Sry Expression Level and Protein Isoform Differences Play a Role in Abnormal Testis Development in C57BL/6J Mice Carrying Certain Sry Alleles

Corresponding author: Eva M. Eicher, 600 Main St., Bar Harbor, ME 04609., eme{at}jax.org (E-mail)

Communicating editor: N. A. JENKINS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Transfer of certain Mus domesticus-derived Y chromosomes (Sry^DOM alleles, e.g., Sry^POS and Sry^AKR) onto the C57BL/6J (B6) mouse strain causes abnormal gonad development due to an aberrant interaction between the Sry^DOM allele and the B6-derived autosomal (tda) genes. For example, B6 XY^POS fetuses develop ovaries and ovotestes and B6 XY^AKR fetuses have delayed testis cord development. To test whether abnormal testis development is caused by insufficient Sry^DOM expression, two approaches were used. First, gonad development and relative Sry expression levels were examined in fetal gonads from two strains of B6 mice that contained a single M. domesticus-derived and a single M. musculus-derived Sry allele (B6-Y^POS,RIII and B6-Y^AKR,RIII). In both cases, presence of the M. musculus Sry^RIII allele corrected abnormal testis development. On the B6 background, Sry^POS was expressed at about half the level of Sry^RIII whereas Sry^AKR and Sry^RIII were equally expressed. On an F₁ hybrid background, both Sry^POS and Sry^RIII expression increased, but Sry^POS expression increased to a greater extent. Second, sexual development and Sry expression levels were determined in XX mice carrying a transgene expressing Sry^POS controlled by POS-derived or MUS-derived regulatory regions. In both cases one B6 transgenic line was recovered in which XX transgenic mice developed only testicular tissue but cord development was delayed despite normal Sry transcriptional initiation and overexpression. For three transgenes where B6 XX transgenic mice developed as females, hermaphrodites, or males, the percentage of XX transgenic males increased on an F₁ background. For the one transgene examined, Sry expression increased on an F₁ background. These results support a model in which delayed testis development is caused by the presence of particular DOM SRY protein isoforms and this, combined with insufficient Sry expression, causes sex reversal. These results also indicate that at least one tda gene regulates Sry expression, possibly by directly binding to Sry regulatory regions.

NORMALLY in mammals, XX individuals develop as females with ovaries, and XY individuals develop as males with testes. Although rare, complete sex reversal (SR) occurs in which XX individuals develop testes and XY individuals develop ovaries. In humans the easiest SR cases to explain are XY females who carry a nonfunctional SRY (sex-determining region, Y chromosome, symbolized as Sry in mice) gene and XX males who carry a normal SRY gene located on their paternally derived X chromosome due to an abnormal meiotic recombination event (reviewed in SCHAFER 1995 ; MCELREAVEY and FELLOUS 1999 ). Several intriguing but unexplained SR conditions occur, however, including XY females and XY hermaphrodites who carry an apparently normal SRY gene and XY females who carry a mutated SRY gene inherited from their father who carried the same mutated SRY allele (reviewed in SCHAFER 1995 ; MCELREAVEY and FELLOUS 1999 ). These cases are reminiscent of what occurs in mice when certain Mus domesticus-derived Sry genes are transferred onto specific inbred strains, such as C57BL/6J (B6).

Standard mouse inbred strains are a composite of two species, M. musculus and M. domesticus (reviewed in BONHOMME 1986 ). Most strains, such as B6, contain a M. musculus (MUS) Y chromosome, but a few, such as AKR/J, contain a M. domesticus (DOM) Y chromosome (BISHOP et al. 1985 ; NISHIOKA and LAMOTHE 1987 ; TUCKER et al. 1992 ). Of interest to gonadal sex determination is the finding that transfer of some DOM Y chromosomes (i.e., Sry alleles) to the B6 strain interferes with testis development (Table 1). For example, B6 XY^POS mice carrying an Sry^DOM allele from wild-derived M. d. poschiavinus mice develop ovaries and ovotestes but not normal testes (referred to as B6-Y^POS sex reversal; EICHER et al. 1982 ; EICHER and WASHBURN 2001 ). B6 XY^AKR mice carrying an Sry^DOM allele from the AKR/J inbred strain have delayed testis cord development but do not develop ovarian tissue (WASHBURN and EICHER 1983 , WASHBURN and EICHER 1989 ; EICHER and WASHBURN 1986 ; NAGAMINE et al. 1987 ). Other Sry^DOM alleles do not interfere with normal testis development when present on B6, an example being B6 XY^BUB mice (Sry allele from the BUB/BnJ inbred strain; WASHBURN et al. 2001 ). These phenomena are highly sensitive to genetic background. That is, an Sry^DOM allele that causes abnormal testicular development in B6 mice may not do so when present on another genetic background (EICHER et al. 1982 ; EICHER and WASHBURN 1986 ; NAGAMINE et al. 1987 ; BIDDLE and NISHIOKA 1988 ). For example, DBA/2J XY^POS (D2 XY^POS) and (D2 x B6)F₁ XY^POS mice develop normal testes. These findings suggest that the ability of a specific Sry allele to function correctly is dependent on other genes, designated tda (testis-determining autosomal) genes, and genetic mapping experiments have supported this hypothesis (EICHER et al. 1996 ). If the proper functioning of the human SRY gene likewise is sensitive to genetic background, this could clarify the unexplained human XY SR conditions noted above.

The simplest mechanism to explain B6-Y^POS SR is that the Sry^POS allele encodes a protein that does not interact correctly with downstream genes if they are derived from the B6 strain (COWARD et al. 1994 ). However, DNA sequence analysis of several DOM Sry alleles revealed that this hypothesis alone is inadequate because no correlation between the open reading frame (ORF) sequence and SR was found (CARLISLE et al. 1996 ; ALBRECHT and EICHER 1997 ). A second possible mechanism to explain B6-Y^POS SR is that expression of the Sry^POS allele is aberrant. Support for this hypothesis was obtained by NAGAMINE et al. 1999 , who assayed Sry expression in mouse fetal gonads from three B6 consomic strains carrying different DOM Sry alleles, Sry^FVB, Sry^AKR, and Sry^TIR. [Sequence analysis indicates that the Sry^TIR and Sry^POS ORFs are identical (ALBRECHT and EICHER 1997 )]. They found that Sry expression was highest in B6 XY^FVB fetal gonads, which develop as normal testes, and lowest in B6 XY^TIR fetal gonads, which develop as ovotestes or ovaries. Moreover, Sry expression was increased in (SWR x B6)F₁ XY^TIR fetal gonads, which develop as normal testes.

Here we report results from two experimental approaches designed to further our understanding of Sry function in B6-Y^POS SR. First, we determined Sry expression levels in fetal gonads from B6 mice carrying a Y chromosome containing both a MUS-derived Sry allele (Sry^RIII) and a DOM-derived Sry allele (Sry^POS or Sry^AKR). This approach differed from that used by NAGAMINE et al. 1999 because it directly compared Sry^DOM and Sry^MUS expression within the same gonad so that the results were independent of the number of Sry-expressing cells. We found that Sry^POS transcript levels were significantly reduced compared to Sry^RIII, whereas Sry^AKR and Sry^RIII transcript levels were equivalent. We then assayed Sry transcript levels on a (D2 x B6)F₁ genetic background previously shown to allow normal testis determination in XY^POS mice (EICHER and WASHBURN 1986 ). Unexpectedly, we found that although both Sry^DOM and Sry^MUS transcript levels increased, Sry^DOM transcript levels increased more than Sry^MUS transcript levels. Together, these data suggest that B6 XY^POS SR is caused, at least in part, by insufficient Sry expression and that one or more tda genes directly regulate Sry expression. These results confirm and extend those of NAGAMINE et al. 1999 .

