Patterns of Variation in the Intergenic Spacers of Ribosomal DNA in Drosophila melanogaster Support a Model for Genetic Exchanges During X-Y Pairing

Corresponding author: Gabriel A. Dover, Department of Genetics, University of Leicester, Leicester LE1 7RH, United Kingdom., genetics{at}le.ac.uk (E-mail)

Communicating editor: R. S. HAWLEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Detailed analysis of variation in intergenic spacer (IGS) and internal transcribed spacer (ITS) regions of rDNA drawn from natural populations of Drosophila melanogaster has revealed contrasting patterns of homogenization although both spacers are located in the same rDNA unit. On the basis of the role of IGS regions in X-Y chromosome pairing, we proposed a mechanism of single-strand exchanges at the IGS regions, which can explain the different evolutionary trajectories followed by the IGS and the ITS regions. Here, we provide data from the chromosomal distribution of selected IGS length variants, as well as the detailed internal structure of a large number of IGS regions obtained from specific X and Y chromosomes. The variability found in the different internal subrepeat regions of IGS regions isolated from X and Y chromosomes supports the proposed mechanism of genetic exchanges and suggests that only the "240" subrepeats are involved. The presence of a putative site for topoisomerase I at the 5' end of the 18S rRNA gene would allow for the exchange between X and Y chromosomes of some 240 subrepeats, the promoter, and the ETS region, leaving the rest of the rDNA unit to evolve along separate chromosomal lineages. The phenomenon of localized units (modules) of homogenization has implications for multigene family evolution in general.

DNA sequences can provide high-resolution information about the genetic variation within and between species (for literature see AVISE 1994 ). However, almost all nuclear components and some key sections of organelle genomes (HOEZEL et al. 1991 , HOEZEL et al. 1994 ) are subject to a variety of mechanisms of turnover that affect the evolution of these sequences. Ignorance of the degree of instability of any genomic region, and the ensuing population consequences, can lead to serious errors of interpretation of the extent to which selection and/or drift are responsible for the genetic variation within and between given taxonomical units. Indeed, selection and drift ignore the effects of genomic turnover on the distribution of sequence variations in populations. Only hard data on the distribution patterns of molecular variation in natural populations, coupled to the analysis of the relevant mutational processes, will allow the forces shaping the genetic variation to be accurately modeled. It is inappropriate to equate the homogenization consequences of turnover with mutation in standard models of population genetics (DOVER 1994 , DOVER 2000 ).

Ribosomal DNA (rDNA) has become the paradigm multigene family regarding the analysis of fundamental genomic processes that influence patterns of variation at the population and taxonomic levels. rDNA also provides an opportunity to assess the interaction of genomic processes with natural selection (WOESE 1991 ; DOVER et al. 1993 ). A full understanding of the molecular evolutionary dynamics of rDNA is not available, in particular for the large intergenic spacer (IGS) and its arrays of subrepeats (see Fig 1), despite its widespread use as a probe.

View larger version (11K): In this window In a new window Download PPT slide   Figure 1. Structure of ribosomal DNA (rDNA) repeat units in Drosophila melanogaster. Boxes indicate 28S, 18S, and 5.8S ribosomal RNA genes and external transcribed spacer (ETS). Solid lines indicate spacer regions: IGS, intergenic spacer region; ITS-1 and ITS-2, internal transcribed spacer regions. TIS, transcription initiation site. 95s, 330s, and 240s indicate the organization of subrepeat units (named after their size in base pairs) that can be present in the IGS region.

In the case of Drosophila melanogaster rDNA, early studies on the distribution of mutations shared between IGSs of X and Y chromosomes plus direct observations of unequal exchanges showed that interchromosomal events were participating in the evolution of the subrepeat arrays embedded in the IGSs, leading to species-specific homogenization patterns characteristic of concerted evolution (COEN and DOVER 1983 ; reviewed in DOVER et al. 1993 ). However, more recent results derived from the internal transcribed spacer (ITS) suggest that these spacers are evolving along separate chromosomal lineages because intrachromosomal unequal exchanges are seemingly much more frequent than interchromosomal unequal exchanges (SCHLOTTERER and TAUTZ 1994 ). In general, if homogenization is confined to a single chromosomal lineage, selection may operate on the whole homogenized array and not on individual units within it, leading to the concerted evolution pattern.

