Copyright Ó 2009 by the Genetics Society of America DOI: 10.1534/genetics.109.106559

Effects of Chromosomal Rearrangements on Transvection at the yellow of Drosophila melanogaster

Sharon A. Ou,* Elaine Chang,† Szexian Lee,† Katherine So,† C.-ting Wu*,1 and James R. Morris*,† *Department of Genetics, Harvard Medical School, Boston, Massachusetts 02115 and †Department of Biology, Brandeis University, Waltham, Massachusetts 02453 Manuscript received June 26, 2009 Accepted for publication August 5, 2009

ABSTRACT Homologous are paired in somatic cells of Drosophila melanogaster. This pairing can lead to transvection, which is a process by which the proximity of homologous can lead to a change in gene expression. At the yellow gene, transvection is the basis for several examples of intragenic comple- mentation involving the enhancers of one acting in trans on the promoter of a paired second allele. Using complementation as our assay, we explored the chromosomal requirements for pairing and transvection at yellow. Following a protocol established by Ed Lewis, we generated and characterized chro- mosomal rearrangements to define a region in cis to yellow that must remain intact for complementation to occur. Our data indicate that homolog pairing at yellow is efficient, as complementation was disrupted only in the presence of chromosomal rearrangements that break #650 kbp from yellow. We also found that three telomerically placed chromosomal duplications, containing 700 or more kbp of the yellow genomic region, are able to alter complementation at yellow, presumably through competitive pairing interactions. These results provide a formal demonstration of the pairing-dependent nature of yellow transvection and suggest that yellow pairing, as measured by transvection, reflects the extent of contiguous homology flanking the locus.

YTOLOGICAL studies of a wide variety of systems Transvection is a process by which the pairing of C are revealingthestrategiesbywhichalargeamount homologous genes results in a change in expression, in of DNA can be organized into an extraordinarily small some situations causing gene activation and in other sit- volume yet still be accurately expressed, replicated, and uations causing gene repression (reviewed by Pirrotta passed through cell divisions. In the somatic cells of Dro- 1999; Wu and Morris 1999; Duncan 2002; Kennison sophila and other dipteran insects, a striking feature of and Southworth 2002). Because it depends on pair- nuclear organization is the extensive amount of pair- ing, transvection can be used as a powerful assay for the ing that occurs between homologous chromosomes. paired state of genes. In this study, we used the protocol This pairing was first noted by Nettie Stevens (Stevens designed by Ed Lewis during his defining studies of 1908) and Charles Metz (Metz 1916) through the transvection at the Ultrabithorax (Ubx) gene (Lewis examination of mitotic nuclei. Somatic pairing of 1954). Lewis began his analyses with the observation homologous chromosomes has now been observed in that certain pairs of Ubx support intragenic Drosophila interphase nuclei using DNA as well as RNA complementation. He predicted that this complemen- in situ hybridization techniques (reviewed by McKee tation depends on the physical pairing of participating 2004; for example, Kopczynski and Muskavitch 1992; alleles and then confirmed this prediction through the Hiraoka et al. 1993; Csink and Henikoff 1998; Fung generation and analysis of chromosomal rearrange- et al. 1998; Gemkow et al. 1998; Sass and Henikoff 1999; ments that disrupted complementation. Interestingly, Bantignies et al. 2003; Ronshaugen and Levine 2004; the vast majority of the rearrangements had at least one Williams et al. 2007; Hartl et al. 2008), assessment of breakpoint in a large chromosomal region between the the frequency of site-specific FLP-mediated recombina- centromere and Ubx on the order of 12 Mbp in size and tion (Golic and Golic 1996a), and methods that mark covering about one-half the arm. This chromosomes with protein tags (Vazquez et al. 2001, region was named the ‘‘critical region’’ and was inter- 2002, 2006). Here, we present our studies using trans- preted as the segment of the chromosome whose vection as our phenotypic assay for chromosomal pairing integrity is important for homolog pairing, and hence in Drosophila. transvection, at Ubx. On the basis of the location of this large critical region between the centromere and Ubx, Lewis (1954) suggested that somatic pairing might 1Corresponding author: Department of Genetics, Harvard Medical School, 77 Avenue Louis Pasteur, Boston, MA 02115. initiate at the centromere and proceed distally toward E-mail: [email protected] the telomere.

Genetics 183: 483–496 (October 2009) 484 S. A. Ou et al.

A number of genes have now been reported to show phenotype of dark cuticular pigmentation (Geyer and transvection in Drosophila, and application of the Lewis Corces 1987; Martin et al. 1989). It is controlled by method to some of these genes has demonstrated several tissue-specific enhancers, including the wing that critical regions can vary greatly in size (reviewed and body enhancers lying upstream of the transcription by Duncan 2002; Kennison and Southworth 2002). start site and the bristle contained within the Like transvection at Ubx, transvection at decapentaplegic single intron (Geyer and Corces 1987). Transvection (dpp)(Gelbart 1982), eyes absent (eya)(Leiserson et al. can occur through two mechanisms at yellow, both of 1994), vestigial (vg)(Coulthard et al. 2005), and one which result in intragenic complementation. In one type of transvection at Abdominal-B (Abd-B)(Sipos et al. mechanism, the enhancers of one allele activate the 1998) is associated with relatively large critical regions, promoter of a second allele in trans when homologous spanning up to a one-third of a chromosome arm or chromosome pairing brings the two alleles close to- more and many megabases proximal to the gene. Con- gether (Geyer et al. 1990; Morris et al. 1999a). Accord- sistent with the large size of these critical regions, it is ing to the other mechanism, enhancers are believed to generally not difficult to isolate rearrangements that dis- bypass a chromatin through a pairing-mediated rupt transvection at these loci. For example, studies of topological change in gene structure (Morris et al. transvection at Ubx (Lewis 1954), dpp (Gelbart 1982), 1998). Here, we describe our work using yellow alleles eya (Leiserson et al. 1994), and Abd-B (Sipos et al. 1998) that support enhancer action in trans. showed that 0.2–0.8% and possibly more of chromo- The assumption that yellow transvection requires somes that had been exposed to 4000–4500 rad of X homolog pairing has rested on a number of studies, rays carried a transvection-disrupting rearrangement. the most important of which addresses the ability of In contrast, transvection at white (w)(Jack and Judd ectopically located yellow genes to support transvection 1979; Smolik-Utlaut and Gelbart 1987; Gubb et al. (Geyer et al. 1990; Chen et al. 2002). In particular, while 1997) and a second example of transvection at Abd-B complementation is observed for pairs of yellow trans- (Hendrickson and Sakonju 1995; Hopmann et al. genes inserted into allelic chromosomal sites, it is not 1995) are associated with small critical regions; only observed between yellow transgenes at nonallelic sites or chromosomal rearrangements that break very close to between an ectopic yellow transgene and yellow at its white or Abd-B are able to disrupt transvection. natural chromosomal location. In addition, it was found Why is transvection more difficult to disrupt for some that while a yellow allele that has been translocated to the genes than for others? A number of explanations have Y chromosome can support complementation in X/Y been suggested (reviewed in Duncan 2002; Kennison flies, it does not do so in X/X/Y flies. Presumably, in the and Southworth 2002; also see Lewis 1954; Gelbart latter case, preferential pairing between the two yellow 1982; Smolik-Utlaut and Gelbart 1987; Wu 1993; alleles on the X chromosomes precludes their pairing Leiserson et al. 1994; Hendrickson and Sakonju 1995; with the yellow allele on the Y (Geyer et al. 1990). Hopmann et al. 1995; Dernburg et al. 1996; Golic and In this article, we use the method of Lewis (1954) to Golic 1996b; Gubb et al. 1997; Fung et al. 1998; Gemkow demonstrate formally that yellow enhancer action in et al. 1998; Sipos et al. 1998; Coulthard et al. 2005). For trans depends on homolog pairing and, in doing so, example, if homolog pairing is mediated by pairing sites further elucidate the chromosomal requirements for and if pairing near such sites is difficult to disrupt, then pairing at the yellow locus. We find that yellow transvec- a gene located in the vicinity of a pairing site would be tion is associated with a small critical region; only chro- expected to have a small critical region. The sizes of mosomal breaks very close to the yellow locus are able critical regions may also reflect gene- or cell-specific to disrupt complementation. Our data also suggest aspects of pairing or transvection. For example, if pair- that transvection at the natural location of yellow can ing is enhanced by an increased amount of time in be disrupted by competitive pairing interactions in the which homologous regions have to find each other, then presence of duplications of the tip of the X chromosome the pairing of genes which are expressed in cells that are containing the yellow locus. Taken together, our findings dividing slowly, if at all, would be expected to be more suggest that somatic pairing of yellow, as assayed by tolerant of chromosome rearrangements, again leading transvection-mediated complementation, requires on to small critical regions. A small critical region would the order of 700 or less kbp of flanking homology. also be expected for genes for which the effects of pair- ing can be imprinted or for which complementation requires only minimal or transient homolog interac- MATERIALS AND METHODS tions (Sabl and Laird 1992; Wu 1993). Our studies use transvection at the yellow (y) gene of Drosophila stocks: The y2, y1#8, y82f29, and y1 alleles have been eyer orris Drosophila to address the chromosomal requirements described elsewhere (G et al. 1990; M et al. 1998, 1999a). Df(1) refers to Df(1) y ac w1118, which carries a for somatic pairing. The yellow gene, located at one end deficiency of the yellow and achaete genes and a mutation in of the X chromosome, is ideal for studies of transvection the white gene (Geyer and Corces 1987). Stocks carrying because it is structurally simple and has an easily assayed P[enh] or P[pro] were provided by P. Geyer (Chen et al. Somatic Homologue Pairing in Drosophila 485

