While heterochromatic gene silencing in cis is often accompanied by nucleosomal compaction, characteristic histone modifications, and recruitment of heterochromatin proteins, little is known concerning genes silenced by heterochromatin in trans. An insertion of heterochromatic satellite DNA in the euchromatic brown (bw) gene of Drosophila melanogaster results in bw^Dominant (bw^D), which can inactivate loci on the homolog by relocation near the centric heterochromatin (trans-inactivation). Nucleosomal compaction was found to accompany trans-inactivation, but stereotypical heterochromatic histone modifications were mostly absent on silenced reporter genes. HP1 was enriched on trans-inactivated reporter constructs and this enrichment was more pronounced on adult chromatin than on larval chromatin. Interestingly, this HP1 enrichment in trans was unaccompanied by an increase in the ²MeH3K9 mark, which is generally thought to be the docking site for HP1 in heterochromatin. However, a substantial increase in the ²MeH3K9 mark was found on or near the bw^D satellite insertion in cis, but did not spread further. These observations suggest that the interaction of HP1 with chromatin in cis is fundamentally different from that in trans. Our molecular data agree well with the differential phenotypic effect on bw^D trans-inactivation of various genes known to be involved in histone modification and cis gene silencing.

---

CHROMATIN structure can influence transcriptional activity by mediating the accessibility of regulatory factors and polymerase to the gene and surrounding DNA. Active genes have a more open or euchromatic structure, while inactive loci have a more condensed nucleosome arrangement that shares many features with constitutively heterochromatic regions of the genome. Constitutive heterochromatin can modify the structure and activity of a euchromatic gene when repositioned next to it by a chromosomal-break event. This change, called position-effect variegation (PEV), was first observed in Drosophila upon isolation of chromosomal a rearrangement that moved the white gene close to the pericentric heterochromatin of the X chromosome and gave a mosaic eye phenotype. Subsequent studies of PEV and related phenomena in other organisms have led to the view that the silencing is due to the progression of heterochromatin along the chromosome to inactivate genes on the same DNA molecule (in cis) (reviewed in WEILER and WAKIMOTO 1995; TALBERT and HENIKOFF 2006).

Studies on cis heterochromatic gene silencing have found that alteration of chromatin structure accompanies silencing by juxtaposed heterochromatin. For example, the regularly spaced, compact array of nucleosomes associated with heterochromatin is also seen over the inactivated gene (WALLRATH and ELGIN 1995). Additionally, mispositioned euchromatin acquires typically heterochromatic histone modification marks. These post-translational modifications of histone tails on their own, or as a code, can recruit a variety of regulatory factors that further specify the chromatin environment (reviewed in MARGUERON et al. 2005). In heterochromatin, histone H3 is often methylated at lysine 9 (H3K9). This methylated residue is a docking site for the nonhistone chromosomal protein HP1 which, among other functions, recruits the H3K9 methyl transferase to perpetuate the methylated state to neighboring nucleosomes, bringing about the linear spreading of heterochromatic structure in cis (BANNISTER et al. 2001; SCHOTTA et al. 2002). One important feature of this spreading model is that the nucleating block of heterochromatin and the gene to be silenced are present on the same DNA molecule.

An intriguing variation on the above-described classic cis-PEV in Drosophila is the dominant PEV of the brown locus, another eye-color gene located in the distal euchromatin of chromosome 2. One of these dominant PEV alleles of bw is bw^D, which has 1.6 Mb of simple satellite heterochromatin inserted in its coding sequence (PLATERO et al. 1998). Expression of the bw gene in bw⁺/bw^D flies is almost totally silenced, with only a few ommatidia showing any expression of the gene. Thus, in addition to disrupting the coding region in cis, the mutant bw^D heterochromatic insertion also inactivates the bw⁺ gene on the opposite chromosome in trans, a phenomena referred to as trans-inactivation or trans-PEV. The linear spread of heterochromatin cannot occur for trans-inactivated genes at the molecular level because the source of heterochromatin and the gene to be inactivated are not present on the same chromosome. Instead, it has been shown that the silencing of the bw⁺ gene is promoted by the mislocalization of the distal euchromatin containing the bw loci to the region of the nucleus containing the centric heterochromatin and a high concentration of the associated binding proteins (CSINK and HENIKOFF 1996; DERNBURG et al. 1996; HARMON and SEDAT 2005). This association of bw^D with the heterochromatic neighborhood is brought about by the tendency of heterochromatin to self-associate and the somatic pairing of the bw⁺- and bw^D-carrying homologs (SAGE and CSINK 2003). As linear spreading is impossible in this situation, how is silencing brought about? One possibility is that nucleosome remodeling factors, histone modifying enzymes, and silencing proteins from the bw^D heterochromatic insert diffuse or reach over to the opposite homolog and bring about inactivation by changing the chromatin structure. This would be aided by location of the chromosomal regions in a heterochromatic compartment rich in silencing proteins and deficient in transcription-promoting factors. Alternatively, trans-silencing might not involve standard chromatin structure changes and gene inactivation might be merely a result of the absence of positive regulatory factors due to the presence of the heterochromatin-associated chromosomes in a specific nuclear compartment.

To test these two possibilities, we analyzed the chromatin structure of transgenes silenced by bw^D in trans and compared it to the structure of transgenes silenced in cis and to endogenous sequences in cis to the bw^D insertion. Transgenes provided unique sequences that were analyzed using nuclease hypersensitivity assays and chromatin immunoprecipitation assays (ChIP) to determine if the chromatin structure of the trans-inactivated genes was altered by the influence of bw^D. While nucleosomal arrangement of one transgene sequence became more compact opposite bw^D, all the histone modifications typically found in heterochromatin (decrease in H3 acetylation and enrichment of methylation on H3K9) were not found to be associated with the transgene in adult or larval cells. A second, different transgene whose reporter is more strongly silenced by bw^D in trans was also analyzed using ChIP, and only a slight decrease in acetylation in adults was found. This lack of change in histone modifications contrasted strongly with an increase in HP1 levels seen on the transgene opposite bw^D. However, this HP1 enrichment was seen mostly in chromatin from adults and was less pronounced in chromatin from larval cells, pointing to the alteration of chromatin structure over development. We also compared the influence of bw^D on a transgenic sequence in trans to its influence on endogenous sequences in cis and found that HP1 spreads linearly in cis from bw^D up to 12 kbp away. While decreased acetylation was found to accompany this HP1 spreading, increased methylation was found only closest to the insert. The moderate changes in HP1 and AcH3 levels in trans show corresponding mild phenotypic suppression by Su(var)205 (the HP1 gene) and Rpd3 (a histone deacetylase gene), respectively, while Su(var)3-9 (a H3K9 methyltransferase gene) does not suppress trans-inactivation at all. Su(var)3-7 is a very strong suppressor of bw^D silencing, and an extra dose of the gene results in additional nucleosomal compaction over bw^D. Taken together, our results indicate that an altered chromatin structure accompanies trans-inactivation, but this change consists mostly of nucleosome compaction and increased presence of HP1, not changes in typical heterochromatic histone modification marks. It also appears that a different set of genes might play key roles in trans-inactivation compared to cis-PEV, which could indicate underlying mechanistic differences between the two kinds of silencing.

Fly culture and material collection:

Flies were grown on standard yeast–cornmeal–molasses medium at 25° unless otherwise stated. The w^–;P{hsp26-Pt}39C-2 (39C-2) line used as a positive control in the nucleosomal compaction and P{hsp26-Pt} ChIP experiments was obtained from L. Wallrath (University of Iowa). w^–;E(bw^D)40/CyO was provided by S. Henikoff (Fred Hutchinson Cancer Research Center, Seattle). P{(ry⁺)Su(var)3-7^+t6.5}T21a/SM6a [extra copy of Su(var)3-7) and Su(var)3-7¹⁴ was provided by P. Spierer (University of Geneva)]. Rpd3³²⁶ [formerly called Su(var)326 or HDAC1³²⁶] was provided by T. Grigliatti (University of British Columbia). All other lines were created in our lab or obtained from the Bloomington Stock Center.

