Comprehensive search for chicken W chromosome-linked genesexpressed in early female embryos from the female-minus-malesubtracted cDNA macroarray

Daisuke Yamada, Yoshiyuki Koyama, Mai Komatsubara, Mayuko Urabe, Masako Mori,Yoichi Hashimoto, Ryosuke Nii, Midori Kobayashi, Ayako Nakamoto, Jun Ogihara,Jun Kato & Shigeki Mizuno*

Department of Agricultural and Biological Chemistry, College of Bioresource Sciences, Nihon University,1866 Kameino, Fujisawa, 252-8510 Japan; Tel/Fax: (81) 466-84-3943; E-mail: s-mizuno@brs.nihon-u.ac.jp*Correspondence

Received 10 August 2004 and accepted for publication by Herbert Macgregor 31 August 2004

Key words: cDNA, chicken, embryo, macroarray, W chromosome

Abstract

In order to seek chicken W chromosome-linked genes expressed signi¢cantly earlier than the time ofgonadal differentiation, female-minus-male-subtracted cDNA macroarrays were prepared from day 2(Hamburger-Hamilton stages 12^13), day 3 (stages 19^20) and day 4 (stages 24^25) embryos. From atotal of 15 744 macroarrayed cDNA clones, 610 clones exhibiting signi¢cantly female-speci¢c expressionwere selected. When each one of the 610 cDNA clones was used as a probe in Southern blot hybridizationwith male or female chicken genomic DNA, 62 clones, grouped into eight (A^H) types according to theirpatterns of hybridization, were considered to be derived from W chromosome-linked genes. When rep-resentative cDNA clones in each type were sequenced, clones derived from two known W-linked genes;SPIN-W and ATP5A1W, and from two hitherto unknown W-linked genes, represented by 2d-2D9and 2d-2F9 clones, were identi¢ed and their localizations on the W chromosome were con¢rmed by £u-orescence in-situ hybridization. The 2d-2D9 sequence has no signi¢cant homology with other genes indatabases but 2d-2F9 has a region which shows partial homology to the consensus sequence of theAAA ATPase superfamily. Both 2d-2D9 and 2d-2F9 sequences are found in contigs of undeterminedchromosome-linkage in the Draft Chicken Genome Sequence.

Introduction

It has been suggested that the evolutionary ori-gins of mammalian XY chromosomes and avianZW chromosomes are unrelated because syntenyanalysis has shown that the chicken Z chromo-some shares a substantial number of genes incommon with the human chromosome 9,

whereas no common genes between the chickenZ chromosome and the human X chromosomehave been found (Schmid et al. 2000). Themale-determining gene, SRY/Sry, located on theY chromosome of most mammalian species isabsent in genomes of chickens of both sexes(Mizuno et al. 2002). It is speculated that afemale or male-determining genetic system

Chromosome Research 12: 741–754, 2004. 741# 2004 Kluwer Academic Publishers. Printed in the Netherlands

unique to the avian species has been acquiredearly in the evolution of the present-day Z andW chromosomes.

Although no de¢nite sex-determining gene hasyet been demonstrated on the avian sex chromo-somes, there are two possibilities: (1) there is afemale-determining gene on the W chromosome,and (2) there is a male-determining gene on theZ chromosome and it triggers testis developmentif its gene dosage or expression level is 2 in ZZembryos as compared with 1 in ZW embryos.The latter situation may be attained by theabsence of total inactivation of one of Z-linkedalleles in male (ZZ) chickens (Kuroda et al.2001, Kuroiwa et al. 2002). At present, a candi-date gene on the chicken W chromosome isASW/Wpkci (O’Neill et al. 2000, Hori et al.2000). The nomenclature of this gene has notbeen ¢nalized but PKCI-W or in particularHINT-W may be more appropriate because itscounterpart gene on the Z chromosome seems tobe a chicken ortholog of mammalian PKCI (agene for protein kinase C-interacting protein) orHINT (a gene for histidine triad nucleotide-binding protein) (Brenner 2002). It has beenproposed that PKCI-W which lacks an HIT(histidine triad) motif essential for the PKCIfunction may form a heterodimer with PKCI-Zin female embryos and thereby inhibit a PKCIfunction (O’Neill et al. 2000, Hori et al. 2000).However, this model has not been proved at thelevel of the protein and the role of PKCI-Z/HINT-Z in the process of sex determinationremains unknown.

A candidate gene for male determination onthe Z chromosome is DMRT1. This gene islinked to the Z chromosome not only in Car-inatae birds but also in emus belonging to Rati-tae (Shetty et al. 2002). DMRT1 is expressed inthe genital ridges of both male and femalechicken embryos and its mRNA level appearedto be about 2 in males and 1 in females accord-ing to whole mount in-situ hybridization(Raymond et al. 1999a). It has been suggestedthat the double dosage of DMRT1 is requiredfor the normal development of testis in malehumans because a single dosage of DMRT1caused by the deletion of 9p24.3 on one of thechromosomes 9 causes XY sex reversal with highfrequency (Raymond et al. 1999b). The DM

domain present in DMRT1 is a DNA-bindingdomain related to those of doublesex in Droso-phila melanogaster and mab-3 in Caenorhabditiselegans. Recently, another DM-containing gene,DMY, was identi¢ed on the Y chromosome ofthe medaka ¢sh as a male-determining gene(Matsuda et al. 2002). In birds, an essential roleof DMRT1 in the normal development of testishas not been proved. Moreover, it has not beendetermined whether the half dosage of DMRT1in females could lead to remarkable female-speci¢c events during the early gonadaldi¡erentiation, such as the female-dominantexpression of 17b-hydroxysteroid dehydrogenasegene and the female-speci¢c expression of aroma-tase gene (Nakabayashi et al. 1998).

