Activity of enzymes and fitness variation

Relation entre l'activite enzymatique et les caracteristiques du fitness telles que la fecondite, le taux de developpement de l'œuf a l'adulte, le taux de developpement embryonnaire, la masse corporelle et la mobilite chez des souches de D. melanogaster selectionnees pour leur taux de developpement


Introduction
The question of adaptive significance of enzyme polymorphisms has recently been oriented to the problem of the phenotypes on which selection might act. This has pointed to the possible role of regulatory gene variation in the processes of evolutionary adaptation (e.g., Ayala and McDonald, 1980; Anderson and Gibson, 1985). Many studies have demonstrated that genetic variation of enzyme activities could be used to distinguish the effects of regulatory genes from those of structural ones (e.g., Gibson, 1970;Ayala and McDonald, 1980;Marinkovi6 et al., 1986). The variation is based on the differences in the amounts of given gene products, which could be explained by differences in regulatory genes, rather than by gene duplication. It has been suggested that variation of regulatory genes may provide an even more important source for adaptive evolutionary change than structural gene variation (Britten and Davidson, 1969;Macintyre and O'Brien, 1976;and others). A large amount of variation of enzyme activities has been documented in Drosophila species, even for monomorphic structural genes (Ward and Herbert, 1972;Mc Donald and Ayala, 1978;Laurie-Ahlberg et al., 1980;Van Delden, 1982;Marinkovi6 et al., 1984bMarinkovi6 et al., , 1986Marinkovid and Ayala, 1986).
In our previous studies, efforts have been focussed on the correlations between rates of preadult development and activity levels of a number of studied enzymes (i.e. G-6pd, 6-Pgd, a-Amy, Adh, a-Gpd, Hk, Idh, Me, Sod) in Drosophila melanogaster and Drosophila subobscura individuals (Marinkovid et al., 1984a, b;Marinkovi6, 1985;Milosevic, 1987). In progeny of wild individuals a significant difference has been found in activity levels between fast-and slow-developing groups. The fastest-developing group of both species had a majority of highly active enzymes. Specific patterns of intercorrelations between enzyme activities in fast, intermediate, or slow preadult developmental classes suggest that different regulatory gene variants with pleiotropic effects on multiple enzymes might influence the variation in developmental dynamics.
In the present paper, different fitness characteristics are investigated to discern multiple relationships between regulatory, metabolic, developmental, and phenotypic levels in D. melanogaster. A continuous 10-generation selection for extremely fast, and slow egg-to-adult developmental rate has been performed, and selected groups of D. melanogaster individuals have been analysed for enzyme activity, fertility, mobility, and body weight. To complete the information about studied correlations, we have also examined a sample descended from a natural population for enzyme activity variation, but from the aspect of differential fertility and body mass of their F i progeny.

Materials and Methods
The selection experiment was initiated with the progeny of about 300 wild D. melanogaster flies caught in June 1984 at Jastrebac Mountain, 150 km South of Belgrade. Starting from more than 2,100 such progeny (G-0 generation), continuous 10-generation selection for extremely fast and slow preadult development was performed under constant laboratory conditions (20°C, relative humidity ca 60%). Five groups of flies were run simultaneously for each line, each in 4 culture bottles with about 200-600 individuals per generation (see Table I). About 10°!° of the fastest (or slowest) developed individuals were transferred to new cultures and allowed to intercross with one an other; 25 such 7-day-old females were randomly chosen per replicate to initiate the following generation. They laid their eggs for 6 h in each of 4 250 cm 3 culture bottles with corn-yeast medium, so that development of their progeny occurred in non competitive conditions. To reduce inbreeding and genetic drift, flies were intercrossed among the 5 fast-line groups (as well as among the 5 slow-line groups), in every second generation. In the first intercrossing generation (G-1 ), 25 males from replicate 1 were placed in a bottle with 25 virgin females from group 2, 25 males from replicate 2 were placed in a bottle with that many virgin females from group 3, and so on. In subsequent intercrossing generations (G-4, G-6, G-8), flies from different replicate cultures were intercrossed such as to provide eventually for interchanges among all replicate cultures.
The obtained enzyme activity values were ajusted by the Lowry test to mg protein per ml solution (Lowry et al., 1951). These adjusted enzyme activity rates are proportional to the relative activities expressed in optical density units, as well as to the values adjusted on the mg of body mass.
