Study of Dosage Compensation in Drosophila
Pei-Wen Chiang, David M. Kurnit

Abstract

Using a sensitive RT-QPCR assay, we analyzed the regulatory effects of sex and different dosage compensation mutations in Drosophila. To validate the assay, we showed that regulation for several genes indeed varied with the number of functional copies of that gene. We then confirmed that dosage compensation occurred for most genes we examined in male and female flies. Finally, we examined the effects on regulation of several genes in the MSL pathway, presumed to be involved in sex-dependent determination of regulation. Rather than seeing global alterations of either X chromosomal or autosomal genes, regulation of genes on either the X chromosome or the autosomes could be elevated, depressed, or unaltered between sexes in unpredictable ways for the various MSL mutations. Relative dosage for a given gene between the sexes could vary at different developmental times. Autosomal genes often showed deranged regulatory levels, indicating they were in pathways perturbed by X chromosomal changes. As exemplified by the BR-C locus and its dependent Sgs genes, multiple genes in a given pathway could exhibit coordinate regulatory modulation. The variegated pattern shown for expression of both X chromosomal and autosomal loci underscores the complexity of gene expression so that the phenotype of MSL mutations does not reflect only simple perturbations of genes on the X chromosome.

FOR the X-linked genes, which differ in dose between male and female, there is a regulatory mechanism of dosage compensation operating such that the resulting expression of these X-linked genes is usually equivalent in the two sexes. In Drosophila, regulation is posited to occur by hypertranscription of loci on the single X chromosome in males (reviewed in Kelley and Kuroda 1995; Bashaw and Baker 1996; Cline and Meyer 1996; Lucchesi 1998; Stuckenholzet al. 1999). The current model for dosage compensation in Drosophila proposes that five MSL (male-specific-lethal) proteins, MSL1, MSL2, MSL3, MLE (encoded by maleless), and MOF (encoded by male-absent on the first), form a heteromeric complex on the male X chromosome. Binding of the MSL complex to the male X chromosome leads to the acetylation of histone H4 at lysine 16 (H4Ac16) and an increase of transcription (reviewed in Kelley and Kuroda 1995; Bashaw and Baker 1996; Lucchesi 1998; Stuckenholzet al. 1999). Further, the twofold upregulation is superimposed on the gene-specific transcription factors that control the amount, time, and tissue of expression. The finding that MOF is a homolog of human histone acetyltransferases TIP60 (HIV Tat-interacting protein of 60 kD) and MOZ (monocytic leukemia zinc finger protein) provides evidence for the role of histone acetyltransferase in dosage compensation (Hilfikeret al. 1997; Guet al. 1998).

Even though the current model of dosage compensation in Drosophila receives wide acceptance, direct experimental data in support of this model are still lacking. A twofold reduction in females of Sgs4, derived from the X chromosome, was reported (Belote and Lucchesi 1980; Breen and Lucchesi 1986). However, when more transcripts were analyzed using Northern blot analysis, a different conclusion was reached (Hiebert and Birchler 1994; Bhadraet al. 1999). In the alternative dosage compensation model proposed by Birchler (1996), the concentration of MSL complex on the X chromosome is not related to dosage compensation. Rather, the concentration of MSL complex on the X chromosome prevents nonspecific binding of the MSL complex to autosomes. The dosage compensation complex cannot accumulate on the male X chromosome in the mutants. Instead, the complex spreads on the autosomes. As a result, autosomal genes are hypertranscribed but the expression of X-linked genes is not changed.

The study of dosage compensation provides a unique opportunity to understand regulation at the chromosomal level. However, the conflicting results in the study of dosage compensation point to the desirability for a rapid and reliable assay to monitor the regulation of given genes. Recently, we developed a sensitive reverse transcriptase-quantitative PCR (RT-QPCR) assay to measure the relative expression of steady-state RNA in Drosophila. The success in the analysis of deficiency flies validated our approach (Chianget al. 1999). The RT-QPCR assay has the following advantages compared with previous quantitative assays of gene expression: (1) Only a small amount of total RNA is needed in the assay, (2) transcripts from any gene (abundant or not) can be analyzed because the assay works over six orders of magnitude, and (3) the reproducibility and resolution of the RT-QPCR assay was documented by studies of different systems in Drosophila (Chianget al. 1999).

