Annotated Bibliography

e

© 2001 Nature Publishing Group http://genetics.nature.com

letter

An abundance of X-linked genes expressed in spermatogonia

P. Jeremy Wang1, John R. McCarrey2, Fang Yang1 & David C. Page1

© 2

0 0 1 N

a tu

re P

u b

li s h

in g

G ro

u p

h

tt p

:/ /g

e n

e ti

c s .n

a tu

re .c

o m

Spermatogonia are the self-renewing, mitotic germ cells of the testis from which sperm arise by means of the differentiation pathway known as spermatogenesis1. By contrast with hematopoietic and other mammalian stem-cell populations, which have been subjects of intense molecular genetic investi- gation, spermatogonia have remained largely unexplored at the molecular level. Here we describe a systematic search for genes expressed in mouse spermatogonia, but not in somatic tissues. We identified 25 genes (19 of which are novel) that are expressed in only male germ cells. Of the 25 genes, 3 are Y- linked and 10 are X-linked. If these genes had been distributed randomly in the genome, one would have expected zero to two of the genes to be X-linked. Our findings indicate that the X chromosome has a predominant role in pre-meiotic stages of mammalian spermatogenesis. We hypothesize that the X chro- mosome acquired this prominent role in male germ-cell develop- ment as it evolved from an ordinary, unspecialized autosome. We identified genes specific to germ cells through ‘cDNA sub- traction’2,3, whereby a pool of transcripts present in one cell type (‘tracer’) is depleted of transcripts shared with other cell types (‘driver’). In our subtraction, tracer cDNA was generated from purified mouse spermatogonia4, whereas driver cDNA was gen- erated from a combination of 11 somatic tissues (heart, brain, lung, liver, skeletal muscle, kidney, spleen, stomach, thymus, skin and germ-cell-depleted KitW-v/W-v testis5).

To validate our cDNA subtraction experiments, we tested whether we had recovered previously identified genes that were known to be expressed in spermatogonia but not in somatic tis- sues. Eight such genes (Mage, Ube1y, Usp9y, Rbmy, Stra8, Ott, Ddx4 and Dazl; Table 1) had been identified during the past decade through the efforts of several laboratories. The extent to which we recovered the eight known genes would provide a mea- sure of our protocol’s adequacy in capturing spermatogonially expressed, germ-cell–specific genes. We determined the nucleotide sequence of 2,235 fragments chosen at random from the cDNA subtraction product. We expected that this collection of sequence fragments would constitute a redundant sampling of a much smaller set of genes. Nucleotide sequence analysis revealed that 409 fragments corresponded to 13 known germ-cell–specific genes, including all 8 genes shown to be expressed in spermatogo- nia in previous studies (Table 1). We recovered five other known germ-cell–specific genes (Table 1) that were not previously reported to be expressed in spermatogonia. We tested and con- firmed their expression in purified spermatogonia by RT–PCR (data not shown; primitive type A and mature type A and B sper- matogonia prepared from prepubertal testes). We recovered no known genes specific to meiotic or post-meiotic germ cells. These results indicated that our spermatogonial cDNA subtraction would provide an efficient and sensitive route to the identification of germ-cell–specific genes expressed before meiosis.

Through further analysis of the remaining subtraction product sequences, we identified 23 novel germ-cell–specific genes. We first identified sequence fragments that were present at least twice among the 2,235 subtraction product sequences and that did not correspond to known genes. By testing these sequence fragments for expression in diverse mouse tissues, we identified novel frag- ments that seemed to be expressed in germ cells, but not in somatic cells of the testis or other organs (Fig. 1). Nucleotide sequencing of cDNA clones, and rescreening of libraries as neces- sary, resulted in full-length cDNA sequences for 23 novel germ- cell–specific genes (Table 2).

Virtual translation of these novel cDNA sequences and com- parison with the previously reported genes indicate that many spermatogonially expressed, germ-cell–specific proteins are involved in transcriptional or post-transcriptional regulation of gene expression. Similarities to well-characterized proteins sug- gest that these proteins include a component of RNA polymerase II transcription initiation complexes (the product of Taf2q; Table 2), a nuclear RNA export factor (Nxf2), a ribonuclease inhibitor (Rnh2), a ring-finger protein (Rnf17), an RNA helicase (Mov10l1), and four proteins with RNA-binding domains (RRM domains in the Dazl (refs. 6,7) and Rbmy products8; tudor domains in the Stk31 and Tdrd1 products). These findings, and particularly the large number of putative RNA regulators, are reminiscent of the large role played by post-transcriptional gene regulation in pre-meiotic germ-cell development in Drosophila melanogaster and Caenorhabditis elegans9,10. Our studies suggest that the same is true of pre-meiotic germ-cell development in mammals.