2002) and made isogenic for the chromosome carrying the transvection-disrupting rearrangements or trans-acting domi- transgene. nant suppressors of the z1 phenotype (reviewed in Pirrotta Culture conditions: Flies were cultured at 25° 6 1° on 1991). standard Drosophila cornmeal, yeast, sugar, and agar medium Screen D differed from screens A, B, and C in that re- with p-hydroxybenzoic acid methyl ester added as a mold arrangements were induced on a chromosome carrying y1#8 inhibitor. In general, three females were mated to three males, instead of y2. Males of the genotype y1#8/Y were irradiated with and crosses were transferred every 1–2 days to avoid crowding. 4500 rad of g-rays and crossed to y2 w1118/y2 w1118 females, where Pigmentation scores: Flies were immobilized by chilling in a w1118 was used to distinguish y2 w1118/y2 w1118 (white-eyed) lightly vial submerged in an ice-water bath and examined on a cold pigmented females from y1#8*/y2 w1118 (red-eyed) noncomple- u owe stage as previously described (W and H 1995). Pigmen- menting females arising from nondisjunction. Approximately tation was scored in 1- to 3-day-old flies on a scale of 1–5, where 20,000 y1#8*/y2 w1118 female progeny were screened. 1 represents the null or nearly null pigmentation level, and 5 We carried out two types of screens to determine whether represents the wild-type or nearly wild-type level. At least 30 transvection at ectopic loci is also associated with a small criti- flies of the relevant genotype from each of two different cal region. In the first type, Df(1)/Y; P[enh]47A/P[enh]47A crosses were scored to determine each phenotypic score. and Df(1)/Y; P[enh]92A/P[enh]92A males homozygous for Mutageneses: Four screens (A–D) were carried out to de- an enhancerless yellow transgene were irradiated with 4500 termine the critical region for yellow transvection at its natural rad of g-rays and crossed to Df(1)/Df(1); P[pro]47A/CyO and location. In screen A, y2/Y males were aged 2–4 days at 25° 6 Df(1)/Df(1); P[pro]92A/TM3 females, respectively, carrying a 1°, irradiated with 4000 or 4185 rad of X rays using a Phillips promoterless yellow transgene. Flies were transferred to fresh X-ray machine at a dose of 465 rad/min and mated to y1#8/y1#8 bottles daily for 4 days, after which the males were discarded. females in bottles. Flies were transferred to fresh bottles daily for 4 days, after which the males were discarded. Approxi- Approximately 10,000 and 26,000 flies carrying the 47A and mately 54,000 y2*/y1#8 female progeny were screened. Screen B 92A transgenes, respectively, were screened. followed a similar protocol, except that the y2/Y males were The second type of screen was similar, except that we used irradiated with 4500 rad of g-rays from a cesium-137 source at a X rays instead of g-rays and transvection at Ubx, manifested dose of 350 rad/min. Approximately 24,000 y2*/y1#8 female as wing-to-haltere transformation, as a positive control for the progeny were screened. generation of rearrangements. In this screen, Df(1)/Y; Cbx Ubx 3 Screen C allowed us to screen simultaneously for rearrange- gl /P[enh ]92A males were irradiated with 4500 rad of X rays ments that disrupt transvection at yellow (located at polytene and crossed to Df(1)/Df(1); P[pro ]92A/TM3 females. Flies chromosome position 1B1–2 on the X chromosome; http:// were again transferred to bottles and brooded daily for 4 days, flybase.org), those that disrupt transvection at white (located at after which the males were discarded. Progeny flies carrying polytene position 3C1–2 proximal to yellow; http://flybase. P[enh]92A and P[pro]92A were screened for reduced yellow org), or those that disrupt transvection at both loci. Trans- complementation, and progeny flies carrying Cbx Ubx but vection at white can be observed in a zeste1 (z1) background, not TM3 were screened for reduced wing-to-haltere trans- which does not affect unpaired white genes but causes paired formation. Approximately 15,000 flies of each genotype were white genes to be repressed, producing flies with eyes that screened. are yellow in color instead of wild-type red ( Jack and Judd For all screens, exceptional flies were backcrossed to con- 1979; reviewed by Pirrotta 1991). Interestingly, mutations at firm the original phenotype, after which mutagenized chro- zeste have also been shown to disrupt transvection-associated mosomes believed to carry informative changes were isolated phenotypes at dpp (Gelbart and Wu 1982), eya (Leiserson and put into stock using standard genetic techniques. et al. 1994), and Ubx (Lewis 1954). Although we have not Southern analyses: Genomic DNA for Southern analyses observed an effect of zeste mutations on yellow transvection was isolated from adult flies as described by Ashburner (J.Morris, unpublished results), a disruptive effect has been (1989) or as described by Rehm (2003). GeneScreen hybrid- reported by Geyer et al. (1990). Therefore, incorporation of ization membranes were used according to the recommenda- the z1 mutation into screen C had the potential of revealing a tion of the manufacturer (NEN), and probes were labeled with critical region that would differ from those obtained in screens the Roche random-primed DNA labeling kit. To determine A and B. the integrity of the yellow locus, genomic DNA was digested in In screen C, y2 z1/Y males were irradiated with 4500 rad two separate reactions, one using HindIII and BamHI and the 1#8 1 1#8 1 of g-rays and crossed to y z /y z females. Approximately other using PstI. The digested DNA was separated on a 0.8% 2 1 1#8 1 17,000 (y z )*/y z female progeny were screened for agarose gel and probed with a 7.7-kbp Sal I fragment encom- reduced pigmentation of the wing and/or body caused by passing the entire yellow gene, including the upstream wing disruption of yellow complementation, as well as for red eyes 9 1 and body enhancers and a region 3 of the transcribed region. caused by disruption of z -mediated repression of white.We Cytology and new orders: Salivary glands of third instar found no females with a phenotype consistent with the dis- larvae heterozygous for the rearranged y2 chromosome and an ruption of transvection at both yellow and white. In contrast, we otherwise structurally normal chromosome bearing y1#8 were found 6 yellow-eyed females with reduced levels of pigmenta- dissected in 45% acetic acid and stained with 2% orcein in tion in both the wing and the body (Table 1). Of these 6, 2 were results equal parts lactic acid and glacial acetic acid. Polytene sterile and the remaining 4 are described in the .We chromosomes were analyzed with phase optics. also found 16 females showing normal levels of complemen- Tentative new orders of YTDs are: tation but red eyes. Of these, 10 were sterile or had too few progeny with which to establish lines. Four of the 6 remaining YTD1: 1A–1B j 78A–100; 61–77D j 1C–20 females produced white-eyed y2 z1*/Y progeny, indicating that YTD2: *1C–1A j (100A7–100C1) j 1D–20; 61–100A7 j 100C1– the white gene of the mutagenized chromosome had been 100F disrupted. This interpretation is consistent with the red-eyed (The asterisk indicates an unknown telomeric sequence.) phenotype of y2 z1*/y1#8 z1 females being due to a structural YTD3: 1A–1B j 40F/41B–36D j 5D–1C j 40F/41B–60; 21–36B j disruption or deletion of white that can prevent z1-mediated 5E–20 repression. The remaining 2 red-eyed females produced red- YTD4: 1A–1B j 28D–21D j 11C–1C j 28F–60; 21A–21C j 99D–61; eyed y2 z1*/Y progeny and may have resulted from white 100–99D j 11D–20 486 S. A. Ou et al.

YTD5:1A1B j 80B98F j 80A61; 21–39E j 1C20; 60–39F j 98F100.

Genomic distances: The estimate of 650 or less kbp for the critical region of yellow was determined by considering the yellow transvection-disrupting rearrangement (YTD) break- points falling between the polytene location of yellow at 1B1– 2 and the polytene subdivisions 1B–D (Table 3), assignment of yellow and the band in polytene subdivision 1D that is most distant from yellow, 1D4, to sequences 240,542–255,278 and 884,615–884,935, respectively (http://flybase.org), and then by determining the distance between yellow and 1D4 in terms of base pairs. The estimates of the sizes of Dp(1;2)y2A, Dp(1;4)y2B, and Dp(1;4)y2C to be #700 kbp, 2.5 Mbp, and 1.1 Mbp, respectively, were determined by considering the break- points of these duplications in polytene subdivisions 1B–C, 3A, and 1E, respectively, and the assignment of the bands in each of these regions that are most distant from yellow to 700,000, 2,500,000, and 1,100,000 bp on the genome map, respectively (http://flybase.org).

