Intragenic Sex-Chromosomal Crossovers of Xmrk Oncogene Alleles Affect Pigment Pattern Formation and the Severity of Melanoma in Xiphophorus

Corresponding author: Manfred Schartl, Physiological Chemistry I, Biocenter, University of Würzburg, Am Hubland, D-97074 Würzburg, Germany., phch1{at}biozentrum.uni-wuerzburg.de (E-mail)

Communicating editor: C. KOZAK

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The X and Y chromosomes of the platyfish (Xiphophorus maculatus) contain a region that encodes several important traits, including the determination of sex, pigment pattern formation, and predisposition to develop malignant melanoma. Several sex-chromosomal crossovers were identified in this region. As the melanoma-inducing oncogene Xmrk is the only molecularly identified constituent, its genomic organization on both sex chromosomes was analyzed in detail. Using X and Y allele-specific sequence differences a high proportion of the crossovers was found to be intragenic in the oncogene Xmrk, concentrating in the extracellular domain-encoding region. The genetic and molecular data allowed establishment of an order of loci over ~0.6 cM. It further revealed a sequence located within several kilobases of the extracellular domain-encoding region of Xmrk that regulates overexpression of the oncogene.

IN the platyfish Xiphophorus maculatus, three sex chromosomes coexist. They have been characterized as X, Y, and W. Over a large geographic range the individual populations are polymorphic for these three chromosomes. A balanced genetic system gives rise to WX, WY, and XX females as well as XY and YY males (ORZACK et al. 1980 ; KALLMAN 1984 ).

The X and Y appear to be very similar. Meiotic recombination has been observed over the entire linkage group, suggesting that the pseudo-autosomal region is still very large (KALLMAN 1975 , KALLMAN 1984 ; MORIZOT et al. 1991 ). Both X and Y carry a set of genetically well-characterized, closely linked loci.

The locus that defines the identity of the gonosome is the sex-determining locus, SD (in Xiphophorus earlier also referred to as SEX; see MORIZOT 1990 ). Nothing is known about the biological nature of SD and how it is related to the mammalian testes-determining factor, SRY. The genetic analysis of sex determination in the platyfish led to the hypothesis that a male-determining gene (M) is present on X, W, and Y, but only the Y-chromosomal allele is active due to suppression of M_X and M_W by autosomal repressors. The W is thought to contain another suppressor specific for M_Y (KALLMAN 1984 ).

The second gene in this region is the so-called pituitary (P) gene that determines the onset of sexual maturation. At present, nine alleles of P have been found. Again, the platyfish populations are polymorphic for P alleles, leading to a wide variety of phenotypes that range from very early to very late maturing animals (KALLMAN 1989 ; SCHREIBMAN et al. 1994 ).

Two pigment pattern loci of X. maculatus also locate to the SD and P-containing region. The red-yellow pattern (RY) locus [this locus has been referred to by other research groups as XANT (see MORIZOT et al. 1993 ), or Ptr (see ANDERS 1992 )] is responsible for different red, brown, orange, and yellow patterns (KALLMAN 1975 ) that are due to local high concentrations of pigment cells of the xanthophore/erythrophore lineage in the iris, on the body, and the fins. This locus appears to consist of a series of closely linked genes (KALLMAN 1975 ). The macromelanophore-determining locus, Mdl, contains the genetic information for a specific type of melanin-containing pigment cells, which are much larger than the normal melanophores, the micromelanophores. Macromelanophores compose certain bold black markings.

The macromelanophore locus has attracted special attention, because it has been shown for some of these patterns that upon certain crossings their expression is enhanced in the hybrids, giving rise to severe melanosis and even malignant melanoma (KALLMAN 1975 ; ANDERS 1991 ; SCHARTL 1995 ). The capacity to develop melanoma is contributed by a dominant oncogene, ONC-Xmrk, which is very closely linked to Mdl (WEIS and SCHARTL 1998 ). ONC-Xmrk codes for a growth factor receptor from the epidermal growth factor receptor subclass of receptor tyrosine kinases (RTKs; WITTBRODT et al. 1989 ). Its oncogenic action is suppressed in the wild platyfish by an autosomal tumor suppressor locus, R. Hybridization experiments set up so that the R-containing autosomes are substituted through introgressive breeding by R-free chromosomes from Xiphophorus fish from other populations or species release ONC-Xmrk from its control, and neoplastic growth of macromelanophores occurs.

The unit of ONC-Xmrk and certain alleles of Mdl is equivalent to the earlier defined Tu-locus sensu Anders (ANDERS 1991 ; see WEIS and SCHARTL 1998 ). Many of these Mdl-ONC-Xmrk combinations exist and each gives a peculiar phenotype of the pigment pattern and the melanoma with respect to the onset, compartment, intensity, and the extension of the pattern on the one hand and the onset, location, and severity of the cancerous disease on the other hand (SCHARTL and WELLBROCK 1998 ). It is unknown whether these features attribute either to Mdl or to the region controlling transcription of ONC-Xmrk, or to structural differences in the transcribed part that alter the biochemical properties of the oncogenic growth factor receptor.

The platyfish populations are highly polymorphic for the multiple alleles of Mdl-ONC-Xmrk, which are grouped in five classes: the spotted dorsal (Sd), the striped (Sr), spotted (Sp), nigra (Ni), and spotted belly (Sb) patterns (GORDON 1948 ). The percentage of macromelanophore-pattern-carrying fish can range from <1% to more than half of the individuals of a population (GORDON and GORDON 1957 ). ONC-Xmrk has so far been found only in conjunction with Mdl. However, it should be noted that in other Xiphophorus species Mdl alleles that are unlinked to ONC-Xmrk exist. These are not melanomagenic (WEIS and SCHARTL 1998 ).