About 80% of the DNA in the chicken W chro-mosome consists of highly repetitive sequencesbelonging to XhoI-, EcoRI- and SspI-families.Non-repetitive DNA sequences in the chicken Wchromosome probably comprise only about10 Mb (Itoh et al. 2001, Mizuno et al. 2002). Atleast 20 active genes are expected to be present inthe latter non-repetitive DNA regions but onlysix genes have so far been identi¢ed and reported(Mizuno et al. 2002).

In the present study, in order to search forhitherto unknown genes on the chicken W chro-mosome which may be involved in the sex determi-nation or in the early process of sexdi¡erentiation, cDNA macroarrays consistingof about 15 000 clones were prepared fromfemale-minus-male-subtracted double-strandedcDNAs of day 2 to day 4 (stages 12^25; Hambur-ger & Hamilton 1951) chicken embryos. Clonesexhibiting signi¢cant female-speci¢c expressionwere selected from these macroarrays and theirlinkage to the W chromosome was ¢rst examinedby Southern blot hybridization with an individualcandidate cDNA clone as a probe and subse-quently by FISH with their corresponding geno-mic clones as probes.

Three known (PKCI-W, SPIN-W andATP5A1W) and two unknown W-linked geneswhich are expressed in as early as day 2 (stages12^13) female embryos were identi¢ed from thesestudies. The latter two genes belong to genomiccontigs whose chromosome linkage is unde-termined in the recently released Draft ChickenGenome Sequence.

742 D. Yamada et al.

Materials and methods

Preparation of female-minus-male-subtracteddouble-stranded cDNA

A whole embryo and extra-embryonic membranewere separated from an individual chickenembryo at day 2 (stages 12–13), day 3 (stages 19–20) or day 4 (stages 24–25). The whole embryoswere immediately frozen at �80�C. The sex of anindividual embryo was determined by thepresence (female) or absence (male) of PCR-amplified W chromosome-specific XhoI-familyrepetitive sequence (Kodama et al. 1987) usingDNA extracted from the extra-embryonicmembrane as a template and a set of primers asdescribed (Clinton 1994). The numbers of sex-determined embryos employed for the extractionof poly(A)þRNA were: 107 each of females (F)and males (M) from day 2, 20 each from day 3and 9 each from day 4 embryos. Total RNA wasextracted with TRIZOL (Invitrogen) acidic phenolreagent and poly(A)þRNA was separated from500 mg each of the total RNA by an oligo(dT)cellulose column using Micro-First Track mRNA2.0 (Invitrogen). A sample (0.5 mg) of each of thepoly(A)þRNA from female or male embryoswas subjected to PCR-based synthesis of double-stranded cDNA using a SMART cDNA synthesiskit (CLONTECH). Ten mg each of the double-stranded cDNA was digested with RsaI and pur-ified by a column of Wizard SV Gel and PCRclean-up system (Promega). A female-minus-male-subtracted double-stranded cDNA was pre-pared using PCR-Select cDNA Subtraction Kit(CLONTECH) in which 15 ng of the adaptor1(2R)-ligated, female cDNA was hybridized suc-cessively with 450 ng and 100 ng of male cDNAand hybrids between adaptor-ligated femalecDNA molecules were amplified by two-stepPCR. The amplified cDNA molecules were diges-ted with RsaI and ligated with SmaI-digested,alkaline phosphatase-treated pBluescript II KSþand transformed E. coli DH5a. The efficiency ofsubtraction of female, male common cDNAmolecules was estimated by comparing levels ofPCR-amplified GAPDH and PKCI-W cDNAfragments between RsaI-digested female cDNA(non-subtraction control) and female-minus-male-subtracted cDNA.

Preparation of cDNA macroarray and selection ofclones exhibiting significantly female-specificexpression

About 5000 each of single colonies of the female-minus-male-subtracted double-stranded cDNApreparations from day 2, day 3 and day 4embryos were picked up and subjected to thecolony-direct PCR with M13(-20) and M13reverse primers. PCR products were heat-dena-tured in 0.4N NaOH, 0.1M EDTA and fixed onaHybond Nþ membrane (8� 12 cm) (AmershamBiosciences) by passing through a vacuum blotterhaving 384 holes, washed successively with 0.4NNaOH and 2� SSC and baked at 80�C for30min. Probes: a female-minus-male or a male-minus-female subtracted cDNA preparation asdescribed above from which adaptors wereremoved by RsaI digestion or a female or maletotal cDNA preparation from different develop-mental stages or PKCI-W cDNA (Hori et al.2000), were labelled using AlkPhos Direct label-ling system (Amersham Biosciences). Hybridiza-tion was carried out according to the protocol ofthe system and hybrids were detected with CDP-Star detection reagent (Amersham Biosciences).Intensity of chemi-luminescence was measured byLumino Imaging Analyzer FAS-1000 (TOYOBO)and significantly female-specific clones were selec-ted by analysing the image with Array-Pro Ana-lyzer Version 4.0 (Media cybernetics). Probeshybridized to a cDNA macroarray were removedby incubating in 0.5% SDS at 65�C for 1 h, fol-lowed by washing in 0.1mol/L Tris-HCl (pH8.0)for 5min, and the array was subjected tore-hybridization.

