Copyright 0 1987 by the Genetics Society of America

X-Y Exchange and the Coevolution of the X and Y rDNA Arrays in Drosophila melanogaster

M. R. Gillings,*?'R. Frankham,' J. Speirst and M. Whalley" "School .f Biological Sciences and tCommonwealth Scientijic and Industrial Research Organization, Division of Food Research, Plant Physiology Group, Macquarie University, New South Wales, 21 09, Australia Manuscript received August 18, 1986 Revised copy accepted March 5, 1987

ABSTRACT The nucleolus organizers on the X and Y chromosomes of Drosophila melanogaster are the sites of 200-250 tandemly repeated genes for ribosomal RNA. As there is no meiotic crossing over in male Drosophila, the X and Y chromosomal rDNA arrays should be evolutionarily independent, and therefore divergent. The rRNAs produced by X and Yare, however, very similar, if not identical. Molecular, genetic and cytological analyses of a series of X chromosome rDNA deletions (bb alleles) showed that they arose by unequal exchange through the nucleolus organizers of the X and Y chromosomes. Three separate exchange events generated compound X-YL chromosomes carrying mainly Y-specific rDNA. This led to the hypothesis that X-Y exchange is responsible for the coevolution of X and Y chromosomal rDNA. We have tested and confirmed several of the predictions of this hypothesis: First, X-YLchromosomes must be found in wild populations. We have found such a chromosome. Second, the X.YL chromosome must lose the YLarm, and/or be at a selective disadvan- tage to normal X' chromosomes, to retain the normal morphology of the X chromosome. Six of seventeen sublines founded from homozygous X. YLbb stocks have become fixed for chromosomes with spontaneous loss of part or all of the appended YL. Third, rDNA variants on the X chromosome are expected to be clustered within the X' nucleolus organizer, recently donated ("Y")forms being proximal, and X-specific forms distal. We present evidence for clustering of rRNA genes containing Type 1 insertions. Consequently,X-Y exchange is probably responsible for the coevolution of X and Y rDNA arrays.

HE ribosomal RNA genes on the X and Y chro- Type 2 (T2) insertions interrupt about 15% of the T mosomes of Drosophila melanogaster show simi- rDNA repeats on both the X and Y chromosomes. larities and differences that pose important problems These insertions are not homologous to T 1 insertions in understanding the evolution of multigene families. and have a slightly different insertion point in the 28s The X and Y chromosomes each carry 200-250 genes rRNA coding region (ROIHA and GLOVER 1980; for the major ribosomal RNAs (TARTOF1975; RI- LONG,REBBERT and DAWID1980). Repeats contain- TOSSA 19'76). These genes are tandemly repeated at ing T1 or T2 insertions do not appear to contribute the nucleolus organizers (RITOSSA and SPIEGELMAN significantly to the production of mature rRNA (LONG 1965). The tandem array consists of alternating spacer and DAWID1979b; LONGet al. 1981). and rRNA coding regions. Spacers vary in size (LONG In the absence of meiotic crossing over in male and DAWID1979a), such variation being found both Drosophila the X and Y chromosomal rDNA arrays among chromosomes, and within single nucleolus or- should be evolutionarily independent and diverge. ganizers (WELLAUER,DAWID and TARTOF1978; BON- Despite the heterogeneity outlined above, the rRNAs CINELLI et al. 1983). A portion of the rDNA repeats is interrupted in the transcribed from the uninterrupted (active) rDNA 28s coding region by one of two types of nonhomolo- repeats on the X and Y chromosomes are very similar. gous insertion sequence. Type 1 (Tl) insertions are The 28s rRNAs are identical (MADENand TARTOF restricted to the X chromosome where they interrupt 1974), and only a single base difference has been about 50% of the rDNA repeats. The DNA sequences reported between the 18s rRNA transcribed from the at rDNA/Tl junctions and the location of sequences X and Y chromosomes (YAGURA,YAGURA and MURA- homologous to T 1 insertions away from the nucleolus MATSU 1979). organizer suggests these elements may be transposable How do the X and Y rDNA arrays coevolve? Re- (KIDDand GLOVER1980; PEACOCKet al. 1981; ROIHA duced rRNA gene copy numbers in the selection lines et al. 1981 ; RAE 1981). of FRANKHAM,BRISCOE and NURTHEN(1978, 1980) were generated by unequal X-Y exchanges (transloca- ' Present address: Biological and Chemical Research Institute, New South Wales Department of Agriculture, Rydalmere, N.S.W., Australia. tions) through the X and Y chromosomal nucleolus

Genetics 116: 241-251 (June, 1987) 242 M. R. Gillings et al. organizers (COENand DOVER1983). This has led to LA LC the hypothesis that the coevolution of the X and Y rwr chromosomal rDNA is mediated by X-Y exchange. This hypothesis has the following predictions: 1. X-Y exchange through the nucleolus organizers must occur. 2. The X-YL product of X-Y exchange must be present in natural populations for the mechanism to be generally applicable. 3. To account for the normal karyotype, the X.YL chromosome must be at a selective disadvantage to the normal X chromosome, and/or the X-YL chro- mosome must be unstable. 4. rDNA variants on the X chromosome should be clustered. The X-Y exchange should result in forms donated by the Y being predominantly on the proxi- mal side of the X nucleolus organizer, while variants

