Interactions Among Dosage-Dependent Trans-Acting Modifiers of Gene Expression and Position-Effect Variegation in Drosophila

Corresponding author: James A. Birchler, Division of Biological Sciences, University of Missouri-Columbia, 117 Tucker Hall, Columbia, MO 65211., birchler{at}biosci.mbp.missouri.edu (E-mail).

Communicating editor: S. HENIKOFF

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We have investigated the effect of dosage-dependent trans-acting regulators of the white eye color gene in combinations to understand their interaction properties. The consequences of the interactions will aid in an understanding of aneuploid syndromes, position-effect variegation (PEV), quantitative traits, and dosage compensation, all of which are affected by dosage-dependent modifiers. Various combinations modulate two functionally related transcripts, white and scarlet, differently. The overall trend is that multiple modifiers are noncumulative or epistatic to each other. In some combinations, developmental transitions from larvae to pupae to adults act as a switch for whether the effect is positive or negative. With position-effect variegation, similar responses were found as with gene expression. The highly multigenic nature of dosage-sensitive modulation of both gene expression and PEV suggests that dosage effects can be progressively transduced through a series of steps in a hierarchical manner.

VARIATION in dosage of chromosomes or chromosomal segments results in alterations of gene expression throughout the nucleus (BIRCHLER 1979 , BIRCHLER 1981 ; BIRCHLER and NEWTON 1981 ; DEVLIN et al. 1982 , DEVLIN et al. 1988 ; SABL and BIRCHLER 1993 ; GUO and BIRCHLER 1994 ). Any one gene can be modulated by several regions of the genome. The range of these changes generally falls within the limits of an inverse or direct correlation between the chromosomal dose and the level of target gene expression. That is, a reduction from two to one copy often causes a twofold increase in expression. An increase in dosage of the same region from two to three copies causes reductions to two-thirds of the normal level. With direct correlations, chromosomal reductions result in 50% of normal expression and increases cause elevations to 150%. Although both types of dosage effects are found, the inverse correlations are more prevalent (BIRCHLER et al. 1994 ; BEGLEY et al. 1995 ; BHADRA and BIRCHLER 1996 ; BHADRA et al. 1997A , BHADRA et al. 1997B ; FROLOV et al. 1998 ).

The dosage effects of some regions have been traced to single loci. Using the white eye color locus as a monitor, numerous second-site modifiers have been recovered that exhibit a dosage effect on white expression and that fall within the range seen with chromosomal aneuploids. A fraction of these have been described in the literature (RABINOW et al. 1991 ; BIRCHLER et al. 1994 ; BHADRA and BIRCHLER 1996 ; BHADRA et al. 1997A , BHADRA et al. 1997B ; FROLOV et al. 1998 ).

A similar situation occurs for position-effect variegation (PEV). PEV is the mosaic expression of genes brought near abnormal junctions of euchromatin and heterochromatin. There are scores of modifiers of PEV that either suppress or enhance the degree of mosaic expression (for reviews, see SPOFFORD 1976 ; WEILER and WAKIMOTO 1995 ; ELGIN 1996 ; HENIKOFF 1996 ). Several of the genes initially identified as dosage modifiers of white are also suppressors of PEV (BIRCHLER et al. 1994 ; BHADRA and BIRCHLER 1996 ; BHADRA et al. 1997B ; FROLOV et al. 1998 ). The modification of gene expression and PEV appears to be related, and indeed several genes affecting PEV have been identified as transcriptional regulators of some type (DORN et al. 1993 ; FARKAS et al. 1994 ; TSCHIERSCH et al. 1994 ; SEUM et al. 1996 ; FROLOV et al. 1998 ).

Despite the fact that several genes or segmental aneuploids produce a dosage effect on a single monitored target gene, larger aneuploids as a general rule do not exceed the limits of inverse and direct correlations of gene expression with the varied dosage (DEVLIN et al. 1982 , DEVLIN et al. 1988 ; BIRCHLER 1992 ; GUO and BIRCHLER 1994 ). Hence, we wished to determine whether the combinations were cumulative, noncumulative, and/or epistatic with regard to both target gene expression and variegation. The results should aid in understanding interacting regulatory pathways affecting the two phenomena.

Two general trends emerge from our results. Increasing the number of mutant modifiers does not usually cause a multiplicative effect on gene expression. Although there are some combinations that are cumulative, the overall tendency is for noncumulative action. Secondly, reducing the dosage of the functional alleles of an increasing number of modifiers produces a greater tendency for an inverse effect to occur.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Flies were maintained on cornmeal-glucose-yeast-agar media at 25°. The genetic mutations and chromosomal rearrangements are described in FLYBASE (http://morgan.harvard.edu/fb.html/).

Genetic recombination:
The cytological locations of the five trans-acting modifiers [Inverse regulator-a^EMS2 (Inr-a, cytological location 46D–47D), Lightener of white (Low, 39E7–F1), Ultrafemale overexpression (Ufo, 47A11), Weakener of white (Wow, 76D5–F), and Modifier of white^EMS1 (Mow, 85F1–86C1)] used in this study were determined earlier (RABINOW et al. 1991 ; BIRCHLER et al. 1994 ; BHADRA and BIRCHLER 1996 ; BHADRA et al. 1997A , BHADRA et al. 1997B ). To analyze the combinations, it was necessary to generate flies carrying two mutations on the same chromosome. Because the modifiers have no morphological phenotypes in adult mutant flies, we initially recombined Inr-a with the dominant marker Tft and the Low mutation with vg^U. The Tft Inr-a/SM6a females were crossed to Low vg^U/In(2LR)Gla males. The Tft Inr-a/Low vg^U daughters were mated with In(2LR)Gla/SM6a males. At least 100 putative recombinant non-Tft non-vg^U offspring were selected for pair mating with the In(2LR)Gla/SM6a balancer stock. Most of these putative recombinants were expected to be Low Inr-a/Balancer. To confirm the genotype, females from each selected stock were individually crossed with Inr-a Tft/SM6a and Low vg^U/In(2LR)Gla males to test for complementation of the recessive lethality of the mutant modifiers. Each stock that complemented neither Inr-a nor Low was selected. Using a similar scheme, we combined the third chromosomal modifiers Wow with Sb and Mow with Ly and recombined them. The double modifier stocks were used to generate the quadruple stock.