RESULTS Our study involved the generation and characteriza- tion of chromosomal rearrangements that disrupt com- plementation between the y2 and y1#8 alleles (Figure 1A). The y2 allele is caused by the insertion of a gypsy retrotransposon, which carries a chromatin insulator that blocks the wing and body enhancers from interact- ing with the promoter, and therefore flies homozygous or hemizygous for y2 are lightly pigmented specifically in Figure 1.—Transvection at yellow at endogenous (A) and 2 the wings and body (Harrison et al. 1989). The y1#8 ectopic (B) locations. (A) The y enhancers are blocked from the promoter by a gypsy insulator. y1#8 has a deletion of the allele is a true null. It is caused by a deletion of the pro- promoter. y2/y1#8 flies show nearly wild-type wing and body moter region, and flies homozygous or hemizygous for pigmentation due to the action of the wing and body en- y1#8 show completely mutant yellow pigmentation in all hancers in trans on the intact promoter. W, wing enhancer; tissues. Importantly, although neither y2 nor y1#8 is able B, body enhancer; Br, bristle enhancer. (B) P[enh] and P[pro] are derived from P[yellowCF] (Chen et al. 2002). to direct significant pigmentation of the wings or body, 2 1#8 P[enh ] is a deletion of the wing and body enhancers. P[pro ] y /y flies show nearly wild-type pigmentation levels has a deletion removing the promoter. Open triangles, loxP 1#8 due to action of the y wing and body enhancers on the sites used to generate the P[enh]; striped triangles, FRT sites y2 promoter in trans (Geyer et al. 1990; Morris et al. used to generate P[pro]. Brackets indicate deleted sequences. 1999a). Conveniently, because the insulator of y2 does 47A is on the right arm the second chromosome; 92A is on not prevent the bristle enhancer from interacting with the right arm of the third chromosome. Shaded boxes, P[enh]; open boxes, P[pro]. Figure is not to scale. its promoter, the bristles of y2 flies are darkly pigmented and can be used in our crosses as a phenotypic marker for the y2 allele. score of (4, 4), where the first number refers to the level Transvection at yellow is associated with a small of wing pigmentation and the second to body pigmen- critical region: Modeling our experiments on those tation. In our screens, we selected exceptional y2/y1#8 carried out by Lewis (1954) at Ubx, we began with the progeny whose pigmentation scores were less than (4, 4). idea that yellow complementation depends on pairing of In general, because yellow is on the X chromosome, only the yellow alleles (Geyer et al. 1990). We then tested this female flies can support yellow complementation. Male idea by generating mutations that disrupt complemen- flies, with only a single X, usually do not show comple- tation and, subsequently, by determining whether the mentation, but are able to do so when they carry an mutations are, in fact, chromosomal rearrangements appropriate duplication of yellow (Geyer et al. 1990) or that have breakpoints on the X chromosome. To score when complementation occurs between two yellow trans- wing and body pigmentation, we used a five-point scale genes that are located at allelic positions on an with 1 representing the null or nearly null phenotype autosome (Chen et al. 2002). and 5 representing the wild-type or nearly wild-type Four screens were carried out (Table 1). In screens A phenotype. According to this scale, y2/y2 flies and y1#8/ and B, y2/Y males were irradiated with X rays and g-rays, y1#8 flies give mutant scores of (1, 11) and (1, 1), respectively, to induce chromosomal rearrangements respectively, while y2/y1#8 flies give a nearly wild-type and then mated to y1#8/y1#8 females. In screen C, y2 z1/Y Somatic Homologue Pairing in Drosophila 487

TABLE 1 Summary of screens for transvection-disrupting rearrangements at yellow

Screen A Screen B Screen C Screen D Type of irradiation X ray g-ray g-ray g-ray y2* y2* y2 z1* y1#8* Genotype of F femalesa 1 y1#8 y1#8 y1#8 z1 y2 w1118 No. of females screened 54,000 24,000 17,000 20,000 Exceptional femalesb 23 5 6 7 Sterile 9 2 2 4 Fertile 14 3 4 3 Darkly pigmented malesc 28 4 13 18 a The mutagenized chromosome is indicated by an asterisk. b Females with lighter wing and/or body pigmentation. c Sterile males with dark wing and body pigmentation. males were irradiated with g-rays and crossed to y1#8 z1/ mosome, as revealed by polytene chromosome spreads, y1#8 z1 females. Finally, in screen D, y1#8/Y males were and unaltered y2 alleles, as demonstrated by Southern irradiated with g-rays and then mated to y2 w1118/y2 w1118 analyses (Table 2; Figure 2). These lines reduce pig- females (see materials and methods for further de- mentation from a score of (4, 4) to as low as (1, 1) in our scription of all four screens). complementation assays (Table 3). Significantly, the On the basis of the frequencies with which transvection- rearrangements in these lines all have a breakpoint in disrupting rearrangements have been recovered for the 1B–D polytene region very close to yellow located at other loci, we anticipated that if yellow were associated 1B1–2 and 110 kbp from the telomere of the X chro- with a large critical region, we should recover YTDs at mosome (Figure 2). In this way, these rearrangements a frequency of two to eight for every 1000 females. To appear to define a region in cis to yellow that is important ensure that our screen would be informative even if we for transvection at yellow. As the simplest interpretation failed to recover disruptive rearrangements, we aimed of these rearrangements is that they disrupt yellow com- to screen tens of thousands of females as well as confirm plementation by disrupting pairing of the yellow geno- the effectiveness of irradiation through the isolation of mic region, these X-linked rearrangements were other types of mutations. We screened 54,000 females designated as YTDs (YTD1–5; Table 3). It should be in screen A, 24,000 females in screen B, and 17,000 noted that YTD3 and YTD5 place yellow in the vicinity females in screen C and obtained a total of 34 excep- of centric heterochromatin, making it possible that tional females with lighter wing and/or body pigmen- the reduced complementation associated with these two tation (Table 1; screen D is discussed below). Importantly, rearrangements is due at least in part to position-effect these females were able to produce dark bristles, indi- variegation (see new orders for these chromosomes in cating that the coding region of the y2 allele on the materials and methods). Consistent with this inter- mutagenized chromosome was functional. Thirteen of pretation, YTD5 is associated with a variegated pattern the exceptional females were sterile or gave too few of bristle pigmentation (Table 3). progeny to allow further studies. Lines were established We placed 13 lines into classes II, III, and IV (Table 2). from the remaining 21 females. Class II consists of 5 lines (2 from screen A, 2 from screen The mutagenized X chromosomes of all lines were B, and 1 from screen C), all of which have a cytologically then analyzed in two ways. First, we determined whether visible rearrangement breakpoint within the 1B–C re- the X or any other chromosome had been rearranged gion of the X but are also altered in the yellow genomic through the cytological analyses of larval polytene chro- region (Table 4). For example, PCR analysis showed that mosomes (Figure 2). Second, we determined the the X chromosome of line 10 has a breakpoint within structural integrity of the yellow locus by carrying out the 250-bp 39 untranslated region (39-UTR) of yellow Southern analysis using genomic DNA and a probe (see materials and methods). The location of chro- covering the entire yellow locus (data not shown). These mosomal disruptions so close to yellow in these class II lines cytological and structural analyses placed the 21 lines precluded our ability to determine whether these lines into five classes (Table 2). The 18 lines falling into compromise complementation by disrupting pairing or classes I–IV are discussed directly below, while the three by disrupting the integrity of the yellow gene itself. lines of class V are described in the final section of Class III consists of five lines (all from screen A) that results. have no obvious chromosomal rearrangements but give Class I is defined by five lines (all from screen A) that abnormal Southern patterns (Table 2). As was the case carry chromosomal rearrangements involving the X chro- with the class II lines, these alterations of the yellow ge- 488 S. A. Ou et al.