Recently, we analyzed for the first time both the IGS and the ITS regions drawn from the same individual flies of natural populations of D. melanogaster (POLANCO et al. 1998 ). The chromosomal distribution of ITS mutations indicated a haplotypic evolution for D. melanogaster rDNA as proposed by SCHLOTTERER and TAUTZ 1994 . However, the patterns of partial or complete homogenization within the ITS region are not the same as homogenization patterns within the IGS region. In contrast to ITS variants, IGS length variants are not distributed to specific chromosomal lineages, but can be shared by several different lineages. The data on IGS point to a multilineage model of evolution for this region. Hence, in one and the same rDNA unit, two different evolutionary progressions are occurring. How has this taken place?

To explain such different evolutionary trajectories, we have proposed a model of exchanges of single-strand fragments of only the IGS regions between X and Y chromosomes during X-Y chromosome pairing (POLANCO et al. 1998 ), on the basis of the achiasmatic pairing process proposed by MCKEE et al. 1992 . This model can explain the shared distribution of IGS mutations between X and Y chromosomes, while leaving the other regions of the rDNA unit to evolve mostly along haplotypic lineages. This model also allows for the presence of chromosome-specific IGS variants, as the whole rDNA does not participate in the X-Y chromosomal pairing (PARK and YAMAMOTO 1995 ).

Fine-grained analyses of the chromosomal distributions of IGS mutations and length variants are needed in order to obtain direct evidence for such a model of IGS fragment exchanges. Here, we provide data of the detailed internal structure of a large number of IGS length variants that have been isolated from specific X and Y chromosomes of isofemale lines collected along a 1000-km transect on the eastern seaboard of Australia. The IGS variants analyzed show chromosomal-linked differences that not only support the above model of restricted exchanges but suggest the precise location of the regions involved.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fly stocks:
Isofemale lines from natural populations of D. melanogaster collected along the east coast of Australia were a gift of Professor L. Partridge (University College, London). See Table 1 for names and locations of the 10 lines used in this study.

Fly stocks whose only source of rDNA is a single X chromosome or a single Y chromosome were constructed from the above Australian lines. X/O males were generated by crossing a male from the stock of interest to C(1)RM, y v f/O virgin females (stock 00400 from Umeå Drosophila Stock Center). After 3 days, the same male was removed and crossed to C(1)DX, y f lz virgin females (stock 06600 from Umeå Drosophila Stock Center) in order to analyze the Y chromosomal rDNA using the female progeny. Complete genotypes used were C(1)RM, y¹ v¹ f¹ / C(1;Y)1, In(1)dl-49, y¹ v¹ f¹ car¹ and C(1)DX, y¹ f¹ / lz³.

Australian isofemale lines and Umeå stocks were routinely transferred to fresh bottles of oatmeal and treacle medium at 3-wk intervals and maintained at 18°.

DNA extraction and PCR amplification of IGS regions:
Genomic DNA was extracted from single flies following the method of ASHBURNER 1989 , with the exception that flies were homogenized directly in the lysis buffer.

Intergenic spacers were amplified by PCR using a forward primer (IGS-F) located in the 3' end of the 28S rDNA gene and a reverse primer [external transcribed spacer (ETS)-R] located within the external transcribed spacer region (see Fig 1) as described elsewhere (POLANCO et al. 1998 ). Evidence for the reproducibility of PCR products is available in RUIZ LINARES et al. 1994 ; BOWEN and DOVER 1995 ; and POLANCO et al. 1998 . When needed, IGS-PCR products were cloned using the pGEM-T Easy vector system (Promega, Madison, WI) following manufacturer's instructions. Cloned inserts were released by EcoRI digestion and were compared in the same gel against the appropriate IGS profile for identification of the specific IGS length variants.

Variants were identified by the same codes given by POLANCO et al. 1998 , which include one letter and two numbers according to their relative gel migration between molecular markers (see Fig 2).

View larger version (58K): In this window In a new window Download PPT slide   Figure 2. Amplification of IGS regions of four selected Australian males (AG-4, LH-4, H-4, and R-4) and single flies carrying their corresponding X or Y chromosomes as the only source of rDNA. IGS length variants are identified by a code of one letter (according to their migration between the molecular marker bands: regions A, B, C, D, and E) and two numbers. Several common length variants are indicated by arrows, as examples.