specific to the X-chromosomal rDNA should be clus- SELECTION tered distally. IOW - - -- htgh -e- 5. X and Y chromosomes in long-established stocks none - should share some rDNA spacer classes, and should not have fixed differences in rRNA coding sequence variants. cantic spa 6. As TI insertions do not exist on the Y chromo- FIGURE1 .-Family tree of the abdominal bristle number selec- some, they may transpose into X chromosomal rDNA tion lines. Broad arrows indicate major X-Y exchange events and fixation of the resultant X.YL. Small arrows indicate lines that have from the nearby heterochromatin or they might even- subsequently lost all or part of the appended YL. tually be eliminated from X chromosomes by the mechanics of X-Y exchange. C(l)DX,y f/X.YL,In(l)dl-49, vof fB/YS 7. The heterochromatin of the proximal X+ and C(l)DX,yf/X.Ys, In(l)dl-49,v'f f B/YL.y3 C(l)DX,y f/X.Ys, I~(~)sc~~sc~~,y sc-/YL-scS1 proximal Ys must be homologous, as this region of the C(l)DX,yf/X-Ys, y uc wa Ct6f/YL.SCS1 Y chromosome is donated to the X chromosome by X- X*YL,y cv V/Y+ Df(Y) kl-1-/C(l)DX,y f Y exchanges. X.YL,y cv V/y+ Df(Y) kl-2-/C(I)DX,y f This paper is concerned with the evaluation of the X.YL,y cv Df(Y)k1-3-/C(l)DX, y f first four of these predictions. We present evidence X.YL, y cv V/Y+ Df (Y)kl-Z-/C(l)DX, y f X.YLYs,y w,/Y+ Df(Y) k1-3-, 4-/C(l)RM,y v bb for the occurrence of three X-Y exchanges through the rDNA, for the loss of part or all of the YL arm Stocks are described by LINDSLEYand GRELL(1968) and from X-YLchromosomes in some lines, for clustering by KENNISON(1982). rDNA variants on the chromosome and for the Generation of specific genotypes: X*/O males were gen- of X erated by crossing males from the stock of interest to presence of an X-YLchromosome in a wild population. C(l)RM,y v bb/O females. Toanalyze Y chromosomal rDNA, males from the stock of interest were crossed to BSY/C(l)DX, MATERIALS AND METHODS y f bb- := females and female progeny utilized. Y chromosome fertility factor complementation tests: Lines: The control and abdominal bristle selection lines Complementation analysis was used to test for the presence analyzed in this paper all arose from an isogenic stock of Ys or YLchromosome arms appended to the X chromo- (derived from the Canberra population) (LATTER1964) into some (X"). Males from each stock were crossed to C(l)DX/ which the fourth chromosome recessive spapol had been Ys or C(I)DX/YLvirgin females. The resultant X*/Ys or X*/ substituted. These stocks are described in detail by FRANK- YLmales were tested for fertility. HAM, BRISCOEand NURTHEN(1 978, 1980), and FRANKHAM YLfertility factor point testing: Mapping of individual (1 980). Their relationships are shown in Figure 1. CD and fertility factors on X.YL chromosomes was accomplished by CF are unselected control lines. All other lines are derived a procedure similar to that above. C(l)DX/DAY)or C(I)RM/ from the low abdominal bristle selection lines LA or LC by DAY) virgin females were crossed to males from the line of continued selection, reverse selection or relaxation of selec- interest, and the F, male progeny tested for fertility by tion. LA and LC fixed for X chromosomal bb alleles that mating to their sisters. That the FP progeny were fathered arose de novo in the lines. LA fixed for an extreme bb allele by the F1 males was confirmed by the segregation of the between generations 33 and 37, while LC fixed for a mild recessive spapo1marker in the FP. bb allele between generations 14 and 21. Mapping of the Stellate control (Ste') region: The pres- Stocks: The following stocks were used in analyses of the ence of Ste", a Y chromosomal locus mapping between kl-1 lines. and kl-2, was scored by inspection of primary spermatocytes of XI0 or X.YL/Omales. The absence of Ste' in X/O males C(l)DX,y f/X.YL,In(l)dl-49, V"'B/Y.' or by loss of the region in X.YL/O males results in the Coevolution of X and Y rDNA 243

Tranrrcribed & ___ Spacer ...._ + rDNA z8s J ~-J7q--fl-J] 28s ...I !...In.I I1 t ll I II I1 H HHE HH HHE HH FIGURE2. Organization of rDNA pDm238 I and related sequences in Drosophila melanogaster. Symbols: Boxes = rRNA coding regions, dotted boxes rDNA/Tl yI.1 I I I I1 = transcribed spacer regions, straight H H c, QA k HUE lines = spacer regions, zig-zag lines = pC225 TI insertion, wavy lines = T2 inser- tion. Restriction enzyme sites: B = BamH1, E = EcoRI, H = HindIII. AAC Fragments held as recombinant plas- mids are indicated below the restric- tion maps.

0 5 TI array 1.. I I I1 I1 I1 -kilobaw Dairs H 6 BH 6H BH appearance of crystals in the spermatocytes of these males beled by nick-translation (RIGBYet al. 1977) using a5*P- (HARDYet al. 1984; LIVAK1984). labeled dCTP. Cytology: Larval brain squashes were prepared after Hybridization was a modification of WAHL,STERN and GATTIand PIMPINELLI(1983) and stained with 0.25 pg/ml STARK'S(1979) procedure, the major change being the Hoechst 33258 (Calbiochem). inclusion of 5% dextran sulfate in the prehybridization. DNA extraction and analysis: Flies (100-1 50) were Hybridization and subsequent washes were under stringent ground in a 2 ml Dounce homogenizer with 1 ml of ROBB'S conditions. Treated filters were used to expose preflashed (1969) phosphate buffered saline (minus Mg salts). A crude X-ray film (Fuji RX) with intensifying screens (LASKEYand nuclear pellet was obtained by low speed centrifugation. MILLS 1977). Transfers were recycled, where necessary, by Nucleic acids were extracted from this pellet by resuspend- stripping the bound probe (MANIATIS,FRITSCH and SAM- ing in 200 PI 1 X SSC (SSC = 0.15 M NaCI, 0.015 M BROOK 1982) before rehybridization with other probes. trisodium citrate), and incubating with 400 PI phenol [O. 1% (w/v) 8-hydroxyquinoline and equilibrated in 1 X TE = 10 RESULTS mM Tris-HCI, 1 mM EDTA pH 7.51, and 200 pl of a lysis solution (0.15 M NaCI, 0.5 M Na perchlorate, 2% SDS, 0.1 Restriction analysis of rDNA in the selection M EDTA, pH 8.0). After centrifugation, the aqueous phase lines: The bb alleles in the two selection lines, LAbb was extracted with 400 pI of chloroform/isoamyl alcohol and LCbb, arose by X-Y exchange through the nucleo- [24:1 (v/v)] and nucleic acids precipitated by the addition of 2 volumes of cold (-20") absolute ethanol. After 0.5 hr on lus organizers. This generated compound X.YL chro- ice, the precipitate was collected, briefly dried under vac- mosomes visible at metaphase (Figure 5) (COENand uum, and resuspended overnight in 100 PI TE. The solution DOVER1983). was brought to 50 mM [Na+]and 10 PI of preboiled RNAase This section deals with the rDNA restriction pat- A (1 mg/ml) were added. The mixture was incubated 