We examined the sex specificity of all 36 spermatogonially expressed, germ-cell–specific genes—in particular, whether they are expressed in the ovary, the site of female germ cells. Eleven genes (four novel genes and seven previously reported genes) are

Table 1 • Known mouse genes expressed in spermatogonia but not in somatic tissues

Gene symbol* Expression Chromosome

Mage testis X Ube1y testis Y Usp9y testis Y Rbmy testis Y Tuba3/Tuba7 testis 6 Stra8 testis 6 Ott testis and ovary X Sycp2 testis and ovary 2 Sycp1 testis and ovary 3 Figla testis and ovary 6 Sycp3 testis and ovary 10 Ddx4 testis and ovary 13 Dazl testis and ovary 17

*For references, see Web Table A.

1Howard Hughes Medical Institute, Whitehead Institute, and Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA. 2Department of Genetics, Southwest Foundation for Biomedical Research, San Antonio, Texas, USA. Correspondence should be addressed to D.C.P. (e-mail: dcpage@wi.mit.edu).

422 nature genetics • volume 27 • april 2001

Gapd

Fshr

Dazi

Rbmy

Fth / 17

Usp26

Tkt/1

Tex11

Tex16

Taf2q

Prame/3

Nxf2

Tex13

Prame/1

Tex17

Stk31

Rnh2

Tex12

Tex18

Tex14

Rnf17

Piwi/2

Mov10/1

Tex20

Tex15

Tex19

Tdrd1

.-,

-

ti

"' Q)

·c ,,, u ,,, 0 -~ ::, Cl E 0 ! .c 1o :,. <ii >, C u ,,, § ~ Q) "'

::, ,,, ~ Q) Q) E Cl 1:'. E C Q) -~ "' -.; C

C Q) ai "' ·.; a. > 3:: -"' :.i2 ,,

0.. 0 ~

C Q) >,

..0 ,,, ! 0 ,,, ,,, :.i2 ,,, .;; .2 .c £

-·-----------· .. ---

-· - -

--- ...

© 2001 Nature Publishing Group http://genetics.nature.com

letter ©

2 0 0 1 N

a tu

re P

u b

li s h

in g

G ro

u p

h

tt p

:/ /g

e n

e ti

c s .n

a tu

re .c

o m

expressed in ovary and in male germ cells (Tables 1 and 2, and Fig. 1). By contrast, the other 25 genes (19 of which are novel) seem to be male-specific (Table 1 and Fig. 1).

We then discovered a strong and unexpected correlation between the sex specificities of the genes and their genomic loca- tions. We ascertained the chromosomal locations of all 36 sper- matogonially expressed, germ-cell–specific genes, 8 of which had been mapped previously (Fig. 2). As expected, the 11 genes that are expressed in both testes and ovaries seemed to be scattered randomly throughout the genome, with 1 such gene mapping to the X chromosome and the other 10 genes distributed among 9 autosomes. By contrast, 13 of 25 male-specific genes mapped to sex chromosomes, with 3 genes (all as previously reported) local- izing to the Y chromosome and 10 genes (9 of which are novel) mapping to the X chromosome. If the 22 non-Y-chromosomal, male-specific genes had been distributed randomly throughout the genome, one would have expected 0–2 such genes to be X- linked. Our finding of 10 X-linked genes is highly unlikely to have occurred by chance (P<10–8), and it indicates a roughly 15-fold enrichment on the X chromosome for male germ-cell–specific, spermatogonially expressed genes. Our mapping and expression studies indicate that, in mammals, the X chromosome has a role in the mitotic stages of spermatogenesis.