At the termination of the selection experiment, several characteristics were measured in both selected groups, most of them simultaneously. The offspring of these selected lines were analysed for the rates of embryonic development, by counting the number of emerged first instar larvae at 2-h intervals. Larvae were hatched from eggs collected at 6-h intervals, in small Petri dishes with cornyeast medium. After that, the larvae pupated inside 200 cm 2 bottles, and the rate of eclosion or total preadult development, was also measured. Each selected line of such experiments consisted of 5 replications. i.e. of more than 2,000 individuals. The randomly collected samples of &dquo;fast&dquo; and &dquo;slow&dquo; flies were tested for fertility at the age of about 5 d from eclosion, and other samples were tested for individual mobility, as well as for longevity. Body weight was also measured individually.
Another year's sample of D. melanogaster flies, F, progeny from the same Jastrebac Mountain natural population, collected in June 1985, were investigated (almost synchronously with the previous studies) for the relationships between some of the analysed fitness characteristics and enzyme activity variation. These characteristics are : female fecundity, body weight and rate of embryogenesis. Here the enzyme assays were performed in smaller samples of 10 flies with certain extreme phenotypes, so that reaction rates of 7 enzymes might relatively differ from reaction rates obtained by previously used homogenates with 100 flies each in our selection experiment.

Results
Fig. 1 presents the average developmental time in 2 lines of 10-generational selection for extremely different rates of egg-to-adult development. Table I gives the numerical results obtained in 5 replicates, from G-0 to G-10. It is evident that selection progress includes some oscillations of the mean developmental times which might be explained by different environmental effects on the selected phenotypes (Botella & Mensua, 1986). However, after G-7, the divergence became relatively established (P < 0.001), and increased up to 60 h between fast and slow lines. A linear regression analysis including all replicates of fast and slow selection lines from G-0 to G-10, led to estimates of heritability H F 00.123 (c i -0.0078, c! -1.2573), and H2 = 0.185 ( Ci = 0.0011, c! -1.3020). Table 11 presents specific activities of 7 studied enzymes in G-0, G-1, G-5, and G-10 0 generations of selection for 2 different rates of preadult development of D. melanogaster. Despite the fact that some enzymes (such as a-Gpol, Adh, and Me), had relatively higher activities than other enzymes (such as Hl!, it can be seen that there is a significant difference in all of the studied enzymes between flies selected to be fast and those to be slow in their development. The combinations of studied enzyme activities are significantly different in 2 developmental groups of flies (measured by x 2 comparisons); this difference is especially pronounced in G-5 and G-10 generations of selection. Decreased activity occurs among flies with longer development, which is pronounced in 5 enzymes in g-1 and G-10 generations, and in 6 out of 7 enzymes in the G-5 generation of divergent selection. Fast/slow ratio is greater than 1 in these 3 sets of generational comparisons, but significantly so only in the G-5 generation, as well as when all comparisons are accumulated (t 2o = 2.43; P < 0.05). Table III presents the analyses of variances in activities of 7 enzymes (A) between and (B) within G-1, G-5, and G-10 generations of selection for fast and slow preadult development of D. melanogaster. In analysis (A), 3 levels of variation were studied : (1) between generations of selection, (2) between developmental lines, and (3) between enzymes controlled by structural loci from the 1st, 2nd, and 3rd chromosomes. There is a significant contribution of the selection process to the observed enzyme activity variation (F 2 , 3 = 20.9; P < 0.02), as well as of the fast and slow developmental phenotypes (F3,!2 2 = 12.4; P < 0.01 The enzyme activity variation of corresponding chromosomal groups of genes turns out not to be significant. In analysis (B), the mean values of 7 studied enzymes are adjusted (with their replicates) within fast and slow selected lines, showing a significant difference in G-5 and G-10 generations of selection. Fig. 2 presents dynamics of embryonic development measured simultaneously in 10 0 replications for each &dquo;fast&dquo; and &dquo;slow&dquo; selected line. This analysis was done in the progeny of the last selected G-10 generation. The average length of embryonic development was 27.9 ± 0.8 h in the &dquo;fast&dquo;, and somewhat longer in the &dquo;slow&dquo; line, i.e. 29.7 ± 0.9 h. There is a marginally significant difference in the dynamics of embryogenesis between these 2 groups of individuals (see also Marinkovid and Tucid, 1981;Smit et al., 1981 ). Fig. 3 presents the longevity studies. The average longevity of the fast line was 29.2 ± 2.7 d, ve. 30.5 ± 3.2 d for the slow line. A x 2 test shows a significant difference in variation of the 2 sets of individuals ( X 2 2 8 = 18.6; P < 0.05).