Our second approach was based on the premise that if B6-Y^POS SR is caused by insufficient Sry expression, then overexpression of Sry^POS would rescue testis development in B6 mice. B6 transgenic mouse lines carrying either a chimeric Sry construct in which Sry^POS expression was regulated by MUS regulatory regions or an Sry^POS genomic DNA clone were produced. In two transgenic lines, B6 XX mice carrying either type of transgene developed testicular tissue exclusively. However, testis cord development was delayed despite normal transcriptional initiation and overexpression of Sry. These data suggest that delayed testis cord development is caused by SRY protein isoform differences that are exacerbated by insufficient Sry expression leading to ovarian tissue development in B6 XY^POS gonads. The above hypothesis is supported by the finding that the MUS Sry^B6 allele is expressed at relatively low levels (LEE and TAKETO 2001 ) without causing delayed testis cord development or SR.

In five Sry B6 transgenic lines, sex reversal was not complete. In two lines, XX transgenic mice developed as females, and in three lines XX transgenic mice developed as females, hermaphrodites, or males. However, for the three transgenes tested, the percentage of XX transgenic males increased on a (D2 x B6)F₁ genetic background. In the one case assayed, transgene expression was increased in F₁ gonads compared to B6 gonads, presumably due to the presence of an enhancer element responsive to genetic background.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Creating B6-Y^POS,Sxr and B6-Y^AKR,Sxr consomic strains:
The Sxr (sex-reversed) Y chromosome rearrangement [formally, Tp(Y)-1Ct); CATTANACH et al. 1971 ] was used to generate mice biallelic for Sry carrying a single copy of a MUS and DOM Sry allele. (The Sxr Sry allele is MUS derived and from the RIII inbred strain.) The Sxr Y chromosome carries a duplication of most of the short arm (Yp), including the Sry gene, at the distal end of the long arm (Yq; EVANS et al. 1982 ; BISHOP et al. 1988 ). This duplicated copy of Sry is adjacent to the pseudoautosomal region and can be transferred to the X chromosome by homologous recombination.

The basic strategy to create the Sry biallelic strains was to first transfer the duplicated Sxr segment from the Sxr Y chromosome to an X chromosome and then to transfer it from the X chromosome onto the Y^POS or Y^AKR chromosome. Because XX^Sxr mice are sterile, we used the T(16;X)16H translocation (T16H) to cause preferential X inactivation of the X^Sxr chromosome: If X inactivation spreads to the Sry gene, these T16H/Sxr mice will develop as females (CATTANACH et al. 1982 ; SIMPSON et al. 1984 ). T16H +/+ Eda^Ta females were mated to XY^Sxr males (Eda^Ta is ectodysplasin-A, also known as tabby). Non-tabby females (i.e., inherited T16H) that inherited Sxr on their paternally derived inactive X chromosome were mated to B6 XY^POS hermaphrodites or to B6 XY^AKR males to generate X^SxrY^POS or X^SxrY^AKR males, respectively. X^SxrY^POS and X^SxrY^AKR males were mated to B6 females, and XY^POS,Sxr and XY^AKR,Sxr male offspring, respectively, were identified (i.e., the Sxr segment was transferred from the X^Sxr chromosome to the Y^POS or Y^AKR chromosome by homologous recombination in the pseudoautosomal region). Thereafter, XY^POS,Sxr and XY^AKR,Sxr males were backcrossed to B6 females. All males used to produce fetal offspring in this study were at backcross generation N10 or greater. Hereafter, the C57BL/6J-XY^POS,Sxr and C57BL/6J-XY^AKR,Sxr strains are referred to as B6-Y^POS,RIII and B6-Y^AKR,RIII, respectively, to reflect the Sry alleles present.

Sry transgene construction:
The Sry^129-POS chimeric transgene is based on a 14.6-kb MUS-derived genomic DNA fragment from a 129 inbred strain (GUBBAY et al. 1990 ) and previously shown to sex reverse XX mice carrying it as a transgene (KOOPMAN et al. 1991 ; EICHER et al. 1995 ; Fig 1). The Sry¹²⁹ open reading frame and a portion of the 3' untranslated region were replaced by a homologous Sry^POS DNA segment as follows: The replacement fragment was PCR amplified using high-fidelity Pfu DNA polymerase (Stratagene, La Jolla, CA), genomic DNA template containing the Y^POS chromosome, and primers Sry-8312 (5'-CCATGTCAAGCGCCCCATGAATGC) and Sry-9816 (5'-AGCTGTTTGCTGTCTTTGTGCTAGCC). (Sry primers are designated by the 5' base using numbering in GenBank entry X67204.) The PCR product was gel purified and cloned into pCR-Script using the manufacturer's protocols (Stratagene). The DNA sequence of individual clones was confirmed prior to further use. The replacement fragment was inserted into the 14.6-kb Sry¹²⁹ genomic DNA clone at the unique HhaI (base pair 8312) and SpeI (base pair 9704) sites. (The Sry ORF is located between base pair 8304 and base pair 9491.) The nucleotide sequences 8304–8312 and 9491–9704 are identical in Sry^POS and Sry¹²⁹ (ALBRECHT and EICHER 1997 ).

View larger version (20K): In this window In a new window Download PPT slide   Figure 1. Diagrammatic representation of the Sry transgenes. (A) Sry129. A 14.6-kb genomic DNA clone isolated from the MUS 129 inbred strain. (B) SryPOS. A 13.5-kb genomic DNA clone isolated from a strain carrying a DOM YPOS chromosome. (C) Sry129-POS. A chimeric 14.6-kb construct in which expression of the DOM SryPOS ORF is controlled by MUS regulatory regions derived from the clone in A. The HMG box DNA-binding domain, GRC, and the transcriptional start site at 8035 are indicated. MUS-derived regions are solid and DOM-derived regions are shaded. The inverted repeats are indicated by the double arrows. Numbering is based on GenBank entry X67204 and is approximate for B and C.

The Sry^POS transgene was derived from a 13.5-kb genomic DNA clone (L961) isolated from a mouse carrying a Y^POS chromosome (GUBBAY et al. 1992 ; Fig 1). Sequence analysis indicated that this clone extends from approximately base pair 2355 to approximately base pair 15,825 relative to the 14.6-kb Sry¹²⁹ clone described above and that a 36-bp deletion was present in the glutamine repeat region downstream of the DOM stop codon. The deleted region was replaced with a wild-type Sry^POS DNA fragment as follows: The replacement fragment was PCR amplified using Pfu DNA polymerase, genomic DNA template containing the Y^POS chromosome, and primers Sry-8386 (5'-CCAGCATGCAAAATACAGAGATCAGC) and Sry-9839 (5-ATGGCATGCTGTATTGACCACAAAGC). The PCR product was digested with SphI, gel purified, and ligated into p961 (a NotI plasmid subclone of L961) digested with SphI (base pair 8389 and base pair 9831 in GenBank entry X67204). Correct orientation was confirmed by restriction digestion and sequence analysis.

Sry transgenic mice:
B6-Sry^129-POS and B6-Sry^POS transgenic mice were produced by micro-injecting the constructs described above, without the plasmid backbone, into fertilized B6 eggs using standard methods (WAGNER et al. 1981 ).