Why should the mammalian X chromosome have such a spe- cialization in spermatogonial function? The mammalian X and Y chromosomes evolved from an ordinary pair of autosomes, a gradual process that began approximately 240–320 million years ago with the emergence of SRY, the sex-determining factor, on one member of that ancestral autosomal pair11,12. Apart from

nature genetics • volume 27 • april 2001

Fig. 1 Expression of 23 novel germ-cell-specific genes in mouse tissues assayed by RT–PCR. RNAs were prepared from spermatogonia of 8-d CD-1 males, testes of 8-d C57BL/6J males, ovaries of adult C57BL/6J females, germ-cell–depleted testes of adult KitW-v/W-v C57BL/6J males5, and other tissues of 8-d C57BL/6J males. The Gapd, Fshr, Dazl and Rbmy served as controls. Gapd is expressed ubiquitously. Fshr is expressed in somatic cells of testis and ovary. The Fshr sig- nal in the spermatogonial lane likely reflects the fact that there is 15% contam- ination (with testicular somatic cells) in the spermatogonial preparation. Rbmy and Dazl are expressed only in germ cells, with Rbmy expressed only in testis8

and Dazl expressed in both testis and ovary6,7. Germ cells are reduced in num- ber, but are not entirely absent, in KitW-v/W-v testes. For genes expressed in sper- matogonia, one expects to see a reduced RT–PCR signal (or none) in KitW-v/W-v

testes as compared with wild-type testes. This was observed for all genes except the somatically expressed controls Gapd and Fshr. For Tex20 and Tex15, faint RT–PCR signals are visible in some somatic tissues. Additional RT–PCR data not shown: all 36 germ-cell-specific genes under study (including the 13 genes listed in Table 1) were found to be expressed in primitive type A spermatogo- nia prepared from 6-d CD-1 mice.

SRY, the ancestral autosome was unlikely to have had an outsized role in testicular development or function. At issue then are the adaptive forces that caused the X chromosome, as it differentiated from the Y chromosome, to accumulate so many genes expressed in early stages of spermatogenesis. No explanation is provided by traditional, prevailing models of mammalian X-chromosome evolution, as these have focused on issues of gene dosage (the emergence of X inactivation) without envisioning or predicting any functional specialization13,14. We will outline two possible explanations, both previously debated in evolutionary biology: sex-chromosome meiotic drive15–17 and sexual antagonism18–20.

Meiotic drive, which has been documented in diverse species, including mice21, refers to mechanisms that result in preferential transmission of certain chromosomes at the expense of their homologs. X and Y chromosomes are considered much more likely to become subject to meiotic drive during evolution than are autosomes15,16. Sex-chromosome meiotic drive skews the transmission of X versus Y chromosomes to gametes, and thus the critical drive genes should be expressed during spermatogen- esis. Perhaps some of the X-linked genes that we report are dri- vers of X transmission or suppressors of Y transmission.

The theory of sexually antagonistic genes, which has been pos- tulated to explain the wealth of spermatogenesis factors on mam- malian Y chromosomes18,22, might also account for our findings on the X chromosome. Studies in Drosophila and other systems have demonstrated the existence of sexually antagonistic genes, which enhance reproductive fitness in one sex but diminish fit- ness in the other sex23. Empirical and theoretical studies indicate that, during evolution, sexually antagonistic genes should accu- mulate preferentially on sex chromosomes18,19. Here, conditions favor the emergence of genes that benefit the heterogametic sex (for example, XY), even when those genes are detrimental to the homogametic sex (XX). Sexual antagonism provides a powerful explanation for the enrichment of dominant genes that benefit males on Y chromosomes, including male-ornamentation genes in guppies and spermatogenesis genes in mammals18,22,24.

Focusing on recessive mutations that enhance reproductive fit- ness in males but diminish it in females, it has been argued that natural selection should favor the emergence of sexually antago- nistic alleles on X chromosomes19. The evolutionary dynamics of such male-benefit mutations were considered when they first appear as rare alleles on X chromosomes as opposed to auto- somes. When they are rare, autosomal recessive alleles would be of no advantage to (heterozygous) males and thus would be unlikely to spread widely in the population. By contrast, X-linked recessive alleles would immediately benefit hemizygous males, greatly increasing the alleles’ likelihood of permeating the popu- lation. Eventually, as an allele’s frequency increased in the popu- lation, female fitness would be diminished by the detrimental

423

© 2001 Nature Publishing Group http://genetics.nature.com

letter

Table 2 • New spermatogonially expressed, germ-cell–specific genes in mouse, and their human orthologs