Adult fly collection:

Flies of the required genotype and sex were collected and aged for 3–6 days, weighed, snap frozen, and stored at –80°. For ChIP studies, to obtain w¹¹¹⁸;P{hsp26-Pt}ab28/+ (ab28/+), males of w¹¹¹⁸;P{hsp26-Pt} ab28/CyO (CSINK et al. 2002) were crossed to w¹¹¹⁸ females and straight-winged males and females were collected. For w¹¹¹⁸;P{hsp26-Pt}ab28/bw^D (ab28/bw^D), males of w¹¹¹⁸;P{hsp26-Pt} ab28/CyO were crossed to w¹¹¹⁸; bw^D females and straight-winged males and females were collected. For w^–;P{hsp26-Pt}ab28/E(bw^D)40 [ab28/E(bw^D)40], males of w¹¹¹⁸;P{hsp26-Pt} ab28/CyO were crossed to w^–;E(bw^D)40/CyO females and straight-winged males and females were collected. The w^–;P{hsp26-Pt}39C-2 flies were raised and collected without further crossing.

To obtain w¹¹¹⁸;P{lacW}chrw/+ (chrw/+) flies, males of w¹¹¹⁸;P{lacW}chrw/CyO (CSINK et al. 2002) were crossed to females of w¹¹¹⁸ and straight-winged males and females were collected. For w¹¹¹⁸;P{lacW}chrw/bw^D (chrw/bw^D) flies, males of w¹¹¹⁸;P{lacW}chrw/CyO were crossed to females of w¹¹¹⁸; bw^D and straight-winged males and females were collected.

To study the effect of PEV modifiers on bw^D silencing, v;bw^D virgins were crossed with Su(var)/Balancer and nonbalancer males were examined. The effect of Su(var)205⁵, Su(var)3-7¹⁴, Su(var) 3-9¹, and Rpd3³²⁶ were compared to male progeny from v;bw^D (female) x Canton-S (male).

For nucleosomal arrangement analysis in the presence of an extra dose of Su(var)3-7, (P{(ry⁺)Su(var)3-7^+t6.5}T21a (referred to in the text as T21a), w⁺/w¹¹¹⁸;T21a,bw^D/CyO males were crossed with w^–;P{hsp26-Pt}ab28/CyO virgin females and w^–;P{hsp26-Pt}ab28/T21a,bw^D (T21a,bw^D) males were collected. For control ab28/bw^D male flies, females from w^–;P{hsp26-Pt}ab28/CyO were crossed with w¹¹¹⁸;bw^D males and w^–;P{hsp26-Pt}ab28/bw^D males were collected. All flies were grown and aged at 18°.

Larval brain and imaginal disc collection:

Larval material was obtained by dissecting the males of the specified genotype. To collect ab28/+ and ab28/bw^D, w^–;P{hsp26-Pt}ab28/BcElp males were crossed to either w¹¹¹⁸ or w¹¹¹⁸;bw^D virgins and non-black-cell male larvae were dissected. To collect w^–; ab28/E(bw^D)40 males, w^–;E(bw^D)40/BcElp males were crossed to w^–;P{hsp26-Pt}ab28/CyO,P{GFP} virgins and third instar male larvae were selected against black cell and GFP markers. 39C-2 flies were grown and third instar larval males were dissected.

For chrw/+ and chrw/bw^D larvae, w¹¹¹⁸;P{lacW}chrw/BcElp males were crossed to either w¹¹¹⁸ or w¹¹¹⁸; bw^D females to get non-black-cell chrw/+ and chrw/bw^D male larvae, respectively. Similarly for w¹¹¹⁸;P{lacW}csc2/+ (csc2/+)and w¹¹¹⁸; P{lacW}csc2/bw^D (csc2/bw^D) larvae (SAGE et al. 2005), w^–;P{lacW}csc2/BcElp males were crossed to w¹¹¹⁸ or w¹¹¹⁸; bw^D females to get non-black-cell csc2/+ and csc2/bw^D male larvae, respectively.

Restriction enzyme accessibility assay:

Nuclei were isolated from 1 g of mixed male and female adult flies frozen at –80° and digested with XbaI as previously described (LU et al. 1993a). After extraction of the nuclear DNA, it was further cut with SalI. The amount of digestion was assayed on Southern blots and radioactively probed with the unique barley cDNA sequences (Figure 1). Blots were exposed to a phosphorimager screen and radioactive counts from hybridization signals were analyzed using ImageQuant software (Molecular Dynamics). Two different XbaI concentrations, 50 and 150 units, were used for checking the restriction site accessibility. These concentrations were arrived upon after empirically checking a number of different concentrations of XbaI. XbaI concentrations >150 units gave a similar percentage of cut values for all genotypes and less or no XbaI did not give any relevant information regarding restriction enzyme accessibility. The percentage cut of the hypersensitivity sites on the Southern blot is calculated as the sum of signal in distal and proximal fragments over the total signal in all three fragments. The percentage of cut value for ab28/+ is set at 100%. The relative percentage of cut values from at least three different restriction digestions from three separate nuclei preps was considered for statistical analysis and a one-tailed Student's t-test was performed to test for significant loss of accessibility.

View larger version (24K): In this window In a new window Download PPT slide   FIGURE 1.— Nucleosomal compaction and trans-inactivation. (A) The P{hsp26-Pt} transposon inserted near the chrw gene, 4.7 kbp proximal to the bwD insertion site on the opposite chromosome. The chromosomal region around bw is depicted by colored boxes representing known and annotated genes drawn to scale. bwD is not to scale. (B) Eye phenotypes of flies of the indicated genotype. All flies are also white– at the endogenous locus on the X chromosome, so silencing is of the hsp70-white reporter gene in P{hsp26-Pt}. P{hsp26-Pt}ab28/+ is the negative control with wild-type eyes. P{hsp26-Pt}ab28/bwD is trans-inactivated. P{hsp26-Pt}ab28/E(bwD)40 is more inactivated than P{hsp26-Pt}ab28/bwD. P{hsp26-Pt}39C-2 contains the transposon inserted in the centric heterochromatin and is silenced in cis (WALLRATH and ELGIN 1995). (C) The P{hsp26-Pt} transposon with the hsp26 promoter driving a unique plant cDNA and the hsp70 promoter driving the white gene. The details of the hsp26-Pt region show a positioned nucleosome flanked by two DNAse hypersensitivity sites (proximal, PDH; distal, DDH) containing XbaI restriction sites and two SalI sites, one each in the hsp26 regulatory region and the plant fragment, used to generate Plant probe (Pt probe) (WALLRATH and ELGIN 1995). (D) Southern blot showing restriction enzyme accessibility of the four genotypes P{hsp26-Pt}ab28/+ (ab28/+), P{hsp26-Pt}ab28/bwD (ab28/bwD), P{hsp26-Pt}ab28/E(bwD)40 [ab28/E(bwD)40], and P{hsp26-Pt}39C-2 (39C-2). The SalI cuts generate a 3-kbp parent fragment, the SalI–XbaI distal cuts generate a 1.2-kbp fragment, and the SalI–XbaI proximal cut gives rise to a 0.8-kbp fragment. On the Southern blot are chromatin digests from the four genotypes with XbaI and SalI. (E) The percentage accessibility of each genotype plotted as bar charts with standard error (n = 3). The value for ab28/+ is set at 100% and relative percentage cuts are calculated for the other three genotypes (MATERIALS AND METHODS). The P-values from a one-tailed Student's t-test for pairs of interest are above the brackets. All three genotypes, ab28/bwD, ab28/E(bwD)40, and 39C-2, are significantly less accessible than ab28/+. The restriction enzyme sites are significantly more protected in the presence of E(bwD)40 as compared to ab28/bwD.

 

Chromatin immunoprecipitation:

Chromatin from adults flies was prepared from 0.2 g of flies of the selected genotype and sex. These were homogenized in 1.8% formaldehyde and crosslinked chromatin was extracted as previously described (NEGRE et al. 2006) except for the following steps. The chromatin extract obtained after Centricon column washes from 0.2 g of flies was utilized for five to six IPs. One aliquot was kept aside as input material and 500 µl of lysis buffer was added to other aliquots. The chromatin was precleared by adding 1 µg of rabbit polyclonal serum (poly IgG) (all antibodies used here were developed in rabbit) and incubated at 4° for 12–16 hr. All IPs were set up with the precleared chromatin with an IgG control kept to serve as a negative control for the precipitation. The preclearing conditions ensured that almost no signal was obtained with the IgG control precipitation, which was done with each experimental set. Different volumes of the resulting ChIP DNA (representing a fold difference, e.g., 2 and 6 or 1, 3, and 9 µl) were used in radioactive duplex PCRs with the specified primer sets for 22–26 cycles. The different DNA concentrations and empirically determined cycle numbers ensured that for each reaction the PCR products were produced while the reaction was not substrate limited, i.e., in the linear range.