Southern blot hybridization to search forW-chromosome-derived cDNA clones andsequencing of their inserts

Female or male chicken genomic DNA preparedfrom blood cells was digested with HindIII orPstI. The digested DNA (20 mg/lane) wassubjected to agarose-gel electrophoresis andSouthern blot hybridization on a Hybond Nþmembrane (Amersham Biosciences) with aninsert of a selected cDNA clone as a probe. Theinsert was amplified by colony-direct PCR, pur-ified by a column of Wizard SV Gel and PCR

Chicken W chromosome-linked genes 743

clean-up system and AlkPhos-labelled as above.Those clones which hybridized to female-specificband(s) or female-specific band(s) plus male,female-common bands were subjected to thesecond-round Southern blot hybridization inwhich genomic DNA preparations were digestedwith HindIII, PstI, KpnI, EcoRI or BamHI.Clones hybridized again as above were groupedaccording to their patterns of hybridization. ThecDNA inserts of clones representing individualgroups were sequenced using Cy5.5 dye terminatorcycle sequencing kit (Amersham Biosciences) andKS HT or T7 HT primer for the sequence inpBluescript II KSþ . Determined sequences weresubjected to homology search against NCBI(National Center for Biotechnology Information)BLAST, DDBJ (DNA Data Bank of Japan)BLAST and the European Bioinformatics Insti-tute (EBI)’s Ensemble Chicken Genome Browser(http://www.ensembl.org/Gallus gallus/).

Isolation of genomic clones for FISH

A BAC library (Hori et al. 2000) prepared frompartially HindIII-digested genomic DNA fromthe female White Leghorn chicken embryo wasscreened by four-dimensional polymerase chainreactions (Kato et al. 2002). A set of PCR pri-mers for the screening was selected from thecDNA sequence of a clone which was suggestedto be derived from a W chromosome-linkedgene. A selected BAC clone was biotinylated bynick translation and used as a probe for fluores-cence in-situ hybridization (FISH) to metaphasechromosome sets prepared from female chickenembryonic fibroblasts as described (Itoh et al.2001). The reaction mixture of FISH also inclu-ded digoxigenin (DIG)-labelled marker probe(s)for identification of sex chromosomes and a sub-region on the W chromosome: EcoRI-family (Wchromosome; Saitoh et al. 1991) plus pFN-1macrosatellite (Z chromosome; Hori et al. 1996)or SspI-family (a subregion on the W-short arm;Itoh & Mizuno 2002). When a BAC clone couldnot be obtained from the above library, a contigof genomic sequence containing the cDNAsequence was searched for from the EBI’sEnsemble Chicken Genome Browser as aboveand its sequence was amplified by PCR from the

female chicken genomic DNA, biotinylated andused as a probe.

Extension of cDNA sequences and sequenceanalysis

The size range of cDNA inserts in selected clonesfrom the cDNA macroarray was generally short(approximately 500–1000 bp) because double-stranded cDNA molecules were digested withRsaI during the course of female-minus-malecDNA subtraction. To get further sequenceinformation, both 50- and 30-extended, double-stranded cDNA was prepared by PCR from thepoly(A)þRNA from day 2 female embryos byapplying the oligo-capping method (Suzuki &Sugano 2001) which required the presence ofboth 50 cap structure and 30 poly (A) tail in theoriginal RNA molecule. A forward primer and areverse primer were selected from the sequence ofa particular cDNA clone. The 50 primer, the 50

nested primer, the 30 primer and the 30 nestedprimer were provided from the GeneRacer kit(Invitrogen). When the extension toward 50-endwas unsuccessful, a primer selected from thesequence of another cDNA clone which was situ-ated in a nearby genomic contig was utilized asdescribed in the text. Sequencing and BLASTsearch were carried out as described above.

Results

Selection of clones exhibiting significantlyfemale-specific expression from thefemale-minus-male subtracted cDNA macroarray

The sexes of individual chicken embryos at day 2to day 4 of incubation were determined by PCRfor the presence (female) or absence (male) ofthe W chromosome-specific XhoI-family repeti-tive sequence in the DNA extracted from theextra-embryonic membrane. Poly(A)þRNA wasprepared from pooled female or male wholeembryos at day 2 (stages 12–13), day 3 (stages19–20) or day 4 (stages 24–25) and converted todouble-stranded cDNA. Female-minus-malesubtracted cDNA was prepared from the double-stranded cDNA as described in Materials andMethods. Approximately 5000 clones were

744 D. Yamada et al.

randomly picked up from the female-minus-malesubtracted cDNA preparation from each devel-opmental stage and cDNA macroarray (384clones per membrane) consisting altogether of15 744 clones was prepared (Table 1).

A set of cDNA macroarray prepared from anindividual stage was hybridized successively withthe female-minus-male subtracted cDNA, themale-minus-female subtracted cDNA preparedfrom embryos of the corresponding develop-mental stage and with the PKCI-W cDNA probe(Hori et al. 2000) (Figure 1). A duplicate set ofthe cDNA macroarray was hybridized succes-sively with the total female and the total malecDNA prepared from embryos of the correspond-ing stage (data not shown in Figure 1). As illu-strated in Figure 1 and summarized in Table 1,altogether 610 clones were judged as clones exhi-biting signi¢cantly female-speci¢c expression. Outof the 610 clones, 532 clones were detected withthe subtracted cDNA probes and 84 clones weredetected with the total cDNA probes. Betweenthe former and the latter clones, only 6 cloneswere overlapped, i.e. 532þ 84� 6¼ 610 clones.