polytene spreads or to mutations at the yellow locus that are not detectable by Southern analyses. It is also possible that one or more of the class IV lines harbor trans-acting mutations in genes that are important for pairing and/or transvection at yellow. Our fourth screen, screen D, was designed to de- termine whether disruption of yellow complementation is sensitive to the particular allele of yellow carried on the rearranged chromosome. This screen resembled screens A, B, and C in that it was designed to recover disruptors of complementation between y2 and y1#8, but differed from them in that it targeted the y1#8 rather than the y2 chromosome for irradiation. Approximately 20,000 progeny females were screened for light wing and/or body pigmentation, and 7 exceptional females were recovered (Table 1). Four of these females were sterile or gave too few progeny for further studies. The re- maining 3 were used to generate mutant lines, each of which was subsequently shown by Southern analysis to be structurally altered in the yellow genomic region. Although we did not characterize these lines by cytology, our Southern analyses indicate that they fall into either class II or class III and, like the other lines in these classes, may be unable to support complementation due to disruption of the yellow locus. Considering all four screens together, it may be noteworthy that all five YTDs were found in screen A, the only screen in which males were irradiated with X rays instead of g-rays. However, as g-rays may be some- what less mutagenic than X rays (Ashburner 1989), it may not be surprising that, given our small sample size of five YTDs, we did not recover YTDs using g-rays. This view is consistent with the observation of Lewis (1954), who found g-rays less effective than X rays at generating transvection-disrupting rearrangements, even though g-rays are capable of generating such rearrangements (Smolik-Utlaut and Gelbart 1987; Hendrickson and Sakonju 1995; Gubb et al. 1997). One way to confirm the effectiveness of both X rays and g-rays in our studies is to assess the recovery of other Figure 2.—Locations of chromosomal breaks. (A) Poly- classes of exceptions, especially those associated with tene sections 1A-1F are labeled on the image of the tip of aura gross chromosomal rearrangements. In particular, Ta- the X chromosome (S et al. 1997). Locations of break- ble 2 shows that class II mutants were recovered in points of the five YTDs as determined by cytology are indi- cated below the figure. (B) Images of unpairing at the tip screen A, which used X rays, as well as in screens B and of the X chromosomes due to chromosomal rearrangements C, which used g-rays. Another class of exceptions that in the five YTDs. For reference, the prominent band 1B3 is may serve as a measure for the efficacy of X rays and indicated on the nonrearranged X chromosome. g-rays is summarized in Table 1. These exceptions were isolated as darkly pigmented sterile males and may have resulted from irradiation-induced rearrangements that nomic region obscured our ability to determine whether attached the tip of the irradiated X chromosome, loss of complementation resulted from disruption yellow carrying y2 or y1#8, to an autosome, followed by loss of pairing. the remainder of the X. In this scenario, dark pigmen- Class IV includes three lines (two from screen A and tation would result from complementation between y2 one from screen C) that are structurally normal by both or y1#8 carried on the translocated X tip and y1#8 or y2, cytological examination and Southern analyses. The respectively, on the maternally derived X chromosome, inability of these lines to support complementation may producing XO flies, which have male characteristics but be due to rearrangements that are not detectable in are sterile due to lack of a Y chromosome. Alternatively, Somatic Homologue Pairing in Drosophila 489

TABLE 2 Classes of mutant lines generated in screens A, B, and C

Class Polytene Southern Screen A Screen B Screen C I Rearranged Normal 5 0 0 II Rearranged Rearranged 2 2 1 III Normal Rearranged 5 0 0 IV Normal Normal 2 0 1 V Duplication NA 0 1 2 Total 14 3 4 Mutant lines were characterized and placed into four classes (I–IV) on the basis of the cytological appearance of their X chromosomes, as determined by polytene analysis, and the integrity of the yellow gene, as determined by Southern analysis. Class V lines carry a duplication of the tip of the X chromosome. Pigmentation scores of y2*/y1#8 class I and II females are provided in Tables 3 and 4. y2*/y1#8 pigmentation scores for the five class III females are (1–2, 2), (4, 2), (4, 2), (2, 2), and (1, 1), and for the three class IV females are (3, 2), (2, 1), and (2, 3). these exceptional males may have been caused by large less than that observed for loci associated with large irradiation-induced intrachromosomal deletions of the critical regions and indicates that the critical region for X, which would produce free duplications of the X yellow transvection is small, on the order of 650 kbp or carrying y2 or y1#8, noted as Dp(1;f)y2 or Dp(1;f)y1#8. Here, less (materials and methods). One explanation for dark pigmentation would be explained by pairing of the the small size of the yellow critical region may be that yellow allele on the Dp(1;f)y2 or Dp(1;f)y1#8 chromosome pairing of yellow alleles depends on a pairing site within with the yellow allele on the maternal X, the male 650 kbp of yellow. In this view, only rearrangements characteristics and sterility resulting from the deletion that break between yellow and the pairing site would be of X material and the absence of a Y chromosome. able to disrupt transvection. Alternatively, proximity to The frequencies with which class II mutants and a telomere may enhance pairing (Pandita et al. 2007), darkly pigmented sterile males were generated can be and, if so, the small critical region of yellow may be the used to determine the relative effectiveness of X rays consequence of the telomeric location of yellow. Finally, (screen A) and g-rays (screens B, C, and D) in our ex- our data indicate that yellow alleles may require up to periments. Screen A produced 2 class II mutants among 650 kbp of contiguous flanking homology to achieve 54,000 females (0.004%) while screens B and C together the level of somatic pairing necessary for an observable produced 3 class II mutants among 41,000 females degree of complementation. (0.007%), and screen A produced 28 dark sterile males Transvection at yellow transgenes may also be asso- (0.05%) while screens B, C, and D together produced 35 ciated with a small critical region: We tested whether such males among 61,000 females (0.06%). On the basis the yellow critical region reflects the presence of a pair- of these observations, the frequencies of generating ing site, proximity to a telomere, or the requirement for rearrangements in our screens are comparable whether a certain amount of contiguous flanking homology by X rays or g-rays are used. applying the Lewis (1954) protocol to yellow transgenes To summarize, our screens of 115,000 females yielded that support transvection at ectopic sites in the genome five lines that belong to class I and are YTDs (Tables 1–3; (Chen et al. 2002). We reasoned that if the critical region Figure 2). This frequency for the generation of putative for yellow is determined by the distance between yellow transvection-disrupting rearrangements (0.004%) is much and the nearest pairing site, then the critical regions for

TABLE 3 Class I lines: YTDs

Name Cytology Pigmentation scores (W, B) YTD1 T(1;3)1B-C;77D-78A (1, 1) YTD2 In(1)1A;1C-D 1 Tp(3;1)100A7;100C1;1C-D (2, 2) YTD3 T(1;2)1B-C;5D-E;36B-D;40F-41B (2, 1) YTD4 T(1;2;3)1B-C;11C-D;21C-D;28D-F;99D (3, 1) YTD5 T(1;2;3)1B-C;39E-F;80A-B;98F (1, 1)a Each of these five YTDs carries a structurally normal y2 gene and a rearrangement breakpoint on the X chro- mosome near yellow. (W, B) refers to the wing and body pigmentation of females on a scale of 1–5, with 5 being wild type or nearly wild type (see materials and methods). Pigmentation scores refer to YTD/y1#8 flies. a The bristle pigmentation of these genotypes is variegated, ranging from dark to light, and is consistent with an intact yellow gene whose expression is not uniform. 490 S. A. Ou et al.

