Proc. Natl. Acad. Sci. USA

Vol. 89, pp. 7591-7595, August 1992

Genetics

Degenerating Y chromosome of Drosophila miranda: A trap for retrotransposons (chromosome structure/larval cuticle protein genes)

MANFRED STEINEMANN AND SIGRID STEINEMANN Institut fur Genetik und Mikrobiologie, Universitat Munchen, Maria-Ward-Str. la, D-8000 Munich 19, Federal Republic of Germany

Communicated by M. M. Green, April 10, 1992 (received for review February 25, 1992)

ABSTRACT In Drosophila mirnda, the larval cuticle protein (Lcp) genes are located on the X2 and Y chromosomes, while in other Drosophila species the Lcp genes are inherited on the autosomes. We chose the D. m yad species as a model system to analyze the molecular bases of Y chromosome degeneration, a phenomenon observed in many species. DNA sequence analysis of the Y chromosomal Lcp gene locus reveals dense clustering of trapped retrotransposons. Once inserted at the Y chromosomal location they cannot easily be eliminated by unequal crossingover, as recombination is a rare event in Drosophila males. In addition, we have uncovered an example of a completely inactive allele on the degenerating Y chromosome. The existence of such inactive Y-specific alleles was originally predicted in H. J. Muller's model for Y chromosome degeneration. We demonstrate that the Y chromosomal Lcp4 allele is no longer transcribed. From the divergence in DNA sequence organization of former homologous chromosome regions we conclude that changes in chromosome structure and destruction of genetic activity in degenerating Y chromosomes are based on one major mechanism, which operates by means of transposable elements.

chromosomal. In D. miranda, the larval cuticle protein gene cluster (here we follow the locus designation Lcp of ref. 18) is located on the X2 and Y chromosomes, while it is autosomally inherited in the two sibling species Drosophila pseudoobscura and Drosophila persimilis. Sequence analysis of the Lcp locus cloned from the X2 and Y chromosomes* reveals a massive accumulation of inserted DNA sequences (ISYs) in the Y chromosomal Lcp region. In situ hybridizations with the TRIM retrotransposon (17) and the ISY3 insertions in fact show a differential distribution of these transposable elements on the X2 and neo-Y chromosomes. By Northern analysis, we show that the Y chromosomal Lcp4 allele is no longer transcribed. Thus, we have uncovered a loss-of-function allele on the degenerating Y chromosome, as was originally predicted in Muller's model for the degeneration of the Y chromosome (1, 2).

MATERIALS AND METHODS Cloning, Standard DNA Techniques, and Sequencing. The X2 and Y chromosomal Lcp regions were cloned and sequenced by standard techniques (16, 19). We sequenced both strands from M13mp18/19 subclones covering the X2 and Y chromosomal Lcp regions according to the protocol supplied with Sequenase (United States Biochemical). Localization of the ISY3 Sequences. For chromosome in situ hybridizations squashing of salivary glands was done as described in ref. 15. To detect ISY3 homologous sequences we used an Xba I fragment containing the ISY10 DNA sequence, which is about 95% identical to part ofthe ISY3 sequence (see Fig. 3). The replicative form (RF) DNA of M13 phage carrying the Xba I fragment was biotinylated with Bio-16-dUTP (Enzo Biochem) by nick translation. The probes were hybridized at 580C overnight and washed at 530C. Signal detection followed the protocol for immunoperoxidase staining supplied with the DETEK I-hrp kit used (Enzo Biochem) and the silver 3,3'diaminobenzidine (DAB) enhancement kit (Amersham). The slides were stained with Giemsa stain. RNA Isolation and Northern Analysis. Undegraded total RNA from larval integument was isolated from 200 handdissected male and female integuments (20). Denaturing agarose gels and Northern analysis were as detailed in ref. 19.

