Science 184(4132): 28-37 (5 Apr 1974)
Enzyme Polymorphism and Metabolism
Polymorphism among enzyme loci is related to metabolic function
George B. Johnson

Since the time of Darwin and Wallace, there have been arguments concerning the evolutionary significance of patterns of natural variation. Evolutionary biologists are now involved in a controversy over the question of whether or not genetic polymorphisms at enzyme loci are maintained by selection. Sufficient experimental data now exist to indicate that they are and to suggest their role in evolutionary processes.

Polymorphism and Selection

The current controversy over the selective significance of enzyme polymorphism has roots extending back several decades in the history of population genetics to the arguments of Fisher and S. Wright concerning the significance of genetic drift (1). Now, however, the same issues are being argued in a different context. In the absence of experimental data on the amounts of genic variation being maintained in natural populations, Kimura and Crow (2) suggested from theoretical considerations that the maintenance of variation should entail an evolutionary cost, or "genetic load," and that because of this, the total amount of polymorphism in natural populations may not be great; excessive genetic load would be expected to result in population extinction (3). However, with the advent of electrophoresis as a common tool for surveying genetic variation of enzyme loci (4), it has become apparent that the amount of polymorphic variation at the enzyme loci of natural populations is quite high (5-14), far higher than could exist if the original genetic load concepts were correct. In view of these results, such concepts could be maintained most simply by assuming that no selectively important differences exist among the electrophoretic variants. Thus, the argument had been advanced (15) that the variant proteins contain only minor differences in tertiary structure. which are sufficient to affect electroporetic mobility but not to affect significantly the functioning of the enzyme. Because electrophoretically different alleles are seen as functionally idntical, they are though to affect the organism's fitness identically, the differences among them thus being nautral to the action of selection.

Experimental evidence is now becoming available (16) which permits a test of the hypothesis that enzyme polymorphisms are selectively neutral. Most reported surveys of electrophoretic variation in natural populations have provided evidence of nonrandom processes. Biogeographic patterns are often reported to be uniform over wide geographic ranges (11, 12, 17), or to reflect parallel patterns of environmental variation (18--22). Analysis of the allele frequencies reported in these studies also reveals nonrandom processes (23), although confusion may arise when data obtained from diverse organisms are pooled (24). The force of such arguments is difficult to evaluate, however, b ecazuse of uncertainty about the possible involvement of migration (25), linkage (26), and founder effect (27). Neither is it clear when data concerning individual loci should be regarded as special cases and when they may be considered as illustrating a more general principle; these individual cases were selection is implicated at an allozyme locus do not necessarily argue powerfully for the generality of selective significance. It seems clear, however, that while some caution is necessary in evaluating the diverse array of information, much of the evidence is against Kimura's hypothesis.

Importance of Substrate Variability

There is a more general line of argument against the hypothesis of selective neutrality, however, which proposes that degrees of enzyme polymorphism reflect physiological function, some functional classes of enzymes being far more variable than others. No combination of linkage, migration, or breeding structure can render this relationship compatible with selective neutrality.

Two hypotheses have been advanced concerning the physiological role of enzyme polymorphism. The first of these, proposed in 1968 by Gillespie and Kojima (28), suggests that levels of enzyme polymorphism may reflect environmental variation in substrates; the second hypothesis, which I proposed in 1971 (29), suggests that these polymorphisms are often associated with regulatory reactions in metabolism.

Gillespie and Kojima (28) pointed out that in cultures of Drosophila ananassae, far less heterozygosity was observed at the loci of enzymes involved in energy production than at other enzyme loci (30). More detailed examination of natural populations by Kojima et al. (8) bore out their original observation (28) and these workers suggested that the greater variability in "nonglucose metabolizing" enzymes might reflect greater variability in their substrates, as many of these substrates originate in the external environment.

Assignment of environmental origin to the substrates of enzymes may be ambiguous when the functional roles of enzymes encompass both groups. Alcohol dehydrogenase and octanol dehydrogenase may both function in lipid metabolism, as well as possibly playing a role in the breakdown of dietary alcohols (31-36); most carboxyesterases probably function in Drosophila to break down secondary compounds of plants, but some esterases undoubtedly have more restricted internal metabolic roles. Kojima and his co-workers (8, 28) assigned loci whose enzymes utilize both classes of substrate to the "external substrate" class. It is important to realize, however, that in these cases the classes are not mutually exclusive.

In the several years since this dichotomy of enzyme types was proposed, there have been numerous electrophoretic investigations of enzymes, and the number of flies that have been collected from natural populations and examined in this way now exceeds 30,000. Extensive work has also been conducted on small vertebrates and on man. Upon analysis, these data strongly support the original hypothesis of Gillespie and Kojima (28). Among 13 Drosophila species, the enzymes of broad specificity, many of which utilize substrates originating from the external environment, are far more variable than those enzymes which utilize specific metabolically produced substrates (37).

Thus, polymorphic variation at enzyme loci correlates with the physiological variability with which the enzymes must interact in functioning. Although the groupings discussed above are broad and not always clearly defined, the general result seems valid for a wide variety of data.


  1. J. Gillespie and K. Kojima, Proc. Natl. Acad. Sci. U.S.A. 61, 582 (1968).
  1. G. Johnson, Nat. New Biol. 243, 151 (1973). If the mean number of alleles observed at a given locus in a given species is considered to be independent of results obtained for other species or at other loci, then for 145 independent samples of "external" loci compared to 147 independent samples of "internal" loci, the Mann-Whitney nonparametric test, corrected for tie rankings, yields a value of z = 8.03, which is significant at P < .001.

CybeRose: The paper goes on to give much detail and discussion of the complexities. The takeaway is that genetic diversity (i.e., polymorphism) is widespread and benetficial. A specimen is not a species. One plant cannot provide more than a small taste of what the broader species has to offer.

Genetic Polymorphism bibliography