Restriction mapping and sequencing of IGS length variants:
The internal subrepeat structure of selected IGS length variants was obtained by digestion of the corresponding clone in three separate reactions using restriction enzymes specific for particular IGS motifs: (i) triple digestion with EcoRI, NheI, and Tsp45I; (ii) double digestion with EcoRI and NheI; and (iii) double digestion with EcoRI and Tsp45I; EcoRI is needed to remove the vector and does not cut in the IGS sequence; Tsp45I cuts in every repeat of the 95-subrepeat (s) family; and NheI cuts in every 240s. The number of 330s repeats and the type of junction region between 330s and 240s were estimated from the triple digestion because no restriction sites for the enzymes are present in those regions. The number of 95s and 240s units was estimated from each of the double digestions once the structure of the central region was known. In contrast to 330s and 240s units, which show near identical length within them, the published sequences for 95s units range from 83 to 100 bp, so restriction mapping leaves a margin of uncertainty as to the precise number of 95s as this is calculated as multiples, giving the approximate size of the bands observed. A mean value of 92 bp has been used to calculate the expected size of the IGS length variants according to their internal structure.

Restriction enzyme AseI has been used in double digestions with EcoRI to reveal the bipartite structure of the 240s family, as the AseI site is not present in type-A 240s units (located on the 3' side) but is found in type-B units (located on the 5' side), according to RUIZ LINARES et al. 1994 .

The internal structure obtained by restriction mapping has been partially confirmed for more than half of the length variants analyzed by sequencing ~800 bp from the reverse primer (ETS-R) by automatic sequencing at the Protein and Nucleic Acids Chemistry Laboratory facility at the University of Leicester (see Table 3).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Are specific IGS length variants shared by isolated X and Y rDNA arrays?
Isofemale lines from 10 Australian natural populations were chosen from the 48 lines analyzed in our previous study (POLANCO et al. 1998 ), one line for each population studied (Table 1). The selection criteria were the presence in the isofemale lines of single X and Y chromosomes and the presence of relatively high numbers of IGS length variants. Selection of stocks with relatively high numbers of IGS variants was necessary to compare the internal structure of a large number of variants shared by chromosomes located in distant populations. Males from these lines were used in crosses designed to obtain fly stocks that had only the rDNA array from the corresponding single Australian X or Y chromosome (see MATERIALS AND METHODS).

DNA was extracted independently, for each isoline, as follows: (i) from the original male used for the crosses; (ii) from a single fly carrying the male X chromosome; and (iii) from a single fly carrying the male Y chromosome. After PCR amplification of the three samples using IGS-F and ETS-R primers, the corresponding IGS profiles were compared in order to assign the chromosomal location of the IGS length variants present in the original male (see Fig 2).

The total number of length variants identified for each male and their chromosomal location are indicated in Table 2. Between 4.8 and 22.2% of the IGS length variants are shared by X and Y chromosomes at the individual male level. However, if we compare the 10 analyzed X chromosomes against the 10 analyzed Y chromosomes among all the isofemale lines, the percentage of shared variants is 64.5%, thus revealing that some length variants located in the X chromosome of any given isofemale line are present in the Y chromosome of other lines, and vice versa.

Two examples are length variants D15 and B98 in Fig 2: D15 is present in male AG-4 only in the X chromosome, but in males LH-4 and R-4 it is present only in the Y chromosome. On the other hand, B98 is present in male AG-4 only in the Y chromosome, but in male LH-4 the same length variant is located only in the X chromosome. In other cases the same length variant is located in both X and Y chromosomes of the same male (as A85 in LH-4 or D55 in AG-4) while it is chromosome specific in other lines (as A85 in H-4 and R-4, or D55 in LH-4, H-4, and R-4).

These results are explicable either because there is an extensive sharing of length variants between chromosomes and across the populations or there is a frequent coincidental production of the same length variants in both X and Y chromosomes (see DISCUSSION).

Do shared IGS length variants consist of the same internal structure?
The internal subrepeat structure of the IGS region of rDNA may differ between nearly identical IGS length variants. If this is the case, then we would be overestimating the degree of sharing of length variants between arrays. To resolve this question, we have cloned 46 IGS regions corresponding to 24 IGS length variants. They have been obtained from seven different chromosomes: four X chromosomes (each from a different population) and three Y chromosomes (each from a different population).