Four Sry^129-POS transgenic lines were recovered and formally designated C57BL/6J-Tg(Sry-129-POS)17Ei, ... 28Ei, ... 94Ei, and ... 121Ei, hereafter referred to as Tg17, Tg28, Tg94, and Tg121. Three Sry^POS transgenic founders were recovered and are formally designated C57BL/6J-Tg(Sry-POS)83Ei, ... 84Ei, and ... 85Ei, hereafter referred to as Tg83, Tg84, and Tg85.

Transgenic line C57BL/6JEi-Y^AKR Tg(Sry-129)2Ei (hereafter Tg2), carrying the original 14.6-kb Sry¹²⁹ Tg (from which the Sry^129-POS Tg was derived), was used as a control for some analyses (Fig 1). All XX Tg2 animals present as males at weaning (WASHBURN et al. 2001 ).

Assessment of sexual phenotype in weaning-age mice:
Animals were classified at weaning as female, male, or hermaphrodite by the appearance of the external genitalia and by the presence of yellow pigmented hairs associated with the mammary glands. These pigmented hairs are present in B6 XX females, absent in B6 XY males, and present in most B6 XY^POS hermaphrodites and in all B6 XY^POS females (EICHER and WASHBURN 2001 ).

For histological analysis, gonads were dissected and fixed in Bouin's fixative, embedded in paraffin, sectioned, and stained with hemotoxylin and eosin using standard procedures.

Fetal gonad analysis:
Fetuses were collected from overnight matings where noon on the day a vaginal plug was observed is designated day 0.5 or from timed early morning matings. For more precise staging of fetuses younger than E13.0 (E, embryonic day) the number of tail somites (ts) posterior to the hind-limb bud was determined: E10.5 is 8 ts, E11.5 is 18 ts, and E12.5 is 30 ts (HACKER et al. 1995 ). E13.0–15.5 fetuses were staged by fore-limb and hind-limb morphology (THEILER 1989 ).

To assess fetal gonad development morphologically, gonads with attached mesonephroi were dissected from E13.5–15.5 fetuses and examined in whole mount using an inverted microscope and transmitted light. This developmental stage was chosen for analysis because a small amount of ovarian tissue is easily visualized in an ovotestis and after this stage the rapid growth of testicular tissue can obscure detection of ovarian tissue (EICHER et al. 1980 ). A tissue sample was saved from each fetus for genotype determination, as described below.

Genotyping:
PCR was used to detect the presence of an Sry^MUS and/or Sry^DOM allele in genomic DNA using one of the following methods: (1) Primers Sry-8207 5'-AGATCTTGATTTTTAGTGTTC and Sry-8677 5'-GAGTACAGGTGTGCAGCTCTA were used to amplify a 470-bp DNA fragment (GUBBAY et al. 1990 ) that was digested with MboI. Three fragments are diagnostic for Sry^MUS alleles (199, 187, and 84 bp) and two are diagnostic for Sry^DOM alleles (271 and 199; EICHER et al. 1995 ) and (2) primers Sry-9431 5'-TGGTGAGCATACACCATACC and Sry-9808 5'-TTGCTGTCTTTGTGCTAGCC were used to amplify a 377-bp DNA fragment that, when digested with NlaIV, produces a 377-bp undigested fragment diagnostic for Sry^MUS alleles and two comigrating fragments diagnostic for Sry^DOM alleles (189 and 188 bp). The Y chromosome was detected by multiplex PCR using primers for the YMT2/B locus (5'-CTGGAGCTCTACAGTGATGA and 5'-CAGTTACCAATCAACACATCAC) in conjunction with primers for the autosomal myogenin gene (5'-TTACGTCCATCGTGGACAGCAT and 5'-TGGGCTGGGTGTTAGTCTTAT) as a control (CAPEL et al. 1999 ).

The Sry^129-POS transgenes were detected by multiplex PCR using the YMT/2B and myogenin primers described above in conjunction with transgene specific primers (5'-GAGGGCATGGTCAGTTGAAC and 5'-CTCAGTGTGGAATTCATCTGC; CAPEL et al. 1999 ). The Sry^POS transgene was detected by multiplex PCR as just described but using different transgene specific primers (5'-CTAATACGACTCACTATAGGGC and 5'-GCTGACTCCATGCACAGGC).

Transgene copy number:
Transgene copy number was determined by semiquantitative PCR using B6 XY^B6 transgenic genomic DNA and the Sry-9431 and Sry-9808 primers. The assay was similar to that described below for semiquantitative RT-PCR except that 20 PCR cycles were employed. The results from at least three independent DNA samples were averaged.

RT-PCR:
Paired urogenital ridges or gonad/mesonephros complexes were dissected and nongonadal and nonmesonephric tissues were removed. The mesonephros was trimmed to the length of the gonad. The gonad and mesonephros were dissected apart in some later developmental stage samples. RNA was extracted from the dissected tissues using the RNeasy mini kit (QIAGEN, Chatsworth, CA). Lysed tissue was stored at -80° in RLT buffer (QIAGEN) until processed. The RNA was DNased during isolation using an on-column protocol (QIAGEN) or after elution from the column using the DNA-free protocol (Ambion, Austin, TX). After elution in 30 µl water, 2 µl of each RNA sample was tested for DNA contamination by PCR amplification (35 cycles) using the Sry-9431 and Sry-9808 primers. Any sample contaminated with DNA was re-DNased, purified, and retested.

One-third of the RNA sample (10 µl) was reverse transcribed at 42° for 1 hr in a 20-µl reaction using the RNA PCR kit (Applied Biosystems, Foster City, CA). Parallel reactions were performed, one with reverse transcriptase (+RT) and one without (-RT). A no-template (H₂O) negative control was included in each experiment. The reverse transcription (RT) reaction (2 µl) was PCR amplified with primers specific for the Hprt gene (5'-CCTGCTGGATTACATTAAAGCACTG and 5'-GTCAAGGGCATATCCAACAACAAAC) as a positive control for the presence of intact RNA (KOOPMAN et al. 1989 ).

Semiquantitative RT-PCR was used to determine the relative expression of the Sry^MUS (Sry^RIII) vs. the Sry^DOM (Sry^POS or Sry^AKR) alleles (BERGSTROM et al. 2000 ). The reverse transcription reaction (4 µl) was amplified by PCR in the presence of [-³²P]dCTP using the Sry-9431 and Sry-9808 primers and restriction digested with NlaIV. The resulting fragments were separated on 2% agarose gels and Southern blotted using standard methods. The amount of radioactivity in each band was determined using Phosphor imaging plates and Image Gauge software (Fuji Medical Systems USA, Stamford, CT).

Sry expression levels were compared to the expression levels of Lhx1 (LIM homeobox protein 1) using a semiquantitative RT-PCR assay. Lhx1 was chosen as the control because it is expressed only in the mesonephric component of the genital ridge and expression is relatively constant during the developmental stages analyzed (BARNES et al. 1994 ; FUJII et al. 1994 ; NAGAMINE et al. 1999 ). Lhx1-specific primers were designed to amplify a 139-bp fragment that spanned a region with no NlaIV restriction sites and contained a 97-bp intron (Lhx1-1660 5'-GGCGAGGAGCTCTACATCATAG and Lhx1-1798 5'-CTTGGGAATCCGGAGATAAAC). The Lhx1 primers were combined with Sry-9431 and Sry-9808 in a multiplex PCR reaction containing [-³²P]dCTP and 2 µl of the RT reaction. The PCR reaction was digested with NlaIV, separated on 3% agarose gels, Southern blotted, and analyzed as outlined above.

The number of PCR cycles corresponding to the exponential amplification phase was determined empirically for each RT-PCR assay (data not shown). Twenty-seven cycles were used for the Sry-only assay and 29 cycles were used for the multiplex Sry/Lhx1 assays. PCR used 1.5 mM MgCl₂ and a 57° annealing temperature.