Changes in chromosome structure and breakdown of genetic activity in evolving Y chromosomes, referred to as chromosome degeneration (1, 2), are observed in many species. Several models have been proposed to explain the phenomena (3-8). Muller (1, 2) assumed that X and Y chromosomes evolved from a pair of originally homologous autosomes (cf. refs. 9-12). However, the molecular aspects of Y chromosome evolution are still an enigma. According to the hypothesis of Muller (1, 2), the differentiation of sex chromosomes is due to the progressive degeneration of the Y chromosome as a consequence of its permanent heterozygosity. Chromosome degeneration is thought to be due to mutation of genes to recessive or completely inactive states. The primeval X and Y chromosomes probably have comparable rates of recurrent mutations. As a consequence, the frequency of recessive lethal mutations should be the same for the X and Y loci. However, such mutations on the Y are effectively neutral because they can never become homozygous. Thus, recurrent mutations from wild-type to recessive states over a sufficiently long period of evolutionary time should lead to the fixation of recessive loss-of-function alleles at most loci. To analyze the molecular bases of degeneration we chose a model system, Drosophila miranda, that is characterized by a pair of heteromorphic sex chromosomes which are evolving from a pair of formerly homologous autosomes. D. miranda shows an exceptional karyotype. Due to the translocation of one of the autosomes to the Y chromosome, a neo-Y chromosome and a monosome, designated X2, were formed (13-17). Thus, formerly autosomal genes are now X2 and Y

RESULTS X2 and Y Chromosomal Lcp Regions. On the basis of known chromosome homologies (16), we decided to analyze the Lcp gene cluster, which encodes the larval cuticle proteins LCP1LCP4 (in the original literature CP1-CP4; for consistency we suggest LCPs). These genes map to 44D on the right arm of chromosome 2 in Drosophila melanogaster (21) and to chroAbbreviations: ISY, DNA insertion sequences detected on the Y chromosome; DY, DNA sequence deletions detected on the Y chromosome; LINE, long interspersed element; LTR, long terminal

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

repeat. *The sequences reported in this paper have been deposited in the GenBank data base (accession nos. M94251, M94252, and M94253). 7591

7592

Genetics: Steinemann and Steinemann

Proc. Natl. Acad Sci. USA 89 (1992)

mosome 3 in D. pseudoobscura and D. persimilis. In D. miranda this cluster is found on the X2 and Y chromosomes (Fig. 1). We cloned the Lcp gene cluster from both chromosomal locations in D. miranda (16). The four Lcp genes clustered on the X2 chromosome contain about 7 kb of sequence information (Fig. 2). The gene arrangement and direction of transcription are similar to those in the D. melanogaster Lcp cluster (22), despite the evolutionary distance of about 30 million years between the two species (23). With the exceptions indicated by the arrows in Fig. 2 we sequenced about 14 kb of the Y chromosomal Lcp region. Alignment of the X2 and Y chromosomal sequences revealed five short deletions on the Y (DYs), and one more extended deletion, DY3, of 221 base pairs (bp), which is found in the coding region of the Y allele of Lcp4 (see below). Even more strikingly, 10 Y chromosomal insertions (designated ISY sequences) that are absent from the X2 region were found (Fig. 2). We have also detected a duplication of the Lcp2 region, including associated ISYs (see below). Except for the tandem duplica.. f#:

Z, i,.

'.LcPf.

.

i

"

B

so

I

Eum

LCP

X2 c

D a

RS H RRXBSXB . .

b XH

S

SRBSS

B

HH

H B H XH X R HRH 1 1 l

H HHR B BXHH . ....

B XHH

**.. .1i1 Il

HBH

[L_ .L-

. I

B

RX

2

B

HXHXHHRHXXHXXHXH

FIG. 1. (A-C) Alignment of homologous polytene chromosomes from salivary gland nuclei of D. melanogaster and D. miranda. (A) Chromosome 2R from D. melanogaster. (B) X2 chromosome from a D. miranda male. (C) Y chromosome from a D. miranda male. The location of the Lcp gene cluster is indicated in A and B. Due to the fact that in situ hybridizations with the Lcp probes label the X2 but not the Y chromosome (16), we are unable to determine the precise location of the Lcp cluster on the latter. (Bars indicate 20 Aum.) (D) Alignment of the X2 and Y chromosomal Lcp regions ofD. miranda. Restriction maps of cloned Lcp regions from the X2 chromosome (a) and Y chromosome (b) ofD. miranda. Fragments showing homology to Lcpl and Lcp2 are indicated by stippled boxes; those showing homology to Lcp3 and Lcp4, by hatched boxes. kb, Kilobases. The restriction map of the Y chromosome indicates the striking DNA rearrangements. Some restriction sites are indicated for orientation: S, Sal I; H, HindIII; R, EcoRI; B, BamHI; X, Xba I.

