Molecular heterogeneity of XY sex reversal in horses

Summary

Male-to-female 64,XY sex reversal is a frequently reported chromosome abnormality in horses. Despite this, the molecular causes of the condition are as yet poorly understood. This is partially because only limited molecular information is available for the horse Y chromosome (ECAY). Here, we used the recently developed ECAY map and carried out the first comprehensive study of the Y chromosome in XY mares (n = 18). The integrity of the ECAY in XY females was studied by FISH and PCR using markers evenly distributed along the euchromatic region. The results showed that the XY sex reversal condition in horses has two molecularly distinct forms: (i) a Y-linked form that is characterized by Y chromosome deletions and (ii) a non-Y-linked form where the Y chromosome of affected females is molecularly the same as in normal males. Further analysis of the Y-linked form (13 cases) showed that the condition is molecularly heterogeneous: the smallest deletions spanned about 21 kb, while the largest involved the entire euchromatic region. Regardless of the size, all deletions included the SRY gene. We show that the deletions were likely caused by inter-chromatid recombination events between repeated sequences in ECAY. Further, we hypothesize that the occurrence of SRY-negative XY females in some species (horse, human) but not in others (pig, dog) is because of differences in the organization of the Y chromosome. Finally, in contrast to the Y-linked SRY-negative form of equine XY sex reversal, the molecular causes of SRY-positive XY mares (5 cases) remain as yet undefined.

Introduction

Male-to-female sex reversal in horses is a developmental disorder in which phenotypic females or female-like animals are genetically males with 64,XY karyotype. The condition has been found in several breeds (Kent et al. 1986, 1988b; Makinen et al. 1999; Bugno et al. 2003; Switonski et al. 2005), and after X chromosome monosomy, it is the most common sex chromosome abnormality in horses (Lear & Bailey 2008; Villagomez et al. 2009). The male-to-female sex reversal has been described in humans (known also as Swyer syndrome) (Michala et al. 2008), cattle, river buffalo, and sheep (Ferrer et al. 2009; Villagomez et al. 2009).

Phenotypes of the XY sex reversal horses can vary in a broad range, from a very feminine to a greatly masculinized mare (Kent et al. 1988b). The majority of ‘feminine’-type animals have normal female external and internal genitalia, and no somatic or behavioural abnormalities (Power 1986; Bowling et al. 1987). Primary cause of infertility of these mares is typically because of ovarian dysgenesis and underdeveloped uterus (Power 1986; Pailhoux et al. 1995; Makinen et al. 1999). The overall phenotype of the ‘feminine’-type XY mares is very similar to those of 64,XO females (Bowling et al. 1987).

Another less frequently observed phenotype shows various degrees of masculinization and virilization (Kent et al. 1986; Long 1988; Lear & Bailey 2008; Villagomez et al. 2009). The ‘masculine’-type XY mares usually have abnormally developed genital tract, and the gonads can range from ovotestes to testicular feminization (Kent et al. 1986; Bowling et al. 1987; Kent et al. 1988a; Howden 2004; Villagomez et al. 2009), leading to male pseudohermaphrodites. All XY female horses studied so far have been found to be infertile, except one that gave birth to a normal 64,XX filly (Sharp et al. 1980).

Most equine male-to-female sex reversal cases are sporadic, or have no supporting pedigree data. However, in some instances, familiar patterns of inheritance, particularly in certain Arabian sire lines, have been observed (Kieffer 1976; Kent et al. 1986; Bowling et al. 1987; Kent et al. 1988b).

Despite the relatively frequent occurrence of XY mares among animals with chromosome abnormalities, the molecular causes of the condition are poorly understood. One of the reasons is the very limited knowledge about the molecular organization of the horse Y chromosome. The only molecular test, initiated in 1995 (Pailhoux et al. 1995), is PCR with the Y-linked male sex-determining gene SRY. According to this, two molecularly distinct types of XY females have been identified. The most prevalent are mares with no SRY gene (Pailhoux et al. 1995; Abe et al. 1999; Makinen et al. 1999; Bugno et al. 2003). Another type, represented by a single case, is an SRY-positive mare (Switonski et al. 2005).

In this study, we carried out the first comprehensive molecular analysis of the Y chromosome in XY sex-reversed mares. This was performed by sequencing the bacterial artificial chromosome (BAC) clones containing the SRY-region (Raudsepp et al. 2004b; Raudsepp & Chowdhary 2008b; Paria 2009) and developing markers that precisely defined the deletion in sex-reversed individuals.

Materials and methods

Animals

The study involved 18 female horses from various breeds. The animals were referred in 2001–2009 to the Texas A&M Molecular Cytogenetic Laboratory because of infertility and/or abnormal genitalia (Table 1). The laboratory received peripheral blood samples from each individual in sodium heparin and EDTA Vacutainers (VACUTAINER™; Becton Dickinson).

Table 1.   Summary of the phenotypes and genotypes of the female horses used in this study.

