Induction of multiple forms of mouse liver cytochrome P-450. Evidence for genetically controlled de novo protein synthesis in response to treatment with beta-naphthoflavone or phenobarbital.

The administration of polycyclic aromatic compounds such as beta-naphthoflavone or 3-methylcholanthrene is known to cause the induction of many liver microsomal monoxygenase activities and the appearance of a distinct cytochrome called P-448 in genetically responsive, but not in nonresponsive, inbred mouse strains. However, the administration of 2,3,7,8-tetrachlorodibenzo-p-dioxin induces these activities and cytochrome P-448 formation to the same extent in both responsive and nonresponsive inbred strains. In contrast, phenobarbital or pregnenolone-16 alpha-carbonitrile induces in both responsive and nonresponsive strains a different profile of enzyme activities and the appearance of cytochrome P-450 (rather than cytochrome P-448). In the present studies, electrophoresis of liver microsomal proteins from inbred C57BL/6N and DBA/2N and recombinant inbred AKXL-38 and AKXL-38A mouse strains revealed the presence of four polypeptides whose relative staining intensity could be correlated with the induction state of the microsomes as determined by enzymatic and spectral methods. Of these four bands, Band 4 (55,000 daltons) was increased whenever spectral measurements revealed an increase in the cytochrome P-448 content due to administration of beta-naphthoflavone or 2,3,7,8-tetrachlorodibenzo-p-dioxin. Administration of pregnenolone-16alpha-carbonitrile caused an increase in Band 3 (54,000 daltons), whereas administration of phenobarbital caused an increase primarily in Band 2 (51,000 daltons) but also smaller increases in Band 1 (49,000 daltons) and Band 4. The changes observed for phenobarbital and pregnenolone-16alpha-carbonitrile were the same for both responsive and nonresponsive strains. The same electrophoretic technique was used to measure the incorporation of radioactive leucine into microsomal proteins. Microsomes were prepared from liver combined from responsive mice (C57BL/6N) treated with beta-naphthoflavone and L-[14C]leucine and nonresponsive mice (DBA/2N) treated with beta-naphthoflavone and L-[3H-4,5]leucine. A significant increase in the 14C/3H ratio was observed for Band 4, and decreases were seen for Bands 1 and 2. In similar experiments with other mice and phenobarbital as the inducing agent with L-[14C]leucine and the vehicle alone with L-[3H-4,5]leucine, the 14C/3H ratio was markedly increased for Band 2, and smaller increases were observed for Bands 1 and 4. These results and other data presented indicate that the increased formation of cytochrome P-448 and P-450 by beta-naphthoflavone and phenobarbital, respectively, is primarily the result of an increased rate of de novo protein synthesis rather than a decreased degradation rate or a conversion of pre-existing polypeptides.


The administration
of polycyclic aromatic compounds such as @naphthoflavone or 3-methylcholanthrene is known to cause the induction of many liver microsomal monoxygenase activities and the appearance of a distinct cytochrome called P-448 in genetically responsive, but not in nonresponsive, inbred mouse strains. However, the administ.ration of 2,3,7,8-tetrachlorodibenzo-p-dioxin induces these activities and cytochrome P-448 formation to the same extent in both responsive and nonresponsive inbred strains. In contrast, phenobarbital or pregnenolone-16a-carbonitrile induces in both responsive and nonresponsive strains a different profile of enzyme activities and the appearance of cytochrome P-450 (rather than cytochrome P-448).
In the present studies, electrophoresis of liver microsomal proteins from inbred C57BL/6N and DBAI'LN and recombinant inbred AKXL-38 and AKXL-38A mouse strains revealed the presence of four polypeptides whose relative staining intensity could be correlated with the induction state of the microsomes as determined by enzymatic and spectral methods. Of these four bands, Band 4 (55,000 daltons) was increased whenever spectral measurements revealed an increase in the cytochrome P-448 content due to administration of &naphthoflavone or 2,3,7,8-tetrachlorodibenzo-p-dioxin. Administration of pregnenolone-16&-carbonitrile caused an increase in Band 3 (54,000 daltons), whereas administrat,ion of phenobarbital caused an increase primarily in Band 2 (51,000 daltons) but also smaller increases in Band 1 (49,000 daltons) and Band 4. The changes observed for phenobarbital and pregnenolone-16cu-carbonitrile were the same for both responsive and nonresponsive strains. The same electrophoretic technique was used to measure the incorporation of radioactive leucine into microsomal proteins. Microsomes were prepared from liver combined from responsive mice (C57BL/6N) treated with fi-naphthoflavone and L-[*'C]leucine and nonresponsive mice (DBA/2N) treated with fi-naphthoflavone and L-[%4,5]leucine. A significant increase in the "C/"H ratio was observed for Band 4, and decreases were seen f'or Bands 1 and 2. In similar experiments with other mice and phenobarbital as the inducing agent with L-["Clleucine and the vehicle alone with L-[3H-4,5]leucine, the "C/3H ratio was markedly increased for Band 2, and smaller increases were observed for Bands 1 and 4. These results and other data presented indicate that the increased formation of cytochrome P-448 and P-450 by p-naphthoflavone and phenobarbital, respectively, is primarily the result of an increased rate of de noun protein synthesis rather than a decreased degradation rate or a conversion of pre-existing polypeptides.
