Genetic regulation of aromatic amine N-acetylation in inbred mice.

A survey among 20 inbred mouse strains revealed large variation (up to approximately 20-fold) for the N-acetylation of p-aminobenzoic acid by blood N-acetyltransferase and for the aromatic amine carcinogen benzidine by both liver and blood N-acetyltransferase. Of 20 strains surveyed, three are classified as slow acetylators (A/J, AHe/J, and X/Gf) and 17 are classified as rapid acetylators (AuSsJ, Castaneous, ST/bJ, C57BL/6J, Molossinus, SF, SWR/J, 129/SV, RF/J, RIII/2J, IsCam, SJL/J, Balb/cJ, C3H/HeJ, CBA/J, AKR/J, and DBA/J). The rapid acetylator strains possessed approximately 10 times greater liver benzidine N-acetyltransferase specific activity than the slow acetylator strains. Intercross and backcross matings of A/J and C57BL/6J mice indicate that a single gene with two major alleles is responsible for differences in N-acetyltransferase activity in blood for p-aminobenzoic acid or the alternate aromatic amine carcinogen aminofluorene, and in liver for aminofluorene. Analysis of 11 recombinant inbred strains derived from matings of A/J with C57BL/6J mice support this conclusion and demonstrate the existence of minor modifying genes that segregate independently of the major N-acetyltransferase gene.

' The abbreviations used are: PABA,p-aminobenzoic acid; A, A/J; AF, aminofluorene; BZ, benzidine; C57, C57BL/6J; INH, isoniazid, Me2S0, dimethylsulfoxide; RI, recombinant inbred, SMZ, sulfamethazine. the fact that no strain differences are apparent with regard to liver PABA N-acetyltransferase activity (Tannen and Weber, 1979). Of nine strains tested, only the A/J mouse lacked blood PABA N-acetyltransferase activity. Low SMZ N-acetyltransferase activity was apparent in either blood or liver of all strains, and thus, no strain differences were seen for this substrate in direct contrast to the case in man or rabbit. Using the A/J and C57BL/6J inbred mouse strains as models of slow and rapid acetylators, respectively, Tannen and Weber (1980a) showed that the expression of PABA acetylation in blood is consistent with simple Mendelian inheritance of two codominant alleles.
In human populations, the acetylation polymorphism plays an important role in determining the variability seen in the incidence of toxic responses to certain arylamine drugs such as INH, procainamide, and hydralazine (Drayer and Reidenberg, 1977). Recent studies have indicated that the acetylation polymorphism may play a role in differential drug toxicity in animal models as well as man. Using the A/J and C57BL/6J mouse model, it was shown that procainamide was acetylated to a lesser degree by the slow acetylator strain (A/J); that A/J mice had a higher incidence of spontaneous antinuclear antibodies than C57BL/6J mice; and that these antibodies could be induced by oral procainamide (Tannen and Weber, 1980b). In C57BL/6J mice, procainamide tended to suppress antinuclear antibody formation. Although it is clear that the ability to N-acetylate procainamide is not the sole factor controlling its ability to induce antinuclear antibodies in this genetic mouse model, it is equally clear that acetylator phenotype does play a role in this regard.
The conversion of compounds such as AF and BZ to metabolities capable of interaction with cellular constituents is important with respect to their carcinogenic properties (Miller, 1970;Miller, 1978). Metabolic activation of aromatic amine carcinogens is a complex process which is generally recognized as involving a number of enzymatically mediated steps, including N-hydroxylation and subsequent esterification. In addition, these compounds may also undergo N-acetylation in a number of different species, including hamster, guinea pig, rabbit, mouse, and rat (Lotlikar and Luha, 1971;Lower and Bryan, 1973;Morton et al., 1979). The dog, incapable of acetylating certain arylamine drugs (Williams, 1959), is also incapable of acetylating these carcinogens (Lower and Bryan, 1973;Poirier et al., 1963).