No	Breed	Phenotype	Karyotype	PCR with SRY	STS content analysis	FISH analysis
1	Thoroughbred	Abnormal cervix, small ovaries (7–12 mm)	64,XYdel	Neg	Deleted Contig I and hetrochromatin	Con I neg, the rest of Y pos
2	Thoroughbred	Small uterus and ovaries	64,XYdel	Neg	Deleted all MSY and PAR	All negative
3	Arabian	Male-like behaviour	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
4	Standardbred	Small uterus and ovaries, not cycling	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
5	Thoroughbred	Small uterus and ovaries, not cycling, aggressive in behaviour	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
6	Quarter Horse	Small uterus and ovaries, not cycling,aggressive in behaviour	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
7	Thoroughbred	Small stature, gonadal dysgenesis, suspected XO	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
8	Thoroughbred	Small ovaries (20 mm), no follicular activity	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
9	Arabian	N/a	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
10	Thoroughbred	Large stature, small ovaries (10 mm)	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
11	Thoroughbred	Immature reproductive tract, small ovaries	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
12	Thoroughbred	Normal stature, small ovaries (10 mm)	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
13	Standardbred	No pregnancies	64,XY	Neg	Deleted ∼21 kb around SRY	All positive
14	Appaloosa	Normal external genitalia, hypoplastic uterus,underdeveloped mammary glands. Shows oestrous behaviour to stallions.	64,XY	Pos	No ECAY deletions	All positive
15	Mixed warmblood	Abnormal: uterus, vulva located ventrally between rear legs at the location of mammary glands.	64,XY	Pos	No ECAY deletions	All positive
16	Standardbred	Male pseudohermaphrodite	64,XY	Pos	No ECAY deletions	All positive
17	Standardbred	Male pseudohermaphrodite	64,XY	Pos	No ECAY deletions	All positive
18	Quarter Horse	Male pseudohermaphrodite	64,XY	Pos	No ECAY deletions	All positive

Cell cultures and chromosome analysis

Cell cultures and chromosome preparations were made using standard cytogenetic methods as previously described (Raudsepp & Chowdhary 2008a). For karyotyping, the slides were stained with Giemsa, the sex chromosomes were identified by C-banding (Arrighi & Hsu 1971), and at least 20 cells were analysed for each technique.

Fluorescence in situ hybridization, FISH

The approximate size of Y chromosome deletions in XY females were determined by FISH using selected BAC clones from the physical map of ECAY (Raudsepp et al. 2004b; Raudsepp & Chowdhary 2008b) (Table 2). BAC DNA was isolated using the Plasmid Midi Kit (Qiagen) according to the manufacturer’s instructions. The BAC DNA was labelled and hybridized to metaphase chromosomes of XY females and control males following our standard FISH protocol (Raudsepp & Chowdhary 2008a). The results were analysed with a Zeiss Axioplan2 fluorescent microscope equipped with Isis V5.2 (MetaSystems GmbH) software.

Table 2.   Markers and BAC clones used for FISH analysis.

ECAY contig	Marker symbol	BAC clone	BAC library	Reference
PAR	PRKXY	164O9	CHORI-241	(Raudsepp & Chowdhary 2008a,b)
V	ZFY	110.3H12	TAMU	(Raudsepp et al. 2004a,b)
IV	AMELY	180P20	CHORI-241	(Raudsepp et al. 2004a,b)
III	NLGN4Y	112E12	CHORI-241	(Paria 2009)
II	YE1	118L7	CHORI-241	(Raudsepp et al. 2004a,b)
II	Y3B12	190M2	CHORI-241	(Raudsepp et al. 2004a,b)
I	KDM5D (SMCY)	60D8	CHORI-241	(Raudsepp et al. 2004a,b)
I	SRY	140M23	CHORI-241	(Raudsepp et al. 2004a,b)
		24I23	CHORI-241	(Raudsepp et al. 2004a,b)
		49.3F11	TAMU	(Raudsepp et al. 2004a,b)
		79.4H1	TAMU	(Raudsepp et al. 2004a,b)
		107.3H9	TAMU	(Raudsepp et al. 2004a,b)

Sequencing

Initial Y chromosome sequence information around the SRY gene was obtained by internal sequencing of two SRY-containing BAC clones, viz., 140M23 and 24I23 (CHORI-241 BAC library: http://bacpac.chori.org/quine241.htm). The first-step sequencing primers were designed in SRY-coding sequence and used for sequencing in both 3′ and 5′ directions. The newly obtained sequences served as templates to design primers for the next step. Such ‘walking’ was repeated several times. Sequencing reactions were carried out in 10 -μl volume with 1 μg of BAC DNA template and BigDye chemistry and resolved on an ABI-3730 capillary sequencer.

Next, the two SRY-containing BAC clones, viz., 140M23 and 24I23, were completely sequenced at Broad Institute of MIT and Harvard (Cambridge, MA, USA) using Sanger sequencing technology. The draft sequence GenBank Accessions are AC214740 and AC215855, respectively. Additionally, a third SRY-containing BAC clone, 107.3H9 from TAMU BAC library (http://hbz7.tamu.edu/homelinks/bac_est/bac.htm), was completely sequenced at Research and Testing Laboratory (Lubbock) following standard Roche 454-sequencing protocols (Roche). A total of 260 150 sequences with an average length of 421 bp were derived and initially assembled using NGen (DNAstar). The contigs were aligned with SeqMan, resulting in a continuous 116 857 -bp sequence (GenBank HM103387). Sequences of the three BACs and all available STSs in the SRY-region (Raudsepp et al. 2004b) were aligned using Sequencher 4.7 (Gene Codes) software and analysed using RepeatMasker (http://www.repeatmasker.org/), GENSCAN (http://genes.mit.edu/GENSCAN.html) and NCBI BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi).