To assay the hypotheses that seek to explain gene expression from the X chromosome in the different sexes, a more sensitive and convenient method should be used to test actual gene expression. To this end, we used TaqMan RT-QPCR to measure the level of steady-state RNA for genes located on both the X chromosome and the autosomes in wild-type flies and in flies with dosage compensation mutations. We demonstrated the use of the technique to detect gene dosage in normal males and females. Our data demonstrate that different loci on both the X chromosome and the autosomes show different patterns of expression in the mutants. The variegated pattern shown for expression of both X chromosomal and autosomal loci underscores the lack of a simple explanation for the anomalies in the dosage compensation mutants. This study of mutants of dosage compensation illustrates the complexity of regulation that may best be comprehended by the elucidation of regulatory pathways where the variation of sets of genes occurs concomitantly.

MATERIALS AND METHODS

Research design: The steady-state RNA (total RNA) from a mix of at least 20 third instar larvae was analyzed for each sample. The standard curve method (user bulletin no. 2; ABI PRISM 7700 sequence detection system) was used for the measurement (Chianget al. 1999). RNA derived from the mix of at least 20 third instar larvae from the Canton-S female fly was used as the standard. The quantity of RNA samples was estimated by spectrophotometry and was diluted to different concentrations. For the standards, four concentrations (5, 2.5, 1.25, and 0.6 ng/μl) were used in the analysis (although knowledge of the absolute amounts of RNA is necessarily approximate, the relative amounts used for the analysis are accurate within the limits of pipetting accuracy). To test the reproducibility of the experiments, two different concentrations of experimental samples (2.5 and 1.25 ng/μl) and three repeats of each dilution were used in the analysis. The experiment was repeated if reproducibility of the three repeats yielded variation >30%. Similarly, the experiment was repeated if a twofold difference (the cutoff was between 0.7 and 1.3) was not observed. During calculation, the numbers 5, 2.5, 1.25, and 0.6 were assigned to the standards and the quantity of specific RNA in the samples was calculated on the basis of relative amount of targets in the samples using the standard as a reference. Since absolute quantitation of RNA is difficult to achieve, a relative ratio approach was used in our calculation. After testing different markers in different wild-type and mutant flies, autosomal markers derived from either RNA polymerase II or the chromatin protein Dspt4 were found to be the most reliable markers for the calculation of relative ratio (data not shown). Therefore, all the measurements in this study were done using a marker from RNA polymerase II as the reference for the calculation of relative ratio. As a result, 1.0 on the ordinate of the tables designates a relative amount of RNA product equal to the relative amount of RNA product obtained using a primer set for RNA polymerase II. The Dspt4 marker was also included in all the measurements to ensure validity. The ratio between Dspt4 and RNA polymerase II was always close to 1 in all the measurements, suggesting the consistency of our analyses. This normalization procedure enabled us to compare accurately the expression of a given gene in two flies.

The primer sets were designed following the rules recommended by the manufacturer (PE Biosystems, Foster City, CA) and are listed in Table 1. The TaqMan Gold RT-PCR reagents were used in our analysis (PE Biosystems). All primer sets were designed to amplify only RNA as they crossed exons. Only the BR-C primer set potentially reacted with both DNA and RNA, as the set was designed so that the first 6 nucleotides of the BR-C forward primer derived from the first exon and the remaining 14 nucleotides derived from the second exon. Using this strategy, the amplification from genomic DNA was still depressed by 100-fold compared with the cDNA reaction, enabling us to distinguish genomic DNA from cDNA.

Flies: yw; mle1/CyO, y+ and the the four hypomorphic mle alleles (mle-AAA, mle-GNT, mle-DQIH, and mle-GET) were provided by Mitzi Kuroda. yw; mleRK/CyO, y+ and Df(1)lD34 were provided by Barry Ganetzky. y1 mof1/Basc was provided by John Lucchesi. C(1)DX, y1, f1/y2Y67g19.1 was provided by Cynthia Bayer. Deficiency flies Df(1)S39/FM6 (deletion of BR-C), Df(1)5D/FM6 (deletion of Dspt6), Df(1)Pgd-kz/FM6 (deletion of 6-Pgd), and Df(1)N-264-105/FM1 (deletion of Sgs4) were ordered from the Bloomington Stock Center. FM7i and C(1)DX, y1, f1/Y were also obtained from the Bloomington Stock Center. The metafemale was produced by crossing an attached-X strain (C(1)DX, y1, f1/Y) marked with yellow (y) to a standard male strain and selected on the basis of gonad size and the color of the mouth hook (y+). The flies were kept at 25° in this study. The difference of the size of gonads between male and female can be easily differentiated at the third instar larvae (the male has a gonad significantly bigger than that of the female). Therefore, in most of our analysis, RNA from third instar larvae was used.