ONC-Xmrk has arisen through a gene duplication event from its corresponding proto-oncogene, INV-Xmrk (ADAM et al. 1993 ). This gene is also located in the same region where ONC-Xmrk, Mdl, RY, P, and SD are residing (SCHARTL 1990 ). Although some suggestions about a gene order have been made (ANDERS et al. 1973 ; KALLMAN 1975 ; SCHARTL 1990 ; SOHN 1991 ), so far a fine mapping has been impossible to achieve due to the extremely low frequency of recombination between these loci. This information, however, is needed for positional cloning approaches employing chromosome walking out of the single cloned sites from this region, the Xmrk genes.

The W chromosome is very different, despite being involved in sex determination. No P factor was identified on the W chromosomes (KALLMAN 1989 ). The wild-type W is also devoid of RY and Mdl alleles (KALLMAN 1975 ) and, as it does not predispose to melanoma, obviously also of ONC-Xmrk. Whether INV-Xmrk is present is unknown at present. Only six cases of W/Y crossovers have been documented with the remaining product always being a W (KALLMAN 1984 ).

In the work described in this article, several mutants, including X/Y crossovers that involved the Mdl/ONC-Xmrk region, were analyzed for structural differences from the wild-type chromosome to understand the resulting differences in the phenotype of the macromelanophore pattern and, most importantly, the melanoma. The structural data were then also used to establish a gene order of this important region of the sex chromosome.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fish:
All fish used in this study (see Table 1) were bred and maintained in the aquarium facilities of the Biozentrum (Würzburg, Germany) under standard conditions. F₁ hybrids of X. maculatus with X. helleri were produced by artificial insemination. The melanoma phenotype was analyzed in F₁ and backcross hybrids with X. helleri (strain 17), while for RFLP linkage analysis F₁ and backcross hybrids with X. gordoni and X. couchianus (strains 13 and 14, respectively) were used. For mapping INV-Xmrk, backcross hybrids of X. variatus with X. helleri were used. The mutant and recombinant sex chromosomes arose either spontaneously or following X-ray mutagenesis (ANDERS et al. 1973 ) and were maintained in the genetic background of the original population from which the wild-type chromosome originated. All were generously made available by A. and F. Anders (Gießen, Germany), except for the Sr crossover 30⁸⁴B (strain 10) and the wild platyfish strain 4, which were kindly supplied by K. D. Kallman, and platyfish strain 3, which was obtained from the Xiphophorus Genetic Stock Center (San Marcos, TX).

DNA extraction and PCR amplifications:
Genomic DNA from pooled organs of individual fish was extracted as previously described (SCHARTL et al. 1995 ). From X. maculatus (Rio Jamapa, ONC^-, strain 3), genomic DNA for PCR was prepared from fin clips according to the protocol of ALTSCHMIED et al. 1997 . As template for amplification of the introns of Xmrk, either phage Sac1-3-1 isolated from a subgenomic library of X. maculatus (Rio Jamapa, Sd, strain 1; see Table 1; ADAM et al. 1991 ), which contains exon 2 to 22 of ONC-Xmrk, or genomic DNA of female X. maculatus (strain 1) was used. PCR amplifications were performed for 35 cycles with a denaturing step at 94° for 30 sec, annealing at 2° below T_m of the respective primer for 30 sec, and an extension at 72° for 30 to 330 sec according to the size of the expected product. In the first cycle, denaturation was for 180 sec and in the last cycle extension time was 180 sec. Allele-specific primers for the X and Y copies of ONC-Xmrk had the following sequences: Xfor 5'-CTTACGTTGAAAGCACGTGA-3', Xrev 5'-AAAGGAGGCTTCATGGAGGG-3', Yfor 5'-TTTGGTGTCTTACTTCTGTG-3', and Yrev 5'-TTCCTCCTACTTGGCTAAAC-3' (COUGHLAN et al. 1998 ). Primers flanking a 1.4-kb deletion in the carboxyterminal domain of the X allele are Ins4 (5'-GCCTCCTGGGAGGACAGCGAC-3') and Ins5 (5'-AGCGAGCCCTGCATCCCGCCG-3'). Products of different size for the X- and Y-ONC-Xmrk allele (X, 5.3 kb; Y, 3.4 kb) from the first intron were generated with primers Hg93 (5'-CTGCAGTCGTCATGGAAACC-3') and Hg96 (5'-CCTCCTGCCGAATCGTTCAG-3') using 1 µl of a mixture of 84:1 units of Taq polymerase (Gibco BRL, Eggenstein, Germany) and Pwo DNA polymerase (AGS, Heidelberg, Germany) in a Taq Extender buffer (Stratagene, Heidelberg, Germany). Differentiation between X- and Y-ONC in the promoter region is possible by using the oligonucleotides Prom2 (5'-CCGCTCCTCCGCGCAGAAAC-3') and Prom3 (5'-AATGACTGGGCAGTGCTAAGG-3'). Sequence information on the 96 primers for amplification of Xmrk introns and exons is available at the Xiphophorus internet web site (http://sprd1.mdacc.tmc.edu/skazianis/mainpage.html/). PCR products were subsequently analyzed on 0.8–1.6% agarose gels.

Single-strand conformation polymorphism (SSCP):
For SSCP analysis PCR products were labeled by incorporation of [³²P]dCTP in 10-µl reactions containing 1 µl 10x PCR buffer (100 mM Tris pH 8.85, 500 mM KCl, 15 mM MgCl₂, 1% Triton X-100, 2 mg/ml BSA), 50 ng genomic DNA, 5 pmol of each primer, 20 µM dATP, dGTP, dTTP, 2 µM dCTP, 1 µCi [³²P]dCTP (3000 Ci/mM), and 0.75 units Taq DNA polymerase. Reactions were overlaid with mineral oil and amplifications carried out as described above. Before loading, labeled PCR products were diluted 15- to 20-fold with formamide-loading dye (95% formamide, 0.05% bromophenyl blue, 0.05% xylene cyanol), denatured for 5 min at 95°, and chilled on ice. Electrophoresis was carried out under two different conditions: samples were either electrophoresed on 6% nondenaturing polyacrylamide gels containing 10% glycerol, 0.5x TBE at 1.8 W constant power overnight at room temperature or on 6% nondenaturing polyacrylamide gels without glycerol at 20 W constant power for 2.5 hr at 4°. Gels were dried on Whatman paper and autoradiographed. Exons larger than 200 bp were digested with an appropriate restriction enzyme after amplification taking 1 µl of PCR product and 1 unit of enzyme in a total volume of 10 µl. SSCP of exon 1 was carried out using the primer Hg91 (5'-GTGCTCAGCATCAGCCGCTG-3'), which is located 3' adjacent to the variable length CTG repeat in exon 1 and primer Hg92 (5'-CCTGAACTCAGTGAAACTGCAG-3'). To discriminate PCR products that are amplified from the proto-oncogenic version of Xmrk, male and female X. maculatus Rio Jamapa fish without ONC-Xmrk were analyzed in parallel.

