Terminal Retrotransposons Activate a Subtelomeric white Transgene at the 2L Telomere in Drosophila

Corresponding author: James M. Mason, Laboratory of Molecular Genetics, National Institute of Environmental Health Sciences, 111 Alexander Dr., P.O. Box 12233, Research Triangle Park, NC 27709-2233., masonj{at}niehs.nih.gov (E-mail)

Communicating editor: S. HENIKOFF

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetically marked P elements inserted into the subtelomeric satellites of Drosophila show repression and variegation of the reporter gene. One such white⁺ reporter, inserted between the subtelomeric satellite and the terminal HeT-A array in the left arm of chromosome 2 (2L), is sensitive to its context; changes in the structure of the telomere region can be identified by changes in eye color. Addition of HeT-A or TART elements to the 2L terminus increases w⁺ expression, and loss of sequence from the end decreases expression. This indicates that the telomeric retrotransposons in Drosophila have an activating influence on the repressed subterminal reporter gene. Changes in eye color due to altered expression of the transgene also allow the detection of interactions between homologous telomeres. The 2L arms that terminate in long HeT-A/TART arrays showed increased expression of the subterminal w⁺ transgene when the terminal repeats on the homologue are absent or markedly shorter. We propose that the chromatin structure of the terminal HeT-A/TART array and the activity of a putative promoter/enhancer element on HeT-A are affected by telomeric interactions. Such trans-activation may reflect control over HeT-A transcription and, thus, transposition activity.

TELOMERES are nucleoprotein structures on the ends of linear chromosomes. They are important for chromosome stability, nuclear architecture, and certain chromosome movements. DNA polymerase cannot replicate a linear chromosome completely. A special mechanism therefore is required to maintain chromosome ends, which in most organisms is accomplished by telomerase, a reverse transcriptase with an internal RNA template (BLACKBURN 1992 ). The length of the telomeric repeat tracts is influenced by a variety of factors either acting on telomerase or on the structure of the telomere itself (GREIDER 1996 ). In the yeast Saccharomyces cerevisiae telomeres are packaged into heterochromatic structures that repress genes placed in their immediate vicinity (ZAKIAN 1996 ; GOTTA et al. 1997 ; GRUNSTEIN 1997 ). Moreover, the length of telomeric repeats is under continuous control and is responsive to levels of a number of different proteins (SHORE 1997 ). Several studies suggest that the processes of telomere length maintenance and telomere silencing in yeast are tightly connected, as demonstrated by the observation that an extended telomeric tract causes increased silencing of a subterminal reporter gene (KYRION et al. 1993 ; PARK and LUSTIG 2000 ). Rap1p binds to double-stranded telomeric DNA (reviewed by SHORE 1994 ) and regulates genetic silencing at telomeres as well as telomere length. Formation of heterochromatic chromatin structures is initiated by Rap1p, which recruits the silencing proteins Sir3p and Sir4p (reviewed by LUSTIG 1998 ; STONE and PILLUS 1998 ).

Telomere length regulation and telomeric silencing are less well understood in Drosophila melanogaster, and very little is known about the cis- and trans-acting factors controlling these processes. The DNA structure of telomeres in D. melanogaster differs significantly from that of yeast and most other eukaryotes (reviewed by MASON and BIESSMANN 1995 ). Drosophila telomeres do not possess arrays of simple telomeric repeats, which could bind proteins like Rap1p. Instead, Drosophila maintains its telomeres by transposition of specific retrotransposons, HeT-A and TART, to chromosome ends. Proximal to the terminal retrotransposon array Drosophila telomeres carry several kilobases of complex satellites, referred to as telomere-associated sequences (TAS), which exhibit sequence similarities among themselves (KARPEN and SPRADLING 1992 ; WALTER et al. 1995 ) and structural similarities to TAS in other eukaryotes (PRYDE et al. 1997 ). Reporter genes exhibit suppressed and variegated expression when inserted into the telomeric region (GEHRING et al. 1984 ; HAZELRIGG et al. 1984 ; LEVIS et al. 1985 ; KARPEN and SPRADLING 1992 ; TOWER et al. 1993 ; ROSEMAN et al. 1995 ; WALLRATH and ELGIN 1995 ; MASON et al. 2000 ), and molecular analyses have shown that these repressed transgenes have inserted adjacent to TAS or within a TAS array (KARPEN and SPRADLING 1992 ; LEVIS et al. 1993 ; CRYDERMAN et al. 1999 ). This suggests that some regions of the Drosophila telomeres are heterochromatic and that TAS plays a role in telomeric silencing, as recently demonstrated directly using a transgenic approach (KURENOVA et al. 1998 ). The genetic control of silencing at telomeres [telomeric position effect (TPE)] differs from the control of classical centromeric position effect variegation (PEV) because, with the exception of certain alleles of the Polycomb-group alleles of Su(z)2 and Psc (CRYDERMAN et al. 1999 ), TPE does not respond to the action of PEV modifiers or developmental silencers (TALBERT et al. 1994 ; ROSEMAN et al. 1995 ; WALLRATH and ELGIN 1995 ; CRYDERMAN et al. 1999 ), suggesting a different molecular mechanism and/or a different set of silencing proteins.

P{w^var} is an insertion of a genomic copy of the white gene into the tip of the left arm of chromosome 2 (2L; GEHRING et al. 1984 ). We report here that it can serve as a sensitive reporter to study telomere dynamics. Individuals with this transgene have orange eyes with a few dark spots, but phenotypic variants arise in the germline with high frequency under some conditions. Molecular analysis of these variants revealed that transgene expression is correlated with alterations in the telomere region. Higher expression is associated with HeT-A and TART additions onto the marked tip and with lower expression with terminal deficiencies. Analysis of the eye color variants described here provides new insight into the nature of telomeric silencing and the regulation of gene activity in telomeric regions. Moreover, genetic interactions were found between the marked telomere and the homologous telomere. We present a new model for telomeric position effect in Drosophila and discuss the contributions of cis- and trans-acting DNA sequences. The studies of P{w^var} described here are an early step in the dissection of the genetic control of telomere stability and maintenance in Drosophila.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila crosses:
Drosophila stocks were maintained and crosses were performed at 25° on cornmeal and molasses medium with dry yeast added to the surface. The w¹; P{w^var} stock (GEHRING et al. 1984 ) was kindly provided by W. Gehring. Other genetic markers and special chromosomes are described by LINDSLEY and ZIMM 1992 . The l(2)gl terminal deficiencies used here have been described in detail more recently (WALTER et al. 1995 ). Eye colors were determined in young individuals, and males were aged for 3 days past eclosion to be photographed. Photographs were taken using a Nikon SMZ-U stereo-microscope with the diaphragm half open.

Except as noted, variants arose spontaneously in one of two stocks, y¹ w^67c23; P{w^var} and y¹ w^67c23; P{w^var} al, or in the crosses involving one of these. Homozygous variant stocks were constructed after crossing to y¹ w^67c23; Sco/SM1.