TABLE 4 ing levels of pigmentation—(31, 4) and (4, 4), respecti- Class II lines vely—while flies carrying just the P[enh ] transgene at these two sites produced relatively low levels of pigmen- 1 Pigmentation tation— (2, 2 ) and (2, 2), respectively (Chen et al. 2002). Name Cytology scores (W, B) To find rearrangements that disrupt transvection, males carrying P[enh] at 47A or 92A and a deletion of Line 6 T(1;4)1B-C;102D-F (4, 1) Line 7 In(1)1B-C;9B;20 (1, 1) the natural yellow gene, Df(1), were irradiated with g-rays Line 8 T(1;2)1B-C;11-12;20;51 (1, 1) and crossed to Df(1)/Df(1) females carrying P[pro ] Line 9 T(1;2;3)1B-C;40-41A;98C-E (1, 3)a at 47A or 92A, respectively, and the resulting progeny Line 10 T(1;2)1B-C;28C (1, 1) were screened for flies with a level of wing and body pigmentation lower than that expected of a comple- Each of these five class II lines carries a structurally disrup- materials and methods ted yellow gene as well as a breakpoint on the X chromosome menting genotype (see for a near yellow. (W, B) refers to the wing and body pigmentation full description of Df(1) and screen). We chose to irra- of females on a scale of 1–5, with 5 being wild type or nearly diate flies carrying P[enh] rather than P[pro] because materials and methods wild type (see ). Lines 6 and 7 are P[enh] retains the bristle enhancer and, by providing a from screen A, lines 8 and 9 are from screen B, and line 10 is from screen C. dark bristle phenotype, provided a convenient marker a The bristle pigmentation of these genotypes is variegated, with which we could follow the irradiated chromosomes ranging from dark to light, and is consistent with an intact yel- in our crosses. Our strategy also permitted us to screen low gene whose expression is not uniform. for disruptive rearrangements in both males and females as the location of the transgenes on the autosomes the transgenes would likely differ in size from each permitted complementation in both sexes. other and from that of yellow at its natural location, We screened 10,000 and 26,000 flies carrying the 47A unless the transgene insertion sites were each within and 92A transgenes, respectively, and found no dark- 650 bp of a pairing site. Alternatively, if the critical bristled noncomplementing flies. We believe that g-- region for yellow at its natural location is determined irradiation was effective in our screens because we did by proximity to a telomere, then the critical regions recover several flies lacking dark bristle pigmentation, of transgenes located far from telomeres should be indicating that the P[enh] transgene had likely been considerably larger and differ from each other. Finally, if disrupted. Furthermore, because our screen involving the critical region for yellow reflects a requirement for the 92A transgenes made use of the TM3 third chromo- a minimum amount of flanking homology, then the some balancer, which carries a recessive mutation of critical regions of transgenes should be small, resem- the ebony gene, we were able to recover four apparent bling that of yellow at its natural location. mutations of ebony. These findings indicate that the fre- The transgenes that we used were generated in an quency of the recovery of transvection-disrupting rear- earlier study from a construct that had been inserted rangements for yellow transgenes at 47A and 92A, as into eight ectopic sites (Chen et al. 2002). This construct determined by our complementation assay, is on the carried a yellow gene that had been modified by target order of ,0.01% (1/10,000) and 0.004% (1/26,000), sites for the Cre and FLP recombinases such that respectively, and predict that the critical regions for separate expression of the recombinases produced two these transgenes is small, as is the critical region for derivatives at each insertion site, one lacking the wing yellow at its natural location. and body enhancers (P[enh]) and another lacking the As no transvection-disrupting rearrangements were promoter (P[pro]) (Figure 1B; Chen et al. 2002). Im- found in the screens using P[enh] and P[pro] trans- portantly, complementation was observed between pairs genes at 47A and 92A, we did a subsequent screen using of allelic P[enh] and P[pro] derivatives at all eight X rays instead of g-rays because X rays were the mutagen ectopic sites, indicating that the genome is generally used in screen A, where YTDs were obtained. This permissive for yellow transvection. screen was also designed to simultaneously recover In our screens for transvection-disrupting rearrange- rearrangements that disrupt transvection at Ubx, giving ments, we chose two pairs of complementing transgenes, us a positive control for the generation of rearrange- one located in the middle of the right arm of chromo- ments. Specifically, we used Cbx Ubx transvection, where some 2 at polytene position 47A and a second located Cbx Ubx/11flies show wing-to-haltere transformation in the middle of the right arm of chromosome 3 at due to the action of the gain-of-function Cbx mutation polytene position 92A (Figure 1B). Transgene pairs at on the wild-type Ubx1 allele in trans; rearrangements these two sites, instead of the other six sites, were chosen can disrupt this interaction, resulting in flies with wings because they supported relatively good viability and, showing less wing-to-haltere transformation (reviewed furthermore, were predicted to give a significant differ- in Duncan 2002). In this screen, we irradiated Df(1)/Y ence in pigmentation between complementing and non- males carrying Cbx Ubx on one third chromosome complementing flies. Specifically, the transgenes at and the enhancerless transgene at 92A on the other positions 47A and 92A produced strong complement- with X rays. These flies were then crossed to Df(1)/Df(1) Somatic Homologue Pairing in Drosophila 491 females carrying the promoterless transgene at 92A. TABLE 5 In this way, we were able to screen for disruption Bithorax transvection-disrupting rearrangements of transvection at yellow by looking for flies that carried the enhancerless transgene and the promoterless Name Location of breakpoint on 3R transgene but showed reduced wing and/or body pigmentation and, simultaneously, screen for disrup- BTD1 82A–B BTD2 87E–F tion of transvection at Ubx by looking for flies that BTD3 84E carried Cbx Ubx but showed reduced wing-to-haltere BTD4 82C–D transformation. BTD5 88A We screened 15,000 flies for reduced cuticular pig- BTD6 82F mentation and another 15,000 for reduced wing- BTD7 83C to-haltere transformation. We found 2 flies showing Seven lines of flies showing reduced wing-to-haltere trans- reduced levels of pigmentation, but they did not trans- formation were analyzed by cytology. The locations of the mit this phenotype. We also found 92 flies showing breakpoint on the right arm of chromosome 3 are indicated reduced wing-to-haltere transformation, giving a 0.6% above. Other breakpoints are not indicated. BTD, bithorax (92/15,000) frequency of recovery, comparable to that transvection-disrupting rearrangement. found in other studies (Lewis 1954). Seven of these were analyzed cytologically, and each was found to have independently of the X chromosome, and cytological a rearrangement between the centromere and Ubx on studies confirmed the presence of duplications (Figure 3). the right arm of chromosome 3 (Table 5). Importantly, the ability of these duplications to affect Our ability to generate rearrangements that disrupt complementation suggested that the y2 duplication can transvection at Ubx but none that disrupt transvection interact with yellow at its natural position on the X chro- between ectopic yellow transgenes suggests that the X mosome. We therefore used these three duplications as rays effectively generated rearrangements and that a tool to better understand how yellow genes pair. the critical region for yellow transvection at ectopic sites One duplication, called Dp(1;2)y2A, reduces the level is small, similar to that for the endogenous yellow gene. of complementation normally observed with y2/y1#8 Taken together, our results are most consistent with females from a score of (4, 4) to (3, 3) (Table 6). It yellow transvection, as determined by complementing consists of the tip of the X chromosome through levels of pigmentation, requiring a minimum amount of polytene subdivision 1B–C attached to the tip of the contiguous flanking homology. right arm of the second chromosome and is #700 kbp in Duplications of y2 can affect yellow complementa- size (materials and methods). We reasoned that the tion: During the course of these studies, we recovered reduction in complementation caused by Dp(1;2)y2A three duplications of the tip of the X chromosome might be due to its ability to disrupt the productive containing y2 (Table 2; class V). The presence of a du- interaction of y2 and y1#8 at the natural yellow location, plication was recognized through crosses of exceptional permitting y2 carried by Dp(1;2)y2A to pair with either the y2*/y1#8 females to y1#8/Y males that produced apparent y2 or the y1#8 allele on the X, being productive only in the complementing male progeny as well as three classes latter case (Figure 4, A and C). of female progeny: complementing, partially comple- To further test whether pairing of Dp(1;2)y2A with y1#8 menting, and noncomplementing. These four classes can be productive with respect to pigmentation, we of progeny indicated that a y2 allele was segregating determined the effect of Dp(1;2)y2A on the homozygous

Figure 3.—(A) Sizes of duplications. Dp(1;2)y2A, Dp(1;4)y2B, and Dp(1;4)y2C are dupli- cations of the tip of the X chromosome that are able to complement a yellow allele on the non- rearranged X chromosome. Breakpoints are indi- cated above the image of the tip of the X chromosome (Bridges 1938). The duplications are #700 kbp (breakpoint at 1B–C), #2.5 Mbp (breakpoint at 3A), and #1.1 Mbp (breakpoint at 1E) in size, respectively. (B) Images of duplica- tions. Dp(1;2)y2A is attached to the tip of the right arm of the second chromosome, while Dp(1;4)y2B and Dp(1;4)y2C are attached to the fourth chro- mosome. 492 S. A. Ou et al.