Statistical analysis:
A two-way analysis of variance (ANOVA) was used to determine if there was a significant effect of fetal age, genetic background, or interaction of these two variables on Sry expression. Analyses were performed using ln-transformed data to better meet the assumptions of ANOVA. Scheffé's F was used for post-hoc multiple comparisons when the ANOVA identified a significant effect. All effects were evaluated using = 0.05.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To assess if ovarian tissue development in B6 XY^POS mice and delayed cord development in B6 XY^AKR mice are caused by insufficient Sry expression, we developed two Sry biallelic B6 lines: One line carried a Y chromosome containing the MUS-derived Sry^RIII allele and the DOM-derived Sry^POS allele (B6-Y^POS,RIII), and the other line carried a Y chromosome containing the Sry^RIII allele and the DOM-derived Sry^AKR allele (B6-Y^RIII,AKR). We reasoned that these B6 lines would allow a direct comparison of the relative expression of two Sry alleles within the same gonad and therefore the results would be independent of the number of Sry-expressing cells. In addition, these Sry biallelic lines would allow us to determine if a single copy of a MUS-derived Sry allele corrected testis development in B6 XY^POS and B6 XY^AKR mice. Previous experiments demonstrated that the presence of a multi-copy MUS-derived Sry¹²⁹ transgene restored normal testis development in B6-Y^POS mice (EICHER et al. 1995 ).

A single copy of Sry^MUS corrects testis development in B6 XY^POS and B6 XY^AKR mice:
All B6 XY^POS,RIII and B6 XY^AKR,RIII mice presented as normal males at weaning. Moreover, gonad differentiation in both types of Sry biallelic fetuses was normal at E13.5–15.5. The 16 B6 XY^POS,RIII fetuses analyzed had two normal testes whereas the 14 B6 XY^POS control sibs had ovaries (N = 22 gonads) or ovotestes (N = 6 gonads; Fig 2). In addition, the 19 B6 XY^AKR,RIII fetuses analyzed had two normal testes whereas 10 of 11 B6 XY^AKR control sibs had testes with delayed cord differentiation. (One B6 XY^AKR fetus had normal testis cord differentiation.) We conclude that the presence of a single copy of an endogenous MUS-derived Sry allele is sufficient to rescue testis differentiation in B6 XY^POS mice and delayed testis cord differentiation in B6 XY^AKR mice. This result also provided the opportunity to perform Sry expression-level experiments using B6 XY^POS,RIII and B6 XY^AKR,RIII gonads to examine relative Sry expression in gonads destined to develop as normal testes.

View larger version (139K): In this window In a new window Download PPT slide   Figure 2. Transmitted light microscope images of E13.5–15.0 fetal B6 gonads and mesonephroi. Testis development in B6 XYPOS fetuses is rescued by a single copy of SryMUS. (A) E14.5 XX ovary. Note the lack of cords and reticular appearance. (B) E14.5 XYB6 testis with testis cords present throughout. (C) E14.5 XYPOS ovary, which looks similar to the XX ovary in A. (D) E14.5 XYPOS ovotestis. Two or three testis cords are present in the center of the gonad and are flanked by ovarian tissue at the cranial and caudal ends. (E) E13.5 XYPOS,RIII testis. Even at this earlier stage, testis cords are present throughout. (F) E14.5 XYPOS,RIII testis appears similar to the control XY testis in B. Testis development is normal in E14.5–15.0 XX Tg85 and XX Tg94 fetuses. (G) E14.5 XX Tg85 testis. (H) E15.0 XX Tg94 testis. Both gonads appear similar to the control XY testis in B with testis cords present throughout. In A–H the gonad is at the top and the mesonephros is below; cranial (anterior) is to the right and caudal (posterior) is to the left.

Sry^POS (DOM) transcript levels are reduced compared to Sry^RIII (MUS) transcript levels between E10.5 and E13.0:
Semiquantitative RT-PCR was used to determine the relative expression of Sry^POS vs. Sry^RIII in urogenital ridges dissected from E10.5–13.0 B6 XY^POS,RIII fetuses. Sry expression normally is first detectable at E10.5 (8-ts stage), peaks at E11.5 (18-ts stage), and is absent by E13.0 (KOOPMAN et al. 1990 ; HACKER et al. 1995 ; JESKE et al. 1995 ). As indicated in Fig 3, the mean ratio (±95% confidence interval) of Sry^POS:Sry^RIII was 0.59 ± 0.04 during this time, indicating that Sry^POS is expressed at a significantly lower level than Sry^RIII. Statistical analysis using ANOVA indicates that this difference is constant between E10.5 and E13.0 (P = 0.958), implying that during this time Sry^POS and Sry^RIII are temporally regulated in a similar manner. We conclude that Sry^POS is expressed at significantly reduced levels relative to Sry^RIII in B6 mice.

View larger version (43K): In this window In a new window Download PPT slide   Figure 3. Analysis of Sry expression in Sry biallelic urogenital ridges. (A) Representative radioactive semiquantitative RT-PCR results for E11.5 SryPOS,RIII and SryAKR,RIII urogenital ridges. Lanes labeled + and - represent samples with and without reverse transcriptase, respectively. Lhx1 expression also was assayed in these samples. Note that SryPOS transcripts are present at lower levels than those of SryMUS, whereas SryAKR and SryMUS are present at about equal levels. (B) Average expression of SryDOM/SryRIII in E10.5–E13.0 urogenital ridges. The average SryDOM:SryRIII ratio, number of samples analyzed (N), and standard error are indicated for each genotype.

Sry transcript level is affected by genetic background:
B6 XY^POS fetuses develop ovaries or ovotestes whereas (D2 x B6)F₁ XY^POS fetuses develop testes (EICHER et al. 1982 ). If B6 XY^POS SR is caused by insufficient Sry expression, expression of Sry should be increased in (D2 x B6)F₁ XY^POS fetal gonads. We analyzed the relative expression of Sry^POS and Sry^RIII in urogenital ridges dissected from E10.5–13.0 (D2 x B6)F₁ XY^POS,RIII fetuses. If the expression of Sry^POS and Sry^RIII were increased equivalently, the ratio of Sry^POS:Sry^RIII would remain at 0.59. However, if the expression of one allele was increased relative to the other, the ratio would change. We found that the mean ratio (±95% confidence interval) of Sry^POS:Sry^RIII was 0.74 ± 0.05 (Fig 3), indicating that Sry^POS is expressed at a significantly lower level than Sry^RIII. The ANOVA indicates that this difference is constant from E10.5 to E13.0 (P = 0.958). However, the ANOVA also indicates that the increased expression of Sry^POS on the F₁ genetic background vs. the B6 background is significant (P < 0.0004). We conclude that the expression of Sry^POS is increased relative to Sry^RIII on a hybrid genetic background.

Sry^POS expression is more sensitive than Sry^RIII to genetic background:
To determine if the expression level of one or both Sry alleles is increased on the F₁ background, the expression level of each allele and Lhx1 were compared. The analysis was conducted using E11.5 (16–20 ts) urogenital ridges because this is the time Sry normally is maximally expressed. As indicated in Fig 4, expression of both Sry^POS and Sry^RIII was increased relative to Lhx1 in 16- to 18-ts gonads from F₁ XY^POS,RIII compared to gonads from B6 XY^POS,RIII fetuses. The ANOVA indicates that the difference between the B6 and F₁ genetic backgrounds is significant (P < 0.003). This result, coupled with the finding that the ratio of Sry^POS:Sry^RIII was increased to 0.74 in F₁ fetal gonads, suggests that Sry^POS is more sensitive than Sry^RIII to genetic background. (These data also confirm that Sry^POS is expressed at lower levels than Sry^RIII.) Additionally, the data suggest that peak Sry expression occurs at an earlier developmental stage (18 ts vs. 19 ts, or 2 hr) in the F₁ background. Whether this small difference in timing is significant is unknown.