Search for cDNA clones derived fromW chromosome-linked genes

The cDNA inserts of 610 clones showingsignificantly female-specific expression wereindividually labelled with alkaline phosphatase

and each clone was used as a probe in the two-step Southern blot hybridization to female ormale genomic DNA of chickens, digested withHindIII or PstI in the first step and HindIII,PstI, KpnI, EcoRI or BamHI in the second step.Sixty-two clones hybridized to female-specificband(s) or female-specific band(s) plus female,male common bands and suggested to be derivedfrom W-chromosome-linked genes (Table 1). Allof these 62 clones had been selected from thecDNA macroarray with subtracted cDNA probes(Table 1). It was noted that most (61/62) of theputatively W-linked clones were obtained fromday 2 embryos (Table 1).

The above 62 clones were grouped into 8 types(A to H) according to their patterns of Southernblot hybridization (Figure 2). Numbers of clonesand clone Nos. belonging to each type areindicated in Table 2.

Homology search with cDNA sequencesof putatively W-linked clones

The above results that the 62 putatively W-linkedcDNA clones were grouped into 8 different typesof Southern blot hybridization patterns do notnecessarily mean that 8 different genes are repre-sented. This is because the double-strandedcDNA prepared from female embryos was diges-ted with RsaI during preparation of the female-minus-male subtracted cDNA. This procedure

Table 1. Selection of clones exhibiting significantly female-specific expression from the female-minus-male subtracted cDNA macroarray

and further selection of candidate clones derived from W chromosome-linked genes.

Days of incubation

and Hamburger–

Hamilton stage

of embryos

Number of

clones

used for the

construction

of cDNA

macroarray

Number of clones exhibiting

significantly female-specific

expression selected with probes

I or II

Number of

clones subjected to

Southern blot

hybridization

Number of

clones suggested to be

derived from

W chromosome-linked genes

I

Subtracted

cDNA probe

II

Total

cDNA probe

Clones

selected

commonly

with I and II

day 2 stages 12–13 5376 170 52 3 219a 61

day 3 stages 19–20 5376 138 18 1 155 1

day 4 stages 24–25 4992 224 14 2 236 0

Total 15744 532 84 6 610 62b

a170þ 52-3; These clones were individually used as probes in Southern blot hybridization.bAll the clones were selected with probe I.

Chicken W chromosome-linked genes 745

yielded relatively short cDNA inserts of approxi-mately 0.5–1 kb. Thus, different clones maycontain cDNA inserts derived from differentregions of the same mRNA molecule and, whenthey are used as probes, different Southern blotpatterns may be produced. Other possible factorsare: (1) the existence of an intron-derivedsequence in a cDNA clone, probably caused bythe presence of pre-mRNA molecules in thestarting poly(A)þRNA preparation, and (2) for-mation of chimeric clones in which different RsaIfragments of double-stranded cDNA are joined.In either case, production of complex patterns ofSouthern blot hybridization may be expected.

Keeping these possibilities in mind, cDNAsequences of representative clones from each ofthe 8 Southern-blotting types were determined

and subjected to homology search against nucleo-tide sequence databases and the Draft ChickenGenome Sequence (Table 2). The results indicatedthat: (1) sequences of cDNA clones of types Cand D matched well with a cDNA sequence ofSPIN-W (a W chromosome-linked spindlin gene)(Itoh et al. 2001), (2) sequence of the cDNA cloneof type H matched well with an intron sequenceof SPIN-W, (3) sequence of the cDNA clone oftype F matched well with a cDNA sequence ofATP5A1W (a W-chromosome-linked gene for a-subunit of ATP synthase (Fridolfsson et al. 1998,Ellegren & Carmichael 2001), but (4) sequences ofcDNA clones belonging to types A, B, E andG did not show signi¢cant homology withcDNA sequences of known W chromosome-linked genes.

Figure 1. Selection of clones exhibiting significantly female-specific expression (red or orange circles) from the female-minus-male

subtracted cDNA macroarray (384 clones/membrane). Examples shown are macroarrays from day 2 embryos. Membrane Nos. and

three different probes hybridized successively are indicated. Results with male and female total cDNA probes are not shown. A set of

41 membranes were prepared for 15 744 subtracted cDNA clones (Table 1). Several clones identified or characterized in this study are

indicated.

746 D. Yamada et al.

cDNA clones belonging to types A, B and E aremost likely derived from the same gene on the Wchromosome

The results of the following four experimentssuggested strongly that cDNA clones belongingto types A, B and E are probably derived fromthe same gene on the W chromosome:

(1) cDNA inserts of individual clones belongingto types A, B and E were amplified by PCRusing primers from the multi-cloning site ofthe vector and subjected to agarose gel elec-trophoresis and Southern blot hybridizationwith the alkaline phosphatase-labelled cDNAprobe prepared from either 2d-1H5 (type A)

or 2d-2D9 (type B) clone (Table 2). The2d-1H5 probe hybridized with all the type-Aclones except for 2d2F11 and also hybridizedwith 2d-1H11 and 2d-2B5 clones of type Ebut did not hybridize with clones belongingto type B and 2d-1G4 clone of type E (datanot shown). On the other hand, the 2d-2D9probe hybridized with all the type B clonesexcept for 2d-1B11, 2d-2A3 and 2d-2E10 andalso hybridized with 2d-1G4 of type E butdid not hybridize with clones belonging totype A (data not shown).

(2) PCR primers were selected from cDNAsequences of following clones: 2d-1H2 and2d-1H5 (type A), 2d-2D7 and 2d-2D9 (typeB), 2d-1G4 and 2d-1H11 (type E), and these

Figure 2. Sixty-two candidate cDNA clones derived from W-chromosome-linked genes are grouped into 8 types (A–H) according to

their patterns of Southern blot hybridization. Individual cDNA clone was used as a probe in hybridization to female (F) or male (M)

chicken genomic DNA digested with different restriction enzymes as indicated. Results with representative cDNA clones are shown.