Sequencing:
For direct sequencing, PCR fragments were gel purified with QIAEX (QIAGEN, Hilden, Germany) and sequenced using the cycle sequencing kit (Pharmacia, Freiburg, Germany). Those fragments that could not be reliably sequenced directly were cloned into pUC18 using the SureClone Ligation kit (Pharmacia), and at least four independent clones were sequenced by Sanger dideoxy sequencing with the Sequenase kit (United States Biochemical, Cleveland). Sequences are deposited in GenBank under accession nos. AF091399 and AF092692, AF092693, AF092694.

Southern analysis:
A total of 4.5 µg of genomic DNA was digested with EcoRI or BglII, separated on a 0.8% agarose gel, and blotted onto nylon membrane (Hybond N+; Amersham Buchler, Braunschweig, Germany). A 0.7-kb BamHI fragment of the first intron of Xmrk (probe b, J. ALTSCHMIED, unpublished results) and the PCR fragment amplified with the primers Hg93/Hg96 (probe c) were used as hybridization probes. The localization of INV-Xmrk was determined using a PCR product as hybridization probe, which was amplified from genomic DNA of a female from strain 1 with the primers Ex1/Jd9 (Ex1, 5'-ATGGAGTTTCTGCGCGGAGG-3'; Jd9, 5'-CAAATTTCTCCTGAACTCACAGC-3') and digested with AvaII. The 900-bp fragment was gel purified and used as probe a. For linkage analysis of INV-Xmrk with ONC-Xmrk, the cDNA probe p17-2 (WITTBRODT et al. 1989 ) encompassing the tyrosine kinase and carboxyterminal domains of the Y-chromosomal ONC-Xmrk was used. The fragments were radiolabeled by random priming according to FEINBERG and VOGELSTEIN 1983 . Hybridization was done at 42° in buffer containing 50% formamide, 5x SSC. Filters were washed at 68° with 0.1x SSC, 1% SDS, and autoradiographed.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Linkage of ONC- and INV-Xmrk:
In a previous study linkage of both ONC- and INV-Xmrk to Mdl^Sd on the X chromosome has been shown (SCHARTL 1990 ). In X. maculatus an EcoRI RFLP for the kinase domain encoding the genomic region of Xmrk allows differentiation of the X- as well as the Y-chromosomal copy of ONC-Xmrk from INV-Xmrk (5, 6.5, and 7 kb, respectively). Cosegregation of ONC- and INV-Xmrk was analyzed in backcross hybrids of X. maculatus (strains 1 and 2) with either X. couchianus or X. gordoni as the recurrent parent. In these two species no ONC-Xmrk is present, and INV-Xmrk is represented by a 10-kb EcoRI fragment in the RFLP analysis. Hence, the X. maculatus INV-Xmrk can be easily distinguished from the X. couchianus or gordoni INV-Xmrk. The segregation analysis revealed the presence of INV-Xmrk also on the Y chromosome and cosegregation of the Y allele of ONC-Xmrk and Mdl^Sr without recombination (in n = 25 backcross segregants; Table 2). The linkage of the X-chromosomal ONC- and INV-Xmrk loci to each other and of both to Mdl^Sp was established in n = 86 backcross segregants without recombination, which confirms our earlier data (Table 2). If all data of this study are combined, the overall recombination frequency between INV- and ONC-Xmrk would be 0.6% (parental, 158; recombinant, 1), which roughly equals 180 kb according to MORIZOT et al. 1991 . For those segregants that were analyzed after they had matured (and thus information on the sex was available), linkage of INV-Xmrk, ONC-Xmrk, and Mdl to SD was established (Table 2). No recombination was observed in our study between any of the Xmrk genes and either Mdl or SD.

To analyze for the presence of INV-Xmrk on W, WY-females (strain 5) that do not have an ONC-Xmrk gene were analyzed. Primers that flank a variable length triplet repeat in exon 1 of both Xmrk genes were designed (SCHARTL et al. 1998 ). Two PCR-products were obtained from 125 females. This indicated two alleles of INV-Xmrk and thus heterozygosity. Males (n = 116) of this strain gave only one PCR product. For RFLP analysis, W/Y females from strains 4 and 7 were crossed with X. couchianus. In 16 F₁ hybrids cosegregation of W and the 7.0-kb EcoRI INV-Xmrk fragment from X. maculatus was found, while 7 F₁ hybrids showed cosegregation of Y, the corresponding Mdl-ONC-Xmrk complex, and the 7.0-kb EcoRI INV-Xmrk fragment. This also confirmed the presence of an INV-Xmrk allele on W.

Structure of the ONC-Xmrk alleles of X. maculatus:
Thus far, information about the genomic structure of ONC-Xmrk was available only for the very 5'-end, the region coding for the tyrosine kinase domain and the carboxyterminus of the receptor (ADAM et al. 1991 ), while for its extracellular and transmembrane part the sequence is known only on the cDNA level (WITTBRODT et al. 1989 ). To establish the full genomic organization, first the exon/intron structure of the missing large genomic region was analyzed (Figure 1). Primers were designed from the ends of each exon and amplification of the intervening sequences was done. This was possible from all missing introns with the exception of the large first intron. The amplified introns were sequenced completely or at least from both ends.