1 hr terns of all extant lines founded from LAbb or LCbb. on ice, when 200 pl of SSC were added and the temperature The initial survey was conducted using the restriction raised to 37" for 10 min. The preparation was extracted as above, ether extracted, precipitated and resuspended as enzyme EcoRI. This enzyme cuts all rDNA repeats above. All manipulations were in 1.5 ml tubes on ice, using once in the 18s rRNA coding region, and also has 15-min microfuge runs, unless otherwise indicated. three internal sites in the T2 insertion, allowing the DNA was digested with HzndIII, EcoRI or BamHI accord- three classes of rDNA repeat (uninterrupted, rDNA/ ing to the manufacturer's instructions. Enzymes were from T1 and rDNA/T2) to be easily separated by agarose BRL (Lyphozymes). XDNA was included in all genomic gel electrophoresis. rDNA repeats interrupted by T2 digests as an internal control. Genomic DNA digests (2-4 pg) were separated by electrophoresis through 0.75% aga- insertions are cleaved into a coding fragment (5.8 kb) rose and transferred to nitrocellulose (S & S BA85) by SMITH and a variable fragment (7-9 kb) largely composed of and SUMMER'S(1980) modification of SOUTHERN'S(1975) spacer sequences (Figures 2 and 3). Uninterrupted original procedure. rDNA repeats produce larger fragments (1 1-1 2 kb), Hybridization probes (Figure 2) were pC225 (T1 rDNA while rDNA repeats containing T 1 insertions produce insertion sequence clone; ROIHAand GLOVER1980) and still larger fragments (17 kb) (Figure 3). The identity pDm238 (rDNA repeat; ROIHAet aE. 1981). A 5.4 kilobase pair (kb) Hind111 fragment, gel purified (TAUTZand RENZ of each fragment class shown in Figure 3 was con- 1983) from pDm238 was used as a probe for genomic spacer firmed using probes specific to coding, spacer, T1 or sequences. This fragment contains all the rDNA spacer and T2 sequences (GILLINGS1986). a small portion of the 18 S rRNA gene (Figure 2). The The rDNA restriction patterns of representative X- presence of 18 S sequences does not interfere with the YL chromosomes are shown in Figure 3. These pat- analysis of spacer variation. EcoRI and Hind111 digests of terns are largely characteristic of Y-chromosomal pC225 and pDm238 were included as flanking marker tracks on all transferred gels. rDNA. Repeats containing T1 insertions are not Gel purified plasmids or restriction fragments were la- found on the Yf chromosomes, and are present in 244 M. R. Gillings et al. Eco R1 vs r DNA probe All LA sublines founded at, or after, generation 108 carry an additional rDNA band at 1 1.2 kb (Figure x+ XYL x+ X*YL Y+ 3, tracks e and h). This band is Y-chromosomal in n a b'c d e' 'h i j'Ik origin, but was not transferred to the X.Y'. by the initial exchanges in LA or LC. A second exchange between the LA X-Y'- and a Y chromosome must have occurred (and fixed) in the main LA line between generations 78 and 108 (see Figure 1). The number of different restriction fragments be- tween l l and 12 kb precludes an absolute assessment of the proportion of X-Y'. rDNA that is Y-chromo- soma1 in origin. These minor size differences are probably caused by polymorphism for the numbers of a 240-base pair (bp) repeat module within the spacer region of each full length repeat (WELLAUER,DAWID and TARTOF1978; BONCINELLIet al. 1983). The same is also true of rDNA containing T2 insertions, the variation in this case being more fully resolved by the EcoRI digest (Figure 3, T2 spacer halves). FIGURE5.--EcoRI digests of genomic DNA probed with rDNA Since the spacer region is apparently polymorphic, sequences (pDm238). Tracks a-j, females from the following lines: the spacers of X+, X-Y'., and Y+ chromosomes were (a) CF, (b) CD. (c) LAbb, (d) HLA78, (e) HLA108. (9 CD, (g) CF, investigated. DNA was digested with the restriction (h) LA 1 OR. (i) HLC78- I, (i) HLC78-2. Tracks k and I, C( J)DX bb-/ enzyme Hind1 I1 and the transferred fragments Y females, Y chromosome from lines (k) HLC78-1, (I) HLC78-2. probed with the HindIII fragment of pDm238 that contains the spacer region (Figure 2). This procedure only small quantities on the X-YL.There is however, confirmed that variation is generated by differing a strong rDNA/TI band in the X+ tracks, demonstrat- numbers of 240-bp subrepeats. Both X-YL and Y' ing that large numbers of these repeats were deleted chromosomes exhibit a ladder of HindIII spacer frag- by the X-Y exchange (Figure 3). ments with a periodicity of about 240 bp (Figure 4). The X+, X.Y'. and Y+ chromosomes do share some The X+ chromosomes of the control lines, CD and rDNA/T2 fragments. All carry the invariant 5.8-kb CF, carry a major spacer fragment of 5.7 kb (Figure coding fragment and the smallest spacer fragment of 4, tracks and b). Consequently, most rDNA repeats 7.4 kb. Above 7.4 kb, the X' and Y+ chromosomes a on the Xfchromosome carry this spacer A range carry rDNA/T2 fragments of differing mobilities. class. The X.YL chromosomes carry elements of both X+ of larger spacers is also present, these being the "long" and Y+ patterns, those elements characteristic of the spacers described by INDIKand TARTOF(1980). The Y being in the majority (Figure 3). Since elements of X+ chromosome does not share any spacer classes with both X+ and Y+ rDNA are present on the X.Y'-, the the Y+. original exchange breakpoints must have been within In contrast to the X+, the Y+ chromosome carries the nucleolus organizers of both X+ and Y+ chromo- five major spacer classes and a similar number of somes (see Figure 7). minor spacer variants (Figure 4, tracks i andj). Con- Analysis of a series of exposures of the autoradi- sideration of the Eco and Hin restriction sites within ographs presented in Figure 3 (and other experiments rDNA (Figure 2) suggests that the doublet observed not shown) yields the following observations. The Xc in the EcoRI digests of Y+ rDNA (Figure 3) is in fact chromosome carries a single uninterrupted rDNA composed of the five major classes, rDNA with 6.25 repeat of 11.6 kb, while the Y+ has a prominent and 6.0-kb