Flux Differences Between Electrophoretic Variants of GGPD from Drosophila melanogaster

Demonstrating that naturally occurring enzyme polymorphisms significantly impact metabolic pathway flux is a fundamental step in examining the possible adaptive significance of such polymorphisms. In earlier studies of the glucose-6-phosphate dehydrogenase (GGPD) polymorphism in Drosophila melanogaster, we used two different methods, exploiting both genotype-dependent interactions with the 6Pgd locus, and conventional steady-state kinetics to examine activity differences between the two common allozymes. In this report we use l-I4Cand 6-'4C-labeled glucose to estimate directly genotype-dependent flux differences through the pentose shunt. Our results show the G6pdA genotype possesses statistically lower pentose shunt flux than G6pd' at 25". We estimate this to be about a 32% reduction, which is consistent with the two former studies. These results reflect a significant responsiveness of pentose shunt flux to activity variation at the GGPD-catalyzed step, and predict that the GGPD allozymes generate a polymorphism for pentose shunt flux. T HE enzyme and polymorphism for glucose-6phosphate dehydrogenase (G6PD; EC 1.1.1.49) in Drosophila melanogaster have been a focus of study in a number of investigations (YOUNG, PORTER and CHILDS 1964). This laboratory has examined the polymorphism in natural populations, focusing on the geographic variation (OAKESHOTT et al. 1983; EANES and HEY 1986), the potential to respond to targeted selection (EANES et al. 1985), assessing potential differences in in vivo (EANES 1984; EANES and HEY 1986) and in vitro function (EANES, KATONA and LONGTINE 1990), and restriction map variation in the G6pd locus region (EANES et al. 1989). Recently, the locus has been cloned, sequenced and shown to have 65% amino acid sequence homology with the human enzyme (GANGULY, GANCULY and MANNING 1985; HORI et al. 1985; FOUTS et al. 1988). G6PD catalyzes the first step in the oxidative pentose phosphate cycle, or pentose shunt. The pathway's role is involved in synthesis of NADPH for lipid synthesis and detoxification, pentose phosphate for nucleic acid synthesis, and triose phosphate for glycogen synthesis. Using radiolabelled glucose and adult D. melanogaster deficient for G6PD activity, GEER, BOWMAN and SIMMONS (1974) confirmed that the pentose shunt is blocked. Subsequent studies by GEER, LINDEL and LINDEL (1 979) estimated that about 40% of reduced NADP is contributed by this pathway, and the rate of lipidogenesis is tightly correlated with pentose shunt enzymatic activities, which can be induced under varying carbohydrate and lipid conditions. GEER et al. (1981) have reviewed the extensive Cknetics 132: 783-787 (November, 1992) nutritional studies on the pentose shunt that have been carried out in D. melanogaster. We have used two different methods to examine the hypothesis that activity differences exist between the two polymorphic allozyme variants of G6PD. Studies by other investigators suggested that the fast and slow electrophoretic alleles, designated here as A and B , possess different i n vitro activities (STEELE, YOUNG and CHILDS 1968; BIJLSMA and VAN DER MEULEN-BRUIJNS 1979; HORI and TANDA 1980, 1981 ; WILLIAMSON and BENTLEY 1983). This was supported by EANES (1984) where dramatic differences were seen in the ability of the G6pdA and G6pdB genotypes to suppress the lethality associated with a 6-phosphogluconate dehydrogenase (6PGD) mutation low in activity, designated 6Pgd'"'. In a G6pdB background this mutation has low viability, while it is nearly normal in the G6pdA background. This difference probably reflects a lower in vivo activity associated with the G6pdA genotype, resulting in a reduced accumulation of 6-phosphogluconate, a potent inhibitor of glycolysis. This observation was repeated in EANES and HEY (1 986) for a series of rare G6pd variants. We proposed that allele-specific differences in suppression reflected two clusters of in vivo activity, depending on whether the rare variant was derived from the A or B common allele. Finally, we have examined, by steady-state kinetics, the in vitro activity of the highly purified G6PD variants (EANES, KATONA and LONGTINE 1990). Our results predict that at 25" the G6pdA genotype possesses about 40% lower activity than G6pdB. While these studies both supported the hypothesis of activity differences, neither directly measured flux. 784 J. Labate and W. F. Eanes CAVENER and CLEGG (1981) compared the relative fluxes between the putative high and low activity dilocus allozyme genotypes at the G6pd and 6Pgd loci in D. melanogaster using the method of WOOD, KATZ and LANDAU (1 963). This method was originally proposed for assessing the apportionment of carbohydrate metabolism between the Embden-Meyerhof and pentose shunt pathways. It is based on the rationale that the first carbon of glucose (1-C) is lost as C 0 2 as it passes through the pentose cycle, while the number six carbon (6-C) is returned to the EM pathway. By using D-[ l-14C]glucose and ~-[6-’~C]glucose separately as substrates in parallel experiments, increasing pentose shunt activity is measured by 1-C/6-C derived I4C count ratios that are less than one. They observed statistically significantly greater pentose shunt flux associated with the proposed higher activity G6pdB6PgdS dilocus genotype, but the experiment was not specifically designed to partition the contribution of the genotypes at the G6pd and 6Pgd loci separately. In this report we extend the study of functional differences to this third independent method and focus on the G6pd genotypes. We estimate that the low activity G6pdA allele possesses about 32% lower pentose shunt flux than the G6pdB allele. MATERIALS AND METHODS Wild-type experimental lines: To generate lines that were either G6pdA or GGpd’, fixed for 6Pgd genotype, and randomized for the wild autosomal background, isofemale lines were collected from Davis Peach Farm, Mt. Sinai, New York, in 1989, full-sib mated and electrophoresed to determine mated genotypes at the G6pd and 6Pgd loci. Ten independent lines homozygous for G6pdA-6PgdF and 23 homozygous lines for G6pdB-6PgdF were isolated and pooled within genotype. Adults from these experimental composite populations provided the larvae for the flux experiments. Flux studies: Larvae were raised in axenic culture at 25 O for the duration of each experiment. All axenic stage transfers were carried out in a laminar flow hood. Adults from the mass lines were allowed to lay eggs for four hours on standard corn meal media covered with a dead yeast paste. Batches of several hundred eggs were collected from the surface by rinsing the eggs into a 20-ml beaker, and dechorionating by two washes in 10 ml of 2.5% sodium hypochlorite, with 0.1% Triton X-100. Next, eggs were transferred into a plastic syringe, where they were surface sterilized with 0.1 % benzyldimethyl-n-hexadecylammonium chloride (ROBERTS 1986), followed by 80% ethanol, and final rinses with sterile water and sterile 0.7% NaCl. Dechorionated sterile eggs were transferred to the surface of 2% agar plates. Newly hatched larvae (less than 20-22 hr old) were transferred into eight dram vials (50 larvae each) containing 5 ml of axenic Sang’s medium C (see ASHBURNER 1989) and 290 mM sucrose to induce pentose shunt activity (CAVENER and CLECG 1981). After 84 hr, samples of 10 larvae were transferred with wooden applicator sticks to 2-ml vials with 1.5 ml of identical medium, and 1 pCi of either D-[ 1-14C] glucose or D-[6-’4C]glucose (Amersham). After 48 hr larvae were collected with forceps and ground in samples of 10 larvae each in 300 liters of 0.05 M phosphate buffer (pH 7.4). Lipid and protein were extracted separately as de0.5 1 1 T