RESULTS

Testing the resolution of quantitative PCR: We have shown that RT-QPCR analysis has the twofold resolution required to detect reductions of Dspt4 and Dspt6 transcripts in deficiency flies (Chianget al. 1999). To further confirm this resolution of RT-QPCR, deficiency flies with a deletion of BR-C, Sgs4, para, or 6-Pgd were analyzed by RT-QPCR (Figure 1). Each of the deficiency chromosomes was paired with the balancer, FM7i. The heterozygous deficiency/FM7i larvae were separated from homozygous FM7i/FM7i larvae on the basis of the color of the mouth hook. As expected, the expression of Sgs4, para, and 6-Pgd was reduced in the deficiency flies Df(1)N-264-105/FM7i, Df(1)lD34/FM7i, and Df(1)Pgd-kz/FM7i, respectively, compared with the expression in FM7i/FM7i. Surprisingly, the expression of BR-C was not reduced in the deficiency fly Df(1)S39/FM7i compared with the FM7i/FM7i female. Since the BR-C region was shown to be deleted in Df(1)S39 (Belyaevaet al. 1980), a DNA QPCR analysis was performed to confirm the deletion of BR-C in Df(1)S39. The relative gene copy number ratio of BR-C/Dspt6 was measured in Df(1)S39/FM7i and FM7i/FM7i. In accordance with the previous electron microscopic analysis (Belyaevaet al. 1980), BR-C was indeed deleted in Df(1)S39 (Figure 1).

View this table:
TABLE 1

List of the PCR primers used in this study

Figure 1.

—Analysis of deficiency flies. Each analysis was done at two different dilutions, with a minimum of four separate analyses for each dilution. Both heterozygous deficiency and homozygous balancer flies were analyzed. The expression of Dspt6 was also analyzed as a control. The order of each analysis is: (1) heterozygous deficiency (white); (2) 1:1 dilution (with an equal volume of water) of heterozygous deficiency (white); (3) homozygous balancer (gray); and (4) 1:1 dilution of homozygous balancer (gray). The copy number of BR-C (DNA) was also analyzed by DNA QPCR in heterozygous deficiency (white) and homozygous normal (gray) flies. In all figures, the standard deviation (95%) is indicated.

To demonstrate gene dosage in Drosophila, 11 primer sets derived from 5 X-linked genes (Sgs4, 6-Pgd, Dspt6, para, and BR-C) and 6 autosomal genes (st, Dspt4, Sgs3, Sgs5, Sgs7, and Sgs8) were analyzed. The expression of these 11 genes was dosage compensated in the male as shown by TaqMan RT-QPCR in wild-type male and female third instar larvae from Canton-S (Figure 2 for the X-linked genes and data not shown for the autosomal genes). Previously, the Lsp-1α protein was shown not to be dosage compensated in male flies (Roberts and Evans-Roberts 1979). To provide more evidence of the resolution of RT-QPCR, the expression of Lsp-1α was analyzed in wild-type male and female third instar larvae from Canton-S. As shown in Figure 2, 30% less Lsp-1α transcript was detected in male flies compared with female flies.

A cross between C(1)DX, y1, f1/Y females and wild-type Canton-S males generates two females, C(1)DX, y1, f1/Y and C(1)DX, y1, f1/X (metafemale), with identical genetic backgrounds except for the sex chromosome. These two females can be separated by the color of the mouth hook. To confirm the genotypes of the third instar larvae from metafemale and female, we analyzed the relative DNA copy number of Dspt6 (an X-linked gene) in metafemale, female, and male against an autosomal gene standard, Dspt4. As shown in Figure 3, the QPCR analysis clearly differentiated the DNA copy number differences of Dspt6 among metafemale (three copies), female (two copies), and male (one copy).