spacers (Hin)comigrating as the "1 1.9"-kb doublet of about 1 1.9 and 1 1.2 kb. The X-YLchro- band (Em), and rDNA with 5.5-, 5.25-, and 5.0-kb mosomes of the LC lines carry the 11.9 kb Y-chro- spacers (Hin)comigrating as the " 1 1.2"-kb band (Eco). mosomal band (Figure 3, tracks i andj). All extant The X-YL chromosomes carry mainly Y rDNA LC lines exhibit this EcoRI restriction profile, with spacer classes. Only small quantities of X+ spacer only minor variations. classes (5.7 and 6.7 kb) remain on the X.Y'-, demon- The rDNA restriction pattern of LA lines founded strating that most X+ rDNA was deleted during the before generation 108 (Figure 3, tracks c and d) is X-Y exchanges. The LC lines all have a similar HindIII similar, but not identical to that of the LC lines. The spacer profile (Figure 4, representative tracks g and LA lines carry different ratios of rDNA/T2 frag- h). LA lines founded before generation 108, have a ments, establishing that the original X-Y exchanges in profile similar, but not identical, to that of the LC LA and LC were independent and different events. lines (Figure 4. tracks c and e). LA lines founded at, Coevolution of X and Y rDNA 245 Hindlll vs spacer probe some morphology. Since LAbb and LCbb fixed for X-YLchromosomes, sublines founded after this fixa- X+ X-YL Y+ 4- 4- tion should all carry X-Y'. chromosomes. To test this abcdefghii supposition, males of each stock in Figure 1 were kb kb crossed to C( l)DX/Y" females. The F, males (X.YL?/ Y")were tested for fertility. Males of each stock were also crossed to C( l)DX/Y'* females. This tested for X. Y' chromosomes and acted as a negative control. No "X" chromosome was fertile against Y', in any of the 95- tests. 8A- X.Y'-? chromosomes from 12 of the 17 LA and LC sublines were fertile against Y'. The Y'. arm of these X-YL chromosomes is therefore unchanged from that 6.7- -6.25 on the normal Y. The remaining five sublines carried X.L'-? chromosomes that were completely sterile against Y'. There are three explanations for the re- covery of X.Y'-?/Y' sterile X chromosomes from lines with known X.YL founders. There may be a point mutation in a Y'. fertility factor, a deletion of one or more fertility factors, or contamination of the stock with normal X+ chromosomes that have subsequently gone to fixation. Contamination can be ruled out, as all the lines are still homozygous for the recessive spaPo' mutant, are electrophoretically monomorphic FIGURE4.-Hindlll digests of genomic DNA probed with a and carry a Y' rDNA restriction pattern on the X-Y'-? Hindlll fragment gel-purified from the spacer region of pDm238. chromosome (GILLINCS1986). Tracks a-h. females from the following lines: (a) CD, (b) CF, (c) To map these mutations or deletions, a series of Y LA3, (d) LA108, (e) HLA78. (9 LA108. (g) HLC78-1, (h) HLC78- chromosomes with known lesions in the fertility fac- 2. Tracks i and j. C(I)DX 66-/Y females. Y chromosomes from the following lines: (i) HLA78. (i) HLAIOR. tors was employed. The principle of the complemen- tation test was to construct X-Y'*?/Y-tester males and or after, generation 108 show evidence of a second assay them for fertility. Rescue of the sterile Y-tester exchange event as suggested by the EcoRI digests. chromosome showed that the X-Y'.? chromosome car- Changes in the restriction pattern after generation ried that YL fertility factor deleted from the Y-tester 108 include the deletion of most of the remaining X' chromosome. spacers (5.7 and 6.7 kb). and the increase in molarity As expected, X-Y'. chromosomes fertile against Y' of a Y+ spacer class (5.0 kb) (Figure 4, track f). Such in the first test were also fertile against all the Y-tester observations can only be explained by a further inter- chromosomes (Table 1). Based on the complementa- chromosomal exchange (X-Y'*-Y') in the main LA line tion tests, the X-Y'.? chromosomes of the three cage between generations 78 and 108. The Hind111 digests lines, LA2, LA3 and LC 1, have lost kl-2, kl-3 and kl- therefore confirm the interpretation suggested by the 5 from the Y'- arm. The reverse selection line HLA78 EcoRI digests. has lost k1-I, kl-2 and k1-5, whereas the reverse selec- In summary, the restriction analysis has shown that tion line HLC78-2 has lost all the Y'- fertility factors the initial events in LA and LC were unequal X-Y from the X.Y'* (Table 1). The X.YL chromosomes that exchanges through the nucleolus organizers. The re- have lost fertility factors from the Y'- will be referred sultant X.YL chromosome carried mainly Y rDNA. A to as X.Df(YL). second exchange (X.Y'+-Y) occurred in the main LA Some of the X.Df(Y'-) chromosomes have lesions line between generations 78 and 108. These events between k1-1 and kl-2, the region containing the Stel- are noted on Figure 1. late control (Ste') locus. The results of tests for this Loss of YLmaterial from X-YL chromosomes: com- locus (in XI0 males) are presented in Table 1. Of the plementation analysis: X-Y exchange through the lines carrying a full Y'- arm, all except LA108 also nucleolus organizers produces compound X-Y'. chro- carry Ste'. LA 108 still carries all the YLfertility factors. mosomes (Figure 7). If such exchanges are to allow Of the X-Df(Y'-) lines, only LA2 and LA3 still carry Y' rDNA to infiltrate the X' nucleolus organizer, the Ste'. All other X-Df(Y'-) lines (LCl, HLA78 and resultant X.YL chromosomes must undergo further HLC78-2) are deficient for Ste'. exchanges with X' chromosomes, or must lose the Y'- Cytological analysis: To confirm the physical dele- arm to regenerate the normal acrocentric X chromo- tion of Y'. sequences, and to map their extent, mitotic 246 M. R. Gillings et al. TABLE 1 breakpoints, retaining the YL telomere on the newly Complementation tests between X or X.YLchromosomes and a generated X.Df(YL).All the YLdeletions were sponta- series of Y chromosomes with lesions in the YLfertility factors neous, and each newly formed X.Df(YL)chromosome replaced the original X.YL in the lines where the Y'. Function events occurred.