Finally, 1 185 bp of the single exon of SRY (Table 3), including the 684 -bp open reading frame, was sequenced from all SRY-positive sex-reversal females using BigDye chemistry. The sequences were aligned with the reference sequence (GenBank AB004572) using Sequencher 4.7 (Gene Codes) software.

Table 3.   Summary data for horse Y chromosome genes and STS markers.

Marker	ECAY region	Primer Sequences 5′–3″	PCR product size, bp	Reference	GenBank Accession No
107-STS1	Con I: SRY-region	F: TGGGCAGTTATGAAAATACGA R: CCTGCACACTGTCCTGTAATTG	121	This study	HM103387
107-STS2	Con I: SRY-region	F: TTTTGTTGGGTCATTTGTGG R: ATCTACATGCTGGGGCAAAC	381	This study	HM103387
107-STS3	Con I: SRY-region	F: TCACTGCAGGAACATTCTGG R: CAAGCAAACAGATTGGAGCA	196	This study	HM103387
107-STS4	Con I: SRY-region	F: CATGGAGGAGAGAGCCTTTG R: TCAGCTGAGGTTTTCCCAAT	425	This study	HM103387
110.3H12-M13	Con V	F: GGGCCAGAATATGCAAGGA R: GATGTGTTTGTGTGCCTCTGTT	182	Raudsepp et al. 2004a,b	BV140834
81F8-SP6	Con III	F: GCAAAGGCTCTGAGAGAGGA R: CCATGCTTCTTGACCAGACA	182	Paria 2009	CU011557
118L7-T7	Con II	F: ATCTGCTCCCCTTTGGTTTT R: CCCCAGATTTACTGCCTTTG	225	Paria 2009	CU025611
126G2-T7	Con II	F: GCAACTTGCACTGATTGTCC R: ATTTGTGTGGAGGGCAGGT	200	This study	n/a
131N23-SP6	Con V	F: GCTTGTCCTAATTTGCCTCTG R: CTTTGAGGACGGGTAAATTG	152	Raudsepp &Chowdhary 2008a,b	CU034701
168I4-T7	Con II	F: CCATTTCAACTCCTGATCCA R: AAGGGTAGGTCAGAGAGGAAGG	101	Raudsepp et al. 2004a,b	BV140781
24I23-T7	Con I: SRY-region	F: ATCTGCTTCGGCCTTCTCTT R: GGTGTACCCTGCTTCTCGT	101	Raudsepp et al. 2004a,b	BV140808, CT942288, AC215855
26B21-SP6	Con I	F: CAAGGAAGCCAGGAAGAGTG R: GTCTCTGGCCCATGAGTGAT	160	Paria 2009	CT951493
3′STS4	Con I: SRY-region	F: ATGTCCAGCTCCTTTTGCTT R: ATGATGTGGGGTCGGTGA	690	This study	AC214740
3′STS5	Con I: SRY-region	F: AAAACAGAGGTTGGCCTTGA R: GCTGTGGACGAAAATAATCCA	602	This study	AC214740
3′STS6	Con I: SRY-region	F: ATTTTTCCCCTTGCATCCTT R: CCTCCAATTTGCGTTTGTTT	838	This study	AC214740
3′STS7	Con I: SRY-region	F: TCGCTGCTTTCTTTAGTGACTC R: CCTGAAGAGCTGTTCTAAATGGA	797	This study	AC214740
3′STS8	Con I: SRY-region	F: CAGGAGCTTTGTGAAAGATGG R: TTGTTGACTCTTGGTCAATGTT	596	This study	AC214740
3′STS9	Con I: SRY-region	F: TGAACCACAAGGGCATTTTT R: GACACGTTAAAATTGGGAAGAA	218	This study	AC214740
3′STS10	Con I: SRY-region	F: TTCTCATAACTGCCCATAGACTACA R: ACGTGGATAATCCCTGCTTG	492	This study	AC214740
3′STS12	Con I: SRY-region	F: GTGATTATGGCCGTTCGAGT R: TGAGACCGCCAACTTAACCT	399	This study	AC214740
3′STS13	Con I: SRY-region	F: TGGAGCTTCCAGAGGAGGTA R: GCCTGCCTAACTGGAAAACA	509	This study	AC214740
419P11-T7	PAR	F: ACAGGGCAGCTCCTACAAGA R: CCCAGAAAAAGCCCGTAAG	151	Raudsepp &Chowdhary 2008a,b	CU267996
49.3F11-M13	Con I: SRY-region	F: TGAAATCACAAAGCTCCAAATG R: GGTTACGAATGGCTCCTTGA	172	Raudsepp et al. 2004a,b	BV140820
5.2A8-M13	Con IV	F: TCAGAATGGAGCTGCCTAAAA R: GGCCCTGGCTTTTCTTTATT	274	This study	n/a
5′STS1	Con I: SRY-region	F: ACTCCTCCCCTCAAACGTG R: ATTTTAAGGAGGGGCAGAGC	150	This study	AC214740, AC215855
5′STS2	Con I: SRY-region	F: GCTCTGCCCCTCCTTAAAAT R: AACGATGATGTTCACCATGAAT	226	This study	AC214740, AC215855
5′STS3	Con I: SRY-region	F: TAGTTTAGCTCGGGGCCAAT R: TGCGATCTGCACTTGCTACT	347	This study	AC214740, AC215855
5′STS4	Con I: SRY-region	F: CAAACAACCCACTTAACCAAA R: CATACCCTATTTTATTTGGATCTGA	177	This study	AC214740, AC215855
5′STS5	Con I: SRY-region	F: TGTACATGTTGGGGAAGTGG R: ATTGAAGGCCTGTTCCAAGA	399	This study	AC214740, AC215855