Chromatin from male third instar larvae was obtained by dissecting out the brain and imaginal disc complex and processed for crosslinking following the method previously described with the following modifications (LEIBOVITCH et al. 2002). Ten disc/brain complexes per antibody were used. Sonication was carried out with the Branson Sonifier 250: power output 3, duty cycle 100%, 3 x 10 sec by chilling the tubes on ice in between. After aliquoting the crosslinked chromatin extract to tubes for independent IPs, one aliquot was kept separate for input and 500 µl immunoprecipitation (IP) buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris–Cl pH 8, 16.7 mM NaCl, 1 mM PMSF, 1 µg/ml aprotinin) was added to the other aliquots. Preclearing was done as described above for adult ChIP. The precleared chromatin was incubated overnight at 4° with antibodies. Salmon sperm DNA block protein A sepharose beads (Upstate) were used to pull down the antibodies. Beads were washed one time each with low salt wash buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris–Cl, pH 8, 150 mM NaCl), high-salt wash buffer (same as low-salt except 500 mM NaCl), LiCl wash buffer (0.25 M LiCl, 1% NP-40, 1% deoxycholate, 1 mM EDTA, 10 mM Tris–Cl, pH 8), and two times with TE, pH 8. For each wash, 1 ml of solution is added, rotated for 5 min in cold, and spun for 3 min at 3000 rpm in cold. After the final wash, 250 µl of freshly prepared elution buffer (1% SDS, 0.1 M NaHCO₃) was added, rotated at room temperature for 15 min, and centrifuged. The eluate was transferred to new tubes and another round of elution was carried out with the beads and the eluates finally combined. A total of 20 µl of 5 M NaCl was added to 500 µl of combined eluates, and histone–DNA crosslinks were reversed by heating at 65° overnight. With this step and following, the input DNA aliquot was also included with the other tubes. A total of 10 µl of 0.5 M EDTA, 20 µl of 1 M Tris–Cl, pH 6.5, and 2 µl of 10 mg/ml proteinase K was added to the combined eluates and incubated for 1 hr at 45°. DNA was recovered by phenol chloroform extraction and ethanol precipitation. DNA was finally resuspended in water and processed as described with adult ChIP.

Antibodies:

The antibodies used in the study were anti-H3 (Abcam), anti-di-AcH3 acetylated at K9 and K14 (Upstate), anti-²MeH3K9 (Abcam), and anti-HP1 (Covance PRB-291C). These anti-²MeH3K9 and anti-HP1 antibodies were tested on polytene chromosomes from bw^D larvae and observed decorating the heterochromatic satellite insertion at 59E. Histone antibodies were checked on Western blots of Drosophila nuclear proteins and each recognized only one band of the expected sizes. The HP1 antibody obtained from Covance is reported to detect multiple bands on the Western blot with proteins from whole larvae. However, in Western blots with homogenates from adult nuclei and chromatin extracts from adults, it recognized only two bands. Other workers have seen similar bands in a variety of samples using other HP1 antibodies (JAMES and ELGIN 1986; EISSENBERG et al. 1992), suggesting some kind of modification of HP1 in different tissues over development. These two bands of 30 and 37 kDa are found together only in samples of extracted nuclear and chromatin homogenates as the source of protein. In whole adults and larval salivary gland extracts, the antibody predominantly recognizes the smaller-molecular-weight band. In a homogenate from third instar larvae dissected diploid tissue, it predominantly recognizes the larger-molecular-weight band. We use extracted chromatin from adults and larvae; therefore it appears that the antibody is specific for HP1 from the chromatin source in our experiments.

Primers:

All primer sets to transgene sequences were tested on genomic DNA lacking transgenes to ensure that they did not amplify endogenous sequences.

P{hsp26-Pt} transgene:

The Plant primer set amplifies 367 bp of the coding region of the plant gene and another primer set called Plant(s) utilizes the same forward primer but a different backward primer to amplify 293 bp (Figure 2A). The Plant(s) primer was developed because the Plant primer worked well together with one of the control primers (Pdi) but amplified the same length of sequence and therefore could not be resolved on the gel. The common forward primer is GTCGCCTACAACACGCTCTT. The reverse primer for Plant is TCGTGGATCCTCGTCTTCTT and for Plant(s) is CGCACTTGTTCATGTTCCAT.

View larger version (31K): In this window In a new window Download PPT slide   FIGURE 2.— Chromatin structure of the hsp26 regulatory region in trans to bwD. (A) Detailed view of the hsp26-Plant region of the transposon showing the Plant primer set amplifying 367 bp of the coding region used in the ChIP assays. Another primer set, Plant(s), utilizes the same forward primer as Plant but uses a different reverse primer to amplify 293 bp of the coding region. (B) Phosphorimager scan of a polyacrylamide gel showing products of the duplex PCR with the Plant primer and an internal control primer using HP1-precipitated DNA from adult flies. The ratio of signal in the Plant band and the control band from precipitated chromatin, normalized to ratios from the input DNA, are plotted as bar charts with standard error, and P-values are calculated to determine significant differences as described in MATERIALS AND METHODS. HP1 levels are significantly higher for all three genotypes, ab28/bwD, ab28/E(bwD)40, and 39C-2, compared to ab28/+ in the transcribed region of the transposon (n = 3). (C) Bar charts with standard error of data from duplex PCR with 2MeH3K9 antibody-precipitated third instar larval DNA (n = 3, where n is three independent chromatin extractions). Methylation level on the transcribed region is slightly higher for the transposon inserted in the centric heterochromatin as compared to ab28/+. No difference in the methyl modification levels is seen with ab28/bwD or ab28/E(bwD)40. (D) Bar charts with standard error of data from duplex PCR with HP1 antibody-precipitated third instar larval DNA (n = 3). HP1 is significantly enriched on 39C-2 in the coding region as compared to the ab28/+ control but no different for ab28/bwD and ab28/E(bwD)40.

 

P{lacW} transgene:

hsp70-mw utilized the chimeric region between the hsp70 termination sequences and regulatory region of mini-white to give a 199-bp product (Figure 3B). Left and right primers are TCCCCGGGAATTCTAGTATGTA and GGAATGTCATTTTGAGTGAGA, respectively. mw(reg) amplified 230 bp of the mini-white regulatory region, covering the transcription start site and going a little into exon 1 (Figure 3B). The left primer is CCAAAACTCCTCTCGCTTCTT and the right primer is TGTTCAGATGCTCGGCAGAT.

View larger version (23K): In this window In a new window Download PPT slide   FIGURE 3.— Chromatin structure of the mini-white regulatory region in trans to bwD. (A) Fly eyes of the genotypes y w67c23; P{lacW}chrw/+ (mw/+, left) and y w67c23; P{lacW}chrw/bwD; P{cos bw+}/+ (mw/bwD, right). The photo is taken from CSINK et al. (2002). See CSINK et al. (2002) for a detailed explanation of the genotypes. (B) The P{lacW} transposon has a mini-white reporter gene and a lacZ enhancer trap. The lacZ has termination (T) sequences from hsp70. A detailed view of the mini-white regulatory region shows two sets of primers designed to amplify ChIP DNA products. Primer hsp70-mw utilizes the chimeric region of hsp70 and mini-white and amplifies 199 bases primarily of mini-white. Primer mw(reg) covers the transcription start (+1) and goes into the first exon of mini-white. (C) Bar charts with standard error of data from duplex PCR with di-AcH3 antibody-precipitated adult fly DNA (n = 3). Lower acetylation levels on the transposon opposite bwD with both primer sets are found. (D) Bar charts with standard error of data from duplex PCR with HP1 antibody-precipitated adult fly DNA (n = 3). Both primers show significantly higher occupation of HP1 on the transposon in the presence of bwD.

 

bw region primers:

bw (exon) is 71 bp from the bw^D insertion site (PLATERO et al. 1998) and amplifies 206 bp of bw sequence. The left primer is CACTCGGAGGCCATCTTATC and the right primer is CTACAGCCTGTCCGCCTACT. bw(UTR) is 10.3 kbp from the insertion and 12 bp from the bw transcription start in the UTR and amplifies 230 bp of the brown gene. Left and right primers are AGTTCGATGGGTTCAGTCAC and GTCTTCCACGCGAACAGT, respectively. cg30181 is 12 kbp from the bw^D insertion in the gene adjacent to bw, CG30181, and amplifies 201 bp. Sequences amplified were left primer, GGACACATTCCACATCTCCT, and right primer, GTGGTCGTAAGCAGTGTGAT.