© 2

0 0 1 N

a tu

re P

u b

li s h

in g

G ro

u p

h

tt p

:/ /g

e n

e ti

c s .n

a tu

re .c

o m

Mouse genes Human orthologs Gene symbol Gene name Expression Chr Comments* Gene symbol Chr

Fthl17 ferritin heavy testis X ferritin, functioning in iron metabolism, FTHL17 X polypeptide-like 17 consists of 24 heavy and light chains

Usp26 ubiquitin specific protease 26 testis X predicted protein contains His and Cys domains USP26 X conserved among deubiquitinating enzymes

Tktl1 transketolase-like 1 testis X homologous to human transketolase TKTL1 Tex11 testis expressed gene 11 testis X novel 947-residue protein TEX11 X Tex16 testis expressed gene 16 testis X novel 1,139-residue protein; rich in serine Taf2q TBP-associated factor, testis X human autosomal homolog TAF2F encodes TAF2Q X

RNA polymerase II, Q a component of TFIID

Pramel3 PRAME (human)-like 3 testis X homologous to human PRAME, encoding a melanoma antigen recognized by cytotoxic T cells

Nxf2 nuclear RNA export factor 2 testis X homologous to Mex67p, NXF1 and NXF2, NXF2 X encoding nuclear RNA export factors

Tex13 testis expressed gene 13 testis X novel 186-residue protein; 2 closely related TEX13A X homologs on human X chromosome TEX13B X

Pramel1 PRAME (human)-like 1 testis 4 homologous to human PRAME Tex17 testis expressed gene 17 testis 4 novel 120-residue protein; calculated pI 9.9 Stk31 serine/threonine kinase 31 testis 6 putative protein kinase with tudor domain (found STK31 7

in RNA-interacting proteins) and coiled coil region Rnh2 ribonuclease inhibitor 2 testis 7 predicted protein contains 6 leucine-rich repeats Tex12 testis expressed gene 12 testis 9 novel 123-residue protein with coiled coil region TEX12 11 Tex18 testis expressed gene 18 testis 10 novel 80-residue protein Tex14 testis expressed gene 14 testis 11 predicted protein contains 2 protein kinase domains TEX14 17 Rnf17 ring finger protein 17 testis 14 a RING finger-containing protein RNF17 13 Piwil2 piwi (Drosophila)-like 2 testis 14 homologous to Drosophila piwi, involved in

germline stem cell renewal and meiotic drive Mov10l1 Mov10 (mouse)-like 1 testis 15 putative RNA helicase MOV10L1 22 Tex20 testis expressed gene 20 testis and ovary 2 novel 188-residue protein; calculated pI 10.2 Tex15 testis expressed gene 15 testis and ovary 8 novel 2785-residue protein TEX15 8 Tex19 testis expressed gene 19 testis and ovary 11 novel 351-residue protein with coiled coil region Tdrd1 tudor domain protein 1 testis and ovary 19 predicted protein contains 4 tudor domains TDRD1 10

*For references, see Web Table B.

effects of homozygosity. This would generate adaptive pressure to limit the gene’s expression to males, through additional muta- tions. Based on this theoretical scenario, it was postulated that X chromosomes should evolve to carry a disproportionate share of male-specific genes functioning in male differentiation19. Our findings are in accord with this prediction.

Our hypothesis that mammalian X chromosomes have pre- eminent roles in early stages of spermatogenesis can now be tested through targeted disruption of the many X-linked mouse genes reported here, and through genetic studies in humans. To facilitate studies in humans, we identi- fied orthologous, full-length human cDNA sequences for 13 of 23 novel Chr. 1 2 3 4 5

CEN

Tex20

Sycp1 Mov10l1

Rnh2

Tex15

Tex12

Tex14

Tex19

Rnf17

Piwil2

Ube1y USP9y Rbmy

I i

Sycp2 Tex17

Pramel1

6 7 8 9 10 11 12 13 14 15 16 17 18 19 X Y

Dazl

Stra8

Stk31 Tex18

Sycp3 Ddx4

Figla

Tuba7

Fthl17

Usp26

Tktl1

Tex11

Tex16

Ott

Taf2q

Nxf2 Tex13

Pramel3

Mage

Tdrd1

mouse genes reported here (Table 2). In all 13 cases, the orthologous human genes are expressed exclusively in testes (or testes and ovaries), presumably in germ cells (Fig. 3), and map to chromo-