5′STS6	Con I: SRY-region	F: TCACCAGTTTGGCTACTGGA R: TCCCCAACATGTACACCTCA	174	This study	AC214740, AC215855
5′STS7	Con I: SRY-region	F: AATGACGACATTTTAAAAGTACAAGTT R: TTTTCCTCATTGGCATCAAG	153	This study	AC214740, AC215855
5′STS8	Con I: SRY-region	F: TTCTAGGTCATCATTGGTGCTC R: TGGCTATCTTCAAAAGACATTAAGAA	848	This study	AC214740, AC215855
5′STS9	Con I: SRY-region	F: GCCATTTTTCCTTTGTTTGC R: CATGGTCGAGCGGTTAAGTT	645	This study	AC214740, AC215855
5′STS10	Con I: SRY-region	F: TGGCAAGTCTTGGCTTTCTT R: TAGAGGCCTTGGGCTTGTAA	188	This study	AC214740, AC215855
5′STS12	Con I: SRY-region	F: GCCCTTCATTCCTTTTCCTT R: AAAAGGCGACTCCAGGGTAT	754	This study	AC214740, AC215855
5′STS13	Con I: SRY-region	F: GGTTTTCCACGTTGTGGTTT R: GGTCAAATTGGGTATGGTCAA	755	This study	AC214740, AC215855
5′STS14	Con I: SRY-region	F: CACTAAACTGCAAAGACATTTTTCAT R: CAGGACGGTGTGCAGAGTAA	192	This study	AC214740, AC215855
5′STS16	Con I: SRY-region	F: TGTGGTGTGTATTGGAGGAGA R: TGGGGACAATGTGTCAGAAA	552	This study	AC214740, AC215855
5′STS17	Con I: SRY-region	F: AGCCCTGCAGATGCTAAGAA R: GAAGCCCAAAAGAGTCCTTG	679	This study	AC214740, AC215855
66M24-SP6	Con III	F: TTGGCTGCTCTCAGAATTGA R: GGATTGTGCATGTCATCTCG	169	This study	CU001288
79.4H1-M13	Con I: SRY-region	F: AGTCCACACCACCACAGTGA R: TGATGGGAGTGGGAGTTTTC	248	Raudsepp et al. 2004a,b	BV140821
90B11-SP6	Con IV	F: TTGCTTGTGGTATTTCCTTGC R: CCTTTCTTCTTTCCTTACATGC	150	This study	BV140791, CU015692
AMELY	Con IV	F: CCAACCCAACACCACCAGCCAA ACCTCCCT R: AGCATAGGGGGCAAGGGCTGCA AGGGGAAT	160	Hasegawa et al. 2000	AB032194
KDM5D	Con I	F: AACAGCGAGCCAATGTTTTT R: GCAAAATTCTGGGAAATCCA	400	Agulnik et al. 1997; Lindgren et al. 2001	EU687564, U52364
NLGN4Y	Con III	F:GGGGATCCATCTTTGTGTTG R:GTCACACAGCAGGCTCTGAC	156	Paria 2009	EU687560
PRKXY	PAR	F: CGCTTCCTGTACATGCTGATGGAG R: TCTTGGCGAACCCGAAGTCGGTGA	229	Raudsepp &Chowdhary 2008a,b	ET052965
RBMY	Con I: SRY-region	F: TTCGGCCTTCTCTTTCACAT R: ACTCAAGCAGCCGAAATGAT	180	Paria 2009	EU687561
SH3B14	Con I	F: GTGACCTCCCAGGAGCTGT R: TCTGCCTATGCTCTGGTGAA	486	Wallner et al. 2004	BV005745
SRY1	Con I	F:TGCATTCATGGTGTGGTCTC R:ATGGCAATTTTTCGGCTTC	200	Paria 2009	EU687565
SRY2	ConI	F: CTTAAGCTTCTGCTATGTCCAGAGTATCC R: GGGGAGGGGAGGTAACATAA	1,185	This study	EU687565, AC214740, AC215855
Y3B12	Con II	F: GGGAGGCACTGGAAAGTACA R: GGTGGAGGAATCAGCTGGAG	400	Wallner et al. 2004	G72338
YE1	Con II	F: CTTCACTCCCGACCAAGAGA R: GTGTGTCGTGCCGTGTTTAC	199	Wallner et al. 2004	BV005727
YM2	Con I	F: TGGTTCAGATGGTGTATTTTGTT R: TTTGCAGCCAGTACCTACCTT	119	Wallner et al. 2004	BV005725
ZFY	Con V	F: TGCACATTTCCTTTAATCT R: GCACATTAAAGAGAAACCTT	342	Lindgren et al. 2001	AF178847 AF133198

Development of STS markers and STS content analysis

Initial analysis of the Y chromosome in XY mares was carried out by PCR using primers for the horse SRY gene and selected STS markers from ECAY maps (Raudsepp et al. 2004b; Raudsepp & Chowdhary 2008b) (Table 3). Thereafter, additional STS primers were designed from masked (RepeatMasker: http://www.repeatmasker.org/) sequences of the three completely sequenced BAC clones. Androgen receptor (AR) from the X chromosome (Raudsepp et al. 2004a) was used as a positive control for all PCR amplifications. A summary of the PCR primers developed and used in this study is presented in Table 3. The same primers were used for sequencing the PCR products and/or for internal BAC sequencing in the SRY-region.