tion, ISY4 (3.1 kb) and ISY5 (about 2.5 kb) are the largest insertions found in the Y Lcp region. These two insertions are positioned between the Y alleles of Lcp2 and Lcp3 and between Lcp3 and Lcp4, respectively. ISY4 represents a retrotransposon, named TRIM (17), which shares homology with the Ifactor of D. melanogaster (24) and retrotransposons belonging to the long interspersed elements (LINEs) family. Both are of the non-long terminal repeat (LTR) type. The TRIM integration is associated with the small deletion DY2. On the other hand the ISY5 insertion, designated TRAM, is not associated with a deletion. TRAM has integrated 433 bp upstream ofthe transcription start site of the Y allele of Lcp4. We sequenced both ends of TRAM, about 1.5 kb in all (unpublished data). This sequence information reveals a target site duplication that flanks perfect LTRs. In addition we find a putative tRNA primer-binding site (PBS), necessary for the initiation of DNA minus-strand synthesis in retroelements (not shown). The PBS of TRAM shows a sequence identical to that in Rous sarcoma virus (RSV), which is complementary to the 3' end of tRNATrP (25). Thus, TRAM represents a second retrotransposon that has integrated within the 7 kb of Lcp sequences present in the autosomal progenitor of the Y chromosome. The retrotransposon insertions TRIM (ISY4) and TRAM (ISY5) have integrated in opposite directions, so that both 5' ends are facing Lcp3. Alignment of the TRIM and TRAM sequences shows no homology. Data base screening (Oct. 1991) with the LTR sequence ofTRAM also revealed no similarity to known retrotransposons. Disregarding the tandem duplication, ISY3 (713 bp) is the third-largest insert (Fig. 3), and is in the Y chromosomal Lcp region positioned between Lcpl and Lcp2. This insert has small open reading frames. All insertions, except for ISY6 (see below), lie outside of the coding regions of the four Lcp Y alleles. The sizes of the smaller inserts are as follows: ISYI, 25 bp; ISY2, 248 bp; ISY6, 11 bp; ISY7, 6 bp; ISY8, 16 bp; ISY9, 141 bp; and ISY10, 357 bp. Data base screening (Oct. 1991) for sequences homologous to ISYI, GAGAGAG CAT CGACGGCGCC, was negative. The significance of the similarity of the underlined motif with the octamer motif present in mammalian transcriptional enhancer elements-e.g., U2 small nuclear RNA enhancer, ATTTGCATGCCC (26)-is unclear. The ISY2 insertion shows homology to a DNA sequence that lies at the 3' end of gene 4 (not shown) and reveals sequence homologies to dispersed repetitive DNA sequences in Drosophila virilis (27) and to a heterogeneous family of mobile sequences in the genome of sea urchins (28). ISY6 contains a DNA sequence duplication of 8 bp (see below) and an additional insert of 3 bp. ISY7 is 6 bp long and is the smallest insertion recorded. An identical 6-bp motif occurs at the beginning of the ISY3 sequence (underlined in Fig. 3). The ISY8 insertion, TTTATCTAAGATAAGT, is bordered by two small deficiencies, DY5 (37 bp) and DY6 (22 bp). The reverse complementary sequence of ISY8 overlaps with one end of ISY3 (underlined in Fig. 3), showing 90.90% identity over 11 bp. ISY9 and ISY10 show DNA sequence identities to ISY3, but rearranged in position, and show in these overlapping regions an accumulation of point mutations (Fig. 3). ISY9 shows 74.4% identity in a 148-bp overlap, and ISY10 reveals 94.7% identity in a 357-bp overlap with ISY3 (Fig. 3). ISY10 is followed by a tandem duplication starting with the ISY3 insertion. The duplication contains an additional copy ofthe Y allele of gene 2 and the TRIM element (unpublished data). Clustered Inserts In the Y Chromosome. The ISY3 insertion element and its derivatives are present in more than 100 copies in the female nucleus (Fig. 4A). However, comparison of the labeling of the former homologues, X2 and Y. reveals a dramatic difference in label density. We find a massive relative accumulation of these sequences on the Y chromosome (Fig. 4B). Preferential localization on the Y chromo-