Fig. 2 Chromosomal locations of 36 spermato- gonially expressed, germ-cell–specific genes in mouse. Genes that seem to be expressed only in testis are shown in blue; genes expressed in both testis and ovary are shown in red; novel genes are boxed. Eight of the genes (Sycp3, Sycp1, Dazl, Rbmy, Ube1y, Usp9y, Mage, Ott) were mapped previously (Table 1); all other genes were mapped by radiation hybrid analy- sis. In the case of gene families residing on a sin- gle chromosome, only one family member is shown (for example, Magea5 is a representative of the X-linked Mage family). The Y chromo- some is shown in proportion to its estimated physical length30; all other chromosomes are drawn on a centiray scale29.

somal regions of known conserved synteny between the mouse and human genomes (Table 2). In particular, we have identified testis-specific, X-linked human orthologs of six of the novel testis-specific, X-linked mouse genes reported here. In the cDNA subtraction experiments reported here, we recovered the mouse homologs of USP9Y, RBMY and DAZ, the three human Y chro- mosomal genes that have been most strongly implicated in male infertility25–27 (Table 1). Perhaps some of the novel X-linked genes will also prove to be sites of mutation in human spermato- genic failure. The stage is set for systematic examination, in both

424 nature genetics • volume 27 • april 2001

-

FTH1 FTHL17 USP26 TEX11 TAF2O NXF2

TEX13A TEX138 STK31 TEX12 TEX14 RNF17 MOV10L1 TEX15 TDRD1

ti

Q) Q)

Cl) N >, C C (,) Q) ::, Cl) ~ ~ C 0 Q) E Cl) ~ "' Q) _Q .>< a. >, e > .~ 0 ::, Cl) ;; 0. J!l 0 (,) ~ --------

© 2001 Nature Publishing Group http://genetics.nature.com

letter ©

2 0 0 1 N

a tu

re P

u b

li s h

in g

G ro

u p

h

tt p

:/ /g

e n

e ti

c s .n

a tu

re .c

o m

Fig. 3 Expression of ortholo- gous human genes assayed by RT–PCR. FTH1 (encoding ferritin heavy chain) is expressed ubiquitously and served as a control. All of the novel human genes seem to be expressed in testis; TEX15 is also expressed in ovary, as is its mouse ortholog (Fig. 1). We also tested eight addi- tional tissues (heart, brain, placenta, lung, liver, skeletal muscle, kidney, pancreas) and detected no expression of the novel genes there (data not shown).

mouse and human, of the postulated role of the X chromosome in early stages of spermatogenesis.

Note: supplementary information is available on the Nature Genet- ics web site (http://genetics.nature.com/supplementary_info/).

Methods Isolation of mouse spermatogonia. We isolated spermatogonia by the Staput method of sedimentation velocity at unit gravity4. Primitive type A spermatogonia were prepared from testes of 6-d CD-1 mice (Charles River Laboratories). Mature type A and type B spermatogonia were isolated from 8-d CD-1 mice. By microscopic examination, at least 85% of the cells in the resulting preparations were spermatogonia, with no more than 15% somatic cell (Sertoli cell) contamination. The spermatogonial preparations contained no spermatocytes, as spermatocytes are not present in the testes of 6-d or 8-d CD-1 mice4.

cDNA subtraction. We carried out three independent subtraction experi- ments, using cDNAs from primitive type A, type A or type B spermatogo- nia as tracer. In all cases, tracer and driver cDNAs were derived from oli- go(dT)-selected RNAs. Germ-cell-depleted testes were from KitW-v/W-v ani- mals. Before subtraction, tracer and driver cDNAs were digested to com- pletion with RsaI. In each of the three experiments, we carried out one round of subtraction using the PCR-select protocol2 (Clontech). To more thoroughly subtract ubiquitous cDNAs, four additional rounds of subtrac- tion were performed using a modified procedure (D. Menke, pers. comm.) as described3. Between rounds of subtraction, we monitored enrichment of Dazl cDNA (germ-cell-specific) and disappearance of Gapd cDNA (ubiq- uitous). Three plasmid libraries (one for each of the three independent experiments) were prepared from the resulting pools of subtracted cDNA fragments. We sequenced (one read only) 800 randomly selected clones from each of the three libraries. Of the 2,400 sequences generated, 165 were of poor quality or derived from the cloning vector, leaving 2,235 sequences for further analysis.