Genomic DNA isolation, field inversion gel electrophoresis, and Southern blotting and hybridizations:
Genomic DNA from various strains was isolated and used for Southern blot hybridizations as described previously (WALTER et al. 1995 ). Hybridizations were done at 68° for 12–16 hr with [³²P]dCTP (New England Nuclear, Boston) random-primed DNA probes in 5x SSC, 50 mM Tris-HCl (pH 7.6), 10x Denhardt's solution (0.2% each of Ficoll, polyvinyl pyrrolidone, and bovine serum albumin), 0.1% SDS, 2.5 mM EDTA, and 0.1 mg/ml sheared and denatured salmon sperm DNA. Filters were washed at 50° in 0.1x SSC, 0.1% SDS for 2 hr and dried. DNA fragments to be used as hybridization probes were purified by gel electrophoresis in low-melt agarose, and 20–50 ng were used for labeling by random priming reaction (Prime-it II, Stratagene, La Jolla, CA). Field inversion gel electrophoresis (FIGE) was performed as described (WALTER et al. 1995 ) in 1% agarose gels in 0.5x TBE at 14° with a Bio-Rad FIGE mapper system, using the built-in program No. 2 (switch time ramp 0.1–0.8 sec, linear shape, forward voltage 180, reverse voltage 120) to obtain optimal separation in the 1- to 50-kb range.

In situ hybridization:
Larval salivary glands were dissected in Ringers and squashed in 45% acetic acid. The chromosomes were rehydrated through a decreasing ethanol series, washed briefly in 2x SSC and incubated in 2x SSC at 80° for 30 min. Chromosomes were denatured in 0.1 N NaOH for 90 sec, followed by a wash in 2x SSC, dehydrated in an increasing ethanol series, and air dried. Hybridization with random-primed digoxigenin-labeled DNA fragments (Boehringer Mannheim, Indianapolis) was carried out at 55° overnight in 5x SSC, 0.02% SDS, 5 mM EDTA, 20 mM Tris HCl, pH 7.6. After several washes in PBS, samples were incubated at 37° with fluorescein-conjugated antidigoxigenin Fab fragments in PBS with 1 mg/ml bovine serum albumin. Preparations were again washed with three changes of PBS, and chromosomes were stained for 2 min in 15 µM propidium iodide and washed in PBS. As probe for HeT-A we used the 5-kb EcoRI fragment from the element 9D4, which contains the entire HeT-A element except for the last 1 kb at the 3' end (BIESSMANN et al. 1992A ). As probe for TART we used the 2.2-kb SacI fragment located on the TART element between nucleotides 434 and 2683 (LEVIS et al. 1993 ). To mark 2L, we included in the hybridization reaction a probe from the dpp gene, which hybridizes to cytological position 22F1-2.

Recombinant phage libraries:
Recombinant phage libraries were generated by ligating 1 µg of SalI-cut genomic DNA of the brown-red variants or the red variant #11-5 with 1 µg of SalI-cut DASH phage (Stratagene). Recombinant phage DNA was packaged with Gigapack II Plus (Stratagene) and plated for screening with random primed ³²P-labeled probes on Escherichia coli XL-1 blue (Stratagene). Positive phage were plaque purified and grown in liquid 30-ml cultures for DNA preparation.

DNA amplification by PCR:
PCR reactions with genomic DNA or with recombinant -phage DNA were done using the Taq polymerase from TaKaRa Shuzo Co. (Otsu, Japan). Reactions of 50 µl usually contained 0.5 mg of genomic DNA, 200 µM each of the dNTPs, 1x PCR buffer (supplied by the manufacturer), 30 ng of each primer, and 2.5 units Taq polymerase. PCR reactions were done in a Crocodile III thermocycler (Appligene Oncor) at an annealing temperature of 5–10° below melting temperature of the primers used, with 1–3 min synthesis at 72°, for 35 cycles. Amplified DNA products were either sequenced directly or cloned into pGem-T-easy (Promega, Madison, WI) for further analysis. The following primers (Genosys Biotechnologies) were used. For their approximate positions and orientation on the P{w^var} transgene and the adjacent HeT-A elements, refer to Fig 3. More exact locations are given in parentheses. Primers HeT-1, -2, and -3 were modeled after consensus sequences from near the 3' end of the elements using GenBank accession nos. M84200, M84201, U06947, U06920, and X77049. Primers HeT-L1 and -L3 were synthesized according to the sequences in HeT-A elements 1 and 2, respectively, which are located at the distal end of the P{w^var} chromosome. Nucleotide positions of white gene primers are indicated according to the accession number of the white gene (X02974). The accession number of the 2L TAS sequence is U35404 and of the P element P25.1, X06779:

• HeT-1: 5'CTGTCTCCGTACCTCCACCAGC3'

  
  

  View larger version (60K): In this window In a new window Download PPT slide   Figure 1. Representative eye color phenotypes of homozygous and heterozygous P{wvar} variants. (A) Brown-red variant #1/+. (B) Brown-red #1/Df(2L)net62. (C) White #23/brown-red #38. (D) Pale #39/brown-red #38. (E) Yellow #40/brown-red #38. (F) P{wvar}/brown-red #38. (G) Homozygous brown-red #1. (H) Orange #2/+. (I) Orange #17/Df(2L)net62. (J) P{wvar}/white #23. (K) Pale #41/P{wvar}. (L) Yellow #40/P{wvar}. (M) Homozygous orange #16. (N) Yellow #9/+. (O) Yellow #40/l(2)gl26. (P) Yellow #42/white #23. (Q) Yellow #40/Pale #43. (R) Homozygous yellow #9. (S) Pale #8/+. (T) Pale #8/Df(2L) net62. (U) Homozygous pale #8. (V) White #23/+. All individuals are males marked with y1 w67c23 on the X chromosome.

  
  

  View larger version (15K): In this window In a new window Download PPT slide   Figure 2. Genetic crosses used to generate y1 w67c23; P{wvar} stocks. Different procedures were used to produce two stocks with the same genetic markers. (A) In generation 1 y1 w67c23 females were crossed to w1; P{wvar} males. F1 progeny were then intercrossed, F2 progeny were mated in single pairs, and cultures that bred true were selected. Thus, recombination was allowed between chromosome 2 homologues. (B) In generation 1 y1 w67c23; Sco/SM1, Cy females were crossed to w1; P{wvar} males. Cy F1 progeny were then intercrossed, and Cy+ F2 progeny were mated en masse. Thus, recombination was not allowed between chromosome 2 homologues.

  
  

  View larger version (14K): In this window In a new window Download PPT slide   Figure 3. Molecular map of the P{wvar} transgene insertion. At the proximal side of the insert, the 3' P-element sequences are contiguous with the 2L-specific TAS. The genomic white transgene contains 2.3 kb of its upstream region, including the eye-specific enhancer (E). On the distal side, the 5' P-element sequences of the original construct, as well as part of the upstream w sequences, are deleted. Attached to nucleotide 5184 of the w gene is an array of two truncated HeT-A elements, joined by oligo(A) tails. The end of the chromosome receded with time (see Fig 4C) and, in February 2000, was 2.8 kb distal to the P{wvar} transgene. Approximate locations and direction of oligonucleotide primers used in PCR reactions are indicated by arrowheads. Important restriction sites are shown, and their nucleotide positions are given according to the white gene map (GenBank accession no. X02974). Locations of fragments used as hybridization probes in genomic blots are indicated by bars.