It has become increasingly evident that several forms of monooxygenase activities in liver microsomes (2-6). More cytochrome P-450 exist and may account for the multiple recently, these forms have been purified and their physical and *This research was supported by Grant BMS71-01195 from the catalytic properties have been partially characterized. Most Sational Science Foundation and Grant AM-10339 from the United likely, the two or three forms of a monooxygenase activity States Public Health Service. A report of portions of this work has been represent two or three spectrally distinct species of CO-binding with the use of the microsomal enzyme inducers, phenobarbital and 3-methylcholanthrene, which preferentially enhance the formation of one or more forms of cytochrome P-450. When certain substrates are metabolized by liver microsomes of rats treated with phenobarbital, the relative amounts of alternate products formed differ from those when rats were treated with 3-methylcholanthrene.
In mice, the 'genetic trait of "aromatic hydrocarbon responsiveness" refers to the capacity for induction of cytochrome P-448 and numerous monoxygenase activities' by various polycyclic aromatic compounds. The trait has been shown (4) to segregate almost exclusively as a single autosomal dominant gene among certain inbred strains. This difference in responsiveness results in differences in the susceptibility to Smethylcholanthrene-initiated tumorigenesis (15,16). in the mutagenicity of 3-methylcholanthrene, 6-aminochrysene, and P-acetylaminofIuorene in uitro (17), and in the suscept.ibility to toxicity caused by acetaminophen (18) and various polycyclic hydrocarbons, lindane, and polychlorinated biphenyls (19). Most likely, these results reflect qualitative and quantitative differences in metabolites formed due to relative amounts of multiple forms of cytochrome P-4tiO in various tissues.?
"Induction" of any monooxygenase activity has been denoted by many investigators as an increased rate of product formation or substrate disappearance (21). Induction may result from an increase in the rate of de nouo synthesis of a protein, in the rate of activation from pre-existing components, a decrease in the rate of degradation, or a combination of these events. It has been difficult to distinguish among these possibilities because of technical problems encountered in attempts to purify the multicomponent membrane-bound enzyme system and because of a lack of knowledge regarding which factor in the system is the rate-limiting component for a given monooxygenase activity. With some purified enzymes to which antibodies have been produced, immunoprecipitable radi0activit.y has been equated with de nom prot.ein synthesis for such inducible enzymes as cu-aminolevulinate dehydratase (22), tyrosine aminotransferase (23,24), glutamine synthetase (25), and @-glucuronidase (26). Differences in enzyme activity presumably due to a conformational change ("activation" or and AKXL-38A-These particular sublines had been chosen for this experiment because AKXL-38 was found to have the nonresponsive allele, and AKXL-38A the responsive allele at the Ah locus (4) by means of skin ulcers following topical treatment with 7,12-dimethylbenz [alanthracene (48). Recombinant inbred sublines can be developed from the cross of two unrelated but highly inbred progenitor strains and maintained independently via strict brother-sister inbreeding since the F, generation (33a, 49, 50). This procedure in each subsequent generation results in ever increasing chances of homozygosity at all loci at which homozygosity is selectively allowed. The resulting subline can be looked upon as a replicable recombinant population which has, at each locus, homozygous alleles from one or the other progenitor strain. Livers from several such mice of each subline were combined for electrophoretic and spectrophotometric studies and numerous enzyme assays ( Fig. 1 and Table I). These combined livers can, therefore, be considered as if one individual responsive F, and one individual nonresponsive F, mouse had been used. The large number of studies described in Table I would not have been possible using the liver of an individual mouse. Fig. 1 is a photograph of a dried 0.75-mm-thick electrophoretic gel. Four distinct bands were resolved in the molecular weight range of 49,000 to 55,000, which are indicated by numbers 1 to 4. The molecular weights (estimated to the nearest 1000) of Bands 1 through 4 are 49,000, 51,000, 54,000, and 55,000, respectively. For /3-naphthoflavone-treated AKXL-38A (position l), an increase in Band 4 is readily apparent, compared with Band 4 in @-naphthoflavone-treated AKXL-38, and control AKXL-38 and control AKXL-38A (positions 2 to 4, respectively). No significant differences in Bands 1, 2, or 3 or in any other bands throughout the gel were  A quantitative increase in Band 4 after @-naphthoflavone or TCDD treatment (Figs. 1 and 2), therefore, appears to be associated with genetically mediated aromatic hydrocarbon responsiveness, or the Ah locus (4). Is this increased amount caused by de nouo synthesis, an altered configuration of existing microsomal proteins, or a block in degradation?