There are essentially two types of patterns: (1) hybridization to female-specific bands (A, B) and (2) hybridization to female-specific

bands and female, male common bands (C–H). The latter-type clone is speculated to have a structurally similar counterpart gene on

the Z chromosome.

Chicken W chromosome-linked genes 747

Table2.

Gro

upingofcD

NA

clones

whicharesu

ggestedto

bederived

from

theW

chro

moso

meinto

8types

from

theirSouthernblottingpattern

san

dtheirco

rresponden

ceto

genes

on

the

Wch

romoso

me

and

genomic

contigs

inthe

Draft

Chicken

Gen

ome

Seq

uen

ce.

Type

Num

berof

clones

Clone

no.a

Approximate

size

ofcDNA

sequenced(kb)

Hom

ologyto

known

W-linked

gene

Hom

ology

(%)

Hom

ologyto

genomiccontig

(chrom

osom

elinkage)

A16

2d-1A10

2d-1D12

2d-1E1

2d-1E3

2d-1F11

0.4

none

5115

.3

(unknown)

2d-1G2

2d-1G10

2d-1G11

2d-1H2

2d-1H5

2d-1H8

2d-2A4

2d-2A5

2d-2B2

2d-2C12

2d-2F11

B28

2d-1B8

2d-1B11

2d-1C7

2d-1C10

2d-1D5

0.5

none

5115

.2

(unknown)

2d-1D7

2d-1D10

2d-1E10

2d-1E11

2d-1F9

2d-1G1

2d-1G7

2d-1H9

2d-1H12

2d-2A3

2d-2B6

2d-2C3

2d-2C11

2d-2D1

2d-2D7

2d-2D8

2d-2D9

2d-2E1

2d-2E2

2d-2E4

2d-2E9

2d-2E10

2d-2F12

C9

2d-1B10

2d-1D11

2d-1F1

2d-1F2

2d-1G5

0.7

SPIN-W

9812

46.8

(W)

2D-1G9

2d-1G12

2d-2A2

2d-2C6

D2

2d-1A1

3d-1H2

0.4

SPIN-W

9912

46.8

(W)

E3

2d-1G4

2d-1H11

2d-2B5

0.8

none

5115

.2

(unknown)

5115

.3

(unknown)

F2

2d-1E6

2d-1G6

0.6

ATP5A1W

9429

56.2

(W)

2136

7.1(W

)

G1

2d-2F9

0.8

none

1279

.1

(unknown)

H1

2d-2D4

0.6

SPIN-W

intron

9637

18.2

(W)

Total

62

aUnderlin

edclones

were

subjected

tosequen

cing.

748 D. Yamada et al.

primers were employed for the four-dimen-sional PCR selection of genomic BAC clonesfrom the female chicken BAC library (Katoet al. 2002). Sequences of all the 6 cDNA

clones were found to be present in the twooverlapping BAC clones 170F6 and 396G1.

(3) When these two BAC clones were used asprobes in FISH to metaphase chromosome

Figure 3. Detection of 2d-2D9 and 2d-2F9 sequences on the chicken W chromosome by FISH. BAC clone 396G1 containing the

genomic sequence of 2d-2D9 (A, B) or genomic contig 1279.1 containing the genomic sequence of 2d-2F9 (C, D) was biotinylated

and hybridized to metaphase sets from female chicken embryonic fibroblasts together with DIG-labelled marker probe(s):

EcoRI-family sequence on the long arm of W chromosome plus pFN-1 macrosatellite sequence on the long arm of Z chromosome

(A, C) or SspI-family sequence on the short arm of W chromosome (B, D). An enlarged image of the W chromosome is shown in the

inset of B and D. Bar indicates 10mm.

Chicken W chromosome-linked genes 749

sets from female chicken embryonic fibro-blasts, they both hybridized to the slightlydistal region from the SspI-family locus onthe short arm of the W chromosome asshown in Figure 3A, B for 396G1.

(4) cDNA inserts of the four clones; 2d-2F11,2d-1B11, 2d-2A3 and 2d-2E10, which werenot hybridized with 2d-1H5 or 2d-2D9 probewere individually labelled with alkaline phos-phatase and used as a probe in Southernblot hybridization to HindIII-digested BACclone 170F6 or 396G1. All the four probeshybridized to a single band of the same

fragment size from either BAC clone (datanot shown)

These results suggested strongly that the 47cDNA clones belonging to types A, B and E(Table 2) were derived from the same Wchromosome-linked gene present in the over-lapping BAC clones 170F6 and 396G1.

A type-G cDNA clone 2d-2F9 is also derived from agene on the W chromosome

A BAC clone containing the cDNA sequence of2d-2F9 (type G in Table 2) could not be isolatedfrom the BAC library employed in this study,which was constructed from the partially HindIII-digested genomic DNA of a female WhiteLeghorn chicken (Hori et al. 2000). Instead, thegenomic sequence of contig 1279.1 (Table 2 andFigure 5A), containing the 2d-2F9 sequence assuggested from the BLAST search was amplifiedby PCR, biotinylated and used as a probe inFISH to metaphase sets from female chickenembryonic fibroblasts. This probe hybridized tothe similar region where the 2d-2D9 probe hybri-dized on the W chromosome (Figure 3C, D).These results suggest strongly that the 2d-2F9gene is located on the short arm of the Wchromosome.