Stock Ykl-1- Ste' a Ykl-2- Ykl-3- Ykl-3,4- Ykl-5- X. Y' chromosome in a wild population: To test if X-U' chromosomes could be found in natural popu- Control lines lations, males from a wild type Sydney population CD Sb N S S S S CF SNS S S S were mated to C(I)DX, jf/ y+ Yk'-'- females, and the FI males tested for fertility. F1 males of one bottle LA sublines LAbb F'+FF F F were fertile, showing that at least one wild male car- LA78 F+FF F F ried an X-YL chromosome, or produced such a ga- LA108 FNF F F F mete. This X.YL chromosome was obtained after sam- HLA78 SNS F F S pling only 110 wild males. The chromosomes of this HLAlO8 F+F F F F stock were cytologically indistinguishable from the X. HLA108/121 F + F F F F LA2 F+S S S S YLof the LA and LC lines. LA 3 F+S S S S Clustering of rDNA repeat types: X-Y exchange LC sublines results in the donation of the proximal Yi nucleolus LCbb F+FF F F organizer to the X.YL. Consequently we would predict LC35 F+FF F F that rDNA/TI repeats should be clustered in the LC78 F+FF F F distal X+ nucleolus organizer. LC 108 F+FF F F LC121 F+FF F F Analysis of BamHI and EcoRI digests of X+ rDNA LC 150 F+FF F F indicates that rDNA repeats containing the major T 1 HLC78-1 F+F F F F insertion are significantly, if not totally, clustered HLC78-2 SNS S S S within the X+ nucleolus organizer. EcoRI digestion of LC 1 FNS S S S X+ rDNA generates a single major band at 17 kb that * Results of cytological tests for Ste', in X.YL/O,X.Df(Y")/O and is homologous to rDNA and TI probes (Figure 6). XI0 males. + = Ste' region present on the X chromosome, N = This band is generated from all rDNA repeats con- needle shaped crystals. S = sterile combination. taining the 5.4-kb T1 insertion. F = fertile combination. BamHI does not cut rDNA or T2 insertion se- quences, but does have two sites in the TI insertion. chromosomes of all lines were examined by Hoechst Hence discrete BamHI bands homologous to rDNA fluorescence microscopy. The results presented here must have T 1 insertions (or unrelated sequences con- agree entirely with the cytological analysis of the Y taining BamHI sites) on either end of the fragment. chromosome published by GATTI and PIMPINELLI The major BamHI fragment homologous to rDNA (1 983), in both the number and placement of Hoechst and T1 sequences is 16.1 kb (Figure 6). The only way fluorescent blocks on the YL, and the location of this fragment can be generated is from two neighbor- particular fertility factors. The numbers used to refer ing rDNA repeats, both containing T1 insertions (Fig- 1' chromosomal landmrrk; in Figure 5 arc :hac io ure 2). Since the 16.1-kb fragment is the only major suggested by GATTIand PIMPINELLI(1 983). band homologous to rDNA and T1 probes, most, if The normal X+ chromosome is an acrocentric with not all, rDNA/TI repeats must be adjacent to other a strong Hoechst positive knob over the centromere. repeats also containing T1 insertions. The numbers The compound sex chromosomes of LAbb and LCbb (and all other X.YL/Ys fertile sublines) clearly dem- of adjacent rDNA/Tl repeats can be gauged by com- onstrate that the appended arm is the YL.The loss of parison with the response to the BamHI fragments KL fertility factors from the X-YLto generate the X. (5.4-4.0 kb) generated from tandem T1 arrays in the DAYL) chromosomes is accompanied by the loss of X heterochromatin (- 100 copies) (HILLIKER,APPELS corresponding regions of the YLarm (Figure 5). The and SCHALET1980). X.Df(YL)chromosome of HLC78-2 has lost all the YL The restriction analyses in this paper also show that fertility factors and consequently the whole YLarm. It other rDNA elements must be clustered within the is morphologically similar to the X+ chromosome (Fig- X+ nucleolus organizer. A 9.5-kb Hind111 spacer class ure 5). is totally missing from the CD X+ chromosome, while The deletions of YL material from the X.YL chro- it is a prominent band in the sister control line, CF mosomes as determined by complementation and cy- (Figure 4, tracks a and b). Such an observation is most tological analysis are summarized at the base of Figure easily explained by a single exchange event deleting 5. Most, if not all the deletions have at least two all of these 9.5-kb spacers, which must then of neces- Yt 2524 232221 20 19 181716 15 14131211 109 8765432 1 x-Y

LA 108 --- LA2 :LA3 m- II.

X*Df CYL) LC1 UI

HLA 78 -0- a N

HLC 78-2 0 -0 mm FIGURE5.-Neuroblast chromosomes from X+, X.YL, and X.DAYL) lines stained with Hoechst 33258. (a) CD female; (b) LAbb male; (c) LCbb female; (d) LA2 male; (e) LA3 female; (0 LCl female; (g) HLA78 female; (h) and (i) HLC78-2 females. Bar = 5 pm. Below the photographs are maps of the Hoechst banding patterns of Y+ and X.YL chromosomes and the genetic functions thereon. Numbers used to indicate chromosomal landmarks are those suggested by CA= and PIMPINELU(1983). Intensity of stippling corresponds to intensity of fluorescence when preparations are stained with 0.25 fig/ml Hoechst 33258. c = centromere. Regions of the YL deleted from the X.YL chromosomes of some lines (as determined by complementation and cytological analysis) are summarized at the base of the figure. 247 24 8 M. R. Gillings et al. X+/X+ control females Eco R1 Bam H1 rDNA T1 rDNA T1 U exchange X-YL : I 2-c I XrDNA YrDNA U 10- of YL XMY9- w : 1 FIGURE7.-Model of X-Y exchange through the nucleolus or- ganizers and subsequent loss of the Y’. arm. organizers of these chromosomes must occur. We have documented three X-Y exchanges involving X and Y rDNA. Two of these events have been described previously (COEN and DOVER 1983), the third was detected by analysis of a more extensive set of lines. The frequency of X-Y exchange through the rDNA and subsequent fixation of the resultant X.YLbb chro- mosome has been estimated at about 2 X per gamete per generation (FRANKHAM,BRISCOE and NURTHEN1980). This is a minimum estimate; rates FIGURE &-Analysis of rDNA clustering in X’/Y’ females. Ge- of exchange through the rDNA may be four times noniic DNA was digested with EcoRl or BnmHl and sequentially probed with rDNA (pDm238) or TI (pC225) sequences. CD fe- higher than this (MADDERN1981). We have also de- males: left hand tracks of each pair, CF females: right hand tracks. tected an X-Y’. chromosome in a natural population, Sizes of fragments are in kilobase pairs. demonstrating that X-Y exchange is not confined to laboratory stocks. sity be clustered, as there is little corresponding re- X-Y exchange through the nucleolus organizers pro- duction in other spacer classes. duces a compound