To generate triple modifier combinations, Low Inr-a/SM6a females were crossed to multiple autosomal balancer males [SM6a/In(2LR)Gla; MKRS/TM3, Ser]. The progeny females Low Inr-a/In(2LR)Gla; +/TM3, Ser were then crossed to +/SM6a, Wow/MKRS males that were produced by crossing Wow/TM3, Ser females with multiple autosomal balancer males. The flies carrying SM6a and TM3 Ser balancer chromosomes were selected and mated with each other to generate a stable Low Inr-a/SM6a; Wow/TM3, Ser stock. Similar types of crosses were used to generate other triple combination flies.

Pigment determination:
To measure eye pigments in flies with the variegating allele, w^m4h, individuals were collected of the appropriate genotype and frozen in liquid nitrogen. The heads were manually separated from the bodies. Forty heads from each genotype were used for each assay as described earlier (BHADRA et al. 1997B ).

RNA extraction and Northern blots for developmental analysis:
The genetic crosses for generation of segregating populations during development were described earlier (BHADRA et al. 1997A , BHADRA et al. 1997B ). The Tubby mutation is a genetic marker used to distinguish segregating populations in larvae. RNA extraction from frozen (-80°) larvae, early and mid-pupae, and adult flies for Northern analysis was performed as described elsewhere (HIEBERT and BIRCHLER 1994 ). Antisense RNA probes (white, scarlet, and rRNA as a loading control) were prepared using an in vitro transcription kit (Ambion), according to the manufacturer's instructions. Hybridization with different labeled RNA probes was carried out as previously described (HIEBERT and BIRCHLER 1994 ). Quantitation of the radioactivity present in each band on Northern filters was performed using a Fuji Bas 2000 phosphorimager.

To eliminate the possibility that the modifiers affected the gel loading control, we have determined the abundance of specific transcripts per unit DNA template in the four segregating classes for each combination. We isolated total nucleic acid (DNA and RNA) using a technique described earlier (RAHA et al. 1990 ; HIEBERT and BIRCHLER 1994 ). Triplicate isolations of the four genotypes in each combination were electrophoresed on 1% agarose. Separate DNase I and RNase A digestions confirmed that the upper and lower bands on ethidium-stained gels corresponded to genomic DNA and rRNA, respectively. In the same gel, a dilution series of identically prepared nucleic acid was electrophoresed. The photographic negatives of an ethidium-stained gel containing triplicate preparations from different modifier combinations were analyzed by laser scanning densitometry using the same parameters as described earlier (HIEBERT and BIRCHLER 1994 ; BHADRA et al. 1997A , BHADRA et al. 1997B ). The measurements of the dilution series were used to establish the standard curve (PAL BHADRA et al. 1998 ). The DNA/18S rRNA and the DNA/28S rRNA ratios were calculated from the densitometric scans. The relative ratios obtained from the males and females of different modifier combinations were not significantly different from the normal males and females (0–24 hr) (PAL BHADRA et al. 1998 ). Therefore, the lack of a significant alteration of rRNA level in any of the tested combinations shows rRNA to be a valid gel loading control.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Our laboratory is interested in second-site dosage-dependent modifiers of gene expression and their involvement with various phenomena. In particular, we are characterizing a collection of trans-acting regulators that have a positive or negative effect on the white eye color gene (RABINOW et al. 1991 ; BIRCHLER et al. 1994 ; BHADRA and BIRCHLER 1996 ; BHADRA et al. 1997A , BHADRA et al. 1997B ; FROLOV et al. 1998 ). The white gene has been genetically and molecularly studied, and many aspects of its cis-acting regulatory features are known (ZACHAR and BINGHAM 1982 ; O'HARE et al. 1983 ). Numerous leaky and hypomorphic alleles at white facilitate characterization of its trans-acting modifiers at the phenotypic level. The five genes that have been selected for this study were described earlier (RABINOW et al. 1991 ; BIRCHLER et al. 1994 ; BHADRA and BIRCHLER 1996 ; BHADRA et al. 1997A , BHADRA et al. 1997B ). Each modifier has a distinct effect on white based on its interaction with different w alleles. Two white alleles, w^carrot (w^crr) and w^spotted (w^sp), were used for an initial test of a combinatorial effect. w^crr is a "point" mutation as determined by the absence of gross rearrangements detected by Southern analysis (ZACHAR and BINGHAM 1982 ). w^sp has a 5' regulatory lesion and shows a reduced level of pigment, lighter in females than males (ZACHAR and BINGHAM 1982 ; O'HARE et al. 1983 ; DAVISON et al. 1985 ). Each modifier has a positive or negative effect on the w^crr phenotype. In contrast, three of the five modifiers alter the w^sp eye color, with Inr-a and Low having no effect on it. Therefore, cis-acting regulatory sequences that are defective in w^sp are required for the Inr-a and Low interaction.