Mutations of mle affect expression of a subset of genes on both the X chromosome and the autosomes: We analyzed one null allele of mle (mle1; Fukunagaet al. 1975; Figure 4, A and B) and one missense mutation of mle (mleRK; Kernanet al. 1991; Figure 4, C and D). Both mle alleles are male lethal at the late third instar to the early pupal stage. Homozygous female larvae are viable for both mle alleles. Homozygous third instar larvae from male and female were separated for each allele and total RNA was isolated from the larvae. Since no significant difference was detected among the heterozygous male, heterozygous female, and homozygous female (data not shown), the analysis was performed using the homozygous male and homozygous female third instar larvae. Both homozygous male and homozygous female larvae have the same genetic background except for the sex chromosome. Therefore, an effect from the mle mutations could be detected in our analysis. The amount of Dspt6 transcript was reduced twofold for both mle alleles in the homozygous male compared with the homozygous female. In contrast, transcripts derived from para and 6-Pgd were not significantly different in the homozygous male compared with the homozygous female. Three- to fourfold reductions of BR-C and Sgs4 transcripts were observed in the homozygous male mleRK larvae compared with the homozygous female mleRK larvae. A twofold reduction was detected in homozygous mle1 male larvae for BR-C and a two- to threefold reduction was detected for Sgs4 compared with homozygous mle1 female larvae. For the three autosomal (chromosome 3L) genes, Sgs3, Sgs7, and Sgs8, a five- to eightfold reduction and a two- to threefold reduction were seen in homozygous mleRK and mle1 males, respectively. For the autosomal 3R gene Sgs5, the reduction was two- to threefold in mle1 and fivefold in mleRK. In contrast, the autosomal 2R gene Dspt4 was not altered in the homozygous male compared with the homozygous female in both mleRK and mle1. Finally, a four- to fivefold increase of the expression of autosomal 3L gene st was detected in both mle1 and mleRK flies.

Figure 2.

—RT-QPCR analysis of wild-type larvae. The order of each analysis is: (1) female (white); (2) 1:1 dilution of female (white); (3) male (gray); and (4) 1:1 dilution of male (gray).

Analysis of mle hypomorphic alleles: We then analyzed four mle alleles generated through site-specific mutagenesis (Richteret al. 1996; Leeet al. 1997; Figure 5). All four alleles were analyzed in third instar larvae on an otherwise mle1 background. A total of seven primer sets (6-Pgd, Dspt6, para, BR-C, Sgs4, Sgs5, and st) were used in the analysis. As shown in Figure 5, A and B, none of the genes analyzed from the AAA and GNT alleles showed twofold variation compared with the homozygous female. The largest variation we observed was an approximate 40-50% reduction of br transcript in the homozygous AAA and GNT males.

We examined the nearly lethal allele GET (Figure 5C) and the less severe sublethal allele DQIH (Figure 5D). A twofold reduction of BR-C, Dspt6, Sgs4, and Sgs5 was observed in the homozygous DQIH male compared with the homozygous female. An approximate 30-40% reduction of the expression of 6-Pgd was observed in the homozygous DQIH male and the expression of para was not changed in the homozygous DQIH male. In contrast with our finding of increased expression for st in mle1, the expression of st did not vary in the homozygous DQIH male. A different picture emerged when the nearly lethal allele GET was analyzed. No significant variation was observed for the expression of 6-Pgd, para, Sgs4, and Sgs5 in the homozygous GET male. An approximate 30-50% reduction of the expression of BR-C and Dspt6 was observed in the homozygous GET male. Finally, an approximate twofold increase of the expression of st was observed in the homozygous GET male.

Analysis of mof mutation: To further our study of MSL variants, mutation of a different MSL gene (mof; Hilfikeret al. 1997) was analyzed (Figure 6, A and B). Since mof is X linked, only three types of larvae could be isolated (mof1/balancer, balancer/Y, and mof1/Y). The expression of Dspt6 in the mof1 hemizygous male was reduced 2-fold and the expression of 6-Pgd, para, and Dspt4 in the mof1 male did not vary significantly when compared to expression in the balancer male. However, a 5-fold reduction of BR-C was seen in the mof1 hemizygous male compared with the balancer male. Ten- to 30-fold reductions of Sgs3, Sgs4, Sgs5, Sgs7, and Sgs8 were detected in the mof1 hemizygous male compared with the balancer male. Finally, a 2-fold increase of the exp2ression of st was detected in the mof1 hemizygous male compared with the balancer male.

Figure 3.

—DNA QPCR analysis of C(1)DX (XX/X; white); y1, f1/X, and C(1)DX (XX/Y; light gray); and y1, f1/Y, and X/Y (X/Y; dark gray). The relative DNA copy number of Dspt6 (an X-linked gene) in metafemale, female, and male against an autosomal gene Dspt4 was measured. For each genotype, no dilution and a twofold dilution of samples are measured.