The internal structures of the IGS regions analyzed were obtained by restriction-enzyme mapping and are indicated in Table 3 and Fig 3. Analysis of the number of subrepeats and the compositional type of the IGS regions reveals the existence of near-identical length IGS variants that have a completely different internal structure, suggesting that they are different variants unrelated by descent. Such is the case for the two B80 variants isolated from different X chromosomes and the B35 and D15 variants isolated from different Y chromosomes. On the other hand, B03 was identical in all three X chromosomes from which it was isolated, and variants B00, B82, D04, and D55, which were isolated from different Y chromosomes, also did not show any internal variability. When we compared the same length variants isolated from X and Y chromosomes, we obtained the same contrasting results: while A85, B25, B35, B82, C30, D00, D15, and D25 showed a different internal structure between X-linked and Y-linked variants, B82 and D55 were identical variants for X and Y chromosomes.

View larger version (23K): In this window In a new window Download PPT slide   Figure 3. Graphic representation of the different regions and subrepeat families of the IGS regions analyzed. The top shows a typical IGS where several units (n) of subrepeat families 95 and 330 are followed by the junction region, then several (n) units of subrepeat family 240, and finally the promoter region. The subrepeat family 95 and the promoter region are always present in each IGS but other regions could be absent (see Table 3). Different types of junction (S, N, and L) and promoter (P, M0, M1, M1*, and M2) have been identified. Promoter regions M1, M1*, and M2 were preceded exclusively by 330s and/or 95s units as indicated in the bottom. All these regions and subrepeat families are made of common sequences that are represented as boxes with different patterns and identified by the codes on top of the boxes; asterisks denote shorter sequences than the reference; straight lines in the promoter regions denote single nonrepeat sequence. Arrows indicate topoisomerase I sites located at the 3' end of every complete sequence "4."

These results indicate that there is another lower level of structural variability within the IGS region of Drosophila rDNA and that identical (or near-identical) IGS length variants cannot be assumed to be related by descent. On the other hand, we have found truly identical IGS length variants, based on their internal structures, which are shared by different X and Y chromosomes, isolated from different and distantly spaced Australian populations of D. melanogaster.

Variants B04 and B95 from Y chromosomes were a special case (see Table 3). Two samples of each variant showed the same number of subrepeats in each of the subrepeat arrays and the same types of junction and promoter regions, yet they showed a different bipartite structure in the 240 subrepeat arrays, as revealed by the AseI digestions (see MATERIALS AND METHODS). Such variants, having identical subrepeat structures, either arose coincidentally by different routes or they are related by descent with subsequent events differentially affecting just the internal structure of the 240-subrepeat array.

Finally, the comparison between all X-linked and all Y-linked IGS length variants analyzed in Table 3 indicates some other differences between them: (i) L-junction regions are only present in X-linked variants; (ii) Y-linked variants have no more than one unit in the 330-subrepeat family; and (iii) arrays with greater than six 95 subrepeats are more frequent in Y-linked IGS spacers (except X-B98 variant). These findings suggest a chromosomal isolation for these two regions of the IGS spacer in D. melanogaster.

Comparison of sequences in the promoter regions:
Twenty-four (8 from X chromosomes and 16 from Y chromosomes) of the 46 IGS regions analyzed by restriction mapping were sequenced from primer ETS-R. The sequences confirmed the validity of the internal structures (see Table 3). The 240-bp sequences from -175 to +95, using the transcription initiation site as +1 position (Fig 3: from sequence 4* in the promoter region to base 70 inside the ETS region), of 22 sequenced IGS regions were manually aligned and compared against the published sequence for D. melanogaster rDNA (TAUTZ et al. 1988 ). Two regions, carrying M0 promoters, were not used for alignments as these contained large deletions.

Eighteen positions show base substitutions or base insertions/deletions and each change is present in only 1 of the 22 comparisons (data not shown). Identical sequences to the published sequence were found in 10 of the IGS regions (6 from Y chromosomes and 4 from X chromosomes). Seven other IGS regions showed a single and different base change, 4 IGS regions showed two polymorphic sites, and only 1 IGS region showed three base changes. Some of these differences in sequence may have been generated during amplification and automatic sequencing procedures as single clones from PCR products have been sequenced. However, the results indicate that the promoter region, which is present in all the X-linked and Y-linked IGS regions analyzed, has an identical structure in all of them and a highly conserved sequence.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The contrasting patterns of evolution of the spacer regions of D. melanogaster rDNA (haplotypic, single-lineage model of homogenization for the ITS vs. multilineage model of homogenization for the IGS; DOVER 1994 ; SCHLOTTERER and TAUTZ 1994 ; POLANCO et al. 1998 ) can be explained if there is a mechanism that allows interchromosomal exchanges at the IGS regions to the exclusion of the ITS spacers, notwithstanding that both spacers are located in the same rDNA repeat unit.