Correlation between cDNA clones and genomicclones for 2d-2D9 and 2d-2F9

A 434-bp band was amplified by PCR from the2d-2D9 cDNA sequence using a set of primersselected from within the cDNA sequence. Thesame-sized band was produced from the femalebut not from the male chicken genomic DNA,BAC clones 170F6 and 396G1 and genomic con-tigs 5115.2–5115.3 (Figure 4A). Similarly, a bandof 1119 bp was produced by PCR using a set ofprimers from the 2d-2F9 cDNA sequence fromthe female but not from the male chicken genomicDNA and the genomic contig 1279.1 (Figure 4B).The size of the band amplified from the genomicsequence was much higher than that of the bandamplified (179 bp) from the 2d-2F9 cDNAsequence (Figure 4B), suggesting the presence ofintron(s) in the genomic sequence. These resultsdemonstrate more directly the presence of 2d-2D9

Figure 4. Correlation between the cDNA clone (2d-2D9 or

2d-2F9) and the genomic clone (BAC clone or genomic contig).

(A) PCR using a set of primers selected from 2d-2D9 cDNA

sequence produced a 434-bp female-specific band (the same

size as from the cDNA clone) from the chicken genomic DNA,

BAC clones 170F6, 396G1 and genomic contigs 5115.2–5115.3

(including a gap region between the two contigs; Figure 5). (B)PCR as in A but using primers from the 2d-2F9 cDNA

sequence produced a 1119-bp female-specific band from the

chicken genomic DNA and contig 1279.1. In this case, the

PCR product from the female genomic DNA or genomic

contig was much longer than the product (179 bp) from the

cDNA clone, suggesting the presence of intron(s).

750 D. Yamada et al.

sequence in contigs 5115.2–5115.3 and in BACclones 170F6 and 396G1 and the presence of 2d-2F9 sequence in the contig 1279.1.

Analysis of extended cDNA sequences of 2d-2D9and 2d-2F9

In order to extend cDNA sequences of 2d-2D9and 2d-2F9 clones toward both 50 and 30 ends,the GeneRacer method (Invitrogen), based onthe oligo-capping method (Suzuki & Sugano

2001), was applied. As illustrated in Figure 5A, a50-extended and a 30-extended cDNA clones wereobtained for 2d-2F9. Sequencing of these twoclones suggested that a full-length cDNA of 2d-2F9 was about 1.2 kb. A PCR product of 776 bpwas obtained from the cDNA preparation of day2 female embryos using 50 and 30 nested primers(Figure 5A), which was then cloned andsequenced (accession number in DDBJ/EMBL/GenBank nucleotide sequence databases isAB188527). In the determined 776-bp cDNAsequence, an open reading frame (ORF) of

Figure 5. Structure of 50 and 30-extended cDNA sequences from original 2d-2F9 (A) or 2d-2D9 (B) cDNA clones. Only one strand of

the double-stranded cDNA sequence obtained is shown. Relative positions of a genomic contig (A) or genomic contigs and BAC

clones (B) are indicated. Sequences of the double-stranded cDNAs: the 776-bp sequence amplified by PCR with nested primers (A)and sequences of 2d-1H5 (the same sequence as 2d-1H2) and 2d-2D9 clones (B) are deposited in DDBJ/EMBL/GenBank nucleotide

sequence databases with accession numbers (AB188532 for 2d-1H5 and AB188526 for 2d-2D9), respectively. In C, it is shown that

approximately 2.5-kb female-specific double-stranded cDNA (as shown in B) is produced by PCR with primers from 2d-1H2 and

2d-2D9 from both genomic DNA (chicken genomic DNA and BAC clones 170F6 and 396G1) and the cDNA template (from day 2

female chicken embryos).

Chicken W chromosome-linked genes 751

753 bp (nucleotide positions 167–919, including astop codon) initiating from a Met codon waspresent. This Met codon was probably the initia-tion codon because there were two in-frame stopcodons in the upstream sequence. The BLASTsearch against nucleotide sequence databaseswith the ORF sequence revealed that a 414-bpN-terminal-side region (nucleotide positions 186–599, encoding 138 amino acid residues) hadabout 70% homology to a part of the 211-residues-long consensus sequence of AAAATPase family proteins (Frohlich 2001) butother regions had no significant homology withother cDNA or protein sequences.*

On the other hand, for 2d-2D9, extensiontoward the 50 cap site was unsuccessful and onlya double-stranded cDNA extended to the 30

poly(A) tail was obtained (Figure 5B). To attainextension of the 2d-2D9 cDNA sequence towardits 50 end, we utilized the information that cDNAsequences of clones belonging to type A werederived from the genomic contig 5115.3, whilethose of clones (including 2d-2D9) belongingto type B were derived from the contig 5115.2(Table 2), yet the two cDNA sequences seemedto be present in the same gene on the W chromo-some as explained above. When PCR was per-formed using a forward primer from the 2d-1H2sequence (type A) and a reverse primer from the2d-2D9 sequence (type B) on the cDNA prepara-tion of day 2 female embryos, a double-strandedcDNA sequence of about 2.5 kb was amp-li¢ed (Figure 5B, C). The cDNA sequences of2d-1H5 (the same sequence as 2d-1H2) and 2d-2D9 clones are deposited in DDBJ/EMBL/Gen-Bank nucleotide sequence databases with acces-sion numbers (AB188532 for 2d-1H5 andAB188526 for 2d-2D9). The BLAST searchdemonstrated that these sequences had no sig-ni¢cant homology with other cDNA sequences indatabases.