Eight double, four triple, and a quadruple combination of modifiers were tested against these two white alleles (Table 1). All the combinations either darken or lighten the eye colors of w^crr and w^sp flies, with the exception of the Inr-a; Wow combination on w^crr and the Low; Mow, Low Inr-a; Wow, and Low Inr-a; Mow combinations on w^sp (Figure 1 and Table 1). The two alleles respond differently to the combinations, and in neither case is the effect of modifier combinations predictable from their separate effects.

View larger version (89K): In this window In a new window Download PPT slide   Figure 1. Phenotypic effect of selected combinations of trans-acting modifiers on two different white alleles and wm4h. The genotypes and white alleles are noted.

zeste-dependent transvection and modifier combinations:
Transvection results in an altered gene expression due to allelic pairing. zeste is a pairing-dependent regulator of the white locus (BINGHAM and ZACHAR 1985 ; BICKEL and PIRROTTA 1990 ; CHEN and PIRROTTA 1993 ). Two doses of white, including its 5' regulatory sequences, are required for a phenotypic effect of zeste¹. The zeste protein binds at the 5' regulatory sequence to suppress white in the eye, resulting in a yellow eye color instead of wild-type red (DAVISON et al. 1985 ; CHEN and PIRROTTA 1993 ). None of the selected modifiers individually affects zeste. To examine whether zeste interacts with various combinations of modifiers, a series of genetic crosses were performed using males from each double, triple, and quadruple combination carrying the w¹¹¹⁸ mutation in the X chromosome and females from a z Dp(1;1)w^61e19/z Dp(1;1)w^61e19 strain (GREEN 1963 ). The eye colors of nonbalancer and balancer males were compared. Data summarized in Table 1 reveal that in most cases the eye color of the balancer and nonbalancer males is similar, with the exception of three combinations: Low; Mow, Wow Mow, and Low; Wow Mow. In Low; Mow males, several patches of brown pigment are distributed throughout the eye (Figure 2 and Table 1). When the Wow mutation is added to this combination, a darker eye color is found along with brown patches (Figure 2). The eye color is the same in the triple combinations and in Wow Mow. Thus, these modifiers also interact nonadditively on the zeste phenotype.

View larger version (52K): In this window In a new window Download PPT slide   Figure 2. Effect of selected modifier combinations on zeste. All flies are z1 Dp(1;1)w61e19/Y. The modifier mutations present are noted.

Position-effect variegation and modifier combinations:
Effect on the variegating white phenotype: In general, dosage-dependent PEV modifiers are separated into two classes, suppressors and enhancers. The effect of multiple modifiers on PEV in most cases is mainly additive (LOCKE et al. 1988 ; ELGIN 1996 ; HENIKOFF 1996 ). That is, mutations in two suppressors more strongly suppress PEV than either one alone. We tested the effect of five modifiers in combinations on position-effect variegation, of which three alone are weak suppressors of PEV (BIRCHLER et al. 1994 ; BHADRA and BIRCHLER 1996 ; BHADRA et al. 1997B ) and two others have no effect on the variegating phenotype (RABINOW et al. 1991 ; BHADRA et al. 1997A ).

When a chromosomal inversion, such as In(1)w^m4h, relocates the white eye color gene next to heterochromatin, the expression of the white gene is silenced in some ommatidia but not in others. To examine the effects of different combinations of modifiers on this phenotype, w^m4h/w^m4h females were crossed with males from multiple modifier stocks. The eye phenotype of each genotype of F₁ males was examined (Figure 1). Table 2 summarizes the results. The degree of suppression fluctuates in various combinations, but in general is noncumulative and epistatic. However, it is interesting to note that inclusion of Inr-a either with Low or Wow intensifies the suppression, although Inr-a alone does not suppress PEV. In contrast, Ufo, which also lacks any effect on variegation, reduces suppression by the Wow Mow combination. These effects of Ufo and Inr-a on position-effect variegation reveal that second-site modifiers that have no effect on their own, participate when they are combined with other genes that act as suppressors. In the absence of Inr-a and Ufo, the three modifiers Low, Wow, and Mow appear to act cumulatively (Figure 1).

The progenitor chromosomes of Low, Ufo, Wow, and Mow were found to have no effect on variegation (BIRCHLER et al. 1994 ; BHADRA and BIRCHLER 1996 ; BHADRA et al. 1997A , BHADRA et al. 1997B ). To determine the existence of any potential preexisting modifier on the Inr-a chromosome, we analyzed the second chromosome from the parental w^a line for its effect on w^m4h. The w^m4h females were mated to males carrying the progenitor second chromosome, together with SM6a, the same balancer chromosome present in the Inr-a stock. The amount of eye pigment in w^m4h; SM6a/+ and w^m4h; +/+ male and females was compared (data not shown) and was not significantly different.

Effect on a variegating yellow rearrangement: In addition to the effect of modifiers on white variegation, we also examined the effect on yellow variegation among the ~80 bristles along the anterior margin of the wing blade in In(1)y^3P flies. These results tested whether the effects of modifier combinations on position-effect variegation are general. An advantage of this study is that variegation events in individual cells can be easily detected and quantitated (KARPEN and SPRADLING 1990 ; BHADRA and BIRCHLER 1996 ; BHADRA et al. 1997B ). After crossing +/Y; M/Balancer (M represents the modifier combination) males to In(1)y^3P females, the y and y⁺ bristles were counted in the experimental In(1)y^3P/Y; M/+ males and their control In(1)y^3P/Y; Balancer/+ brothers. For each genotype, the number of normal black (yellow⁺) bristles on the anterior margin of one wing from each of 10 males was scored (Table 3).

The results reveal that various modifier combinations alter yellow variegation. In most combinations, the effect on the yellow variegation closely resembles the effect obtained with w^m4h. However, two combinations, Inr-a; Wow and Low Inr-a; Mow, have less effect on y^3P than on w^m4h. Overall, however, the results support the conclusion that most of the combinations act generally on PEV rather than specifically on w^m4h.