• HeT-2: 5'CCCCAAACTCACCMCATGYAATG3'

• HeT-3: 5'GCTTCCAGCGACTCGGTGCTTCCG3'

• HeT-L1: 5'GGTACGTTGGCTGGGAGGTAAATTGG3'

• HeT-L3: 5'CATTCTTATTGAATTTTCCTTTCATTGCAGG3'

• white-B: 5'GGCAAAGGTAAGTCCCCAGCC3' [nucleotide (nt) 5726–5706]

• white-BR: 5'GGCTGGGGACTTACCTTTGCCG3' (nt 5706–5727)

• white-C: 5'GCATTATCAATCTTCATCATCGG3' (nt 6807–6785)

• white-D: 5'GCTAGGTAACGCTACAAACGGTGG3' (nt 7493–7470)

• 2L-Sat: 5'CATCTCGTTCATCCGCCACC3' (nt 161–142)

• P-3': 5'GCATACGTTAAGTGGATGTCTCTTGCCG3' (nt 2946–2973)

DNA sequencing and sequence analysis:
Sequences of fragments subcloned in pBluescript were determined by automated DNA sequencing using the Prism Ready Reaction DyeDeoxy Terminator kits from Applied Biosystems (Foster City, CA). Samples of 0.5 µg of double-stranded supercoiled plasmid DNA template, prepared by alkaline/SDS minipreparation, were sequenced with 3.2 pmol of primer in 20-µl reactions for 25 cycles, each with 96° for 10 sec, 50° for 5 sec, and 60° for 4 min. The reaction mixtures were then run through individual Sentri Sep columns (Princeton Separations), and fragments were separated and analyzed using an ABI 377 automated fluorescence sequencer. GenBank searches were done with BLAST (ALTSCHUL et al. 1990 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Phenotype and genetic instability of the telomeric P{w^var} transgene:
The variegating P-element insert in the 2L telomere, termed P{w^var}, originated in one of the early Drosophila germline transformation experiments (GEHRING et al. 1984 ). The original P-element construct used for transformation contained a 12-kb genomic fragment of the w gene (accession number of the white gene is X02974), including the transcribed region (nt 7460–13428) and 2.9 kb of upstream sequences beginning at the EcoRI site at nt 4587, which contains the eye-specific enhancer (between nt 6300 and 5500; LEVIS et al. 1985 ; PIRROTTA et al. 1985 ). The telomeric P{w^var} insertion arose during remobilization of a euchromatic integrant from cytological position 82A to the tip of 2L. The eye color phenotype in the original w¹; P{w^var} strain was orange, similar in color to w^apricot, with a few small red spots (see Fig 1). A similar phenotype was seen when the transgene was homozygous or heterozygous with a standard sequence chromosome. Such behavior is unusual; with the exception of another insert of a full-length w⁺ gene, P{w⁺, ry⁺}A4-4 (LEVIS et al. 1985 ), other w⁺ (mini-white) variegating telomeric inserts exhibit much darker eye color in homozygotes than in hemizygotes (WALLRATH and ELGIN 1995 ).

The eye color phenotype of P{w^var} in the original w¹ stock was fairly stable, and variations from the orange eye color were rarely observed. However, after outcrossing (in the autumn of 1996) to substitute genetic markers, individuals with exceptional phenotypes began to appear. Most exceptions had a brown-red eye color when homozygous; rarer variants exhibited white, light orange, or an intermediate eye color when homozygous. Genetic analysis revealed that the phenotypic changes mapped to the second chromosome. The new variants were crossed with y w^67c23; Sco/SM1 to make stocks. Progeny y w^67c23; Sco/variant females from these crosses were also outcrossed to determine the position of the new mutation relative to Sco and the 2L tip. While the number of recombinants from any one variant was small, all variants appear to be genetically inseparable from w⁺ itself. Thus, the exceptional P{w^var} derivatives reflect genetic instability in the vicinity of the insertion and provide a basis for using P{w^var} as a model to study telomere dynamics.

To determine the rate of appearance of brown-red exceptions, brother-sister, single-pair matings were made from stocks with different genetic backgrounds (Table 1). The original strain produced exceptions with a frequency of 1.7 x 10^-4 (0.8 x 10^-4/gamete). The frequency of brown-red exceptions from y w^67c23; P{w^var} was 1.3 x 10^-2 (0.6 x 10^-2/gamete), 75-fold higher. Results from a different set of crosses (data not shown) suggested that a genetic factor on chromosome 2 might be responsible for this difference. We therefore established a new y w^67c23; P{w^var} stock using the SM1 second chromosome balancer to prevent recombination between the P{w^var} chromosome and its homologue (Fig 2B). The frequency of exceptions from this new stock did not differ significantly from that of the original.

Taken together, the exceptions from these two sets of crosses (Table 1, rows 1 and 3) represent five independent events. All of these exceptions carried new variant chromosomes that were homozygous white, as well as the original P{w^var}; none carried a typical brown-red variant chromosome. Thus, there were no brown-red variants among 63,000 tested chromosomes [2 x (17,895 + 13,608)]. Given free recombination on chromosome 2 when y w^67c23; P{w^var} was first established (Fig 2A), the genetic factor from w¹; P{w^var} that decreases the frequency of exceptions may be segregating in y w^67c23; P{w^var}. Thus, the frequency of exceptions measured here (Table 1, row 2) is probably an underestimate. Most of these variants were homozygous brown-red, but a few (2 of 55 independent exceptions) were homozygous white. The variants continued to change, producing secondary and higher-order phenotypic derivatives.

Phenotypes of eye color variants and their genetic interactions:
Phenotypically, the homozygous variants at the lighter end of the spectrum fell into a rough continuum, while variants at the darker end showed discrete phenotypes. For purposes of discussion, we have divided the variants into six classes according to their eye color phenotype as homozygotes: white, pale, yellow, orange, brown-red, and red. As shown in Fig 1 and discussed below, these variant chromosomes interact with each other and with a wild-type or deficient 2L telomere in unexpected ways. While not reported separately here, all variants were, of necessity, crossed to y w^67c23; Sco/SM1 when stocks were made. The Sco and SM1 chromosomes both interacted with the variants in the same way as the wild-type chromosome. Also, in the crosses described below, four short terminal l(2)gl deficiencies were used (9, 26, 258, and 334), as well as the longer Df(2L)net62. Not all variants were crossed with all deficiencies; most were tested with either l(2)gl²⁶ and/or Df(2L)net62, and a few variants of each class were tested with multiple deficiencies. As all these tests gave essentially the same result, they are not reported separately.

White variants have no pigment in the eye and appear to be w null when homozygous or heterozygous with a wild-type 2L telomere. Many white variants were recovered because they induce dark-orange to brown, severely mottled eyes when the homologue carries P{w^var} or a similar orange variant (see below).

Most pale homozygotes have off-white or cream-colored eyes with a few yellow-to-orange spots; some have completely white eyes with spots. In heterozygotes with a wild-type chromosome or a white variant chromosome the eyes have less pigment and may be indistinguishable from white. A few pale variants were recovered from individuals with almost wild-type eyes. These exceptions very rarely appeared in orange eye color stocks. In two such cases studied, the exceptions with red color were heterozygous for a pale variant chromosome and a brown-red variant chromosome.