In experiments designed to answer this question, radioactive leucine was administered during the time period of maximal increase of cytochrome P-448 or P-450 content in response to treatment with p-naphthoflavone or phenobarbital, respectively. In a typical experiment, L-["Clleucine was given to one and L-[4,5-SH]leucine to another mouse between 8 and 16 hours following injection of inducer or corn oil. At the end of the radioactive pulse, the mice were sacrificed, livers were combined, and "C/"H ratios of resultant microsomal proteins on electrophoretic gels were determined.
In order to recover sufficient quantities of radioactivity in microsomal polypeptides from individual slab gel slices, we used 3.0-mm-thick slabs, and tested 50, 75, 100, and 150 rg of protein applied in individual wells. We found that 100 to 150 pg of protein caused indistinct borders for many of the bands, and that 50 wg/well produced the most distinct pattern. We also found that 50 fig of protein is from liver microsomes from a B6 mouse treated with phenobarbital 16 hours, and 615 PCi of L-["C]leucine 8 hours, before being killed. The trit.ium radioact.ivity is from microsomes from a control B6 mouse treated with corn oil 16 hours, and 1.6 mCi of L-[4,5-3H]1eucine 8 hours, before being killed. B, "CPH ratio of each gel slice shown in Fig. 5A. C, radioactivity in "C/3H ratio correspond exceedingly well with the qualitative increases visualized in liver microsomes from phenobarbitaltreated B6 and D2 mice (Fig. 1, positions 10 and 11, respectively, and Fig. 2). These data indicate that phenobarbital causes several forms of cytochrome P-450 (apparently including P-448) to increase primarily as a result of de nouo synthesis and not as a result of alteration in pre-existing components or a decrease in degradation rates. In Fig. 5, B and D, and to a lesser extent in Fig. 4B, a significant increase in the "C/"H ratio is evident in the molecular weight region of about 79,000 (about 52 to 58 mm from the origin in different experiments); this increase may represent NADPH-cytochrome P-450 reductase (54-58). If this is true, phenobarbital, and P-naphthoflavone to a lesser degree, cause increases in the rate of de nouo synthesis of this enzyme. Another significant increase in the 1'C/9H ratio can be seen at or near the front in Figs. 4B, 5B, and 50. This region corresponds to all polypeptides having molecular weights of less than approximately 20,000. We are uncertain what this increase might represent; however, it may reflect important small induction-specific proteins formed in response to /3naphthoflavone or phenobarbital treatment. Cytochrome b, is another possibility.
Comparison of Control B6 and 02 or Phenobarbital-treated B6 and 02 Mice-No electrophoretic differences were visualized between control B6 and D2 mice (data not illustrated) or between phenobarbital-treated B6 and D2 mice (Fig. 1,  positions 10 and 11). In order to confirm these visual observations and to evaluate the reliability of the radiometric technique, we performed double label experiments in which both strains were treated either with corn oil or with phenobarbital. Fig. 6 shows typical expttrimental results, indicating that no significant differences in leucine incorporation were seen between these two strains after control or after phenobarbital treatment. The normal degree of experimental variation in radioactivity of individual gel slices can also be appreciated in Fig. 6. We have also performed experiments in which 2-hour radioactive pulses were used, and found that the results were qualitatively very similar to those obtained with an El-hour radioactive leucine pulse for both fl-naphthoflavone-and phenobarbital-treated mice. DISCUSSION We have shown in this report that P-naphthoflavone and phenobarbital, which induce microsomal monooxygenase activities and total cytochrome P-450 content, exert their effect primarily as a result of de nouo synthesis of cytochrome apoproteins rather than as a result of stabilization or of conversion from pre-existing polypeptides. Similar previous studies with liver microsomes from rats treated with phenobarbital (59, 60) or 3-methylcholanthrene (59) are consistent with our conclusion; however, the four electrophoretic bands described in this report were not as well resolved in the earlier studies. Recent improvements in electrophoretic techniques (37-39, 61-63) and in the partial purification of several forms of cytochrome P-450 (5, 6, 64) have resulted in the demonstration of three or more forms of cytochrome P-450 having varying substrate specificities and physical properties (5, 6). In this study, we have (a) combined the improved electrophoretic techniques with previous double label radioisotopic methods; (b) examined genetically mediated differences in inbred strains and recombinant inbred sublines of mice in response to inducers; and (c) demonstrated that three different classes of monooxygenase inducers, exemplified by fi-naphthoflavone or TCDD, pregnenolone-16a-carbonitrile, and phenobarbital, each cause increases in specific microsomal proteins presumed to be apoproteins of cytochromes P-450. Both P-naphthoflavone and TCDD increase Band 4 (molecular weight, 55,000), pregnenolone-16a-carbonitrile enhances Band 3 (molecular weight, 54,000), and phenobarbital primarily increases Band 2 (molecular weight, 51,000) but also causes significant increases in Band 1 (molecular weight, 49,000) and Band 4. The molecular weights of these multiple forms of mouse liver Lower, livers from phenobarbital-treated I36 and D2 mice were combined, after administration of radioactive leucine as described above. The time periods between administration of corn oil, phenobarbital, or radioactive leucine were the same as those in Fig. 5. PhBarb, phenoharhital.