_ i . -ois_

Genetics: Steinernann and Steinernann

Proc. Natl. Acad. Sci. USA 89 (1992)

7593

1 kb LJ

Sa

X2

C I

HHCSa II I

C 1E

I

GU

H XhH

IN

-,

_

Sa

1

Gene 2

Gene 1

TRIM

ISV4

R C

H

El

I I

Sa XhXh E1

I

B

MEM

B> Gene 4

Gene 3

.....:,DUPLICATION

ISY2

ISY5

ISY3

: : IS1...I .............

a

s~~~~~~~ 0 %f'7 ISY7.

z

y

c

HHCSa .. 11

I

. I .

Gene 1 DY1

Sa

C

Sa

H

0

XhH

H

El mm

I

Gene 2 DY2

.............

_

ISY6

Isyl

A

B RC Sa I 0 I0 Ia I

XhXh U

El

Gene 3

DeY

>

n n

DY4 A

some is also shown by the TRIM retrotransposon (17). The observed biases in distribution, in favor of the Y chromosomal location, suggest that the accumulation of transposable elements on the Y chromosome is not disadvantageous, and thus selection cannot eliminate it. A Loss-of-Function Lcp4 Aflele on the Y Chromosome. The largest deletion analyzed in the Y chromosomal Lcp region occurs within the coding sequence of the Y allele of Lcp4. Alignment of the sequences of X2 and Y alleles of Lcp4 uncovers a deletion of 221 bp on Y (Fig. 5). The deletion is flanked by the DNA sequence duplication, CGGAATTT, most probably representing the footprint left behind by an insertion/excision event. The origin of the additionally inserted TTT is unclear (Fig. 5). The deletion extends beyond

Y........... Genx

~~\DY6

the TAG stop codon of the X2 allele but ends 65 bp before the poly(A) addition signal. The sequence of the Y chromosomal allele contains a potential TAG stop codon (Fig. 5 underlined). The truncated Y allele Lcp4 still contains 5' upstream consensus regulatory sequences, such as the CAAT and the TATA motifs, together with the short intron and the poly(A) addition signal (Fig. 5). If this allele is still functional we should expect, in addition to the mRNA transcribed from the X2 allele (about 550 nucleotides), a second truncated species of about 330 nucleotides, 221 nucleotides shorter, which is derived from the Y allele. Northern analysis of total RNA isolated from third-instar larval integuments of females and males reveals only one mRNA species of the size expected for a transcript of the X2 allele (Fig. 6). Therefore, the Y