It was desirable to examine younger larvae to minimize any effects of dying that might confound analyses of older larvae. Younger y1 mof1/Y males (first and second instar) larvae can be separated from the mix of Basc/Y and y1 mof1/Basc larvae on the basis of the color of the mouth hook. Since the BR-C and Sgs genes are expressed later, beginning at the third instar, only 6-Pgd, para, Dspt6, and st could be analyzed in these two groups of early larvae (Figure 6C). Similar to the analysis of third instar larvae, the expression of Dspt6 was reduced twofold in younger y1 mof1/Y and the expression of para was not changed. However, the expression of 6-Pgd was reduced twofold in younger y1 mof1/Y in contrast to the lack of significant change observed in the third instar larvae. Further, a less than twofold (40-50%) reduction of the expression of st was seen in the younger larvae in contrast to the twofold increase of expression in the third instar male. Since the expression of 6-Pgd, para, Dspt6, and st was analyzed only in the third instar larvae from wild-type Canton-S (Figure 2), the expression pattern of these four genes at the younger developmental stage was also analyzed in wild type. The FM7i female was crossed to the wild-type Canton-S male. From this cross, the female offspring was y+ and the male offspring was y-. Therefore, the first and second instar male and female larvae could be separated from each other on the basis of the color of mouth hook. The RT-QPCR was performed on the younger wild-type samples and the expression of 6-Pgd, para, Dspt6, and st was similar between male and female in all cases (Figure 6D). This control confirmed the suitability of these analyses in these young larvae.

DISCUSSION

Validity of the RT-QPCR methodology as used to investigate the dosage compensation model: It is postulated that the transcriptional product of the majority of genes, be they autosomal or X chromosomal, is equivalent in males and females (reviewed in Kelley and Kuroda 1995; Cline and Meyer 1996). As shown in Figure 2, our sensitive RT-QPCR analysis agrees with this assertion, as dosage compensation between wild-type males and females can be seen at the regulatory level for five X-linked genes. The one case where dosage compensation was known not to exist (Lsp-1α; Roberts and Evans-Roberts 1979) did not show dosage compensation by RT-QPCR. These collective data are consistent with the hypothesis that RT-QPCR accurately quantitates transcriptional product, verifying the concept that the steady-state amount of product of a given gene is usually (but not always) equivalent in males and females. Taken together with our previous publication (Chianget al. 1999), the work confirms the ability of the TaqMan QPCR and RT-QPCR methodologies to discern twofold dosage differences at both the DNA and RNA levels in Drosophila. These studies also substantiate the ability of the method to work with whole flies as opposed to using only selected tissues from a fly.

No simple rule for effect of MSL mutations on regulation: Our initial hypothesis was that application of the TaqMan QPCR and RT-QPCR methodologies would allow us to detect simple rules pertaining to the various entities comprising the MSL pathway. Surprisingly, our analyses demonstrated that there is no simple rule that can explain the effects of the MSL network on regulation. These analyses showed that both autosomal and X chromosomal genes are expressed in patterns that could not be predicted in a straightforward manner. There are genes on autosomes whose expression is altered and genes on the X chromosome that do not show dosage compensation, and the quantitative anomalies are often not consistent with the chromosomal dosage and/or dosage compensation.

Relative effects on gene expression for the various mle and mof mutants: Since males carrying homozygous mle lethal alleles and mof1 hemizygous males can survive through the late third instar to early pupal stage, the amount of steady-state mRNA could be analyzed in third instar larvae. If lethality reflected gene expression in the homozygous males, one would expect the effect on the male X chromosome to be as follows: mle1 = mof1GET > DQIH > GNT > AAA. (The effect from mleRK is more difficult to predict since a dominant negative effect could occur.) Instead, we observed a different order of negative effects on the expression of X-linked genes: mof1 > mleRK > mle1DQIH > GET > GNT = AAA. The limited numbers of X-linked genes analyzed and the ultimate resolution of RT-QPCR could be responsible for the similar patterns observed in GNT and AAA alleles in our analysis. Even so, it is clear from our analysis that the expression patterns we observed are not simply related to the lethal effects caused by the mle or mof mutations.

