Genetic Determination of Amylase Synthesis in the Mouse*

Extensive genetic variation in structure and rate of synthesis of pancreatic amylase has been identified among strains of mice. The relative rate of synthesis of amylase varies from 0.15 in strain YBR to 0.26 in strain C3H. The number of electrophoretic isozymes of pan- creatic amylase varies between one and four. In each strain with multiple amylase isozymes, a characteristic quantitative distribution of protein among the iso- zymes is observed. Isozyme proportions are determined by the relative rates of synthesis of each component. Congenic lines with different amylase phenotypes have been established. Genetic analysis reveals the close linkage of cis-acting sites determining rate of synthesis and electrophoretic mobility of mouse pancreatic amylase. Genetic variation affecting the rate of synthesis of specific proteins can used to identify the chromosomal locations and modes of action of regulatory elements in the mammalian genome. Mouse amylases may be useful gene products for such studies. The enzymes exhibit extensive genetic variation (1-4), they are synthesized in abundance in specific differen-tiated tissues, and they can he readily purified for biochemical analysis. In the mouse, pancreatic amylase and the homologous parotid enzyme encoded by closely linked duplicated loci are monomeric enzymes of molecular

Genetic variation affecting the rate of synthesis of specific proteins can he used to identify the chromosomal locations and modes of action of regulatory elements in the mammalian genome. Mouse amylases may be useful gene products for such studies. The enzymes exhibit extensive genetic variation (1-4), they are synthesized in abundance in specific differentiated tissues, and they can he readily purified for biochemical analysis.
In the mouse, pancreatic amylase and the homologous parotid enzyme are encoded by closely linked duplicated loci (2,5 ) . Both are monomeric enzymes of molecular weight approximately 56,000. The physical properties and developmental expression of mammalian amylases have recently been reviewed (6,7 ) .
Both quantitative genetic variation and qualitative (electrophoretic) variation affecting mouse pancreatic amylase have been described. Although the majority of inbred mouse stocks have only a single electrophoretic form of pancreatic amylase, double-banded phenotypes have been found among wild mice (2) and in a few inbred stocks (8,9). Improved electrophoretic techniques have recently made possible the resolution of up to four components of pancreatic amylase from some of these double-banded animals (4, 10). Quantitative variation is evident within the multi-banded animals, as the different isoen- zymes often occur in unequal amounts, resulting in electrophoretic patterns with "skew" proportions of the bands. Within a particular stock, the relative proportions of isozymes are constant, whereas marked differences are found among animals of different origin. Breeding experiments have demonstrated that this quantitative variation is under strict genetic control so that a given isozyme ratio is inherited as a Mendelian unit.
This study was initiated to determine whether the proportions of pancreatic amylase isozymes found in a given stock are a direct result of differences in production rate or whether other factors, such as differences in rate of degradation or secretion, may be involved. We have investigated the genetic determination of amylase expression in several mouse strains and will describe genetic variation affecting in vivo rates of amylase synthesis in the mouse pancreas.

RESULTS
Two Pancreatic Amylase Isozymes in Strain YBR-Most strains of mice contain a single electrophoretic isozyme of pancreatic amylase (8)(9)(10). YBR is one of the exceptions, with two pancreatic amylase isozymes demonstrable by agar gel electrophoresis.' The two isozymes, designated A, and BI, can also be separated by electrophoresis at pH 8.1 in polyacrylamide gels (Fig. 1). There is a consistent difference in the concentration of the two isozymes in individuals of this strain.
When pancreatic homogenates or purified amylase from YBR mice are electrophoresed and the gels stained for amylase activity, the Bl form predominates. However, the specific activities of the purified AI and BI isozymes do not differ.:' The relative amounts of protein in the two isozymes in pancreatic homogenates was quantitated by densitometry of protein-stained gels (Miniprint Fig. 1). Approximately 60% of the amylase protein is associated with the BI isozyme; the ratio of To determine the basis for the unequal concentrations of these isozymes, we compared their relative rates of synthesis. The numbers below the gels are the overall percentages of amylase synthesis in these lines. When a percentage is not shown, it is approximately 25%.
Amylase was labeled in vivo for 15 min with ['Hlleucine. The radiolabeled amylase was purified as described under "Materials and Methods." The purity of the amylase preparations was evaluated by electrophoresis in the presence of sodium dodecyl sulfate (Miniprint Fig. 2). A single protein band, of the expected molecular weight, is present. Isozymes AI and B1 were separated from the purified labeled amylase by electrophoresis on cylindrical gels at pH 8.1. When the gels were sliced and counted, two peaks of radioactivity corresponding in position to isozymes AI and B1 were evident (Miniprint Fig.  3). From the amount of radioactivity associated with each peak, the relative rate of synthesis of the two amylases could be calculated. In eight experiments, the per cent of label associated with the B, isozyme was 58 f 3% (Table I). Varying the labeling period between 2 min and 90 min did not change this percentage. Since this value is very close to that determined by densitometry of the protein-stained gels,'it appears that the unequal proportions of isozyme AI and B1 in YBR pancreas is determined by their unequal rates of synthesis.