View larger version (22K): In this window In a new window Download PPT slide   Figure 4. Average Sry/Lhx1 expression level vs. developmental age in B6 XYPOS,RIII and (D2 x B6)F1 XYPOS,RIII E11.5 urogenital ridges. Expression of both SryPOS and SryRIII increases in the F1 genetic background. As expected, Sry expression peaks at 18–19 ts and then decreases. However, peak expression occurs later in B6 urogenital ridges. For B6 XYPOS,RIII, two samples were analyzed at 16 ts, three at 17 ts, five at 18 ts, three at 19 ts, and six at 20 ts. For (D2 x B6)F1 XYPOS,RIII, four samples were analyzed at 16 ts, two at 17 ts, two at 18 ts, and one at 19 ts.

Sry^AKR and Sry^RIII are expressed at equivalent levels in E10.5–13.0 gonads:
In contrast to B6 XY^POS gonads, B6 XY^AKR gonads develop as normal testes, but have delayed testis cord differentiation. The relative expression of Sry^AKR vs. Sry^RIII was determined by semiquantitative RT-PCR using RNA from E10.5–13.0 B6 XY^AKR,RIII urogenital ridges. The mean ratio (±95% confidence interval) of Sry^AKR:Sry^RIII is 1.02 ± 0.1 (Fig 3), indicating that Sry^AKR and Sry^RIII are expressed at equivalent levels. As indicated by the ANOVA, the relative expression ratio was constant throughout this time (P = 0.958). We conclude that Sry^AKR and Sry^RIII are expressed at essentially equivalent levels and are similarly regulated temporally. These results suggest that delayed testis cord development in B6 XY^AKR fetuses is not caused by insufficient or delayed Sry expression.

Sry^B6 expression is lower than Sry^AKR expression:
Recent data indicated that Sry^B6 (MUS) is expressed at lower levels than Sry^AKR (DOM; LEE and TAKETO 2001 ). We assayed the relative expression of Sry in E11.5 (18-ts stage) B6 XY^B6 and B6 XY^AKR gonads using Lhx1 as a control. The average Sry:Lhx1 ratio was 0.95 in B6 XY^AKR gonads (N = 5 gonad pairs) and 0.38 in B6 XY^B6 gonads (N = 10 single gonads). In (D2 x B6)F₁ XY^B6 gonads (N = 16 single gonads) the average Sry:Lhx1 ratio was 0.43. Thus, the expression of Sry^B6 (MUS), like that of Sry^RIII (MUS), is less sensitive to genetic background than the expression of Sry^POS (DOM). Our results confirm those of LEE and TAKETO 2001 and emphasize the fact that Sry alleles are expressed at distinct levels.

Sry^POS transcripts are present at later developmental stages than Sry^RIII:
Previous results indicated that expression of Sry persisted longer in B6 XY^TIR gonads, which develop abnormally, than in B6 XY^B6 (LEE and TAKETO 1994 ) and B6 XY^FVB (NAGAMINE et al. 1999 ) gonads, which develop normally. (Y^FVB is DOM derived.) However, in an F₁ hybrid background where XY^TIR gonads develop into normal testes, Sry^TIR expression was not prolonged. Between E10.5 and E13.0 Sry^POS is present at 60% of the Sry^RIII level in B6 XY^POS,RIII gonads. However, after E13.0, the situation is reversed and Sry^POS is present at higher levels than Sry^RIII. For example, Sry^POS, Sry^RIII, and Lhx1 transcript levels were determined by RT-PCR in seven E13.5 B6 XY^POS,RIII gonads with attached mesonephroi (three pairs and four single complexes). The average Sry^POS:Lhx1 ratio was 0.03 whereas the average Sry^RIII:Lhx1 ratio was 0.007, indicating that at this stage Sry^POS is present at about four times the level of Sry^RIII. These data suggest that expression of Sry^POS persists longer than expression of Sry^RIII in B6 XY^POS,RIII gonads.

Transgenic overexpression of Sry^POS rescues testis determination:
The comparative Sry expression results suggested that B6-Y^POS SR is caused, at least in part, by insufficient Sry^POS expression. If this hypothesis is correct, transgenic overexpression of Sry^POS in B6 XX mice should initiate normal testis determination. Two different Sry^POS transgenic constructs were employed to test this hypothesis (Fig 1). The first was a genomic DNA clone isolated from the M. d. poschiavinus Y chromosome. Analyses of three B6 transgenic lines (Tgs 83–85, Table 2) carrying this construct (Sry^POS) are presented. The second construct was derived from the original 14.6-kb Sry clone but contained the DOM Sry^POS ORF in place of the MUS Sry¹²⁹ ORF so that expression of Sry^POS was controlled by MUS-derived regulatory regions. Analyses of four B6 transgenic lines (Tgs 17, 28, 94, and 121, Table 2) carrying this construct (Sry^129-POS) are presented.

At weaning, 75% of B6 XY^POS mice present as normal females and 25% present as hermaphrodites (EICHER and WASHBURN 2001 ). In contrast, for both the Sry^POS and Sry^129-POS constructs, one transgenic line from each was identified in which all XX transgenic (XX Tg) mice presented as normal males at weaning (Sry^POS-Tg85 and Sry^129-POS-Tg94). Histological examination of testes from three XX Tg85 and seven XX Tg94 adult males demonstrated the presence of Sertoli and Leydig cells, lack of ovarian tissue, and absence of germ cells (data not shown). (The absence of germ cells is expected in XX males due to the lack of a Y chromosome and presence of two X chromosomes.)

We then examined gonads from B6 XX Tg85 and XX Tg94 E14.5–15.5 fetuses to determine if ovarian tissue was present during fetal development. All XX Tg85 (N = 24) and XX Tg94 (N = 26) gonads developed testicular tissue exclusively (Fig 2). In contrast, ovarian tissue is readily visible in all gonads from E14.5–15.5 B6 XY^POS fetuses (EICHER and WASHBURN 2001 ).

Semiquantitative RT-PCR analysis revealed that Tg85 and Tg94 were overexpressed relative to Sry^B6 in B6 XY Tg fetal gonads at the 18-ts stage (E11.5), the timepoint when Sry is normally maximally expressed: Tg85 was expressed threefold greater and Tg94 fivefold greater than the endogenous Sry^B6 allele (Fig 5). We conclude that overexpression of Sry^POS allows normal testes to develop in E14.5 B6 fetuses.

View larger version (40K): In this window In a new window Download PPT slide   Figure 5. Both Tg85 and Tg94 are overexpressed relative to the endogenous SryB6 allele in E11.5 urogenital ridges. Representative radioactive semiquantitative RT-PCR results. Lanes labeled + and - represent samples with and without reverse transcriptase, respectively. Tg85 is expressed three times and Tg94 five times greater than the endogenous SryB6 allele.

Testis cord development is delayed in XX Tg85 and XX Tg94 fetal gonads:
At E14.5–15.5 B6 XY^AKR fetal gonads are normal appearing testes but at developmental stages prior to E14.5, testis cord development is delayed relative to B6 XY^B6 gonads. We examined testis cord differentiation in E13.5 XX Tg85 and XX Tg94 fetuses to determine if testis development was normal. Similar to B6 XY^AKR, all XX Tg85 (N = 10) and XX Tg94 (N = 14) gonads had delayed testis cord development (Fig 6).