The results presented here show that several specific length variants can be further classified into different types according to the internal organization of the various subrepeat arrays within them. Additionally, the 240-subrepeat array can be further classified according to the numbers of "A" and "B" type repeats that make up the bipartite structure of the subrepeat array, (RUIZ LINARES et al. 1994 ). Given the existence of the same, or very similar, IGS lengths having different internal structures, any assessment of the extent of sharing between X and Y rDNA arrays based only on IGS lengths would overestimate the true situation. Nevertheless, our current analysis of the distribution patterns of variant structures within given IGS lengths reveals extensive sharing between chromosomes. The variety and complexity of structures under scrutiny argues against a coincidental occurrence of the same variants in different locations. Fluctuations in the numbers of copies of repeats by unequal crossing over at three separate IGS subrepeat arrays (95s, 330s, and 240s) decrease the chance occurrence of precisely the same number of subrepeats in each array in different IGSs. Patterns of sharing at the sub-IGS level for IGSs of the same overall length are more acceptably explicable in terms of genetic exchanges between X-X and X-Y chromosomes. Where and how do these exchanges occur?

Is there a mechanistic link between rDNA exchanges and transcription?
Recent studies have demonstrated that the IGS region is not only involved in rDNA transcription but also in X-Y chromosome pairing with the strength of pairing being proportional to the number of 240 subrepeats (MCKEE et al. 1992 , MCKEE et al. 1998 ; MCKEE 1996 ; REN et al. 1997 ).

Each 240 subrepeat contains a functional RNA polymerase I promoter sequence (COEN and DOVER 1983 ; GRIMALDI and DI NOCERA 1988 ). Mutations involving only a few bases suffice to completely disable both the transcriptional and the pairing functions of 240 subrepeats (MCKEE 1998 ), indicating a direct link between the mechanisms of transcription and meiotic pairing.

Another notable feature of 240 subrepeats is the presence in each of them of a high-affinity cleavage site for topoisomerase I (CHRISTIANSEN et al. 1987 ), which might be involved in X-Y achiasmatic pairing for the formation of a stable heteroduplex, known as a "hemicatenane" (MCKEE et al. 1992 ). Because Drosophila rRNA transcripts can be initiated at the 240s "spacer promoters" (MURTIF and RAE 1985 ; TAUTZ and DOVER 1986 ), it is likely that these regions are frequently single stranded. It has been suggested that 5' ends of nascent transcripts could base pair with complementary regions of a homologous strand in the early stages of homologous alignment (MCKEE 1998 ). So the link between pairing and transcription mechanisms may have several connections: the promoter sequences at the 240 subrepeats might align the chromosomes at these regions using the nascent rRNA as a bridge, and then topoisomerase I cleavage sites could induce the chromatids to intertwine in order to maintain chromosome pairing.

Although topoisomerase I sites occur both at the 240 subrepeats and the 330 subrepeats (Fig 3, box 4), our results reveal that only topoisomerase I sites located in the 240s seem to participate in exchanges. This is to be expected if topoisomerase I activity is related to transcriptional activity.

Fragment exchanges restricted to the 240 subrepeats would mean that the rest of the IGS regions would be under the same evolutionary processes as the ITS spacers and ribosomal genes, with more frequent intra- than interchromosomal exchanges taking place, leading to a haplotypic model of homogenization. Evidence for this comes from the detection of chromosome-specific mutations, for example, the type I insertions within the 28S gene and ITS deletions, both of which are only present on X chromosomes (TARTOF and DAWID 1976 ; ENGLAND et al. 1988 ; POLANCO et al. 1998 ).

Promoter homogenization:
There is one further problem to be considered. Our results show chromosome-specific variability in copy number or structure at the 95- and 330-subrepeat families and the junction regions. Differences between X-linked and Y-linked IGS variants have also been mapped exclusively to the 95s and 330s (WILLIAMS et al. 1987 ; MARKOVA et al. 1997 ; REN et al. 1997 ). However, our sequence analysis of 22 different IGS regions revealed that the true promoter sequence adjacent to the ETS is highly homogenized both within and between X-linked and Y-linked rDNA arrays. How is it that the true promoter region is extensively shared between all chromosomes while the region between the 28S gene and 240 subrepeats shows chromosome-specific differences when, according to our model, both regions flanking the 240s are not expected to participate in X-Y chromatid exchanges?