All PCRs done are duplex PCRs utilizing one of the primers against the transgene sequences or the endogenous sequences in cis to bw^D, in combination with either of two control primers. Of the two control primers used in our studies, one amplifies 277 bp of GpDH, a gene on the right arm of the second chromosome, and another amplifies 367 bp of Pdi, a gene on the third chromosome. Both these genes are not affected by the presence or absence of bw^D near the end of the left arm of the second chromosome. The combination of primer sets with GpDH worked better for some experiments and the combination with Pdi worked better for the others. The specific control primers used with each transgenic/genetic primer set are the following: GpDH, left primer—TGTACTGCGCATTGGAAAAG; and right primer—TACTACTTGGCTCCGCGATT. Pdi amplifies TTCGTCTTTGGTGTGAGCAG and AACTCGAAGATGCGGGTATG with the left and right primers, respectively. The following set of primers was used in duplex PCRs: P{hsp26-Pt} transgene in the ab28 set for H3, AcH3, and HP1 Plant primer plus GpDH and for ²MeH3K9, Plant(s) plus Pdi were used. For the P{lacW} transgene for H3, AcH3, and HP1, both hsp70-mw and mw(reg) were used with GpDH. For ²MeH3K9, both of the primers were used in duplex with Pdi. For cis studies, bw(exon), bw(UTR), cg30181, and hsp70-mw were used with Pdi for H3, AcH3, HP1, and ²MeH3K9.

Gels and quantification:

Each ChIP experiment for each genotype was independently replicated at least three times, each time from new starting material. The duplex PCR products were resolved on a 6% native polyacrylamide gel. The dried gels were exposed to phosphorimager screens, scanned, and the radioactive signal from each band was measured using Image Gauge software (Fuji Film). After background subtractions for each band, the ratio for the experimental/control band was calculated. The ratios were normalized to the input ratio for each genotype and summed for all the genotypes being compared for one antibody. The ratio for each genotype was divided by the sum and these summed values were used for statistical analysis. Cell bar charts with standard error were plotted and a one-tailed Student's t-test analysis was performed to analyze significant increase or decrease in histone modification or protein. P-values <0.05 are considered significant.

To study the effects of heterochromatic silencing in trans, fly lines were developed and collected in our lab that had transgenes inserted on the wild-type chromosome near the bw locus at band 59E. The reporter genes in these transgenes were silenced to various degrees under different conditions (CSINK et al. 2002; SAGE et al. 2005). The transgenic insertions provide us with unique sequences to look for chromatin structure changes in trans and dissect them out from changes in cis due to bw^D. Two different transgenes in three setups were used in the experiments detailed below. The bw gene is transcribed distal to proximal and the bw^D insertion site is in the eighth exon of the gene (PLATERO et al. 1998) (Figure 1A).

P{hsp70-w-hsp26-Pt} transgene [referred to from here on as P{hsp26-Pt}] has an hsp70 promoter driving the white gene and a hsp26 promoter ahead of a plant gene (WALLRATH and ELGIN 1995) (Figure 1C). In the fly line P{hsp26-Pt}ab28 (referred to as ab28), the transgene is present near the chrw locus, 4.7 kbp proximal to the bw^D insertion on the opposite chromosome (Figure 1A). The hsp70-white transgene is partially trans-inactivated in ab28/bw^D adults and phenotypically shows an eye with a gradient of pigment (Figure 1B). In Northern blots, the hsp26-plant reporter gene also shows reduced RNA levels, indicating trans-inactivation (CSINK et al. 2002).

The P{lacW} transgene has an enhancer trap lacZ reporter gene and a mini-white eye color gene (Figure 3B). The mini-white construct contains the endogenous white gene with some of its upstream regulatory and intronic sequences deleted (BIER et al. 1989). When present at the same location as P{hsp26-Pt} transgene in ab28, the P{lacW} transgene is very strongly inactivated (Figure 3A) as compared to P{hsp26-Pt} transgene (Figure 1B) in the presence of bw^D. The phenotype of P{lacW}chrw/bw^D (referred to as chrw/bw^D) is closer to the classic "salt and pepper" phenotype seen with trans-inactivation with only a few scattered red–orange speckles on a predominantly white background. In the absence of bw^D, the mini-white gene gives a bright-orange eye color at this insertion position (CSINK et al. 2002) (Figure 3A).

The third chromosome assayed here is a derivative of P{lacW}chrw with a flanking deletion that removes 17 kbp of sequences distal to the chrw locus, including the bw gene on the wild-type homolog (Figure 4A). The phenotypic inactivation of the mini-white gene in P{lacW}csc2/bw^D (referred to as csc2/bw^D) is additionally enhanced when compared to chrw/bw^D, which is attributed to a small enhancer of inactivation sequence brought closer to the transgene (SAGE et al. 2005).

View larger version (27K): In this window In a new window Download PPT slide   FIGURE 4.— Analysis of sequences in cis to bwD detects acetylation, methylation, and HP1 differences. (A) The P{lacW} transposon inserted near the chrw locus, 4.7 kbp from bwD on the opposite chromosome. A distal flanking deletion on the wild-type chromosome removes 17 kbp of genomic sequences, including the bw gene, giving rise to the csc2 chromosome shown in the top line. Therefore, csc2/bwD has the transposon and bwD in the same location as in chrw/bwD, but changes the distal cis sequences for the transposon. Three primers sets were designed to study cis changes in chromatin structure by the bwD heterochromatic insert. The first primer, bw (exon), is 71 bp from the bwD insertion. This primer set will most likely pick up the simple sequence satellite sequences in the heterochromatic insert. bw (UTR) is 10.3 kbp from the insertion, 12 bp from the transcription start site of the bw gene. The primer set cg30181 is 12 kbp from the bwD insertion in the gene CG30181 next to bw, in its first exon, 400 bp from its transcription start site. (B–D) Bar charts with standard error of data from duplex PCR with (B) 2MeH3K9, (C) di-AcH3, and (D) HP1 antibody-precipitated larval DNA (n = 3). Methylation level is higher in the presence of bwD with bw (exon) primer set but not farther away in the bw gene or in the gene CG30181 next to bw. HP1 levels are higher in the presence of bwD with all three cis sequences amplifying primer sets. Higher HP1 levels are seen 12 kbp from the bwD insertion.

 
An important strength of this general experimental design is minimal perturbation of the system over different sets of experiments. Keeping the transgenes and bw^D in the same location, this system gives us a chance to directly compare different trans-inactivated sequences at one insertion point and also to analyze the cis sequences under influence from the same heterochromatin. This allows us to build a complete picture of the effect of heterochromatin on chromatin structures of homologous sequences and to draw conclusions about differences in cis and in trans effects.

Effect of bwD on nucleosomal arrangement of a P{hsp26-Pt} transgene:

The P{hsp26-Pt} transgene was analyzed for influence of heterochromatic silencing in cis and found to undergo nucleosomal compaction (WALLRATH and ELGIN 1995). The hsp26 promoter has a positioned nucleosome flanked by two DNaseI hypersensitivity sites that are essential for establishing the potentiated state of the promoter. These hypersensitivity sites harbor XbaI restriction enzyme sites, which serve as a readout of the chromatin structure surrounding them, being more accessible in euchromatin than in heterochromatin (Figure 1C). To analyze the restriction enzyme accessibility of a trans-inactivated transgene and to compare the effects of heterochromatin in trans and in cis, the hsp26 promoter was analyzed in the presence and absence of bw^D in trans.

P{hsp26-Pt}39C-2 (referred to as 39C-2) has the transgene inserted in the centric heterochromatin of the second chromosome. Both reporters are very strongly inactivated (Figure 1B and WALLRATH and ELGIN 1995) and undergo nucleosomal compaction. Therefore 39C-2 was used as a positive control for our analysis. We also examined the nucleosomal compaction of P{hsp26-Pt}ab28/E(bw^D)40 [referred to as ab28/E(bw^D)40]. The E(bw^D)40 chromosome was obtained in a screen for modifiers of bw^D, shows an enhanced bw^D phenotype, and was found to result from a chromosomal rearrangement that moved the bw^D region closer to (but not next to) the pericentric heterochromatin (TALBERT et al. 1994). E(bw^D)40 also inactivates the white reporter in P{hsp26-Pt}ab28 to a greater degree than the unrearranged bw^D chromosome (Figure 1B). Data from eye pigment assays showed less pigment with ab28/E(bw^D)40 (0.61, SE ± 0.007, n = 5) as compared to ab28/bw^D (0.67, SE ± 0.004, n = 5) (CSINK et al. 2002).