Developmental alteration of white and scarlet mRNA:
An initial characterization of the five modifiers revealed that they altered the transcript levels of white and two other functionally related genes, scarlet and brown, either positively or negatively. These three eye color genes are involved in transport of pigment precursors from the hemolymph into the appropriate cell types (DREESEN et al. 1988 ; TEARLE et al. 1989 ). Each modifier is distinct from the others. Wow modulates the target gene expression either positively or negatively, depending on the developmental stage or tissue (BIRCHLER et al. 1994 ). Two modifiers, Ufo and Low, affect a wide variety of target loci. Ufo in particular shows an extensive overexpression of target genes and eliminates the sexual equivalence between males and females in some developmental stages (BHADRA et al. 1997A , BHADRA et al. 1997B ). Mow and Low have a sex-influenced effect. The effect of Inr-a is limited to the transcripts of white, brown, and scarlet primarily in larvae and pupae (RABINOW et al. 1991 ).

In double combinations: To discover whether the modifiers interact similarly on different target genes, we measured the steady-state levels of white and scarlet transcripts in the two segregating classes for males and females for each modifier combination. Total RNA was isolated from each genotype in four developmental stages: third instar larvae, early pupae (0–24 hr), middle pupae (24–48 hr), and eclosed adults (0–24 hr). The Northern profiles of each combination and their quantitative results are presented in Figure 3 and Table 4 and Table 5.

View larger version (37K): In this window In a new window Download PPT slide   Figure 3. Abundance of white transcripts in 12 selected combinations of modifiers at the larval stage. Northern blots of each genotype were hybridized with a 32P-labeled antisense RNA of white and then reprobed with antisense rRNA as a gel loading control (BHADRA et al. 1997A , BHADRA et al. 1997B ). M, heterozygous combination of the modifiers. The respective probes are noted.

In larvae, three of the five modifier mutations, Inr-a, Ufo, and Wow separately increased white transcripts to the twofold level or more, while in all double combinations the effects are more limited (Figure 3 and Table 4), especially for Ufo. Mow lowers white transcript levels even when combined with any of the three modifiers that individually increase it, but raises the white transcript level when combined with Low.

The combined effect of multiple modifiers on scarlet transcripts is distinct from their effect on white (Table 5). The influence of Mow in the double combinations is not as strong as found on white, with the exception of the Inr-a Mow combination. In some instances, such as Low; Wow, the combined effects were intermediate between the individual effects (Table 5).

We also measured the white and scarlet transcript abundance in two different pupal stages. The larval-pupal transition significantly influenced the effect of the modifiers (Table 4 and Table 5). In some cases, combinations that exhibited a direct correlative effect in larvae switch to an inverse effect, while inversely acting combinations in larvae act as direct correlative modifiers in pupae. Overall, the modulation in pupae is greater than in larvae. Similar to other stages of development, the extreme overexpression of scarlet transcripts in pupae caused by Ufo is eliminated when Ufo is associated with the Wow mutation. The mode of action of several double combinations is similar in the two pupal stages. In the Inr-a; Wow mid-pupal combination, white transcripts are reduced, while their respective individual effects each exhibit a significant elevation. In general, the double combinations exhibit noncumulative or epistatic relationships among the modifiers at the pupal stage (Table 4 and Table 5).

The effect of double combinations on white transcript levels is minimal at the adult stage and the effects on scarlet are mostly noncumulative (Table 4 and Table 5). Three of the five modifiers, Low, Wow, and Mow separately reduced white transcript levels, while such reduction is eliminated in their combinations. In several cases, variation of the white and scarlet transcripts at the adult stage is significantly different from that of the pupal stage (Table 4 and Table 5). This result suggests that the developmental transition from pupa to adult also plays a role in modulating the combined effect of multiple modifiers on target genes.

In triple combinations: The quantitative effect on white transcripts was measured in each of four triple combinations (Table 4 and Table 5). In two of these, Low Inr-a; Wow and Low; Wow Mow, white transcripts are overexpressed relative to their closest double combinations in larvae (Table 4). In the Low Inr-a; Wow combination, this effect also persists at the pupal stage, whereas, in the Low Inr-a Mow and Inr-a Wow Mow combinations, the level of white mRNA is considerably less elevated or reduced (Table 4). The pattern of scarlet transcripts in triple combinations closely resembles the effect on white (Table 5). However, there are instances where white transcript levels are elevated but scarlet is significantly reduced, and vice versa.

In the quadruple combination: The relative abundance of white and scarlet transcripts in the quadruple combination is increased to a lesser degree than in certain triple combinations (Figure 3 and Figure 4 and Table 4 and Table 5). The modulation for the quadruple modifiers is limited within a threefold range throughout development.

View larger version (42K): In this window In a new window Download PPT slide   Figure 4. Northern blots of total cellular RNA probed with white and scarlet antisense RNA of segregating populations of a quadruple modifier combination in four different developmental stages (see MATERIALS AND METHODS). The bottom panel of each blot shows reprobing with rRNA as a gel loading control. M, heterozygous combination of the modifiers. The name of the antisense RNA probes and the developmental stages are noted.

Sexual equivalence and modifier combinations:
We also analyzed the difference of the combined effect of multiple modifiers between males and females. In a few combinations, the transcript levels of each gene (white and scarlet) exhibit sex-specific effects in larvae and adults (Figure 5). Six out of 13 combinations show a sex-specific effect on white in early pupae, while only the quadruple combination exhibits a differential level of scarlet expression between males and females.