X.Y’. chromosome. To account for Spacer classes on the Y+ chromosome must also be the normal X chromosome morphology, the X.YL significantly clustered. Since some spacer classes on chromosome must be unstable, and/or be at a selective the Y+ (5.25 and >6.25 kb) were not transferred to disadvantage to the normal X+ chromosome. Six of the X-Y‘- (Figure 4). these spacer classes must be seventeen lines founded from homozygous X Y‘- restricted to the distal Y+ nucleolus organizer. Those stocks have become fixed for X.YLchromosomes with Y+ spacer classes present in large numbers on the X. loss of all or part of the YL arm (X-DJY’.) chromo- Y’. must be primarily located in the proximal Y+ nu- somes). This shows that the X.Y’. chromosome is cleolus organizer. prone to spontaneous loss of the Y’.. Fixation of the resultant X-DJY”) chromosomes suggests that the X. DISCUSSION Y‘, was eliminated by natural selection. There is now A model illustrating the events by which X-Y ex- direct evidence that the X-Y’. chromosome is at a change may mediate coevolution of the rDNA arrays selective disadvantage to the same chromosome lack- is shown in Figure 7. Such exchanges result in the ing the Y’. arm (R. FRANKHAM.unpublished data). donation of Y rDNA, the proximal Ys heterochroma- X-Y exchange should result in recently donated tin, the Y centromere and the Y’, to the newly formed rDNA (Y forms) being predominantly on the proximal X-Y’.chromosonle. Infiltration of Y rDNA into the Xc side of the X nucleolus organizer and X forms being rDNA pool can then occur by X-Y’--X+ exchange or clustered in the distal X nucleolus organizer. Since by loss of the Y‘. arm to generate a pseudo X chro- rDNA repeats containing T 1 insertions are only found niosome. Evidence concerning predictions arising on X chromosomes, this class of repeats should be from this model is reviewed below. clustered in the distal X nucleolus organizer. Analysis If X-Y exchange is responsible for the coevolution of rDNA/Tl repeats on the X+ chromosomes of our of X and Y rDNA, exchange through the nucleolus control lines shows that the rDNA/Tl repeats are Coevolution of X and Y rDNA 249 clustered. Published analysis of other X+ chromo- chromosomes do share rDNA spacer classes (WEL- somes is, however, equivocal. Various authors report LAUER, DAWIDand TARTOFF1978; COEN, THODAY clustering of repeats containing T 1 insertions (REN- and DOVER1982; BONCINELLIet aE. 1983), internal KAWITZ-POHL,GLATZER and KUNZ 1981; SHARP, transcribed spacers (COEN, STRACHANand DOVER GANDHIand PROCUNIER1983; KALUMUCKand PRO- 1982) and classes of T2 insertions (ROIHAet al. 1983) CUNIER 1984; SALZANOand MALVA1984), while oth- so this prediction may be satisfied. ers maintain repeats containing T 1 insertions are ran- YAGURAet al. (1979) reported a single base differ- domly distributed within the Xf nucleolus organizer. ence in the 18 S rRNA transcribed by X and Y chro- The lack of change in rDNA restriction pattern dur- mosomes. The X and Y chromosomes used in their ing rDNA magnification/reduction (DE CICCOand work also came from unrelated stocks, so this may GLOVER1983; PALUMBO,ENDOW and HAWLEY1984) simply reflect stock differences in 18 S rRNA coding or compensation (DUTTONand KRIDER 1984) has sequences. Further, the 28 S rRNAs transcribed by X been used to argue that different rDNA repeats are and Y chromosomes are reputed to be identical randomly dispersed within the X+ chromosomal nu- (MADENand TARTOF1974). cleolus organizer. However, these interpretations are The prediction above does not preclude the exist- subject to assumptions about the mechanisms of mag- ence of rDNA variants unique to either the X or Y nification and compensation that are still open to chromosomes of wild-type stocks. We would predict question. In fact, M. R. GILLINGS(unpublished data) that such unique variants should lie in regions not has also failed to find changes in restriction patterns regularly participating in X-Y exchange (i.e., in the during magnification and compensation experiments distal portion of X and Y nucleolus organizers). using one of our lines, even though rDNA variants As T 1 insertions do not exist on the Y chromosome, carried by this line are significantly clustered (GILL- we must predict that they transpose into X chromo- INGS 1986). somal rDNA from the nearby T1 arrays in the X Analysis of EcoRI and BamHI rDNA restriction heterochromatin, otherwise they would eventually be patterns led HAWLEYand TARTOF(1983) to suggest eliminated by the mechanics of X-Y exchange. No that rDNA/T 1 repeats were randomly distributed. evidence for transposition of T1 sequences from the However, they did not confirm that their high molec- X chromosomal heterochromatin to the rDNA has ular weight rDNA bands were in fact homologous to been found in our X.YL lines (GILLINGS1986), con- type 1 probes. We have shown that most of the high flicting with expectations. However, the YL arm may molecular weight material is not significantly homol- be inhibiting transposition, thereby also preventing ogous to type 1 sequences. Close examination of HAW- transposition into Yf rDNA in X’/Y’ males. Alterna- LEY and TARTOF’Sdata shows that the 17-kb EcoRl tively, the rate of transposition may be too low to have band and the 16.1-kb BamHI band account for equal been detected, insertions may have occurred but not proportions of the total rDNA response. Since the 17- altered the restriction pattern by a detectable amount, kb EcoRI bands are generated from rDNA repeats or “dysgenic” crosses may be required to mobilize containing the 5.4-kb T1 insertion, but the 16.1-kb these elements. BamHI band is only produced from clusters of these The proximal heterochromatin of the Xf and Ys repeats, the equality of response to the Bum and Eco must show homology, as this region of the Y chromo- bands in HAWLEYand TARTOF’Sexperiment shows some is donated to the X+ chromosome during X-Y that most rDNA/Tl repeats are adjacent to other exchange. X and Y chromosomes do share satellite similar repeats. Despite assertions to the contrary, sequences in these regions (PEACOCKet al. 1978; STEF- rDNA/T 1 insertions are most probably clustered in FENSEN, APPELSand PEACOCK1981). However, se- all X+ nucleolus organizers. In situ hybridization in- quences influencing nucleolar dominance in D.melan- dicates that most rDNA repeats containing type 1 ogaster x D. simulans hybrids have been mapped to insertions are located in the distal region of the Xf the proximal X heterochromatin, but to a different nucleolus organizer (HILLIKERand APPELS1982), as region of the Y chromosome (DURICAand KRIDER predicted by the X-Y exchange hypothesis. 