STS content analysis was carried out on the genomic DNA isolated from EDTA-stabilized peripheral blood of the 18 XY females (Table 1). DNA from ‘Bravo’, the donor for CHORI-241 BAC library, served as the male control. Similarly, DNA from ‘Twilight’, the donor for the genome sequence project (Wade et al., 2009), served as the female control. STS content analysis for each animal was carried out in 10 -μl volume duplicate PCR with 0.25 units Taq polymerase (JumpStart RedTaq, Sigma Aldrich), 50 mM KCl, 10 mM Tris–HCl (pH 8.4), 0.2 mM dATP, dCTP, dTTP and dGTP, 0.3 μM of each primer, and 50 ng of genomic DNA as template. The PCR products were resolved in 2% agarose gels containing ethidium bromide.

Results

Cytogenetic analysis

Cytogenetic analysis of Giemsa-stained metaphase spreads revealed that all 18 female horses (Table 1) had normal diploid chromosome number (2n = 64) with the XY sex chromosome complement, which is typical of males. This was confirmed by C-banding, which clearly showed the presence of one X chromosome and one Y chromosome in all cells analysed (Fig. 1a–d). Animals 1 and 2 had extremely small Y chromosomes (Fig. 1c, d), compared to other XY females and male controls. In contrast, the Y chromosome in Animal 15 was larger than usual with a prominent and extended heterochromatic region (Fig. 1b).

Figure 1.  C-banded metaphase spreads (a–b) and Giemsa-stained karyotypes (c–d) of Animal 6 (a), Animal 15 (b), Animal 1 (c), and Animal 2 (d).

The presence of the Y chromosome was further confirmed by FISH using microdissected horse Y chromosome (Raudsepp & Chowdhary 1999) as a probe (Fig. 2a–b). Hybridization results confirmed the cytogenetic observations that the Y chromosomes in Animals 1 and 2 were abnormally small (Fig. 2b). Moreover, FISH with 8 BAC clones evenly distributed along the ECAY contig map (Table. 2, Fig. 3a) showed that the two rudimentary Y chromosomes are molecularly different. The Y chromosome in Animal 1 gave FISH signals with all contig II–V and PAR BACs (Fig. 2c, d) and was negative only for contig I clones containing SRY and KDM5D (alias SMCY). In contrast, no FISH signals with any of the probes were observed in Animal 2, indicating that the rudimentary Y chromosome is comprised of little or no euchromatin (Fig. 3b). In the remaining 16 animals, FISH signals were observed with all MSY and PAR BACs, including the five overlapping clones in the SRY-region (Table 2, Fig. 2e, f; Fig. 3c). Taken together, the cytogenetic and FISH analyses detected massive Y chromosome deletions over 10–15 Mb size (Raudsepp et al. 2004b) in two animals, whereas the gross organization of the Y chromosomes in all other XY mares appeared to be normal.

Figure 2.  Fluorescence in situ hybridization with microdissected Y chromosome painting probe (a–b), PAR gene GYG2 (c), MSY markers YE1 (green), and ZFY (red) (d), and BAC clones from SRY-region 24I23 (green) and 107.3H9 (red) (e–f). Hybridizations were carried out on metaphase chromosomes of Animal 15 (a and f), Animal 1 (b–d) and Animal 9 (e) (see Table 1 for details).

Figure 3.  Schematic showing various ECAY deletions in XY females. (a). Composite map of ECAY showing the heterochromatic region, five BAC contigs (I–V) in the male-specific region of Y (MSY), and the pseudoautosomal region (PAR). Markers that were used for FISH (bold and shaded grey) and PCR analysis are shown at the top of each region. Gene symbols are in italics. (b). Massive ECAY deletions (dark grey bars) in two XY females. (c). Contig of 5 overlapping BAC clones in the SRY-region. Clones demarcated in bold were completely sequenced. (d). Detailed map of a 146 432 -bp region of MSY showing the size and marker content of the deletion, the two directed repeats (yellow and green arrows R1 and R2), and MSY (blue bars).

Sequence analysis of the Y chromosome

Analysis by PCR with SRY primers divided the 18 XY females into two molecularly distinct groups: SRY negative (Animals 1–13, Table 1) and SRY positive (Animals 14–18). All XY mares were positive for the AR control gene. Next, we used a panel of 20 MSY and PAR markers (Fig. 3a) and studied by PCR their presence or absence in individual animals. As expected, all tested markers were negative in Animal 2, confirming the FISH results that the small Y chromosome is exclusively heterochromatic. In Animal 1, only markers from contigs II–V and the PAR were present. No amplification by PCR was observed with the 4 markers in contig I: YM2, KDM5D, SH3B14, and 26B21-SP6 (Fig. 3a). Thus, the PCR and FISH results for Animal 1 were in complete agreement. Similarly, PCR analysis confirmed and refined FISH results for the remaining animals. All 20 ECAY markers were positive in Animals 3–18, showing that regardless of the presence or absence of SRY, the overall integrity of the Y chromosome was retained.