ISY3

TCTAGAGTGTAAAATCTCAACCAGATCGTATAATTATTATAGCCAGAATCAAGAAAACAA

ISYlO

TCTAGAGTGTAAAATCTGAACCAGATCGTATAATTATTATAGCCAGAATCAQGAAAACAA

ISY3

TTTCATTCTTTCTCGCTTTGTCTCTCTCTAACACACAGGTTTCATGGTCGGCTTTGCCAA

ISY9

--

---

60

120

-----._

ISYlO

TTTCATTCTTTCTCGCT8tTGTCTCTCTCTAACACACAGGTTTCATGGTCGGTTTTGCCAA

ISY3

TTGCAAAATATGAGTTCAAGGATCTCAGAACCTATAAGAGCCAGAGCAACCAAATTTGGT

ISY1O

TTGCAAAATATGAGTTCAAGGATCTCAGAACCTQTAAGAGCCAGAGCAACCAAATTTGGT

ISY3

ATCCACACTCCTGTGATATCGGACCTTGACCA:TTTCGTGTCCAAATTTCTCCACAccccc

ISY9

2D

FIG. 2. Schematic view of the X2 and Y chromosomal Lcp regions. The drawing is based on DNA sequence information for both areas. Unsequenced DNA stretches are indicated by arrows below the map. Large open arrowheads indicate the direction of transcription. Alignment of the DNA sequences from the Y and X2 regions reveals a dense cluster ofinsertions (ISYs) in the Y chromosomal Lcp region (stippled boxes). Most of the insertions are bordered by small deletions (DY), indicated by triangles. For orientation some restriction sites are indicated. The insertions are shown without restriction sites. Restriction sites: Sa, Sac I; H, HindI1; R, EcoRI; B, BamHI; Xh, Xho I; C, Cla I.

180

240

-----------------------------

ISYlO

ATCCACACTCCTGTGATATCGGACCTTGACCG:TTTCGTGTCCAAATTTC'''CCACACCCCC

ISY3

TTCCGCCCCCGCAAAGGACGAAAATCTGGGGCATCT!ACAAATCTCAGAGACTATTAAGGc

ISYM

TTCCGCCCCCGCAAAGGACGAAAATCTGGGGCATCQCACAAATCTCAGAGACTATTAAGGC

ISY3

TAGAGTAACCAAATTTGGTATCCGCACTTCTGT:T TCGTCTCACTATAATACCC

ISY9 ISY10

-ACTTCTGTOA-ATOTCACT.CTACAAAAATATTT TCAOATCGCCCC ACCACGGCCACAAAGGACGAAAATCTGTTCGCATCCAAACTAAATATTACAA

TA

T

300

360

ISYlO TAGAGTAACCAAATTTGGTAT-CSCACTTCTGTTC9GCCTO-CTCAAA-AC TATAA-

ISY3 ISY9 Isy l

C AG-AA'.T..T TOG C C C C-A-C C C CA T T C C G C C C CAA C AA AGG &C GAAAA T C T G TT-G C A T C C A C AA CAAiAAtCTTTGCCCCGCCCCCTTcCGCCCC£ACAAAGGGCGAAAATCTGTGGCATCCACAA

420

TTTTGCACAT TCGAOAAAAACTA'AAAACGCAGAATCATAGATAACGACC-ATATCTATCAGA

480

a0

ISY3

ISY9 ISY3

"ATACGAAAAAACTOAAAACGCAGAATC.iTAGAG

ATGACTATATOT-

TTGCTGAATCTGGACCAGATCAGATAATTTTTATAGCCAAAAGGAACAAATCAATTTGCA

540

ISY3

CTGGCTACGCAGCCCCCGACGTCACGCTCAGACTGATTTTCTGTCTCTCTCGCACGCACT

600

ISY3

CCTTGTCGTGTCGTTTAATATTAGCGGCGTCTGCCGGAGGAAAGCCATACTGAATTAGTA

660

Isylo

ISvY

ISMY1----l-o-----------------ISY3

------------------------

TCGGGTATAACTGTAGAGTTGCGGTGTCCGCAGCAACTCACAACGTTCCCCCTC

ISY9

ISy

------------------------------------------------------

FIG. 3. Alignment of the ISY9 and ISYJO sequences with ISY3. For alignment, gaps are inserted (-). Point mutations in the aligned sequences are shaded. The bp numbers refer to the ISY3 sequence. Sequence identities between the ISY7 insertion, AAAAT and the reverse complement of ISY8, ACTTAfCTTAGATAAA, with the ISY3 sequence are underlined. ,

714

Proc. Nati. Acad. Sci. USA 89 (1992)

Genetics: Steinemnann and Steinemnann

7594

A

i

a

V141,

is .t

,

"' "

it

-".:.-;I.-.-,

i.,1 ,1 ,16

",

104

.