In human and rabbit populations, certain aromatic amine carcinogens such as AF, BZ and 2-naphthylamine are acetylated by the same polymorphic N-acetyltransferase that is responsible for the acetylation of arylamine drugs (Glowinski et al., 1978). This has led to the speculation that rapid and slow acetylator populations may differ in susceptibility to aromatic amine carcinogen-induced chemical carcinogenesis, paralleling the situation for the drug-induced toxicities (Glowinski et al., 1980). One preliminary report suggests that slow acetylators may be at greater risk of urinary bladder tumors from  L~~~~ et al., 1979), distilled water and allowed to hemolyze for 3 min. One-half ml of although further studies in appropriate human chilled 50 mM potassium phosphate buffer, 1 m~ dithiothreitol, pH are necessary to establish this. In an attempt to develop new ~i~~~-~~b a l s were w e d by decapitation, livers removed and animal models for the N-acetylation polymorphism and homogenized by hand in 4 volumes of 50 m~ potassium phosphate thereby possibly assess its effect on tumor susceptibility, these buffer, 1 mM dithiothreitol, pH 7.4, in a ground glass homogenizer. studies were undertaken to analyze the genetic basis of strain Homogenates were centrifuged a t 1 0 , m X g for 20 min, and the differences in aromatic amine carcinogen N-acetylation in resulting at 105~000 x g for h. inbred mice.
Determination of Acetylating Activity in Vitro 7.4, was added and the hemolysate kept on ice.

EXPERIMENTAL PROCEDURES
Chemicals a n d Reagents Reagent grade sulfamethazine (free acid) was obtained from Nutritional Biochemicals, Cleveland, OH; p-aminobenzoic acid from Sigma; acetyl coenzyme A, lithium salt from P & L Biochemicals, Milwaukee, WI; [~cetyl-~H]coenzyme A (3 Ci/mmol) and Omnifluor from New England Nuclear; 2-aminofluorene from K & K Labs, Plainview, NY; and benzidine from Matheson, Coleman-Bell, Cincinnati, OH. All other chemicals and reagents were reagent grade.

Animals
Mice used in the strain survey (Table I) were obtained from The Jackson Laboratory, Bar Harbor, ME, through the help of Dr. Loren Skow. Mice of the X/Gf strain were donated by Dr. Anna Goldfeder, Columbia University, NY. Fl, F2, and backcross animals used for genetic analysis were bred at The University of Michigan, Department of Pharmacology from parental A/J and C57BL/6J mice received from The Jackson Laboratory. The recombinant inbred strains of mice designated BXA and AXB, derived by inbreeding the F2 generation of crosses C57BL/6J X A/J (Taylor, 1976) were obtained through the generosity of Dr. Muriel Nesbitt, Department of Biology, University of California, San Diego, CA.
All animals were housed 2 to 8/standard shoe box cage with food (sulfa-free Purina Mouse Chow) and water ad libitum. Bedding consisted of cedar wood shavings which were changed once per week. Animals were maintained on a light-dark schedule with lights on from 8 AM to 8 PM. All mice were 8-24 weeks of age at the time of experimentation.

Tissue Preparation
Blood-Whole blood was collected by orbital sinus puncture in a 0.075-ml heparinized hematocrit tube. Fifty pl were added to 0.5 ml of Colorimetric Assay for Acetylation ofp-Aminobenzoic Acid The amount of PABA N-acetyltransferase activity was determined by the procedure of Hearse and Weber (1973), which is a micromodification of the diazotization procedure of Bratton and Marshall (1939). Ail incubations were carried out at 37 "C in capped polyethylene microtest tubes (0.4 m l ) (Kew Scientific, Columbus, OH). The reaction mixtures (0.09 m l ) contained 0.05 ml of suitably diluted enzyme, aqueous CoASAc (0.02 ml of a 10 m~ solutioh), and an aqueous substrate solution (0.02 ml of PABA, 0.2 mM). Control tubes contained no CoASAc. Following diazotization (Hearse and Weber, 1973), absorbance was measured at 540 nm in 1-ml cuvettes of 1-cm light path against a water blank in a Beckman 35 spectrophotometer. The extent of acetylation was obtained by subtracting the experimental reading from the control reading. Under the conditions of this assay a decrease of 1 absorbance unit corresponds to the acetylation of 6.08 nmol of PABA. Specific activities are expressed as nanomoles of acetylated product formed/minute/milligram of protein, unless otherwise indicated.