View larger version (45K): In this window In a new window Download PPT slide   Figure 5. Summary of the effects of multiple modifiers on white and scarlet transcripts in four developmental stages. The mode of interaction between different modifiers and their target loci is derived from Table 4 and Table 5. +, increase; -, reduction; and , no effect on transcript levels. Boxes represent sexual dimorphism between male and female values.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this study, we have generated double, triple, and quadruple combinations of dosage-dependent modifiers of the white eye color locus to gain information on their interaction properties. Taken altogether, there are examples of cumulative effects, cancellation by opposite effects, and epistatic interactions of one gene with another. In some cases, the combination of two directly correlative effects becomes an inversely correlative one, and vice versa. However, despite the individual examples of specific types of interactions, considering all four developmental stages in both sexes for the two target genes studied, the overall trend is for noncumulative action. For example, although Low, Inr-a, Wow, and Mow all individually increase white expression in the mid-pupal stage between ~two- and threefold, the quadruple combination still only increases white expression two- to threefold rather than the >16x level expected from independent action.

The second general trend, considering only Inr-a, Wow, Mow, and Low, for which a complete data set exists, is that with an increased number of modifiers in the combination, the likelihood increases that the net result is an inverse correlation between the dosage of the functional modifier alleles and target gene expression. With single modifiers, 26 of 64 data points (40% of the total, Ufo excluded) show an inverse effect. In the double combinations, 40% (38 out of 96) of the 96 data points (Ufo; Wow and Ufo; Mow not included) indicate an inverse response. However, with the triple and quadruple combinations, 55% (44 out of 80) of the 80 data points are now inverse effects (Figure 5). A row by column chi-square analysis of the null hypothesis that the single plus double vs. triple plus quadruple arrays are the same is rejected (² = 4.26; d.f. = 1; P < 0.05). These two general trends are consistent with the results from larger aneuploids in which the predominant dosage effect is an inverse one (DEVLIN et al. 1988 ; BIRCHLER et al. 1989 ), and that the magnitude of the effects more or less remain within the limits of an inverse correlation. Despite this potential connection from these data, it is probable that in many larger aneuploids more than four modifiers for one target gene are varied.

Interactions:
Effect of zeste: An interesting interaction involves the effect on zeste. The zeste gene affects pairing-dependent expression of white. When zeste¹ is present with two copies of white, a lemon eye color results rather than the normal brick red. None of the white modifiers alone alters the eye color when zeste¹ is modifying white. However, two double combinations—Low; Mow and Wow Mow—suppress the zeste-white interaction. The triple combination of Low; Wow Mow also suppresses zeste, but interestingly the Inr-a; Wow Mow returns to normal.

Effect on PEV: Of the five loci alone, Low, Wow, and Mow act as suppressors of PEV. We tested the combinations on two variegating rearrangements, w^m4h and y^3P, to examine the generality of the effects, particularly since they were selected as modifiers of white. The results on the two rearrangements are quite similar.

The generalities formulated for the effects on gene expression also apply to PEV, primarily, in that the combinations do not exhibit strictly cumulative interactions. It is true in the case of PEV that the triple combinations are the strongest group of suppressors, but among these the two greatest both involve Low and Mow, which is the strongest dual combination. The quadruple combination drops back into the range of the single modifiers as did the effects on gene expression. Although Inr-a alone has no effect on variegation, in combination with Wow or Low a greater suppression occurs than with either alone. On the other hand Ufo, which has no effect by itself, will dilute the suppression caused by Wow or Mow. In contrast to the effects on gene expression, no combinations were found which switched the effect on PEV from suppression to enhancement.

Why so many dosage-dependent modifiers of gene expression and PEV?
There are scores of dosage-dependent modifiers of PEV and gene expression. For PEV, HENIKOFF 1996 provided insight on this issue by noting that some modifiers are structural components of heterochromatin and others are regulatory genes that might control them. Indeed, many suppressors and enhancers of PEV have been cloned, and while they are a heterogeneous collection, all are reasonable candidates for having an involvement in the process as chromatin components or their regulators (EISSENBERG et al. 1992 ; GARZINO et al. 1992 ; REUTER and SPIERER 1992 ; TSCHIERSCH et al. 1994 ; DE RUBERTIS et al. 1996 ; TYLER et al. 1996 ). The same could be noted for dosage modifiers of gene expression (e.g., DORN et al. 1993 ; FARKAS et al. 1994 ; SEUM et al. 1996 ; FROLOV et al. 1998 ). Indeed, there appears to be a significant overlap between the two.

The above-mentioned examples, as well as others (BOTAS et al. 1982 ; WEINTRAUB 1993 ; KENNISON 1995 ; CUBADDA et al. 1997 ; CREWS 1998 ), suggest that a disproportionately high fraction of regulatory genes are dosage sensitive. Over evolutionary time, genes involved with many aspects of gene expression from signal transduction to transcriptional regulation apparently have evolved to be rate limiting in the diploid state and as such would exhibit a dosage effect when assayed genetically. A hierarchy of dosage-sensitive regulation would contribute to the multigenic nature of these systems. If the regulatory genes most intimately involved with the expression of a structural gene are dosage dependent and in turn are controlled by dosage-dependent regulators and so on, the higher-order regulatory genes will still be dosage dependent on the expression of the monitored structural gene. Because a dosage effect can be transduced through a series of dosage-dependent steps, this situation could at least partially explain why so many dosage modifiers of gene expression and PEV exist. While individual genes in a hierarchical series might all affect the ultimate target gene, one would also expect examples of epistasis and an overall noncumulative effect in combinations, as found in this study.

