa protein was found in heterozygote microsomes.
Cytochrome c reductase activity in liver microsomes isolated from adult heterozygous mice was approximately 50% that found in wild-type microsomes (Table 2), while EROD activity was decreased to 77% of wild-type. Cytochrome c reductase activity in microsomes prepared from either wild-type or heterozygous embryos (E11.5) was approximately 10% that of adult liver microsomes and was undetectable in microsomes isolated from Type I CYPOR -/embryos. EROD activity was also undetectable in embryonic CYPOR -/-microsomes (Table 2) Although the authentic CYPOR translation initiation site has been deleted in the disrupted allele, the presence of a 66 kDa protein in heterozygous and Type I homozygous embryos indicates that translation must have initiated at an alternate start site. Analysis by 5'-RACE of mRNA from CYPOR +/-embryos (not shown) reveals, in addition to the wild-type transcript, an alternative transcript containing Exon 1 spliced to Exon 3. Exon 3 contains two potential translation initiation sites, with initiation at the ATG at position 107 of the wild-type protein producing a protein of 65 kDa, remarkably similar to the size observed. This protein would retain binding sites for the isoalloxazine, but not the phosphate, of FMN, as well as the FAD and NADPH domains (27). The absence of the membrane-binding domain, essential for formation of a productive catalytic complex with cytochrome P450 (6), and its cytoplasmic localization argues against the ability of the 66 kDa protein to support microsomal P450 activities.
Although it is not known which CYPOR-dependent activities are essential for embryonic development, detection of CYPOR transcripts in two-cell mouse embryos suggests an essential role in the early stages of embryogenesis (18). The question of whether the survival of CYPOR -/-embryos to as late as E13.5 is the result of residual activity of the cytoplasmic protein or some other compensatory activity remains. Possible compensatory mechanisms include both the presence of an alternate electron donor which substitutes for the electron transfer functions of CYPOR, such as adrenodoxin/adrenodoxin reductase (28) or cytochrome b 5 (9, 10), or alternate sources of metabolites produced by CYPOR-dependent enzymes, for example, maternal sources of cholesterol (29). Regardless of the source, these pathways are insufficient to support normal embryonic development beyond mid-gestation. Teratogenicity has been associated with retinoic acid excess as well as deficiency, with major targets being the craniofacial region, heart, skeleton, limbs, eye, and central nervous system (30). Several catabolic pathways, including those mediated by CYP26A1, catalyze by guest on July 16, 2017 http://www.jbc.org/ Downloaded from spatially-and temporally-regulated breakdown of retinoic acid (12). Although CYP26A1 knockouts are lethal between E9.5 and birth and display some phenotypes in common with the CYPOR knockout, including exencephaly and arrested development, the predominant phenotypes, caudal truncations and vertebrate transformations, differ from those observed here (31, 32).
Cholesterol, in addition to being an essential component of eucaryotic cell membranes, has been recently shown to be necessary for activity of the sonic hedgehog protein, a mediator of patterning in the vertebrate embryo (33). CYPOR is required at two points in the biosynthesis of cholesterol: squalene monoxygenase catalyzes formation of squalene-2, 3-epoxide (5) while CYP51 catalyzes lanosterol demethylation (11). The structural defects and increased friability observed in the CYPOR -/-embryos are consistent with defects in cholesterol biosynthesis and are reminiscent of those observed in a human genetic disorder of cholesterol biosynthesis, desmosterolosis (29). Defective neural tube closure and developmental delay have been observed in a knockout of another enzyme involved in cholesterol biosynthesis, squalene synthase; however, lethality in these mice occurred somewhat earlier (E9.5) (34).
In view of the redundancy observed in many cellular functions, it is surprising that there is not an efficient alternative electron donor to the cytochromes P450. Although evolutionarily ancient and present in multiple copies in plants (8), CYPOR has been shown to exist as a single copy gene in the mouse (19). This situation may be different in primates, including humans, where CYPOR has been localized to a region of Chromosome 7 that is duplicated (35); this may provide a means to insure against the catastrophic consequences of CYPOR deletion.