Our RT-QPCR analysis indicated that the interpretation of regulation of the dosage compensation mutants is complex. The twofold reductions of BR-C, Sgs4, and Dspt6 in mle1 and DQIH mutants supported the current model. However, we did not observe the expected two-fold reduction for the X-linked loci, para and 6-Pgd, both of which remain dosage compensated in mutant males [para may be a special case since mle may participate in its splicing and editing (Hanrahanet al. 2000)]. Analysis of younger mof1 mutants (that are not sick) confirmed the analysis on Dspt6 and para observed at the later third instar stage. Another unexpected finding from our analysis was that the reduction of transcripts from both BR-C and its downstream Sgs genes (including the Sgs4 gene on the X chromosome) was greater than twofold in both the mof1 and mleRK male flies. The greater than twofold reduction suggested that the regulation of BR-C on the single X chromosome is influenced by these mutations.

Since the younger y1 mof1/Y larvae were still healthy and all four genes analyzed were dosage compensated at the younger stage of wild-type flies, the effects we observed in younger y1 mof1/Y males reflect the effect of the mof mutation on gene expression.

The similarity of the mle and mof data on the expression patterns of the genes we examined (cf. Figures 4 and 6) favors the hypothesis that the functions of these genes are alike. This provides evidence for the upregulation model (in which MLE and MOF act together to upregulate the male X; reviewed in Kelley and Kuroda 1995; Bashaw and Baker 1996; Cline and Meyer 1996; Lucchesi 1998; Stuckenholzet al. 1999), rather than the inverse model (in which the absence of MOF should differ from the absence of MLE that sequesters active MOF; Birchler 1996).

Interconnected regulation networks: We postulate that some of the effects we observed resulted from interconnected regulation networks. In particular, expression of the Sgs genes (on both the autosomes and the sex chromosome) correlated with the amount of BR-C transcript in the mutants. This correlation was seen in a variety of mutants. The most severe reduction was observed in the mof1 mutant. A 5-fold reduction of BR-C, and 10- to 30-fold reductions of Sgs3, Sgs4, Sgs5, Sgs7, and Sgs8 were detected in the mof1 hemizygous male compared with the balancer male. A 3- to 4-fold reduction of BR-C and 3- to 8-fold reductions of Sgs3, Sgs4, Sgs5, Sgs7, and Sgs8 were detected in mleRK homozygous males compared with homozygous females. The reduction of expression of autosomal Sgs genes is ∼2- to 3-fold in the mle1 and DQIH homozygous males, where the reduction of BR-C is 2- to 3-fold. Thus, our analysis delineates a strong correlation between BR-C expression and expression of the Sgs regulation units in a variety of flies. Presumably this reflects the regulation network documented previously, in which BR-C mutations were shown to decrease the induction of the intermolt genes Sgs3, Sgs4, and Sgs5 (Guay and Guild 1991; Karimet al. 1993). Thus, the data is compatible with BR-C transcript levels controlling the transcript levels of the Sgs genes.

Figure 4.

—RT-QPCR analysis of mle1 (A and B) and mleRK (C and D). The order of each analysis is: (1) homozygous female (white); (2) 1:1 dilution of homozygous female (white); (3) homozygous male (gray); and (4) 1:1 dilution of homozygous male (gray). Analysis of the X chromosome genes are shown in A and C and analysis of the autosomal genes are shown in B and D.

Figure 5.

—RT-QPCR analysis of AAA (A), GNT (B), GET (C), and DQIH (D) on an otherwise deficient (mle1) background. The order of each analysis is: (1) female (white); (2) 1:1 dilution of female (white); (3) male (gray); and (4) 1:1 dilution of male (gray).

Figure 6.

—RT-QPCR analysis of mof1. (A and B) Third instar larvae. (C) Young mof1 larvae. (D) Young larvae from wild type. (A and B) The order of each analysis is: (1) mof1/Basc (indicated as mof/+; white); (2) 1:1 dilution of mof1/Basc (white); (3) Basc/Y (black); (4) 1:1 dilution of Basc/Y (black); (5) mof1/Y (gray); and (6) 1:1 dilution of mof1/Y (gray). Basc indicates the presence of a balancer chromosome.

Developmental effects of MSL mutations: The effects on regulation of a gene can vary at different developmental stages. The expression of the autosomal gene st provides a different picture of altered regulation than the BR-C and Sgs genes do in that st is upregulated in general, rather than downregulated, in the MSL mutants. An approximately fourfold increase of expression relative to the wild type was observed in the mle1 and mleRK homozygous male mutants. An approximately twofold increase was observed in the mof1 hemizygous mutant. In contrast to several other genes analyzed, the expression of st was changed in GET mutants but not in DQIH mutants. When younger mof1 larvae were analyzed, a small (40-50%) reduction of the expression of st was seen in younger male larvae in contrast to the twofold increase of expression in third instar males. Therefore, altered regulation of st in these mutants is developmentally regulated, manifesting different regulation asymmetries between the sexes at different developmental stages. This underscores the complexity of effects caused by these mutations.