One possibility is that the region surrounding the true promoter at the start of transcription is under intense selective pressure, eliminating any new variants, as they become homogenized. If this were the case then we would not expect Y-based true promoter regions to extensively share sequences with the X-based true promoter regions, given that the X-based rDNA is preferentially expressed (WILLIAMS et al. 1987 ; CLARK et al. 1990 ). Y-based promoters would be expected to decay over time.

Furthermore, true promoters exhibit the phenomenon of "molecular coevolution" (DOVER and FLAVELL 1984 ). This is the observation that promoters are not refractory to the occurrence and homogenization of mutations within them. Differences in promoter sequences do exist between closely related species. These have led to the spread of compensatory changes, probably promoted by selection, in the genes coding for Pol I transcription factors.

We are left, therefore, with the problem of shared promoter sequences between X- and Y-based promoters. The answer may lie in the presence of a putative topoisomerase I site that we have located at the 5' end of the 18S rRNA gene (Fig 4). This putative site is 1279 bp apart from the topoisomerase I site of the nearest 240 subrepeat in a typical P-promoter. This distance is very similar to that between five 240 subrepeats (1200 bp), the minimum number of 240s needed for a detectable stimulation of X-Y pairing. The more active true promoter sequence is located in this 1279-bp region, meaning that all the elements needed for the exchange of these regions (from the nearest 240 subrepeat to the beginning of 18S gene) would be present.

View larger version (25K): In this window In a new window Download PPT slide   Figure 4. Topoisomerase I recognition sequences in the rDNA of Tetrahymena thermophila and Drosophila melanogaster. IGS sites are described in BONVEN et al. 1985 and are present in both 240 and 330 repeats; topoisomerase I cleavage sites are marked with an arrow. The Drosophila 18S gene site is a putative recognition site of high homology with the IGS site; the cleavage site is located at position 60 from the 5' end of the gene. A and B are the interaction regions of the enzyme and the DNA described for Tetrahymena topoisomerase I. Region A includes positions -5 through -1 on the noncleaved strand and positions -7 through +2 on the scissile strand, relative to the cleavage site. Region B is delimited to positions 6–11 in both strands (CHRISTIANSEN et al. 1993 ).

Genetic exchanges between promoter regions might also explain the maintenance of unexpressed Y chromosome rDNA arrays, that is, without selection pressures. Finally, the localized exchanges could explain the unexpected bipartite distribution of A and B type 240 subrepeats described by RUIZ LINARES et al. 1994 . Such bipartite structures cannot easily be explained by random unequal crossings over, but could be expected to arise if the B type subrepeats located close to the ETS are transferred together and in higher frequency than the A type subrepeats by our proposed mechanism.

Unequal crossing over at work in the IGSs:
Unequal crossing over has long been held responsible for copy number fluctuations of the IGS internal subrepeats and of the whole rDNA unit leading to the homogenization and spread (molecular drive) of new variants through sexual populations (DOVER 1982 ).

However, unequal crossing over alone cannot explain why different regions of the IGS show different evolutionary patterns. The bizarre and ultravariable structure of Drosophila IGS regions can only be understood if additional and highly localized mechanisms of genetic exchange are at work. The proposed model of localized exchanges based on the evidence of IGS variability and the role of the IGS in X-Y chromosome pairing can go some way toward explaining the observations.

Our analysis of rDNA shows that the forces underlying concerted evolution are complex and that each multigene family can consist of regions with different evolutionary histories. Such complexity also has implications for the use of rDNA or other gene families in the construction of phylogenies. Without detailed knowledge of the behavior or misbehavior of the genes in question, errors of interpretation can occur.

	ACKNOWLEDGMENTS

This research is supported by grant no. GR3/09750 of the National Environmental Research Council of the United Kingdom awarded to G. A. Dover.

Manuscript received November 24, 1999; Accepted for publication March 27, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ASHBURNER, M., 1989 Protocol 48, pp. 108–109 in Drosophila: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

AVISE, J. C., 1994 Molecular Markers, Natural History and Evolution. Chapman and Hall, London.

BONVEN, B. J., E. GOCKE, and O. WESTERGAARD, 1985  A high affinity topoisomerase I binding sequence is clustered at DNAase I hypersensitive sites in Tetrahymena R-chromatin. Cell 41:541-551[Medline].