View larger version (181K): In this window In a new window Download PPT slide   Figure 6. Transmitted light microscope images of E13.5 fetal B6 gonads. Testis cord differentiation is delayed in XX Tg85 (SryPOS) and XX Tg94 (Sry129-POS) gonads. (A) XX ovary. (B) XYB6 testis. (C) XYAKR testis. Note that the caudal pole is undifferentiated and lacks testis cords (arrow). (D) XX Tg94 testis. (E) XX Tg85 testis. Both types of XX SryPOS transgenic testes appear similar to the XYAKR testis with delayed cord differentiation at the caudal pole (arrows). (F) XX Tg2 testis, which appears similar to the control XY testis in B with testis cords present throughout. This result indicates that the 14.6-kb transgenic construct, which shares regulatory regions with Tg94, contains the regulatory elements necessary for normal testis differentiation. In A–F the gonad is at the top and the mesonephros is below; cranial (anterior) is to the right and caudal (posterior) is to the left.

Because delayed testis cord development could be caused by delayed initiation of transgene expression, we assayed transgene expression in urogenital ridges from fetuses at E10.5, the time when endogenous Sry expression is first initiated (KOOPMAN et al. 1990 ; HACKER et al. 1995 ; JESKE et al. 1995 ). Tg94 transcripts were detected in 7-ts-stage XY Tg fetuses whereas endogenous Sry transcripts were not yet detectable (three pairs of urogenital ridges, 35 PCR cycles). Tg94 expression also was detected in 9-ts XX Tg and 11-ts XY Tg fetuses (one pair of urogenital ridges each). Expression of Tg94 was clearly higher than in the endogenous Sry^B6 allele in the 11-ts XY Tg sample. Tg85 expression was detected in 9-ts-stage fetuses (two pairs of XX Tg and one pair of XY Tg urogenital ridges) and transgenic Sry expression was clearly higher than in the endogenous Sry^B6 in the XY Tg sample. We conclude that testis cord development is delayed in XX Tg85 and XX Tg94 fetal gonads despite normal transcriptional initiation and overexpression of the transgenes.

Because the Sry Tg constructs might be missing regulatory elements necessary for the initiation of normal, nondelayed testis cord development, we examined testis cord development in E13.5 B6 XX Tg fetuses from an Sry transgenic line, Tg2, carrying an intact 14.6-kb Sry¹²⁹ (MUS) Tg. As shown in Fig 6, in contrast to the delayed testis cord development observed in E13.5 B6 XX Tg85 (DOM), B6 XX Tg94 (DOM), and B6 XY^AKR fetuses, testis cord development in E13.5 B6 XX Tg2 (MUS) and B6 XY^AKR Tg2 fetuses was complete (N = 12 gonads). Because the Tg2 and Tg94 constructs contain the same MUS-derived regulatory regions, we conclude that delayed testis cord development in B6 XX Tg94 fetuses is not caused by the absence of a critical regulatory region(s).

External sexual phenotype of XX Tg mice and transgene expression level are sensitive to genetic background:
Of the seven Sry transgenes analyzed, Tg85 and Tg94 were the only ones in which 100% of the B6 XX Tg offspring were completely sex reversed (Table 2). In contrast, at weaning 56% of B6 XX Tg17 mice, 9% of B6 XX Tg28 mice, and 18% of B6 XX Tg121 mice presented as male. No B6 XX Tg83 (N = 84) or B6 XX Tg84 (N = 71) mice presented as males.

Because the XX Tg females are fertile, we intercrossed hemizygotes from the Tg28, Tg83, and Tg84 lines to determine if these transgenes caused XX SR when homozygous. Insertion of the transgene created recessive lethal mutations in the Tg28 and Tg83 lines (as suggested by underrepresentation of transgenic offspring in the intercross) so that the phenotype of Tg homozygotes could not be examined. From Tg84 intercrosses, two XX Tg SR males were present among the 37 XX Tg offspring. Because known XX Tg84/+ mice are not sex reversed, we conclude that two copies of Tg84 can cause XX sex reversal. The homozygous phenotypes for Tg17 and Tg121 were not examined because B6 XX Tg heterozygotes are sometimes sex reversed.

Because B6-Y^POS SR is highly sensitive to genetic background, we produced F₁ hybrid Tg mice by mating B6 Tg carriers to D2 and C3H/HeSnJ (C3H) mice and examined the external sexual phenotype of XX Tg mice at weaning (Table 2). In the three transgenic lines tested (both Sry^POS and Sry^129-POS), the phenotype of XX Tg mice was modulated by genetic background. For example, at one extreme, 81% (13/16) of the (C3H or D2 x B6)F₁ XX Tg83 mice presented as males whereas all (N = 84) B6 XX Tg83 mice presented as females. For Tg84, different F₁ hybrid backgrounds gave different results: All 23 (D2 x B6)F₁ XX Tg84 mice were female whereas 7 (C3H x B6)F₁ XX Tg84 mice were female and 11 were male. Surprisingly, none of the F₁ XX Tg84 mice were obvious hermaphrodites. We conclude that the external sexual phenotype and, by inference, testis determination of XX Tg mice is sensitive to genetic background.

Semiquantitative RT-PCR was used to determine if an increase in transgenic RNA transcript levels correlated with sex reversal in F₁ XX Tg mice. We analyzed (D2 x B6) Tg83 E11.5 gonads because Tg83 seemed to be the most sensitive to genetic background. Tg83 expression was compared to Lhx1 expression in gonads from 16- to 21-ts fetuses. As illustrated in Fig 7, initial (16- to 17-ts) expression was similar in both backgrounds. However, at the 18- to 21-ts developmental stage, Tg83 expression was increased in the F₁ background. The data presented in Fig 7 represent average expression, and not all of the XX Tg 83 gonads are destined to develop as testes. Therefore, the difference in expression between the B6 and F₁ genetic backgrounds probably is greater than represented. This idea is supported by the relatively large range of Tg83 expression obtained for these gonads (data not shown). These data suggest that expression of Tg83 is sensitive to genetic background, a finding that correlates with the external sexual phenotype observed in F₁ XX Tg83 mice.

View larger version (17K): In this window In a new window Download PPT slide   Figure 7. Expression of Tg83 is increased on a (D2 x B6)F1 genetic background. Average Sry/Lhx1 expression vs. developmental age in B6 Tg83 and (D2 x B6)F1 Tg83 E11.5 urogenital ridges. Initial (16- to 17-ts) expression was similar in both backgrounds. However, at the 18- to 21-ts developmental stage, Tg83 expression was increased on the F1 background. The data presented represent average expression, and not all XX Tg 83 gonads are destined to develop as testes. Therefore, the difference in expression between the B6 and F1 genetic backgrounds probably is greater than represented. The number of gonads analyzed for each stage is indicated.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Transfer of certain M. domesticus-derived (DOM) Y chromosomes (Sry^DOM alleles) onto specific inbred strains, such as B6, causes abnormal testis determination (EICHER et al. 1982 ; EICHER and WASHBURN 1986 ; NAGAMINE et al. 1987 ; BIDDLE and NISHIOKA 1988 ). The degree of abnormality depends on the particular Y chromosome transferred. The experiments presented were designed to elucidate the mechanism of Sry^DOM misfunction when present on the B6 genetic background.