Sequence analysis. Of the 2,235 sequence fragments, 409 corresponded to 13 previously reported germ-cell-specific genes (142 to Mage, 11 to Ube1y, 2 to Usp9y, 44 to Rbmy, 10 to Tuba3/Tuba7, 2 to Stra8, 45 to Ott, 16 to Sycp2, 3 to Sycp1, 3 to Figla, 8 to Sycp3, 21 to Ddx4 and 102 to Dazl). Among the remaining 1,826 sequence fragments, we searched electronical- ly for redundancies and identities to known genes. We found 98 unique, novel sequence fragments that were each recovered at least twice. We tested each of these 98 sequences for germ-cell specificity by RT–PCR on the 14 tissues shown in Fig. 1. Of the 98 sequences, 45 were found to be expressed in spermatogonia and wild-type testis, but not in somatic tissues including KitW-v/W-v testis, indicating that they are germ-cell specific. After full- length cDNA sequences were assembled, these 45 sequence fragments were found to derive from a total of 23 different genes. Of the original set of 2,235 sequence fragments, 546 corresponded to these 23 novel genes (8 to Fthl17; 29 to Usp26; 38 to Tktl1; 66 to Tex11; 2 to Tex16; 132 to Taf2q; 57 to

nature genetics • volume 27 • april 2001

Pramel3; 13 to Nxf2; 5 to Tex13; 4 to Pramel1; 3 to Tex17; 2 to Stk31; 6 to Rnh2; 29 to Tex12; 4 to Tex18; 2 to Tex14; 8 to Rnf17; 16 to Piwil2; 36 to Mov10l1; 7 to Tex20; 71 to Tex15; 6 to Tex19; 2 to Tdrd1).

cDNA cloning. Full-length mouse cDNA sequences were composites derived from subtracted cDNA clones, 5´- and 3´-RACE products, and clones isolated from conventional cDNA libraries that were prepared from adult testes (Clontech, Stratagene and one library of our own construc- tion). We identified orthologous human sequences by searching GenBank using mouse cDNA sequences. We obtained full-length human cDNA sequences by screening a cDNA library prepared from adult testes (Clon- tech).

Radiation hybrid mapping. Using PCR, we tested genomic DNAs from the 93 cell lines of the mouse T31 radiation hybrid panel (Research Genetics) for the presence of each gene28. PCR conditions and primer sequences have been deposited at GenBank. Analysis of the results positioned the genes with respect to the radiation hybrid map of the mouse genome constructed at the Whitehead/MIT Center for Genome Research29 (http://www- genome.wi.mit.edu/mouse_rh/index.html). Chromosomal mapping data of human genes were retrieved from GenBank and confirmed by RH map- ping using the GeneBridge 4 panel (Research Genetics; data not shown).

Expression analysis. Total RNAs were prepared using TRIzol reagent (Gib- co BRL); poly(A)+ RNAs were subsequently isolated using a QuickPrep Micro mRNA purification kit (Amersham Pharmacia Biotech). For each of the 14 tissues shown in Fig. 1, reverse transcription primed with either ran- dom hexamers or oligo (dT)18 was carried out in bulk, using poly(A)

+

RNA (70 ng) from spermatogonia and poly(A)+ RNA (200 ng) from each of the other tissues. RT products were diluted to a final volume of 200 µl, 5 µl of which was used in each PCR amplification. PCR conditions and primer sequences have been deposited at GenBank.

GenBank accession numbers. cDNA sequences for mouse genes: Fthl17, AF285569; Mov10l1, AF285587; Nxf2, AF285575; Piwil2, AF285586; Pramel1, AF285578; Pramel3, AY004873; Rnf17, AF285585; Rnh2, AF285581; Stk31, AF285580; Taf2q, AF285574; Tdrd1, AF285591; Tex11, AF285572; Tex12, AF285582; Tex13, AF285576; Tex14, AF285584; Tex15, AF285589; Tex16, AF285573; Tex17, AF285579; Tex18, AF285583; Tex19, AF285590; Tex20, AF285588; Tktl1, AF285571; and Usp26, AF285570.

cDNA sequences for human genes: FTHL17, AF285592; MOV10L1, AF285604; NXF2, AF285596; RNF17, AF285602 and AF285603; STK31, AF285599; TAF2Q, AF285595; TDRD1, AF285606; TEX11, AF285594; TEX12, AF285600; TEX13A, AF285597; TEX13B, AF285598; TEX14, AF285601; TEX15, AF285605; and USP26, AF285593.