When PCR was performed using primers from2d-1H2 and 2d-2D9 sequences (Figure 5B), an

about 2.5-kb female-speci¢c band was producedfrom both cDNA (from day 2 female embryos)and genomic sequences (Figure 5C). We speculatethat the 2d-1H2/2d-2D9 region may correspondto a large exon transcribed into the 30-untrans-lated region or the 2d-2D9 RNA is a non-codingRNA. Further cloning and sequencing ofupstream regions are required to elucidate thenature of the 2d-2D9 gene.

Discussion

Evaluation of the approach from the cDNAmacroarray

Altogether 5 cDNA clones derived from the Wchromosome-linked genes were identified in thisstudy from the female-minus-male subtractedcDNA macroarray constructed from the day 2 today 4 chicken embryos. Three clones were fromknown W-linked genes, PKCI-W, SPIN-W andATP5A1W, and two clones, represented by2d-2D9 and 2d-2F9, were from genes whose W-linkage has hitherto been unknown. We pre-dicted that at least 20 genes were present in theabout 10-Mb non-repetitive DNA region on thechicken W chromosome, among which six geneshave been identified (Mizuno et al. 2002, Reed &Sinclair 2002). In the present search, we set con-ditions that cDNA clones to be selected werederived from W chromosome-linked genes whichwere expressed in day 2 (stages 12–13) to day 4(stages 24–25) chicken embryos; i.e. develop-mental stages significantly earlier than the onsetof gonadal differentiation (day 6 to day 7 orstages 28–30) (Nakabayashi et al. 1998). The factthat five W chromosome-linked genes includingtwo new genes were identified in this studysuggests that the present cDNA macroarray-based selection was a proper approach for find-ing genes expressed under the above definedconditions.

*Note added in proof The latest BLAST search indicates that the deduced amino acid sequence of the 2d-2F9 cDNA clone shows

high homology to the C-terminal sequences of chEST83114 50(Accession No. BU134069) (93%) and predicted: Gallus gallus similar to

valosin precursor (Accession No. LOC430766) (82%), suggesting that 2d-2F9 corresponds to the C-terminal region of a VCP

(valosin-containing protein)-related protein belonging to the AAA ATPase superfamily.

752 D. Yamada et al.

Why was it that nearly all cDNA clones fromW-chromosome-linked genes were selected withsubtracted cDNA probes and obtained from day2 embryos?

The present selection of clones exhibiting sig-nificantly female-specific expression from thefemale-minus-male-subtracted cDNA macroarrayrevealed two distinct features:

(1) All of the 62 clones were selected by hybridi-zation with female-minus-male-subtractedcDNA probes but not with total cDNAprobes from female embryos (Table 1).

(2) All the clones except for 3d-1H2 (later identi-fied as SPIN-W) were obtained from thecDNA macroarray prepared from day 2(stages 12–13) embryos (Tables 1, 2).

The first feature may be interpreted as that thosecDNA clones exhibiting significantly female-specific expression were derived from mRNAspecies of relatively low-copy number. The cDNAsubtraction method employed in this study wasprimarily intended to concentrate female-specificcDNA species but also expected to some extentto equalize copy numbers of individual cDNAspecies. Thus, female-specific cDNA clones oflow copy number are expected to be selectedmuch more efficiently with female-minus-male-subtracted (female-enriched) cDNA probesthan with total female cDNA probes.

The second feature was a rather surprise tous because it indicates that several W-chromo-some-linked genes are expressed in femaleembryos as early as stages 12^13; i.e. stageseven earlier than the time (day 3 or aroundstage 20) of di¡erentiation of the genital ridgeon the median surface of the mesonephros(Romano¡ 1960). The exact reason why most ofthe W-chromosome-derived cDNA clones areobtained from day 2 but not from day 3 andday 4 embryos is unknown. We prepared andcompared female-minus-male subtracted cDNAmacroarrays from two di¡erent pools of day 3embryos but the number of W-chromosome-derived cDNA clones obtained was zero fromone and one (as presented in this paper) fromthe other macroarray. We speculate that con-

centrations of transcripts from several W chro-mosome-linked genes are relatively high in theday 2 female embryos but their concentrationsbecome markedly lower along with tissue di¡er-entiation and embryonic growth in subsequentstages of development. Consequently, the num-ber of those W-chromosome-derived cDNAclones both in the macroarray and in the probemay become undetectably low.

The two new W chromosome-linked genes

The present sequence-level studies suggest thata gene represented by a cDNA clone 2d-2F9 islikely to be a member of the AAA ATPasesuperfamily. On the other hand, a possiblefunction of the gene represented by a cDNAclone 2d-2D9 has not been suggested from thepresent sequence analysis. The latter cDNAspecies made up a great majority (47 out of 62)of the W-chromosome-derived clones selectedfrom the day 2 female embryos (Table 2). Wespeculate that the 2d-2D9-cDNA sequence ishighly W-chromosome specific judging fromSouthern blotting patterns for types A and B(Figure 2) and thus its cDNA molecules areenriched substantially after the process offemale-minus-male subtraction and/or that the2d-2D9 mRNA species are relatively abundant(yet to the level only selected with the female-minus-male-subtracted cDNA probe) in the day2 embryos and the equalization as mentionedabove was insufficient during the process ofcDNA subtraction.

Characterization of various aspects of functionand expression of 2d-2D9 and 2d-2F9 genes andtheir counterpart genes, if any, on the Z chromo-some in early embryonic stages is the subject ofour ongoing studies. Current progress of thechicken gene mapping and sequencing projects istruly remarkable but assignment of sequences andcontigs on the W chromosome has often beenuncertain because the majority of DNA on thechicken W chromosome consists of highly repeti-tive sequences (Mizuno et al. 2002) and numbersof available linkage map markers are small.Under such circumstances, a comprehensivecDNA-level approach such as the one carried outin this study may provide useful complementaryinformation.