Blood-Whole blood lysates (as described under "Tissue Preparation") were diluted 2.25-fold with 50 mM potassium phosphate buffer, 1 mM dithiothreitol, pH 7.4, and used as enzyme source. Incubations were carried out for up to 3 min.
Liver-Liver cytosol was diluted 20-fold with 50 mM potassium phosphate buffer, 1 m~ dithiothreitol, pH 7.4. Incubations were carried out for up to 1.5 min.
polypropylene micro sample tubes (1.5 m l ) (Kew Scientific, Columbus, OH). The final concentrations of BZ, AF and SMZ in the assay were 0.25 mM and that for CoASAc was 0.5 mM. BZ and AF were dissolved in M e 8 0 and SMZ in 50 m M potassium phosphate buffer, pH 7.4. All incubations were carried out at 37 "C. Incubations were terminated by addition of 1 ml of N-ethylmaleimide (2 mM) in ethylene dichloride. After extraction of acetylated product, an aliquot of the organic layer was counted. Under the conditions of this assay, 5500 dpm corresponded to the acetylation of 1 nmol of substrate and 0.042 nmol were detectable. Values were corrected for both the counting efficiency of the counter (42.5%) and the extraction efficiency of each substrate (Glowinski et al., 1978). Specific activities are expressed as nanomoles of acetylated product formed/minute/milligram of protein, unless otherwise indicated. Blood-Whole blood lysates (as described under "Tissue Preparation") were used as source of enzyme. Incubations were carried out for up to 12 min.
Liuer-Cytosol was diluted 2-fold with 50 m M potassium phosphate buffer, 1 m M dithiothreitol, pH 7.4. Incubations were carried out for up to 6 min for BZ and AF. For SMZ, incubations were carried out for up to 15 min.
Treatment of Inheritance Data For Figs. 1, 2, and 4, additive inheritance of N-acetyltransferase genes in both parental strains was used as a fist approximation. The calculated value for N-acetyltransferase activity in the F1 generation was thus half-way between the two parental mean values. For classification of animals according to genotype and for comparison of observed and expected segregation ratios, a line was drawn at the midpoint of the means between the A/J and the calculated F1 values, and also between the F1 and C57BL/6J values.
Protein Determinations Blood protein concentrations were determined by the method of Warburg and Christian (1941). Liver protein concentrations were determined by the modified method of Lowry et al., 1951.

RESULTS
Strain Survey of N-Acetyltransferase Actiuity-Twenty selected inbred strains of mice were surveyed for initial Nacetylation rates of a number of aromatic amine substrates in liver and blood (Table I). This data reveals wide interstrain variation in N-acetylation in these tissues. A/J, AHe/J, and X/Gf mice have no detectable blood BZ N-acetyltransferase activity, while the other strains tested have clearly distinguishable activity for this substrate (0.020-0.060 nmol/min/ mg). Liver BZ N-acetyltransferase, also highly variable among these strains, correlates significantly with blood BZ activity (p < 0.001). A/J and AHe/J mice have very low activity (0.058 and 0.032 nmol/min/mg) while the other strains tested have considerably higher activities for this substrate. The blood Nacetyltransferase activity pattern of PABA follows that of BZ. A/J, AHe/J, and X/Gf strains exhibit very low blood Nacetyltransferase activity (0.045-0.112 nmol/min/mg) and the other strains exhibit activity which is -10-fold higher.