1978). Close examination of the data in their paper The X-Y exchange hypothesis predicts that X and Y suggests that such dominance is more easily explained chromosomes in long established stocks should share by dependence on the rDNA region itself. This is in some rDNA spacer classes, coding sequence classes accord with what is now known about the behavior of and classes of type 2 insertions. The X+ and Y+ chro- the rDNA spacer in cross-species transcription systems mosomes studied here carry no spacer classes in com- (COENand DOVER 1982; KOHORN and RAE 1982; mon. While this made the X-Y exchange easy to detect, MILLER, HAYWARDand GLOVER1983; REEDERand it conflicts with the prediction. However, these sex ROAN 1984). Whatever the case, DURICAand KRI- chromosomes came from unrelated stocks, perhaps DER’S experiments merit reinvestigation. explaining this anomaly. In other stocks, X and Y The 1.688 satellite of D.melanogaster was originally 250 M. R. Gillings et al. reported as being located primarily on the sex chro- in rDNA spacers of Drosophila melanogaster. Nucleic Acids Res. mosomes (PEACOCKet al. 1978). A more recent report 10 7017-7026. COEN,E. S. and G. A. DOVER,1983 Unequal exchanges and the using a cloned 1.688 sequence as a probe suggests coevolution of X and Y chromosomal rDNA arrays in Drosoph- that this satellite may be restricted to the proximal X ila melanogaster. Cell 33: 849-855. heterochromatin (D. L. BRUTLAGpersonal communi- COEN, E. S., T. STRACHANand G. DOVER, 1982 Dynamics of cation, cited in HILLIKERand APPELS1982), an obser- concerted evolution of ribosomal DNA and histone gene fam- vation that would be fatal to the X-Y exchange hy- ilies in the melanogaster species subgroup of Drosophila. J. Mol. Biol. 158 17-35. pothesis as presented in Figure 7. If this were the COEN,E. S., J. M. THODAYand G. DOVER,1982 Rate of turnover case, for the hypothesis to still be tenable, the coevo- of structural variants in the rDNA gene family of Drosophila lution of X and Y rDNA would have to proceed by melanogaster. Nature 295: 564-568. further exchange of X-YL chromosomes with normal DE CICCO,D. V. and D. M. GLOVER,1983 Amplification of rDNA X chromosomes, thereby transferring some Y rDNA and T1 sequences in Drosophila males deficient in rDNA. Cell 32: 1217-1225. into the X nucleolus organizer. However, the resolu- DURICA,D. S. and H. M. KRIDER,1978 Studies on the ribosomal tion of in situ hybridization to Drosophila mitotic chro- RNA cistrons in interspecific Drosophila hybrids. 11. Hetero- mosomes is less than ideal, and the discrepancies be- chromatic regions mediating nucleolar dominance. Genetics tween cloned probes and those purified from gra- 89 37-64. dients may be caused by sequence complexity of the DUTTON,F. L. and H. M. KRIDER,1984 Ribosomal RNA cistrons of X-chromosomes clonally derived from D. melanogaster lab 1.688 satellite. While adjacent repeats are highly ho- oratory populations: redundancy, organization and stability. mologous, independent 1.688 satellite clones can have Genetics 10% 405-42 1. as little as 60% homology (MIKLOSand GILL 1982). If FRANKHAM,R., 1980 Origin of genetic variation in selection lines. variants of the 1.688 satellite were present on the X pp. 56-68. In: Selection Experiments in Laboratory and Domestic and Y chromosomes divergence of these sequences Animals, Edited by A. ROBERTSON.Commonwealth Agricul- tural Bureaux, Slough. would be expected, with resulting complications in FRANKHAM,R., D. A. BRISCOE and R. K. NURTHEN, 1978 hybridization analysis. Unequal crossing over at the rRNA locus as a source of quan- This paper presents evidence that the coevolution titative genetic variation. Nature 272: 80-81. of X and Y rDNA proceeds via X-Y exchange. Are FRANKHAM,R., D. A. BRIsCoEand R. K. NURTHEN,1980 Unequal there other mechanisms that could explain this co- crossing over at the rRNA tandon as a source of quantitative genetic variation. Genetics 95: 727-742. evolution? While the processes of natural selection GATTI, M. and S. PIMPINELLI,1983 Cytological and genetic anal- and gene conversion undoubtedly influence rDNA ysis of the Y chromosome of Drosophila melanogaster. I. Orga- evolution, neither can adequately explain the coevo- nization of the fertility factors. Chromosoma 88: 349-373. lution of X and Y rDNA. It is highly unlikely that GILLINGS,M. R., 1986 Evolution of ribosomal RNA genes on the selection could fix the same mutants within X and Y X and Y chromosomes of Drosophila melanogaster. Ph.D. thesis, Macquarie University. rDNA arrays, and the presence of T1 rDNA inser- HARDY,R. W., D. L. LINDSLEY,K. J. LIVAK,B. LEWIS, A. L. tions on the X, but not the Y chromosome argues SILVERSTEIN,G. L. JOSLYN,J. EDWARDSand S. BONACCORSI, against gene conversion operating between these 1984 Cytological analysis of a segment of the Y chromosome chromosomes. We suggest that occasional X-Y ex- of Drosophila melanogaster. Genetics 107: 59 1-6 10. change events allow Y rDNA to infiltrate the Y chro- HAWLEY,R. S. and K. D. TARTOF,1983 The ribosomal DNA of Drosophila melanogaster is organized differently from that of mosome nucleolus organizer via an X.YL intermedi- Drosophila hydei. J. Mol. Biol. 163: 499-503. ate. We have confirmed several predictions arising HILLIKER,A. J. and R. APPELS,1982 Pleiotropic effects associated from this hypothesis. Other predictions are currently with the deletion of heterochromatin surrounding rDNA on being evaluated. the X chromosome of Drosophila. Chromosoma 86: 469-490. HILLIKER,A. J., R. APPELSand A. SCHALET,1980 The genetic We are grateful to the following for the provision of Drosophila analysis of D. melanogaster heterochromatin. Cell 21: 607-6 19. stocks: The Mid-America Drosophila Stock Center, Bowling Green INDIK,2. K. and K. D. TARTOF,1980 Long spacers among ribo- State University, Ohio; Division of Biology, California Institute of somal genes of Drosophila melanogaster. Nature 284 477-479. Technology and the Department of Zoology, Indiana University, KALUMUCK,K. E. and J. D. PROCUNIER,1984 Non-selective am- Bloomington, Indiana. Drs. R. APPELS,E. S. COENand D. GLOVER plification of ribosomal DNA repeat units in compensating kindly supplied us with recombinant plasmids. This investigation genotypes of Drosophila melanogaster. Biochem. Genet. 