Next, we investigated whether the deletion of SRY is the only molecular signature of the Y chromosome in the 11 SRY-negative females (Animals 3–13). We hypothesized that if only the single exon of SRY (Hasegawa et al. 1999) is deleted, it should be possible to design a simple PCR test to distinguish between normal males and XY SRY-negative females. We reasoned that if primers are designed to flank the deletion, distinctly different-sized PCR products would be amplified from normal males and XY females. Such tests will show the size of deleted segments in sex-reversed females and will be a useful alternative to the current presence- or absence-based PCR analysis with SRY primers.

To develop this test, we carried out a detailed STS content analysis with a panel of 34 markers in a 146 -kb region around SRY that was developed from three completely sequenced BAC clones (Fig. 3d). The analysis revealed that the deletion is not restricted to the SRY exon and includes at least 21 500 bp around it (Fig. 3d). We discovered that the deletion is flanked by approximately 28 kb of directed repeats (R1 and R2 arrows in Fig. 3d) that are present both in normal males and XY females. As the duplicated regions share a high degree of sequence similarity (approximately 96–100%), it was not possible to determine by PCR whether one or both segments are present in sex-reversed animals. As shown in Fig. 3d, the unique MSY-specific sequences (blue shades), which are suitable for primer design, are located tens of kilobase pairs away from the deletion. Thus, it was not possible to design primers for an alternative PCR test to distinguish between Y chromosomes in normal males and SRY-negative females.

Sequence analysis of the 146 432 -bp genomic segment around the deletion (Fig. 3d) identified three MSY genes –SRY, RBMY, and ATP6V0CY. The single-copy SRY is part of the deletion, a multicopy gene RBMY (Paria 2009) is located in both directionally duplicated segments on either side of the deletion (Fig. 3d), and ATP6V0CY is a newly found Y-linked gene in horses.

Finally, FISH and STS content analyses did not detect any molecular differences between the Y chromosomes of SRY-positive XY females and normal males. Likewise, sequences of the single exon of SRY in all 5 SRY-positive females were 100% identical with each other and with the male reference sequence (data not shown). Therefore, the detailed analysis of the Y chromosome in XY females showed that the molecular causes of SRY-positive male-to-female sex reversal in horses are not Y-linked and remain as yet undefined.

Discussion

To our knowledge, this study represents the first comprehensive molecular characterization of the Y chromosome in XY sex-reversed female horses. Using recently available molecular tools for MSY and the PAR (Raudsepp et al. 2004b; Raudsepp & Chowdhary 2008b; Paria 2009), we considerably refined the underlying molecular aetiology of the SRY-negative (Pailhoux et al. 1995; Abe et al. 1999; Makinen et al. 1999, 2001; Bugno et al. 2003) form of sex reversal in horses.

Molecular heterogeneity of ECAY in SRY-negative XY females

A combination of cytogenetic, FISH, STS content, and sequence analysis revealed that the SRY-negative form of sex reversal is a molecularly heterogenous condition. Deletions that lead to the loss of SRY can be massive, including megabase pairs of DNA (Fig. 3b), or limited to a region around SRY, removing tens of kilobase pairs of DNA (Fig. 3d). Notably, none of the deletions is restricted to the SRY sequence only. Furthermore, as illustrated by Animals 1 and 2 (Table 1), the massive deletions differ from each other both in size and content (Fig. 3b). The loss of entire ECAY euchromatin in Animal 2 makes this case genetically very similar to X monosomy. This is in agreement with the animal’s phenotype (Table 1) and supports earlier observations that mares with XY and XO sex chromosome complements are frequently indistinguishable by overall appearance (small stature), behaviour (abnormal oestrus behaviour), and gonadal phenotype (underdeveloped ovaries and uterus) (Chandley et al. 1975; Bowling et al. 1987; Long 1988). However, as none of the earlier cytogenetic studies report about finding an abnormally small Y chromosome, the molecular cause of the XO-like phenotype in XY mares is probably not because of the massive loss of Y euchromatin. Female gonadal dysgenesis might rather be because of the absence of the second X chromosome.