Relationship to the genetic control of quantitative traits:
As noted above, most mutations characterized in genetically studied organisms are recessive (see, for example, FLYBASE http://morgan.harvard.edu/fb.html/; ORR 1991 ; THATCHER et al. 1998 ). Yet the control of quantitative traits is governed disproportionately by genes that exhibit additivity in hybrids between the parental extremes (TANKSLEY 1993 ; LIU et al. 1996 ). Typically, quantitative traits are governed by multiple factors, which nevertheless often exhibit epistasis (TANKSLEY 1993 ; DAMERVAL et al. 1994 ; CLARK et al. 1995 ; MACKAY 1996 ; CLARK and WANG 1997 ; LAURIE et al. 1997 ). The parallels to the multiple dosage-sensitive modifiers of gene expression, as exemplified here by the effect on the w^crr phenotype, are striking.

In a simplistic view, a phenotypic characteristic is controlled by a biochemical pathway. One enzymatic step in that pathway will be rate limiting by definition under a certain set of conditions and hence will exhibit a dosage effect on the phenotype when genetic variation is present at the locus encoding that enzyme. If this gene is governed by a dosage-dependent hierarchy, the phenotype will be governed by multiple dosage-sensitive factors (e.g., BYRNE et al. 1996 ) that overall are not cumulative. To the extent that epistasis and interactions occur among the regulatory loci, such as was found in this study, these effects would be reflected in the phenotype. Of course, multiple biochemical pathways contributing to the phenotype would complicate the genetics further.

Relationship to dosage compensation:
With dosage compensation, increasing the number of copies of a particular chromosomal segment leads to less product per copy of a gene included in that segment. Because the structural gene being monitored is increased in copy number at the same time, an overall equalization of total output per cell results. In addition, genes elsewhere in the genome may also be downregulated, leading to a reduction of their total product (inverse correlation) (BIRCHLER 1979 , BIRCHLER 1981 , BIRCHLER 1996 ; DEVLIN et al. 1982 ; BIRCHLER et al. 1989 , BIRCHLER et al. 1990 ; GUO and BIRCHLER 1994 ). These two responses appear to be related in that the dosage compensation results from the cancellation of a structural gene dosage effect by the inverse effect produced simultaneously (BIRCHLER 1981 ; BIRCHLER et al. 1990 ). If the same set of dosage-dependent modifiers are responsible for both dosage compensation, as well as the inverse effects on loci elsewhere in the genome, then they must not be generally cumulative and must favor an inverse dosage effect. While our study has analyzed in detail only four of the several modifiers of white, the results are consistent with these requirements on the whole.

	FOOTNOTES

We dedicate this work in memory of Dr. MALAYA KR. PAL.

	ACKNOWLEDGMENTS

We thank the Bloomington Stock Center, Indiana University, for providing fly stocks. We thank the members of Birchler lab for helpful discussions and comments on the manuscript. This work is supported by a National Science Foundation grant to J.A.B.

Manuscript received December 17, 1997; Accepted for publication June 8, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BEGLEY, D., A. M. MURPHY, C. HIU, and S. I. TSUBOTA, 1995  Modifier of rudimentary p1; mod(r)p1, a trans-acting regulatory mutation of rudimentary. Mol. Gen. Genet. 248:69-78[Medline].

BHADRA, U. and J. A. BIRCHLER, 1996  Characterization of a sex influenced modifier of gene expression and suppressor of position-effect variegation in Drosophila. Mol. Gen. Genet. 250:601-613[Medline].

BHADRA, U., M. PAL BHADRA, and J. A. BIRCHLER, 1997a  A trans-acting modifier causing extensive overexpression of genes in Drosophila melanogaster. Mol. Gen. Genet. 254:621-634[Medline].

BHADRA, U., M. PAL BHADRA, and J. A. BIRCHLER, 1997b  A sex-influenced modifier in Drosophila that affects a broad spectrum of target loci including the histone repeats. Genetics 146:903-917[Abstract].

BICKEL, S. and V. PIRROTTA, 1990  Self-association of the Drosophila zeste protein is responsible for transvection effects. EMBO J. 9:2959-2967[Medline].

BINGHAM, P. M. and Z. ZACHAR, 1985  Evidence that two mutations, w^DZL and z¹, affecting synapsis-dependent genetic behavior of white are transcriptional regulatory mutations. Cell 40:819-825[Medline].

BIRCHLER, J. A., 1979  A study of enzyme activities in a dosage series of the long arm of chromosome 1 in maize. Genetics 92:1211-1229[Abstract/Free Full Text].

BIRCHLER, J. A., 1981  The genetic basis of dosage compensation of Alcohol dehydrogenase-1 in maize. Genetics 97:625-637[Abstract/Free Full Text].

BIRCHLER, J. A., 1992  Expression of cis-regulatory mutations of the white locus in metafemales of Drosophila melanogaster. Genet. Res. 59:11-18[Medline].

BIRCHLER, J. A., 1996  X chromosomal dosage compensation in Drosophila. Science 272:1190[Medline].

BIRCHLER, J. A. and K. J. NEWTON, 1981  Modulation of protein levels in chromosomal dosage series of maize: the biochemical basis of aneuploid syndromes. Genetics 99:247-266[Abstract/Free Full Text].

BIRCHLER, J. A., J. C. HIEBERT, and M. KRIETZMAN, 1989  Gene expression in adult metafemales of Drosophila melanogaster. Genetics 122:869-879[Abstract/Free Full Text].

BIRCHLER, J. A., J. C. HIEBERT, and K. PAIGEN, 1990  Analysis of autosomal dosage compensation involving the Alcohol dehydrogenase locus in Drosophila melanogaster. Genetics 124:677-686[Abstract].

