From Sequence to Chromosome: The Tip of the X Chromosome of D. melanogaster

See allHide authors and affiliations

Science  24 Mar 2000:
Vol. 287, Issue 5461, pp. 2220-2222
DOI: 10.1126/science.287.5461.2220


One of the rewards of having a Drosophila melanogaster whole-genome sequence will be the potential to understand the molecular bases for structural features of chromosomes that have been a long-standing puzzle. Analysis of 2.6 megabases of sequence from the tip of the X chromosome ofDrosophila identifies 273 genes. Cloned DNAs from the characteristic bulbous structure at the tip of the Xchromosome in the region of the broad complex display an unusual pattern of in situ hybridization. Sequence analysis revealed that this region comprises 154 kilobases of DNA flanked by 1.2-kilobases of inverted repeats, each composed of a 350–base pair satellite related element. Thus, some aspects of chromosome structure appear to be revealed directly within the DNA sequence itself.

Fewer than 90 years have elapsed since Alfred H. Sturtevant presented the world with the first-ever genetic map of six visible markers on the X chromosome of Drosophila(1). Now that the sequence of almost the entire euchromatic genome of Drosophila has been determined (2), we have the opportunity to study the function of each gene. In addition, we can study the relation between DNA sequence and chromosome structure.

The European Drosophila Genome Project (EDGP) (3) has determined the sequence of a contiguous segment of DNA extending some 2.6 Mb from the tip of the X chromosome, that is, from subdivision 1A to subdivision 3C on the standard polytene chromosome map (4). We predict the existence of 273 protein-coding genes in this region (one gene every 9.6 kb), which is of some sentimental, as well as much scientific, interest to geneticists. It extends from a position 120-kb distal to theyellow locus to 150-kb proximal to the whitelocus, whose mutation was the first clearly visible mutation found inDrosophila and whose study led to the discovery of sex-linked inheritance and, hence, to the proof of the chromosome theory of heredity (5).

In the region sequenced, we have identified 17 transposon insertions. The most common element was roo (six copies), but six other retroviral-like elements, two LINE-like elements, an element with inverted repeat ends (S-element), and a foldback (FB) element were also found. The overall density of insertions (one insertion every 155 kb) is similar to that in the Adh region [one every 170 kb (6)]. This is of some interest because the tip of the X is a region of low genetic recombination and, on theoretical grounds, might have been expected to accumulate transposable elements. However, it has been suspected that transposon insertion might reduce fitness. Because the Xchromosome is hemizygous in the male fly, it is subject to stronger selection pressures. This would lead to the prediction of a lower frequency of insertions on the X (7). Our finding that the overall transposable element densities in the Xtip and region 35 to 36 are comparable argues against the maintenance of element copy number by negative selection.

The first physical map of any genome was the description of the polytene chromosomes of Drosophila (8). These chromosomes arise from endoreduplication resulting in a large number of parallel fibers with each fiber representing a single haploid chromosome. They are characterized by an aperiodic pattern of dark-stained bands and light-stained interbands, reflecting differences in the extent of DNA packing. These patterns are colinear with the genetic map, as proven by Bridges (9) and must be remarkably stable because they are recognizable in species that have diverged many millions of years ago (10). Our contiguous sequence covers 102 of the 5072 polytene chromosome bands. The average DNA content per band is 26.2 kb, similar to that estimated by Sorsa (11) for the genome as a whole but less than that estimated (49 kb/band) for the Adhregion (6). Little is known about the mechanism that determines the banding pattern of polytene chromosomes. Although the answer must be based on the DNA sequence, any satisfactory solution to this problem will require further understanding of the DNA-protein interactions inherent to chromosome structure. We can now suggest an explanation for a long-standing observation concerning the morphology of the polytene X chromosome tip.