Yellow variants have yellow eyes with small orange spots when homozygous and pale eyes with yellow spots in heterozygotes with a wild-type 2L telomere or a white variant chromosome. The increase in color may represent an additive dosage effect, although we have not measured pigment directly. This effect of gene dosage clearly distinguishes pale and yellow variants from orange variants, which resemble the original insert in that they display the same eye color in homozygotes and heterozygotes with a wild-type chromosome.

Brown-red eye color variants are the most frequent class. Heterozygotes with a wild-type 2L telomere have the same eye orange color as found in the original strain, but homozygotes have a brown-red, mottled eye color. Moreover, a similar brown-red eye color is expressed whether the homologue carries a yellow, an orange, another brown-red variant, or a complete deletion of the 2L telomere (e.g., l(2)gl²⁶; MECHLER et al. 1985). In heterozygotes with pale or white variant chromosomes the eye color is red, almost wild type. It appears that the dark eye color in these heterozygotes is caused by activation of the telomeric white gene by the telomere structure of the homologue (see below).

In most stocks secondary eye color changes occurred over time, indicating continued instability. As an example, we analyzed here the stock red #11-5. During the isolation of a brown-red variant as described above, an orange chromosome was also isolated, and the orange variant (called #11-5-orange) was kept in stock. After about 2 years a bright red eye color variant appeared in the orange stock, which was isolated and called red #11-5.

In addition to these germline mutations, somatic mosaics appeared among the P{w^var} orange homozygotes with a frequency of 5 x 10^-3 per eye. The mosaic eyes had large, dark, internally mottled sectors (not shown). The character of w expression in these mosaics is clearly different from the fine-grained variegation of the P{w^var} transgene. The appearance of mosaic sectors in homozygous individuals cannot be explained by mitotic recombination, although the size and the typical clonal shape of the sectors suggests that they result from a genetic event. They were not analyzed further.

Molecular structure of P{w^var} and of eye color variants derived from it:
Aided by our previous molecular analysis of the 2L telomeric region (WALTER et al. 1995 ), we first determined the molecular structure of the P{w^var} insertion (Fig 3). The centromere-proximal side was amplified from genomic DNA by PCR with primers P-3' and 2L-Sat and the resulting fragment was sequenced. We found that the 3' inverted repeat of the P element was directly attached to nt 24 of the 2L TAS. It appears that the last 4 bp of the P-element inverted repeat, CATG, are missing. Thus, the element is joined to the 2L TAS at CAG. The sequence of the 0.6-kb PCR product obtained from genomic DNA with primers HeT-1 and white-B identified the distal end of the transgene and revealed that 597 bp of w sequences from the original construct and all of the 5' P-element terminal repeat sequences were missing. A HeT-A element (HeT-A #1) was attached with its oligo(A) tail directly to nt 5184 in the w sequence, leaving the eye-enhancer region (E) intact. Thus, the P{w^var} transgene is oriented such that the reporter is transcribed in the distal-to-proximal direction. Various combinations of primers, indicated by arrowheads in Fig 3 and used in PCR reactions with genomic P{w^var} DNA, produced the predicted sized fragments from the w gene, indicating that no major rearrangements had occurred within the transgene. Moreover, primers HeT-1, -2, and -3, modeled after a HeT-A consensus sequence, at various distances upstream of the oligo(A) tail, in combination with primers white-B or -C, also produced fragments consistent with the map. The sequence of these PCR fragments showed that HeT-A #1 (Fig 3) has the typical structure of HeT-A (BIESSMANN et al. 1992B ; DANILEVSKAYA et al. 1992 ). The exact structure of the HeT-A array at the end of the original P{w^var} chromosome could not be determined, but was deduced from the structure of the brown-red variants derived from it (see below).

The position of the P{w^var} chromosome terminus and of restriction sites within the terminal HeT-A array was determined by Southern blot analyses after cutting genomic DNA with various restriction enzymes that have known sites in the w transgene (Fig 4). Blots were hybridized with several probes from the w transgene (shown in Fig 3). BamHI, HindIII, and SpeI have sites in the HeT-A array, while the apparent distal sites of five other 6-cutter restriction enzymes (PstI, XhoI, SstI, NruI, ClaI, and SalI) occurred very close together within 500 bp, indicating that the chromosome end is positioned 2.8 kb distal to the junction with the w transgene. During the course of our studies, from 1996 to 2000, we realized that the end of the 2L chromosome was receding. This is not surprising, since terminal nucleotide loss of 50–100 bp per sexual generation has previously been documented (BIESSMANN and MASON 1988 ; LEVIS 1989 ). An example of the receding telomere at the P{w^var} chromosome is shown in Fig 4C. XhoI-digested genomic DNA from the original w¹; P{w^var} stock or the outcrossed y w^67c23; P{w^var} stock, both harvested in August 1998 and in February 2000, was hybridized to probe 2a to visualize the size of the terminal restriction fragment. Clearly, the terminal fragment is shortening with time in both stocks and even splits into several polymorphic bands in the most recent isolate of the w¹; P{w^var} stock. Terminal recession occurs at 75 bp/sexual generation, as previously reported (BIESSMANN and MASON 1988 ; LEVIS 1989 ). Together with similar determinations done in 1997, we extrapolate that the length of the HeT-A array was between 5.5 and 6.5 kb in the fall of 1996, when the eye color variants were first detected (see Fig 5).

View larger version (78K): In this window In a new window Download PPT slide   Figure 4. Genomic Southern blots of P{wvar} and its yellow, orange, brown-red, and red eye color variants. DNA from P{wvar} and its derivatives was cut with various restriction enzymes as indicated and hybridized to various DNA probes, as indicated in Fig 3. All stocks are homozygous for the 2L subtelomeric w+ transgene and carry the w67c23 null allele on the X chromosome, except for the original P{wvar} stock, which carries w1. (A) Comparison of P{wvar} and its color derivatives with the w67c23 null allele. DNA was cut with SstI or ClaI and hybridized to probe 2b. Hybridizing fragments from the transgene are marked with arrows. (B) Comparison of P{wvar} with its red-brown and orange derivatives. DNA was cut with SpeI, PstI, BamHI, and XhoI and hybridized to probe 1. (C) Receding terminal XhoI fragment of P{wvar} over time. DNA from the original w1; P{wvar} (lanes 1 and 2) and from y1 w67c23; P{wvar} (lanes 3 and 4) was prepared from stocks in August 1998 (lanes 1 and 3) and again in February 2000 (lanes 2 and 4), cut with XhoI, and hybridized to probe 2a. (D) DNA from single orange-eyed flies of orange stock #2 and #4 was cut with BamHI and hybridized to probe 1 to reveal terminal DNA length polymorphism. The shorter DNA fragments in stock #2 are associated with a lighter orange eye color. (E) Southern blot of FIGE gel hybridized with probe 2a to determine the length of the terminal DNA addition in the brown-red variants #20 and KR3-3 and in the variant red #11-5. Relevant fragments are indicated by arrows. XhoI cuts within the terminal HeT-A in the brown-red variants and NruI (which usually has no site in HeT-A) reveal a very long terminal DNA extension, which is longer in KR3-3 than in brown-red #20. In contrast, XhoI and NruI indicate only an 6.5-kb HeT-A addition distal to the P{wvar} transgene in the variant red #11-5.