BOWEN, T. and G. A. DOVER, 1995  PCR amplification of intergenic spacers in the ribosomal DNA of Drosophila melanogaster reveals high levels of turnover in lengths and copy-number of spacers in geographically separated populations. Mol. Ecol. 4:419-427[Medline].

CLARK, A. G., F. M. SZUMSKI, and E. M. S. LYCKEGAARD, 1990  Population genetics of the Y chromosome of Drosophila melanogaster: rDNA variation and phenotypic correlates. Genet. Res. 58:7-13.

COEN, E. S. and G. A. DOVER, 1983  Unequal exchanges and the co-evolution of X and Y rDNA arrays in D. melanogaster.. Cell 73:849-855.

CHRISTIANSEN, K., B. J. BONVEN, and O. WESTERGAARD, 1987  Mapping of sequence-specific chromatin proteins by a novel method: topoisomerase I on Tetrahymena ribosomal chromatin. J. Mol. Biol. 193:517-525[Medline].

CHRISTIANSEN, K., A. B. D. SVEJSTRUP, A. H. ANDERSEN, and O. WESTERGAARD, 1993  Eukaryotic topoisomerase I-mediated cleavage requires bipartite DNA interaction. J. Biol. Chem. 268:9690-9701[Abstract/Free Full Text].

DOVER, G. A., 1982  Molecular drive: a cohesive mode of species evolution. Nature 299:111-117[Medline].

DOVER, G. A., 1994  Concerted evolution, molecular drive and natural selection. Curr. Biol. 4:1165-1166[Medline].

DOVER, G. A., 2000 Dear Mr. Darwin: Letters on The Evolution of Life and Human Nature. Weidenfeld & Nicolson, London.

DOVER, G. A. and R. B. FLAVELL, 1984  Molecular coevolution: DNA divergence and the maintenance of function. Cell 38:623-624.

DOVER, G. A., A. R. LINARES, T. BOWEN, and J. M. HANCOCK, 1993  The detection and quantification of concerted evolution and molecular drive. Methods Enzymol. 224:525-541[Medline].

ENGLAND, P. R., H. W. STOKES, and R. FRANKHAM, 1988  Clustering of rDNA containing type-1 insertion sequence in the distal nucleolus organizer of Drosophila melanogaster. Implications for the evolution of X-rDNA and Y-rDNA arrays. Genet. Res. 51:209-215.

GRIMALDI, G. and P. P. DI NOCERA, 1988  Multiple repeated units in Drosophila melanogaster ribosomal DNA spacer stimulate rRNA precursor transcription. Proc. Natl. Acad. Sci. USA 85:5502-5506[Abstract/Free Full Text].

HOEZEL, A. R., J. M. HANCOCK, and G. A. DOVER, 1991  Evolution of the cetacean mitochondrial D-loop region. Mol. Biol. Evol. 8:475-493[Abstract].

HOEZEL, A. R., J. V. LOPEZ, G. A. DOVER, and S. J. O'BRIEN, 1994  Rapid evolution of a heteroplasmic repetitive sequence in the mitochondrial DNA control region of carnivores. J. Mol. Evol. 39:191-199[Medline].

MARKOVA, B. A., R. S. MIRONOVA, M. L. GRANTCHAROVA, O. G. GEORGIEV, and E. P. SEMIONOV, 1997  Complex alterations of the ribosomal gene spacers in mutant sc⁸ of Drosophila melanogaster.. Chromosoma 106:361-368[Medline].

MCKEE, B. D., 1996  The license to pair: identification of meiotic pairing sites in Drosophila. Chromosoma 105:135-141[Medline].

MCKEE, B. D., 1998  Pairing sites and the role of chromosome pairing in meiosis and spermatogenesis in male Drosophila. Curr. Top. Dev. Biol. 37:77-115[Medline].

MCKEE, B. D., L. HABERA, and J. A. VRANA, 1992  Evidence that intergenic spacer repeats of Drosophila melanogaster rRNA genes function as X-Y pairing sites in male meiosis, and a general model for achiasmatic pairing. Genetics 132:529-544[Abstract].

MCKEE, B. D., K. WILHELM, C. MERRILL, and X. REN, 1998  Male sterility and meiotic drive associated with sex chromosome rearrangements in Drosophila: role of X-Y pairing. Genetics 149:143-155[Abstract/Free Full Text].