The genotype 39C-2 that is cis-inactivated showed only 20% accessibility when the transgene was inserted in the centric heterochromatin. The two experimental genotypes ab28/bw^D and ab28/E(bw^D)40 were significantly less accessible than the genotype with the transgene in the absence of bw^D (ab28/+). ab28/bw^D showed 15% more protection while E(bw^D)40 was 33% more protected (Figure 1E). The additional phenotypic inactivation seen in ab28/E(bw^D)40 compared to ab28/bw^D was also significantly different at the molecular level, reflected in ab28/E(bw^D)40 being 10% more protected than ab28/bw^D (Figure 1E). These results indicate that nucleosomal compaction accompanies trans-inactivation and that heterochromatin can influence chromatin structure in trans to bring about its compaction. E(bw^D)40 does bring about additional compaction over bw^D, corroborating its stronger phenotypic inactivation. Nonetheless, it is noteworthy that the magnitude of overall compaction in trans is less as compared with silencing in cis.

Analyzing histone modifications and nonhistone protein association of the P{hsp26-Pt} transgene in trans in adult flies:

Histone tail modification changes accompany genetic inactivation in cis. Modifications such as ²MeH3K9 are enriched in heterochromatin while acetylation of H3 is found to decrease in heterochromatin and increase in euchromatin (reviewed in MARGUERON et al. 2005; TALBERT and HENIKOFF 2006). To determine if there were any changes in histone modifications of the transgene when inactivated by bw^D in trans, ChIP assays were carried out on chromatin from adult flies. All four of the genotypes compared for nucleosomal arrangement were used in ChIP with antibodies against euchromatic (di-AcH3) and heterochromatic marks (²MeH3K9) and HP1. Histone occupancy was also assayed with a general anti-H3 antibody.

ChIP was carried out on whole adult flies and duplex radioactive PCRs were run utilizing the primers that amplified the transcribed plant sequences (MATERIALS AND METHODS and Figure 2A). A representative gel of PCR products is shown along with some of analyzed data (Figure 2B), while all of the data from each antibody are shown in Table 1. Histone occupancy was checked with anti-H3 and was not found to be any different for the transcribed regions in ab28/bw^D and 39C-2 when compared to ab28/+ but was slightly reduced on the transgene opposite E(bw^D)40. Neither the euchromatic mark di-AcH3 nor the heterochromatic mark ²MeH3K9 differed when heterozygous with bw^D. These marks were also unchanged in E(bw^D)40 and 39C-2 when compared to ab28/+ (Table 1). It was fairly surprising to observe a lack of differential histone modification levels for 39C-2.

View this table: In this window In a new window   TABLE 1 Chromatin immunoprecipitation of hsp26-pt near heterochromatin

 
HP1 is a nonhistone chromosomal protein usually enriched in heterochromatin (reviewed in EISSENBERG and ELGIN 2000). Mutations in the Su(var)205 that encodes the HP1 protein are moderate suppressors of bw^D phenotype (SASS and HENIKOFF 1998). In polytene chromosome squashes, HP1 is found to bind the bw^D heterochromatin but does not detectably bind the trans-inactivated wild-type homolog (PLATERO et al. 1998). The binding pattern of HP1 might be different in diploid cells as polytene nuclei are larger in size. Therefore, it might be possible for HP1 to physically cross over to the homologous chromosome in a smaller nuclear space to bring about chromatin structure changes. In the presence of bw^D, HP1 levels are found to be 58% higher on the coding region of the transgene. HP1 is about twofold higher on 39C-2 and also significantly higher for E(bw^D)40 when compared with ab28/+ (Figure 2B and Table 1). Taken together, these data indicate that HP1 may be directly involved in trans-inactivation.

Developmental comparison of chromatin structure in larvae:

The lack of substantial difference in the histone modification marks observed in the various silenced loci was unexpected. To determine if these marks were present at an earlier stage and were lost in the adult stage, we did ChIP analysis of third instar larval chromatin. All previously used histone modification antibodies and anti-HP1 were used for ChIP, and transcribed region primers in duplex PCR were utilized to detect enrichment or the loss of it in trans-inactivated genes. In our larval studies, we have used only dissected central nervous system (CNS) and imaginal disc complexes. Unlike the bulk of the larval structures, these contain diploid nuclei, not polytene. Indeed, these are the cells that give rise to most of the adult structures, as the majority of the rest of the larval (polytene) tissue undergoes histolysis during metamorphosis. Hence, for an examination of preadult chromatin it is most appropriate to use the CNS and imaginal discs. Additionally, it is especially important that our study use only diploid tissue, because a high level of bw^D association with the centric heterochromatin compartment that promotes bw^D trans-inactivation is observed only in diploid tissue (TALBERT et al. 1994; CSINK and HENIKOFF 1998; THAKAR et al. 2006). Also, polytene tissue has severely underreplicated chromocentric heterochromatin, which may alter some heterochromatic effects.

²MeH3K9 levels were very slightly significantly higher on the heterochromatin-inserted transposon in 39C-2 than on the euchromatic transposon in ab28/+ (Figure 2C). No difference in methylation was observed on the transposon opposite bw^D or E(bw^D)40 (Figure 2C). Histone occupancy of H3 and the euchromatin-enriched modification mark AcH3 were not found to be any different for any of the genotypes (Table 1). As seen in adults, HP1 levels were significantly high in P{hsp26-Pt} in 39C-2 as compared to ab28/+ (Figure 2D). But unlike in the adult stage, HP1 levels were not any different for the coding region of the transgene in the presence of bw^D or E(bw^D)40. ChIP results for P{hsp26-Pt} from two developmental stages suggest that methylation of H3K9 does not play a role in trans-inactivation. Perhaps HP1 interacts with the DNA by itself or through another modification mark or protein partner.

Chromatin structure of a strongly trans-inactivated transgene, P{lacW}:

To determine whether the lack of histone modification changes is a general feature of trans-inactivation and is not transgene specific, we utilized another transgene, P{lacW}, in ChIP experiments. We were additionally motivated to study P{lacW} present in exactly the same location as P{hsp26-Pt} in ab28 because the P{lacW} transgene is very strongly trans-inactivated as compared to P{hsp26-Pt} (compare Figures 1B and 3A) (CSINK et al. 2002). With a stronger phenotype, it might be experimentally possible to detect changes in histone modifications. Analyzing the regulatory region of the mini-white gene also provides us with an opportunity to study the chromatin structure of a different type of promoter and compare it to the potentiated hsp26 promoter region. The mini-white promoter is weaker than even the constitutive expression of the heat-shock promoters. Expression of mini-white promoter is tissue and developmentally specific, while the hsp26 promoter can be expressed in most tissues and stages of development (GLASER et al. 1986). Additionally, the hsp26 promoter is believed to be in a potentiated state in all tissues, unlike the white gene (LU et al. 1993b). This multitude of differences between the two promoters could lead them to behave dissimilarly even under the influence of the same heterochromatin in trans.

Two different primer sets in the mini-white reporter, one covering the transcription start site and a second farther upstream, were developed to study the P{lacW} transgene in the chrw locus (MATERIALS AND METHODS and Figure 3B). ChIP on chromatin from adult flies found that AcH3 levels were significantly lower with both primer sets for the transgene over bw^D (Figure 3C). H3 and ²MeH3K9 levels were not found to be any different (Table 2). HP1 levels were significantly higher with both primer sets for chrw/bw^D compared to chrw/+ (Figure 3D). Therefore, AcH3 modifications play a modest role in trans-inactivation, and with a stronger inactivation phenotype, it is possible to experimentally detect molecular differences.

View this table: In this window In a new window   TABLE 2 Chromatin immunoprecipitation of mini-white (mw) over bwD

 
ChIP with third instar larval diploid tissue was also done for P{lacW}chrw, but H3, AcH3, and ²MeH3K9 levels were not any different between the two genotypes. HP1 was significantly higher for chrw/bw^D with both primer sets, albeit only modestly (Table 2). Therefore, again we see differential histone modification patterns of the same genetic locus over development.

Chromatin structure of endogenous sequences in cis to the bw^D heterochromatic insertion:

As we were puzzled by the lack of, or small change in, the histone modification marks, especially the ²MeH3K9 in cis silencing P{hsp26-Pt}39C-2, we wanted to determine if we could detect any changes of this type in cis. We further wanted to analyze the effects of bw^D heterochromatin on cis sequences to ascertain the association of constitutive heterochromatin properties with bw^D.