Primer sequences and PCR conditions for mouse RH mapping: Figla, G65193; Magea5, G65194; Ddx4, G65195; Ott, G65196; Sycp2, G65197; Sycp3,G65198; Stra8,G65199; Tuba3, G65200; Tuba7, G65201; Fthl17, G65202; Mov10l1,G65203; Nxf2,G65204; Piwil2, G65205; Pramel1, G65206; Pramel3, G65331; Rnf17, G65207; Rnh2, G65208; Stk31, G65210; Taf2q, G65211; Tdrd1, G65212; Tex11, G65213; Tex12, G65214; Tex13, G65215; Tex14, G65216; Tex15, G65217; Tex16, G65218; Tex17, G65219; Tex18, G65220; Tex19, G65221; Tex20, G65222; Tktl1, G65223; and Usp26, G65224.

Primer sequences and RT–PCR conditions for mouse genes: Gapd, G65758; Fshr, G65759; Dazl, G65760; Rbmy, G65761; Fthl17, G65778; Mov10l1, G65779; Nxf2, G65780; Piwil2, G65781; Pramel1, G65762; Pramel3, G65782; Rnf17, G65763; Rnh2, G65783; Stk31, G65784; Taf2q, G65785; Tdrd1, G65786; Tex11, G65787; Tex12, G65788; Tex13, G65789; Tex14, G65790; Tex15, G65791; Tex16, G65792; Tex17, G65793; Tex18, G65794; Tex19, G65795; Tex20, G65796; Tktl1, G65797; Usp26, G65798.

Primer sequences and RT–PCR conditions for human genes: FTH1, G65764; FTHL17, G65765; MOV10L1, G65766; NXF2, G65767; RNF17, G65799; STK31, G65768; TAF2Q, G65769; TDRD1, G65770; TEX11, G65771; TEX12, G65772; TEX13A, G65773; TEX13B, G65774; TEX14, G65775; TEX15, G65776; USP26, G65777.

Acknowledgments We thank D. Menke for developing the subtraction protocol; H. Skaletsky for statistical advice and bioinformatics support; and A. Arango, D. Berry, A.

425

e

© 2001 Nature Publishing Group http://genetics.nature.com

letter ©

2 0 0 1 N

a tu

re P

u b

li s h

in g

G ro

u p

h

tt p

:/ /g

e n

e ti

c s .n

a tu

re .c

o m

Bortvin, D. Charlesworth, B. Charlesworth, A. Chess, A. Clark, C. Disteche, L. Goldmakher, D. Haig, M. Handel, R. Jaenisch, T. Kawaguchi, F. Lewitter, B. Lahn, A. Lin, D. Menke, T. Rasmussen, W. Rice, S. Rozen and S. Silber for comments on the manuscript. Supported by National Institutes of Health. P.J.W. was the recipient of a Lalor Foundation fellowship.

Received 27 December 2000; accepted 7 March 2001.

1. de Rooij, D.G. & Grootegoed, J.A. Spermatogonial stem cells. Curr. Opin. Cell Biol. 10, 694–701 (1998).

2. Diatchenko, L. et al. Suppression subtractive hybridization: a method for generating differentially regulated or tissue-specific cDNA probes and libraries. Proc. Natl. Acad. Sci. USA 93, 6025–6030 (1996).

3. Lavery, D.J., Lopez-Molina, L., Fleury-Olela, F. & Schibler, U. Selective amplification via biotin- and restriction-mediated enrichment (SABRE), a novel selective amplification procedure for detection of differentially expressed mRNAs. Proc. Natl. Acad. Sci. USA 94, 6831–6836 (1997).

4. Bellve, A.R. Purification, culture, and fractionation of spermatogenic cells. Methods Enzymol. 225, 84–113 (1993).

5. Morrison-Graham, K. & Takahashi, Y. Steel factor and c-kit receptor: from mutants to a growth factor system. Bioessays 15, 77–83 (1993).