MURTIF, V. L. and P. M. M. RAE, 1985  In vivo transcription of rDNA spacers in Drosophila. Nucleic Acids Res. 13:3221-3239[Abstract/Free Full Text].

PARK, H. S. and M. T. YAMAMOTO, 1995  The centric region of the X chromosome rDNA functions in male meiotic pairing in Drosophila melanogaster.. Chromosoma 103:700-707[Medline].

POLANCO, C., A. I. GONZALEZ, A. DE LA FUENTE, and G. A. DOVER, 1998  Multigene family of ribosomal DNA in Drosophila melanogaster reveals contrasting patterns of homogenization for IGS and ITS spacer regions: a possible mechanism to resolve this paradox. Genetics 149:243-256[Abstract/Free Full Text].

REN, X., L. EISENHOUR, C. HONG, Y. LEE, and B. D. MCKEE, 1997  Roles of rDNA spacer and transcription unit-sequences in X-Y meiotic chromosome pairing in Drosophila melanogaster males. Chromosoma 106:29-36[Medline].

RUIZ LINARES, A., T. BOWEN, and G. A. DOVER, 1994  Aspects of nonrandom turnover involved in the concerted evolution of intergenic spacers within the ribosomal DNA of Drosophila melanogaster.. J. Mol. Evol. 39:151-159[Medline].

SCHLÖTTERER, C. and D. TAUTZ, 1994  Chromosomal homogeneity of Drosophila ribosomal DNA arrays suggests intrachromosomal exchanges drive concerted evolution. Curr. Biol. 4:777-783[Medline].

TARTOF, K. and I. B. DAWID, 1976  Similarities and differences in the structure of X and Y chromosome rRNA genes of Drosophila.. Nature 263:27-30[Medline].

TAUTZ, D. and G. A. DOVER, 1986  Transcription of the tandem array of ribosomal DNA in Drosophila melanogaster does not terminate at any fixed-point. EMBO J. 5:1267-1273[Medline].

TAUTZ, D., J. M. HANCOCK, D. A. WEBB, C. TAUTZ, and G. A. DOVER, 1988  Complete sequences of the rRNA genes of Drosophila melanogaster.. Mol. Biol. Evol. 5:366-376[Abstract].

WILLIAMS, S. M., G. R. FURNIER, E. FUOG, and C. STROBECK, 1987  Evolution of the ribosomal DNA spacers of Drosophila melanogaster: different patterns of variation on X and Y chromosomes. Genetics 116:225-232[Abstract/Free Full Text].

WOESE, C. R., 1991 The use of rRNA in reconstructing evolutionary relationships, pp. 1–24 in Evolution at Molecular level, edited by R. K. SELANDER, A. G. CLARK and T. S. WHITTAM. Sinauer Associates, Sunderland, MA.

This article has been cited by other articles:

				D. E. Stage and T. H. Eickbush Sequence variation within the rRNA gene loci of 12 Drosophila species Genome Res., December 1, 2007; 17(12): 1888 - 1897. [Abstract] [Full Text] [PDF]

				T. H. Eickbush and D. G. Eickbush Finely Orchestrated Movements: Evolution of the Ribosomal RNA Genes Genetics, February 1, 2007; 175(2): 477 - 485. [Abstract] [Full Text] [PDF]

				C. Li and R. C. Wilkerson Intragenomic rDNA ITS2 Variation in the Neotropical Anopheles (Nyssorhynchus) albitarsis Complex (Diptera: Culicidae) J. Hered., January 1, 2007; 98(1): 51 - 59. [Abstract] [Full Text] [PDF]

				K. T. Averbeck and T. H. Eickbush Monitoring the Mode and Tempo of Concerted Evolution in the Drosophila melanogaster rDNA Locus Genetics, December 1, 2005; 171(4): 1837 - 1846. [Abstract] [Full Text] [PDF]

				C. E. Perez-Gonzalez, W. D. Burke, and T. H. Eickbush R1 and R2 Retrotransposition and Deletion in the rDNA Loci on the X and Y Chromosomes of Drosophila melanogaster Genetics, October 1, 2003; 165(2): 675 - 685. [Abstract] [Full Text] [PDF]

				C. E. Perez-Gonzalez and T. H. Eickbush Dynamics of R1 and R2 Elements in the rDNA Locus of Drosophila simulans Genetics, August 1, 2001; 158(4): 1557 - 1567. [Abstract] [Full Text] [PDF]