The acetylation of PABA by mouse liver cytosol reveals a different pattern. This activity is higher in all strains tested than that seen in the livers of either man or rabbit. In contrast to the situation in mouse blood, however, the interstrain variation in liver PABA activity is relatively small (17.3-35.4 nmol/min/mg). The A/J and AHe/J mouse strains appear to be on the low side of this range and are only 1-2-fold lower than the other strains tested. Liver SMZ N-acetyltransferase activity is at the limit of sensitivity for this assay and thus, no differences between strains were detected. It is clear that mice can be classified as rapid or slow acetylators from blood BZ or blood PABA N-acetyltransferase activities since either is correlated with liver BZ activity. despite the fact they have high blood PABA activity; the SF strain has high blood PABA and blood BZ activities, yet has a low liver PABA activity; and the X/GF strain, which has low activity like the "slow" acetylator A/J and AHe/J strains, for both blood PABA and BZ N-acetyltransferase, exhibits intermediate liver BZ activity. Thus, there are both tissuespecific and substrate-specific differences between strains in the expression of N-acetyltransferase activity.
Inheritance of N-Acetyltransferase Activity in A / J a n d

C57BL/6J
Mice-Mice of the A/J (slow acetylator) and C57BL/6J (rapid acetylator) strains were chosen for the analysis of inheritance of N-acetyltransferase. AF, an alternate substrate for BZ, was used for the inheritance studies. Assays of N-acetyltransferase with AF in tissue preparations from the two strains revealed no differences from BZ with respect to N-acetylation. All animals were phenotyped for PABA and AF activity in blood several days later they were killed and liver N-acetyltransferase determinations were performed using the same two substrates. Thus, the same mice are represented in each of Figs. 1-4. Fig. 1 shows the segregation of the N-acetylation capacity for PABA in blood of A/J, C57BL/6J, AC57F1, AC57F2, and backcross animals. There is -16-fold difference between the A/J (mean 100 pmol/min/mg) and C57BL/6J (mean 1660 pmol/min/mg) parental strains, and no overlap between the two. The mean blood PABA N-acetyltransferase activity in AC57F1 hybrids is 1160 pmol/min/mg which is clearly intermediate (the calculated value is 880 pmol/min/mg) (see "EXperimental Procedures"), but nearer the activity of the C57BL/6J parental strain. Identification of the phenotypically "slow" acetylator among AC57F2 and backcross animals presented no problem as this phenotype is clearly distinct from the others. However, overlap between the intermediate and "rapid" phenotypes makes definitive assignment of some animals uncertain. The 46 animals in the AC57F2 generation segregate into three groups in a ratio of 1ow:intermediate:high of 1013:23, which obviously departs from the expected 1:2:1 ratio (x* = 16.1, p < 0.001). Although the overlap in the C57 x AC57F1 backcross animals makes phenotype assignment rather imprecise, both sets of backcross animals appear to segregate in a 1:l ratio (C57 X AC57F1: 2 = 0.81, p 0.5; A X AC57F1: 2 = 0, p < 0.5), and are thus not significantly different from that which is expected. Fig. 2 shows the segregation of N-acetylation capacity for A F in blood of A/J, C57BL/6J, AC57F1, AC57F2, and backcross animals. A pattern similar to that seen for blood PABA is observed. The AF N-acetyltransferase activity in A/J mice is, without exception, nondetectable by the conditions of the assay. C57BL/6J mice have a mean activity of 110 pmol/min/ mg with a 2-fold variation around this mean, yet the two phenotypes do not overlap. The mean AF N-acetyltransferase activity in AC57F1 hybrids is 65 pmol/min/mg which suggests the absence of dominance effects; the calculated value is 55 pmol/min/mg (see "Experimental Procedures"). However, there is considerable overlap between the F1 hybrids and the C57BL/6J parents. The AC57F2 animals exhibit wide variation and a definitive assignment of phenotype of every animal is again difficult. In addition to overlap between the "rapid" animals (2 = 0 , p < 0.5), which agrees with the expected ratio.
Liver PABA N-acetyltransferase activities in A/J, C57BL/ 6 J , AC57F1, AC57F2, and backcross animals are shown in Fig.  3. The pattern of activity for PABA in liver is clearly different than that for PABA in blood (Fig. 1). Although all liver PABA activities appear to be unimodal, there is a significant difference between the A/J (mean of 18.8 nmol/min/mg) and C57BL/6J (mean of 22.6 nmol/min/mg) parental types, which will be shown definitively in the RI strains (Table 11).