View larger version (29K): In this window In a new window Download PPT slide   Figure 5. Molecular maps determined by genomic Southern blots, PCR amplification, cloning, and sequencing of relevant fragments from P{wvar} and derivatives. All variants studied here derive from the P{wvar} stock either by terminal deletion, erosion, or terminal attachment of new HeT-A elements to the chromosome end. Orange variants have shortened terminal HeT-A arrays, and pale and yellow variants have lost all HeT-A sequences and part of the w+ transgene. Brown-red variants carry up to 65 kb of terminal HeT-A distal to the transgene and preserve the preexisting array of three truncated HeT-A element 3' UTRs. The exact structure of the long DNA extension cannot be determined and therefore is left blank. The red-eyed variant #11-5 has a much shorter terminal addition, consisting of the preexisting HeT-A element #1. An apparently full-length HeT-A element (HeT-A #2 new) has transposed onto the 66-bp shorter (compared to the brown-red variants) HeT-A#1 element.

Genomic Southern blotting, together with cloning and sequencing, was used to establish the molecular maps of the white, pale, yellow, orange, brown-red, and red variants derived from the original P{w^var} insertion (Fig 5). With hybridization probe 3 (see Fig 3) or a fragment from the 2L TAS, no changes were detected on the proximal side of the transgene in any of the variants studied (data not shown), suggesting that all phenotypic changes are determined by events distal to the w transgene. Changes at the distal end of the transgene were detected by hybridizing genomic Southern blots with probes 2b (Fig 4A) and 1 (Fig 4B). Probe 2b also hybridizes to a 4.0-kb SstI and an 14-kb ClaI fragment derived from the X chromosome location of the remaining w gene sequences in the w^67c23 deficiency (Fig 4A).

In yellow variant #9 and other yellow and pale variants, the chromosome end was positioned very close to the 5' start site of w. This terminal deficiency has lost all terminal HeT-A elements as well as the eye enhancer. As expected from the rate of terminal nucleotide loss, pale variant #8, which had the shortest 2L chromosome distal to the transgene (not shown), lost all eye color shortly after the original mapping was done, due to loss of the promoter region and part of the first exon of the w reporter gene. In the white variants the chromosome end had receded to various positions within the white transgene.

While the orange variants ranged in color from light to dark orange, all variegate with small, red spots. Without frequent selection to maintain a constant eye color, most of the orange variant stocks accumulated individuals with lighter and darker background eye color. Our results suggest that the shorter chromosomes result in lighter orange eye color. The genomic Southern blots in Fig 4A and Fig B, show three representative orange variants (#2, #11, and #12). The stock of orange variant #2 contains two different 2L chromosomes, the shorter being more prominent (Fig 4A). Genomic DNA from this stock showed strong hybridization of probe 2b to a 7.0-kb SstI and an 8.0-kb ClaI band, while the 9.4-kb SstI and 10-kb ClaI bands hybridized less strongly, reflecting a ratio of 3:1 of light:dark orange-eyed flies in the stock. By contrast, the stock of orange variant #11 showed strong hybridization to 9.4-kb SstI and 10-kb ClaI bands and weaker to 7.0-kb SstI and 8.0-kb ClaI bands, reflecting the 1:3 ratio of light:dark orange-eyed flies in the stock. It also shows (Fig 4B, probe 1) strong 3.0-, 4.0-, 4.1-, and 6.0-kb bands, and weaker 2.3-, 2.4-, 2.5-, and 5.0-kb bands when cut with SpeI, PstI, BamHI, and XhoI, respectively. Since PstI and XhoI do not cut in the HeT-A element array attached at the distal end of the P{w^var} transgene, the fragments generated by these enzymes reflect the position of the chromosome termini in these orange variants. The stock of light-orange variant #12 contains only the shorter version of the 2L chromosome. A genomic Southern blot with BamHI-cut DNA from individual orange flies (Fig 4D, probe 1) indicated considerable length polymorphism of terminal fragments, which cluster around the shorter and longer 2L chromosomes. Most of the flies were heterozygous, which may contribute to the range of orange eye colors in the stocks. Consistent with its eye color phenotype, the variant #11-5-orange had a short terminal fragment resembling orange #12, but red #11-5 differed significantly from the brown-red variants with all of the enzymes tested (not shown, but see map in Fig 5).

Genomic Southern blots showed that all brown-red variants had much larger SstI, ClaI, PstI, and XhoI bands than P{w^var}, suggesting a terminal DNA addition (Fig 4A and Fig B). It should be noted that the sizes of the SpeI and BamHI fragments (Fig 4B) remain unchanged between P{w^var}, the brown-red variants (brown-red #6 shown here), and the long version of the orange stocks (orange #11 shown here), indicating structural conservation of at least the proximal 2.5 kb of the HeT-A array in flies with different eye colors.

While the positions of the chromosome ends in the orange and yellow eye color variants could be determined on blots from regular agarose gels by the apparent colocalization of distal restriction sites (see above), the size of the terminal DNA additions in the brown-red and red variants was larger. Thus, we separated XhoI- and NruI-cut genomic DNA on an FIGE and hybridized to probe 2a. The autoradiograph in Fig 4E shows the same 11- to 12-kb Xho fragment for both brown-red variants #20 and KR3-3 (indicating a XhoI site in the terminal addition) and for very large NruI fragments, which differ between the two brown-red stocks. Since NruI does not usually cut in HeT-A and TART elements, these fragments are probably terminal. Thus, 55 kb of DNA is present distal to the w reporter gene in brown-red variant #20 (and others, not shown) and 65 kb in the independently isolated brown-red variant KR3-3. These long terminal extensions were confirmed by a blot of DNA digested with the 8-cutter restriction enzyme PmeI (not shown), which has a site located 12 kb proximal to the subtelomeric w transgene in l(2)gl (WALTER et al. 1995 ). By contrast, red #11-5 showed a much shorter terminal fragment with NruI. Given the relative positions of XhoI and NruI on the w transgene (see Fig 3), the chromosome end in red #11-5 appears to be located only 6.5 kb distal to the HeT-A/white junction.

To determine the molecular structure of the proximal end of the terminal addition in the brown-red variants, the 13-kb SalI fragment extending from the SalI site at nt 11866 in the w transgene to a SalI site in the distal DNA extension was cloned from four brown-red variant stocks into -phage. Sequences of these fragments revealed the presence of an array of three partial HeT-A elements joined in the usual way by their oligo(A) tails (Fig 5). Elements #1 and #2 consist entirely of 3' noncoding sequences of 2.1 and 1.9 kb, respectively; element #3 may be a complete 5-kb HeT-A element, but could only be analyzed up to the SalI site, which lies within the single open reading frame (ORF) of full-size HeT-A elements (BIESSMANN et al. 1992A ; DANILEVSKAYA et al. 1992 ). This organization was identical in all brown-red stocks analyzed and therefore probably existed in the original P{w^var} strain, from which the brown-red variants arose (see Fig 5).

