Drosophila stocks:

Unless otherwise cited, D. melanogaster mutations were described previously (Lindsley and Zimm 1992). The P{lacW}ci^Dplac allele (Eaton and Kornberg 1990) is a P{lacZ^P\T.Ww^+mC amp^R ori = lacW} construct inserted at an 8-bp sequence at 79894 (GTCTGTAC) in AE03854.3, which is ∼3 kbp upstream from the ci locus on chromosome 4 (J. Locke, D. Bushey and S. Hanna, unpublished observations). The Harwich wild-type strain has been previously described (Kidwell and Kidwell 1976; Anxolabehere et al. 1988) as an inbred P strain containing many P elements and producing a P cytotype. The Oregon-R strain lacks P elements (M strain) as tested by Southern transfers and does not mobilize P{w⁺} marked transposons. P{ry⁺ SalI}89D is a type I P repressor element that has a frameshift mutation introduced into the SalI site at the beginning of exon 3 (Karess and Rubin 1984; Rio et al. 1986; Gloor et al. 1993). The Su(var)3-7^+t6.5 transgene was provided by P. Spierer and originally described by Reuter et al. (1990). T1 and BX2 have seven P{lacW} inserts in a tandem array. Both P{lacW} arrays were originally generated by Dorer and Henikoff (1994) and provided to us by S. Ronsseray. KP-D(Cy) and KP-U(TM6B) stocks were also provided by S. Ronsseray. Our PCR and Southern blot analysis found that these stocks have four to six KP elements segregating with a Cy Bl vg chromosome for KP-D(Cy) or with a TM6B chromosome for KP-D(TM6). The P{hsp26-pt-T}ci^2-M1021.R and P{hsp26-pt-T}2-M010.R insertions were originally isolated by Sun et al. (2000). The vg^21-3 allele was originally described in Williams et al. (1988) and supplied by R. Hodgetts. All stocks containing w⁺ transgenes were tested in a white⁻ [w^67c23(2)] background to permit the assessment of w⁺ transgene expression.

Cultures were grown at 21° (unless otherwise stated) on standard yeast/cornmeal medium (each liter contains 8 g agar, 18 g Torula yeast, 72 g cornmeal, 96 ml fancy molasses, 2.8 ml propionic acid, and 2.8 g methyl paraben dissolved in 10 ml of 95% ethanol).

Mutagenesis:

Three different screens used gamma irradiation, ethyl methanesulfonate (EMS), or P element dysgenesis to generate mutants that dominantly suppress P-element-dependent silencing [Su(PDS)]. In the first screen, y w; P{lacW}ci^Dplac males were exposed to 40 Gy of γ-radiation and crossed to y w; P{ry⁺ SalI}89D Sb/TM6B; ci¹ ey^R females. We selected for suppressor mutants that caused increased numbers of colored ommatidia in P{ry⁺ SalI}89D Sb progeny, which normally had predominately white eyes with a few colored ommatidia. Flies with a putative suppressor mutation, which allowed w⁺ expression when P{ry⁺ SalI}89D was present, were backcrossed to the original y w; P{ry⁺ SalI}89D Sb/TM6B; ci¹ ey^R stock to confirm that the mutant phenotype transmitted. We also screened for enhancer mutations in the progeny that inherited TM6B, which normally had uniformly colored eyes. Flies with a putative enhancer mutation were backcrossed to y w; P{lacW}ci^Dplac to confirm the mutant phenotype transmitted. Males displaying a mutant phenotype were crossed successively to a w; CyO/Xa and w; TM6B/Ly stocks to map the mutation to chromosome 2 or 3 by testing for segregation from the balancer and to stock the mutant with the appropriate balancer.

In the second screen, y w; dp; e; P{lacW}ci^Dplac males were starved for 8 hr and then placed overnight in a fly vial containing a piece of Whatman filter paper soaked with 600 μl of a solution containing 25 mm EMS, 1% sucrose, and 0.1% bromophenol blue (Ashburner 1989a). Males with blue abdomens from bromophenol blue ingestion were mated to w; Sb P{ry⁺ SalI}89D/TM3, Ser; ci¹ ey^R females. We screened for a suppressed eye phenotype in the Sb progeny and an enhanced eye phenotype in progeny with wild-type bristles. The mutants were stocked using a w; CyO/Bl; Sb P{ry⁺ SalI}89D/TM2 stock.

In the last screen, P{PZ} inserts, P{PZ}05137⁰⁵¹³⁷ and P{PZ}CtBP⁰³⁴⁶³, which flank Su(var)3-7, were mobilized to generate mutations within 87E. To generate mutations, y w; Sb e P{Δ2-3}99B)/P{PZ} ry⁵⁰⁶ (where P{PZ} is either P{PZ}05137⁰⁵¹³⁷ or P{PZ}CtBP⁰³⁴⁶³) males were crossed to y w; P{ry⁺ SalI}89D; P{lacW}ci^Dplac females. Progeny with wild-type bristles and an increased number of colored ommatidia were backcrossed to a y w; P{ry⁺ SalI}89D; P{lacW}ci^Dplac stock to confirm the mutant phenotype transmits and then stocked as described in the second screen.

Genetic mapping:

Su(PDS) mutations on chromosome 2 were mapped by crossing w; mutant/wg^Sp-1 Bl¹ L^rm Bc¹ Pu² Pin^B females to w; P{ry⁺ SalI}89D; P{lacW}ci^Dplac males. The frequencies of recombination events with wg^Sp-1 Bl¹ and L^rm were used to determine the genetic location of the Su(PDS) mutations.

The Su(PDS) phenotype of Su(var)3-7^P9 was mapped by crossing w; Su(var)3-7^P9/Ki¹ rs¹ Ubx¹ H² Dr¹ heterozygous females to y w; P{ry⁺ SalI}89D; P{lac W}ci^Dplac males. Although Dr¹ flies had reduced eye size, the relative amount of red pigment in the eye tissue could still be assessed.

Staining imaginal discs for β-galactosidase activity from P{lacW}ci^Dplac:

We followed the method described by Ashburner (1989b). Imaginal discs were dissected and washed in PBS (137 mm NaCl, 2.7 mm KCl, 10 mm Na₂HPO₄, 1.8 mm KH₂PO₄, pH 7.4) and then fixed in 0.75% glutaraldehyde 1× PBS. The discs were washed in 1× PBT (PBS + 0.05% Triton-X). All discs were stained in X-gal staining buffer (10 mm Na₂HPO₄, 150 mm NaCl, 1 mm MgCl₂, 5 mm potassium ferricyanide, 5 mm potassium ferrocyanide, 0.2% 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside, pH 7.2) over the same 8-hr time period at 37°. Then the discs were washed in PBT and mounted in disc mountant (90% glycerol, 0.1 m Tris-HCl, pH 8) solution. We used the Tubby (Tb) phenotype to distinguish larvae having Su(var)3-7 mutations from progeny that inherited TM6B-Tb, Antp^Hu e¹ Tb. Photographs were taken with a Zeiss Axiophot photomicroscope.

Adult eye photographs:

Photographs of adult male flies that were aged between 5 and 9 days were taken with a Zeiss Axiophot photomicrosope using either a Zeiss 35-mm film camera or a Nikon Coolpix 995 digital camera.

Eye pigment assay:

The extent of w⁺ gene activity was determined by measuring the amount of brown eye pigment. Measurements were made using the method of Ephrussi and Herold (1944) as modified by Locke et al. (1988). The heads were recovered from 5- to 9-day-old adult flies stored at −70°. Three replicate samples containing either 10 or 20 heads, depending on the assay, of each genotype were extracted in 500 μl of acidified methyl alcohol (1% v/v HCl in methanol). After 36–48 hr of continuous agitation to extract the pigments, 2.5 μl of 1% H₂O₂ was add to each tube and then incubated at room temperature for 90–120 min. The absorbance at 470 nm was measured in an LKB Ultraspec II spectrophotometer or a Genova Life Science analyzer.