The measurement of fertility, which is clearly an important component of fitness in Drosophila, comprises female fecundity measured as number of eggs produced by a single female per 24 h. Female fecundity was insignificantly greater in the slow line (34.8 ± 4.2) compared to the fast-line flies (29.1 ± 3.2).
Mobility of adult flies was analysed among other fitness components, after the selection proceeded. About 400 individuals were investigated by means of the model of a double maze with 5 chambers (Kerid, 1981 Table IV presents the results of such an experiment, where samp:es of adult flies were placed simultaneously in the starting chambers and allowed to move through the next chambers at 3-min intervals. It can be seen that flies selected for extremely fast egg-to-adult development moved farther in the maze than the slow group. The observed distribution along the maze, on 3 successive days, was analysed by the appropriate Chi-square method, which gave a significant difference between fast and slow groups. Here it should be mentioned that in an earlier experiment with D. subobscura, individuals with the slowest embryonic development were more mobile than those with the fastest development (Marinkovid and Milosevic, 1983). Table V presents the averages of adult body weights that were measured in the l0th generation of selection for fast and slow preadult development. The observed differences were marginally significant, and it might be concluded that the slowest group of flies had a larger body mass, compared to the fastest long-term selected individuals. The average body weight of flies with extreme rates of egg-to-adult development in the G-1 generation, in an earlier study, was found to be equal or higher in the fast group of Drosophila individuals (Marinkovi6 et aL, 1984b).
A separate investigation was conducted with non selected groups of D. melanogaster individuals on the relationship between naturally occurring variation of some adaptively significant traits, and the variation of enzyme activities as a metabolic property. These individuals were from the same natural population as the flies from the 10-generational selection experiment, but their parents were collected in the following season, i.e. summer 1985. Fig. 4 shows the average levels of enzyme activities in such groups with fast, medium, and slow embryonic development. The observed differences between developmental phenotypes were found to be significantly correlated to enzyme activity variation for Adh, a-Gpd, ldh, Me, and Hk. However, specific associations of activity levels for 7 studied enzymes could be observed among flies with a fast, intermediate, or slow embryonic rate of development, pointing to quite complex genetic-physiological relationships. Fig. 5 presents the average activities of 7 enzymes/mg protein/ml in samples of D. melanogaster flies, a progeny of wild females, that differed in average body mass. Three classes acording to body weight were obtained, each with 10 individuals, with minimal (x = 0.72 mg), medium (x = 0.89 mg), and maximal (x = 1.1 mg) weight. As can be seen from the figure, most of the enzymes were found to vary independently of body mass. Yet variations of 6-Pgd and Hk showed a marginally significant increase in the heaviest males. Only a-Gpd variation corresponds to body mass, i.e. males with minimal body weight had higher average activity of this enzyme per unit of body mass, than those with maximal body weight (x 2 = 217.8; df = 6, P < 0.001). Fig. 6 shows the variation of enzyme activities with respect to differential female fecundity. The experiment was performed on 3 samples, containing 10 females each, again the progeny of wild parents collected in summer 1985, that were tested for egg production individually at 24-h intervals. A group of such females that had produced 18 $ eggs on average was designated as &dquo;minimal&dquo; fecundity, a group with 32 eggs as &dquo;medium&dquo; fecundity, and 61 eggs as &dquo;maximal&dquo; fecundity group. The minimal fecundity group had significantly higher average enzyme activities of Adh, a-Gpd, G-6pd, idh, and Me. The medium fecundity group was very similar to the minimal, except for G-6pd and ldh. This result might be explained as the possible consequence of the egg production processes at metabolic level. Also, a large amount of variation was observed in assays of D. melanogaster females, and it was preferable to use males for enzyme assay procedures (Stam and Laurie-Ahlberg, 1982). In this study we analysed males in all other experiments. Assuming that observed differences in fecundity versus enzyme activity have some adaptive meaning, that might be the possible force maintaining polymorphism of structural and regulatory genes in this species.