Liver AF N-acetyltransferase activities in A/J, C57BL/6J, AC57F1, AC57F2, and backcross animals are shown in Fig. 4. A/J (mean of 134 pmol/min/mg) and C57BL/6J (mean of 1260 pmol/min/mg) parental types exhibit -10-fold difference in the levels of AF N-acetyltransferase, and there is no overlap between them. There is -2-fold variation in the liver AF Nacetyltransferase activity in the C57BL/6J parental types. This is comparable to that seen for AF and PABA activities in blood (Figs. 1 and 2). The mean activity in AC57F1 animals is 103 pmol/min/mg which is consistent with additive inheritance (the calculated value is 697 pmol/min/mg) (see "Experimental Procedures"). However, there is some overlap FIG. 4. Inheritance of liver aminofluorene N-acetyltransferase activity in A/J, C57BL/6J, AC57F1, AC57F2, and backcross animals. Liver cytosol, prepared as described under "Experimental Procedures," was diluted 2-fold. Aminofluorene N-acetyltransferase activity was determined by the radioassay at 37 "C for up to 6 min. Final concentrations of CoASAc and AF were 0.5 and 0.25 mM, respectively. Initial velocities were calculated by extrapolating the time-activity curves back to time zero. Dotted lines were drawn as described under "Experimental Procedures." between the intermediate type and both the low and high parental types which makes classification of individuals in the F2 and backcross generations more difficult. AC57F2 animals segregate into three classes with a ratio of 1ow:intermediate: high of 10:28:8; this ratio is not significantly different from the expected 1:2:1 ratio (x' = 2 . 3 5 ,~ < 0.5). Despite the overlap of the AC57F1 animals with both parental types, backcross animals clearly segregate into the expected 1:l ratio. A ratio of 9 1 2 was observed in the C57 X AC57F1 animals (2 = 0.81, p < 0.5), and a ratio of 1O:lO was observed in the A X AC57F1 animals (2 = 0, p < 0.5).
Thus, the distributions of phenotypes in the F1, F2, and backcross generations of A/J and C57BL/6J mice are consistent with the hypothesis that AF N-acetyltransferase activity in liver and blood and PABA N-acetyltransferase activity in blood are all controlled by a single major gene. The fact that in some instances, the observed ratios deviate from that which is expected provides evidence for the existence of modifying influences.
Inheritance of N-Acetyltransferase Activity in A/J and C57BL/6J Recombinant Inbred Strains- Table I1 summarizes N-acetyltransferase activities for PABA in liver and blood of A/J, C57BL/6J, and AC57F1 mice, as well as 11 RI strains derived by inbreeding the F2 generation of A/J and C57BL/6J crosses. The RI strain distribution patterns can be divided into two principal classes with respect to blood PABA N-acetyltransferase. Five strains (BXA-1, AXB-2, -5, -6, and -17) have high activity resembling the C57BL/6J progenitor, while the other six strains have low activity resembling the A/J progenitor. These results provide additional evidence for

I1
Liver and blood p-aminobenzoic acid N-acetyltransferase activity in A / J , C57BL/6J, AC57Fl and A / J X C57BL/6J recombinant inbred mouse strains Details of preparation of 105,000 X g liver cytosols and blood hemolysates and of N-acetyltransferase assays are described under ''Experimental Procedures."   the single gene inheritance hypothesis of blood PABA Nacetyltransferase. The pattern of liver PABA N-acetyltransferase is inconsistent with additive inheritance, although A/J animals have a slightly lower value (18.8 nmol/min/mg) than do C57BL/6J animals (22.6 nmol/min/mg). However, the RI strains that resemble the A/J parent for blood PABA Nacetyltransferase activity (BXA-3, -6, -15, AXB-3, -4 and -9) appear to have -2-fold lower liver activity for this substrate than the RI strains that resemble the C57BL/6J parent, with the exception of RI strain BXA-3. Thus, there may be modifying genes operating in A/J mice that enhance liver PABA N-acetyltransferase activity. Table I11 summarizes AF N-acetyltransferase activities in liver and blood of A/J, C57BL/6J and AC57F1 mice as well as 11 RI strains. For both tissues, the RI strains can be divided into two distinct classes. Five