TABLE 6 Pairing competition in the presence of y2 duplications

Pigmentation Duplication Genotype scores (W, B) None y2/y2 (1, 11) y1#8/y1#8 (1, 1) y2/y1#8 (4, 4) y1/y1#8 (1, 1) Df(1)/y1#8 (1, 1) y1#8/Y (1, 1) Dp(1;2)y2Ay1#8/y1#8; 1/Dp(1;2)y2A (2, 2) y2/y1#8 ; 1/Dp(1;2)y2A (3, 3) y1/y1#8; 1/Dp(1;2)y2A (2, 2) Df(1)/y1#8 ; 1/Dp(1;2)y2A (4, 31) y1#8/Y; 1/Dp(1;2)y2A Dark pigmentationa y1#8/Y; Dp(1;2)y2A/ Light pigmentationa Dp(1;2)y2A Dp(1;4)y2By1#8/y1#8; 1/Dp(1;4)y2B (3, 3) y2/y1#8; 1/Dp(1;4)y2B (2, 1) Dp(1;4)y2Cy1#8/y1#8; 1/Dp(1;4)y2C (3, 3) y2/y1#8; 1/Dp(1;4)y2C (2, 1) a Pigmentation scale is used only to score female wing and body pigmentation. The Df(1)y ac w1118 chromosome is indi- Figure 4.—Model for pairing competition in the presence cated as Df(1). (W, B) refers to the wing and body pigmenta- of duplications of y2. Shaded diamonds represent y2 and open tion of females on a scale of 1–5, with 5 being wild type or diamonds represent y1#8. Arrow thickness corresponds to nearly wild type (see materials and methods). strength of pairing interactions. Numbers indicate pigmenta- tion scores of the wing and body, respectively. (A) In the ab- sence of a duplication, y2 and y1#8 in their normal location on y1#8/y1#8 as well as the heterozygous y1/y1#8 genotypes. In the X chromosome are able to pair and give a pigmentation 1#8 contrast to y2/y1#8, neither of these genotypes supports score of (4, 4). (B) In the absence of a duplication, two y alleles in their normal location pair but do not complement, complementation, giving us an opportunity to detect giving a pigmentation score of (1, 1). (C) The presence of 2 2 productive interactions between the y of Dp(1;2)y A Dp(1;2)y2A disrupts the pairing of y2 and y1#8 on the X chromo- and y1#8 as increases in pigmentation. Consistent with some, reducing the pigmentation score to (3, 3). (D) The our model, we found that the presence of Dp(1;2)y2A presence of Dp(1;2)y2A disrupts the pairing of the two non- 1#8 1#8 1#8 complementing y alleles on the X chromosome, permitting increased the pigmentation scores of both y /y and 2 2 1#8 1 1#8 the pairing between the y allele of Dp(1;2)y A and y and, y /y flies from (1, 1) to (2, 2) (Table 6; Figure 4, B therefore, complementation corresponding to a pigmentation and D). score of (2, 2). (E) The presence of Dp(1;4)y2B or Dp(1;4)y2C, Finally, we determined the effect of Dp(1;2)y2A when it which appear to be stronger competitors of pairing compared is not competing with another yellow allele in pairing to Dp(1;2)y2A, disrupts the pairing of y2 and y1#8 on the X chro- interactions by placing it in females carrying a yellow allele mosome to a relatively high degree, reducing the pigmentation score to (2, 1). (F) The presence of Dp(1;4)y2B or Dp(1;4)y2C at the natural yellow locus in trans to Df(1) or in males disrupts the pairing of the two noncomplementing y1#8 alleles carrying only a single X chromosome that are therefore on the X chromosome, resulting in a relatively high level of hemizygous for the yellow locus. Remarkably, Dp(1;2)y2A pairing between the y2 allele of the duplications and y1#8 increases the pigmentation score of Df(1)/y1#8 flies from and, therefore, a relatively high level of complementation, cor- (1, 1) to (4, 31). Similarly, male y1#8/Y flies carrying responding to a pigmentation score of (3, 3). Dp(1;2)y2A also show dark pigmentation. Interestingly, however, male y1#8/Y flies with two copies of Dp(1;2)y2A have light pigmentation (Table 6). These results suggest X chromosome through polytene subdivision 3A to the that, when present in only one copy, a y2 allele carried by fourth chromosome at polytene subdivision 102F, and Dp(1;2)y2A can pair well with an unpaired yellow allele on Dp(1;4)y2C attaches the tip of the X through polytene the X and compete effectively for paired alleles on the X. subdivision 1E to the fourth chromosome at 102C. On However, when homozygous, Dp(1;2)y2A becomes much the basis of the Drosophila genome sequence, Dp(1;4)y2B less effective in pairing with X-linked yellow alleles. and Dp(1;4)y2C are #2.5 Mbp and #1.1 Mbp in size, The two other duplications carrying y2, called Dp(1;4)y2B respectively (see materials and methods). and Dp(1;4)y2C, were also tested for their ability to affect We find that the pigmentation scores of y2/y1#8 flies complementation. Both these duplications are attached are reduced from (4, 4) to (2, 1) by both Dp(1;4)y2B and to the fourth chromosome and are much larger than Dp(1;4)y2C (Table 6; Figure 4, A and E). This reduction Dp(1;2)y2A (Figure 3). Dp(1;4)y2B attaches the tip of the in pigmentation is in contrast to the higher score of (3, Somatic Homologue Pairing in Drosophila 493