Znterstrain Differences in Total Amylase
Biosynthesis-Amylase is an abundant gene product, and a major component of pancreatic protein. We calculated the relative rate of amylase synthesis by dividing the radioactivity incorporated into amylase by that incorporated into total pancreatic protein after in vivo labeling with ["Hlleucine as described under Relative synthesis ofpancreatic amylase a n d of individual isozymes in mice of various genotypes A, Pancreatic amylase was isolated from individual animals after a 15-min period of incorporation of [3H]leucine in vivo as described under "Materials and Methods." Isozymes were separated by electrophoresis -at pH 8.1. After staining with Coomassie blue, gel slices containing individual isozymes were solubilized and their radioactivity determined. The counts per min associated with each band was divided by the total in all bands on the gel to give the percentages reported here. The amylase from each individual was analyzed on duplicate or, in most cases, on triplicate gels. B, Pancreatic amylase was purified from animals after a IO-min in vivo labeling with ['Hlleucine. The amount of trichloroacetic acid-precipitable radioactivity in amylase was divided by the amount in pancreatic homogenates to give the values for relative rate of synthesis. Values represent means f S.E. (n) where n is the number of animals assayed. n.d., not done.
"Materials and Methods." In strain YBR, amylase accounts for 15% of protein synthesized during a 15-min labeling period (Table I). This is unusually low for inbred mouse strains (10). Strain C3H/HeHa, which is typical, incorporates 26% of ['HI leucine into amylase during a similar labeling experiment ( Table I). The results in the congenic line C3H.AmyYBH demonstrate that this quantitative trait is encoded by a site linked to the amylase structural gene region. The expression of this genetic trait is additive in heterozygotes (Table I).
The magnitude of the interstrain difference can be explained by a 2-fold elevation of the absolute rate of amylase synthesis in strain C3H, assuming that the rates of synthesis of the nonamylase proteins do not differ. This assumption seems justified especially in the congenic C3H.AmyYBR line, which differs from C3H only in the amylase gene region. (See Miniprint for discussion and additional data on the interstrain difference.) A cis-Acting Site Determines the Znterstrain Difference-The isozyme pattern of C3H/HeHa mice is the common mouse pattern with a single pancreatic amylase isozyme, designated A*, which can be separated from the YBR isozymes by electrophoresis at pH 8.1 (Fig. 1). In heterozygotes from crosses of C3H and YBR mice, all three parental isozymes can be distinguished. We were therefore able to examine the rate of synthesis of each isozyme in heterozygotes.
If the two chromosomes in the heterozygotes contributed equally to amylase production, we would expect to find 50% of the heterozygote's amylase in the A2 isozyme, with the remainder divided between B1 and AI. However, when the protein concentration of the three isozymes was determined by densitometry, 70% of the amylase was of the AZ mobility, with 16% and 14% of the protein in isozymes B1 and AI, respectively. The relative rate of synthesis of the A2 isozyme was also twice as great as the sum of the AI and B, isozymes (Table I). It is evident that the protein concentrations of the three isozymes are determined by their individual rates of synthesis. The isozyme from the strain with the 2-fold greater rate of amylase synthesis is also synthesized in 2-fold greater Genetic Determination of Amylase Synthesis in the Mouse 375 quantities in the heterozygote, indicating that the trait is determined by a cis-acting site. Additional Amylase Phenotypes-several unique amylase phenotypes, differing in number and proportion of isozymes, were identified in a feral population. The characterization of congenic strains with these phenotypes is described in the accompanying Miniprint (Table I). All the investigated pancreatic amylase chromosome regions determine two types of information: the relative proportions of different isozymes and the overall rate of synthesis.
Expression of Isozymes in Newborn Mice-No major shift in the proportions of pancreatic amylase isozymes occur in the mouse after birth (see miniprint).

