In studies involving segmental trisomy, designed to locate structural gene loci for enzymes, it was observed that trisomy for certain small regions in Drosophila and Datura caused reductions in the expression of some enzymes to a lower limit of approximately 67% of the euploid control (References are in the first report). The phenomenon appears to be the same as that which has been described for the dosage series produced by TB-1La in maize. This literature does not, however, contain any data on the effect of segmental monosomy on enzyme expression. The fact that monosomy (in maize at least) increases the expression of some enzymes above the euploid level allows one to eliminate the possibility that the trisomy effect on enzyme level is a result of reduced vigor in these aneuploids, since monosomics are less vigorous than the euploids. In other words, the change in expression of these enzymes does not correlate with the vigor of the plant; it does, however, negatively correlate with the ploidy of particular chromosomal regions. The order of enzyme activity from greatest to least is monosomic, disomic, trisomic, tetrasomic; for vigor the relationship is disomic, trisomic, tetrasomic, monosomic.

Several trends become apparent when one takes an overview of these data: 1) Almost every enzyme (about 15 different ones) studied is affected by some trisomic regions, which cause reductions to about 67% of the euploid. The only exception is isocitrate dehydrogenase in Drosophila. In this case, several regions caused slight increases (Rawls and Lucchesi, 1974, Genetical Research 24:59); 2) A single enzyme can be affected by multiple regions; 3) These regions are scattered throughout the chromosomes and do not appear to be restricted in their location.

The precise role of these regions in the expression of enzymes is, of course, not known. The facts that they are found in higher eukaryotes as diverse as maize, Datura and Drosophila, that they affect a wide spectrum of enzyme types, and that there are many such regions, suggest that these effects are a manifestation of a central mechanism of gene expression. Some possibilities to be considered are as follows:

1) Inverse effect (IE) regions contain genes that encode enzymes of certain bio- chemical pathways that produce metabolites that affect the expression of other pathways at some level. If a metabolite affects the rate limiting process of gene expression inversely to the concentration of the metabolite, the observed values would be realized. A difficulty for this hypothesis is the fact that there are multiple regions affecting any one enzyme. This would require that each metabolite affecting a particular gene would be operating in a very similar manner.

2) IE regions contain genes that encode for enzymes that degrade specifically certain enzymes or messenger RNA's. This seems unlikely since trisomic vs. disomic values do not usually surpass the .67 lower level. Since different enzymes are undoubtedly expressed at different levels, a change in the level of every individual degradative enzyme would not necessarily produce an inverse effect. It should, however, be considered and rigorously tested.

3) IE regions contain genes which compete against the one which is being monitored.

4) IE regions contain genes whose products are repressors. This has been proposed by Smith and Conklin. There can be little argument that IE regions produce a negative effect on the enzyme monitored. But repressors as we know them from prokaryotes do not show different levels of repression dependent upon their dosage. If repressors of this type were present in the normal euploid, one would not expect to be able to detect any enzyme. Repressors that function in a different manner remain a possibility.

5) IE regions are genes whose products are negative modulators of enzyme expres- sion such that the concentration of the modulator determines the level of enzyme expression. In order to produce such a close fit to an exact inverse effect in relation to their concentration, they must affect the rate limiting process of enzyme expression. If one believes that IE regions underlie the phenomenon of dosage compensation in Drosophila, they probably affect the process of transcription. Problems involved with the possibility that IE regions produce negative modulators are: a) The contrast of enzyme expression in an aneuploidy vs. ploidy series (to be discussed more fully later); b) In maize, multiple regions that affect ADH expression have been found (Birchler, unpublished). Yet no cis dominant, constitutive mutant of Adh has been discovered in a large unselective screen (Schwartz, personal communication, and Birchler, unpublished). This may be due to the difficulty, for unknown reasons, of inducing such a mutant; c) Some data suggest that IE regions act in a positive manner under certain circumstances (to be discussed more fully later).

6) IE regions contain genes that encode positive effectors of enzyme expression, but act in a negative way under the conditions in which they have been detected. The reasons for suggesting this possibility will be discussed in the section on dosage compensation.