In this sample of non selected flies (which corresponds to the G-0 generation in our selection experiment), the evidence of allozyme frequencies for studied gene-enzyme systems was also obtained. Only a-Gpdh and Hk-2 loci turned out to be polymorphic (the frequencies of their 2 commonest alleles are about 0.7 and 0.3), while other loci were almost monomorphic (Adh : 0.98 vs. 0.02; 6-Pgdh : 0.97 vs. 0.03), or completely monomorphic (G-6pdh, Me, Idh). This might confirm the hypothesis that the differences in enzyme activity of Adh, 6-Pghd, G-6pdh, Me, and ldh found between lines selected for fast and slow rates of development might be due to the differences in regulatory genes. However, in our previous study of F, progeny from another D. melanogaster population (from Titova Mitrovica, September 1984, Marinkovi6 et aL, 1986 it was found that the differences in genotypic constitution between &dquo;fast&dquo; and &dquo;slow&dquo; individuals are highly significant at the a-Gpdh, 6-Pgdh, Adh (as well as at Sod, Aldox and Acph), suggesting that they could also be attributed in this case to structural variation at the corresponding loci (see also Cavener, 1983).

Discussion
Rate of development in Drosophila is proposed as an important component of fitness (Dobzhansky et al., 1964), which includes activities of many genes, and has a relatively low heritability (e.g., Brncic and Budnik, 1974;De Oliviera and Cordeiro, 1981;Marinkovic and Tucic, 1981 ).
In the G-0, G-1, G-5, and G-10 generations of our selection for fast and slow preadult development, the assay procedures for 7 enzymes were applied in corresponding groups of D. melanogaster individuals. As has been already described in Results, the combinations of studied enzyme activities were found to be ostensibly different in 2 selected groups of flies, which was especially pronounced in the G-5 and G-10 generations of this selection. Greater activity of a majority of enzymes was found in the fast group and, vice versa, lower activity was observed in the slow groups of flies. These results, as well as our previous studies (e.g., Marinkovi6 et aL, 1984bMarinkovi6 et aL, , 1986 might confirm the hypothesis that greater enzyme activity may speed up the growth rate of some individuals, and that a lower activity, on an average, slows down the growth rate. Specific patterns of intercorrelations between enzyme activities definitely exist in each of the preadult developmental classes. Associations of enzyme activity variation and other fitness components have also been investigated in this analysis. Many studies have shown the importance of regulatory gene variation, influencing enzyme activities (e.g. Britten and Davidson, 1969;Gillespie and Langley, 1974;Stein and Stein, 1976;McDonald and Ayala, 1978; and so on). The difference in the amount of enzymes might be due to a modification of the rate of synthesis of a polypeptide, or to the rate of degradation by specific binding of macromolecules at control sites adjacent to the structural loci. Cluster et al. (1987) found a larger rDNA activity in D. melanogaster flies with extremely fast preadult development, and with 3 out of 4 studied enzymes showing a greater activity than among individuals with the slowest egg-to-adult development. There is also experimental evidence that interactions between chromosomes, through regulation of the rate of transcription of certain structural genes, or through post-transcriptional or post-translational processes, should be taken into account (McClin 7 1965;Maclntyre and O'Brien, 1976;McDonald et al., 1977;Cochrane and Richmond, 1979, etc.).
Significant difference in activity rates of most studied enzymes in our analysis could be observed in samples with different preadult developmental rates (Table II), as well as with different body weight (Fig. 5), female fecundity (Fig. 6), and rate of embryonic development (Fig. 4). This suggests that dynamics of analysed enzymatic processes could be the basis of differences in fitness components of studied D. melanogasterflies. In the last generation of selection several fitness characteristics were compared, and some were specifically associated with the rate of egg-to-adult development, as well as with the enzyme activity variation. Since there is no single conclusion that might be applied for all 7 enzyme activity variants, we will discuss the observed results separately.