cytochromes P-450 that increase in response to these specific inducers are similar, but not identical, to those found in rat (5) and rabbit (6) liver. Band 4, which is induced in mice by ,&naphthoflavone and TCDD (and by 3-methylcholanthrene),5 has a greater apparent molecular weight than the principal band induced by phenobarbital.
The same relative order of molecular weights has been observed for rat (5) and rabbit (6) liver microsomes.
The evidence indicating that phenobarbital enhances, to some degree, the band presumed to be apocytochrome P-448 in mouse liver is consistent with several other observations of 66 ducer, TCDD, was shown (33) to induce several monooxygenase activities and to increase cytochrome P-448 content in D2 mice to the same extent as that found in B6 mice, further indicating that D2 mice have the capacity to form P-448. Radioactivity in the region of the gel corresponding to proteins having a molecular weight between about 49,000 and 55,000 ranged between 34 and 39% of the total microsomal radioactivity.
This result was the same for all samples from either control or /$naphthoflavone-treated B6 and D2 mice, but was between 48 and 55% of the total microsomal radioactivity for all samples from phenobarbital-treated B6 and D2 mice. Thus, compared with controls, phenobarbital treatment caused about a 40% increase in leucine incorporation into the cytochrome P-450 proteins relative to the total microsomal proteins, whereas p-naphthoflavone treatment produced no such increase. These data are also consistent with the "C/"H pattern seen in Fig. 4B (compared with the patterns seen in Fig. 5, Band D), where it appears that @-naphthoflavone causes a genetically mediated increase specifically in the de nouo synthesis of P-448 apoprotein to the exclusion of de nouo synthesis of one or more other forms of cytochrome P-450. How such a balance in de nouo synthesis exists among these different forms of microsomal cytochromes requires further investigation.
It is extremely unlikely that such a preferential synthesis of one cytochrome form instead of several other forms could be controlled at the translational level; more likely, some mechanism of processing or compartmentalization (67) is involved. For example, each of the P-450 apoproteins under normal conditions may fit properly into the microsomal membrane mosaic. However, although each of the apoproteins is still synthesized during @naphthoflavone stimulation, several apoproteins now may be excluded under this new membrane condition.