.-m-&

FIG. 4. Chromosome in situ

hybridization with an ISY3-homologous probe. (A) Female poly-

tene, nucleus from D. miranda salf~~.' ~ivary glands. The ISY3 and ISY3identical sequences are localized

B

at more than 100 loci in the D. miranda female karyotype. (B) Y chromosome from D. miranda. The polytenized degenerating Y

chromosome in the male karyotype shows a dense clustering of labeled loci. (Bars indicate 20

of a euchromatic chromosome into a heterochromatic one. We have used the small multigene family of the Lcp genes as a genetic test system to monitor the molecular changes that have occurred over the course of the divergent evolution of the X2 and Y alleles. The most obvious difference between the Lcp gene clusters on the X2 and Y chromosomes is the dense clustering of inserted DNA elements in the Y Lcp region. We uncovered two retrotransposons, TRIM (3.1 kb) and TRAM (about 2.5 kb), and eight additional, smaller, insertions within the Y region corresponding to the 7-kb X2 Lcp locus. The third large insertion, 713 bp, is the ISY3 element. This element has short open reading frames and no

chromosomal Lcp4 allele is no longer transcribed. Thus, we can unequivocally show that the Y Lcp4 gene represents a specific loss-of-function allele. The accumulation of such alleles was originally predicted by Muller's Y chromosome degeneration model (1, 2).

DISCUSSION The unique karyotype of D. miranda offers the opportunity to analyze the evolutionary changes that occur during Y chromosome degeneration. Two main phenomena characterize the degeneration of the Y chromosome: (i) destruction of genetic information on the Y chromosome and (ii) conversion IIIIA *'I X2..A -A i AA I A:I AAA A. AI T. NA ~ Y -A A T A A _1 A T I T '-AI A

Y

Gi A A I' AN

1A (

K2 'I T

A

Al

A

(I A A A A II

AA A

-AA

-I

X2

((III, A-

1

A A

'