For these experiments, we utilized the fly line csc2, which carries a deletion of the bw gene and other nearby sequences on the csc2 chromosome (SAGE et al. 2005). Primers to endogenous sequences within this deletion will amplify only regions on the opposite homolog when the csc2 chromosome is heterozygous with either a wild-type chromosome or bw^D. Three primers to sequences of increasing distance from the bw^D insertion site were designed so that spreading of any chromatin structural changes in cis could be tracked (Figure 4A). To directly compare the effect of bw^D on trans sequences in this same setup, primer set hsp70-mw(reg) against P{lacW} was utilized for PCR with the same ChIP DNA as used with the cis primer sets (Figure 3B).

ChIP was carried out on third instar larval diploid tissue as described above with antibodies against euchromatic and heterochromatic histone modification markers. In contrast to our previous results in trans, the chromatin immediately next to the bw^D insertion site (71 bp away) amplified by primer set bw(exon) was significantly enriched for ²MeH3K9 in the presence of bw^D as compared with csc2/+ (Figure 4B). However, similar to our results in trans, no difference in methylation levels was observed for the more distant primer sets, bw (UTR) or cg30181 (Figure 4B), nor in trans on the P{lacW} transgene on csc2 (Table 3). ChIP was carried out with di-AcH3 antibodies and PCR amplification was done with all four primers—bw (exon), bw (UTR), cg30181 in cis, and hsp70-mw(reg) in trans. Acetylation levels were significantly lower for csc2/bw^D compared to csc2/+ (Figure 4C), with all three primer sets in cis, but not for the trans primer (Figure 4C and Table 3).

View this table: In this window In a new window   TABLE 3 Chromatin immunoprecipitation of endogenous sequences in cis to bwD

 
ChIP studies for HP1 occupation found that HP1 was much higher in the presence of bw^D with all three cis primer sets (Figure 4D). This indicates that HP1 is present at the bw^D heterochromatin and can spread in cis. This cis-spreading can be detected at least 12 kbp from the heterochromatic block. In contrast, when the regulatory region of the transgene in trans was amplified with the primer set hsp70-mw(reg), no significant difference in HP1 levels could be detected in the presence of bw^D (Table 3). This does not agree with the trans results obtained with P{lacW} in chrw/bw^D and leads us to believe that the surrounding cis sequences also play a role in determining the chromatin structure in trans (also see SAGE et al. 2005).

Genetic interaction of PEV modifiers with trans-inactivation:

There are estimated to be 50–150 genes in the fly genome that act as second site modifiers of PEV. Mutations in genes that act as suppressors [Su(var)s] or enhancers [E(var)s] of PEV usually encode proteins involved in nucleation, maintenance, or modification of chromatin. Many dominant modifiers of PEV are characterized by dosage sensitivity that either suppresses or enhances PEV in a single dose (haplo-suppressors) and have reciprocal effects in three doses (triplo-enhancers). Examples of this class of modifiers include the loci Su(var)205 that encodes the HP1 protein, Su(var)3-9 that codes for the H3K9-specific methyltransferase, and the Su(var)3-7 loci that give rise to a zinc-finger protein. All three of these proteins localize to chromocentric heterochromatin of polytene chromosomes as well as to other scattered euchromatic sites, both shared and unique.

The altered chromatin structure of genes inactivated by heterochromatin in trans, in terms of histone modifications and levels of associated proteins, appears to be different from what is usually seen in cis-PEV. Therefore, to determine the degree of similarity between the two kinds of PEV, we wanted to test the phenotypic effects of some known cis-PEV modifier mutations on bw^D inactivation. Previous work had found that some Su(var) loci that encode proteins known to play a role in cis-PEV have a different phenotypic effect on trans-inactivation (SASS and HENIKOFF 1998). Su(var)205, Su(var)3-9, and Rpd3 (a histone deacetylase) are very strong suppressors of the classic cis-PEV allele w^m4 (GRIGLIATTI 1991; REUTER and SPIERER 1992). However, Su(var)205-5 and Rpd3 were found to be only moderate suppressors of the trans-inactivation phenotype while Su(var)3-9 did not affect the phenotype at all (Figure 5). In contrast, Su(var)3-7, which is a weaker cis-PEV suppressor (REUTER and SPIERER 1992 and our data not shown), was found to be a very strong suppressor of bw^D inactivation (Figure 5). Therefore, it appears that different proteins are key players in the two kinds of PEV.

View larger version (117K): In this window In a new window Download PPT slide   FIGURE 5.— Effect of PEV modifiers on bwD trans-inactivation. Su(var)3-7/+ is a very strong suppressor of bwD inactivation, but Su(var)205 and Rpd3 are only mild suppressors while Su(var)3-9 does not suppress trans-inactivation at all. Known functions and products of the Su(var)s are in parentheses below the genotypes. The wild-type chromosomes do not carry balancers and the + chromosome in the v; +/bwD control is from the Canton-S wild-type line.

 
These genetic data agree with our molecular results very well. The moderate phenotypic suppressors of trans-inactivation, Su(var)205-5 and Rpd3, do show a corresponding small change at the molecular level in HP1 distribution and decreased acetylation on the trans-inactivated transgenes. Additionally, the lack of effect of Su(var)3-9 on bw^D trans-inactivation correlates well with the unchanged levels of ²MeH3K9 on the trans-inactivated sequences.

Effect of Su(var)3-7 on bw^D inactivation and nucleosomal arrangement of the P{hsp26-Pt} transgene:

An extra copy of Su(var)3-7 enhances the inactivation of the hsp70-w reporter in the P{hsp26-Pt} transgene in ab28/bw^D (Figure 6A). This can be seen as less pigment in the eyes of the ab28/bw^D flies carrying three doses of Su(var)3-7 (referred to as ab28/bw^D,T21a) when compared with ab28/bw^D, which has only two doses of the gene. Restriction enzyme accessibility of XbaI sites in the P{hsp26-Pt} transgene opposite bw^D are further reduced by 10% in the presence of an extra copy of Su(var)3-7 when compared with ab28/bw^D (Figure 6B).

View larger version (47K): In this window In a new window Download PPT slide   FIGURE 6.— Effect Su(var)3-7 on trans-inactivation. (A) An extra dose of Su(var)3-7 enhances trans-inactivation. (Left) The trans-inactivated P{hsp26-Pt} transgene in the presence of bwD. (Right) The bwD inactivated transgene with an extra copy of Su(var)3-7 carried in the transgene P{Suv3-7}T21a. (B) Bar graph with standard error of the nuclease accessibility of each genotype (n = 3). See Figure 1 for transposon and probe details. The value for ab28/+ is set at 100% and relative percentage cuts for the other genotypes are calculated. The P-value from a Student's t-test for pairs of interest is above the bracket. In the presence of an extra copy of Su(var)3-7, the restriction enzyme accessibility was significantly lower as compared to ab28/bwD. 39C-2 is a positive control.

 

In the earliest example of cis-inactivation of the white gene by pericentric heterochromatin of the X chromosome, the dark staining pattern of the permanently condensed heterochromatin was seen to spread over to the inactivated gene in polytene chromosomes. With evidence from various systems, the linear spread of heterochromatic properties and proteins from a nucleating block of heterochromatin along the chromatin fiber to the inactivated nearby gene has come to be the favored molecular explanation of silencing by heterochromatin in cis (reviewed in TALBERT and HENIKOFF 2006). However, this model is unsatisfactory or inadequate for silencing by heterochromatin in trans. Our results in this work revealed both similarities and differences in the chromatin structure of genes silenced by heterochromatin in cis and in trans.

Nucleosomal compaction:

Restriction enzyme accessibility assays of the hsp26 promoter in the P{hsp26-Pt} transgene revealed increased protection of the restriction enzyme sites in the presence of bw^D, indicating enhanced nucleosomal compaction. E(bw^D)40 enhances inactivation of the transgene and further decreases the restriction enzyme accessibility. However, the magnitude of protection is much less than cis-inactivated transgene in 39C-2. This could be because the heterochromatic effects have to reach in trans from the opposite chromosome and at a distance of 4.7 kbp from the bw^D insertion (Figure 1A). Proteins have to diffuse out farther away in trans as compared to spreading linearly in cis, which could result in relatively less compaction as compared with cis-inactivation. Another factor influencing this chromatin structure alteration could be that the transgene inserted in the pericentric heterochromatin of the second chromosome is always present in the centric heterochromatic environment. On the other hand, the association of bw^D with the heterochromatic compartment is not established until hours into G₁ of the cell cycle and often not until the cell begins to differentiate (CSINK and HENIKOFF 1998; THAKAR and CSINK 2005). Therefore, the transgene in the pericentric heterochromatin is exposed earlier and longer to a heterochromatic environment.