Systematic mapping of the horse Y chromosome and the development of Y-linked markers started only recently (Raudsepp et al. 2004b). This is why all earlier molecular studies of XY mares have been limited to a few Y-linked genes, viz., SRY (Pailhoux et al. 1995; Abe et al. 1999; Makinen et al. 1999, 2001; Bugno et al. 2003), AMELY, STS-Y and ZFY (Makinen et al. 2001; Bugno et al. 2003). These tests confirmed the absence of SRY, and the presence of the three other markers, but had no knowledge about the relative order or map location of these genes. Therefore, the present study is the first in which the Y chromosomes of XY mares were analysed using detailed map information (Raudsepp et al. 2004b; Raudsepp & Chowdhary 2008b) and partial sequence data for ECAY. This allowed fine demarcation of ECAY deletions (Fig. 3) and showed that all SRY-negative mares, except Animals 1 and 2, share the same large 21.5 -kb deletion that includes SRY (Fig. 3d) but no other known functional MSY genes. The results are in agreement with the common idea that the female-like phenotype or the lack of male phenotype in SRY-negative mares is primarily because of the loss of SRY (Pailhoux et al. 1995; Makinen et al. 1999). However, the cause of female gonadal dysgenesis, another characteristic feature of both the XY and XO mares (Table 1), is not so clear. In females with X monosomy, haploinsufficiency for certain PAR genes has been considered as the cause of gonadal dysgenesis (Ellison et al. 1997; Blaschke & Rappold 2001). In this study, using SRY and a panel of 20 Y-linked markers (Fig. 3a), we showed that, except for the SRY-region, the rest of the Y chromosome is undisturbed. Thus, in contrast to XO females, the SRY-negative XY mares have two normal copies of the PAR. Therefore, given that the molecular causes of gonadal dysgenesis in XY and XO females are the same, they are more likely associated with incomplete dosage of some non-PAR Xp genes that escape X inactivation.

Possible molecular mechanisms of Y chromosome deletions

Previously, it has been proposed that SRY-negative sex-reversal syndrome in horses is because of abnormal X-Y recombination in male meiosis, resulting in the loss of SRY from the Y to the X chromosome (Abe et al. 1999; Makinen et al. 2001; Bugno et al. 2003). The assumption was based on observations in humans where such uneven crossovers are common and give rise to 46,XY SRY-negative females and 46,XX SRY-positive males (Rosser et al. 2009). However, it must be noted that the molecular organization of the human and horse Y chromosomes are very different. In humans, the SRY is located only 35 kb away from the pseudoautosomal boundary (PAB) (Skaletsky et al. 2003), which makes uneven X-Y meiotic exchange feasible. In contrast, the horse SRY is located in the very proximal part of MSY (Raudsepp et al. 2004b), far from the PAB (Fig. 3a), and its involvement in abnormal meiotic exchange is unlikely (Raudsepp & Chowdhary 2008b).

A more plausible explanation of Y deletions lies in the intrinsic nature of Y chromosome sequences. A typical feature of the haploid and non-recombining Y chromosome is the presence of massive and numerous reversed (palindromic) and directional repeats (Skaletsky et al. 2003; Lange et al. 2009; Hughes et al. 2010). Sequence similarity between palindrome arms or repeated units might be as high as 99%–100%, facilitating inter-and intra-chromatid gene conversion and recombination (Lange et al. 2009). As such genetic exchanges help to maintain Y chromosome structural and functional integrity, the majority of important spermatogenesis genes are located in palindrome arms or directional repeats (Skaletsky et al. 2003; Lange et al. 2009). However, this mechanism also has a downside: inter-chromatid recombination can frequently lead to duplications and deletions which, in humans, are associated with spermatogenic failure, sex reversal, and Turner syndrome (Lange et al. 2009). In this study, we show that in the horse Y chromosome, the SRY-region is surrounded by two almost 100% identical directional repeats (Figs 3d & 4a). Not coincidentally, both repeated segments contain one copy of a known spermatogenesis-related gene, RBMY (Elliott 2004). It is possible that the deletion of the SRY-region, as seen in SRY-negative XY mares, is the result of an inter-chromatid recombination between the two repeats (Fig. 4b). Theoretically, this inter-chromatid exchange should remove from the Y chromosome of XY females not only SRY but also one of the directed repeats (Fig. 4c). However, this can be proven only by direct Y chromosome sequencing, and not by PCR. If our assumption is correct, the SRY-region is expected to become duplicated on another chromatid (Fig. 4c) and should be present in the Y chromosomes of paternal male full- or half-sibs of the XY females. Possibly, duplication of SRY and triplication of repeats have no significant phenotypic effects, and the carriers remain undetected. Further, we infer that the molecular mechanism of the two massive deletions in Animals 1 and 2 might be quite similar to that presented in Fig. 4, with the only difference that recombination events have taken place between more distantly located repeats resulting in larger deletions.

Figure 4.  Schematic of the likely mechanism of SRY deletion in XY females (adopted from Lange et al. 2009). (a). Double-stranded break (DSB) occurs in one of the directed repeats (R2) flanking SRY. (b). DSB is repaired by inter-chromatid homologous recombination using R1 as a template to repair DSB in R2. (c). Homologous recombination between sister chromatids of ECAY deletes SRY and R2 from one chromatid and moves them to another.

Taken together, the key factors causing SRY-negative sex reversal are Y-linked, although some X-linked genes might also contribute to the phenotype. Thus, the primary mutation, sporadic or genetically predisposed, is derived paternally. Indeed, as shown in earlier studies, there are certain sire lines that produce more XY females than expected by chance (Kieffer 1976; Kent et al. 1986; Bowling et al. 1987; Kent et al. 1988b). However, the 13 SRY-negative XY mares analysed in this study were from different breeds (Table 1) and families and were therefore not informative regarding the pedigree analysis.