1- Al A A AllA lAAIlC A'C i( V 1A~~liiTI~l-AIII.ACGIAAAI 1l(AAI(J

.A'I

(A

C

A T C

A

Al

1hFIG. 5. DNA sequence align-

I

ment of the X2 and Y

Tll

T(iIATA

A

AT

Al

I

11I 22IIIII1111Il Cli C~~~~~ll CIII [T( -T IT (111 A AA 22 IC11-A A 'I (1 Al IT'i A A 1CIIC Ai 7 -I

A A il T~ll; 1(1..A (ICC *A 11(1 Al

CI A Gl

X2 C (.C

Ii

U

22CIY

([CTA (I A111

X2 [l-l

I

-N 22 1 -Al-ll G

2

~~~~ T

A

-

IICIIACI

I

-AC *AAi

I ACA IT

-Al

I

A

A(-A

i-I

C i

A

l

r

IA

TA T

-

.:i-* 1

lA

T

k

X2

AAl IA 'Ii T TIA -NAl -i(3 A A GTTT6G ACTCIGGAA'T' 'I 'T T A T T TIC Q -A j ( A iA A*. -A Ai 22IIIAl 1AICII--

YA

A

A

A

-

-A11111I A C CI T FlU A.-1 UCUAl C 111 UA ICI *I- T-111T- i C C

IY-A I' I P 'I A III-'I II

The sense strands are aligned in the 5'-to-3' orientation. The alignment begins 230 bp upstream of

A T'.AAI.1.N i.A

Al

IAA~11( 1.TY

AArAA1

D1-A I '-A -A-A

I A AAA TC-A-AA

~IAAI

chromoso-

AlAI(Imal alleles of Lcp4 of D. miranda. AlA Il~l~l

A

A -J IAAIIIIT

A A GlAiA-A IllA-AIA

22 A.T C 'I\.Al V

A~lGA A

iA

A AA A

IAA T1 TC AA., A' A T I Al AA.(I1 (I TATAAACIIA I X2111(1 IA V 1 11-A Al; (:11 A Ill All ~~~~~~~~~ V Ci TA

AA-T-IAl A

.AITl..AA~AIAAIA A A T A A1 AA* A1

C A C -N(A ACI': (i 'II T

X2

i

l-

-

TlT

the transcription start (+ 1). Consensus sequences suchasthestar of transcription, the CAAT box,

the TATA box, and poly(A) addi-

tion motifs are boxed and shaded. The intron is bordered by vertical lines. Translation initiation and stop signals are underlined. Point mutations are shaded. For optinmal alignment, gaps are inserted. We reveal aDNA sequence duplica-

tion, most probably representing a

duplication (boxed and ~~~~~~~~~~~~~targetsiteand of shaded),

adeletion, DY3,

221lbpwithin the coding region.

Genetics: Steinernann and Steinernann female

male

kb -7.4

-5.3 -28

-1 .9 -1 .6 -1

.0

-0-6

_-0.4 -0.3

FIG. 6. Northern analysis of total RNA isolated from male and female

third-instar larval integuments. About 10 g±g of total RNA was applied to each slot on a denaturing 1.2% agarose/formaldehyde gel. The Northern filter was hybridized with the 32P-labeled HindIII-BamHI DNA fragment (cf. Fig. 2) that encompasses Lcp3 and Lcp4 from the X2 region. In both sexes, only the mRNA species derived from the X2 allele is observed.

LTRs, and, unlike the other two large insertions, it is not associated with a target site duplication. The ISY3 and ISY2 insertions have to be further analyzed for features characterizing non-LTR retrotransposons. The small ISYJ insertion reveals, besides the indicated motifs of unclear significance, no homology to known sequences. ISY7, ISY8, ISY9, and ISY10 share homology with the ISY3 sequence (Fig. 3). We assume that these short sequences are either products of extensive rearrangements or remnants of previously excised elements. The latter possibility is more likely, because ISY8, ISY9, and ISY10 contain overlapping sequence motifs. Of interest is the accumulation of point mutations in this common DNA region (Fig. 3). Our results for the Lcp locus indicate that the major forces driving the evolution of a degenerating Y chromosome are retrotransposons or other transposable elements that produce insertions and excisions of different kinds. If this is correct, these elements must be responsible for both of the main phenomena (see above) associated with Y chromosome degeneration. On the basis of the structural features of the 221-bp deletion in Lcp4-e.g., the putative associated target site duplication-we conclude that the null allele is generated by insertion/excision mutagenesis. Alignment of the 5' upstream sequences of the X2 and Y alleles of Lcp4 reveals complete identity from the transcription start (+ 1) up to bp -109 (see Fig. 5). From the structural features one would assume that the truncated allele should still be transcribed. On the other hand, cis regulatory effects exerted by the TRAM retroelement, which flanks the 5' region and/or the ISY inserts together with the duplication in the 3' downstream sequence, might be responsible for the inhibition of transcription (see below). Several mutations associated with transposable elements have been identified in Drosophila (for review see ref. 29). The loss-of-function allele detected at the Lcp4 locus represents an example of a Y-specific null allele and the breakdown of genetic information on the degenerating D. miranda Y chromosome. The biased distribution of the TRIM element and the dramatic increase in the number of the ISY3 identical sequences inserted on the Y chromosome constitute good evidence for the involvement of these elements in the evolutionary process of the destruction of genetic information during Y chromosome degeneration. We believe that at the chromosomal level the observed rearrangements are reflected in an increase in the number of heteropycnotic chromosome sections. Thus, we assume that inserted DNA sequences must be involved in the change in chromosome structure. It is thought that D. miranda diverged from D. pseudoobscura/D. persimilis some 5 million years ago (cf. ref. 30). Assuming the neo-Y/X2 chromosome system arose at or near the time of splitting, this means there have been a lot of generations for the neo-Y to have gone completely

