Rare mutations of the DMBT1 gene in human astrocytic gliomas

The Deleted in Malignant Brain Tumors 1 gene (DMBT1) has been proposed as a tumor suppressor gene candidate in human brain tumors, based on the observation of homozygous deletions affecting the DMBT1 region or part of the gene. In order to support this hypothesis, we performed a mutational analysis of the entire coding region of DMBT1, employing SSCP analysis and direct DNA sequencing in a series of 79 astrocytic gliomas. Five somatic mutations were detected. Two mutations, one of which resulted in an amino acid exchange, occurred in glioblastomas. One pilocytic astrocytoma carried two missense mutations and another pilocytic astrocytoma contained a somatic mutation, not affecting the presumed protein. In addition, 21 of the 27 single nucleotide polymorphisms identified in this study have not been recognized previously. The data indicate, that small mutations are not a frequent finding in gliomas.

Basal transcription activity of the dyskeratosis congenita gene is mediated by Sp1 and Sp3 and a patient mutation in a Sp1 binding site is associated with decreased promoter activity

The multisystem disorder dyskeratosis congenita (DKC) is caused by mutations in the DKC1 gene. The protein dyskerin is a component of the box H+ACA small nucleolar RNAs (snoRNAs) and is also functionally associated with the RNA component of the human telomerase. The majority of mutations are missense mutations, although single examples of non-coding mutations have been described. One of these is a point mutation in a putative Sp1 binding site in the 5'-upstream region of the DKC1 gene which presumably represents the promoter region of the gene. In this report, we compare the promoter sequences of both the human and mouse genes and provide a first functional characterisation of the human DKC1 promoter. This includes a characterisation of the disease-associated implications caused by the mutation identified in one patient. By reporter gene analysis, functional regions of the DKC1 promoter were delineated. The core promoter region critical for basal level of transcription was found to lie at -10 to -180. Bandshift- and supershift experiments clearly demonstrated a mutual binding of transcription factors Sp1 and Sp3 to two of five putative GC-box/Sp1-binding sites located within the core promoter region. An additional GC-box interacts only with the Sp1 transcription factor. Further, we provide evidence that the DKC1 mutation in one of the Sp1 binding sites results in reduced promoter activity.

Pore membrane and/or filament interacting like protein 1 (POMFIL1) is predominantly expressed in the nervous system and encodes different protein isoforms

We have isolated and characterized a novel differentially spliced gene predominantly expressed in the nervous system, which encodes protein isoforms with significant homology to the alpha-actinin protein superfamily, the Caenorhabditis elegans UNC-53 protein and weak homology to the nuclear membrane protein POM121. Similar to POM121 the primary structures show a hydrophobic region that is likely to form one or more adjacent transmembrane segment(s). Indirect immunofluorescence with antibodies against a synthetic peptide gave staining of the nucleus. Target experiments with EGFP (enhanced green fluorescent protein)-fusion proteins confirmed the nuclear localization. Two further members of this gene family could be isolated. All three pore membrane and/or filament interacting like (POMFIL) genes are differentially expressed in neuronal tumor cell lines. In 40% of tested primary neuroblastomas expression of POMFIL1 is strongly reduced and after brain injury POMFIL1 protein expression is upregulated, indicating that POMFIL1 is involved in the process of neuron growth and regeneration, as well as in neural tumorigenesis.

A mutation hot spot for nonspecific X-linked mental retardation in the MECP2 gene causes the PPM-X syndrome

We report here the genetic cause of the X-linked syndrome of psychosis, pyramidal signs, and macro-orchidism (PPM-X) in a three-generation family manifesting the disorder as a mutation in the methyl-CpG binding-protein 2 (MECP2) gene in Xq28. The A140V mutation was found in all affected males and all carrier females in the family. To date, descriptions have been published of two patients with independent familial mental retardation (MR) and two patients with sporadic MR who harbor this specific mutation in the MECP2 gene. This strongly suggests that A140V is a hot spot of mutation resulting in moderate to severe MR in males. A simple and reliable PCR approach has been developed for detection of the hot spot A140V mutation to prescreen any other unexplained cases of MR before further extensive mutation analyses.

Identifying splits with clear separation: a new class discovery method for gene expression data

We present a new class discovery method for microarray gene expression data. Based on a collection of gene expression profiles from different tissue samples, the method searches for binary class distinctions in the set of samples that show clear separation in the expression levels of specific subsets of genes. Several mutually independent class distinctions may be found, which is difficult to obtain from most commonly used clustering algorithms. Each class distinction can be biologically interpreted in terms of its supporting genes. The mathematical characterization of the favored class distinctions is based on statistical concepts. By analyzing three data sets from cancer gene expression studies, we demonstrate that our method is able to detect biologically relevant structures, for example cancer subtypes, in an unsupervised fashion.