strains (BXA-1, AXB-2, -5, -6, and -17) have high AF activity resembling the C57BL/6J by guest on July 10, 2020 http://www.jbc.org/ Downloaded from parent, while the other six strains have low activities for this substrate resembling the A/J parent. The strain distribution pattern is identical to that seen for blood PABA N-acetyltransferase activity (Table 11). Thus, the probability is high that there is a significant relationship between the expression of N-acetyltransferase activity for these two substrates in both liver and blood. These results also provide evidence for the existence of modifying genes. The liver AF N-acetyltransferase activity in the AXB-3 strain (0.059 and 0.061 nmol/min/ mg), is approximately one-half that of the A/J parent (0.134 nmol/min/mg), although it is still of the low type (Table 111). As in the case for liver PABA activity (Table 11), A/J mice may possess a modifying gene which acts to enhance this activity. The C57BL/GJ-like RI strains (BXA-1, AXB-5, -6, and -17) possess higher blood AF N-acetyltransferase activities (0.162 to 0.173 nmol/min/mg) than the C57BL/6J progenitors, although none are as high as that seen in the AXB-2 strain. This may provide evidence for the existence of modifying genes operating in the C57BL/6J mouse which lowers its AF activity in blood.

DISCUSSION
The concept that N-acetylation may be capable of modulating the susceptibility to carcinogenesis is supported by a number of observations. Dogs develop tumors of both liver and bladder upon administration of acetylated aromatic amines, but only urinary bladder tumors when the unacetylated parent compound is administered (Poirier et al., 1963). Oral administration of 4-aminobiphenyl leads to primarily urinary bladder tumors (Walpole et al., 1954;Deichmann et al., 1958) while administration of 4-acetylaminobiphenyl leads to tumors of both liver and bladder (Jabara, 1963). In contrast, oral administration of 2-acetyl-naphthylamine fails to induce urinary bladder tumors (Conzelman and Flanders, 1972), even though the unacetylated parent compound 2-naphthylamine is a well established canine urinary bladder carcinogen (Deichmann and Radomski, 1963). Thus, it is possible that the capacity for N-acetylation may play a role in the susceptibility to aromatic amine-induced carcinogenesis.
Previous studies of N-acetylation of carcinogenic aromatic amines have dealt with the capacity of a species to perform this reaction, but comparatively little attention has been paid to host factors affecting individual variability of this metabolic pathway in any species. One major goal in surveying inbred mouse strains was to detect intraspecies variation in N-acetylation capacity which could be developed into new genetic models of the human acetylator polymorphism. These models may facilitate investigations of genetic differences in susceptibility to aromatic amine-induced carcinogenesis in animals and perhaps in man. They may also be useful in studies of genetic differences in arylamine drug metabolism and related toxicities.
A survey of 20 inbred strains of mice (Table I) indicates that high N-acetyltransferase activity is more prevalent than low N-acetyltransferase activity by a ratio of 6:l. Two of the low strains (A/J and AHe/J) have a common origin. The X/Gf mice are known for their low susceptibility to a number of potent carcinogenic agents, including the aromatic amine acetylaminofluorene (Goldfeder, 1974), and it would be interesting to learn more about other possible biochemical similarities of X/Gf to the A/J and AHe/J strains.
It is clear from Table I that there is a significant correlation between blood PABA and blood BZ activities ( p e 0.001).
Distributions of these activities suggest the presence of two acetylator phenotypes. This is especially noteworthy as the A/J, AHe/J, and X/Gf strains stand alone in having no blood BZ activity, and in having blood PABA activity which is at least 10-fold less than that seen in the other strains. The occurrence of genetically controlled differences in blood PABA N-acetyltransferase activity has been demonstrated in rapid and slow acetylator rabbits (Szabadi et al., 1978).