BIRCHLER, J. A., U. BHADRA, L. RABINOW, R. LINSK, and A. T. NGUYEN-HUYNH, 1994  Weakener of white (Wow), a gene that modifies the expression of the white eye color locus and that suppresses position-effect variegation in Drosophila melanogaster. Genetics 137:1057-1070[Abstract].

BOTAS, J., J. MOSCOSO DEL PRADO, and A. GARCIA-BELLIDO, 1982  Gene-dose titration analysis in the search of trans-regulatory genes in Drosophila. EMBO J. 1:307-310[Medline].

BYRNE, P. F., M. D. MCMULLEN, M. E. SNOOK, T. A. MUSKET, and J. M. THEURI et al., 1996  Quantitative trait loci and metabolic pathways: genetic control of the concentration of maysin, a corn earworm resistance factor, in maize. Proc. Natl. Acad. Sci. USA 93:8820-8825[Abstract/Free Full Text].

CHEN, J. D. and V. PIRROTTA, 1993  Multimerization of the Drosophila zeste protein is required for efficient DNA binding. EMBO J. 12:2075-2083[Medline].

CLARK, A. G. and L. WANG, 1997  Epistasis in measured genotypes: Drosophila P-element insertions. Genetics 147:157-163[Abstract].

CLARK, A. G., L. WANG, and T. HULLEBERG, 1995  P-element induced variation in metabolic regulation in Drosophila. Genetics 139:337-348[Abstract].

CREWS, S. T., 1998  Control of cell lineage-specific development and transcription by bHLH-PAS proteins. Genes Dev. 12:607-620[Free Full Text].

CUBADDA, Y., P. HEITZLER, R. P. RAY, M. BOUROUIS, and P. RAMAIN et al., 1997  u-shaped encodes a zinc finger protein that regulates the proneural genes achaete and scute during the formation of bristles in Drosophila. Genes Dev. 11:3083-3095[Abstract/Free Full Text].

DAMERVAL, C., A. MAURICE, J. M. JOSSE, and D. DE VIENNE, 1994  Quantitative trait loci underlying gene product variation: a novel perspective for analyzing regulation of genome expression. Genetics 137:289-301[Abstract].

DAVISON, D., C. CHAPMAN, C. WEDEEN, and P. M. BINGHAM, 1985  Genetic and physical studies of a portion of the white locus participating in transcriptional regulation and in synapsis-dependent interactions in Drosophila adult tissues. Genetics 110:479-494[Abstract/Free Full Text].

DE RUBERTIS, F., D. KADOSH, S. HENCHOZ, D. PAULI, and G. REUTER et al., 1996  The histone deacetylase RPD3 conteracts genomic silencing in Drosophila and yeast. Nature 384:589-591[Medline].

DEVLIN, R. H., D. G. HOLM, and T. A. GRIGLIATTI, 1982  Autosomal dosage compensation in Drosophila melanogaster strains trisomic for the left arm of chromosome 2. Proc. Natl. Acad. Sci. USA 79:1200-1204[Abstract/Free Full Text].

DEVLIN, R. H., D. G. HOLM, and T. A. GRIGLIATTI, 1988  The influence of whole arm trisomy on gene expression in Drosophila. Genetics 118:87-101[Abstract/Free Full Text].

DORN, R., V. KRAUSS, G. REUTER, and H. SAUMWEBER, 1993  The enhancer of position-effect variegation of Drosophila, E(var)3-93D, codes for a chromatin protein containing a conserved domain common to several transcriptional regulators. Proc. Natl. Acad. Sci. USA 90:11376-11380[Abstract/Free Full Text].

DREESEN, T. D., D. H. JOHNSON, and S. HENIKOFF, 1988  The brown protein of Drosophila melanogaster is similar to the white protein and to components of active transport complexes. Mol. Cell. Biol. 8:5206-5215[Abstract/Free Full Text].

EISSENBERG, J. C., T. C. JAMES, D. M. FOSTER-HARTNETT, V. NGAN, and S. C. ELGIN, 1992  The heterochromatin-associated protein HP-1 is an essential protein in Drosophila with dosage-dependent effects on position-effect variegation. Genetics 131:345-352[Abstract].

ELGIN, S. C. R., 1996  Heterochromatin and gene regulation in Drosophila. Curr. Opin. Genet. Dev. 6:193-202[Medline].

FARKAS, G., J. GAUSZ, M. GALLONI, G. REUTER, and H. GYURKOVICS et al., 1994  The Trithorax-like gene encodes the Drosophila GAGA factor. Nature 371:806-808[Medline].

FROLOV, M. V., E. V. BENEVOLENSKAYA, and J. A. BIRCHLER, 1998  Regena (Rga), a Drosophila homolog of the global negative transcriptional regulator CDC36 (NOT2) from yeast, modifies gene expression and suppresses position-effect variegation. Genetics 148:317-329[Abstract/Free Full Text].

GARZINO, V., A. PEREIRA, P. LAURENTI, Y. GRABA, and R. W. LEVIS et al., 1992  Cell lineage-specific expression of modulo, a dose-dependent modifier of variegation in Drosophila. EMBO J. 11:4471-4479[Medline].

GREEN, M. M., 1963  Unequal crossing over and the genetical organization of the white locus of Drosophila melanogaster. Z. Verebungsl. 94:200-214.

GUO, M. and J. A. BIRCHLER, 1994  Trans-acting dosage effects on the expression of model gene systems in maize aneuploids. Science 266:1999-2002[Abstract/Free Full Text].

HENIKOFF, S., 1996  Dosage dependent modification of position-effect variegation in Drosophila. BioEssays 18:401-409[Medline].