One novel and two recurrent missense DKC1 mutations in patients with dyskeratosis congenita (DKC)

X-linked dyskeratosis congenita (DKC) is a progressive multisystem disorder most severely affecting tissues with a high cellular turnover such as skin, mucous membranes, and blood. Most patients die of bone marrow failure, although the chances of succumbing to various types of cancer and pulmonary disease are also high. DKC is caused predominantly by missense mutations in the DKC1 gene linked to Xq28. Some of the clinical features are reminiscent of premature ageing and this agrees with recent indications that DKC could be a telomere maintenance disorder. There is considerable variability in the type, severity, and age at onset of the various anomalies. Recognition of this has increased with the finding that patients with Hoyeraal-Hreidarsson syndrome (HHS) who exhibit severe neurological problems in addition to early-onset pancytopenia, also bear mutations in the DKC1 gene. For these reasons, and compounded by the range of mutations, phenotype-genotype correlations and accurate assessments of prognosis have not been possible. To complement the present data, we here report on three new cases of DKC and their mutations. One is a novel mutation in the exon 3 (K43E). The other two represent a frequently recurring mutation in exon 11 (A353V) and a less frequently recurring mutation in the exon 3 (T49M).

Association studies of the HOPA dodecamer duplication variant in different subtypes of autism

The HOPA gene in Xq13 is coding for a protein involved in a nuclear thyroid receptor complex. Previous studies suggested association of the dodecamer duplication in the OPA-repeat region in exon 43 (according to the genomic database sequence) with autism, mental retardation, and schizophrenia/hypothyroidism. We determined the frequency of this 12 bp duplication variant in a sample of 155 patients divided in different subtypes of autism, 278 parents of those patients, and 157 control individuals. The allele frequency of the duplication variant was not significantly different between autistic patients, their parents, and the control group. Therefore, it is unlikely that this 12 bp duplication variant of the HOPA gene has major relevance to the susceptibility to different subtypes of autism at least in this German patient sample. In addition, we identified a third variant with a 15 bp deletion in the OPA-repeat region, recently described by another group, in one autistic patient. This third allele was also present in the patient's nonautistic mother and sister, who are heterozygous for this variant, but could not be detected in any other individual genotyped in this study. Expression analysis revealed transcription of all three allelic variants in lymphoblastoid cell lines. Furthermore, we identified a new splice variant that utilizes an additional 9 bp of the 3' intron subsequent to exon 39. Both alternative transcripts are coexpressed in all fetal and adult tissues examined.

Transcription profiling of renal cell carcinoma

AIMS

Our aim was to prepare a comprehensive catalogue of the changes in gene expression accompanying the development and progression of renal cell carcinoma, and to correlate these with histo-pathological, cytogenetic and clinical findings.

METHODS

mRNA samples from paired neoplastic and non-cancerous human kidney tissue were labeled and hybridized in duplicate against high-density cDNA arrays. Two array technologies were used: 31,500-element transcriptome-wide nylon arrays for hybridization with 37 radioactively labelled sample pairs, and 4200-element kidney- and cancer-specific glass microarrays for hybridization with 19 fluorescently labelled sample pairs.

RESULTS

We identified more than 1700 cDNA clones that show differential transcription levels in kidney tumor tissue compared to normal kidney tissue. The functional classification of 389 annotated genes provided views of the changes in the activities of specific biological processes in renal cancer. Among the biological processes with a large proportion of up-regulated genes we found cell adhesion, signal transduction, and nucleotide metabolism. Down-regulated processes included small molecule transport, ion homeostasis, and oxygen and radical metabolism. Furthermore, we explored the feasibility of molecular diagnosis for renal cell tumors using cDNA microarrays on glass slides, investigating the association of transcription levels with tumor type, progression, and a putative prognostic variable. The experimental data is available from the GEO gene expression database (http://www.ncbi.nlm.nih.gov/geo; accession no. GSE3), and a comprehensive presentation of the results is available in the web supplement (http://www.dkfz-heidelberg.de/abt0840/whuber/rcc).

CONCLUSION

Transcription profiling using high-density cDNA arrays is a powerful method with the potential to improve cancer diagnosis and prognosis. The identification and classification of differentially transcribed genes, as described in our study, is the beginning of a more complete understanding of kidney cancer.