3) produced by y2/y1#8 females carrying Dp(1;2)y2A, from each other when tested for their impact on yellow suggesting that Dp(1;4)y2B and Dp(1;4)y2C are relatively transvection, suggesting that duplication location and/ more effective at disrupting pairing at the natural yellow or size can influence the extent to which duplications locus. Consistent with this interpretation, we find that compete for pairing partners. These findings are in the pigmentation of y1#8/y1#8 flies is increased from (1, agreementwiththoseofGeyer et al. (1990), who studied 1) to (3, 3) in the presence of Dp(1;4)y2B or Dp(1;4)y2C, a duplication of the yellow genomic region attaching y2 to in contrast to the lower (2, 2) score of y1#8/y1#8 flies the Y chromosome. carrying Dp(1;2)y2A (Table 6; Figure 4, B and F). The critical regions for yellow are not consistent with Although our data rest on just three duplications, they a role for pairing sites: The small size of the critical suggest that duplication location and/or size can in- region that we found for transvection at the natural fluence pairing competition. More specifically, the location of yellow was reinforced by our studies using greater ability of Dp(1;4)y2B and Dp(1;4)y2C compared yellow transgenes. Although we found no rearrange- to Dp(1;2)y2A to compete for pairing with alleles on the X ments that disrupted transvection at the two ectopic may be the result of their attachment to the fourth locations tested, the number of flies screened—10,000 rather than the second chromosome or of the larger for transgenes at 47A and 41,000 for transgenes at sizes of Dp(1;4)y2B and Dp(1;4)y2C. 92A—is informative. In particular, for genes with large critical regions spanning up to a one-third of a chro- mosome arm or more, the percentage of irradiated DISCUSSION chromosomes found to carry transvection-disrupting Our studies were aimed toward a better understand- rearrangements is comparatively large, ranging from ing of homolog pairing at yellow. Using a technique 0.2% to 0.8% or more when chromosomes are treated devised by Lewis (1954), we show that yellow transvec- with 4000–4500 rad of X rays (Lewis 1954; Gelbart tion, as assayed by intragenic complementation between 1982; Leiserson et al. 1994; Sipos et al. 1998). Thus, if y2 and y1#8, is dependent on pairing. In addition, we find transvection at the two ectopic locations of yellow were that this transvection is extremely difficult to disrupt associated with large critical regions, we should have by chromosomal rearrangements: only five YTDs were recovered on the order of 20–801 YTDs for transvection recovered from 115,000 female flies screened in our at 47A and on the order of 82–3281 YTDs for trans- studies. As these rearrangements all have at least one vection at 92A. Instead, we failed to recover a YTD for breakpoint close to yellow, our data further suggest that transvection at these ectopic locations, even though we the critical region for yellow is small. We assume from obtained disruptors of transvection at Ubx at a frequency this that all rearrangements that broke the X chromo- of 0.6%. These data suggest that the critical regions for some further from yellow allowed enough pairing to ectopic yellow transvection are small and similar in size to produce flies with complementing levels of pigmenta- that for transvection at the natural yellow locus. tion and therefore were not recovered in our screens. Our findings argue against the yellow critical region Cytological analysis showed that all five YTDs had being determined by proximity to a telomere or by breakpoints less than 650 kbp from yellow (Figure 2; distance from a nearby hypothetical pairing site, as both materials and methods). Because the chance of transgene pairs were located far from any telomere having all five breakpoints fall, at random, within a and it is unlikely that both would have coincidentally 650-kbp interval is 1 3 109 [(650 kbp/40 Mbp)5], inserted very near pairing sites. Instead, our data are assuming that the X chromosome contains approxi- consistent with yellow requiring #700 kbp of contiguous mately one-fifth of the genome, or 40 Mbp (Bosco flanking homology to support a level of pairing-mediated et al. 2007), we believe that 650 kbp is a good estimate complementation that can be detected in our visual of the critical region for yellow transvection as assayed by assays. It is possible that the location of the 47A and 92A complementation between y2 and y1#8. transgenes in the middle of chromosomal arms contrib- In addition to the five YTDs, we also recovered three uted to our inability to recover transvection-disrupting duplications of the tip of X chromosome containing the rearrangements because a mid-chromosome location y2 allele (Figure 3). These duplications were identified would, a priori, permit pairing to be established or by their ability to disrupt complementation at the nat- stabilized either proximally or distally to the transgene. ural yellow locus and are likely exerting their influence In this situation, disruption of transgene complemen- by acting as competitive pairing partners (Figure 4). As tation could require that the transgenes be flanked by the smallest of these duplications is #700 kbp in size, two chromosomal breakpoints, each within 650 kbp of it appears that this amount of the yellow genomic region the yellow transgene. Such events would be very rare. is sufficient to support a level of homology searching Our analysis of yellow transgenes is consistent with a and pairing that permits yellow enhancer action in trans. previous study demonstrating that the size of the critical This interpretation is consistent with our estimate of region associated with white gene transvection also does 650 kbp as an upper limit for the size of the yellow not depend on the chromosomal location of white critical region. Interestingly, these duplications differ (Smolik-Utlaut and Gelbart 1987). Transvection at 494 S. A. Ou et al. white occurs in a z1 background, where the mutant zeste Finally, the small size of some critical regions may protein leads to suppression of the paired w1 genes. reflect a means by which two alleles become paired and/ Using insertional translocations that placed a w1 gene or are maintained in a paired state as a consequence, within the large dpp critical region, Smolik-Utlaut at least in part, of their being individually localized to a and Gelbart (1987) found that rearrangements that specific nuclear address or nuclear compartment, such had breakpoints proximal to both white and dpp and as a transcription factory. This potential mechanism that disrupted dpp transvection did not disrupt white of homolog association has been proposed for both transvection. A similar outcome was found using a z1-mediated repression of white (Davidson et al. 1985; P-element insertion of w1 in the large critical region Wu 1993; Cook 1997; Xu and Cook 2008) and PcG- for Ubx. That is, the critical region for white was small mediated silencing (Mateos-Langerak and Cavalli regardless of its location within the large critical region 2008). Here, we suggest that the smaller size of some of another gene, suggesting that pairing of white,as critical regions may reflect a correspondingly larger assayed by the z1 phenotype, is not determined by a contribution of this mechanism to pairing and trans- pairing site (Smolik-Utlaut and Gelbart 1987). In- vection. If the element being targeted to a nuclear stead, these observations, together with our findings for location lies close to the gene, then only breakpoints yellow, suggest that the critical regions of at least some falling near enough to the gene to separate it from such genes may reflect gene-specific features, such as the an element would be able to disrupt transvection. On tissue in which a gene is expressed and the cell cycle the other hand, if the targeted element is the gene itself, length of that tissue, the degree or amount of pairing then it may be that transvection can be disrupted only by that is required for a gene to respond to pairing, the rearrangement breakpoints that are near enough to the capacity of a gene to remember the paired state, and/or gene to overcome the localization of the gene to the the strength of the interaction between the promoter of correct nuclear compartment. Alternatively, a rearrange- a gene and its enhancers (for a more in-depth review, ment could introduce a foreign element that pulls the see Duncan 2002; also see Lewis 1954; Gelbart 1982; gene to another nuclear location. Note that, because Smolik-Utlaut and Gelbart 1987; Wu 1993; Leiserson some nuclear compartments are formed by the coales- et al. 1994; Hendrickson and Sakonju 1995; Hopmann cence of multiple chromosomal regions, a gene targeted et al. 1995; Dernburg et al. 1996; Golic and Golic to such a compartment may have numerous opportuni- 1996b; Gubb et al. 1997; Fung et al. 1998; Gemkow et al. ties to enter the compartment, making transvection at 1998; Sipos et al. 1998; McKee et al. 2000; Coulthard such a gene relatively more permissive of rearrangements et al. 2005). For example, the critical region of yellow may and, therefore, more likely to be associated with a small be small because the yellow wing and body enhancers critical region. In contrast, a gene targeted to only a may be especially capable of long-distance, productive, single nuclear address would have limited opportunities and/or prolonged interactions. to colocalize with its homolog, would be less tolerant of Another determinant of the size of a critical region rearrangements, and would have a larger critical region. may be the sensitivity of a specific transvection-mediated In light of these arguments, nuclear compartmentaliza- phenotype to transcript levels (reviewed by Duncan tion in organisms where extensive homolog pairing has 2002; also see Smolik-Utlaut and Gelbart 1987; not been observed may be the equivalent of somatic Sipos et al. 1998; Morris et al. 1999b). For example, a pairing (Wu 1993; Cook 1997; Xu and Cook 2008). phenotype that can be achieved with low transcript levels In addition to the different sizes of critical regions, may remain unaltered by a pairing-disruptive rearrange- researchers have observed a propensity of the chro- ment if the rearrangement does not completely abolish mosomal breakpoints that disrupt transvection to fall pairing. In this case, a low but productive amount of proximal to the gene (reviewed in Duncan 2002; pairing resulting from transient homolog interactions Kennison and Southworth 2002). Again, arguments in each cell or from the cumulative amount of pairing that consider chromosomal positioning may be relevant. across a cell population may be sufficient to produce a For example, a proximal location of the breakpoints phenotype that approximates that of a fully paired would be expected if colocalization of homologous genotype. In contrast, a phenotype that requires high chromosomes to the same nuclear territory facilitates transcript levels may be sensitive to all rearrangement homology searching or the targeting of alleles to the breakpoints that reduce gene expression, even if by just same nuclear position. Rearrangements that reattach a a small amount, and consequently be associated with a gene to another centromere could antagonize pairing large critical region. Transvection at yellow may be partic- by placing that gene in a nuclear territory that is ularly difficult to disrupt because wild-type pigmenta- different from that of its homolog, as has also been tion can be achieved with only 33% of the normal suggested by Coulthard et al. (2005). In light of this transcript level, as assayed in whole organisms (Morris interpretation, it is noteworthy that the y2 alleles of et al. 2004; Lee et al. 2006), which limits visual assays to Dp(1;2)y2A, Dp(1;2)y2B, and Dp(1;2)y2C seem able to pair the recovery of only those breakpoints close enough to with a yellow allele on the X chromosome even though yellow to reduce transcript levels below this threshold. they are attached to autosomal centromeres. Such Somatic Homologue Pairing in Drosophila 495 pairing may reflect the ability of telomeric sequences to Chen, J. L., K. L. Huisinga,M.M.Viering,S.A.Ou, C.-t.Wu et al., more easily find their homologs, as compared to 2002 Enhancer action in trans is permitted throughout the Drosophila genome. Proc. Natl. Acad. Sci. USA 99: 3723–3728. sequences located more centrally in chromosomes, Cook, P. R., 1997 The transcriptional basis of chromosomal pairing. because of the greater capacity of telomeres to traverse J. Cell Sci. 110: 1033–1040. oulthard olan ell illiker nuclear space and/or because the Rabl configuration of C , A. B., N. N ,J.B.B and A. J. H , 2005 Transvection at the vestigial locus of Drosophila melanogast- chromosomes after cell division promotes the interac- er. Genetics 170: 1711–1721. tion of telomeres (Pandita et al. 2007). Csink, A. K., and S. Henikoff, 1998 Large-scale chromosomal The proximal location of rearrangement breakpoints movements during interphase progression in Drosophila. J. Cell Biol. 143: 13–22. has also been interpreted to indicate that pairing prog- Davidson, D., C. H. Chapman,C.Wedeen and P. M. Bingham, resses in a centromere-to-telomere direction, with the 1985 Genetic and physical studies of a portion of the white proximal endpoint of the critical region marking the locus participating in transcriptional regulation and in synapsis- ewis dependent interactions in Drosophila adult tissues. Genetics 110: location of a pairing site (L 1954). In contrast, 479–494. fluorescence in situ hybridization (FISH) analyses of Dernburg, A. F., K. W. Broman,J.C.Fung,W.F.Marshall, developing embryos indicate that pairing initiates at J. Philips et al., 1996 Perturbation of nuclear architecture by long-distance chromosome interactions. Cell 85: 745–759. multiple sites along a chromosome in random order Duncan, I. W., 2002 Transvection effects in Drosophila. Annu. Rev. (Csink and Henikoff 1998; Fung et al. 1998; Gemkow Genet. 36: 521–556. et al. 1998; Vazquez et al. 2001). This difference in in- Fung, J. C., W. F. Marshall,A.Dernburg,D.A.Agard and J. W. Sedat, 1998 pairing in Drosophila terpretation may result from differences between the melanogaster proceeds through multiple independent initiations. experimental approaches, the first relying on pheno- J. Cell Biol. 141: 5–20. typic outcomes in the adult fly of changes in gene Gelbart, W. M., 1982 Synapsis-dependent allelic complementation at the decapentaplegic gene complex in Drosophila melanogaster. expression caused by chromosomal rearrangements, Proc. Natl. Acad. Sci. USA 79: 2636–2640. and the second using direct labeling of chromosomes to Gelbart, W. M., and C.-t.Wu, 1982 Interactions of zeste mutations reveal the initiation of pairing in embryonic nuclei. with loci exhibiting transvection effects in Drosophila melanogaster. Alternatively, it may indicate that the level of pairing Genetics 102: 179–189. Gemkow,M.J.,P.J.Verveer andD.J.Arndt-Jovin, 1998 Homologous necessary for transvection requires multiple steps and association of the bithorax-complex during embryogenesis: conse- that different experimental approaches address differ- quences for transvection in Drosophila melanogaster. Development ent steps. For example, pairing may involve homology 125: 4541–4552. Geyer, P. K., and V. G. Corces, 1987 Separate regulatory elements sensing followed by homolog alignment, pairing initia- are responsible for the complex pattern of tissue-specific and de- tion, pairing establishment, pairing maintenance, and velopmental transcription of the yellow locus in Drosophila mela- perhaps even a refinement or intensification of pairing nogaster. Genes Dev. 1: 996–1004. Geyer, P. K., M. M. Green and V. G. Corces, 1990 Tissue-specific prior to the event of transvection. If so, FISH analyses transcriptional enhancers may act in trans on the gene located of embryonic nuclei may be addressing pairing initia- in the homologous chromosome: the molecular basis of transvec- tion while rearrangement studies may be defining the tion in Drosophila. EMBO J. 9: 2247–2256. Golic, K. G., and M. M. Golic, 1996a Engineering the Drosophila parameters required for later stages of pairing. In this genome: chromosome rearrangements by design. Genetics 144: light, the proximal location of rearrangement break- 1693–1711. points for some genes may correspond to a step of pair- Golic, M. M., and K. G. Golic, 1996b A quantitative measure of the mitotic pairing of alleles in Drosophila melanogaster and the influ- ing, perhaps that of establishment, that progresses in a ence of structural heterozygosity. Genetics 143: 385–400. centromere-to-telomere direction. Gubb, D., J. Roote,J.Trenear,D.Coulson and M. Ashburner, 1997 Topological constraints on transvection between white We thank P. Geyer for many years of close collaboration and fly genes within the transposing element TE35B in Drosophila mela- stocks; J. Bateman for suggesting the simultaneous screen for nogaster. Genetics 146: 919–937. disruption of yellow and Ubx transvection; the Wu lab for insightful Harrison, D. A., P. K. Geyer,C.Spana and V. G. Corces, 1989 The discussion and support; W. Bender, J. Lee, N. Perrimon, R. Pruitt, L. gypsy retrotransposon of Drosophila melanogaster: mechanisms of Raftery, and F. Winston for thoughtful input; A. Moran for generous mutagenesis and interaction with the suppressor of Hairy-wing technical assistance; and the Bloomington Stock Center for fly stocks. locus. Dev. Genet. 10: 239–248. This study was supported by grants GM085169 and GM61936 from the Hartl, T. A., H. F. Smith and G. Bosco, 2008 Chromosome align- National Institutes of Health to C.-t.W. and by DUE-0736995 from the ment and transvection are antagonized by condensin II. Science National Science Foundation to J.R.M. and S. Lovett. 322: 1384–1387. Hendrickson, J. E., and S. Sakonju, 1995 Cis and trans interactions between the iab regulatory regions and abdominal-A and abdominal- LITERATURE CITED B in Drosophila melanogaster. Genetics 139: 835–848. Hiraoka, Y., A. F. Dernburg,S.J.Parmelee,M.C.Rykowski,D.A. Ashburner, M., 1989 Drosophila: A Laboratory Manual. Cold Spring Agard et al., 1993 The onset of homologous chromosome pair- Harbor Laboratory Press, Cold Spring Harbor, NY. ing during Drosophila melanogaster embryogenesis. J. Cell Biol. Bantignies,F,C.Grimaud,S.Lavrov,M.Gabut and G. Cavalli, 120: 591–600. 2003 Inheritance of Polycomb-dependent chromosomal inter- Hopmann, R., D. Duncan and I. Duncan, 1995 Transvection in the actions in Drosophila. Genes Dev. 17: 2406–2420. iab-5,6,7 region of the bithorax complex of Drosophila: homol- Bosco, G., P. Campbell,J.T.Leiva-Neto and T. A. Markow, ogy independent interactions in trans. Genetics 139: 815–833. 2007 Analysis of Drosophila species genome size and satellite Jack, J., and B. Judd, 1979 Allelic pairing and gene regulation: a DNA content reveals significant differences among strains as well model for the zeste-white interaction in Drosophila melanogaster. as between species. Genetics 177: 1277–1290. Proc. Natl. Acad. Sci. USA 76: 1368–1372. Bridges, C. B., 1938 A revised map of the salivary gland X-chromosome Kennison, J. A., and J. W. Southworth, 2002 Transvection in Dro- of Drosophila melanogaster. J. Hered. 29: 11–13. sophila. Adv. Genet. 46: 399–420. 496 S. A. Ou et al.