Histone modifications and HP1 in bw^D trans-inactivation:

While the complete lack of difference in H3K9 dimethylation levels on the transgenes over bw^D was unexpected, this biochemical evidence agrees with the genetic observation that mutations in the Su(var)3-9 gene, encoding the H3K9-specific methyltransferase in flies, fail to suppress trans-inactivation by bw^D (Figure 5 and SASS and HENIKOFF 1998). In contrast, mutations in the HP1 gene are moderate suppressors of bw^D (Figure 5) and HP1 levels are usually enriched on the trans-inactivated transgenes. These data support the idea that HP1 has some role to play in trans-inactivation, but does not act through the ²MeH3K9 modification mark, which is considered the primary chromatin-docking site for HP1 in flies (BANNISTER et al. 2001; JACOBS et al. 2001). However, there are a number of studies in Drosophila that question the requirement of ²MeH3K9 for HP1 recruitment and function on chromatin. For example, HP1 is present on many locations on polytene chromosomes that do not overlap with ²MeH3K9 staining (FANTI et al. 2003). Other studies have also shown that HP1 can act in both an MeH3K9-dependent and -independent manner (DANZER and WALLRATH 2004). Therefore, it appears that the role of HP1 in trans-inactivation is independent of its interaction with ²MeH3K9 histone modification. HP1 could interact with DNA on its own, through an unknown trans silencing-specific histone modification mark or by interaction with other proteins. HP1 has also been shown to interact dynamically with chromatin and compete for heterochromatic binding sites with other proteins. Fluorescence recovery after photobleaching experiments found that HP1 is constantly moving on and off the chromatin, but this transiently unbound HP1 nevertheless remains localized to the heterochromatic compartment. As the local concentration of HP1 would be higher in this region, it may bind to lower-affinity partners in a dynamic manner and promote silencing (CHEUTIN et al. 2003; FESTENSTEIN et al. 2003). This suggests that location to the heterochromatic compartment may allow binding of HP1 to certain sequences or proteins, which would remain unbound in a euchromatic neighborhood.

Even though the silencing of the hsp70-white transgene is enhanced in ab28/E(bw^D)40 relative to ab28/bw^D flies, the amount of HP1 does not appear to increase. It is possible that this result indicates that HP1 plays no role in trans-inactivation of the hsp70-white transgene. However, it is likely that a large number of factors play a role in trans-inactivation and it is possible that different sets of factors play different absolute or relative roles in the trans-inactivation brought about by the two chromosomes. It should be kept in mind that E(bw^D)40 may silence in a somewhat different manner than an unrearranged bw^D chromosome. The rearrangement of the chromosome that produces the enhanced phenotype likely promotes earlier association with the heterochromatic environment, as somatic pairing of the rearranged and unrearranged chromosomes brings it closer to centric heterochromatin in the space of the nucleus (TALBERT et al. 1994). It is possible that other, earlier-acting factors are responsible for the enhanced silencing of ab28/E(bw^D)40, while the role of HP1 remains unchanged.

Developmental variation of chromatin structure of trans-inactivated genes:

By examining the same sequences in two different stages of organismal development, we made a number of interesting observations. Decreased acetylation of mini-white sequences was seen only in chromatin from adults (Figure 3C). Enrichment of HP1 on trans heterochromatin was observed only on P{hsp26-pt}ab28 in adults (Table 1) and was approximately fourfold higher on mini-white sequences in adults than in larvae (Table 2). This greater HP1 enrichment in adults is not surprising, as a number of changes take place during differentiation that may promote the formation of heterochromatin and deposition of HP1 and stabilize trans-inactivation. First, heterochromatinization and gross nuclear chromatin condensation is generally increased with exit from the cell cycle and differentiation (FRANCASTEL et al. 2000). Second, HP1 is only transiently associated with the centric heterochromatin during the cell cycle (PLATERO et al. 1998; DORMANN et al. 2006), and more dividing cells make up the larval diploid tissue than the adult. Third, there is a general decrease in large-scale chromatin dynamics in differentiating cells that may contribute to the stability of the heterochromatic compartment and the HP1 resident there (THAKAR et al. 2006). Finally, bw^D heterochromatic associations are not set up until the latter part of G₁. Often the length of G₁ is not sufficient to allow association in undifferentiated cells, so association is delayed until a cell exits from the cell cycle during differentiation (CSINK and HENIKOFF 1998; THAKAR and CSINK 2005). Such differentiated cells, long past cell cycle exit, would make up a larger proportion of adult cells.

Fundamental differences in the developmental regulation of the two transgenes used in this study may shed light on the differing results obtained with the two different transgenes (Tables 1 and 2). In general, the P{lacW} transposon shows more enrichment of heterochromatic marks than P{hsp26-Pt} when heterozygous with bw^D. This agrees with the observation that P{lacW}chrw is more strongly silenced than P{hsp26-Pt}ab28 (CSINK et al. 2002) and the observation that early expressing genes are more resistant to trans-inactivation (SAGE et al. 2008). The potentiated state of the heat-shock promoters are set up very early in development, while the mini-white promoter is turned on later in development, is weaker, and is tissue specific.

Effect of bw^D heterochromatin in cis:

The bw^D insertion consists of a large block of simple satellite sequence containing AAGAG. While it binds HP1 and other Su(var) proteins in polytene chromosomes, unlike centric heterochromatin, it appears to be uninterrupted by middle repetitive, transposon-derived sequences (PLATERO et al. 1998; DELATTRE et al. 2000). The insertion is in the distal euchromatin of an otherwise unrearranged second chromosome well away from the telomeric heterochromatin. This somewhat unique arrangement prompted us to find out to what extent it was similar to ordinary centric heterochromatin and to what extent its heterochromatic properties spread in cis. Hypoacetylation of H3 appears to spread into distal sequences, indicating more compact structures in cis of bw^D. HP1 was enriched on the sequences distal to the insert up to at least 12 kbp away (Figure 4 and Table 3). H3K9 methylation is higher only on the closest sequence. This may reflect methylation on the actual heterochromatic insert instead of spreading of the mark, as the primer set is only 71 bp from the insertion site (Figure 4A). On the more distal sequences, once again we see HP1 enrichment in the absence of ²MeH3K9 enrichment as discussed above. The absence of ²MeH3K9 spreading in cis of bw^D could be because of interaction of HP1 with a different level of H3K9 modification (mono- or tri-) or a different histone modification mark. Alternatively, the region around bw is especially gene rich, so it is likely that there are sequences that act as boundaries to the spread of chromatin structure.

Effect of Su(var)3-7 on bw^D trans-inactivation:

Both phenotypic and molecular studies indicate that Su(var)3-7 is a key player in trans-inactivation. The gene encodes a protein with seven zinc fingers, which are motifs known to bind DNA (CLEARD et al. 1995). However, the function of Su(var)3-7 is still not very clear. On polytene chromosomes, Su(var)3-7 shows a similar, but not identical, localization pattern to pericentric heterochromatin and euchromatin arms as HP1 (DELATTRE et al. 2000). Both genetic and biochemical experiments have shown that Su(var)3-7 interacts with HP1 (CLEARD et al. 1997; DELATTRE et al. 2000). Evidence indicates that Su(var)3-7 does not recruit HP1 to heterochromatin (JAQUET et al. 2002). However, ectopically delocalized HP1 can recruit Su(var)3-7 (DELATTRE et al. 2000). These results suggest a direct link between Su(var)3-7 and HP1 in PEV and genomic silencing. As our results indicate, HP1 does not utilize its interaction through the ²MeH3K9 mark in trans-inactivation. Instead, it could make use of its functional association with Su(var)3-7 in trans-inactivation. The strong modification by Su(var)3-7 of the bw^D phenotype suggests a role for the protein and its interactions in trans-inactivation.

Overall, our results reveal differences in chromatin structure of genes silenced by heterochromatin in trans from what is seen in cis-inactivation. Some of the key players appear to be different in the two kinds of PEV. These results suggest some mechanistic differences between the two kinds of silencing. Unraveling the involvement of proteins like Su(var)3-7 and their interactions in trans-inactivation will add to our understanding of the mechanism behind heterochromatic gene silencing in trans.