A structural gene on the X chromosome could have positive and/or negative regulators on the same chromosome. Similarly, a structural gene on an autosome could also have positive and/or negative regulators on the X chromosome. Any specific effect on the X chromosome could then trigger a chain reaction for the expression of other genes on the X chromosome or on autosomes. The regulation of the Sgs genes by BR-C was shown previously (Guay and Guild 1991; Renaultet al. 2001) and could explain the interrelationships we observed among these genes. The factor(s) directly or indirectly involved in the regulation of st is not known. For example, at least one of the factors could be on the X chromosome and this unknown factor could be regulated developmentally. The presence of developmentally regulated factors was further suggested by our analysis of 6-Pgd. No significant variation of 6-Pgd levels was observed in third instar larvae of the mof1 mutant but a twofold reduction was observed in the younger larvae of the mof1 mutant. The point that ensues from this analysis is that it is necessary to ascertain the behavior of a gene (and when applicable, its interconnected regulation network pathway) throughout development to describe the effects of that gene (and when applicable, its network). For example, the more than twofold reduction of BR-C in some dosage compensation mutants might result from the effect of an interconnected regulation network of BR-C and its positive regulator(s).

Implications for understanding development: The importance of the interconnected regulation networks, with its resultant complexities, was overlooked in the past:

  1. Understanding the role of aneusomic genomes in tumor cells or in human genetic diseases (e.g., Down syndrome) will involve the complexity of interconnected regulation networks. In other words, the etiology of some features of Down syndrome may not be caused simply by the expected 1.5-fold increase of steady-state transcriptional product from genes on chromosome 21 (Kurnit 1979). The disturbance of regulation of genes on other chromosomes may be involved as well.

  2. A more detailed molecular analysis of different functional domains of the dosage compensation complex could be initiated on the basis of our analysis. The expected twofold reduction of Dspt6, BR-C, and Sgs4 transcripts provides an opportunity to study the functional domains of MLE in the mle1 background. As illustrated in our study, the function of helicase activity in MLE could be separated from the lethal effect caused by a mutation of mle. Without the helicase activity, the hypertranscription activity of the MSL complex is still intact. Similar analyses may be extended to other msl genes.

  3. No two genes share a common expression pattern in all the mutants we analyzed. Therefore, regulation happens at the level of each individual gene. Indeed, our analyses show that it is not possible to predict the dosage of a gene in a given MSL mutant. The situation is further complicated by the finding that dosage varies at different developmental stages. The fact that young mutant larvae can give different results than older mutant larvae is additional evidence for the regulation network idea, as indirect effects may be amplified with time and/or proximity to death.

Desirability of global analyses: With the demonstration that the MSL pathway yields more complex regulation of gene expression than was previously thought, this requires more complicated analyses to identify pathways affected by the MSL genes. In turn, indirect effects of the MSL mutants can be expected, potentially extending to unexpected effects affecting entire pathways. Thus, a more global comparison such as microarray analysis (Schenaet al. 1995; Churchill 2002) is indicated to outline the genes that belong to potential pathways (e.g., the Sgs genes regulated by BR-C gene) influenced by the MSL genes. Indeed, this technology has been used to study sex determination in Drosophila (Churchill and Oliver 2001; Jinet al. 2001). Greater detail for the members of an indicated pathway can then be obtained by analyses using methodologies such as those presented herein that can potentially detect small differences. Although our data demonstrated that regulation of a gene cannot be deduced a priori, it can be measured using RT-QPCR analysis in a given case. With the success of such analyses, more MSL alleles could be analyzed and the effect of the genes constituting the MSL pathway on expression of a given gene can be unraveled.

Acknowledgments

We thank James Birchler for his support of the initial phase of the work. Mitzi Kuroda provided the different mle flies used in this study. Kathy Welch aided our statistical analyses. This study was supported by a grant from the state of Michigan (MEDC-410).

Footnotes

  • Communicating editor: K. Golic

  • Received March 3, 2003.
  • Accepted June 27, 2003.

LITERATURE CITED

View Abstract