ADH enzyme is one of the most investigated models of specific metabolic and adaptive significance. The 2 common alleles, AdhFand Adhs, influence the difference in enzyme concentration that has been presumed to be responsible for the observed variation in enzyme activity of Adh FF and SS structural genotypes, other than catalytic properties of protein products (Gibson, 1972). However, a regulatory gene has been mapped closely linked to the structural locus (Thompson et aL, 1977). Activity variation in Adh was measured during selection for extremely fast and slow egg-to-adult development, and significant differences were found between the developmental phenotypes with respect to analysed generation. However, fast/slow ratio in G-5 generation is greater (1.64) than in G-10 (0.68). A significant positive correlation was found between Adh and a-Gpd activity in the G-5 generation of selection, as well as in samples of differentially fertile females. Cavener and Clegg (1981) found that relative fitness of a-Gpd genotypes is strongly dependent upon the corresponding genotypes at the Adh locus. The lack of such correlation in the G-10 generation of selection, for example, might be attributed to interactions of other genes.
Among 7 analysed enzymes in D. melanogaster assays, a-Gpd is one of the most active. This is a cytoplasmic enzyme with several metabolic functions, very important in flight metabolism (O'Brien, 1972;Zera, 1981 During selection, the enzyme was somewhat more active in the fast line (fast vs. slow = 1.07 in G-5 and 1.11 in G-10, respectively). Significantly higher activity has been observed in females of minimal fecundity ver sus those of maximal (ratio 1.30), as well as in flies of minimal body weight (1.38).
A structural gene for the Hk enzyme, as Adh and a-Gpd, has been mapped on the second D. me/anogaster chromosome. HK is known as a polymorphic, glucose-metabolizing enzyme (Ayala et aL, 1972). Variation of its activity closely corresponds to the variation of Adh and a-Gpd in different developmental phenotypes during 10-generational selection, yet a group of males with maximal body weight has significantly higher activity.
It has been suggested that enzymes involved in the glucose metabolism cycle in Drosophila, such as Hk, ldh, and Me, tend to have lower variability (Gillespie and Kojima, 1968;Kojima et al., 1970). However, in the present study a high level of enzyme activity variation was found for these enzymes. ldh and Me structural genes were mapped on the third D. melanogasterchromosome, but there is no evidence of a correlation between the activity level of these 2 enzymes. ME tends to be more active in the fast selection line, i.e. fast/slow ration G-5 was 1.36 and 1.68 in G-10. Idh was significantly more active in the &dquo;slow&dquo; line (F/S ratio 0.82 in G-5, and 0.71 in G-10 generation). Both enzymes are highly active in the minimal fecundity class of females (ratio for Me is 1.70, and 2.66 for Idh). ' G-6-Pdh and 6-Pgdh are located on the first D. melanogaster chromosome, and both are known as polymorphic loci in this species. A strong epistatic interaction has been found between common variants of the loci (Bijlsma, 1978). There are differences between individuals that are hemiand homozygous for a common, versus null G-6-Pdh allele, but not upon the allelic state of 6-Pgdh (Hughes and Lucchesi, 1977). It is suppo-sed that enzyme activity levels might be influenced by closely linked genes that are acting as regulatory genes (Bijlsma and Van Delden, 1977). In our present study the variation of G-6pd and 6-Pgd has been very often alternative.
Product-moment correlation coefficients were calculated for enzyme activity averages in samples of D. melanogaster individuals with differential fitness properties. Table V summaries such relationships between 4 enzymes. The activities of some of the studied enzymes are significantly positively associated (e.g. ldh, and Me, ldh and a-Gpd, Me and a-Gpd, Adh and ldh, etc.). On the other hand, there is marginally significant negative association between Adh and G-6pd, as well as G-6pd and 6-Pgd.
Dissimilarity in the activity pattern of studied enzymes in samples with different preadult developments, body weights and female fecundities suggests that dynamics of analysed enzymatic processes could be the basis of differences in fitness components of D . melanogaster flies. The variability of corresponding regulatory genes and their interactions might be a model system for understanding some aspects of adaptive significance of enzyme polymorphism in Drosophila as well as in other organisms. Even in enzymes coded by monomorphic loci (and these are our study G-6pdh, Me, ldh, and almost so 6 -Pgdh and Adh), a large amount of variation could be maintained by balancing selection that was acting at the regulatory gene level.
The present results demonstrate that complex relationships between fitness characteristics and developmental, metabolic, and genetic properties could be evaluated. This knowledge may significantly change our understanding of how individual organisms and populational systems respond to evolutionary forces, and how complex genetico-physiological adaptations are built up during the processes of organic evolution.