V. Genetic Background and Selection O N Different Carbohydrates

Frequency changes in amylase allozymes and patterns of tissue-specific expression of amylase have been monitored in laboratory populations of Drosophila pseudoobscura maintained on media in which the only carbohydrate source was maltose or starch. Nonrandom changes occurred in patterns of expression, whereas no patterns in allozyme frequency changes were discernible. The nature of the pattern changes was similar to an identical study done on populations derived from a natural population several hundred miles from the population used in the present experiments. However, in the previous study nonrandom changes in allozyme frequencies were also noted. Evidently, selection on the Drosophila amylase system differs depending upon the genetic background of the population. Furthermore, the evolutionary dynamics of structural gene variants and those regions controlling its expression may be independent, a result consistent with DNA sequence data. HAT natural selection molds adaptations in populations is nearly universally T accepted. However, the specific nature of the genetic variants contributing to adaptations is less well understood. Presently, there is controversy over the relative roles of changes in genes ( i .e . , amino acid substitutions) compared with changes in their regulation of expression in adaptive evolution ($ WILLS 1973; WILSON 1976). In an attempt to shed some light on the problem, we embarked on a research program designed to compare the evolutionary dynamics of naturally occurring polymorphisms of both types. The stimulus for this program was the seminal work of ABRAHAM and DOANE (1978), who clearly documented genetically determined variants in the regulation of a-amylase expression in the midguts of Drosophila. Electrophoretic variants (allozymes) of the structural gene, Amy, are also widespread, thus allowing such comparative studies. Previous studies in this series include confirmation of the genetic control of tissue-specific midgut patterns of expression in D. pseudoobscura (POWELL and LICHTENFELS 1979), a survey of geographic variation in frequencies of midgut patterns and allozymes (POWELL 1979) and interspecific comparisons (POWELL, RICO and ANDJELKOVIC 1980). Most relevant to the present report is a study of the effects of maintaining laboratory populations on medium containing different carbohydrates (POWELL and ANDJELKOVIC 1983). Replicas of the same founding Genetics 106: 625-629 April, 1984. 62 6 J. R. POWELL AND G. D. AMATO population were maintained for more than 2 yr on medium in which the only carbohydrate was starch (the substrate of amylase) or maltose (the product of amylase). Nonrandom changes in frequencies of both Amy alleles and midgut patterns of expression were observed. In particular, on starch medium the Amy F allele was favored, and a pattern of expression of amylase confined to the very anterior of the midgut was favored. On maltose medium, no consistent changes in either polymorphism were detected. A total of 14 populations, seven on each medium, was studied; all were derived from the same natural population, Bryce Canyon, Utah. T o test the robustness of the conclusion that these polymorphisms respond to selection in certain environments, we have repeated the experiments using a different base population. The results are both similar and different. MATERIALS AND METHODS The base population for the present experiments was a set of 11 isofemale lines begun by flies captured in the Chiricahua Mountains in southern Arizona during the summer of 1980; WYATT ANDERSON kindly sent us these strains. This sight is some 425 miles south of Bryce Canyon. D. pseudoobscura in this area, as in Bryce Canyon, are nearly monomorphic for the AR third chromosome. All 1 1 strains were checked at least twice to assure that they were homozygous AR; during the course of the experiments, samples were taken from the cages and chromosomes were reexamined. N o third chromosome gene arrangements other than AR were ever found. Details of the experimental design can be found in POWELL and ANDJELKOVIC (1983), and the description here will be brief. The 11 isofemale lines were mixed about equally and used to begin laboratory populations in cages which support a population of more than 3000 adults; at least 500 flies from the mixed population per cage were the founders. Two cages, designated I"' and I P , had medium in which the only carbohydrate sourcs was starch. Two cages, designated I"" and 11", had medium in which the only carbohydrate source was maltose. Periodically, egg samples were taken on standard cornmeal-molasses medium, and flies were developed on this medium at low densities and optimal temperature. Adults emerging were assayed for Amy allele frequencies by polyacrylamide electrophoresis, and the pattern of expression along the midgut was determined as described by ABRAHAM and DOANE (1978). Designation of midgut patterns follows DOANE (1980) and POWELL and LICHTENFELS (1979). The anterior midgut (AMG) is divided into three regions designated 1, 2 and 3; the posterior midgut (PMG) is divided into two regions, 1 and 2. Presence of activity is indicated by the number and absence by a zero. Thus, pattern 120-00 would have activity in the two most anterior regions but not elsewhere.