To analyze the structure of the HeT-A array in red #11-5, the 11-kb SalI fragment from the SalI site in white to the one in the distal DNA extension was cloned and analyzed. The sequence of HeT-A #1 was identical to that of the proximal-most element in other orange and brown-red stocks and was also attached to nucleotide 5184 of the w transgene, indicating that no change had occurred in the proximal part of the HeT-A array. However, the distal end of HeT-A #1 in red #11-5 was 66 bp shorter than in the brown-red variants, deleting the SpeI site present in P{w^var} and its long orange and red-brown variants (see above). A new HeT-A element was attached with its oligo(A) tail to the distal end of the shortened HeT-A #1. This new element differed in restriction map and sequence from HeT-A #2 in the red-brown variants and therefore represents an independent transposition event. Sequences at the distal SalI site in the second element are located in the ORF of this element, suggesting that the second HeT-A element was very likely full size (6 kb) at the time of its transposition. This structure of red #11-5 is consistent with the origin of this red-eyed variant. We propose that a long HeT-A extension of its red-brown predecessor stock was removed by a terminal deletion event, giving rise to the orange-eyed intermediate with a truncated HeT-A array, which then became subject to terminal erosion. Transposition of a full-sized HeT-A element then gave rise to the terminal structure in the red #11-5 stock.

The molecular structure of the distal part of the 55- to 65-kb DNA additions in the brown-red variants could not be determined by cloning or PCR. We therefore used in situ hybridization to polytene chromosomes with probes from HeT-A and TART, the two known telomeric retrotransposons involved in telomere elongation in D. melanogaster. The P{w^var} chromosome was used for comparison (Fig 6). As expected from our molecular analysis, the 2L tip of P{w^var} hybridized only weakly to HeT-A (Fig 6A), and not to TART (Fig 6D). The brown-red variants showed increased HeT-A hybridization at their 2L tips (Fig 6B and Fig C), but also a strong TART hybridization (Fig 6E and Fig F). This suggests that the 2L tips in these variants have acquired more TART than HeT-A elements, probably in an intermingled arrangement similar to the one described at the 3R tip (LEVIS et al. 1993 ).

View larger version (65K): In this window In a new window Download PPT slide   Figure 6. FISH to polytene salivary gland chromosomes from P{wvar} and its derivatives. (A and D) P{wvar}. (B and E) Brown-red variant #5. (C and F) Brown-red variant KR3-3. A HeT-A probe was used in A–C, and a TART probe in D–F. All hybridizations include a dpp probe, which labels 22F1-2, for comparison.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this article we have used the subtelomeric transgene P{w^var} as a convenient and sensitive marker to detect molecular changes that occur at the 2L telomere. The present findings contribute to our understanding of several aspects of telomere biology in Drosophila and demonstrate an influence of telomeric retrotransposons on telomeric position effect. We detected HeT-A-mediated stimulation of the subtelomeric w⁺ transgene, which we interpret as a chromatin-altering influence of telomeric HeT-A elements. We were also able to quantitate HeT-A transposition to this chromosome end and identified a potential trans-acting factor, or factors, that influence transposition frequency. Finally, the brown-red eye color variants of P{w^var}, with their extended terminal HeT-A/TART arrays, enabled us to detect genetic interactions between homologous telomeres causing altered expression levels of the subtelomeric reporter gene. On the basis of these observations, we present here a new model for telomeric position effect in Drosophila and discuss the contributions of cis- and trans-acting DNA sequences.

P{w^var} as a sensitive detector of telomere dynamics:
The molecular structure of the P{w^var} transgene and the surrounding sequences at the 2L telomere were determined. At the distal end of the insert, 597 bp of the original construct as well as the 5' P-element sequences were deleted. Loss of the original HeT-A array on the target chromosome and any adjoining TAS repeats may have occurred at the time of insertion (GEHRING et al. 1984 ), or these sequences might have been lost by a terminal deletion or terminal recession after the integration of the P-element construct into the TAS array. Subsequent attachment of HeT-A elements to the truncated transgene may have then restored the telomere (BIESSMANN et al. 1990 ). The structure of P{w^var} is different from other analyzed P-element insertions into TAS regions, which preserved the integrity of the original construct (KARPEN and SPRADLING 1992 ; LEVIS et al. 1993 ; CRYDERMAN et al. 1999 ), but a similarly 5' truncated P-element insertion, NA-P(1A), has recently been identified at the X chromosome telomere (MARIN et al. 2000 ).

This unique structure may render the w⁺ transgene, which is transcribed in the distal-to-proximal direction, sensitive to naturally occurring or induced alterations at the 2L tip. Thus, the w⁺ transgene serves as a convenient sensor for structural changes at the telomere and allows the detection and quantitation of different genetic events. Through selection of phenotypic changes that occurred in P{w^var} stocks and crosses, followed by molecular characterization of the variants, we found that terminal deficiencies reaching the white transcription unit occur spontaneously with a rate of 1.3 x 10^-4/gamete/generation. These are likely to be similar in origin to the terminal deficiencies found at the tip of 3R using another w⁺ transgene (LEVIS 1989 ). New transpositions to the tip of 2L occur with a rate of 0.6%/gamete/generation, which is similar to the 1% rate calculated for HeT-A transposition onto the broken end of the X chromosome using a much more cumbersome system (BIESSMANN et al. 1992A ). Using the present assay, however, transposition can be identified in a single individual. Finally, the gradual change of eye color in homozygous stocks, from orange to yellow to pale to white, reflects the progressive loss of terminal DNA due to incomplete DNA replication (BIESSMANN and MASON 1988 ; BIESSMANN et al. 1992A ).

Using this system, we were able to detect a second-site genetic factor that affects telomere elongation. This genetic factor, which was present in the original P{w^var} stock on chromosome 2, may act as a repressor of HeT-A transposition because it eliminates the darker eye color variants that arise from HeT-A transposition. It is easy to imagine how artificial selection in favor of the repressor of transposition may have been applied. Any effort to maintain the phenotypic stability of the P{w^var} stock by elimination of the brown-red variants would have increased the frequency of the repressor of transposition.

Cis-acting factors of telomeric position effect:
All telomeric P-element integrations that are partially silenced and exhibit variegated expression of the reporter gene have occurred in TAS (KARPEN and SPRADLING 1992 ; LEVIS et al. 1993 ; CRYDERMAN et al. 1999 ). To our knowledge, no P-element insertion has yet been found in the terminal HeT-A/TART array, even though it is often larger than most subterminal TAS arrays, nor have insertions been identified in the single copy region that lies immediately proximal to TAS at 2L (WALTER et al. 1995 ) and at other telomeres (LEVIS et al. 1993 ). It is possible that insertions into the retrotransposon arrays are either completely silenced or not silenced at all; both cases would prevent detection of the transgene in a screen for variegating inserts. Alternatively, the HeT-A/TART array may be packaged in such a way that P-element insertion is reduced in these regions. There are two reasons to suggest that HeT-A elements are not packaged in a heterochromatic configuration or cause heterochromatic silencing. First, HeT-A transpositions onto a terminal X chromosome deficiency broken in the 5' upstream region of the yellow gene did not cause silencing or variegation of the adjacent yellow gene (BIESSMANN et al. 1990 ). Second, addition of HeT-A and TART elements onto the tip of the P{w^var} chromosome, as described here, cause an increase, not a decrease, in white expression, suggesting that HeT-A functions as a suppressor of TPE. By contrast, it appears likely that silencing of the telomeric P-element insertions is caused by the heterochromatic nature of TAS. First, recent evidence suggests that the 2L TAS, when placed between the eye-specific enhancer and promoter of a w⁺ reporter gene, is able to silence the reporter in an orientation-dependent manner (KURENOVA et al. 1998 ). Second, as described here, a fully terminal w⁺ transgene without distal HeT-A or TART elements is silenced as much or more than the same transgene with an adjacent HeT-A array.