Proc. Natl. Acad. Sci. USA 89 (1992)

7595

heterochromatic. This would imply that the heterochromatization of Y chromosomes is a fairly slow process. Assuming that the accumulation of transposable elements on the Y chromosome is not disadvantageous and thus selection cannot eliminate it, a simple model can account for at least some aspects of the ongoing evolutionary process leading to a completely inert and heterochromatic Y chromosome. The Y-specific loss-of-function allele at the Lcp4 locus retains essential elements of the 5' upstream regulatory region, while a segment of the coding region is truncated. It could help the genome, if similar disabled alleles having intact regulatory motifs, followed by truncated or rearranged coding regions, were hindered in binding transcription factors (TFs). Thus, already-disrupted coding regions of Y alleles would be prevented from being transcribed. This repression of the binding of TFs could be achieved by a switch in chromosome structure from a euchromatic to a heterochromatic state. The switch could be induced by the binding of heterochromatinassociated proteins (31) to the inserted DNA sequences. Thus, in cases where redundant binding motifs cannot be saturated with TFs under the normal "two-gene dose" conditions, the additional TFs could raise the level of transcriptional activity in the corresponding X2 alleles. We thank R. Schuh and H. Jdckle for helpful discussion and P. Hardy for critical reading of the manuscript. We are especially indebted to Herbert Jackle for his generous support and kind hospitality. Our work was supported by grants from the Austrian Fonds zur Forderung der Wisseuschaftlichen Forschung (P5413/ P5987), the Deutsche Forschungsgemeinschaft (St266/2-1), and Sonderforschungsbereich 190. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

24. 25. 26. 27.

28. 29. 30. 31.

Muller, H. J. (1918) Genetics 3, 422-499. Muller, H. J. (1932) Am. Nat. 66, 118-138. Fisher, R. A. (1931) Biol. Rev. 6, 345-358. Hamilton, W. D. (1967) Science 156, 477-488. Nei, M. (1970) Am. Nat. 104, 311-322. Lucchesi, J. C. (1978) Science 202, 711-716. Charlesworth, B. (1978) Proc. Natl. Acad. Sci. USA 75, 5618-5622. Rice, W. R. (1987) Genetics 116, 161-167. Mittwoch, W. (1967) Sex Chromosomes (Academic, New York). Ohno, S. (1967) Sex Chromosomes and Sex-Linked Genes (Springer, Berlin). White, M. J. D. (1973) Animal Cytology and Evolution (Cambridge Univ. Press, Cambridge, U.K.), 3rd Ed. Bull, J. J. (1983) Evolution of Sex Determination Mechanisms (Benjamin/Cummings, Menlo Park, CA), pp. 248-269. Dobzhansky, T. (1935) Genetics 20, 377-391. MacKnight, R. H. (1939) Genetics 24, 180-201. Steinemann, M. (1982) Chromosoma 86, 59-76. Steinemann, M. & Steinemann, S. (1990) Chromosoma 99, 424-431. Steinemann, M. & Steinemann, S. (1991) Chromosoma 101, 169179. Lindsley, D. L. & Zimm, G. (1990) Drosoph. Inf. Serv. 68, 382. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY), 2nd Ed. Fristrom, J. W., Hill, R. J. & Watt, F. (1978) Biochemistry 19, 3917-3924. Snyder, M., Hirsh, J. & Davidson, N. (1981) Cell 25, 165-177. Snyder, M., Hunkapiller, M. M., Yuen, D., Silvert, D., Fristrom, J. & Davidson, N. (1982) Cell 29, 1027-1040. Throckmorton, L. H. (1975) in Handbook of Genetics, ed. King, R. C. (Plenum, New York), Vol. 3, pp. 421-469. Fawcett, D. H., Lister, C. K., Kellett, E. & Finnegan, D. J. (1986) Cell 47, 1007-1015. Schwartz, D. E., Tizard, R. & Gilbert, W. (1983) Cell 32, 853-869. Rosales, R., Vigneron, M., Macchi, M., Davidson, I., Xiao, J. H. & Chambon, P. (1987) EMBO J. 6, 3015-3025. Zelentsova, E. S., Vashakidze, R. P., Kraev, A. S. & Evgen'ev, M. B. (1985) Mol. Biol. 20, 577-583. Cohen, J. B., Hoffman-Liebermann, B. & Kedes, L. (1985) Mol. Cell. Biol. 5, 2804-2813. Green, M. M. (1980) Annu. Rev. Genet. 14, 109-120. Aquadro, C. F., Weaver, A. L., Schaeffer, S. W. & Anderson, W. W. (1991) Proc. Natl. Acad. Sci. USA 88, 305-309. James, T. C. & Elgin, S. C. R. (1986) Mol. Cell. Biol. 6,3862-3872.