To measure the amount of pigment produced by In(1)w^m4, w^m4 progeny, mutant male flies were crossed to In(1)w^m4, w^m4; Xa/CyO; Xa/TM2 females. Then males with In(1)w^m4, w^m4; CyO/?; TM2/? were crossed to In(1)w^m4, w^m4 females. If the mutation was on chromosome 2, then Curly (CyO) flies were compared to their counterparts that had wild-type wings and inherited the gene of interest; if the mutation was on chromosome 3, then flies with mutant halteres (TM2, Ubx) were compared to their counterparts that inherited the gene of interest and had wild-type halteres.

Complementation of lethal and maternal effect phenotypes:

Mutant alleles were tested to determine if they complemented recessive lethal or maternal effect phenotypes associated with the 87E region. In each complementation test, at least 50 flies were counted in a balancer class that was expected to occur with frequency equal to that of the class that had the mutant combination. After at least 50 flies in a balancer class were counted, if the heterozygous mutant combination did not occur, the mutant allele combination was considered lethal (−). If the mutant combination was <5% of the balancer class, the combination was labeled semilethal (s). If the combination occurred more frequently than 5%, the combination was considered viable (+).

Mutation detection and DNA sequencing:

Base pair changes in Su(var)205 and Su(var)3-7 mutant alleles were detected using Denaturing HPLC with the WAVE DNA fragment analysis system (Transgenomics, Omaha, NE). Overlapping sections of each gene were amplified by PCR using template DNA from mutant/+ heterozygous flies. Fragments containing heteroduplexes between mutant and wild-type sequences could be detected using this system and then sequenced.

For sequencing and study, Df(2L)TEAa-ll, Su(var)205^P4, and Su(var)205^P5 were balanced over a CyO-pAct-GFP balancer chromosome (Reichhart and Ferrandon 1998). Using the pAct-GFP marker, hemizygotes could be isolated by crossing the Su(var)205/CyO-pAct-GFP stocks to Df(2L)TEAa-II. When the hemizygotes did not develop into larvae, embryos were isolated on the basis of a PCR test for pAct-GFP because the pAct-GFP gene was not expressed at high levels in early embryo development. Su(var)3-7 mutant alleles were sequenced as hemizygotes over Df(3R)126c (B-3009). All mutations were confirmed by reading from both strands.

Genetic screens for dominant mutations that suppress PDS:

When P{ry⁺ SalI}89D is present, the w⁺ gene of P{lacW}ci^Dplac variegates and the eye is mostly white with a few patches of red ommatidia (Figure 1, A and I). We screened for dominant mutations that suppress this P-element-dependent silencing [Su(PDS)]. We scored 4742 progeny from γ-irradiated males and recovered two Su(PDS) mutants on chromosome 3 as well as nine translocations, which presumably suppress PDS by a position effect and are described elsewhere (J. Locke, D. Bushey and S. Hanna, unpublished observations). We also scored 29,601 progeny from the EMS-treated males and isolated two mutations on chromosome 2 and seven mutations on chromosome 3 that suppressed PDS (Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.—

Photographs depicting mutations in Su(var)205 and Su(var)3-7 suppressing w⁺ variegation from P{lacW}ci^Dplac when P{ry⁺ SalI}89D is present. Photographs were taken of male progeny from a cross between y w; P{ry⁺ SalI}(89D); P{lacW}ci^Dplac females and a male with the indicated mutation/balancer. Male progeny that inherited the mutation were compared to progeny that inherited the balancer to determine if the variegation is suppressed. All eye pigment production is dependent on the P{lacW}ci^Dplac chromosome. Alleles tested for a Su(PDS) phenotype are as follows: (A) CyO, (B) Su(var)205^P4, (C) Su(var)205^P5, (D) Su(var)205², (E) Su(var)205⁴, (F) Su(var)205⁵, (G) wg^Sp-1 Bl¹ L^rm Bc¹ Pu² Pin^B, (H) Su(var)205⁵* L^rm Bl¹, (I) TM6B, (J) Su(var)3-7^P9, (K) Su(var)3-7^P25, (L) Su(var)3-7¹⁴, (M) Su(var)3-7⁹, (N) Su(var)3-7^7.1, (O) Su(PDS)P80, and (P) Su(PDS)P86. The CyO and TM6B balancers produce a wild-type phenotype and the eye is predominately white.

In addition to suppressor mutations, the γ and EMS screens also allowed us to search for dominant enhancer mutations that silence w⁺ expression from P{lacW}ci^Dplac in the absence of P elements. No enhancer mutations were recovered from these screens. We did isolate two mutants with completely white eyes, and since they mapped to the P{lacW}ci^Dplac chromosome we assumed them to be null mutants in the w⁺ gene of the P{lacW}ci^Dplac insert. These mutants were not considered further. Although these screens did not isolate any trans-acting enhancer mutants, we did recover two spontaneous P{lacW}ci^Dplac mutants that variegate in the absence of other P elements (see below).

Mutations in Su(var)205 produced a Su(PDS) phenotype:

The two Su(PDS) mutants on chromosome 2 were recessive lethal and failed to complement each other. The Su(PDS) phenotype in one of our mutants [Su(var)205^P4] mapped 8.6 cM to the right of wg^Sp-1 (2-21.9), which is the approximate location of Su(var)205 (2-31.1). Both Su(PDS) mutations also failed to complement the adult recessive lethal phenotype of Su(var)205², Su(var)205⁵, and Df(2L)TEAa-ll, which is deficient for the Su(var)205 locus. Therefore our mutants were alleles of Su(var)205, which encodes HP1, and we called our new alleles Su(var)205^P4 and Su(var)205^P5 (Figure 1, B and C).

The Su(var)205^P4/Df(2L)TEAa-11 hemizygotes developed into third instar larvae but did not survive to become adults. Sequencing the Su(var)205^P4 allele, using template isolated from these larvae, found a nonsense mutation at position Q64*. In contrast, no Su(var)205^P5/Df(2L)TEAa-11 heterozygotes developed to the larval stage. This would be expected if the Su(var)205^P5 allele is a deletion that includes an adjacent gene necessary for embryo development. Molecular analysis confirmed the absence of Su(var)205 because PCR tests, using two different nonoverlapping primer pairs from the 5′ and 3′ gene ends, could not amplify sequence from Su(var)205^P5 homozygotes.

Our recovery of Su(var)205 mutant alleles suggested we test extant alleles at this locus for a Su(PDS) phenotype. Both Df(2L)TEA-11 and Su(var)205² produced a Su(PDS) phenotype (Figure 1D). The Su(PDS) phenotype in Su(var)205² genetically mapped 9.3 cM to the right of wg^Sp-1 [∼31.2 cM (N = 450)], which is coincident with the Su(var)205 locus (Sinclair et al. 1989). However, Su(var)205⁴ and Su(var)205⁵ did not produce a Su(PDS) phenotype (Figure 1, E and F). We examined Su(var)205⁵ further by recombining the chromosome with a homolog marked with wg^Sp-1 Bl¹ L^RM Bc¹ Pu¹ and Pin^B, which does not affect PDS at P{lacW}ci^Dplac (Figure 1G). We isolated recombinant chromosomes that consistently produced a Su(PDS) phenotype (Figure 1H) and sequencing of two recombinants confirmed they retained the Su(var)205⁵ allele. The recombination data were consistent with a single PDS enhancer locus near L (2-98; N = 183) that neutralized the Su(PDS) effect of the Su(var)205⁵ mutation. This putative enhancer was not pursued further.

Chromosome 3 Su(PDS) mutations map to Su(var)3-7 and two other loci:

The alleles P9 and P13 were the first Su(PDS) mutants isolated on chromosome 3 in our initial screen using γ-mutagenesis (Table 1). In addition to the Su(PDS) phenotype, P13 is recessive lethal and P9 has a maternal effect. Homozygous P9 mothers produce homozygous P9 progeny that stop development as first instar larvae as defined by their mouth hooks (Ashburner 1989a). A wild-type chromosome in either mother or progeny allows development to proceed to adulthood. This phenotype is consistent with the P9 mutation occurring in a gene necessary for development and whose product can be supplied either maternally or zygotically. P13 fails to complement this maternal effect because P13/P9 heterozygotes produced by P9/P9 homozygous mothers do not develop into adults.