Liver N-acetyltransferase activity for SMZ, the prototype polymorphic substrate in both rabbit and man, was not detectable in mice and thus, no interstrain variation was observed. The level of liver N-acetyltransferase activity as measured by either BZ or PABA is quite different from that of SMZ. Liver BZ N-acetyltransferase activity is bimodally distributed with A/J and AHe/J animals exhibiting an -8-fold lower activity than the lowest of the rapid acetylator strains. It is interesting to note that the differences in SMZ N-acetylation capacity between rapid and slow human liver samples are approximately the same (Glowinski et al., 1978). The pattern of liver PABA N-acetyltransferase activity is similar, although the differences are 2-fold or less depending on the strains compared.
Liver BZ N-acetylation appears to correlate well with blood N-acetyltransferase activity for either PABA or BZ, with some exceptions. Such strains as ST/bJ and RII1/2J, which have approximately one-half the liver BZ activity of the other rapid acetylator strains, or as the X/Gf strain which has -4 times the liver BZ activity of the other slow acetylator strains, may prove useful in studies of variant forms of the N-acetyltransferase molecule.
The results shown in Table I are in complete agreement with those of Tannen and Weber (1979), although only four strains of inbred mice were surveyed using SMZ and PABA in that study. Their values for liver PABA N-acetyltransferase are slightly lower which may be due to the fact that the enzyme preparation used for assay was a 30,000 X g homogenate in contrast to the 105,000 X g cytosol used for the studies reported here.
The data on the inheritance of N-acetyltransferase in A/J and C57BL/6J mice presented in Figs. 1-4 shows that F1 are intermediate, F2 are divisible into three classes corresponding to the two parental types and an intermediate type, and backcross animals fall into the parental and intermediate types. This is consistent with the conclusion that a single major gene with two alleles is responsible for this activity in blood (AF and PABA) and liver (AF). The gene symbol Nat has been proposed for this variation with Nat' as the common rapid acetylator allele and Nat" as the slow acetylator allele.2 The data in Fig. 1 on the inheritance of blood PABA activity is in agreement with that reported by Tannen and Weber (1980a). The data in the RI strains (Tables I1 and 111) provide further support for the hypothesis of single gene inheritance.
The 2-fold variation in N-acetyltransferase activity in either blood or liver of the C57BL/6J parental types, irrespective of substrate, must be due to environmental influences or the presence of modifying genes. The RI strain data (Tables I1 and 111) indicate that an appreciable part of this variation may be attributable to the presence of modifier genes and that these genes segregate independently of the major Nat gene. While liver PABA N-acetyltransferase appears unimodal in all offspring of the appropriate crosses (Fig. 3), data in the RI strains (Table 11) reveals the existence of two groups of animals. RI strains BXA-3, -6, -15, and AXB-3, -4, and -9, which can be classified as A/J types from their blood PABA N-acetyltransferase data, all exhibit liver PABA N-acetyltransferase activity -2-fold lower than that seen in the C57BL/GJ-like strains. This is consistent with the presence of a modifier in A/J mice, increasing the apparent liver PABA N-acetyltransferase activity, which is not present in these RI strains. The data in Tables I1 and I11 also support the presence of modifier genes in the C57BL/6J parental types, but their effect is in the opposite direction of that seen in A/J mice.
The concept of repeatedly passing a selected gene from one inbred strain into the genome of another strain has recently been reviewed (Nebert, 1980). The ultimate congenic "inbred line" (developed after 120 generations) will have the selected gene "fured" and theoretically in more than 99.99% genome of the host background strain. Placement of the N-acetyltransferase-C57BL/GJ allele into the low acetylator strain and placement of the N-acetyltransferase-A/J allele into the high acetylator strain would clearly be of interest with respect to the question of modifier genes. The development of these lines within this laboratory is now in progress. They may provide an improved means to evaluate the effect of differences in acetylator phenotype on aromatic amine-induced tumorigenesis, and also to study the possible effect of genetic modifiers on this process.