22: was supported by the Australian Research Grants Scheme and 453-465. Macquarie University Research Grants. M.R.G. was the recipient KENNISON,J. A., 1982 Analyses of Y linked mutations to male of a Commonwealth Postgraduate Research Award. sterility in Drosophila melanogaster. Genetics 103: 219-234. KIDD, S. J. and D. M. GLOVER,1980 A DNA segment from D. melanogaster which contains five tandemly repeating units ho- LITERATURE CITED mologous to the major rDNA insertion. Cell 19103-119. KOHORN,B. D. and P. M. M. RAE, 1982 Non-transcribed spacer BONCINELLI,E., A. BORGHESE,F. GRAZIANI,A. LA MANTIA,A. sequences promote in vitro transcription of Drosophila ribo- MANZI,C. MARIANIand A. SIMEONE,1983 Inheritance of the somal DNA. Nucleic Acids Res. 106879-6886. rDNA spacer in D. melanogaster. Mol. Gen. Genet. 189: 370- LASKEY,R. A. and R. D. MILLS, 1977 Enhancedautoradiographic 374. detection of '*P and Iz5l using intensifying screens and hyper- COEN,E. S. and G. A. DOVER,1982 Multiple Pol 1 initiation sites sensitised film. FEBS Lett. 82: 314-316. Coevolution of X and Y rDNA 25 1 LATTER,B. D. H., 1964 Selection for a threshold character in RIGBY, P. J. W., M. DIECKMAN,C. RHODES and P. BERG, Drosophila. I. An analysis of the phenotypic variance on the 1977 Labelling deoxyribonucleic acid to high specific activity underlying scale. Genet. Res. 5 198-210. in vitro by nick translation with DNA polymerase 1. J. Mol. LINDSLEY,D. L. and E. H. GRELL, 1968 Genetic variations of Biol. 113: 237-25 1. Drosophila melanogaster. Carnegie Inst. Wash. Publ. 627. RITOSSA,F., 1976 The bobbed locus. pp. 801-846. In: Geneticsand LIVAK,K. J., 1984 Organization and mapping of a sequence on Biology of Drosophila, Vol. lb, Edited by M. ASHBURNERand E. the Drosophila melanogaster X and Y chromosomes that is NOVITSKI.Academic Press, London. transcribed during spermatogenesis. Genetics 107: 6 1 1-634. RITOSSA,F. and S. SPIEGELMAN,1965 Localization of DNA com- LONG,E. 0. and 1. B. DAWID,1979a Restriction analysis of spacers plementary to RNA in the nucleolus organizer region of Dro- in ribosomal DNA of Drosophila melanogaster. Nucleic Acids sophila melanogaster. Proc. Natl. Acad. Sci. USA 53: 737-745. Res. 7: 205-215. ROBB,J. A., 1969 Maintenance of imaginal discs of Drosophila LONG,E. 0. and 1. B. DAWID,1979b Expression of ribosomal melanogaster in chemically defined media. J. Cell. Biol. 41: 855- DNA insertions in Drosophila melanogaster. Cell 18: 1185- 876. 1196. ROIHA,H. and D. M. GLOVER,1980 Characterisation of complete LONG, E. O., M. COLLINS, B. I. KIEFER and I. B. DAWID, type I1 insertions in cloned segments of ribosomal DNA from 1981 Expression of the ribosomal DNA insertions in bobbed Drosophila melanogaster. J. Mol. Biol. 140 341-355. mutants of Drosophila melanogaster. Mol. Gen. Genet. 182: ROIHA, H., J. R. MILLER, L. C. WOODSand D. M. GLOVER, 377-384. 1981 Arrangements and rearrangements of sequences flank- LONG,E. O., M. L. REBBERTand I. B. DAWID,1980 Structure ing the two types of rDNA insertion in D. melanogaster. Nature and expression of ribosomal RNA genes of Drosophila melano- 290: 749-753. ROIHA, H., C. A. READ, M. J. BROWNEand D. M. GLOVER, gaster interrupted by type 2 insertions. Cold Spring Harbor 1983 Widely differing degrees of sequence conservation of Symp. Quant. Biol. 45 667-672. the two types of rDNA insertion within the melanogaster species MADDERN,R. H., 1981 Exchange between the ribosomal RNA subgroup of Drosophila. EMBO J. 2: 721-726. genes of the and Y chromosomes of Drosophila melanogaster X SALZANO,G. and C. MALVA,1984 Non-random loss of uninter- males. Genet. Res. 38: 1-7. rupted ribosomal DNA repeating units upon induction of a MADEN,B. E. H. and K. D. TARTOF,1974 Nature of the ribosomal bobbed mutation. J. Mol. Biol. 177: 189-200. RNA transcribed from the X and Y chromosomes of Drosophila SHARP,2. D., V. V. GHANDIand J. D. PROCUNIER,1983 X melanogaster. J. Mol. Biol. 90 51-64. chromosome nucleolus organizer mutants which alter major MANIATIS,T., E. F. FRITSCHand J. SAMBROOK,1982 Molecular type I repeat multiplicity in Drosophila melanogaster. Mol. Gen. Cloning: A Laboratory Manual. Cold Spring Laboratory, Cold Genet. 190 438-443. Spring Harbor, New York. SMITH,G. E. and M. D. SUMMERS,1980 The bidirectional transfer MIKLOS,G. L. G. and A. G. GILL, 1982 Nucleotide sequences of of DNA and RNA to nitrocellulose or diazobenzyloxymethyl highly repeated DNAs; compilation and comments. Genet. Res. paper. Anal. Biochem. 109 123-129. 39: 1-30. SOUTHERN,E. M., 1975 Detection of specific sequences among MILLER,J. R., D. C. HAYWARDand D. M. GLOVER,1983 DNA fragments separated by gel electrophoresis. J. Mol. Biol. Transcription of the “non-transcribed” spacer of Drosophila 98: 503-5 17. melanogaster. Nucleic Acids Res. 11: 1 1- 19. STEFFENSEN,D. M., R. APPELS and W. J. PEACOCK,1981 The PALUMBO,G., S. A. ENww and R. S. HAWLEY,1984 Reduction distribution of two highly repeated DNA sequences within of wild type X chromosomes with the Yb6-chromosome of D. Drosophila melanogaster chromosomes. Chromosoma 82: 525- melanogaster. Genet. Res. 43: 93-98. 541. PEACOCK,W. J., A. R. LOHE,W. L. GERLACH,P. DUNSMUIR,E. S. TARTOF,K. D., 1975 Redundant genes. Ann. Rev. Genet. 9 355- DENNISand R. APPELS,1978 Fine structure and evolution of 385. DNA in heterochromatin. Cold Spring Harbor Symp. Quant. TAUTZ,D. and M. RENZ, 1983 An optimized freeze-squeeze Biol.45: 1121-1135. method for the recovery of DNA from agarose gels. Anal. PEACOCK,W. J., R. APPELS, S. ENDOWand D. GLOVER,1981 Biochem. 132: 14-19. Chromosomal distribution of the major insert in Drosophila WAHL,G. M., M. STERNand G. R. STARK,1979 Efficient transfer melanogaster 28s rRNA genes. Genet. Res. 37: 209-214. of large DNA fragments from agarose gels to diazobenzloxy- RAE, P. M., 1981 Coding region deletions associated with the methal paper and rapid hybridization using dextran sulphate. major form of rDNA interruption in Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 76 3683-3687. Nucleic Acids Res. 9 4997-5010. WELLAUER,P. K., I. B. DAWID and K. D. TARTOF,1978 X and Y REEDFX,R. H. and J. G. ROAN,1984 The mechanism of nucleolar chromosomal ribosomal DNA of Drosophila: comparison of dominance in Xenopus hybrids. Cell 38: 39-44. spacers and insertions. Cell 14: 269-278. YAGURA,T., M. YAGURAand M. MURAMATSU,1979 Drosophila RENKAWITZ-POHL,R., K. H. GLATZERand W. KUNZ, 1981 has different ribosomal RNA sequences on and Ribosomal RNA genes with an intervening sequence are clus- melanogaster X Y chromosomes. J. Mol. Biol. 133: 533-547. tered within the X-chromosomal ribosomal DNA of Drosophila hydei. J. Mol. Biol. 148: 95-101. Communicating editor: M. T. CLEGG