Kopczynski, C. C., and M. A. Muskavitch, 1992 Introns excised Rehm, E. J., 2003 Inverse PCR and cycle sequencing of P element from the Delta primary transcript are localized near sites of Delta insertions for STS generation. Berkeley Drosophila Genome Pro- transcription. J. Cell Biol. 119: 503–512. ject. http://www.fruitfly.org/about/methods/inverse.pcr.html Lee, A. M., and C.-t.Wu, 2006 Enhancer-promoter communica- Ronshaugen,M.,andM.Levine, 2004 Visualization of trans-homolog tion at the yellow gene of Drosophila melanogaster:diversepro- enhancer-promoter interactions at the Abd-B Hox locus in the moters participate in and regulate trans interactions. Genetics Drosophila embryo. Dev. Cell 7: 925–932. 174: 1867–1880. Sabl, J. F., and C. D. Laird, 1992 Epigene conversion: a proposal Leiserson, W. M., N. M. Bonini and S. Benzer, 1994 Transvection with implications for gene mapping in humans. Am. J. Hum. at the eyes absent gene of Drosophila. Genetics 138: 1171–1179. Genet. 50: 1171–1177. Lewis, E. B., 1954 The theory and application of a new method of Sass, G. L., and S. Henikoff, 1999 Pairing-dependent mislocaliza- detecting chromosomal rearrangements in Drosphila melanogaster. tion of a Drosophila brown gene reporter to a heterochromatic Am. Nat. 88: 225–239. environment. Genetics 152: 595–604. Martin, M., Y. B. Meng and W. Chia, 1989 Regulatory elements in- Saura, A. O., A. J. Saura and V. Sorsa, 1997 Electron micrograph volved in the tissue-specific expression of the yellow gene of Dro- maps of Drosophila melanogaster polytene chromosomes. http:// sophila. Mol. Gen. Genet. 218: 118–126. www.helsinki.fi/saura/EM/xdiv1.html. Mateos-Langerak,J.,andG.Cavalli, 2008 Polycomb group Sipos, L., J. Mihaly,F.Karch,P.Schedl,J.Gausz et al., proteins and long-range gene regulation. Adv. Genet. 61: 45– 1998 Transvection in the Drosophila Abd-B domain: extensive 66. upstream sequences are involved in anchoring distant cis-regulatory McKee, B. D., 2004 Homologous pairing and chromosome dynam- regions to the promoter. Genetics 149: 1031–1050. ics in meiosis and mitosis. Biochim. Biophys. Acta 1677: 165– Smolik-Utlaut, S. M., and W. M. Gelbart, 1987 The effects of 180. chromosomal rearrangements on the zeste-white interaction in McKee, B. D., C. S. Hong and S. Das, 2000 On the roles of hetero- Drosophila melanogaster. Genetics 116: 285–298. chromatin and euchromatin in meiosis in Drosophila: mapping Stevens, N., 1908 A study of the germ cells of certain Diptera with chromosomal pairing sites and testing candidate mutations for reference to the heterochromosomes and the phenomena of syn- effects on X-Y nondisjunction and meiotic drive in male meiosis. apsis. J. Exp. Zool. 5: 359–374. Genetica 109: 77–93. Vazquez, J., A. S. Belmont and J. W. Sedat, 2001 Multiple regimes Metz, C. W., 1916 Chromosome studies on the Diptera. II. The of constrained chromosome motion are regulated in the inter- paired association of chromosomes in the Diptera, and its signif- phase Drosophila nucleus. Curr. Biol. 11: 1227–1239. icance. J. Exp. Zool. 21: 213–279. Vazquez, J., A. S. Belmont and J. W. Sedat, 2002 The dynamics of Morris, J. R., J. L. Chen,P.K.Geyer and C.-t. Wu, 1998 Two modes homologous chromosome pairing during male Drosophila mei- of transvection: enhancer action in trans and bypass of a chro- osis. Curr. Biol. 12: 1473–1483. matin insulator in cis. Proc. Natl. Acad. Sci. USA 95: 10740– Vazquez, J., M. Mu¨ller,V.Pirrotta and J. W. Sedat, 2006 The 10745. Mcp element mediates stable long-range chromosome-chromosome Morris, J. R., J. Chen,S.T.Filandrinos,R.C.Dunn,R.Fisk et al., interactions in Drosophila. Mol. Biol. Cell 17: 2158–2165. 1999a An analysis of transvection at the yellow locus of Drosophila Williams, B. R., J. R. Bateman,N.D.Novikov and C.-t.Wu, melanogaster. Genetics 151: 633–651. 2007 Disruption of topoisomerase II perturbs pairing in Dro- Morris, J. R., P. K. Geyer and C.-t. Wu, 1999b Core promoter sophila cell culture. Genetics 177: 31–46. elements can regulate transcription on a separate chromosome Wu, C.-t., 1993 Transvection, nuclear structure, and chromatin pro- in trans. Genes Dev. 13: 253–258. teins. J. Cell. Biol. 120: 587–590. Morris, J. R., D. A. Petrov,A.M.Lee and C.-t. Wu, 2004 Enhancer Wu, C.-t., and M. Howe, 1995 A genetic analysis of the Suppressor 2 choice in cis and in trans in Drosophila melanogaster: role of the pro- of zeste complex of Drosophila melanogaster. Genetics 140: 139–181. moter. Genetics 167: 1739–1747. Wu,C.-t., and J. R. Morris, 1999 Transvection and other homology Pandita, T. K., C. R. Hunt,G.G.Sharma and Q. Yang, 2007 Regu- effects. Curr. Opin. Genet. Dev. 9: 237–246. lation of telomere movement by telomere chromatin structure. Xu, M., and P. R. Cook, 2008 The role of specialized transcription Cell. Mol. Life Sci. 64: 131–138. factories in chromosome pairing. Biochim. Biophys. Acta 1783: Pirrotta, V., 1991 The genetics and molecular biology of zeste in 2155–2160. Drosophila melanogaster. Adv. Genet. 29: 301–348. Pirrotta, V., 1999 Transvection and chromosomal trans-interaction effects. Biochim. Biophys. Acta 1424: M1–M8. Communicating editor: T. C. Kaufman