Chicken W chromosome-linked genes 753

Acknowledgements

This work was supported by Grants-in-Aid forScientific Research on Priority Areas (2) from theMinistry of Education, Culture, Sports, Scienceand Technology, Japan to S. Mizuno. We thankTomoko Arai, Takumi Nagashima, SayakaHabata and Tatsuya Fukui of Nihon Universityfor selection of BAC clones.

References

Brenner C (2002) Hint, Fhit, and GalT: Function, structure,

evolution, and mechanism of three branches of the histidine

triad superfamily of nucleotide hydrolases and transferases.

Biochemistry 41: 9003–9014.

Clinton M (1994) A rapid protocol for sexing chick embryos

(Gallus g. domesticus). Anim Genet 25: 361–362.

Ellegren H, Carmichael A (2001) Multiple and independent

cessation of recombination between avian sex chromosomes.

Genetics 158: 325–331.

Fridolfsson AK, Cheng H, Copeland NG et al. (1998)

Evolution of the avian sex chromosomes from an ancestral

pair of autosomes. Proc Natl Acad Sci USA 95: 8147–8152.Frohlich KU (2001) An AAA family tree. J Cell Sci 114:

1601–1602.

Hamburger V, Hamilton HL (1951) A series of normal stages in

the development of the chick embryo. J Morphol 88:49–92.

Hori T, Suzuki Y, Solovei I et al. (1996) Characterization of

DNA sequences constituting the terminal heterochromatin

of the chicken Z chromosome. Chromosome Res 4: 411–426.Hori T, Asakawa S, Itoh Y, Shimizu N, Mizuno S (2000)

Wpkci, encoding an altered form of PKCI, is conserved

widely on the avian W chromosome and expressed in early

female embryos: implication of its role in female sex

determination. Mol Biol Cell 11: 3645–3660.

Itoh Y, Mizuno S (2002) Molecular and cytological char-

acterization of SspI-family repetitive sequence on the

chicken W chromosome. Chromosome Res 10: 499–511.

Itoh Y, Hori T, Saitoh H, Mizuno S (2001) Chicken spindlin

genes on W and Z chromosomes: transcriptional expression

of both genes and dynamic behavior of spindlin in

interphase and mitotic cells. Chromosome Res 9: 283–299.Kato J, Hattori T, Ohba S et al. (2002) Efficient selection of

genomic clones from a female chicken bacterial artificial

chromosome library by four-dimensional polymerase chain

reactions. Poultry Sci 81: 1501–1508.

Kodama H, Saitoh H, Tone M, Kuhara S, Sakaki Y, Mizuno

S (1987) Nucleotide sequences and unusual electrophoretic

behavior of the W chromosome-specific repeating DNA

units of the domestic fowl, Gallus gallus domesticus.

Chromosoma 96: 18–25.

Kuroda Y, Arai N, Arita M et al. (2001) Absence of Z-

chromosome inactivation for five genes in male chickens.

Chromosome Res 9: 457–468.

Kuroiwa A, Yokomine T, Sasaki H et al. (2002) Biallelic

expression of Z-linked genes in male chickens. Cytogenet

Genome Res 99: 310–314.

Matsuda M, Nagahama Y, Shinomiya A et al. (2002) DMY is

a Y-specific DM-domain gene required for male develop-

ment in the medaka fish. Nature 417: 559–563.

Mizuno S, Kunita R, Nakabayashi O et al. (2002) Z and W

chromosomes of chickens: studies on their gene functions in

sex determination and sex differentiation. Cytogenet Genome

Res 99: 236–244.

Nakabayashi O, Kikuchi H, Kikuchi T, Mizuno S (1998)

Differential expression of genes for aromatase and estrogen

receptor during the gonadal development in chicken

embryos. J Mol Endocrinol 20: 193–202.

O’Neill M, Binder M, Smith C et al. (2000) ASW: a gene

with conserved avianW-linkage and female specific expression

in chick embryonic gonad. Dev Genes Evol 210: 243–249.Raymond CS, Kettlewell JR, Hirsch B, Bardwell VJ, Zarkower

D (1999a) Expression of Dmrt1 in the genital ridge of mouse

and chicken embryos suggests a role in vertebrate sexual

development. Dev Biol 215: 208–220.

Raymond CS, Parker ED, Kettlewell JR et al. (1999b) A

region of human chromosome 9p required for testis

development contains two genes related to known sexual

regulators. Hum Mol Genet 8: 989–996.

Reed KJ, Sinclair AH (2002) FET-1: a novel W-linked, female-

specific gene up-regulated in the embryonic chicken ovary.

Gene Expr Patterns 2: 83–86.

Romanoff AL (1960) The genital system. In: The Avian

Embryo. Structural and Functional Development. New York:

Macmillan Company, pp. 816–820.

Saitoh Y, Saitoh H, Ohtomo K, Mizuno S (1991) Occupancy

of the majority of DNA in the chicken W chromosome

by bent-repetitive sequences. Chromosoma 101: 32–40.

Schmid M, Nanda I, Guttenbach M et al. (2000) First report

on chicken genes and chromosomes 2000. Cytogenet Cell

Genet 90: 169–218.

Shetty S, Kirby P, Zarkower D, Graves J A M (2002) DMRT1

in a ratite bird: evidence for a role in sex determination and

discovery of a putative regulatory element. Cytogenet

Genome Res 99: 245–251.

Suzuki Y, Sugano S (2001) Construction of full-

length-enriched cDNA libraries. The oligo-capping method.

Methods Mol Biol 175: 143–153.

754 D. Yamada et al.