View larger version (10K): In this window In a new window Download PPT slide   Figure 1. (A) Schematic maps of the X- and Y-chromosomal ONC-Xmrk loci. Numbers give the size of the EcoRI fragments in kilobases. E, EcoRI sites. Homologous sites are connected by lines. (B) Intron/exon arrangement and location of allele-specific landmarks. Exons are indicated by black bars, open symbols above arrows point to structural characters diagnostic for the X locus, while solid symbols are for Y-chromosomal markers. Triangles represent insertions/deletions, rectangles are identified sequence differences, diamonds indicate positions for allele-specific PCR differences, and circles are SSCP markers. Primers (arrows) and probes (hatched boxes) used for detecting allele-specific differences between X and Y are as follows: 1a,1b/2, Prom3,4/Prom2; 3/4, Hg91/Hg92; 5/6, Hg93/Hg96; 7/8, Hg65/Hg66; 9/10, Hg69/Hg70; 11/12, Hg73/Hg74; 13a/14a, Xfor/Xrev; 13b/14b, Yfor/Yrev; 15/16, Hg81/Hg82; 17/18, Hg85/Hg86; 19/20, Ins4/Ins5; a, Ex1/Jd9 PCR product, digested with AvaII; b, 0.7-kb BamHI fragment; c, Hg93/Hg96 PCR product.

Comparison of exon/intron arrangement and exon borders with that of the closely related chicken epidermal growth factor-receptor (EGF-R) gene revealed high conservation with a few exceptions. First, intron 12, which divides subdomains III and IV of the extracellular domain of the chicken EGF-R (CALLAGHAN et al. 1993 ), is absent in ONC-Xmrk and INV-Xmrk as detected by sequencing across the deduced exon 12/13 border of phage Sac1-3-1, cosmids containing different alleles (ONC, INV) of Xmrk, as well as PCR products from genomic DNA of several X. maculatus genotypes (strains 8, 11, and 12). Thus the Xmrk gene is composed of 27 (Y-ONC-Xmrk, INV-Xmrk) or 26 exons (X-ONC-Xmrk) instead of the 28 of the chicken EGF-R gene (Figure 1). To facilitate structural comparison to other RTKs, the fused exon in Xmrk is denominated 12/13. For exons 9 and 10, 11 and 12, and 16 and 17 the corresponding exon/intron borders are shifted by 1 or 3 bp.

The design of primers matching to intron ends made it possible to analyze for sequence differences between the corresponding exons of the X and the Y allele of ONC-Xmrk by SSCP analysis. The ONC-Xmrk SSCP bands could be differentiated from the corresponding INV-Xmrk exons by comparison to the pattern obtained from fish that possess only INV-Xmrk but not ONC-Xmrk. For exons 1, 15, 19, 23, and 25 informative SSCP patterns were obtained (Figure 2), recognizable by an additional band appearing only in the hemizygous male fish. Besides SSCP, strategic sequencing revealed informative base changes in exons 1 and 17. Further allele-specific landmarks for each of the ONC-Xmrk alleles, namely large size differences due to insertions/deletions, are detectable by PCR. They were found in the 5' promoter region, the first intron, and the 3' carboxyterminal domain. A multiplex PCR detects polymorphic base pairs between X-ONC and Y-ONC in intron 22 and sequences unique for Y-ONC in intron 25. In summary, 20 structural characters that are specific either for the X or the Y allele were identified (Figure 1).

View larger version (39K): In this window In a new window Download PPT slide   Figure 2. Informative SSCP for exon 23, amplified with the primers Hg81/Hg82 and digested with HpaI. (1) X. maculatus Sd, X/X female; (2) X. maculatus SdSr, X/Y male; (3) X. maculatus Sr crossover 3084B, X/X female; (4) X. maculatus DrLi, X/X female; (5) X. maculatus/X. helleri DrLi (mut) F1 hybrid female; (6) X. maculatus Sb, Y/Y male; (7, 8) X. maculatus without ONC-Xmrk, X/X female, X/Y male; (9) X. maculatus N1, Y/Y male; (10) X. maculatus Sp4, Y/Y male; (11) X. maculatus N2, Y/Y male; (12) X. maculatus Sd, X/X female.

Mutant analysis and structure of mutant ONC-Xmrk alleles:
With the availability of X- and Y-chromosomal ONC-Xmrk-specific molecular characters (see Figure 1) an analysis of sex chromosomal crossover mutants was possible. Both mutant X chromosomes (strains 9 and 10; see Table 1) that carry the Mdl^Sr from Y are the result of an X/Y crossover that did not affect the wild-type phenotype of any of the sex chromosomal loci. The Sr macromelanophore pattern is unchanged in purebred platyfish and in F₁ as well as in backcross hybrids with X. helleri. The RY-locus allele Dr remained linked to the X-chromosomal SD in both cases. In the DrSr strain the structure of ONC-Xmrk is the same as that for the Y-chromosomal wild-type allele, indicating that the crossover took place outside ONC-Xmrk. In the Sr crossover 30⁸⁴B strain, analysis of ONC-Xmrk for the allele-specific characters revealed that the 5' markers up to intron 1 are diagnostic for the Y allele and that all further downstream markers give the pattern for the X allele. This indicates that obviously an intragenic crossover of the X- and Y-chromosomal allele took place in a region between the 3'-end of intron 1 and exon 15 (Figure 3B).

View larger version (11K): In this window In a new window Download PPT slide   Figure 3. Schematic map of the ONC-Xmrk loci of the sex chromosomal crossover mutants. (a) DrLi; (b) DrLi (mut), Sr crossover 3084B, Sb; (c) Sp4, N1; and (d) N2. For symbol explanations see Figure 1. Hatched symbols represent X. variatus-specific landmarks. n.d., strategic sequencing of exon 17 was not performed for Sp4 and N2.

The Mdl^Sb-carrying Y chromosome (strain 8; see Table 1) originates most probably from fish taken from the Rio Tonto (KALLMAN 1975 ). Its linked ONC-Xmrk produces severe malignant melanoma originating from the spotted belly pattern, even in F₁ hybrids with X. helleri. In the case of the Sb chromosome the 3' portion of the Xmrk oncogene was exchanged against X-ONC sequences by a crossover event that must have taken place somewhere between the 3'-end of intron 1 and exon 15 (Figure 3B). It is assumed that this happened in the wild before the strain was collected.

The Y chromosome with the Mdl^Sp4 allele (strain 4; see Table 1) leads to malignant melanoma, comparable to the well-studied Sp¹ tumors. However, it shows a very pronounced gene-dosage effect. Whereas heterozygous nonhybrid platyfish have only scattered small spots, in the homozygotes the entire peduncle is bold black. Analysis of ONC-Xmrk indicates a combination of 5' upstream sequences of the X allele and the 3'-part of Y-ONC (Figure 3C). The gene shares a polymorphic nucleotide in exon 1 with Y-ONC. However, it shows an additional base exchange different from all other ONC alleles, except N1. This possibly represents a sequence polymorphism specific to these strains. Southern experiments revealed that the first intron is neither identical to Y-ONC as would have been expected nor to the X-ONC copy. The markers for Y-ONC were found for the 3'-end from exon 23 onward. Thus the composite structure of this ONC-Xmrk allele could not be resolved exactly.

The Mdl^N1-carrying Y chromosome (strain 6; see Table 1) was isolated from the Belize River. Its associated ONC-Xmrk is very similar to Xmrk of Sp4 (Figure 3C); it also has the nucleotide in exon 1, which is different from all other ONC alleles.

The Mdl^N2 chromosome (strain 7; see Table 1) is a spontaneous mutation of the N1 chromosome. In parental platyfish the large blotches encoded by Mdl^N1 are extended in N2 to a coverage of the whole body side by macromelanophores, and in the hybrids with X. helleri a more severe melanoma phenotype is observed. The 3'-part of the Xmrk oncogene showed all markers of the Y allele (Figure 3D). Exon 1 again shows the unusual nucleotide exchange like the parental chromosome. The 5'-flanking promoter region upstream (nt -44) of the TATA box, however, is totally different compared with all other known X and Y sequences.

The DrLi X chromosome (strain 11; see Table 1) arose obviously by a crossover of the X. maculatus X chromosome harboring Mdl^Sd-ONC-Xmrk and the X. variatus X chromosome carrying Mdl^Li-ONC-Xmrk. The RY allele Dr linked to Mdl^Sd was retained (ANDERS et al. 1973 ). The macromelanophore pattern of these fish, however, does not resemble either Sd or Li. The pigment cells are dispersed over the whole body at high density. The pattern is most similar to Sr'' (see below). Unlike Li, which only leads to melanoma in hybrids with X. helleri that are homozygous for Mdl^Li-ONC-Xmrk, fish with a DrLi X chromosome develop melanoma even in F₁ and readily in heterozygous backcross hybrids with X. helleri. In the molecular analysis the breakpoint region could be narrowed down to the 5'-part of the first intron of the Xmrk oncogene (Figure 3A).

A second mutation (DrLi mut) occurred when a male platyfish (X^DrLi/Y^Sr'') was crossed to a X. helleri female. One of the offspring did not exhibit the strongly enhanced macromelanophore expression typical for either Sr'' or DrLi hybrids, but a phenotype like F₁ of wild-type Sr. Adults of the pedigree with this new mutant chromosome are indistinguishable from wild-type Sr. Interestingly, a variable fraction of neonates of this strain develops severe melanosis and even invasive melanoma over the course of the first two months. Then all lesions gradually regress over a period of two to three months until the normal Sr phenotype is reached. What happened to the recombined Xmrk gene was a substitution of all 5'-X. variatus sequences by those of Y^Sr'' by another crossover. As in the case of the Sr crossover 30⁸⁴B and of Sb, the breakpoint region is localized between the 3'-end of the first intron and exon 15 (Figure 3B).

Sr'' is an X-ray-induced mutation of the wild-type Mdl^Sr-ONC-Xmrk carrying Y (ANDERS et al. 1973 ). The Sr pattern in these platyfish is considerably enhanced through an increase in numbers of macromelanophores that spread all over the body but are arranged following the basic reticular pattern preformed by the margins of the scale pouches. While the wild-type Sr pattern is only slightly enhanced in hybrids and does not give rise to tumors, Sr'' leads to melanosis and melanoma already in F₁ hybrids with X. helleri. The SSCP analysis of all exons, as well as sequencing, PCR, and Southern blot analysis for the size-specific landmark positions of the Xmrk oncogene of Sr'', did not reveal differences in the wild-type Y-ONC. However, the breakpoint for the DrLi mut chromosome where the 5' region of Sr'' was transferred to the X-ONC shows that the mutation responsible for the Sr'' phenotype must lie 3' to that breakpoint, because the transferred 5' region restored the normal adult wild-type phenotype of the Sr pattern.

Localization of INV-Xmrk:
The structure of the DrLi X chromosome that combines complementary portions of the X chromosomes of two different Xiphophorus species provided the chance to determine the position of the Xmrk proto-oncogene relative to ONC-Xmrk. An RFLP that differentiates between INV-Xmrk and ONC-Xmrk of X. maculatus and X. variatus, respectively, was found in BglII digestions of genomic DNA using the 0.7-kb BamHI fragment from intron 1 as a probe and rehybridization of the filter with a PCR probe (Ex1/Jd9) that spans a region 5' of the 0.7-kb BamHI fragment. This region harbors a 4.5-kb transposon-like element in INV-Xmrk of X. maculatus, which is absent in ONC-Xmrk and in INV-Xmrk of the other species (J. ALTSCHMIED and J. N. VOLFF, unpublished results). In fish that are homozygous for the DrLi X chromosome, only the INV-Xmrk band specific for X. maculatus was detected (Figure 4). This clearly indicated that INV-Xmrk maps to the part of the DrLi X chromosome that is derived from X. maculatus and thus is downstream of ONC-Xmrk.