By analyzing the molecular structure of various eye color variants derived from P{w^var}, we realized that all phenotypic changes were caused by changes in DNA organization occurring distal to the transgene. In general, the number of retroelements at the end of the chromosome influences the level of w⁺ transgene expression. For instance, in the case of the orange variants, a correlation exists between the length of the HeT-A array and eye color: two tandem truncated 3' nontranslated HeT-A fragments cause darker eye color than one. Moreover, the brown-red variants and the even darker variant red #11-5 also support the idea that terminal HeT-A (and possibly TART) elements have an activating influence on the subterminal transgene. Although Drosophila terminal retrotransposons are structurally very different from the telomerase-generated telomeric poly(G_1-3T) tracts in S. cerevisiae, and they do not play an integral part in the capping function (MASON et al. 1984 ; BIESSMANN and MASON 1988 ), as do the poly(G_1-3T) tracts in yeast (SANDELL and ZAKIAN 1993 ), they may be considered equivalent in their function of telomere elongation. However, while the terminal sequences in S. cerevisiae repress a subterminal transgene, and extended poly(G_1-3T) tracts have a stronger silencing effect than shorter ones (KYRION et al. 1993 ), the influence of HeT-A on telomeric silencing is the opposite. An explanation for this difference lies in the structure and the transposition mechanism of the HeT-A element itself, as a promoter activity is located in the 3' end of these retroelements (DANILEVSKAYA et al. 1997 ). Thus, HeT-A has a natural ability to activate downstream sequences, and the HeT-A promoter activity might influence subtelomeric transgenes that are transcribed from distal to proximal, such as P{w^var} (this article), A4-4 (LEVIS 1989 ), and 39C-5 (CRYDERMAN et al. 1999 ), perhaps by a readthrough mechanism.

However, a more complex scenario seems likely. Because the new HeT-A and TART additions, which are correlated with substantial darkening in eye color, occur 6 kb distal to the subtelomeric transgene, the possibility of a readthrough from the HeT/TART promoters seems diminished. Moreover, the telomeric structure of red #11-5, which consists of one partial and one possibly full-length HeT-A element, suggests that the strong activating influence of the terminal HeT-A array in red #11-5 on the w transgene cannot be attributed solely to the promoter activity in the 3' untranslated regions (UTRs). The presence of a probably full-length HeT-A element distinguishes red #11-5 from all of the orange variants, which have only a short, second truncated HeT-A 3' UTR in its place. Why would a full-length HeT-A element have such a strong activating influence on a downstream reporter gene? Several possible explanations exist. First, w⁺ activity may be increased in red #11-5 because its promoter is farther from the repressive effects of TAS. This seems unlikely, because the brown-red variants have new HeT-A elements that are, in general, farther from TAS, but have weaker effects on w⁺ activity than does red #11-5. Second, full-length HeT-A elements may have a boundary sequence that buffers at least part of the 3' promoter (DANILEVSKAYA et al. 1997 ) from the silencing effects of TAS. This hypothesis does not explain the difference between red #11-5 and the brown-red variants in w⁺ activation. The simplest explanation would be the presence of a strong enhancer/promoter on the HeT-A element that resides farther upstream on the element than the promoter detected close to the 3' oligo(A) tail (DANILEVSKAYA et al. 1997 ). Such enhancer/promoter activity might also explain the brown-red eye color variants. In these variants, new HeT-A and TART elements have attached to the preexisting array of three truncated 3' noncoding HeT-A regions. As these newly attached elements are farther away from the transgene in the brown-red variants than in red #11-5, their influence on the w reporter gene might be weaker, resulting in the less intense brown-red phenotype. Experiments to identify this postulated enhancer/promoter activity in transgenic reporter constructs are currently underway.

A new model for telomeric position effect variegation:
Telomeric silencing differs from centromeric position effect variegation. The mechanisms responsible for heterochromatic gene silencing and variegated expression are not fully understood, but two primary models, the heterochromatic spreading model and the intranuclear compartmentalization model, have been proposed to explain these phenomena (for a discussion see HENIKOFF 1995 ). On the basis of results reported here, we propose a different model for TPE, which takes into account the specific structure of the Drosophila telomere. This model postulates that TAS represses transcription toward the HeT-A/TART array, while the terminal HeT-A elements promote transcription proximally. Thus, a reporter gene inserted between the two arrays, or within TAS, is subject to these competing influences. We propose that the variegation of telomeric reporter genes is caused by a competition between the "repressive" force of TAS and the "opening" force of HeT-A elements. The situation is certainly more complex than the simple readthrough mechanism from HeT-A promoters, because other subterminal P elements in TAS regions, in which the reporter gene is transcribed in the opposite direction, also variegate (P. GEYER, personal communication). If binding of a protein complex to the postulated enhancer/promoter on HeT-A were sufficient to transiently open the heterochromatic structure of the subtelomeric region, access of the transcription machinery to the reporter gene promoter would be facilitated, and the orientation of the transgene would become less relevant. This idea is also consistent with the observation that transcription at telomeres in yeast has a suppression effect on TPE (RENAULD et al. 1993 ; SANDELL et al. 1994 ). The model also explains the previously reported observation (LEVIS 1989 ; SHEEN and LEVIS 1994 ) that deletions of the distally located TAS and the terminal HeT-A/TART array from the 3R tip of the chromosome carrying P{ry⁺ w⁺}A4-4 relieve repression on the white reporter gene.

We think it unlikely that variegation occurs by somatic HeT-A transposition to the chromosome ends which, by analogy with the red-eyed variants described here, might result in somatic clones that express the w⁺ transgene more strongly. Variants that show consistently high numbers of dark spots do not give evidence on Southerns of somatic HeT-A additions, as might be expected if HeT-A transposition were elevated in all somatic tissues (data not shown). On the other hand, somatic HeT-A transposition may produce the mosaic dark sectors that are occasionally seen.

Expression level of the subterminal transgene depends on the structure of the homologous telomere:
The brown-red variants of P{w^var}, which have a 50- to 60-kb array of HeT-A/TART elements attached at the 2L telomere, respond to disturbances at the telomere of the homologous chromosome. While they show repressed orange eye color when the homologue contains an undisturbed (wild-type) telomere, alteration or deletion of the 2L telomere of the homologous chromosome results in darkening of the eye color. This effect is specific to alterations of the telomere itself, because insertions into and deletions of the adjacent l(2)gl gene that do not affect the TAS array have no effect (M. D. GOLUBOVSKY, A. Y. KONEV and J. M. MASON, unpublished results). If the telomeres of the two homologues interact with each other to assess their integrity, a "disturbed" telomere could stimulate enhancer/promoter activity of the HeT-A elements on the homologous telomere, thus resulting in increased transcription from the white reporter gene and darker eye color. A similar trans-activation of P{w⁺, ry⁺}A4-4 by 3R terminal deficiencies has been reported (LAURENTI et al. 1995 ). Such telomeric communication may serve to regulate HeT-A promoter activity and consequently might stimulate HeT-A transposition and chromosome end elongation in response to HeT-A arrays that are too short. This might be a mechanism by which Drosophila controls the length of its telomeres. The mechanism of this trans-activation remains unknown, but it might depend on homologous pairing, as do transvection (GELBART and WU 1982 ; WU and MORRIS 1999 ) and euchromatic gene silencing at Polycomb response element sites (KASSIS et al. 1991 ; KASSIS 1994 ; GINDHART and KAUFMAN 1995 ), or it may depend on telomere-telomere interactions of a more global nature. Further studies will address these possibilities.