To determine if DOM Sry alleles are expressed at different levels or in different temporal patterns from those of MUS Sry alleles, we developed two B6 mouse lines that each carry a single DOM Sry allele (POS or AKR) and a single MUS Sry allele (RIII). Gonads in B6 XY^POS,RIII and B6 XY^AKR,RIII mice are phenotypically normal testes. This finding confirms and extends results demonstrating that transgenic overexpression of a MUS Sry¹²⁹ allele rescues SR in B6-Y^POS mice (EICHER et al. 1995 ) and delayed testis cord development in B6-Y^AKR mice (data presented here). Also, the data indicate that the DOM Sry^POS and Sry^AKR alleles do not act as dominant-negative alleles. It is interesting to note that, to our knowledge, all Sry transgenic lines capable of sex reversing XX mice contain multiple insertions of the Sry gene (KOOPMAN et al. 1991 ; EICHER et al. 1995 ; WASHBURN et al. 2001 ). The reasons for this are not clear, but it may be that expression of a transgene is dependent on the chromosomal site of insertion and on the presence of critical cis-acting regulatory elements. In the case of the B6 XY^POS,RIII and B6 XY^AKR,RIII mice, a single copy of a MUS-derived Sry gene was present and able to correct testicular abnormalities. We suggest that the reason this single copy of Sry functioned normally was that it was present on a segment of the mouse Y chromosome that normally contains it.

If the misexpression hypothesis is correct, expression of the Sry^POS allele should be more "abnormal" than that of the Sry^AKR allele. This was, in fact, the case: On the B6 background, the DOM Sry^POS allele was expressed at 59% of the MUS Sry^RIII allele whereas the DOM Sry^AKR allele and the MUS Sry^RIII allele were expressed at equal levels. Moreover, if the misexpression hypothesis is correct, expression of Sry^POS would be more "normal" on a hybrid genetic background known to rescue B6-Y^POS sex reversal. This, too, was the case: The Sry^POS allele was expressed at 74% of the MUS Sry^RIII allele on a (D2 x B6)F₁ genetic background. Because relative Sry expression was measured in genital ridges destined to develop as normal gonads and independent of the number of Sry-expressing cells, we conclude that Sry expression per cell is reduced. The results, however, do not exclude the possibility that the number of Sry-expressing cells also is reduced.

The fact that the relative expression of Sry^POS/Sry^RIII and Sry^AKR/Sry^RIII was constant between E10.5 and E13.0 suggests that the temporal expression of DOM and MUS alleles is similar during this time. Therefore, it is unlikely that delayed Sry expression is responsible for either SR in B6 XY^POS gonads or delayed testis development in B6 XY^AKR gonads. These results are consistent with those of LEE and TAKETO 1994 and NAGAMINE et al. 1999 .

After E13.0, expression of Sry^POS persisted longer than expression of Sry^RIII. This result implies that Sry^POS expression is downregulated more slowly than Sry^RIII expression. However, we cannot exclude the possibility that the Sry^POS transcript is more stable than the Sry^RIII transcript. We suggest that if persistent expression is due to inefficient downregulation of Sry^POS expression, then the same regulatory elements that prevent efficient upregulation of Sry^POS expression may be identical to those that prevent efficient downregulation.

The relative expression results were confirmed by measuring expression of the individual Sry alleles against expression of a control gene (Lhx1). These data indicated that expression of both the DOM and MUS Sry alleles was increased on the hybrid genetic background, but the expression of the DOM allele was increased to a greater extent. This result suggests that the Sry^POS allele is more sensitive to genetic background than the Sry^RIII allele. It is likely, therefore, that at least one tda gene affects Sry expression and that this interaction is direct. The simplest model is that one or more tda genes is a transcription factor that controls Sry transcription by directly interacting with the Sry promoter. However, other models are possible. For example, a tda gene could interact with the Sry transcript and affect its stability or localization. Further functional studies are needed to test these models.

We found that in B6 XY^AKR,RIII fetal gonads the DOM Sry^AKR and MUS Sry^RIII alleles were expressed at equal levels. The question of whether the Sry^RIII allele initiates normal testis determination is complicated by the fact that we analyzed testis development in XX^Sxr fetuses where random X inactivation can affect the expression of the Sry^RIII allele. However, 32 of the 40 B6 XX^Sxr gonads examined between E13.25 and E14.5 were normal testes without delayed testis cord development. (The remaining 8 gonads were ovotestes.) This result suggests that in the absence of significant inactivation of the X^Sxr chromosome, the Sry^RIII expression level is sufficient to initiate normal testis development on the B6 background. Because Sry^AKR and Sry^RIII are expressed at equivalent levels yet B6 XY^AKR gonads have delayed testis cord development, delayed testis cord development cannot be attributed solely to insufficient Sry^AKR expression. Rather, delayed testis cord development probably is caused by reduced translation of the Sry^AKR transcript, by reduced stability of the SRY^AKR protein isoform, or by reduced ability of the SRY^AKR protein isoform [which is approximately half the size of the MUS SRY protein isoform (COWARD et al. 1994 )] to initiate testis development. These results are consistent with the finding that Sry^B6 (MUS) is expressed at lower levels than Sry^AKR is yet testis development in B6 XY^B6 mice is normal (LEE and TAKETO 1994 and results herein). Furthermore, these results suggest that at least one tda gene participates in the sex-determination cascade downstream of or in parallel with Sry.

Overall, the Sry expression analysis indicates that B6-Y^POS SR is caused by insufficient Sry^POS expression and that delayed testis cord development in B6 XY^AKR mice is caused by reduced efficiency of the SRY^AKR isoform. If this model is correct, then overexpression of Sry^POS in B6 mice would rescue SR but might not rescue delayed testis cord development. Two different transgenic constructs were used to test this hypothesis: an Sry^POS genomic DNA clone and a chimeric construct in which expression of the Sry^POS ORF was controlled by Sry^MUS regulatory regions (Sry^129-POS). Two B6 transgenic lines, one from each type of construct, were established in which all XX transgenic progeny developed testes. However, testis cord development was delayed in both lines despite overexpression of Sry^POS and normal transcriptional initiation from the transgenes. These results suggest that testes develop when Sry^POS is expressed at relatively high levels; however, overexpression is not sufficient to correct delayed testis cord development. The transgenic results support a model where delayed testis cord development is caused by the presence of particular DOM SRY protein isoforms that cause SR when underexpressed. The fact that (D2 x B6)F₁ XY^POS fetuses develop normal testes without evidence of delay (EICHER et al. 1996 ) indicates that delayed testis cord development requires that at least one tda gene be homozygous for the B6 allele.

To our knowledge, all Sry^DOM ORFs analyzed have a stop codon in the glutamine repeat region downstream of the HMG box, which means that SRY^DOM proteins are about half the size of SRY^MUS proteins (COWARD et al. 1994 ; CARLISLE et al. 1996 ; ALBRECHT and EICHER 1997 ). However, the predicted "half-size" SRY^DOM protein isoform alone is not sufficient to account for either sex reversal or delayed testis development when present on the B6 genetic background because some Sry^DOM alleles, such as Sry^FVB and Sry^BUB, initiate normal testis development when on the B6 background (BIDDLE and NISHIOKA 1988 ; NAGAMINE et al. 1999 ; WASHBURN et al. 2001 ). Different Sry^DOM protein isoforms that differ in the number of glutamines encoded by the third glutamine repeat cluster (GRC-3) have been identified. No correlation is found between the number of glutamines in GRC-3 and sex reversal (CARLISLE et al. 1996 ; ALBRECHT and EICHER 1997 ). However, it is possible that the number of glutamines in GRC-3 plays a role in whether a given Sry^DOM allele causes delayed testis development when on the B6 background. Sry^AKR, which causes delayed testis development, has 13 glutamines in GRC-3 while Sry^FVB and Sry^BUB, which initiate normal testis development, have 12. The situation is complicated by the fact that different Sry alleles are expressed at different levels. For example, the Sry^FVB allele is expressed at higher levels than the Sry^AKR allele is (NAGAMINE et al. 1999 ).