DISCUSSION
The observation that up to four pancreatic amylase isozymes are synthesized in homozygous mouse strains suggests that at least four structural genes may be active in these lines. Previous studies by heat inactivation and peptide mapping also revealed molecular heterogeneity of amylase within inbred strains (4). We have studied two kinds of quantitative elements encoded by cis-acting sites closely linked to the structural amylase genes.
Mice with one to four pancreatic amylase isozymes were identified by polyacrylamide gel electrophoresis of inbred stocks and feral populations. From some of the animals we have produced congenic lines on a C3H/As background. The lines with multiple amylase forms have distinctive proportions of isozymes in the pancreas. Within a line each amylase isozyme is synthesized in vivo at a unique rate which determines its contribution to the total amylase protein. In heterozygotes produced by crossing congenic lines with C3H, the parental ratios among isozymes are preserved. The three components of phenotypic variation encoded by the amylase gene region in congenic lines are summarized in Fig. 2.
Two alternative explanations for multiple amylase isozymes in homozygous mice are possible: either they are encoded by duplicated copies of the amylase structural gene, or they result from processing of a single precursor nucleic acid or peptide. Two lines of evidence suggest that the isozymes do not result from post-translational modification. First, the isozyme patterns of heterozygotes are co-dominant, rather than identical with those of either parent; parental patterns are expected in individuals heterozygous a t loci encoding modifying enzymes (15)(16)(17). Second, the consistent quantitative proportions of isozymes within individuals of a given stock is difficult to attribute to an enzymatic modification.
Strains C3H and YBR differ with regard to the basal rate of in vivo amylase synthesis. In C3H, 25% of the radioactivity incorporated into pancreatic proteins is associated with amylase, while the corresponding value in YBR is 15%. These high rates of synthesis are not unexpected, since amylase is known to account for 24% of pancreatic protein in the guinea pig (18) and up to 30% of the protein in chicken pancreas (19). In C3H X YBR heterozygotes, amylase synthesis accounts for 20% of the pancreatic protein synthesis and two-thirds of the labeled amylase is the C3H type. Thus, the interstrain difference in amylase synthesis appears to be determined genetically by a cis-acting site, with additive expression in heterozygotes. The concomitant transfer of the YBR electrophoretic phenotype and rate of synthesis during the 10 generations of backcrossing to produce the congenic C3H. AmyYBH strain demonstrates the close linkage or identity of the sites determining structural and quantitative characteristics of this enzyme.
It is possible that differences in the primary structure of the amylase isozymes could directly influence their rates of synthesis. However, examples of such effects are rare. We have studied two congenic lines, C3H.Amyw3 and C3H.AmyCE, which contain four amylase isozymes of corresponding electrophoretic mobilities, yet the isozymes are synthesized in different proportions in the two lines (see miniprint). It therefore seems unlikely that the quantitative differences are determined by the primary structure of the isozymes.
Both qualitative and quantitative variation in pancreatic amylase as represented here by the interstrain differences in patterns of isozyme expression, may be explained by variation in the number of amylase structural genes (20) or by variation at cis-acting sites (see discussion in miniprint). At present we are not able to distinguish between genecopy and regulation-sequence models. However, the functional variation we have found associated with the amylase gene region in these mouse strains make them ideal material for further molecular analysis.