	FOOTNOTES

¹ Present address: Division of Evolutionary Theory, Institute of Science and Technology History, Russian Academy of Sciences, St. Petersburg 199034, Russia.
² Present address: Molecular Biology and Virology Laboratory, The Salk Institute, La Jolla, CA 92037.

	ACKNOWLEDGMENTS

Janice Yao, Leila Bozorgnia, and Aneeta Maheshwari helped with the molecular analyses. We thank Robert Levis for sharing unpublished information and for the TART probe. We especially thank Karen Sutton and Alexandra Amriz for their expert automated sequence analyses. This work was supported by the U.S. Public Health Service grant GM-56729 to H.B.

Manuscript received November 6, 2000; Accepted for publication April 9, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALTSCHUL, S. F., W. GISH, W. MILLER, E. W. MYERS, and D. J. LIPMAN, 1990  Basic local alignment search tool. J. Mol. Biol. 215:403-410[Medline].

BIESSMANN, H. and J. M. MASON, 1988  Progressive loss of DNA sequences from terminal chromosome deficiencies in Drosophila melanogaster. EMBO J. 7:1081-1086[Medline].

BIESSMANN, H., J. M. MASON, K. FERRY, M. D'HULST, and K. VALGEIRSDOTTIR et al., 1990  Addition of telomere-associated HeT DNA sequences "heals" broken chromosome ends in Drosophila. Cell 61:663-673[Medline].

BIESSMANN, H., L. E. CHAMPION, M. O'HAIR, K. IKENAGA, and B. KASRAVI et al., 1992a  Frequent transpositions of Drosophila melanogaster HeT-A transposable elements to receding chromosome ends. EMBO J. 11:4459-4469[Medline].

BIESSMANN, H., K. VALGEIRSDOTTIR, A. LOFSKY, C. CHIN, and B. GINTHER et al., 1992b  HeT-A, a transposable element specifically involved in healing broken chromosome ends in Drosophila melanogaster. Mol. Cell. Biol. 12:3910-3918[Abstract/Free Full Text].

BLACKBURN, E. H., 1992  Telomerases. Annu. Rev. Biochem. 61:113-129[Medline].

CRYDERMAN, D. E., E. J. MORRIS, H. BIESSMANN, S. C. R. ELGIN, and L. L. WALLRATH, 1999  Silencing at Drosophila telomeres: nuclear organization and chromatin structure play critical roles. EMBO J. 18:3724-3735[Medline].

DANILEVSKAYA, O. N., D. A. PETROV, M. N. PAVLOVA, A. KOGA, and E. V. KURENOVA et al., 1992  A repetitive DNA element, associated with telomeric sequences in Drosophila melanogaster, contains open reading frames. Chromosoma 102:32-40[Medline].

DANILEVSKAYA, O. N., I. R. ARKHIPOVA, K. L. TRAVERSE, and M. L. PARDUE, 1997  Promoting in tandem: the promoter for telomere transposon HeT-A and implications for the evolution of retroviral LTRs. Cell 88:647-655[Medline].

GEHRING, W. J., R. KLEMENZ, U. WEBER, and U. KLOTER, 1984  Functional analysis of the white⁺ gene of Drosophila by P-factor-mediated transformation. EMBO J. 3:2077-2085[Medline].

GELBART, W. M. and C. T. WU, 1982  Interactions of zeste mutations with loci exhibiting transvection effects in Drosophila melanogaster. Genetics 102:179-189[Abstract/Free Full Text].

GINDHART, J. G. and T. C. KAUFMAN, 1995  Identification of polycomb and trithorax group responsive elements in the regulatory region of the Drosophila homeotic gene sex combs reduced. Genetics 139:797-814[Abstract].

GOTTA, M., S. STRAHLBOLSINGER, H. RENAULD, T. LAROCHE, and B. K. KENNEDY et al., 1997  Localization of Sir2p: the nucleolus as a compartment for silent information regulators. EMBO J. 16:3243-3255[Medline].

GREIDER, C. W., 1996  Telomere length regulation. Annu. Rev. Biochem. 65:337-365[Medline].

GRUNSTEIN, M., 1997  Molecular model for telomeric heterochromatin in yeast. Curr. Opin. Cell Biol. 9:383-387[Medline].

HAZELRIGG, T., R. W. LEVIS, and G. M. RUBIN, 1984  Transformation of white locus DNA in Drosophila: dosage compensation, zeste interaction, and position effects. Cell 36:469-481[Medline].

HENIKOFF, S., 1995  Gene silencing in Drosophila. Curr. Top. Microbiol. Immunol. 197:193-208[Medline].

KARPEN, G. H. and A. C. SPRADLING, 1992  Analysis of subtelomeric heterochromatin in the Drosophila minichromosome Dp1187 by single-P element insertional mutagenesis. Genetics 132:737-753[Abstract].

KASSIS, J. A., 1994  Unusual properties of regulatory DNA from the Drosophila engrailed gene: three "pairing-sensitive" sites within a 1.6-kb region. Genetics 136:1025-1038[Abstract].

KASSIS, J. A., E. P. VANSICKLE, and S. M. SENSABAUGH, 1991  A fragment of engrailed regulatory DNA can mediate transvection of the white gene in Drosophila. Genetics 128:751-761[Abstract].

KURENOVA, E., L. CHAMPION, H. BIESSMANN, and J. M. MASON, 1998  Directional gene silencing induced by a complex subtelomeric satellite from Drosophila. Chromosoma 107:311-320[Medline].

KYRION, G., K. LIU, C. LIU, and A. J. LUSTIG, 1993  RAP1 and telomere structure regulate telomere position effects in Saccharomyces cerevisiae. Genes Dev. 7:1146-1159[Abstract/Free Full Text].

LAURENTI, P., Y. GRABA, R. ROSSET, and J. PRADEL, 1995  Genetic and molecular analysis of terminal deletions of chromosome 3R of Drosophila melanogaster. Gene 154:177-181[Medline].

LEVIS, R. W., 1989  Viable deletions of a telomere from a Drosophila chromosome. Cell 58:791-801[Medline].

LEVIS, R., T. HAZELRIGG, and G. M. RUBIN, 1985  Effects of genome position on the expression of transduced copies of the white gene of Drosophila. Science 229:558-561[Abstract/Free Full Text].

LEVIS, R. W., R. GANESAN, K. HOUTCHENS, L. A. TOLAR, and F. M. SHEEN, 1993  Transposons in place of telomeric repeats at a Drosophila telomere. Cell 75:1083-1093