6. Cooke, H.J., Lee, M., Kerr, S. & Ruggiu, M. A murine homologue of the human DAZ gene is autosomal and expressed only in male and female gonads. Hum. Mol. Genet. 5, 513–516 (1996).

7. Reijo, R. et al. Mouse autosomal homolog of DAZ, a candidate male sterility gene in humans, is expressed in male germ cells before and after puberty. Genomics 35, 346–352 (1996).

8. Elliott, D.J. et al. An RBM homologue maps to the mouse Y chromosome and is expressed in germ cells. Hum. Mol. Genet. 5, 869–874 (1996).

9. Rongo, C. & Lehmann, R. Regulated synthesis, transport and assembly of the Drosophila germ plasm. Trends Genet. 12, 102–109 (1996).

10. Seydoux, G. & Strome, S. Launching the germline in Caenorhabditis elegans: regulation of gene expression in early germ cells. Development 126, 3275–3283 (1999).

11. Foster, J.W. & Graves, J.A. An SRY-related sequence on the marsupial X chromosome: implications for the evolution of the mammalian testis-determining gene. Proc. Natl. Acad. Sci. USA 91, 1927–1931 (1994).

12. Lahn, B.T. & Page, D.C. Four evolutionary strata on the human X chromosome. Science 286, 964–967 (1999).

13. Graves, J.A.M., Disteche, C.M. & Toder, R. Gene dosage in the evolution and function of mammalian sex chromosomes. Cytogenet. Cell Genet. 80, 94–103 (1998).

14. Jegalian, K. & Page, D.C. A proposed path by which genes common to mammalian X and Y chromosomes evolve to become X inactivated. Nature 394, 776–780 (1998).

15. Frank, S.A. Divergence of meiotic drive-suppression systems as an explanation for sex-biased hybrid sterility and inviability. Evolution 45, 262–267 (1991).

16. Hurst, L.D. & Pomiankowski, A. Causes of sex ratio bias may account for unisexual sterility in hybrids: a new explanation of Haldane’s rule and related phenomena. Genetics 128, 841–858 (1991).

17. Charlesworth, B., Coyne, J.A. & Orr, H.A. Meiotic drive and unisexual hybrid sterility: a comment. Genetics 133, 421–432 (1993).

18. Fisher, R.A. The evolution of dominance. Biol. Rev. 6, 345–368 (1931). 19. Rice, W.R. Sex-chromosomes and the evolution of sexual dimorphism. Evolution

38, 735–742 (1984). 20. Hurst, L.D. & Randerson, J.P. An eXceptional chromosome. Trends Genet. 15,

383–385 (1999). 21. Herrmann, B.G., Koschorz, B., Wertz, K., McLaughlin, K.J. & Kispert, A. A protein

kinase encoded by the t complex responder gene causes non-mendelian inheritance. Nature 402, 141–146 (1999).

22. Saxena, R. et al. The DAZ gene cluster on the human Y chromosome arose from an autosomal gene that was transposed, repeatedly amplified and pruned. Nature Genet. 14, 292–299 (1996).

23. Rice, W.R. Sexually antagonistic genes: experimental evidence. Science 256, 1436–1439 (1992).

24. Brooks, R. Negative genetic correlation between male sexual attractiveness and survival. Nature 406, 67–70 (2000).

25. Sun, C. et al. An azoospermic man with a de novo point mutation in the Y- chromosomal gene USP9Y. Nature Genet. 23, 429–432 (1999).

26. Ma, K. et al. A Y chromosome gene family with RNA-binding protein homology: candidates for the azoospermia factor AZF controlling human spermatogenesis. Cell 75, 1287–1295 (1993).

27. Reijo, R. et al. Diverse spermatogenic defects in humans caused by Y chromosome deletions encompassing a novel RNA-binding protein gene. Nature Genet. 10, 383–393 (1995).

28. McCarthy, L.C. et al. A first-generation whole genome-radiation hybrid map spanning the mouse genome. Genome Res. 7, 1153–1161 (1997).

29. Van Etten, W.J. et al. Radiation hybrid map of the mouse genome. Nature Genet. 22, 384–387 (1999).

30. Bergstrom, D.E. et al. The mouse Y chromosome: enrichment, sizing, and cloning by bivariate flow cytometry. Genomics 48, 304–313 (1998).

426 nature genetics • volume 27 • april 2001