It is obvious that none of these is completely satisfactory or even that all possibilities have been considered. There is no precedent for this phenomenon and an understanding of the mechanism involved will require much more research.

Regardless of their mechanism of action, the existence of presumptive IE regions on all the chromosomes of Drosophila that affect sets of genes on all the chromosomes is sufficient to explain a huge volume of data and observations on dosage compensation, autosomal sexual dimorphic mutants and sex determination. Dosage compensation in Drosophila appears to fit into the total picture of inverse effects Multiple regions which, when trisomic, reduce certain enzymes to about 67% of the euploid occur on the X and chromosomes 2 and 3. Thus if IE regions on the X affect structural genes on the X, males would have structural gene reductions to 50% of the female value, but since IE regions are only present once, the expression of the genes it controls would be approximately doubled. Hence, the affected genes would have a net expression in males nearly equal to females. In other words, an increase, due to the IE region, to approximately 200% of the female value is cancelled by a reduction of the number of structural genes to half that of females. Since suggestions of IE regions show up in data with such regularity, it is difficult to avoid such an explanation. This hypothesis will explain the observed compensation in metafemales [3/2 structural genes X 2/3 (the reciprocal of the dosage ratio of inverse regions) = 1.001 and in triploid intersexes vs. triploid females [2/3 structural genes in intersexes X 3/2 (the reciprocal of the dosage of inverse regions) = 1.00] (Muller, 1948, The Harvey Lectures; Lucchesi and Rawls, 1973, Genetics 73:459).

The IE regions hypothesis predicts that duplications of structural genes in males would produce more net gene expression than an extra dose (total 3) in females. Likewise, deficiencies of structural genes in females would have less net expression than in males. This is expected because the reduction of IE regions in males would allow each individual gene greater expression. This is the case (Muller, 1950, The Harvey Lectures).

The IE regions hypothesis does not readily account for the fact that triploid flies have approximately 150% enzyme expression per cell as diploids. That is, in a triploid, the IE regions would be increased 3/2 over the diploid and the structural genes also would be increased. This would give an hypothetical gene expression [3/2 structural genes X 2/3 (the reciprocal of the IE regions ratio)] equal to the diploid. This is not the case, as reported in the literature (Lucchesi and Rawls, 1973, Bioch. Genetics 9:41). The basis for this seeming paradox is unknown. One way to reconcile it would be to consider that when all the IE regions of a cell are increased they act in a positive manner. But when only one of several IE regions for a particular gene is increased, there is an inhibition of the total group. Other suggestions that IE regions may act positively under some conditions are: 1) The data of Rawls and Lucchesi show that IDH is exceptional in that it has multiple positive effecting regions instead of negative effecting regions; 2) If one examines the data of Lucchesi, Rawls and Maroni (1974, Nature 248:564), Abraham and Lucchesi (Genetics 78:1119), and Ananiev and Gvozdev (1974, Chromosoma 45:193-201), s/he will find that the dosage of the X chromosome produces a slight positive "dosage effect" on autosomal genes in Drosophila larvae. The dosage of the X shows an inverse effect on some autosomal genes when adult flies are assayed.

Another way to reconcile the aneuploidy/ploidy paradox is to consider that the processes that control cell volume are independent of IE regions. Thus triploid cells would have correspondingly larger cells. Since the concentration of the IE region products (3/2 as much in 3/2 the volume) would be equal to each gene present in the diploid, so the increase in structural genes (3/2) is directly expressed. These are only two ways to reconcile this apparent paradox.

Another explanation of dosage compensation deserves discussion. It postulates that the activity of X-linked genes is limited by gene products produced by autosomal genes (Maroni and Plaut, 1973, Chromosoma 40:361). Since the autosomal level in males and females is equal, the females' two X's would be limited to the same level as the males' one X. This hypothesis suffers from the fact that screens for dosage dependent regions affecting X linked genes have failed to show any such regions (Rawls and Lucchesi, 1974, Genetical Research 24:59). In the data, however, are values consistent with the presence of IE regions. The limiting factor hypothesis does not account for IE regions and their roles.