HIEBERT, J. C. and J. A. BIRCHLER, 1994  Effects of the maleless mutation on X and autosomal gene expression in Drosophila melanogaster. Genetics 136:913-926[Abstract].

KARPEN, G. and A. C. SPRADLING, 1990  Reduced DNA polytenization of a minichromosome region undergoing position-effect variegation in Drosophila. Cell 63:97-107[Medline].

KENNISON, J. A., 1995  The Polycomb and trithorax Group proteins of Drosophila: trans-regulators of homeotic gene function. Annu. Rev. Genet. 29:289-303[Medline].

LAURIE, C. C., J. R. TRUE, J. LIU, and J. M. MERCER, 1997  An introgression analysis of quantitative trait loci that contribute to a morphological difference between Drosophila simulans and Drosophila mauritiana. Genetics 145:339-348[Abstract].

LIU, J., J. M. MERCER, L. F. STAM, G. GIBSON, and C. C. LAURIE, 1996  Genetic analysis of a morphological shape and difference in the male genitalia of Drosophila similans and D. mauritiana. Genetics 142:1129-1145[Abstract].

LOCKE, J., M. A. KOTARSKI, and K. D. TARTOF, 1988  Dosage dependent modifiers of position-effect variegation in Drosophila and a mass action model that explains their effect. Genetics 120:181-198[Abstract/Free Full Text].

MACKAY, T. F. C., 1996  The nature of quantitative genetic variation revisited: lessons from Drosophila bristles. BioEssays 18:113-121[Medline].

O'HARE, K., R. LEVIS, and G. M. RUBIN, 1983  Transcription of the white locus in Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 80:6917-6921[Abstract/Free Full Text].

ORR, H. A., 1991  A test of Fisher's theory of dominance. Proc. Natl. Acad. Sci. USA 88:11413-11415[Abstract/Free Full Text].

PAL BHADRA, M., U. BHADRA, and J. A. BIRCHLER, 1998  Role of multiple trans-acting regulators in modifying the effect of the retrotransposon, copia, on host gene expression in Drosophila. Mol. Gen. Genet. in press.

RABINOW, L., A. T. NGUYEN-HUYNH, and J. A. BIRCHLER, 1991  A trans-acting regulatory gene that inversely affects the expression of the white, brown and scarlet loci in Drosophila. Genetics 129:463-480[Abstract].

RAHA, S. F., F. MERANTE, G. PROTEAU, and J. K. REED, 1990  Simultaneous isolation of total cellular RNA and DNA from tissue culture cells using phenol and lithium chloride. Genet. Anal. Tech. Appl. 7:173-177[Medline].

REUTER, G. and P. SPIERER, 1992  Position-effect variegation and chromatin proteins. BioEssays 14:605-612[Medline].

SABL, J. and J. A. BIRCHLER, 1993  Dosage dependent modifiers of white alleles in Drosophila melanogaster. Genet. Res. 62:15-22[Medline].

SEUM, C., A. SPIERER, D. PAULI, J. SZIDONYA, and G. REUTER et al., 1996  Position-effect variegation in Drosophila depends on the dose of the gene encoding the E2F transcriptional activator and cell cycle regulator. Development 122:1949-1956[Abstract].

SPOFFORD, J. B., 1976 Position-effect variegation in Drosophila, pp. 955–1019 in Genetics and Biology of Drosophila, Vol. IC, edited by M. ASHBURNER and E. NOVITSKI. Academic Press, London.

TANKSLEY, S. D., 1993  Mapping polygenes. Annu. Rev. Genet. 27:205-233[Medline].

TEARLE, R. G., J. M. BELOTE, M. MCKEOWN, B. S. BAKER, and A. J. HOWELLS, 1989  Cloning and characterization of the scarlet gene of Drosophila melanogaster. Genetics 122:595-606[Abstract/Free Full Text].

THATCHER, J. W., J. M. SHAW, and W. J. DICKINSON, 1998  Marginal fitness contributions of nonessential genes in yeast. Proc. Natl. Acad. Sci. USA 95:253-257[Abstract/Free Full Text].

TSCHIERSCH, B., A. HOFMANN, V. KRAUSS, R. DORN, and G. KORGE et al., 1994  The protein encoded by the Drosophila position effect variegation suppressor gene Su(var)3-9 combines domains of antagonistic regulators of homeotic gene complexes. EMBO J. 13:3822-3831[Medline].

TYLER, J. K., M. BULGER, R. T. KAMAKAKA, R. KOBAYASHI, and J. T. KADONAGA, 1996  The p⁵⁵ subunit of Drosophila chromatin assembly factor 1 is homologous to a histone deacetylase-associated protein. Mol. Cell Biol. 16:6149-6159[Abstract].

WEILER, K. S. and B. T. WAKIMOTO, 1995  Heterochromatin and gene expression in Drosophila. Annu. Rev. Genet. 29:577-605[Medline].

WEINTRAUB, H., 1993  The myoD family and myogenesis: redundancy, networks and thresholds. Cell 75:1241-1244[Medline].

ZACHAR, Z. and P. M. BINGHAM, 1982  Regulation of white locus expression: the structure of mutant alleles at the white locus in Drosophila melanogaster. Cell 30:529-541[Medline].

This article has been cited by other articles:

				N. Malmanche and D. V. Clark Identification of Trans-dominant Modifiers of Prat Expression in Drosophila melanogaster Genetics, August 1, 2003; 164(4): 1419 - 1433. [Abstract] [Full Text] [PDF]

				U. Bhadra, M. Pal-Bhadra, and J. A. Birchler Histone Acetylation and Gene Expression Analysis of Sex lethal Mutants in Drosophila Genetics, June 1, 2000; 155(2): 753 - 763. [Abstract] [Full Text]