The SRY-positive male-to-female sex reversal

The SRY-positive females are phenotypically distinct from the SRY-negative mares, as shown by a single previous report (Switonski et al. 2005) and this study. These mares display features that are not typically observed in SRY-negative animals, such as stallion-like behaviour, large body size, abnormal external genitalia, elevated blood testosterone levels, and male rather than female gonadal dysgenesis (Table 1). Based on this, it is possible that many more SRY-positive XY females have been described earlier, but because of the lack of molecular tools, they were not properly identified (Kieffer 1976; Kent et al. 1986, 1988a,b; Crabbe et al. 1992; Howden 2004). Therefore, the ‘masculine’-type equine XY sex reversal might be more prevalent than the single previous report (Switonski et al. 2005) and the 5 cases described in this study.

Detailed molecular analysis of the Y chromosome showed that the ‘feminine’ (SRY-negative) and ‘masculine’ (SRY-positive) XY sex reversals differed genetically more than just the absence or presence of SRY. While SRY-negative animals showed various ECAY deletions, the Y chromosome of SRY-positive mares was the same as in normal males; even the unusually large Y chromosome in Animal 15 was molecularly normal by PCR, FISH, and sequence analysis, and the abnormal size could be attributed to substantially enlarged heterochromatic region (Fig. 1b). We infer that the SRY-positive condition is likely not Y-linked. However, human studies suggest that a Y-linked ‘growth’ gene might be responsible for the larger than usual body size (adult height) (Ogata & Matsuo 1992; McDonough 2003) of XY females. Further, it has been proposed that testicular feminization and male pseudohermaphroditism in SRY-positive mares might be caused by a mutation in the X-linked androgen receptor (AR) gene (Crabbe et al. 1992; Howden 2004; Switonski et al. 2005). To date, there is no experimental evidence to support this. It is also possible that, similarly to humans, the SRY-positive sex reversal is a genetically heterogeneous disorder (Sarafoglou & Ostrer 2000). In humans, some cases are Y-linked, showing different missense and frameshift mutations in the SRY-coding region (Salehi et al. 2006; Shahid et al. 2008; Marchina et al. 2009). However, almost 80% of 46,XY women have no mutations in SRY or other Y-linked genes, indicating the involvement of X-linked or autosomal factors. The few different autosomal (Barbaro et al. 2009; Biason-Lauber et al. 2009; Schimmer & White 2010) and X-linked (Sarafoglou & Ostrer 2000) mutations described in XY women do not give consistent answers, and the underlying genetic causes of SRY-positive sex reversal in humans and horses alike remain poorly understood.

Occurrence of XY sex reversal in horses and other species

Male-to-female XY sex reversal, especially the SRY-negative form, is a relatively frequent chromosomal abnormality in horses. Incidence of XY mares in our cytogenetic analysis practice is as high as 26% among all chromosomally abnormal animals studied during the period between 2001 and 2009. This is in line with earlier data showing that 12%–30% of cytogenetic abnormalities in horses count for XY females (Bowling et al. 1987; Power 1990). The condition has also been described in humans (Sarafoglou & Ostrer 2000; Michala et al. 2008), where about 10%–20% XY women are SRY negative, and 80–90% have normal SRY. Both SRY-positive and SRY-negative XY females have been found in cattle (Ferrer et al. 2009; Villagomez et al. 2009) and mouse (Arnold & Chen 2009). In contrast, surprisingly few and exclusively SRY-positive ‘masculine’-type XY females have been reported for other species, such as river buffalo (Di Meo et al. 2008), dog (Nowacka-Woszuk et al. 2007; Whyte et al. 2009), and sheep (Ferrer et al. 2009). Notably, no XY sows have been found in pigs (Villagomez et al. 2009). This disparity might be because of uneven cytogenetic sampling of populations in different species. However, given the well-organized cytogenetic screening system of domestic species, especially pigs, in France and many other European countries (Ducos et al. 2008), this is unlikely. Instead, we hypothesize that the prevalence of Y-linked XY SRY-negative condition is related to the diversity of the organization of the Y chromosome in different species (Skaletsky et al. 2003; Raudsepp et al. 2004b; Murphy et al. 2006; Hughes et al. 2010). For example, SRY is a single-copy gene in human (Skaletsky et al. 2003), mouse (Lundrigan & Tucker 1997), and horse (Paria 2009). This implies that any Y chromosome rearrangement that causes the loss of SRY will result in an SRY-negative condition. In contrast, species like cat (Pearks Wilkerson et al. 2008), rabbit (Geraldes & Ferrand 2006), rat (Turner et al. 2007), and several other rodents (Lundrigan & Tucker 1997) have multiple copies of SRY. Consequently, deletion of one SRY copy leaves other copies intact to carry out their function. Therefore, one plausible reason why species like pigs, dogs or sheep have no SRY-negative XY females might be the presence of multicopy SRY. It could also be that SRY location in relation to inverted and directional repeats is different across species, thus facilitating SRY deletions in some species but not in others. It is anticipated that the ongoing mapping and sequencing projects for the mammalian Y chromosomes will soon provide better explanations for these ideas. Likewise, the availability of genome-wide analysis tools, such as SNP-chips and tiling arrays, are expected to improve the discovery of genes and rearrangements underlying the SRY-positive form of sex reversal in horses and other species.

Acknowledgements

This work was supported by USDA grant 2006-04801 and CVMBS LINK Endowment for Equine Genomics. The University of Kentucky Agricultural Experiment Station number for this paper is 10-14-077.

Conflict of interest

The authors have declared no potential conflicts.