Not all of the transgenes produced exclusively male XX Tg progeny, and for several the percentage of male XX Tg progeny was increased on a hybrid genetic background. For the one transgene examined, the increase in male XX Tg progeny was correlated with increased expression of the Sry transgene. The results indicate that the transgenes contain a DNA element that controls Sry expression level and is sensitive to genetic background. We suggest that this element is likely to directly interact with a tda gene. Furthermore, this control element is present in the region of minimal overlap between the two types of transgenes (i.e., between 2355 bp and 14,625 bp). Future experiments are focused on identifying the Sry expression control element.

We are intrigued by the finding that all (D2 x B6)F₁ XX Tg84 mice are female whereas approximately half of the (C3H x B6)F₁ XX Tg84 mice are female and the remainder are male. This result nicely illustrates the fact that sex determination in mice is exquisitely sensitive to genetic background. We do not know if the difference between the D2 and C3H inbred strains is due to different alleles of the tda genes previously mapped or to differences in novel tda genes. Molecular identification of the tda genes will clarify this.

As noted in the Introduction, several intriguing but unexplained SR conditions are found in humans, including XY females and XY hermaphrodites who carry an apparently normal SRY gene and XY females who carry a mutated SRY gene inherited from their carrier father. We hypothesize that these human SR conditions are like B6-Y^POS SR and are caused by conditionally insufficient SRY expression. Therefore, it is possible that the human homologs of tda genes implicated in B6-Y^POS SR play a role in these and other human SR conditions.

	FOOTNOTES

¹ Present address: Genetics Program, Department of Medicine, Boston University Medical School, 715 Albany St., E325, Boston, MA 02118.

	ACKNOWLEDGMENTS

We thank members of The Jackson Laboratory Microinjection and Microchemistry services for their technical assistance. Appreciation is expressed to Jason Stockwell of The Jackson Laboratory Computational Biology Resource for statistical analysis of the data and to Robin Lovell-Badge (MRC, NIMR, London) for providing the L961 Sry^POS clone. We are grateful to Luanne L. Peters and Timothy P. O'Brien for critical reviews of the manuscript. This work was funded by National Institutes of Health research grant GM-20919 (to E.M.E.), fellowships GM-16726 (to K.H.A) and HD-08492 (to M.Y.), and by a National Cancer Institute CORE grant CA34196 (to The Jackson Laboratory).

Manuscript received October 7, 2002; Accepted for publication January 28, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALBRECHT, K. H. and E. M. EICHER, 1997  DNA Sequence analysis of Sry alleles (subgenus Mus) implicates misregulation as the cause of C57BL/6J-Y^POS sex reversal and defines the SRY functional unit. Genetics 147:1267-1277.[Abstract]

BARNES, J. D., J. L. CROSBY, C. M. JONES, C. V. WRIGHT, and B. L. HOGAN, 1994  Embryonic expression of Lim-1, the mouse homolog of Xenopus Xlim-1, suggests a role in lateral mesoderm differentiation and neurogenesis. Dev. Biol. 161:168-178.[Medline]

BERGSTROM, D. E., M. YOUNG, K. H. ALBRECHT, and E. M. EICHER, 2000  Related function of mouse SOX3, SOX9, and SRY HMG domains assayed by male sex determination. Genesis 28:111-124.[Medline]

BIDDLE, F. G. and Y. NISHIOKA, 1988  Assays of testis development of the mouse distinguish three classes of domesticus-type Y chromosome. Genome 30:870-878.[Medline]

BISHOP, C. E., P. BOURSOT, B. BARON, F. BONHOMME, and D. HATAT, 1985  Most classical Mus musculus domesticus laboratory mouse strains carry a Mus musculus musculus Y chromosome. Nature 315:70-72.[Medline]

BISHOP, C. E., A. WEITH, M. G. MATTEI, and C. ROBERTS, 1988  Molecular aspects of sex determination in mice: an alternative model for the origin of the Sxr region. Philos. Trans. R. Soc. Lond. B Biol. Sci. 322:119-124.[Medline]

BONHOMME, F., 1986  Evolutionary relationships in the genus Mus.. Curr. Top. Microbiol. Immunol. 127:19-34.[Medline]

CAPEL, B., K. H. ALBRECHT, L. L. WASHBURN, and E. M. EICHER, 1999  Migration of mesonephric cells into the mammalian gonad depends on Sry.. Mech. Dev. 84:127-131.[Medline]

CARLISLE, C., H. WINKING, D. WEICHENHAN, and C. M. NAGAMINE, 1996  Absence of correlation between Sry polymorphisms and XY sex reversal caused by the M. m. domesticus Y chromosome. Genomics 33:32-45.[Medline]

CATTANACH, B. M., C. E. POLLARD, and S. G. HAWKES, 1971  Sex-reversed mice: XX and XO males. Cytogenetics 10:318-337.[Medline]

CATTANACH, B. M., E. P. EVANS, M. D. BURTENSHAW, and J. BARLOW, 1982  Male, female and intersex development in mice of identical chromosome constitution. Nature 300:445-446.[Medline]

COWARD, P., K. NAGAI, D. CHEN, H. D. THOMAS, and C. M. NAGAMINE et al., 1994  Polymorphism of a CAG trinucleotide repeat within Sry correlates with B6.Y^DOM sex reversal. Nat. Genet. 6:245-250.[Medline]

EICHER, E. M. and L. L. WASHBURN, 1986  Genetic control of sex determination in mice. Annu. Rev. Genet. 20:327-360.[Medline]

EICHER, E. M. and L. L. WASHBURN, 2001  Does one gene determine whether a C57BL/6J-Y^POS mouse will develop as a female or as an hermaphrodite? J. Exp. Zool. 290:322-326.[Medline]

EICHER, E. M., W. G. BEAMER, L. L. WASHBURN, and W. K. WHITTEN, 1980  A cytogenetic investigation of inherited true hermaphroditism in BALB/cWt mice. Cytogenet. Cell Genet. 28:104-115.[Medline]

EICHER, E. M., L. L. WASHBURN, J. B. WHITNEY, III, and K. E. MORROW, 1982  Mus poschiavinus Y chromosome in the C57BL/6J murine genome causes sex reversal. Science 217:535-537.[Abstract/Free Full Text]

EICHER, E. M., E. P. SHOWN, and L. L. WASHBURN, 1995  Sex reversal in C57BL/6J-Y^POS mice corrected by a Sry transgene. Philos. Trans. R. Soc. Lond. B Biol. Sci. 350:263-269.[Medline]

EICHER, E. M., L. L. WASHBURN, N. J. SCHORK, B. K. LEE, and E. P. SHOWN et al., 1996  Sex-determining genes on mouse autosomes identified by linkage analysis of C57BL/6J-Y^POS sex reversal. Nat. Genet. 14:206-209.[Medline]

EVANS, E. P., M. D. BURTENSHAW, and B. M. CATTANACH, 1982  Meiotic crossing-over between the X and Y chromosomes of male mice carrying the sex-reversing (Sxr) factor. Nature 300:443-445.[Medline]

FUJII, T., J. G. PICHEL, M. TAIRA, R. TOYAMA, and I. B. DAWID et al., 1994  Expression patterns of the murine LIM c