We genetically mapped the Su(PDS) phenotype of P9 to 3-50.9 (N = 1205) and tested deficiencies in this region for a Su(PDS) phenotype. Deficiencies that include 87E (Df(3R)126c, Df(3R)ry615, and Df(3R)ry27) produced a Su(PDS) phenotype. As well, these deficiencies failed to complement the recessive lethality of P13 and maternal effect of P9. These results are consistent with P13 being a deletion that includes the 87E region and the P9 mutation being a mutation that affects a single gene in 87E.

We mobilized two P-element inserts (P(PZ)05137⁰⁵¹³⁷ and P{PZ}CtBP⁰³⁴⁶³) in the 87E region to find a local insertion that would cause the Su(PDS) phenotype. We screened 19,407 and 18,835 flies generated from males producing transposase and containing either P(PZ)05137⁰⁵¹³⁷ or P{PZ}CtBP⁰³⁴⁶³, respectively. Four new Su(PDS) mutants were isolated, but none of the chromosomes that had these mutations had a P{PZ} insert (Table 1). With the four new Su(PDS) mutations, we had a total of 13 Su(PDS) mutants on chromosome 3.

Su(var)3-7 was a candidate gene in the 87E region because it encoded a heterochromatic modifier. Consequently, we tested three previously isolated Su(var)3-7 mutants produced by site-directed mutagenesis (Seum et al. 2002) and all three failed to complement the P9 maternal effect and caused a Su(PDS) phenotype (Table 1). Su(var)3-7⁹, being a hypomorphic allele, has a weaker Su(PDS) phenotype than Su(var)3-7^7.1 and Su(var)3-7¹⁴, which are reported to be amorphic alleles (Figure 1, K–M). These results confirmed that the P9 and P13 mutations could cause the Su(PDS) phenotype by interrupting Su(var)3-7 expression. Consequently, we adopted the allele designations Su(var)3-7^P9 and Su(var)3-7^P13 for P9 and P13, respectively. As discussed later, molecular characterization found a mutation in Su(var)3-7^P9 confirming this designation.

By crossing the 11 remaining Su(PDS) mutants on chromosome 3 to homozygous Su(var)3-7^P9 females, we determined whether these mutants zygotically complemented the Su(var)3-7^P9 mutation. We expect that only those progeny produced by homozygous Su(var)3-7^P9 females that paternally inherit a functional Su(var)3-7⁺ allele will develop into adults. Of the 11 Su(PDS) mutants, nine Su(PDS) mutations failed to complement Su(var)3-7^P9 zytogically (Table 1, Zygotic). These 9 mutants in addition to Su(var)3-7^P9 and Su(var)3-7^P13 define the complementation group that have mutations in Su(var)3-7.

Of the 11 Su(PDS) mutants affecting Su(var)3-7, six alleles had a mutation within the 87E region that affects only Su(var)3-7 because they were viable over Su(var)3-7^P13 (Table 1, Deficiency). We tested whether these six alleles also had a maternal effect. In our test for a maternal effect, females hemizygous for a particular mutant [mutant/Su(var)3-7^P13] were backcrossed to Su(var)3-7^P13/TM6B males (Table 1, Maternal). If the Su(var)3-7 mutant alleles had a maternal effect, then the hemizygous mothers should not produce adult progeny that lack both a maternally and a zygotically functional Su(var)3-7 allele [Su(var)3-7⁻/Su(var)3-7^P13]. If any Su(var)3-7⁻/Su(var)3-7^P13 progeny survived to become adults, they had to be to due to a maternal contribution because the previous zygotic test demonstrated that any zygotic contribution from these alleles was not enough to allow development to proceed to the adult stage. We observed that all the Su(var)3-7 alleles that failed to act zygotically also had a maternal effect (Table 1, Maternal). In the case of Su(var)3-7^P25, Su(var)3-7^P43, or Su(var)3-7^P71 the maternal effect was incomplete and some of the expected Su(var)3-7⁻/Su(var)3-7^P13 progeny (<5%) developed into adults.

Molecular analysis of our Su(var)3-7 mutants, identified using the genetic evidence described above, confirmed that these alleles had a mutation affecting Su(var)3-7. Southern blot analysis showed that Su(var)3-7^P12 had a deletion within Su(var)3-7 and sequencing identified the missing bases (Table 1). Southern blot and sequencing determined that a Juan element was inserted downstream from the longest identified transcript for Su(var)3-7^P43. Since the insertion at this location was not expected to interrupt Su(var)3-7⁺ function, we searched for an additional sequence change within the Su(var)3-7 coding region.

Using denaturing HPLC to screen ∼600-bp fragments spanning Su(var)3-7, we detected base pair changes in Su(var)3-7^P9, Su(var)3-7^P43, and Su(var)3-7^P49. DNA sequencing of amplified fragments from hemizygotes identified the mutations within these alleles (Table 1). Mutations in Su(var)3-7^P25 and Su(var)3-7^P71 were identified after sequencing the entire coding region in hemizygotes (Table 1). In total, all the putative Su(var)3-7⁻ alleles had a mutation within the Su(var)3-7 gene that changed the amino acid coding sequence (Table 1).

In addition to the Su(var)3-7 complementation group, there were two other complementation groups corresponding to two other alleles: Su(PDS)P80 and Su(PDS)P86 (Table 1). These mutations complemented the Su(var)3-7^P9 maternal effect phenotype and produced a weak Su(PDS) phenotype. The genes affected by these mutations have not been identified and these mutations are not discussed further.

Transgene Su(var)3-7^+t6.5 rescued the Su(PDS) and developmental phenotypes of Su(var)3-7^P9:

We tested the effect of Su(var)3-7⁺ dosage on w⁺ silencing at P{lacW}ci^Dplac by using a Su(var)3-7⁺ transgene. The Su(var)3-7 gene was originally identified by transforming Drosophila with a series of genomic restriction fragments from 87E and selecting fragments that enhanced variegation of In(1)w^m4, w^m4 (Reuter et al. 1990). The smallest construct that enhanced variegation was called P{[ry⁺] Su(var)3-7 6.5 kb}. This 6.5-kb transgenic fragment contains the Su(var)3-7⁺ gene as well as an additional gene, called Ravus, which shares sequence similarity with Su(var)3-7⁺ (Delattre et al. 2002). We attribute the effects of the P{[ry⁺] Su(var)3-7 6.5 kb} construct to the Su(var)3-7⁺ gene within this construct rather than to Ravus, because Ravus does not modify hPEV and, unlike Su(var)3-7, its protein product does not associate with the chromocenter (Delattre et al. 2002). In the following experiments, we used a P{[ry+] Su(var)3-7 6.5 kb} insert on chromosome 2, referred to as Su(var)3-7^+t6.5, to increase the Su(var)3-7⁺ gene dosage (Figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.—

Increasing the Su(var)3-7⁺ gene dose acts to silence w⁺ expression from P{lacW}ci^Dplac when P{ry⁺ SalI}89D is present. (A) Genetic cross to produce the progeny observed in B and C. (B) Pigment accumulation when P{ry⁺ SalI}89D and P{lacW}ci^Dplac is present at 21°. As indicated at the bottom of the graph, Su(var)3-7⁺ dosage increases from 1, 2 (transgene), 2, to 3 doses. Measurements are taken from three samples with 10 male heads each. Error bars represent standard error. (C) Photographs of eyes from flies with increasing Su(var)3-7⁺ dosage. Each row of eyes is collected from flies raised at the rearing temperature indicated. The Su(var)3-7⁺ dosage increases in each column from 1, 2 (transgene), 2, and 3 as indicated in B.

A single cross (Figure 2A) generated progeny with three different (1, 2, and 3) Su(var)3-7⁺ dosages (Figure 2, B and C). When two functional copies of Su(var)3-7⁺ were combined with P{lacW}ci^Dplac and P{ry⁺ SalI}89D, the eyes were predominately white, indicating w⁺ was silenced in most of the ommatidia at 21°. A mutation affecting one of the two Su(var)3-7 copies suppressed this silencing and the eye was predominately red colored. When the number of functional Su(var)3-7⁺ copies returned to two in a Su(var)3-7^P9/+ heterozygote by crossing in the Su(var)3-7^+t6.5 transgene, the w⁺ transgene was silenced and produced a predominately white eye (Figure 2, B and C). Increasing the Su(var)3-7⁺ dosage from two to three copies does not produce a significant difference at 21° because the w⁺ transgene is predominately silenced at two copies. However, increasing the rearing temperature increases the number of red ommatidia. At 29°, increasing the Su(var)3-7⁺ dosage to three copies significantly reduces the number of red ommatidia compared to two doses at the same temperature (Figure 2C). These results indicate that Su(var)3-7⁺ acts dose dependently to silence w⁺ expression from P{lacW}ci^Dplac as it does in other heterochromatic regions.

The Su(var)3-7^+t6.5 transgene also rescued the developmental phenotype of Su(var)3-7^P9 both paternally and maternally. Homozygous Su(var)3-7^P9 progeny, produced by Su(var)3-7^P9 homozygous mothers, develop into adults when they inherited Su(var)3-7^+t6.5 paternally. Progeny that did not inherit Su(var)3-7^+t6.5 did not develop into adults. Also, homozygous Su(var)3-7^P9 mothers that had one copy of Su(var)3-7^+t6.5 produced Su(var)3-7^P9 homozygous offspring that developed into adults that had no functional Su(var)3-7⁺ gene. Without Su(var)3-7^+t6.5, Su(var)3-7^P9 homozygous mothers cannot produce homozygous adult offspring that lack a functional Su(var)3-7⁺ allele.

Mutations in Su(var)3-7 did not interrupt the ability of P{ry⁺ SalI}89D to modify other repressor-sensitive alleles:

The presence of P{ry⁺ SalI}89D can modify the phenotypic expression from other repressor-sensitive alleles, such as vg^21-3 and P-lacZ. This modification may occur because the P repressor protein, which P{ry⁺ SalI}89D produces, interacts directly with the P promoter. On the basis of this model, the mutations in Su(var)3-7 should not interrupt this direct interaction. In regards to vg^21-3, the allele combination vg^21-3/vg^B produces a strong vestigial wing phenotype (Figure 3A), which is suppressed by P{ry⁺ SalI}89D (Figure 3B). We observed that this suppression was not changed by the presence of Su(var)3-7^P9 (Figure 3C). Therefore, mutations in Su(var)3-7 do not interrupt P-element suppression at vg^21-3 and, presumably, other loci by inhibiting P{ry⁺ SalI}89D activity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.—

Mutations in Su(var)3-7 do not interrupt P{ry⁺ SalI}89D suppression of vg^21-3 or repression of P-lacZ. (A) w; vg^21-3/vg^B; Sb P{ry⁺ SalI}89D/TM6B, Ubx. (B) w; vg^21-3/vg^B; Su(var)3-7^P9/TM6B, Ubx. (C) w; vg^21-3/vg^B; Sb P{ry⁺ SalI}89D/Su(var)3-7^P9. (D) Wing imaginal discs were stained for β-galactosidase activity to determine if mutations in Su(var)3-7 affect P-lacZ expression from P{lacW}ci^Dplac. Changes in genotype of the wing imaginal discs are indicated at the top of each column and to the left of each row. All imaginal discs inherited P{lacW}ci^Dplac. The imaginal discs on the left (+) are from progeny of w; P{lacW}ci^Dplac females while the imaginal discs on the right (P{ry⁺ SalI}89D) are from progeny of w; P{ry⁺ SalI}89D; P{lacW}ci^Dplac females. These females were crossed with Su(var)3-7⁺, Su(var)3-7^P9, or Su(var)3-7^P13 males balanced with TM6B, Antp^Hu e Tb. Imaginal discs were collected from Tb+ larvae, which contain the indicated mutation.

In somatic cells, P{ry⁺ SalI}89D reduces β-galactosidase activity when the P-element promoter is fused to the lacZ gene (P-lacZ; Lemaitre and Coen 1991). In the absence of P{ry⁺ SalI}89D, the P-lacZ expression pattern from P{lacW}ci^Dplac mimicked ci expression and the anterior portion of the wing imaginal disc was stained (Figure 3D, left; Schwartz et al. 1995). When P{ry⁺ SalI}89D is present, P-lacZ expression from P{lacW}ci^Dplac was completely repressed and no staining occurred in the wing imaginal disc (Figure 3D, right). Mutations in Su(var)3-7 [Su(var)3-7^P9 or Su(var)3-7^P13] did not interrupt this repression, because P{ry⁺ SalI}89D continued to repress P-lacZ activity when these mutations were present. This indicates the Su(var)3-7⁺ product acts directly at P{lacW}ci^Dplac to silence w⁺ expression and is not necessary to repress expression from another repressor-sensitive allele at the same location. The P-lacZ repression and w⁺ silencing occur through two different mechanisms that are triggered by the presence of the P repressor protein.

Mutations in Su(var)3-7 suppress silencing at P{lacW}ci^Dplac caused by both type I and type II elements:

Thus far, we have shown that mutations in Su(var)3-7 suppress silencing caused by P{ry⁺ SalI}89D, which is a single type I repressor element. We also tested whether mutations in Su(var)3-7 suppress silencing at P{lacW}ci^Dplac caused by chromosomes that have KP elements, which are type II elements. Both KP-D(Cy) and KP-U(TM6B) chromosomes have multiple KP elements and cause w⁺ silencing at P{lacW}ci^Dplac when there are two functional Su(var)3-7⁺ alleles (Figure 4, right). In contrast, when flies inherited Su(var)3-7^P9 and had only one functional Su(var)3-7⁺ allele, w⁺ was expressed predominately throughout the eye when these KP chromosomes were present (Figure 4, left). Mutations in Su(var)3-7 also suppressed silencing at P{lacW}ci^Dplac that occurs in crosses with P strains, such as Harwich and π2 (data not shown). These P strains have autonomous P elements and their derivatives. These results demonstrate that both types of P elements are acting through a common mechanism that is affected by Su(var)3-7⁺ dose to cause silencing at P{lacW}ci^Dplac.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.—

Su(var)3-7^P9 suppresses w⁺ silencing caused by KP elements at P{lacW}ci^Dplac . Genotypes are indicated above each column and at the left of each row. All flies inherited a single P{lacW}ci^Dplac chromosome. Eyes on the left inherited the Su(var)3-7^P9 allele and have one functional copy. Eyes on the right inherited two functional copies. Eyes in the top row inherited KP-U and in the bottom row inherited KP-D. The flies that inherited KP-U chromosome were produced by crossing w; TM6B, Ant^Hu e Tb KP-U/Ly females to w; e/Su(var)3-7^P9; P{lacW}ci^Dplac males. The flies that inherited KP-D chromosome were produced by crossing w; Cy Bl vg KP-D/l(2)NS females to w; TM6B/Su(var)3-7^P9; P{lacW}ci^Dplac males.

Su(var)3-7⁺ dose dependently silences w⁺ expression from P{lacW}ci^Dplac in the absence of other P elements:

When there are two Su(var)3-7⁺ gene copies, w⁺ expression from P{lacW}ci^Dplac produces a uniformly colored eye in the absence of other P elements. Using the Su(var)3-7^+t6.5 transgene, the Su(var)3-7⁺ gene number was increased to three copies and this resulted in w⁺ variegation (Figure 5B). Pigment analysis provided a quantitative measure for w⁺ expression and found a significant decrease in w⁺ expression from P{lacW}ci^Dplac as the Su(var)3-7⁺ dose increased (Figure 5A). This indicates SU(VAR)3-7 acts in this region and the presence of P elements enhances only the heterochromatic silencing. The P elements do not have to produce a product that recruits SU(VAR)3-7 to P{lacW}ci^Dplac because SU(VAR)3-7 already acts at this insert in a dose-dependent manner.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.—

The effect of Su(var)3-7⁺ dosage on pigment production from a variety of w⁺ transgenes in the absence of P elements. (A) The insert tested is indicated at the bottom of each set of bars. The insert P{hsp26-pt-T}2-M010.R is shortened to 2-M010.R. The Su(var)3-7⁺ dosage increases from 1 to 2 (transgene), 2, and 3 doses from left to right for each w⁺ transgene. The average pigment value was obtained from an average of three samples each containing 20 heads from flies reared at 21°. Error bars indicate standard error. Asterisk above data set indicates ANOVA analysis found a significant (P < 0.05) difference between the samples taken at the three different Su(var)3-7⁺ dosages. (B) Pigment production is dependent on the w⁺ transgene indicated to the left. From left to right the Su(var)3-7⁺ dosage increased by 1, 2 (transgene), 2, and 3 as indicated in A above each column.

As controls, we also tested whether Su(var)3-7⁺ gene dosage had an effect on w⁺ expression from two other inserts where PDS did not occur: P{lacW}3-76a and P{hsp26-pt-T}2-M010.R. Both inserts produce a uniformly colored eye, which does not variegate when other P elements are present. P{lacW}3-76a is on the X chromosome and P{hsp26-pt-T}2-M010.R is inserted in a more distal region on chromosome 4 (102B; Sun et al. 2000). When there were three copies of Su(var)3-7⁺ there was no visible effect on w⁺ expression from P{lacW}3-76a while w⁺ expression from P{hsp26-pt-T}2-M010.R produced a variegated eye phenotype (Figure 5B). Therefore, P{lacW}ci^Dplac behaved differently from other w⁺ transgenes in euchromatic regions, such as P{lacW}3-76a, because it was sensitive to Su(var)3-7⁺ dosage. However, this sensitivity was not unique to P{lacW}ci^Dplac because the other insert tested on chromosome 4 was also sensitive to Su(var)3-7⁺ dosage. Hence, other factors as well as the sensitivity to Su(var)3-7⁺ dose determine whether PDS will occur at a P insert.

Chromosomal translocations that moved P{lacW}ci^Dplac to more distal positions from the centromere suppressed PDS (J. Locke, D. Bushey and S. Hanna, unpublished observations). This suppression might act by moving P{lacW}ci^Dplac to a typical euchromatic location where Su(var)3-7⁺ no longer acts dose dependently to repress w⁺ expression. We tested whether increasing the Su(var)3-7⁺ dosage reduced w⁺ expression from these translocations as it did from the original P{lacW}ci^Dplac allele. Eye pigment assays provided a quantitative measure to determine if an increase in Su(var)3-7⁺ dose repressed w⁺ expression. In M strains, increasing the dosage to three copies of Su(var)3-7⁺ resulted in reduced pigment levels in those flies with the translocations [T(2;4)DB47F and T(2;4)DB57A in Figure 5A]. Therefore, according to pigment analysis, these translocations continued to behave like other inserts on chromosome 4 and an increased Su(var)3-7⁺ dosage reduces w⁺ expression.

Although pigment analysis indicated that the translocations were still sensitive to Su(var)3-7⁺ dosage, visual inspection indicated that the increased Su(var)3-7⁺ dosage did not cause w⁺ variegation with the translocations as it did with the original P{lacW}ci^Dplac allele. At three Su(var)3-7⁺ doses, w⁺ expression variegated from P{lacW}ci^Dplac but expression was uniform from the translocations that included P{lacW}ci^Dplac [T(2;4)DB47F and T(2;4)DB57A; Figure 5B]. Our quantitative pigment analysis could not measure the stochastic differences in w⁺ expression level that occurred between the ommatidia. The fact that the translocations did not variegate indicated that a change in chromosome position interfered with the ability of the Su(var)3-7⁺ product to cause w⁺ silencing, which produces the variegated phenotype. According to these results, the Su(var)3-7⁺ product has two activities, one that reduces the level of expression and another that causes stochastic silencing of w⁺ expression. Changing the chromosomal position interfered with the stochastic silencing but not the ability of Su(var)3-7⁺ to reduce w⁺ expression.

Two P{lacW}ci^Dplac derivatives had a w⁺ variegated phenotype in the absence of other P elements:

During our examination of gene silencing at P{lacW}ci^Dplac, we recovered two spontaneous mutants that had a variegated eye phenotype instead of the usual full red eye. We include them here because they indicate that the P{lacW}ci^Dplac transgene is sensitive to silencing in the absence of P elements. Genetic analysis found that in both cases the variegated phenotype of the two alleles segregated with the P{lacW}ci^Dplac chromosome and both stocks lacked P elements (other than the transgene itself). The alleles, called P{lacW}ci^E1 and P{lacW}ci^E2, produced eyes that displayed pigment variegation when heterozygous over a wild-type chromosome 4 (Figure 6). When either P{lacW}ci^E allele was combined with P{ry⁺ SalI}89D, the w⁺ expression was completely silenced as was evident from a completely white eye.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.—

Mutations in Su(var)205 and Su(var)3-7 suppress the variegated P{lacW}ci^E alleles. The P{lacW}ci^E allele being tested is indicated to the left. The Su(var) mutation being tested is indicated at the top. These male flies were generated by crossing w; P{lacW}ci^E females to males with mutant/balancer.

Su(PDS) mutants suppressed the P{lacW}ci^E alleles:

Su(var)205 and Su(var)3-7 mutants suppressed the variegated eye phenotype from the P{lacW}ci^E alleles just as they did in PDS. All the Su(var)205 mutant alleles weakly suppressed the P{lacW}ci^E alleles (Figure 6). In addition, Su(var)3-7^P9 and Su(var)3-7^P13 strongly suppressed the P{lacW}ci^E alleles (Figure 6). These results indicated that the same dose-dependent heterochromatic modifiers that suppress silencing at P{lacW}ci^Dplac in the presence of P elements also suppress variegation of the P{lacW}ci^E alleles. Therefore, the cis sequence changes in the P{lacW}ci^E alleles (see below) appear to trigger equivalent heterochromatic changes as the trans-acting P elements.

A gypsy element is responsible for both P{lacW}ci^E alleles:

Southern blot analysis suggested that both P{lacW}ci^E1 and P{lacW}ci^E2 had an insertion ∼1 kb from P{lacW}ci^Dplac toward RpS3A (data not shown). PCR amplification and DNA sequencing found that these alleles each had a gypsy element insertion (Figure 7). These gypsy inserts occur at slightly different positions and opposite orientations. Since each independent gypsy insertion coincides with the presence of the variegated eye phenotype, we believe they cause the w⁺ variegation in both P{lacW}ci^E alleles.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.—

Diagram of P{lacW}ci^E alleles. P{lacW}ci^E1 has a gypsy element inserted between a 4-bp duplication (TACA) starting at 81,294 (AE003854). P{lacW}ci^E2 has a gypsy element inserted between a 4-bp duplication (TATA) starting at nucleotide at 80,747 (AE003854). Arrows indicate the direction of transcription.

Many gypsy-induced mutations can be suppressed by su(Hw) and enhanced by mod(mdg4) (Gdula et al. 1996). tested the su(Hw) combinations su(Hw)²/su(Hw)³, su(Hw)²/su(Hw)⁵, su(Hw)²/su(Hw)⁸, and su(Hw)³/su(Hw)⁸ and found they either weakly suppressed or did not affect the variegated eye phenotype produced by either P{lacW}ci^E allele (data not shown). Also, the su(Hw)³/su(Hw)⁵ combination did not alter the variegated phenotype of P{lacW}ci^E1. However, su(Hw)²/su(Hw)³, su(Hw)³/su(Hw)⁵, su(Hw)²/su(Hw)⁵, and su(Hw)²/su(Hw)⁸ did suppress the lozenge¹ (lz¹) eye phenotype in males, which was a known su(Hw)-dependent mutation (Modolell et al. 1983). We also tested an allele of modifier of mdg4 [P{PZ}mod(mdg4)⁰³⁵⁸²] and found it weakly enhanced silencing. With either of the su(Hw) or mod(mdg4) mutations, the phenotype was too weak to map genetically and thereby determine whether it segregated with the associated mutation. It was possible that other modifiers on the same chromosome as the mutation tested were responsible for the weak changes to w⁺ expression. The weak and inconsistent suppression of variegation by the su(Hw) combinations indicated that the variegated phenotype produced by the P{lacW}ci^E alleles did not result from the insulator function previously ascribed to gypsy element insertions.

Transvection occurred between w⁺ transgenes inserted in the ci regulatory region:

In the absence of other P elements, P{lacW}ci^Dplac produced a full red eye, but when heterozygous with either P{lacW}ci^E allele, the eye was variegated (Figure 8A). This demonstrated that both P{lacW}ci^E alleles trans-silenced w⁺ expression from a homolog containing the original P{lacW}ci^Dplac allele. We tested whether the P{lacW}ci^E alleles could also trans-silence w⁺ expression from four different P{lacW}ci^Dplac translocations: T(3;4)DB63E, T(3;4)DB62B, T(2;4)DB47F, and T(3;4)DB66C (J. Locke, D. Bushey and S. Hanna, unpublished observations). Except for T(3;4)DB66C, the P{lacW}ci^E alleles did not trans-silence expression from w⁺ in the translocations of P{lacW}ci^Dplac. With T(3;4)DB66C, the trans-silencing effect was weaker than that with the original P{lacW}ci^Dplac allele. Figure 8, B and C displays our results with T(3;4)DB63E, which is a typical example. Together, these results show that the translocations inhibited the trans-silencing and this indicated that either position or pairing dependence contributed to the trans-silencing. As well, we tested whether the presence of the P{lacW}ci^E alleles caused variegation from other P{lacW} inserts on other chromosomes and did not observe any change in pigment production from the other w⁺ transgenes (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.—

Trans-silencing between P{lacW}ci derivatives and other w⁺ inserts in the ci region. (A) The P{lacW}ci^E alleles trans-silence w⁺ expression from P{lacW}ci^Dplac. Pigment produced depended solely on the P{lacW} allele(s) inherited. The abbreviation for each P{lacW} allele is shown at the bottom. Plus indicates a wild-type chromosome 4 that does not have a P{lacW} insertion. The different males were produced from a cross between w; P{lacW}ci^Dplac females and males with a P{lacW}ci^E1, P{lacW}ci^E2, or wild-type chromosome 4. (B) P{lacW}ci^E1 does not trans-silence w⁺ expression from the P{lacW}ci^Dplac translocation, T(3;4)DB63E, as strongly as the original P{lacW}ci^Dplac allele. Flies in the first two columns inherited P{ry⁺ SalI}89D and those in the last two columns inherited TM3, Ser. The translocation is in combination with either ey^D or P{lacW}ci^E1 as indicated. Eyes are from male siblings in a cross between w; T(3;4)DB63E/TM6B males and w; Sb P{ry⁺ SalI}89D/TM3; P{lacW}ci^E1/ey^D females. (C) P{lacW}ci^E2 does not trans-silence expression from the P{lacW}ci^Dplac translocation, T(3;4)DB63E, as strongly as the original P{lacW}ci^Dplac allele. Genotypes and labeling are the same as in B but P{lacW}ci^E2 is tested rather than P{lacW}ci^E1. Eyes are from male siblings from a cross between w; T(3;4)DB63E/TM6B males and w; Sb P{ry⁺ SalI}89D/TM3; P{lacW}ci^E2/ey^D females. (D) Trans-silencing between P{lacW}ci^2-M1021.R and P{lacW}ci alleles with and without P{ry⁺ SalI}89D present. Flies in the first two columns inherited P{ry⁺ SalI}89D and flies in the last two columns inherited TM3, Ser. The chromosome 4 combinations are indicated at the top. The inherited P{lacW}ci allele is indicated to the left. Eyes shown in each row are from siblings produced from a cross between w; Sb P{ry⁺ SalI}89D/TM3; P{lacW}ci^2-M1021.R/M^57g females and w; P{lacW}ci allele/ey^D, where the P{lacW}ci allele is P{lacW}ci^Dplac, P{lacW}ci^E1, or P{lacW}ci^E2 as listed to the left.

Combining PDS and the trans-silencing effects caused stronger w⁺ silencing than did their individual effects. By itself, P{ry⁺ SalI}89D caused only weak variegation of the hsp70-w⁺ transgene in P{hsp26-pt-T}ci^2-M1021.R (Figure 8D, middle left; J. Locke, D. Bushey and S. Hanna, unpublished observations). Also, individually the P{lacW}ci^E alleles weakly trans-silenced the w⁺ (P{hsp26-pt-T}ci^2-M1021.R) transgene (Figure 8D, middle right). However, when both P{ry⁺ SalI}89D and one of the P{lacW}ci^E alleles were present, the w⁺ (P{hsp26-pt-T}ci^2-M1021.R) transgene was silenced in many ommatidia (Figure 8D, left). Since neither P{ry⁺ SalI}89D nor the P{lacW}ci^E alleles individually caused strong variegation, the predominant silencing resulted from the combination of trans-silencing and PDS.

The same combined effect occurred with the four translocations of P{lacW}ci^Dplac we tested previously for trans-silencing. In each case, the presence of the P{lacW}ci^E alleles or P{ry⁺ SalI}89D sometimes caused weak w⁺ variegation depending on the translocation but in combination more ommatidia were white (Figure 8, B and C).

Su(PDS) mutants suppressed In(1)w^m4, w^m4 variegation:

If our Su(PDS) mutants were deficient for the same function as Su(var) alleles, then they should also suppress hPEV. Our Su(PDS) mutants were crossed to an In(1)w^m4, w^m4 stock to determine if they suppressed w^m4 variegation. The Su(var)205 mutants consistently suppressed the w^m4 variegated eye phenotype, but the Su(var)3-7 mutants did not consistently suppress silencing compared to controls. Pigment analysis of our Su(PDS)-derived mutants provided a quantitative comparison and showed that both Su(var)205 mutant alleles strongly suppressed w^m4 variegation (Figure 9). However, our Su(var)3-7 mutant alleles were inconsistent in their ability to suppress variegation (Figure 9). Both Su(var)3-7^P9 and Su(var)3-7^P12 suppressed w^m4 variegation, while Su(var)3-7^P25, Su(var)3-7^P43, Su(var)3-7^P47, and Su(var)3-7^P71 did not suppress variegation. We noted that Df(3R)126c and Df(3R)ry615, which deleted Su(var)3-7(87E), also did not produce a Su(var) phenotype. Since deficiencies that remove 87E are reported to be strong suppressors of w^m4 variegation (Reuter et al. 1987), we assume that other antagonistic PEV modifiers within our stocks are masking the Su(var) phenotype that should be produced by the mutations in Su(var)3-7.

• Download figure
• Open in new tab
• Download powerpoint

Figure 9.—

Pigment analysis testing whether the Su(PDS) mutants suppress In(1)w^m4 variegation in males. The crosses compare pigment production between siblings that inherit a mutation against those that inherit a CyO or TM2, Ubx balancer chromosome (solid bars). As a control, these balancer chromosomes (black bar) were compared to wild-type (Oregon-R) chromosomes (shaded bar) in the TM2, Ubx and CyO labeled columns. The balancer chromosomes did not cause a substantial increase in pigment production. The other pairs of columns compare a mutant (shaded bar) to balancer (solid bar). Measurements are the average of three samples each containing 20 heads. Error bars show the standard error between the three samples. Asterisk indicates the mean value is significantly different from that of the siblings that inherit the balancer (P < 0.05) according to a t-test.

w⁺ variegation from the P{lacW} arrays were sensitive to the same heterochromatic modifiers that suppress PDS:

In addition to P{lacW}ci^Dplac, P elements also enhance variegation from the P{lacW} arrays, BX2 and T1 (Josse et al. 2002). In the P{lacW} arrays, expression from the w⁺ transgenes variegates, and mutations in Su(var)205 suppressed this variegation (Dorer and Henikoff 1994, 1997). Therefore, the same factors that modified variegation at these P{lacW} arrays also modified expression from P{lacW}ci^Dplac. We wanted to extend this list and test whether the P{lacW} arrays were sensitive to Su(var)3-7⁺ dosage. We found that increasing the Su(var)3-7⁺ dosage strongly enhanced variegation from the BX2 and T1 arrays (Figure 10A). Therefore, silencing at the P{lacW} tandem repeats involves a similar set of chromatin proteins that cause silencing at P{lacW}ci^Dplac . These common chromatin proteins could explain why P elements enhance silencing at both loci.

• Download figure
• Open in new tab
• Download powerpoint

Figure 10.—

Transgene arrays are also sensitive to Su(var)3-7 dosage and insensitive to Su(var)3-9 dosage. (A) Increasing the Su(var)3-7⁺ dose enhances variegation from T1 and BX2. Pigment production is dependent on either T1 or BX2 as indicated to the left. The Su(var)3-7⁺ dosage increased from 1, 2 (transgene), 2, and 3 doses as indicated at the top. The male progeny depicted were generated by crossing w; Bl/Su(var)3-7^+t6.5; Su(var)3-7^P9/TM2, Ubx males to females that had the indicated P{lacW} array. (B) Testing whether Su(var)3-9 mutations suppress variegation from w⁺ transgenes that are affected by PDS. The transgene tested is indicated at the left and the Su(var)3-9 allele is indicated at the top. The male progeny depicted were generated by crossing males with the indicated Su(var)3-9 allele to females with the indicated w⁺ reporter gene(s).

Our Su(PDS) screen identified only a subset of the many Su(var) loci. To confirm that other Su(var) modifier loci did not suppress w⁺ variegation from P{lacW}ci and the P{lacW} repeats, we tested Su(var)3-9 mutants. We chose Su(var)3-9 because it strongly suppressed variegation of In(1)w^m4, w^m4; it was frequently found in screens for heterochromatic modifiers; and its protein associates with the chromocenter like SU(VAR)3-7 and HP1 (Schotta et al. 2003). Su(var)3-9¹ and Su(var)3-9² weakly suppressed variegation at T1 and BX2 when compared to their wild-type siblings (Figure 10B). Neither Su(var)3-9 allele modified w⁺ variegation from the P{lacW}ci^Dplac when P{ry⁺ SalI}89D or the enhanced P{lacW}ci^E alleles were present. These results confirmed that a subset of Su(var) genes dose dependently suppressed variegation at loci where PDS occurs. These P{lacW}ci alleles may appear insensitive to other Su(var) mutations because the dosages tested were not limiting their activity in this region or because their product did not act within this region to cause silencing.

A subset of Su(var) loci silence w⁺ expression from P{lacW}ci:

In contrast to genetic screens that have identified ∼150 Su(var) and E(var) modifiers of w^m4 (Wallrath 1998; Schotta et al. 2003), our genetic screen repeatedly identified mutations in four loci that suppress PDS at P{lacW}ci^Dplac. Of the four loci, two had mutations in Su(var)205 and Su(var)3-7 while the other two were not identified. Both in vivo and in vitro evidence indicate that the SU(VAR)3-7 and the HP1 protein product of Su(var)205 directly interact and are part of a complex that causes gene silencing (Cleard et al. 1997; Delattre et al. 2000).

Similar evidence indicates that HP1 also interacts with SU(VAR)3-9 (Hwang et al. 2001) and mutations in Su(var)3-9 suppress variegation that occurs at transgenes inserted in telomeric and centromeric positions (Donaldson et al. 2002). However, PDS at P{lacW}ci is insensitive to mutations in Su(var)3-9. This insensitivity to Su(var)3-9⁺ dose probably results from the composition of chromatin structure on chromosome 4. The Su(var)3-9⁺ product is a histone methyltransferase that is necessary for appropriate histone methylation at the chromocenter (Schotta et al. 2002). However, mutations in Su(var)3-9 do not disrupt the chromosome 4 methylation pattern (Schotta et al. 2002) or HP1 binding to chromosome 4 (Greil et al. 2003).

Mutations in Su(var)205 and Su(var)3-7 show a contrast in their ability to suppress variegation in the w^m4 and P{lacW}ci^Dplac assays. Both previously isolated Su(var)205 mutations and those isolated in our screen weakly suppress PDS at P{lacW}ci^Dplac but strongly suppressed w^m4 variegation (Eissenberg et al. 1990). In contrast, Su(var)3-7 has the opposite effect in that these mutations have a strong Su(PDS) phenotype, but weakly suppressed w^m4 variegation in our tests. Although previous assays using w^m4 have isolated mutations affecting the Su(var)3-7⁺ locus (Reuter et al. 1990), these screens did not isolate single base pair mutations within Su(var)3-7. Targeted mutagenesis of Su(var)3-7 has been used to isolate mutations within this gene (Seum et al. 2002). In our P{lacW}ci^Dplac-based screen, which used EMS to generate random mutations, single base pair changes in Su(var)3-7 were the most common mutation. Screens using w^m4 have identified many mutations in Su(var)205, which produces a very strong Su(var) phenotype (Eissenberg et al. 1990). Other assays studying heterochromatic silencing at other loci have also preferentially isolated different heterochromatic modifiers (Cryderman et al. 1999; Balasov 2002; Donaldson et al. 2002). This bias presumably reflects the different components of the heterochromatic structure found at each variegating locus.

Our mutations in Su(var)3-7 offer further insight into the protein domains needed for function. These mutations are point mutations in the zinc finger consensus sequences, deletions resulting in the production of a truncated product or a deficiency for the entire gene. SU(VAR)3-7 has seven zinc fingers. In vitro experiments have demonstrated that proteins containing at least two of the seven zinc fingers from the N-terminal half of SU(VAR)3-7 can bind DNA (Cleard and Spierer 2001). However, we found that point mutations affecting either zinc finger 4 or zinc finger 5 [Su(var)3-7^P25, Su(var)3-7^P43, and Su(var)3-7^P71] strongly suppress PDS. Therefore these zinc fingers are necessary for SU(VAR)3-7-dependent silencing in vivo. We also found a mutation that affected only the C-terminal region and not the zinc fingers. Su(var)3-7^P12 has a deletion that removes the BESS motif, which is necessary for SU(VAR)3-7 self-association (Jaquet et al. 2002). Therefore self-association, as well as zinc fingers 4 and 5, is necessary for SU(VAR)3-7 to silence w⁺ from P{lacW}ci^Dplac. The mutations we recovered support the conclusion that both the N terminus and the C terminus must be present for SU(VAR)3-7 to enhance heterochromatic silencing (Jaquet et al. 2002).

Mutations in Su(var)3-7 are not recessive lethal, but do cause a recessive maternal effect that stops development during the larval stage (Seum et al. 2002). Su(var)3-7 alleles can be tested for the ability to zygotically rescue this phenotype or maternally produce it. None of our Su(var)3-7 mutants could zygotically rescue the maternal effect caused by homozygous Su(var)3-7^P9 females. Even hypomorphic mutants, such as Su(var)3-7⁹ (Seum et al. 2002), fail to zygotically rescue the maternal effect. Thus, testing for zygotic rescue using Su(var)3-7^P9 provided a sensitive test to identify mutations in Su(var)3-7. The fact that the Su(var)3-7^+t6.5 transgene could rescue Su(var)3-7^P9 maternally and zygotically confirms that mutations within Su(var)3-7 cause this phenotype. Although these mutations interrupt zygotic activity, alleles with mutations in one of the zinc fingers, such as Su(var)3-7^P25, Su(var)3-7^P43, and Su(var)3-7^P71, have a weak maternal effect and hemizygous mothers [Su(var)3-7⁻/Deficiency] produced Su(var)3-7⁻ progeny [Su(var)3-7⁻/Deficiency] that developed into adults at a reduced frequency. In comparison, Su(var)3-7⁻ alleles encoding truncated proteins, such as Su(var)3-7^P9 and Su(var)3-7^P49, have a strong maternal effect and hemizygous mothers never produce Su(var)3-7⁻ adult progeny.

P elements enhance heterochromatic silencing at P{lacW}ci:

The simplest explanation as to why P elements trigger w⁺ silencing only at P{lacW}ci^Dplac is that there is a local interaction between the P repressor protein, which binds to sequences in the P insert, and the chromatin structure at this location. There is a predisposition for heterochromatin to form and silence w⁺ expression at this location. This predisposition is evident because heterochromatic silencing at this location can occur in the absence of P elements under two different circumstances. Increasing the Su(var)3-7⁺ dosage to three copies causes w⁺ silencing at P{lacW}ci^Dplac in the absence of other P elements. As well, we isolated local gypsy insertions that coincide with w⁺ variegation. This suggests that a heterochromatin complex already acts in this region and the P repressor is not causing its de novo formation.

Two factors may be necessary for heterochromatin formation in this region (J. Locke, D. Bushey and S. Hanna, unpublished observations). First, the region is AT rich and has repetitive sequence. This repetitive sequence may act analogously to the repeats found in a P{lacW} array that are necessary for heterochromatic silencing within these arrays. Second, proximity to the centromere is necessary for PDS to occur (J. Locke, D. Bushey and S. Hanna, unpublished observations). Proximity to heterochromatic regions enhances silencing within such regions by possibly sequestering such regions in one compartment in the nucleus (Csink and Henikoff 1996; Dernburg et al. 1996). The heterochromatin formed at P{lacW}ci^Dplac may be stabilized by the constitutive heterochromatin at the centromere. This could explain why w⁺ expression variegates from P{lacW}ci^Dplac at three Su(var)3-7⁺ doses, but w⁺ expression from translocations of P{lacW}ci^Dplac remains uniform at the same dose. The translocations may not be silenced because moving the P{lacW}ci^Dplac to a different genomic location prevents interactions with the centromere that stabilize heterochromatin formation.

By itself, sensitivity to Su(var)3-7⁺ dosage is not sufficient to cause PDS. An increased Su(var)3-7⁺ dosage causes variegation at both P{lacW}ci^Dplac and P{hsp26-pt-T}2-M010.R. However, P elements do not cause PDS at P{hsp26-pt-T}2-M010.R, which is inserted at 102B. P elements can silence P{hsp26-pt-T} inserts, because at P{hsp26-pt-T}ci^2-M1021.R, which is 140 bp away from P{lacW}ci^Dplac at 101F, w⁺ expression variegates when other P elements are present.

Sun et al. (2000) compared the chromosome structure at P{hsp26-pt-T}2-M010.R and P{hsp26-pt-T}ci^2-M1021.R. Both of these inserts have been previously defined as occurring in a euchromatic-like region, because they uniformly express w⁺ throughout the eye in the absence of other P elements (Sun et al. 2000). However, P{hsp26-pt-T}ci^2-M1021.R behaves more like an insert in a heterochromatic region because it is less accessible to endonuclease digestion and heat-shock induction (Sun et al. 2000). Therefore, the insert location of P{hsp26-pt-T}ci^2-M1021.R could be considered to be heterochromatic despite being uniformly expressed. This contrasts with P{hsp26-pt-T}2-M010.R, which is accessible to endonuclease digestion and heat-shock induction, as are the other P{hsp26-pt-T} inserts on chromosome 4 that do not variegate. This difference in chromatin structure, as well as the distance to the centromere, may explain why PDS occurs at P{hsp26-pt-T}ci^2-M1021.R but not at the w⁺ transgenes inserted at other locations, such as P{hsp26-pt-T}2-M010.R.

Do P elements encode a chromatin modifier protein?

Since P strains do not modify expression from w^m4, P elements are not considered to encode general modifiers of hPEV. However, P elements are precipitating heterochromatic silencing at P{lacW}ci^Dplac and, traditionally, only increasing Su(var) gene dosage or changing chromosomal position could trigger such an event. This scenario is complicated by the fact that although the P repressor protein causes silencing at P{lacW}ci^Dplac, it appears to increase transcription from P transgenes inserted at other locations (Lemaitre and Coen 1991; Josse et al. 2002; J. Locke, D. Bushey and S. Hanna, unpublished observations). Other heterochromatic modifiers, such as Su(var)205, can also reduce or increase expression depending on the location of the variegating locus (Cryderman et al. 1999; Balasov 2002; Piacentini et al. 2003).

Thus far, P elements have been shown to modify expression only from w⁺ transgenes inserted within P constructs. This means that a sequence within the P element may be necessary for the P elements to affect expression in trans. This differs from classic heterochromatic modifiers that act on many adjacent loci along a chromosome and on different rearrangements without requiring a specific sequence. Although we have not tested this directly, we expect that P elements would not cause silencing of a w⁺ transgene inserted in this region that lacked flanking P-element sequence. Furthermore, we have not formally tested whether P elements act dose dependently to cause silencing at P{lacW}ci^Dplac. Heterochromatic modifiers act dose dependently to silence expression from heterochromatic regions. Whether or not the P repressor protein should be classed with other heterochromatic modifiers depends on further biochemical and genetic characterization. However, mutations within other transposable elements, such as nomad, can modify variegation from w^m4. A mutation within a single nomad insertion, which reduces transcript expression, enhances heterochromatic silencing at w^m4 and bw^VDe2 (Whalen et al. 2003). Therefore, transposable elements in general could encode factors that modify heterochromatin structure.

Thus far, we have assumed that the P repressor protein is the trans-acting factor produced by P elements that causes silencing at P{lacW}ci^Dplac. However, we have not rigorously tested whether mutations within the coding sequence of the P repressor affect its ability to cause silencing.

A model for chromatin structural changes at P inserts:

We have provided evidence that both P elements and changes in cis-DNA sequence can enhance heterochromatic silencing at single P inserts in somatic cells. These interactions imply that there is an evolutionary relationship between the formation of heterochromatin and transposable element biology. Heterochromatin may have evolved as a means to prevent transposable element mobilization and minimize the detrimental effects caused by repetitive sequences (Dimitri and Junakovic 1999).

In P-element biology, heterochromatin formation at P inserts in the germ line is linked to the production of P cytotype (Ronsseray et al. 2003). P inserts that produce P cytotype-like effects occur either at the telomeres or in the P{lacW} repeats where heterochromatin formation occurs (Ronsseray et al. 1998; Stuart et al. 2002). Thus far, a P insert in pericentromeric heterochromatin has not been identified that produces P cytotype-like effects (Ronsseray et al. 1998). Since heterochromatin forms at P{lacW}ci^Dplac in somatic tissues, this may suggest that P inserts in this region could produce P cytotype-like effects if this heterochromatin also forms in the germ line. However, we have not tested this possibility. Nevertheless, the unique interactions that enhance heterochromatin formation at the P{lacW}ci^Dplac insert, which occurs near centromeric heterochromatin, may also occur at other locations in the genome and in the germ line. Such interactions do occur in the P{lacW} arrays because P elements enhance silencing from the transgene repeats (Josse et al. 2002) and transvection occurs between the arrays and a transgene inserted at a homologous position (Dorer and Henikoff 1997).

At P{lacW}ci^Dplac, both P elements in trans and transvection from the P{lacW}ci^E alleles enhance heterochromatic silencing. If such interactions exist in the germ line, then this would predict P elements and transvection could contribute indirectly to P cytotype by causing heterochromatin formation at P inserts. Experiments with P cytotype support this statement, because they have shown that the presence of paternally inherited P elements enhances the ability of P constructs inserted in telomeric regions to produce P cytotype-like effects (Ronsseray et al. 1998). These paternally inherited P elements cannot reproduce P cytotype-like effects by themselves. However, by contributing P repressor and truncated derivatives, they could enhance heterochromatin formation at certain maternally derived P elements. In addition, transvection between homozygous P inserts should result in a stronger P cytotype because such interactions would enhance heterochromatin formation. Indeed, this appears to occur because passing P inserts through heterozygous females attenuates their ability to produce P cytotype compared to their homozygous counterparts (Niemi et al. 2004).