The problems arising from the use of a labeled amino acid which can be reutilized, compared with the use of a labeled amino acid such as guanidinoarginine which is not recycled, have been reported (68)(69)(70). The mean half-life for various microsomal membrane protein components has been estimated to range from about 2 days (70) to 4 days (71); cytochrome b, apoprotein has an estimated half-life of about 3.5 days (72), and membrane-bound NAD glycohydrolase has an apparent half-life of about 18 days (73). Thus, we believe that our results, based on the use of labeled leucine for pulses of 2 or 8 hours, represent changes primarily in de nouo synthesis, and that problems with reutilization of labeled amino acids would only be encountered with longer pulses (i.e. 24 hours or more). Our data showing an increase in the "C/*H ratio following fi-naphthoflavone or phenobarbital treatment indicate that these compounds act by increasing the rate of synthesis of one or more microsomal proteins. Although a decrease in the rate of protein degradation would result in an accumulation of newly synthesized protein, we believe that it is unlikely that the corresponding "C/3H ratio would increase after only 2 to 8 hours of radioisotope labeling. Therefore, if decreased protein degradation rates do contribute in a minor way to our observed increase in 'C/W ratio, the experiments described in this report would probably not detect such an effect. The possibility of changes in amino acid precursor pool size after either &naphthoflavone or phenobarbital treatment was not examined in the present study. The intracellular free amino acid pool is greatly changed by the administration of such drugs as actinomycin D (74), but to our knowledge, no one has looked for such a change after treatment with microsomal inducers such as polycyclic aromatic compounds or phenobarbital. Other investigators (75,76) have described a biphasic decrease in rat liver microsomal radioactive hemoproteins. 3-Methylcholanthrene treatment causes a 3-to 4-fold increase in the slow phase component (half-life, 46 to 48 hours) and does not affect the fast phase component (half-life, 7 to 8 hours). The increased amount of any given electrophoretic band will reflect the net change in rate of de nouo synthesis and the presumably unchanged degradation rate constant. If we assume that similar differences in half-lives also exist in mice, the limitations of our experiments may be better understood by several simple calculations. Consider the possibility that the apoprotein in Band 4 has a half-life 7 times longer than that in Bands 1, 2, and 3. If we assume that the steady state level of all four bands is equal in control mice, then for every 7 molecules of apoproteins in Bands 1, 2, and 3 being synthesized, only 1 molecule of Band 4 hemoprotein would be synthesized in order to maintain steady state conditions. This means that 1 molecule out of every 22 synthesized (or 4.5%) of total synthesis is required for Band 4 apoprotein. A 4-fold increase, for example, in de nouo synthesis for Band 4 apoprotein would result in the synthesis of 4 molecules out of every 25 (or 16% of total synthesis). If we assume that the rate of degradation is unchanged after P-naphthoflavone treatment, it is, therefore, possible that a marked increase in the rate of synthesis of the apoprotein having a 7 times longer half-life would be difficult to detect within the limits of experimental variability. Also, a relatively smaller net increase in de nouo synthesis of the specific apoprotein being degraded 7 times more slowly than the other three apoproteins is needed to account for the same (or greater) increase in total cytochrome P-450 apoprotein content. These thoughts are summarized in Fig. 7. In the control mouse, 10 units of apoprotein exist, a relatively small amount being associated with Band 4. Sixteen hours after phenobarbital treatment, we found about a 40% increase in leucine CONTROL El incorporation into the 49,000 to 55,000 molecular weight region (Fig. 5), and about a 40% increase in total P-450 content (Fig. 3); a primary increase in Band 2 and secondary increases in Bands 1 and 4 occurred. Hence, the phenobarbital-treated mouse is shown with 14 units of hemoproteins with the major increase (2 units) being in Band 2. In the fi-naphthoflavone-treated B6 mouse, we found no increase in leucine incorporation into the 49,000 to 55,000 molecular weight region, a marked increase in leucine incorporation only for Band 4 (Fig. 4), and yet, about a 50% increase in total CO-binding hemoprotein content (Fig. 3). Either a marked (9) increase in Band 4 occurs concomitantly with decreases in Bands 1, 2, and 3, (see Fig. 7) or an increase (from 1 to 6 units) in Band 4 occurs without the concomitant decreases in the other bands. From the standpoint of visualizing the four bands (Fig. l), this latter possibility appears more probable. From the standpoint of enzymatic activity, the former possibility may also be possible. In the liver of 3-methylcholanthrene-or &naphthoflavone-treated rabbits ('77), lethoxycoumarin 0-deethylase-specific activity is significantly lower. This result suggests that, at least in the case of rabbit liver, certain cytochrome forms of cytochrome P-450 may, in fact, diminish in concentration after polycyclic aromatic compound treatment (77). We have not found such a decrease in any hepatic monaoxygenase activities from 3-methylcholanthrene-or BNF-treated mice.
Another microsomal protein which is linked in function and may be also linked architecturally in the membrane (78) is epoxide hydrase. Whereas this enzyme activity is not inducible in 3-methylcholanthrene-treated B6 or D2 mice, following 3-methylcholanthrene treatment epoxide hydrase is induced more than 2-fold in phenobarbital-treated B6 mice and less than 50% in phenobarbital-treated D2 mice (79). This differential response to phenobarbital between B6 and D2 mice was not readily apparent for any polypeptide band seen in Fig. 7. Further, we may conclude that Band 1, which increases to some degree after phenobarbital treatment, does Fm. 7. Hypothetical scheme attempting to interpret the results of this report. Each sguaro denotes a stoichiometric unit of electrophoretic Band 4. 3, 2, or 1. The terms F&K, 54K, 5lK. and 49K denote molecular weight S5.000, 54,000, 51,000, and 49,000, respectively.
l'his scheme is admittedly simplified in that each electrophoretic band may represent two or more hemoprotein enzyme active-sites having dif-