We thank G. Cavalli (Centre National de la Recherche Scientifique, Montpellier) and D. Gilmour (Pennsylvania State University) for sharing procedures for adult and dissected larval ChIP, respectively, before publication and for additional helpful advice. We thank L. Wallrath (University of Iowa), S. Henikoff (Fred Hutchinson Cancer Research Center, Seattle), P. Spierer (University of Geneva), and T. Grigliatti (University of British Columbia) for some of the fly lines used in this study. This work was supported by an American Cancer Society grant (RSG-00-073-04-DDC) to A.K.C.

¹ Present address: Laboratory of Cellular and Developmental Biology, NIDDK, National Institutes of Health, Bethesda, MD 20892.

BANNISTER, A. J., P. ZEGERMAN, J. F. PARTRIDGE, E. A. MISKA, J. O. THOMAS et al., 2001 Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature 410: 120–124.[CrossRef][Medline]

BIER, E., H. VAESSIN, S. SHEPHERD, K. LEE, K. MCCALL et al., 1989 Searching for pattern and mutation in the Drosophila genome with a P-lacZ vector. Genes Dev. 3: 1273–1287.[Abstract/Free Full Text]

CHEUTIN, T., A. J. MCNAIRN, T. JENUWEIN, D. M. GILBERT, P. B. SINGH et al., 2003 Maintenance of stable heterochromatin domains by dynamic HP1 binding. Science 299: 721–725.[Abstract/Free Full Text]

CLEARD, F., M. MATSARSKAIA and P. SPIERER, 1995 The modifier of position-effect variegation Su(var)3-7 of Drosophila: there are two alternative transcripts and seven scattered zinc fingers, each preceded by a tryptophan box. Nucleic Acids Res. 23: 796–802.[Abstract/Free Full Text]

CLEARD, F., M. DELATTRE and P. SPIERER, 1997 SU(VAR)3-7, a Drosophila heterochromatin-associated protein and companion of HP1 in the genomic silencing of position-effect variegation. EMBO J. 16: 5280–5288.[CrossRef][Medline]

CSINK, A. K., and S. HENIKOFF, 1996 Genetic modification of heterochromatic association and nuclear organization in Drosophila. Nature 381: 529–531.[CrossRef][Medline]

CSINK, A. K., and S. HENIKOFF, 1998 Large-scale chromosomal movements during interphase progression in Drosophila. J. Cell Biol. 143: 13–22.[Abstract/Free Full Text]

CSINK, A. K., A. BOUNOUTAS, M. L. GRIFFITH, J. F. SABL and B. T. SAGE, 2002 Differential gene silencing by trans-heterochromatin in Drosophila melanogaster. Genetics 160: 257–269.[Abstract/Free Full Text]

DANZER, J. R., and L. L. WALLRATH, 2004 Mechanisms of HP1-mediated gene silencing in Drosophila. Development 131: 3571–3580.[Abstract/Free Full Text]

DELATTRE, M., A. SPIERER, C. H. TONKA and P. SPIERER, 2000 The genomic silencing of position-effect variegation in Drosophila melanogaster: interaction between the heterochromatin-associated proteins Su(var)3-7 and HP1. J. Cell Sci. 113(Pt. 23): 4253–4261.[Abstract]

DERNBURG, A. F., K. W. BROMAN, J. C. FUNG, W. F. MARSHALL, J. PHILIPS et al., 1996 Perturbation of nuclear architecture by long-distance chromosome interactions. Cell 85: 745–759.[CrossRef][Medline]

DORMANN, H. L., B. S. TSENG, C. D. ALLIS, H. FUNABIKI and W. FISCHLE, 2006 Dynamic regulation of effector protein binding to histone modifications: the biology of HP1 switching. Cell Cycle 5: 2842–2851.[Medline]

EISSENBERG, J. C., and S. C. ELGIN, 2000 The HP1 protein family: getting a grip on chromatin. Curr. Opin. Genet. Dev. 10: 204–210.[CrossRef][Medline]

EISSENBERG, J. C., G. D. MORRIS, G. REUTER and T. HARTNETT, 1992 The heterochromatin-associated protein HP-1 is an essential protein in Drosophila with dosage-dependent effects on position-effect variegation. Genetics 131: 345–352.[Abstract]

FANTI, L., M. BERLOCO, L. PIACENTINI and S. PIMPINELLI, 2003 Chromosomal distribution of heterochromatin protein 1 (HP1) in Drosophila: a cytological map of euchromatic HP1 binding sites. Genetica 117: 135–147.[CrossRef][Medline]

FESTENSTEIN, R., S. N. PAGAKIS, K. HIRAGAMI, D. LYON, A. VERREAULT et al., 2003 Modulation of heterochromatin protein 1 dynamics in primary mammalian cells. Science 299: 719–721.[Abstract/Free Full Text]

FRANCASTEL, C., D. SCHUBELER, D. I. MARTIN and M. GROUDINE, 2000 Nuclear compartmentalization and gene activity. Nat. Rev. Mol. Cell Biol. 1: 137–143.[CrossRef][Medline]

GLASER, R. L., M. F. WOLFNER and J. T. LIS, 1986 Spatial and temporal pattern of hsp26 expression during normal development. EMBO J. 5: 747–754.[Medline]

GRIGLIATTI, T., 1991 Position-effect variegation: an assay for nonhistone chromosomal proteins and chromatin assembly and modifying factors. Methods Cell Biol. 35: 587–627.[Medline]

HARMON, B., and J. SEDAT, 2005 Cell-by-cell dissection of gene expression and chromosomal interactions reveals consequences of nuclear reorganization. PLoS Biol. 3: e67.[CrossRef][Medline]

JACOBS, S. A., S. D. TAVERNA, Y. ZHANG, S. D. BRIGGS, J. LI et al., 2001 Specificity of the HP1 chromo domain for the methylated N-terminus of histone H3. EMBO J. 20: 5232–5241.[CrossRef][Medline]

JAMES, T. C., and S. C. R. ELGIN, 1986 Identification of a nonhistone chromosomal protein associated with heterochromatin in Drosophila melanogaster and its gene. Mol. Cell. Biol. 6: 3862–3872.[Abstract/Free Full Text]

JAQUET, Y., M. DELATTRE, A. SPIERER and P. SPIERER, 2002 Functional dissection of the Drosophila modifier of variegation Su(var)3-7. Development 129: 3975–3982.[Abstract/Free Full Text]

LEIBOVITCH, B., Q. LU, L. BENJAMIN, Y. LIU, D. GILMOUR et al., 2002 GAGA factor and the TFIID complex collaborate in generating an open chromatin structure at the Drosophila melanogaster hsp26 promoter. Mol. Cell. Biol. 22: 6148–6157.[Abstract/Free Full Text]

LU, Q., L. L. WALLRATH and S. C. R. ELGIN, 1993a Using Drosophila P-element-mediated germ line transformation to examine chromatin structure and expression of in vitro-modified genes. Methods Mol. Genet. 1: 333–357.

LU, Q., L. L. WALLRATH, H. GRANOK and S. C. R. ELGIN, 1993b (CT)-n (GA)-n repeats and heat shock elements have distinct roles in chromatin structure and transcriptional activation of the Drosophila hsp26 gene. Mol. Cell. Biol. 13: 2802–2814.[Abstract/Free Full Text]

MARGUERON, R., P. TROJER and D. REINBERG, 2005 The key to development: Interpreting the histone code? Curr. Opin. Genet. Dev. 15: 163–176.[CrossRef][Medline]

NEGRE, N., J. HENNETIN, L. V. SUN, S. LAVROV, M. BELLIS et al., 2006 Chromosomal distribution of PcG proteins during Drosophila development. PLoS Biol. 4: e170.[CrossRef][Medline]

PLATERO, J., A. CSINK, A. QUINTANILLA and S. HENIKOFF, 1998 Changes in chromosomal localization of heterochromatin binding proteins during the cell cycle in Drosophila. J. Cell Biol. 140: 1297–1306.[Abstract/Free Full Text]

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

SAGE, B. T., and A. K. CSINK, 2003 Heterochromatic self-association, a determinant of nuclear organization, does not require sequence homology in Drosophila. Genetics 165: 1183–1193.[Abstract/Free Full Text]

SAGE, B. T., J. L. JONES, A. L. HOLMES, M. D. WU and A. K. CSINK, 2005 Sequence elements in cis influence heterochromatic silencing in trans. Mol. Cell. Biol. 25: 377–388.[Abstract/Free Full Text]

SAGE, B. T., M. D. WU and A. K. CSINK, 2008 Interplay of developmentally regulated gene expression and hete