Nor does this hypothesis account for sexual dimorphism of autosomal genes. In the vast majority of sexual dimorphisms of autosomal mutants, the males are more like wild type than are females even though the gene dosage is the same in the two sexes (Goldschmidt, 1955, Science; Smith and Lucchesi, 1969, Genetics 61:607). If positive factors on the X controlled these autosomal genes, males would be either less like or equal to wild type as females, depending upon whether the positive factor itself is not dosage compensated or dosage compensated, respectively.

If dosage compensation is to be explained as the natural consequence of a cancellation of structural gene dose and inverse effect, certain predictions would follow:

1) There should be IE regions on the X that affect autosomal genes. Since the inverse regions are reduced in males, the affected autosomal genes would be expressed at greater levels in males than in females. Phenotypic and quantitative studies have shown autosomal genes to be expressed more strongly in males (Smith and Lucchesi). Biochemical studies show some indication that this is so (Lucchesi, Rawls, Maroni, 1974, Nature 248:564). However, the degree of the phenomenon could have been obscured in these studies by correcting enzyme levels per mg protein, since the protein per cell in males and females may change in the same direction as the autosomal genes tested.

2) Regardless of the composition of the X in various species of Drosophila, there should be dosage compensation. Abraham and Lucchesi (Genetics 78:1119) present evidence that this is true, although their data may be obscured by correction against total protein in males and females.

The biochemical studies may not reveal very accurately the true nature of the phenomenon. Since the hypothesis predicts that many autosomal genes will have increased expression in males, the total protein/cell ratio may be different in males and females. Thus "correcting" enzyme levels by protein per extract may cancel out to some degree the level of increase. Also these studies are complicated by the fact that males and females possess different anatomies. Undoubtedly, many enzymes are expressed differently in these different tissues and organs. For these reasons, biochemical comparisons of male and female flies should be interpreted with great circumspection.

It may be that IE regions on the X would be indirectly responsible for sex determination in Drosophila. Since it appears that IE regions themselves show a dosage effect, they would increase the level of expression of many autosomal genes in adult males as compared to females. In triploids the increase of 2X3A over 3X3A would not be as great as 1X2A over 2x2A and hence an intersex would result. Genes such as transformer, intersex, and masculinizer that convert females to malelike with little or no effect on males may be major inverse "genes" (or other genes that affect their action) that have mutated to a hypomorphic allele. One would predict from such an hypothesis that they would raise many enzyme levels in genetic females but not in males. Komma (1966, Genetics 54:497) claimed transformer raised the level of G6PDH in females but not in males. 6PGDH was not affected in either. Smith and Lucchesi showed that transformer increased the expression of autosomal glass mutants in females more than in males. Since genetic males already have presumably many inverse effect regions varied in relation to normal females, varying another, even though major, might not produce as radical a difference. In females, the transformer gene would be the first inverse "gene" varied and as a result a larger increase would be seen. The mutant doublesex, which converts both males and females to intersexes, is not readily understood by this hypothesis, although it also produces a greater gene expression in genetic females than in males.

Evidence that dosage compensation, sexual dimorphism of autosomal mutants and sex determination are affected by the same mechanism comes from studies of the effect of temperature on these processes. Lee (1968, Genetical Research 11:115) presented evidence that at high temperature, dosage compensation of an X linked gene was lost, whereas it occurred at low temperature. Smith and Lucchesi showed that sexual dimorphism of autosomal glass mutants was less evident at high temperature than at low. Finally, Dobzhansky (1930, Am. Naturalist 64:261) showed that low temperatures caused intersexes to become more male-like while high temperatures had a feminizing effect. If low temperatures enhance the inverse effect and high temperatures decrease it (there are no data available), and if the inverse effect is responsible directly or indirectly for all three phenomena, these results would be found.

In conclusion, it should be stated that much research is needed to clarify the nature of the inverse effect. Its apparent presence in such diverse higher organisms as maize and Drosophila suggests that it has a central importance in gene expression and possibly differentiation. This report hopefully has pointed out this possibility.

James A. Birchler

Return to the MNL 51 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage