EP1220895A1

EP1220895A1 - Novel gene encoding a dna methyltransferase, dnmt3l

Info

Publication number: EP1220895A1
Application number: EP00969592A
Authority: EP
Inventors: Kai Krohn; Ulla Aapola; Hamish Walter and Eliza Hall SCOTT; Stylianos Antonarakis; Nobuyoshi Shimizu; Jun Kudoh; Pärt PETERSON
Original assignee: Finnish Immunotechnology Ltd
Current assignee: AAPOLA, ULLA; Antonarakis Stylianos; Krohn Kai; SCOTT, HAMISH; Shimizu Nobuyoshi; Finnish Immunotechnology Ltd
Priority date: 1999-10-15
Filing date: 2000-10-13
Publication date: 2002-07-10
Also published as: AU7926400A; WO2001027249A1

Abstract

The present invention relates to a novel gene, a novel protein encoded by said gene, a corresponding mRNA of the gene, a mutated form of the gene and to diagnostic and therapeutic uses of the gene or a mutated form thereof. More specifically, the present invention relates to a novel gene which is useful in the characterization and mutation search, and in the development of the diagnosis and therapy of hereditary diseases linked to locus 21q22.3 and in the development of the diagnosis and therapy of tumors. The novel gene and protein are also useful in the diagnosis and therapy of infertility and in the development of means for the control of reproduction.

Description

NOVEL GENE ENCODING A DNA METHYLRANSFERASE, DNMT3L

Field of the invention

The present invention relates to a novel gene, a novel protein encoded by said gene, a corresponding mRNA of the gene, a mutated form of the gene and to diagnostic and therapeutic uses of the gene or a mutated form thereof. More specifically, the present invention relates to a novel gene which is useful in the characterization and mutation search, and in the development of the diagnosis and therapy of hereditary diseases linked to locus 21q22.3 and in the development of the diagnosis and therapy of tumors. The novel gene and protein are also useful in the diagnosis and therapy of infertility and in the development of means for the control of reproduction. Background of the invention

Human chromosome 21 (HC21 or 21 q) is the smallest chromosome, and its long arm constitutes only about 1% of the human genome. Be- cause of its small size HC21 has been considered as a model chromosome for the Human Genome Project coordinated by the National Institute of Health in the U.S.A. HC21 is currently under intensive study because of trisomy 21 that causes Down syndrome and many other monogenic disorders, which have been mapped thereto [Antonarakis, S. E., Genomics 51 (1998) 1-16]. Estimates of the number of genes on 21 q range from 300-350 genes obtained by computational gene prediction [Gardiner, K. et al., Am. J. Hum. Genet. 63 (1998) 249, abstract] and 560-800 genes by counting expressed sequence tag (EST) sequences (Schuler et al., Science 274 (1996) 540-546] to 700-1000 genes by the chromosomal size estimation (Antonarakis, supra). However, only 112 of the HC21 genes have been cloned to date (Swiss-Prot protein databank sequences; http://www.expasv.ch/cai-bin/lists7humchr21.txt).

Gene mapping of HC21 is rapidly progressing based on transcription unit identification using exon trapping, cDNA selection and EST mapping [Cheng et al., Genomics 23 (1994) 75-84; Tassone et al., Hum. Mol. Genet. 4 (1995) 1509-1518; Yaspo et al., Hum. Mol. Genet. 4 (1995) 1291-1304; Chen et al., Genome Res. 6 (1996) 747-760; Schuler et al., supra; Guimera et al., Genomics 45 (1997) 59-67; Ohira et al., Genome Res. 7 (1997) 47-58; Dah- mane et al., Genomics 48 (1998) 12-23]. Subtelomeric region of HC21 is gene-rich [Bemardi, G., Annu. Rev. Genet. 23 (1989) 637-661 ; Gardiner et al., EMBO J. 9 (1990) 1853-1858; Tassone et al., Am. J. Hum. Genet. 51 (1992) 1251-1264; Graig and Bickmore . Nat. Genet. 7 (1994) 551], approximately 48 % of HC21 genes have been identified from distal 21q22.3 which accounts for only 15 % of the long arm length (Swiss-Prot protein databank; http://www.expasy.ch/cgi-bin/lists7humchr21.txt).

One of the largest gene families in human genome is the zinc fin- ger protein family that contains many members involved in regulation of transcription processes. The zinc finger domains have been categorized into subtypes based on their order of cysteine (C) and histidine (H) residues. In addition to the two most frequently occurring zinc finger motifs, C2C2 and C2H2, several other subtypes including LIM [C2HC5; Sanchez-Garcia, I. and Rab- bitts, T. H., Trends Genet. 10 (1994) 315-320], RING [C3HC4; Borden, K.L. and Freemont, P.S., Curr. Opin. Struct. Biol. 6 (1996) 395-401] and PHD [C4HC3; Aasland, R. et al., Trends Biochem. Sci. 20 (1995) 56-59] have been described. Zinc fingers are thought to mediate protein-DNA and protein-protein interactions. The transcription mapping has resulted in the identification of several disease- causing genes on 21q22.3, such as cystatin B (CSTB) in progressive myoclonus epilepsy (EPM1) [OMIM #254800, Pennacchio et al., Science 271 (1996) 1731-1734; Lalioti et al., Nature 386 (1997) 847-851], AIRE in autoimmune polyendocrinopathy candidiasis ectodermal dystrophy (APECED) syndrome [OMIM #240300, Nagamine et al., Nature Genetics 17 (1997) 393- 398; Finnish-German APECED Consortium, 1997], and cystathionine beta- synthase gene in homocystinuria [OMIM #236200; Gu et al., Am. J. Hum. Genet. 49 (1991 ) 406]. transmembrane serine protease 3 (TMPRSS3) in nonsyndromic hereditary deafness DFNB10 [OMIM *605316; Scott, H. et al., (2000) The American Society of Human Genetics, program nr 16, 2000 annual meeting], and COL18A1 in Knobloch syndrome [OMIM *267750; Sertie, A. L. et al., Hum. Mol. Genet. 9 (2000) 2051-2058]. However, other hereditary disorders of yet unknown defective gene, have been linked to 21q22.3, such as holoprocencephaly [OMIM *236100; Muenke, M. et al., Am. J. Hum. Genet. 57 (1995) 1074-1079], nonsyndromic hereditary deafness DFNB8 [OMIM *601072; Veske, A. et al., Hum. Mol. Genet. 5 (1996) 165-168], and bipolar affective disorder [Straub, R. E. et al., Nat. Genet. 8 (1994) 291-296].

DNA methyltransferases catalyse the transfer of a methyl group from S-adenosyl-methionine to cytosine residue within the CpG dinucleotide. CpG-rich regions, i.e. CpG islands, are often associated with the promoters of genes and, usually, they are not methylated [Bird, Nature 321 (1986) 209- 213]. DNA methylation is believed to be important in the regulation of gene expression in embryonic development, genomic imprinting, X chromosome in- activation, and the hypermethylation of CpG islands of tumor suppressor genes has been associated with cancer development [for review, see Tajima, S. and Suetake, L, J. Biochem 123 (1998) 993-999]. ]. DNA methylation of CpG islands suppresses the gene expression, since the presence of 5- methylcytosine blocks the binding of sequence-specific transcription factors [Tate and Bird, Curr. Opin. Genet. Dev. 3 (1993) 226-231] and recruits methyl- binding proteins and associated chromatin remodeling factors, e.g. histone deacetylases [Jones et al., Nature Genet. 19 (1998) 187-191].

Abnormal methylation patterns, local de novo hypermethylation and genome-wide hypomethylation, are frequently observed early in tumori- genesis [Baylin et al., Trends Genet. 16 (2000) 168-174]. Hypermethylation of tumor suppressor genes, accompanied by silencing and a growth advantage for the cell, has been shown in cases of retinoblastoma \pRb; Stirzaker et al., Cancer Res. 57 (1997) 2229-2237], renal carcinoma [VHL; Herman et al., Proc Natl Acad Sci 91 (1994) 9700-9704], melanoma [p16INK4b Ruas and Peters, Biochim. Biophys. Acta 1378 (1998) F115-F177], hematologic malignancies \p15INK4b; Herman et al., Cancer Res. 54 (1996) 722-727], colorectal calrci- noma [hMLH1; Proc. Natl. Acad. Sci. 98 (1998) 6870-6875], and APC [Hiltunen ef al., Int. J. Cancer 70 (1997) 644-648] and breast cancer [BRCA1; Dobrovic and Simpfendorfen, Cancer Res. 57 (1997) 3347-3350].

The link between DNA methylation, histone deacetylase and sequence-specific DNA binding activity, and a growth-regulatory pathway that is disrupted in most of cancer cells was established by Robertson et al. [Nat Genet. 25 (2000) 338-342]. They showed that DNMT1 co-purifies with the retinoblastoma (Rb) tumor suppressor gene product, E2F1 , and HDAC1. DNMT1 cooperates with Rb to repress transcription from promoters containing E2F- binding sites. In normal cells CpG methylation plays role in X-chromosome ac- tivation [Panning and Jaenisch, Cell 93 (1998) 305-308] and imprinting [Li et al., Nature 366 (1993) 362-365] and it has been shown to increase with age [Issa ef al., Nature Genet. 7 (1994) 536-540]. In colon cancer, age-related CpG island methylation in normal cells followed by intensification of methylation in cancer cells explains some hypermethylation events and, according to Toyota and Issa [Electrophoresis 21 (2000) 329-333], may provide a mechanistic link between aging and cancer formation. So far, four separate genes, DNMT1, 2, 3A and 3B [Bestor, T. H. et al., J. Mol. Biol. 203 H 988) 971-983; Okano, M. et al., Nucleic Acids Res. 26 (1998) 2536-2540; Yoder, J. A. and Bestor, T-H, Hum. Mol. Genet. 7 (1998) 279-284; Yoder, J. A. and Bestor, T-H, Hum. Mol. Genet. 7 (1998) 279- 284; Okano, M. et al., Nat. Genet. 19 (1998; 219-220] encoding DNA cytosine- 5-methyltrasferases have been cloned from mammalian cells. In addition, another gene DNMT2 (Okano, M. et al., Nucleic Acids Res. 26 (1998) 2536- 2540; Yoder, J. A. and Bestor, supra] has been identified but, although containing all the conserved methyltransferase motifs, it has not been shown to possess methyltransferase activity. DNMT1 is expressed constitutively in proliferating cells and acts as a maintenance methyltransferase, which methylates cytosines in hemi-methylated CpG sites after DNA replication. DNMT3A and DNMT3B are highly expressed in early embryogenesis. They are essential for embryonic development and responsible for de novo methylation seen during embryogenesis.

Short discription of the invention

We have now isolated a novel zinc finger containing gene, designated as DNA cytosine-5-methyltransferase like gene or DNMT3L on locus 21q22.3, between the previously described KIAA0653 [Ishikawa, K. et al., DNA Res. 5 (1998) 169-176] and AIRE (Nagamine et al., supra; Finnish- German APECED Consortium, supra] genes. The novel gene encodes a DNA cytosine-5-methyltransferase like protein or DNMT3L, which shows high similarity with human and mouse DNA cytosine-5-methyltransferase 3 family (DNMT3) members [Okano, M. et al., Nat Genet. 19 (1998) 219-220; Xie, S. et al. Gene 236 (1999) 87-95], the proteins involved in de novo methylation of CpG islands in genome.

The novel gene and protein are useful for further characterization, including mutation search, and for the development of the diagnosis and therapy of hereditary diseases linked to locus 21q22.3 and for the development of the diagnosis and therapy of tumors. The novel gene and protein are also useful in the diagnosis and therapy of infertility and in the development of means for the control of reproduction, especially in the development of means for contraception in males.

An object of the invention is to provide means, which are useful in a diagnostic method and a gene therapeutic method in the diagnosis and treatment of hereditary diseases linked to locus 21q22.3 of the human genome, such as, HPE1 , and in the development of the diagnosis, therapy and prophylaxis of tumors, especially testicular and ovarian tumors and tumors of thymus. Additionally, an object of the invention is to provide means, which are useful in the diagnosis of infertility, especially male infertility, and in the development of contraceptive means for males.

The present invention relates to an isolated DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or an functionally equivalent isolated DNA sequence hybridizable thereto, and to a corresponding mRNA sequence, the DNA sequence and the corresponding mRNA sequence being associated to diseases linked to locus 21q22.3 of the human genome, such as HPE1 , and/or to tumors, especially tumors of the testis, ovary tumors and thymus.

The present invention also relates to a protein comprising the amino acid sequence id. no. 2 or a fragment or variant thereof having func- tionally equivalent properties, the protein sequence being associated to diseases linked to locus 21q22.3 of the human genome, such as HPE1 , and/or or to tumors, especially tumors of the testis, ovary tumors and thymus.

The present invention further relates to a method for the diagnosis of diseases linked to locus 21q22.3, such as HPE1 , comprising the detection, in a biological specimen, of the presence or absence of a DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent DNA-sequence hybridizable thereto or a corresponding mRNA sequence, the DNA sequence and the corresponding mRNA sequence being associated with said diseases linked to locus 21q22.3, such as HPEL

The present invention further relates to a method for the diagnosis of tumors, especially those of the testis, ovary and thymus, comprising the detection, in a biological specimen, of the presence or absence of a DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent DNA-sequence hybridizable thereto, or a corresponding mRNA sequence, the DNA sequence and the corresponding mRNA sequence being associated with said tumors, especially tumors of the testis, ovary tumors and thymus.

The present invention further relates to a method for the diagnosis of infertility, especially male infertility, comprising the detection, in a biological specimen, of the presence or absence of a DNA sequence comprising the se- quence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent DNA-sequence hybridizable thereto, or the detection of a corresponding mRNA sequence.

The present invention further relates to a method for the diagnosis of diseases related to immune maturation and to regulation of immune response, comprising the detection, in a biological specimen, the presence or absence of a DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent isolated DNA- sequence hybridizable thereto, or the detection of a corresponding mRNA se- quence.

The present invention further relates to the use of the above- identified DNA-sequences and the corresponding mRNA sequences in the diagnosis and/or therapy of diseases linked to locus 21q22.3 and/or diseases related to immune maturation and to regulation of immune response and/or in the diagnosis and treatment of tumors and/or infertility.

The present invention further relates to a method for the diagnosis of diseases linked to locus 21q22.3, such as HPE1 , comprising detecting in a biological specimen the presence or the absence of a protein comprising the sequence id. no. 2 or a fragment or variant thereof having functionally equiva- lent properties, the protein being associated with said diseases.

The present invention further relates to a method for the diagnosis of tumors, especially those of the testis, ovary and thymus, comprising detecting in a biological specimen the presence or the absence of a protein comprising the sequence id. no. 2 or a fragment or variant thereof having function- ally equivalent properties, the protein being associated with said tumors.

The present invention further relates to the use of the above- identified protein or a fragment or variant thereof having functionally equivalent properties in the diagnosis of diseases linked to locus 21q22.3, such as HPE1 , and/or tumors, such as those of the testis, ovary and thymys. The present invention further relates to the use of the above- identified DNA sequences in gene therapy or for the preparation of a pharmaceutical composition useful in a gene therapy method of diseases linked to locus 21q22.3, such as HPE1 ,

The present invention further relates to a method for the treatment of diseases linked to locus 21 q22.3, such as HPE1 , comprising administering to a mammalian, including humans, a pharmaceutical composition containing a DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent DNA-sequence hybridizable thereto.

The present invention further relates to a method for the treatment and/or prevention of tumors, especially those of the testis, ovary and thymus, comprising administering to a mammalian, including humans, a pharmaceutical composition containing a DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent DNA- sequence hybridizable thereto. The present invention further relates to a method for the male contraception comprising administering to a mammalian, including human male a pharmaceutical composition containing a mutated form of a DNA sequence having the sequence id. no. 1 or a functional fragment or variant of said mutated form, or a functionally equivalent DNA-sequence hybridizable to said mutated form.

For the present purposes, the term variant includes any variants or mutated forms of the DNA sequence, or the corresponding mRNA, if applicable, including insertion, deletion, dimeric and like mutants.

Brief description of the drawings Figure 1 shows (A) the localization of the DNMT3L gene on chromosome 21q22.3. Idiogram of the human chromosome 21 (top). The positions of DNA markers between D21S1460 and D21S171 are shown (middle). The positions of the genes located in the genomic region are indicated by arrows. DNMT3L maps centromeric to AIRE and telomeric to KIAA0653 gene (bottom); (B) a schematic overview of gene modeling results, genomic structure of DNMT3L and PCR-cloning strategy of DNMT3L. Exons predicted by Genscan program are shown as white boxes (top). The twelve exons of the DNMT3L are shown in the middle. Trapped exons are indicated as arrows, and RT-PCR and RACE products are indicated as lines (bottom). Figure 2 shows the similarity of DNMT3L with DNMT3A and

DNMT3B proteins. Cysteine-rich region of DNMT3L is underlined, and structural domains I, IV and VI of the DNMT3 family are boxed. Proteins were aligned by CLUSTALW and printed by using BOXSHADE. Identical amino acids are shown in black and conserved amino acids in grey. Figure 3 shows the alignment of ADD and PHD zinc finger motifs by ClustalW. Conserved cysteine residues involved in zinc finger formation are shown by black boxes. In the consensus sequence, the lower case represents residues that are predominant and # indicates highly conserved hydrophobic- ity. Amino acids identical to DNMT3L are shown by capital letters. ATRX mutations according to Gibbons, R. J., et al. [Nat. Genet. 17 (1997) 146-148] are shown below the alignment. Accession numbers of the sequences shown are DNMT3L (AF194032), H. sapiens DNMT3A (AAD33084), M. musculus Dnmt3a (AAC40177), H. sapiens DNMT3B (AAD53063), M. musculus Dnmt3b (AAC40178), D. reήo Dnmt3 (AAD32631), H. sapiens ATRX (AAC51657), M. musculus Atrx (AAC08741), H. sapiens Mi-2 (CAA60384), D. melanogaster Mi-2 (AAD17276), H. sapiens KIAA1113 (BAA83065), H. sapiens TIF1a (AAD17258) and H. sapiens AIRE (BAA23990).

Figure 4 shows the predicted three-dimensional structure of the ADD domain. On top is the first half of the motif. Zn-binding cysteines are in yellow and the residues corresponding to ATRX mutation sites are in magneta (P66 top and F68 below). Zn atoms are in red. On the bottom is the structure of the PHD type motif. Zn-binding cysteines are in yellow and the cysteine residues corresponding to ATRX mutation sites are in green. ATRX mutation positions are in magneta, I98 left, G127 top, and S124 bottom.

Figure 5 shows the expression pattern of DNMT3L mRNA ana- lyzed by PCR. (A) The 261-bp of DNMT3L cDNA fragment containing exons 1 and 2 were amplified from 27 different cDNAs in MTC panels (Clontech) using a primer pair 1 F/2R. Lane 1 is a negative control without template cDNA. Lanes 2-28, 29-32, and 33-36 are PCR products using 0.2 ng, 20 pg, and 2 pg of each of template MTC panel cDNA, respectively. (B) Control PCR of the 983-bp G3PDH cDNA using 2 pg of each of the template MTC panel cDNA (lanes 2-289) or no template (lane 1).

Figure 6 depicts a Western blot analysis and the subcellular localization of DNMT3L. (A) Western blot analysis of pcDNMT3L and pcDNA extracts with a monoclonal antibody against a myc tag. The arrow indicates the expressed DNMT3L protein. The protein markers are in kDa. (B) subcellular location of the transiently transfected DNMT3L cDNA into Cos-1 cells by immunofluorescence with an anti-myc antibody and (C) the DAPI staining of the same cells.

Fig. 7 shows the genomic sequence of DNMT 3L (7A-H) Detailed description of the invention

The present invention is based on studies aiming for the identification and characterization of further genes on human chromosome 21 and especially on 21 q22.3. For the isolation of the DNMT3L cDNA and gene da- tabase analysis of the 25 kb of genomic sequence of 21 q22.3 clone 28B11 using analysis program BLAST was performed. This revealed six previously trapped exons (HMC34E06/HMC43H08, HC21 Exc32,

HMC34A04/HC21 Exc31 , mdl24a08, mdl24b12 and HMC34D02; Chen et al., supra; Kudoh et al., supra). To further analyse this genomic locus, the gene modeling and exon predicting programs Genscan, Grail II and NIX were used. Both Genscan and NIX predicted a novel gene that overlapped with the trapped exons and covered more than 13 kb of genomic DNA (Figure 1 B). The RACE and RT-PCR screening of thymus and placenta cDNA libraries and testis and fetal liver mRNAs was performed using the primers specific to com- puter predicted and trapped exons, and several independent overlapping DNMT3L cDNA fragments were isolated and sequenced (Figure 1 B). The analysis of the PCR products revealed a sequence of 1705 bp that contained an open reading frame, starting with methionine codon at nucleotide 485 and terminated by TAA stop codon at nucleotide 1646 thereby coding a putative protein of 387 amino acids. The gene is divided into 12 exons over 16 kb of genomic sequence and coincided almost exactly with the computer predicted structure. The differences in the isolated DNMT3L from the gene modeled by in silico approach were the lack of one exon following the exon 1 , the presence of an additional 3'- terminal exon, and shorter exons 5 and 1 1. The translation initiation codon ATG at the nucleotide 485 of the cDNA is located in the beginning of the exon 2 leaving the first 477 bp exon most likely as a non- coding 5' untranslated region (UTR). This reading frame has also a stop codon TGA upstream in exon 1. However, there is, in addition, a short open reading frame of 52 amino acids that may start earlier from alternative ATG codon in exon 1 and is terminated by TGA codon in exon 2. Although both ATG codons have almost complete flanking Kozak consensus sequence (Kozak, M., supra) the presence of CC dinucleotide at -3 and -4 positions of downstream ATG (vs. TT at upstream ATG) has been shown to be more favorable for translation initiation signal [lida, Y. and Masuda, T., Nucleic Acids Res. 17 (1996) 3313- 3316]. At the 3'-UTR, a polyadenylation signal (TATAAA) sequence was present at 61 bp upstream from the cDNA end. All DNMT3L exons follow the rule of canonical splice sites and have a range in their size from 45 bp to 477 bp whereas the length of the introns is variable from 82 bp (intron IV) to 2918 bp (intron VIII). One splice variant was also detected in the RT-PCR amplified products. The sequencing of one fragment amplified with primers 6F and 9R revealed an alternatively spliced transcript lacking exon 8 of the cDNA. This changes the reading frame in exon 9 resulting in 21 unrelated amino acid residues followed by stop codon. Another splicing event was observed in human EST clones (see below) where the codon of serine residue is spliced out using another 5' splice site at exon 12. The calculated relative mass of the predicted DNMT3L protein is

43.7 kDa with a theoretical pi of 5.6. There is no strong signal peptide sequence or a hydrophobic transmembrane-like sequence, suggesting that the protein is intracellular. Although PSORT II program predicted the protein location into cytoplasm rather than to nucleus (61 % vs. 13%), the protein se- quence features a potential nuclear localization signal (NLS), RRRK at amino acid 156-159. In its N-terminal part DNMT3L contains a highly cysteine-rich (CR) region spanning 92 amino acids (aa 53-145). The protein has also several potential phosphorylation and myristylation sites.

Search against translated EST databank with TBLASTN identified several mouse ESTs showing high similarity to DNMT3L sequence. The mouse EST clones AA895770 and AA413561 had 65% identity with amino acids 12 to 76 and 73% identity with amino acids 58 to 211 , respectively, on N-terminal part. The third mouse EST (AA919800) showed 57% identity with amino acids 299 to 380 on C-terminal region of the DNMT3L. The sequencing of the corresponding I.M.A.G.E. clones (Ids 0778424, 1295481 and 1295738) did not add any further data to that available in Genbank. The overlapping regions of the ESTs and strong identity with DNMT3L on amino acid level, however, suggested that the three fragments most likely belong to the mouse homologue of DNMT3L cDNA covering altogether 289 of 378 residues of the human protein. Three human ESTs (AW014181 , AI335740 and AI032151 ) were found from the sequence databanks showing identity with 3'-prime end of the DNMT3L.

Interestingly, the protein similarity search also showed that the amino acids 35 to 199 on N-terminal part shared a 43-45% identity with mouse and human cytosine-5-methyltransferase 3 gene family members DNMT3A (human: Aad33084, mouse: AAC40177) and DNMT3B (human: AAD53063, mouse: AAD40178; Xie, S. et al., supra; Okano et al., Nat Genet. 19 (1998) 219-220; Figure 2), having the highest homology rate within CR region. The similarity between DNMT3L and DNMT3 proteins continued further with slightly lower identity, 31-35%, from amino acid 216 to the C-terminal end of the DNMT3L. The DNMT3L protein did not share full identity with the invariant amino acid residues in the five highly conserved motifs [I, IV, VI, IX, X; Lauster et al., J. Mol. Biol. 206 (1989) 305-312; Kumar et al., supra] in the catalytic domain of DNMT family proteins, though, there was similarity in the conserved motifs I, IV and VI. The CR region of the DNMT3/Dnmt3 proteins and DNMT3L showed significant similarity to ATRX, a member of the SNF2 protein family with ATPase and helicase domains (Figure 3).

In order to further characterize the CR region on DNMT3L, the protein and translated EST homologies with DNMT3L residues 48 - 149 containing CR region were searched. Several proteins belonging to human and mouse Dnmt3 and ATRX family and related ESTs originating from other organisms such as D. melanogaster (AA941484) and Danio rerio (zebrafish; AI584618, AI436984) were detected with good statistical support (E values, expected frequency of false positives, less than 4e-05). Although a putative ATRX homologue from Arabidopsis thaliana (AAC14034) was found, no sta- tistically supported matching proteins were found from yeast and Caenorhab- ditis elegans. The homology comparison revealed the conserved nature of cysteines and histidines in CR region and suggested the presence of a novel type of zinc-finger motif in these proteins (Figure 3). Thus, a name ADD type zinc finger' (for ATRX, DNMT3, DNMT3L) to this 100 amino acid zinc finger domain is proposed. The ADD domain consists of two subparts, 1 ) a C2C2 type zinc finger and 2) closely located domain of imperfect PHD zinc finger where histidine within C4HC3 motif is replaced by cysteine (C4C4). The spacing between the two subdomains was conserved containing identical and conserved amino acids. As a distinct feature, the C2C2 type subdomain contains a highly conserved histidine and proline residues. To study whether the two subdomains of ADD zinc finger might occur independently, further searches with separate subparts were performed. Both motifs gave the same top matching sequences as with the full structure without new similarities except for the PHD zinc finger subset. This PHD subset has been found in several nuclear proteins including Mi-2 [Seeling, H. P. et al., Arthritis Rheum. 38 (1995) 1389-1399], TIF1 [Venturini, L. et al., Oncogene 18 (1999) 1209-1217] and AIRE (Figure 3). Thus, the homology searches suggested that the two motifs are more likely occurring together than independently, probably forming a novel type zinc finger motif.

The three dimensional structure of the ADD domain was modeled based on the LIM zinc finger structure (Konrat, R. et al., supra) (Figure 4.) For modeling the N-terminal C2C2 motif, one residue was deleted from the loop between the zinc binding cysteines in the LIM finger. All four cysteines were conserved. In the latter imperfect PHD finger motif four changes to the polypeptide chain length, as compared to the LIM finger, were needed. A de- letion of eight residues was introduced into the loop in the first zinc finger. Two residue insertions were added between the histidine and cysteine on the latter half of the first zinc finger and another two residues were added between the two fingers. Three residues were added into the loop of the second finger. The zinc binding residues were conserved except for histamine, which was substi- tuted by a cysteine. The distance between the two sub-domains of the ADD structure was 19 residues and too long to be accurately predicted. The modeled structure had good stereochemistry according to Procheck program analysis.

Based on the genomic sequence data the DNMT3L is located on 21q22.3 between the markers D21 S25 and D21 S146 and its orientation of expression is from telomere to centromere (Figure 1A). The first exon of the gene is 24 kb apart from the first exon of AIRE, a defective gene in an autosomal recessive disorder called APECED (Nagamine, K. et a/., supra; Finnish- German APECED Consortium, supra). The genomic region between the DNMT3L and AIRE is enriched with repeat sequences covering in total 55% of this sequence. A mouse cosmid sequence (AF073797) containing also the 5' prime region of the corresponding mouse DNMT3L. According to the mouse EST homologies and NIX program analysis, the exon structure of the first six exons of mouse DNMT3L highly resembles the human counterpart. The only difference is a shorter exon 2, whereas the putative protein is started from the initiation codon ATG in exon 1. Interestingly, the genomic region between mouse DNMT3L and AIRE genes is much shorter than on human chromosome, ca 6.5 kb, containing less repeat sequences (21 %). The search of the human and mouse intermediate genomic sequences for computer predicted exons and EST homologies did not reveal any apparent transcripts in this region, except for one trapped exon (HC21 EXc143, D86108) located 5.2 kb, te- lomeric to DNMT3L in human sequence, however, without open reading frame. On the other flanking side, approximately 5.4 kb centromeric to DNMT3L locates a gene, called KIAA0653 [Ishikawa, K. et al., DNA Res. 5 (1998) 169-176]. The KIAA0653 belongs to the immunoglobulin super-family and the coded polypeptide shares similarity with butyrophilin protein, a membrane associated receptor expressed in mammary gland [Banghart, L. R. et al., J. Biol. Chem. 273 (1998) 4171 -4179] and with CD80 and CD86, central cosignaling molecules in T-cell stimulation pathway (Slavik, J. M. er a/., Immunol. Res. 19 (1998) 1-24]. Northern blots with 23 adult and four fetal tissues were used to identify the expression pattern of the DNMT3L. No detectable expression was found in Northern blots even after a prolonged long exposure using the 5'-end fragment of the cDNA containing the two first exons (data not shown). This suggests a low level of expression of the DNMT3L gene. The expression of DNMT3L was further studied by RT-PCR using 0.2 ng of each of cDNA from twenty-seven human tissues as templates; a specific 261 bp 1 F/2R PCT fragment was amplified in some of the tissues studied, i.e. ovary, testis, thymus, and fetal thymus (Figure 5). When a decreased amount of template cDNAs (20 pg and 2 pg each) from the positive tissues (testis, ovary, thymus and fetal thymus), only testis cDNA gave the 1 F/2R PCT fragment (Figure 5). These results indicate that the amount of DNMT3L mRNA in testis is at least 100-fold more than in other tissues. However, several other tissues should express DNMT3L mRNA at the level lower than that in ovary, thymys and fetal thymus. because the central and 3' region of the DNMT3L cDNA were amplified by PCR using increased an amount of mRNA or cDNA from other tissues, such as fetal liver or placenta.

To determine the subcellular location of the DNMT3L, cDNA was cloned into a eukaryotic expression vector under CMV promoter and trans- fected into cos-1 cells. In immunofiuorescence with the anti-myc tag antibody the expression of DNMT3L was not found to locate to the specific cellular structures as the staining was observed diffusely distributed throughout the cytoplasm and cell nucleus (Figure 6). A western blot analysis showed that the anti-myc antibody recognised a protein with a molecular mass of 45 kDa (Figure 6) in agreement with the predicted molecular mass (43.7 kDa). The most distinct domain in DNMT3L is a novel type combined zinc finger motif that we here call ADD zinc finger. According to homology modeling, the ADD domain forms three typical zinc fingers where three zinc atoms are bound to twelve cysteines. In addition to DNMT3/Dnmt3 family members, the ADD domain is an important structural feature of the ATRX protein [Gibbons, R. J. et al, Nat. Genet. 17 (1997) 146-148]. ADD zinc finger domain consisting of a C2C2-motif and an imperfect C4C4 PHD zinc finger- motif. PHD finger is a zinc-finger like C4HC3-motif found in transcriptional co- factors and in many nuclear proteins. It has been suggested that PHD zinc fingers are related to chromatin-mediated transcriptional regulation through DNA or protein interactions (Aasland et al., supra). Several missense mutations in ATRX gene ADD domain have been previously shown to cause ATR-X syndrome, a X-linked alpha- thalassemia/meπtal retardation disorder (Gibbons, R. J. et al, supra). The mutations affecting the conserved cysteines destroy the metal binding: three of the nine missense mutations described in ATR-X change the critical cysteine to another amino acid and one missense mutation introduces a new cysteine which in all cases disrupt the zinc finger structure of the domain (Gibbons, R. J. et al., supra; Figure 4). Mutations of Zn coordinating residues are also known to cause other diseases such as X-linked agammaglobunemia due to impaired metal binding and structural alterations [Vihinen, M. et al., FEBS Lett. 413 (1997) 205-210]. Other ATRX mutations in the ADD domain may affect the organisation of the zinc finger or affect the function of the motif.

DNMT3L is likely to be involved in mechanisms of gene expression or genome integrity. Interestingly, within the imperfect PHD zinc finger, DNMT3L is also similar to its neighbouring gene product AIRE, which has two PHD finger domains. The 5' regions of DNMT3L and AIRE gene are 24 kb apart but may share some regulatory regions and can be expressed coordi- nately on chromosome. DNMT3L appears to be expressed predominantly in the gonads and thymus. Similarly, the expression of AIRE is almost undetect- able by Northern blotting and is expressed in a rare set of cells in thymic me- dulla (Heino er a/., Biochem. Biophys. Res. Com. 257 (1999) 821-825]. Given the structural similarities, it is possible that the two proteins might be involved in common cellular mechanisms. The function of AIRE remains unknown but is most likely a transcriptional regulator involved in the regulation of thymic gene expression (Nagamine et al., supra, Peterson et al., Immunology Today 19 (1998) 384-386]. According to the diagnostic methods of the invention, the presence or the absence of the DNMT3L gene can be detected from a biological sample by any known detection method suitable for the detection, i.e. methods based on detecting the gene (or DNA) and/or those based on detecting the gene expression products (mRNA or protein). Such methods are easily recognized by those skilled in the art and include in situ hybridizations, such as fluorescence in situ hybridization (FISH) and mRNA in situ hybridization, and Southern analysis, RT-PCR, Northern and Western analyses, immunohisto- chemistry, and other immunoassays. Preferable methods are those suitable for use in routine clinical laboratories, such as FiSH and immunohistochemistry.

In some embodiments of the invention, known detection methods suitable for detecting mutations can be used. Such methods include the method described by Saiki et al. [Proc. Natl. Acad. Sci USA 86 (1989) 6230- 6234) utilizing hybridization to an allele specific oligonucleotide probe, or modifications thereof; the method described by Newton, C. R. et al. [Nucl. Acids Res. 17 (1989) 2503-2516] using the DNA sequences or DNA-fragments of the invention as probes; the solid phase minisequencing method described by Syvanen er al. [Genomics 8 (1990) 684-692] in which use is made of a biotinylated probe; or the oligonucleotide ligation method described by Lande- gren, U. et al. [Science 24J_. (1988) 1077-1080]. Methods include the denaturing gradient gel electrophoresis (DGGE) [Fischer, S.G. and Lerman, L.S., PNAS 80 (1983) 1579-1583] or a modification of this method, constant dena- turant gel electrophoresis (CDGE) [Hoving et al., Genes Chromosomes Cancer 5 (1992) 97-103]. The mutation separation principle of DGGE and CDGE is based on the melting behavior of the DNA double helix of a given fragment.

In the diagnostic methods of the invention the biological sample can be any tissue or body fluid containing cells, such as blood, e.g. umbilical cord blood, separated blood cells, such as lymphocytes, B-cells, T-cells etc., biopsy material, such as fetal liver, thymus, testis or ovary biopsy, sperm, sa- liva, etc. The biological sample can be, where necessary, pretreated in a suitable manner known to those skilled in the art.

In therapy, an altered form of DNMT3L or antisense oligonucleotide against the DNMT3L can be used therapeutically in any technique presently available for gene therapy to prevent the progression of a proliferating disease. In particular, tumor cell growth may be slowed down or even stopped by such therapy. Such techniques include the ex vivo and in situ therapy methods, the former comprising transducing or transfecting an altered DNMT3L in a vector or antisense oligonucleotides containing cells to the patient, and the latter comprising inserting the altered gene or oligonucleotide into a carrier, which is then introduced into the patient. The carrier vector used can be a retrovirus vector, an adeno virus vector, an adeno associated virus (AAV) vector or a eucaryotic vector. Depending on the disease to be treated, a transient cure or a permanent cure may be achieved. Alternatively, poly- or monoclonal antibodies can be used to suppress the function of the DNMT3L protein, and thus tumor cell growth may be slowed down or even stopped. Antibodies against DNMT3L could also be used to carry other agents, such as cytotoxic substances, to the tumor cells. Such agents could then be used to kill specifically the cancer cells.

The isolated DNMT3L sequence or a functional fragment or variant thereof, or an functionally equivalent isolated DNA sequence hybridizable thereto, or the corresponding mRNA sequence according to the invention can be used in a suitable gene therapeutic method for the control or cure of infertility in mammals, including humans, especially in males, and for the regulation of immune responses and maturation.

Similarly, pharmaceutical compositions containing the above sequences can also prepared using suitable pharmaceutical techniques, which are well known to persons skilled in the art.

The present invention provides means for a more reliable, rapid and easier diagnosis and/or treatment of various hereditary diseases and means for the regulation of gene expression, the methylation of the regulatory regions of other gene, organogenesis and embryogenesis as well as of the re- production and maturation of immune response, including disturbances in any of the above biological processes.

The invention is now elucidated by the following non-limiting examples.

Example 1

Sequencing and analysis of the genomic sequence

The genomic sequencing of the region was performed by the sequencing team in the Laboratory of Genomic Medicine, Keio University School of Medicine, Tokyo, who is carrying out the sequencing of the human chromo- some 21 q22.3 within the Japanese Science and Technology Corporation Human Genome Sequencing Project (http://www-alis-tokyo.jst.go.jp/HGS/top.- html). The exon trapping experiments from HC21 specific cosmids has been described earlier (Chen, H. M. et al., supra). The computer software programs Genscan [http://gnomic.stanford.edu/~chris/GENSCANW.html; Burge, C. and Kariin, S., J. Mol. Biol. 268 (1997) 78-94], Grail II [http://grail.lsd.ornl.gov/Grail- 1.3/; Xu, Y. et al., Genet. Eng. (NY) 16 (1994) 241-253] and NIX (http://www.hgmp.mrc.ac.uk/Registered/Webapp/nix/) were employed for gene modelling and exon prediction in further analysis of the genomic sequence. The exon specific primers were designed according to the sequences of trapped and predicted exons. Example 2

Isolation of cDNA fragments and sequence analysis The primers used in RT-PCR and RACE experiments are shown In Table 1. Testis, fetal liver and placenta mRNAs and thymus cDNA library were purchased from Clontech. The mRNAs were synthesised to the single- stranded cDNA using AMV reverse transcriptase and Marathon cDNA Amplification Kit (Clontech) prior to amplification by PCR. The cDNA fragment 1 F/2R was amplified from testis, and fragments 2F/5R and 8F/10R were amplified from placenta whereas RT-PCR fragment 4F/8R was amplified from thymus. RT-PCR reaction mixtures (25 μl) contained three microliters of 1 :250 diluted cDNA, PCR buffer (10mM Tris-HCI (pH 8.8), 50mM KCl, 0.08 % Nonidet P40), 1.5-1.75 mM MgCI₂, 0.2 mM dNTP, 5 pmol of each primer and 0.2 U Thermus aquaticus DNA polymerase (Fermentas). The 1 F/2R cDNA amplification was performed in a mixture containing Expand HF buffer with 1.5 mM MgCI₂, 0.2 mM dNTP, 5 pmol of each primer, 1 M betaine and 0.35 U Expand high fidelity PCR enzyme (Boehringer Mannheim). The RT-PCR reactions were started with denaturation at 94°C for 3 min, followed by 40 cycles at 94°C for 20 s, 64°C for 30 s, and 72°C for 2 min. The 5'- and 3'-RACEs were performed by the Marathon cDNA Amplification Kit (Clontech) following the instructions provided by the manufacturer. The 5'- and 3'-RACE fragments (testis and fetal liver) were first amplified with primers 2R AP1 and 10F/AP1 and subsequently with nested primers 1 R/AP1 and 10F2/AP2, respectively. The cDNA fragments were gel-purified, subcloned into either a pBluescriptll SK+ (Stratagene) or a pCRII-TOPO (Invitrogen) vectors, and sequenced by dye deoxy terminator cycle-sequencing method with AmpliTaq/FS DNA polymerase (Perkin Elmer) and then analyzed by an automatic DNA sequencer (Applied Biosystems, 310 or 377). cDNA sequences were compaired against public databases with

BLASTn program [Altschul, S. F. et al., J. Mol. Biol. 215 (1990) 403-410].

Table 1. Primers for RT-PCR and RACE experiments

Name Sequence Position

1 F 5'-GGCCTGGATCCTTCCAGCTCATTC-3' 330-353 (sequence id. no. 3)

1 R 5'-TGCTGCCACGGCTCCGGGTACAGC-3' 449-472

(sequence id. no. 4) 2F 5'-CATCCCCATGGCGGCCATCCCAGC-3' 478-501

(sequence id. no. 5) 2R 5'-CTCTGCCTGTCCCGGGTGAAACGG-3' 567-590

(sequence id. no. 6) 4F 5'-TCTGCTGCGGAAGTCTCCAGGTTC-3' 645-668

(sequence id. no. 7) 5R 5'-ATCAGGGTTTCCGCAGATGAGCAG-3' 797-820 (sequence id. no. 8)

6F 5'-GCTTCGAGTGTGTGGATAGCCTGG-3' 837-860

(sequence id. no. 9) 8F 5'-AGCTGACGAGTTTGGGCTTTTTGG-3' 1089-1112

(sequence id. no. 10) 8R 5'-CCGGGTCAGAACCACTTTCCAAAA-3' 1107-1130

(sequence id. no. 1 1 ) 9R 5'-TGGGAGGACGGTCACAGGTGTGGC-3' 1230-1253

(sequence id. no. 12) 10F 5'-GTTCCACCGGTTCCTGCAGTACGC-3' 1270-1293 (sequence id. no. 13)

10F2 5'-TGCAGTACGCACGGCCCAAGCCAG-3' 1284-1307

(sequence id. no. 14) 10R 5'-CCTGGGGCTGCCTGGCTTGGGCCG-3' 1295-1318

(sequence id. no. 15) AP1 5'-CCATCCTAATACGACTCACTATAGGGC-3'

(sequence id. no. 16) AP2 5'-ACTCACTATAGGGCTCGAGCGGC-3'

(sequence id. no. 17)

Example 3 Protein sequence analysis and computer modeling

For protein analysis programs 0-Blast, ProtParam, PSORT II and ProfileScan at http://www.expasy.ch/tools/ were used. Alignments of protein domains were prepared with Clus- talW at http://transfac.gbf- braunschweig.de/dbsearch/clustalw.html.

The zinc finger of the DNMT3L protein was modeled based on the structure of LIM 2 domain of CRP2 protein [Protein Data Bank entry 1 QLI; Konrat, R. et al., J. Biol. Chem. 272 (1997) 12001], The single zinc finger motif terminal to PHD domain was modeled based on the C-terminal half of the two zinc fingers containing LIM 2 domain. The model was built using the programs Insightll and Discover (Molecular Simulations, Inc., San Diego, CA). Insertions and deletions were modeled by searching the loops from a selection of PDB entires. The model was defined by energy minimization with Discover program in a stepwise manner by using Amber force field. All heavy atoms were first fixed and then the side chains of built loops were freed followed by the backbone of loops. Finally, only the C atoms of conserved regions were constrained. Zinc binding amino acids were constrained in all minimization steps. The model was evaluated with PROCHECK [Laskowski, R. A. et al., J. Appl. Cryst. 26 (1993) 283-291 ; Morris, A. L. et al., Proteins 12 (1992) 345-354). Example 4

Northern blot and multiple tissue cDNA panel analysis Northern blots containing 2 μg of poly(A)⁺ RNA from 23 different adult and four fetal human tissues were purchased from Clontech (Human MTN Blot1-4 and Human Fetal MTN Blot). The membrane was hybridized with ³²P labelled 261 bp (5'-end, nucieotides 330-590) or 670 bp (3'-end, 1008- 1677) fragments of DNMT3L cDNA. Washing conditions were O.l xSSC, 0.5% SDS at 65°C and the filters were exposed to a Phosphoimager's image plate for 4 days. Expression analysis of the DNMT3L gene was also performed using the Human Multiple Tissue cDNA (MTC) panels (Clontech; I, II, fetal, and immune system panels) containing cDNAs from 27 human tissues. The amplification of the DNMT3L cDNAs were carried out with a primer pair 1 F/2R (Table 1 ) using approximately 0.2 ng, 20 pg and 2 pg of each of MTC panel cDNA as template. As a control, PCRs were performed with glyceraldehyde-3- phosphate dehydrogenase (G3PDH) control amplimer set using 2 pg of each of MTC panel cDNA as template. Example 5

Transfections and subcellular localization The full-length cDNA of DNMT3L was amplified with primers 5ΕcoF (5'-TTTGAATTCATGGCGGCCATCCCAGCCCTG; sequence id. nr. 18) and 3'HindR (5'-TTTAAGCTTCATGTATAAAGAGGAAGTGAG; sequence id. nr.19), digested with EcoRl and Hind\\\ restriction enzymes and cloned into pcDNA3.1 Myc-His vector (Invitrogen). The cloned fragments were verified by sequencing. Transfections of the pcDNMT3L and pcDNA constructs were performed using Qiagens's SuperFect transfection reagent. Cos cells were transfected according to the manufacturer's instructions, and after 48 hours the ceils were fixed and stained with anti-myc antibody (invirtogen) for subcellular localisation. In Western blotting after 48 hour the transfected cell were harvested and cell extracts were separated by electrophoresis on a 10% SDS/PAGE gel, transferred to nitrocellulose, and probed with an anti-myc antibody.

Claims

1. An isolated DNA sequence characterized by comprising the sequence id. no. 1 or a functional fragment or variant thereof, or an functionally equivalent isolated DNA sequence hybridizable thereto, or a corresponding mRNA thereof.

2. An isolated DNA sequence or a corresponding mRNA thereof according to claim 1 , characterized in that it is associated to diseases linked to locus 21q22.3 of the human genome.

3. An isolated DNA sequence or a corresponding mRNA thereof according to claim 1 , characterized in that it is associated to tumors.

4. An isolated DNA sequence or a corresponding mRNA thereof according to claim 3, characterized in that it is associated to tumors of the testis, ovary tumors and thymus.

5. An isolated DNA sequence or a corresponding mRNA thereof according to claim 3, characterized in that it is associated to diseases related to immune maturation and in regulation of immune response.

6. An isolated DNA sequence or a corresponding mRNA thereof according to any one of claims 1 to 5, characterized by having the sequence according to sequence id. no 1 or a functional fragment thereof.

7. A protein characterized by comprising the amino acid sequence id. no. 2 or a functional fragment or variant thereof.

8. A protein according to claim 7, characterized in that it is associated to diseases linked to locus 21q22.3 of the human genome.

9. A protein according to claim 7, characterized in that it is associated to tumors.

10. A protein according to claim 8, characterized in that it is associated to tumors of the testis, ovary tumors and thymus.

11. A protein according to claim 7, characterized in that it is asso- dated to diseases related to immune maturation and in regulation of immune response

12. A protein according to any one of claims 7 to 11 , characterized by having the amino acid sequence id. no. 2.

13. A protein according to any of claims 7 to 12, characterized by having distinct structural motifs, including the ADD finger motif.

14. A method for the diagnosis of diseases linked to locus 21q22.3, characterized by detecting in a biological specimen the presence or absence of a DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent DNA-sequence hy- bridizable thereto, or a corresponding mRNA thereof.

15. A method for the diagnosis of tumors, characterized by detecting in a biological specimen the presence or absence of a DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent DNA-sequence hybridizable thereto, or a corre- sponding mRNA thereof.

16. A method of claim 15, characterized in that the tumors of the testis, ovary tumors and thymus are detected.

17. A method for the diagnosis of infertility, especially male infertility, characterized by detecting in a biological specimen the presence or ab- sence of a DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent DNA-sequence hybridizable thereto, a corresponding mRNA thereof.

18. A method for the diagnosis of diseases related to immune maturation and in regulation of immune response, characterized by detecting in a biological specimen the presence or absence of a DNA sequence comprising the sequence id. no. 1 or a functional fragment or variant thereof, or a functionally equivalent isolated DNA-sequence hybridizable thereto, or a corresponding mRNA thereof.

19. A method according to any one of claims 14 to 18, character- ized in that DNA techniques are used for the detection.

20. A method for the diagnosis of diseases linked to locus 21q22.3, characterized by detecting in a biological specimen the presence or absence of a protein comprising the sequence id. no. 2 or a functional fragment or variant thereof.

21. A method for the diagnosis of tumors, characterized by detecting in a biological specimen the presence or absence of a protein comprising the sequence id. no. 2 or a functional fragment or variant thereof.

22. A method of claim 21 , characterized in that the tumors of the testis, ovary tumors and thymus are detected.

23. A method for the diagnosis of infertility, especially male infertility, characterized by detecting in a biological specimen the presence or ab- sence of a protein comprising the sequence id. no. 2 or a functional fragment or variant thereof.

24. A method for the diagnosis of diseases related to immune maturation and in regulation of immune response, characterized by detecting in a biological specimen the presence or absence of a protein comprising the sequence id. no. 2 or a functional fragment or variant thereof.

25. The use of the DNA sequence or a corresponding mRNA thereof according to any one of claims 1 to 6 in diagnosis or in therapy.

26. The use of the protein according to any one of claims 7 to 12 in diagnosis and in therapy.

27. The use of the DNA sequence or a corresponding mRNA thereof according to any one of claims 1 to 6 for the preparation of a medicament useful in a gene therapy method of diseases linked to locus 21q22.3 of the human genome and/or associated to tumors and/or diseases related to immune maturation and in regulation of immune response and/or infertility.

28. The use of the DNA sequence or a corresponding mRNA thereof according to any one of claims 1 to 6 in the treatment of diseases linked to locus 21q22.3 of the human genome and/or associated to tumors and/or diseases related to immune maturation and in regulation of immune re- sponse and/or infertility.