View larger version (45K): In this window In a new window Download PPT slide   Figure 4. Southern hybridization of BglII-digested genomic DNA of wild-type and mutant Xiphophorus fish. (A) Filter probed with the 0.7-kb BamHI fragment (probe b) common to ONC and INV copies of Xmrk. (B) Filter reprobed with the Ex1/Jd9 PCR product (probe a), digested with AvaII. Different fragments for INV-Xmrk were detected in X. maculatus with both probes because of the presence of an inserted transposon-like element in the region covered by probe a (J. ALTSCHMIED and J. DUSCHL, unpublished results), which is absent in the INV copies of the other Xiphophorus species. Arrows point to the allele-specific fragments; * indicates a band that is most probably due to a partial restriction digest; arrowheads point to the X. maculatus INV-Xmrk-specific fragments. Filters were hybridized under stringent conditions (50% formamide) and washed with 0.1x SSC, 1% SDS at 68°.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Intragenic crossovers have as yet rarely been described (MARCH et al. 1993 ; KIM et al. 1994 ) because of the difficulty in detecting their products when allele-specific landmarks are scarce and the crossover does not result in an altered phenotype. Determination of the genomic structure of both X- and Y-chromosomal alleles of ONC-Xmrk from the Rio Jamapa platyfish made it possible to identify characteristic landmarks for distinguishing between the two alleles. These could then be used to analyze several mutants of the Mdl-Xmrk complex locus. Regarding the relatively low number of crossovers affecting the whole chromosomal region, the proportion of intragenic crossovers in the ONC-Xmrk gene was surprisingly high. The breakpoints all lie within the 5' portion of the gene somewhere between the end of exon 1 and the beginning of the kinase domain-encoding part. Although it was impossible to find the exact crossover points on the single nucleotide level for each mutant, it appears that despite this obvious concentration in a certain defined region at least the DrLi breakpoint is different from the DrLi (mut), 30⁸⁴B, and Sb group as well as from the Sp4, N1, and N2 group.

The finding that crossovers are concentrating in a defined region in the 5'-part of the Xmrk oncogene may point to a recombination hotspot. The analysis of the available sequence of this region, however, has not yet revealed obvious GT-rich sequences that have been proposed to be a prerequisite in DNA strand exchange protein-mediated DNA recombination (TRACY et al. 1997 ). A similar unexpectedly high recombination frequency has been described for the Duchenne muscular dystrophy gene (DMD; GRIMM et al. 1989 ). Contrary to the DMD gene, crossovers of ONC-Xmrk exhibit phenotypic effects leading to new macromelanophore patterns probably due to the new combination of regulatory elements of different alleles from both the Mdl and the Xmrk genes.

Combining the molecular data from the crossover mutants with the phenotype of the mutants and phenogenetic data from X/Y crossovers outside ONC-Xmrk allowed us to establish a fine map of this region of the sex chromosomes. The pattern of macromelanophores encoded by the Mdl locus in all intragenic crossover mutants is that of the parental chromosome that contributed the 5' portion of the respective ONC-Xmrk locus; e.g., a wild-type Sr pattern is seen in the Sr crossover 30⁸⁴B mutant. This places the sequences responsible for the Mdl phenotype 5' of ONC-Xmrk. Similarly, the RY and SD locus can be placed 3' of ONC-Xmrk. Crossover chromosomes with the 3' part of the X-chromosomal allele have the RY pattern of the wild-type X chromosome. The 3'-part of ONC-Xmrk is also diagnostic for the linked female or male SD locus, indicating location of SD also downstream of ONC-Xmrk. A possible exception is the Sb chromosome, which is a Y, but has an X-chromosomal 3'-end of ONC-Xmrk. As this chromosome was isolated from a wild population, its precursors are unknown and this exceptional structure might be reasonably explained by a double crossover.

The gene order of RY and SD with respect to ONC-Xmrk can be determined from crossovers where the whole pigmentary gene-containing part was transferred to the other sex chromosome, e.g., the RY^Dr-Mdl^Sd-Xmrk from the X to the Y chromosome (KALLMAN 1975 ). Thus the gene order is Mdl-5' ONC-Xmrk 3'-RY-SD (see Figure 5).

View larger version (5K): In this window In a new window Download PPT slide   Figure 5. Putative gene order on the sex chromosome of X. maculatus. SD, sex-determining locus; RY, locus for red-yellow pattern; Mdl, macromelanophore-determining locus; CEN, centromere; TEL, telomere. The localization of INV-Xmrk is 3' of ONC-Xmrk, but the exact position cannot yet be determined.

INV-Xmrk is also 3' of ONC-Xmrk, but exact placement in this linkage group was not possible with the currently available material. The genetic distance that can be calculated from our data would be ~0.6 cM; however, this is based on the one recombinant in all 159 fish analyzed. Analysis of other crosses will reveal more data, which finally should allow precise fine mapping of INV-Xmrk.

Referring to ANDERS et al. 1973 , it appears reasonable to suggest that the whole linkage group is oriented in a way that SD is toward the centromere and Mdl toward the telomere. The X/Y crossover rate over the whole segment between SD and Mdl is very low, usually in the range of 1% (GORDON 1937 ; KALLMAN 1975 ; K. D. KALLMAN, personal communication, and our own unpublished data). As this whole complex is located in the subtelomeric region of the sex chromosome (I. NANDA, M. SCHMIDT and M. SCHARTL, unpublished results), where recombination frequency and physical map distance are not expected to be considerably biased, it might cover ~300 kb (according to MORIZOT et al. 1991 ) or less. This should allow cloning of the whole set of genes by a chromosome walk starting from the available ONC- and INV-Xmrk genomic clones.

So far nothing is known regarding what determines the profound differences in the pathophysiological phenotype of the melanoma arising from the different Mdl-Xmrk alleles. From the fact that the Sr crossover 30⁸⁴B has the identical macromelanophore pattern phenotype in the X. maculatus genetic background and in hybrids as the wild-type Sr, it can be deduced that all the genetic elements determining the time of onset, the body compartment of melanoma appearance, and malignancy of this Mdl-Xmrk complex are located 5' of exon 15 of the oncogene. This is in agreement with earlier data suggesting that pattern information is contained within Mdl (WEIS and SCHARTL 1998 ).

The malignancy of the melanoma is highly correlated with the amount of ONC-Xmrk transcripts (WITTBRODT et al. 1989 ; ADAM et al. 1991 ; MAUELER et al. 1993 ). Therefore, the regulatory elements controlling the allele-specific overexpression of ONC-Xmrk in hybrid melanoma have to be searched in the region 5' of exon 15. This is also confirmed by the DrLi (mut) phenotype, which is—at least in adult fish—like the wild-type Sr (from Mdl^Sr-Xmrk on the Y chromosome of the Rio Jamapa platyfish, strain 1) and is, as judged from all landmarks, structurally like the Sr crossover 30⁸⁴B. It also has its crossover point somewhere between intron 1 and exon 15. Interestingly, the 5' portion of the DrLi (mut) locus stems from the Sr'' chromosome. Sr'' fish are characterized by an extremely malignant melanoma in hybrids and by an overabundance of macromelanophores in nonhybrid X. maculatus. SSCP analysis of all exons revealed no mutation in Sr''. Thus a mutation in the nontranslated portion of ONC-Xmrk that leads to an enhanced expression of Xmrk appears as the most likely cause for the Sr'' phenotype, although such a mutation was not detected with the tools used in this study. The structure of the DrLi (mut) ONC-Xmrk allele would place such a mutation 3' of its crossover point, restricting the localization of the putative regulatory element to a region downstream of this breakpoint up to exon 15.

The Mdl^DrLi-Xmrk chromosome gives its carrier a macromelanophore phenotype that is neither characteristic for the paternal Mdl^Sd-Xmrk (that contributed the 3' portion) nor for the Mdl^Li-Xmrk from which the 5' region and exon 1 is derived. The novel phenotype might then be due to the inappropriate interaction of the 5' regulatory region from X. variatus and a regulatory element located somewhere in intron 1 or in the following introns up to exon 15. Alternatively, the crossover process may have occurred within such an element and has compromised its structure and function. This information should be useful for molecularly identifying the regulatory elements of ONC-Xmrk in the future.

	ACKNOWLEDGMENTS

We thank Jean-Nicolas Volff for critically reading the manuscript and for discussions, Ute Hornung for technical assistance, Georg Schneider, Hugo Schwind, and Petra Weber for breeding of the fish, Christine Moeller for help in preparing the manuscript, and Steven Kazianis (Austin, TX) and an anonymous reviewer for many useful suggestions. Founder fish for some of our stocks were obtained from Annerose and Fritz Anders, Gießen, Klaus D. Kallman, New York, and the Xiphophorus Genetic Stock Center at the Southwest Texas State University, San Marcos, TX. This work was supported through grants to M.S. supplied by the Deutsche Forschungsgemeinschaft through Sonderforschungsbereich 465 and 165, the European Commission (CI1* CT94-0021, FAIR PL97-3796), and the Fonds der Chemischen Industrie.

Manuscript received July 23, 1998; Accepted for publication October 21, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAM, D., W. MÄUELER, and M. SCHARTL, 1991  Transcriptional activation of the melanoma inducing Xmrk oncogene in Xiphophorus. Oncogene 6:73-80[Medline].

ADAM, D., N. DIMITRIJEVIC, and M. SCHARTL, 1993  Tumor suppression in Xiphophorus by an accidentally acquired promoter. Science 259:816-819[Abstract/Free Full Text].

ALTSCHMIED, J., U. HORNUNG, I. SCHLUPP, J. GADAU, and R. KOLB et al., 1997  Isolation of DNA suitable for PCR for field and laboratory work. BioTechniques 23:228-229[Medline].

ANDERS, F., 1991  Contributions of the Gordon-Kosswig melanoma system to the present concept of neoplasia. Pigm. Cell Res. 3:7-29.

ANDERS, F., 1992  A gene alone may not make a tumor. J. Aids Res. Hum. Retrovirus. 8:833-853.

ANDERS, A., F. ANDERS and K. KLINKE, 1973 Regulation of gene expression in the Gorden-Kosswig melanoma system, pp. 33–63 in Genetics and Mutagenesis of Fish, edited by J. H. SCHROEDER. Springer-Verlag, New York.

CALLAGHAN, T., M. ANTCZAK, T. FLICKINGER, M. RAINES, and M. MYERS et al., 1993  A complete description of the EGF-receptor exon structure: implication in oncogenic activation and domain evolution. Oncogene 8:2939-2948[Medline].

COUGHLAN, T., M. SCHARTL, I. HOPE, and A. STEWART, 1998  PCR-based sex test for the platyfish Xiphophorus maculatus.. J. Fish Biol. in press.

FEINBERG, A. P. and B. VOGELSTEIN, 1983  A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132:6-13[Medline].

GORDON, M., 1937  Genetics of Platypoecilus. 3. Inheritance of sex and crossing over of the sex chromosomes in the platyfish. Genetics 22:376-392[Free Full Text].

GORDON, M., 1948 Effects of five primary genes on the site of melanomas in fishes and the influence of two color genes on their pigmentation, pp. 216–268 in The Biology of Melanomas, Vol. 4, New York Academy of Science, New York.

GORDON, H. and M. GORDON, 1957  Maintenance of polymorphism by potentially injurious genes in eight natural populations of the platyfish, Xiphophorus maculatus.. J. Genet. 55:1-44.

GRIMM, T., B. MULLER, M. DREIER, E. KIND, and T. BETTECKEN et al., 1989  Hot spot of recombination within DXS164 in the Duchenne muscular dystrophy gene. Am. J. Hum. Genet. 45:368-372[Medline].

KALLMAN, K. D., 1975 The platyfish, Xiphophorus maculatus, pp. 81–149 in Handbook of Genetics, Vol. 4, edited by R. C. KING. Plenum Press, New York.

KALLMAN, K. D., 1984 A new look at sex determination in poeciliid fishes, pp. 95–171 in Evolutionary Genetics of Fishes, edited by B. J. TURNER. Plenum Publishing, New York.

KALLMAN, K. D., 1989 Genetic control of size at maturity in Xiphophorus, pp. 163–184 in Ecology and Evolution in Livebearing Fishes (Poeciliidae), edited by G. K. MEFFEE and F. F. SNELSON. Prentice-Hall, Englewood Cliffs, NJ.

KIM, Y. S., M. H. KIM, Y. K. CHOI, C. H. KIM, and D. S. LEE, 1994  Allelic divergence in the human insulin gene provides evidence for intragenic recombination events in the non-coding regions: evidence for existence of new alleles. Mol. Gen. Genet. 245:146-151[Medline].

MARCH, R. E., W. PUTT, M. HOLLYOAKE, J. H. IVES, and J. U. LOVEGROVE et al., 1993  The classical human phosphoglucomutase (PGM1) isozyme polymorphism is generated by intragenic recombination. Proc. Natl. Acad. Sci. USA 90:10730-10733[Abstract/Free Full Text].

MÄUELER, W., A. SCHARTL, and M. SCHARTL, 1993  Different expression patterns of oncogenes and proto-oncogenes in hereditary and carcinogen-induced tumors of Xiphophorus. Int. J. Cancer 55:288-296[Medline].

MORIZOT, D. C., 1990 Use of fish gene maps to predict ancestral vertebrate genome organization, pp. 207–234 in Structure, Function, and Use in Biology and Medicine, edited by Z. I. OGITA and C. L. MARKERT. Wiley-Liss, New York.

MORIZOT, D. C., S. A. SLAUGENHAUPT, K. D. KALLMAN, and A. CHAKRAVARTI, 1991  Genetic linkage map of fishes of the genus Xiphophorus (Teleostei: Poeciliidae). Genetics 127:399-410[Abstract].

MORIZOT, D. C., J. HARLESS, R. S. NAIRN, K. D. KALLMAN and R. B. WALTER, 1993 Linkage maps of non-salmonid fishes, pp. 4318–4325 in Genetic Maps: Nonhuman Vertebrates, Vol. IV, Ed. 6, edited by STEPHEN J. O'BRIEN. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

ORZACK, S., J. J. SOHN, K. D. KALLMAN, S. A. LEVIN, and R. JOHNSTON, 1980  Maintenance of the three sex chromosome polymorphism in the platyfish, Xiphophorus maculatus.. Evolution 34:663-672.

SCHARTL, M., 1990  Homology of melanoma-inducing loci in the genus Xiphophorus. Genetics 126:1083-1091[Abstract].

SCHARTL, M., 1995  Platyfish and swordtails: a genetic system for the analysis of molecular mechanisms in tumor formation. Trends Genet. 11:185-189[Medline].

SCHARTL, M., and C. WELLBROCK, 1998 Polygenetic inheritance of melanoma in Xiphophorus, pp. 167–187 in Human Polygenic Diseases: Animal Models, edited by T. A. DRAGANI. Harwood Academic Publishers, Chur, Switzerland.

SCHARTL, M., B. WILDE, I. SCHLUPP, and J. PARZEFALL, 1995  Evolutionary origin of a parthenoform, the Amazon Molly P. formosa, on the basis of a molecular genealogy. Evolution 49:827-835.

SCHARTL, M., B. WILDE, and U. HORNUNG, 1998  Triplet repeat variability in the signal peptide sequence of the Xmrk receptor tryosine kinase gene in Xiphophorus fish. Gene in press.

SCHREIBMAN, M. P., M. SCHARTL, K. D. KALLMAN and L. MAGLIULO-CEPRIANO, 1994 Molecular approaches to study the genetic regulation of the fish reproductive system, pp. 343–351 in Perspectives in Comparative Endocrinology, edited by K. G. DAVEY, R. E. PETER and S. S. TOBE. Natl. Res. Council of Canada, Ottawa.

SOHN, J. J., 1991  Does Xmrk an interesting spot? Genetics 128:655-656[Medline].

TRACY, R. B., J. K. BAUMOHL, and S. C. KOWALCZYKOWSKI, 1997  The preference for GT-rich DNA by the yeast Rad51 protein defines a set of universal pairing sequences. Genes Dev. 11:3423-3431[Abstract/Free Full Text].

WEIS, S. and M. SCHARTL, 1998  The macromelanophore locus and the melanoma oncogene Xmrk are separate genetic entities in the genome of Xiphophorus. Genetics 149:1909-1920[Abstract/Free Full Text].

WITTBRODT, J., D. ADAM, B. MALITSCHEK, W. MÄULELER, and F. RAULF et al., 1989  Novel putative receptor tyrosine kinase encoded by the melanoma-inducing Tu locus in Xiphophorus. Nature 341:415-421[Medline].

This article has been cited by other articles:

				I. Nanda, U. Hornung, M. Kondo, M. Schmid, and M. Schartl Common Spontaneous Sex-Reversed XX males of the Medaka Oryzias latipes Genetics, January 1, 2003; 163(1): 245 - 251. [Abstract] [Full Text] [PDF]

				M. Schartl, U. Hornung, H. Gutbrod, J.-N. Volff, and J. Wittbrodt Melanoma Loss-of-Function Mutants in Xiphophorus Caused by Xmrk-Oncogene Deletion and Gene Disruption by a Transposable Element Genetics, November 1, 1999; 153(3): 1385 - 1394. [Abstract] [Full Text]