The tip of the polytene X chromosome inDrosophila and related species is characterized by a bulbous structure that often exhibits an unusual arrangement of chromosome bands. This arrangement was noted by those who pioneered the study of polytene chromosomes, including Bridges (12) and Offermann (13). Normally, in polytene chromosomes ofDrosophila, each band crosses the entire width of the chromosome and lies perpendicular to its long axis. In the bulbous region in division 2 of the X chromosome, the bands within the cytological interval 2B3-8 are often roughly parallel to the chromosome's long axis (Fig. 1). This pattern is unique to the wild-type chromosomes of Drosophila. Both Bridges and Offermann concluded that it resulted from local duplications of polytene chromosome bands, in particular from reverse duplications. These could, in molecular terms, have represented large-scale duplications up to 25 kb, although Bridges and Offerman did not have the tools to resolve this. Banding patterns not dissimilar to this are seen in chromosome aberrations that have been interpreted as reverse duplications (14) and are common in the genomes of some other Drosophila species (15).

Figure 1

The 2B region of the X chromosome. (A) A polytene chromosome to which the cosmid clone 9E2 has been hybridized in situ. This ∼40-kb clone apparently covers six polytene chromosome bands and the hybridization signals are concentrated laterally. This is seen with many independent cosmid clones hybridizing to this region, including 196F3 and 9D2. (B) A reproduction of the drawing of this region by Bridges (12). He depicted the bands 2B3-4 and 2B7-8 as being connected. (C) A molecular map of this region (180 kb) showing the inverted 1.688 satellite DNA-like repeats distal tobr (within cosmid 196F3) and proximal to a6 (within cosmid 9D2). Four genes in this region were known previously (br,dor, a6, b6); the others have been computationally predicted.

To determine whether this unusual polytene structure might reflect the organization of the DNA in the 2B3-8 interval, we carried out in situ hybridization studies with more than 50 independent cosmid clones previously localized to this region by overlapping restriction endonuclease digestion patterns (3). Independent overlapping clones from this region give the striking pattern of in situ hybridization seen in Fig. 1. Rather than lying transverse to the chromosome as it would normally, the hybridization appears to be restricted to the lateral parts of the 2B bulb. Thus, it seems that the exceptional banding pattern in region 2B of the polytene chromosomes has its counterpart in the patterns of in situ hybridization.

We know from genetic studies that region 2B includes thebroad (br) gene. This gene encodes a family of zinc finger proteins that possess common and unique exons (16). A consequence of its molecular organization is that the br region has a complex pattern of complementation between mutant alleles: It was originally defined as four mutually complementing, lethally mutable loci. The region is expressed as an “early” ecdysone puff in salivary gland polytene chromosomes (17). We had no difficulty in assembling contiguous DNA sequence from the DNA sequence of 12 cosmid clones spanning the 2B3-8 region and including the br locus. Thus, we could not account for the unusual polytene chromosome structure or for in situ hybridization patterns of region 2B cosmids by the presence of any large-scale sequence repeat.

To determine whether there might be any other aspects of DNA sequence responsible for both of these features, we used computational methods (18) to look for both direct (i.e., abab ) and reverse (i.e., abba) repeat motifs in a 1.4-Mb region that spans from polytene chromosome region 1D (690 kb distal tobr) to 3A1 (720 kb proximal to br). We find at a position some 82 kb distal to broad a 1.2 kb sequence composed of three and a half tandem repeats of a 350 base pair (bp)–element. A corresponding 3.5 repeat of the same element is found in inverted orientation some 44 kb proximal to broad. This 350-bp element is related to the 1.688 satellite repeat, of which three dispersed subfamilies have been described previously (19). They are scattered in units of 1 to 4 and are at sites largely restricted to the X chromosome, where it has been speculated that they have a role in dosage compensation. The two internally repetitive 1.2-kb inverted repeats in region 2B are located precisely where they could define the ends of the inverted repeat band region suggested by Bridges and Offermann some 60 years ago.

Our findings suggest that inverted repeats of DNA can influence the architecture of chromosomes even when they are widely separated, in this case by 154 kb of sequences that have no obvious repetitive structure within them. Factors that influence the three- dimensional organization of chromosomes within the nucleus are poorly understood. We suggest that some of them might be recognized within the long-range sequence of the DNA itself. Because sequence repeats of this type are not uncommon within eukaroytic genomes, this could be one general means influencing the organization of chromosomal domains.


View Abstract

Navigate This Article