Physical and genetic characterization reveals a pseudogene, an evolutionary junction, and unstable loci in distal Xq28

A large portion of human Xq28 has been completely characterized but the interval between G6PD and Xqter has remained poorly understood. Because of a lack of stable, high-density clone coverage in this region, we constructed a 1.6-Mb bacterial and P1 artificial chromosome (BAC and PAC, respectively) contig to expedite mapping, structural and evolutionary analysis, and sequencing. The contig helped to reposition previously mismapped genes and to characterize the XAP135 pseudogene near the int22h-2 repeat. BAC clones containing the distal int22h repeats also demonstrated spontaneous rearrangements and sparse coverage, which suggested that they were unstable. Because the int22h repeats are involved in genetic diseases, we examined them in great apes to see if they have always been unstable. Differences in copy number among the apes, due to duplications and deletions, indicated that they have been unstable throughout their evolution. Taking another approach toward understanding the genomic nature of distal Xq28, we examined the homologous mouse region and found an evolutionary junction near the distal int22h loci that separated the human distal Xq28 region into two segments on the mouse X chromosome. Finally, haplotype analysis showed that a segment within Xq28 has resisted excessive interchromosomal exchange through great ape evolution, potentially accounting for the linkage disequilibrium recently reported in this region. Collectively, these data highlight some interesting features of the genomic sequence in Xq28 and will be useful for positional cloning efforts, mouse mutagenesis studies, and further evolutionary analyses.

Deleted in Malignant Brain Tumors 1 is a versatile mucin-like molecule likely to play a differential role in digestive tract cancer

Deleted in Malignant Brain Tumors 1 (DMBT1) has been proposed as a candidate tumor suppressor gene for brain, lung, and digestive tract cancer. In particular, alterations of the gene and/or a loss of expression have been observed in gastric, colorectal, and esophageal carcinomas. Initial evidence has accumulated that DMBT1 may represent a multifunctional protein. Because the consequences of a loss of DMBT1 function may be different depending on its original function in a particular tissue, we wondered if it is appropriate to assume a uniform role for DMBT1 in digestive tract carcinomas. We hypothesized that a systematic characterization of DMBT1 in the human alimentary tract would be useful to improve the understanding of this molecule and its role in digestive tract carcinomas. Our data indicate that the expression pattern and subcellular distribution of DMBT1 in the human alimentary tract is reminiscent of epithelial mucins. Bovine gallbladder mucin is identified as the DMBT1 homologue in cattle. An elaborate alternative splicing may generate a great variety of DMBT1 isoforms. Monolayered epithelia display transcripts of 6 kb and larger, and generally show a lumenal secretion of DMBT1 indicating a role in mucosal protection. The esophagus is the only tissue displaying an additional smaller transcript of approximately 5 kb. The stratified squamous epithelium of the esophagus is the only epithelium showing a constitutive targeting of DMBT1 to the extracellular matrix (ECM) suggestive of a role in epithelial differentiation. Squamous cell carcinomas of the esophagus show an early loss of DMBT1 expression. In contrast, adenocarcinomas of the esophagus commonly maintain higher DMBT1 expression levels. However, presumably subsequent to a transition from the lumenal secretion to a targeting to the ECM, a loss of DMBT1 expression also takes place in adenocarcinomas. Regarding DMBT1 as a mucin-like molecule is a new perspective that is instructive for its functions and its role in cancer. We conclude that DMBT1 is likely to play a differential role in the genesis of digestive tract carcinomas. However, although DMBT1 originally has divergent functions in monolayered and multilayered epithelia, carcinogenesis possibly converges in a common pathway that requires an inactivation of its functions in the ECM.

Prognostic impacts of cytogenetic findings in clear cell renal cell carcinoma: gain of 5q31-qter predicts a distinct clinical phenotype with favorable prognosis

To evaluate the prognostic significance of cytogenetic findings in clear cell renal cell carcinoma (RCC), cytogenetic results of 118 primary RCCs were evaluated in relation to classical indicators of prognosis and overall survival. Losses in 3p (98.3%) were most prevalent and included 32 (27.6%) monosomies of chromosome 3 and 84 (72.4%) structural aberrations involving 3p, of which 36 were unbalanced translocations, der(3)t(3;5)(p11-p22;q13-q31), resulting in duplication of 5q sequences. Patients with gain of 5q31-qter resulting from either polysomies or structural rearrangements of 5q, the most frequent of which was der(3)t(3;5), had a significantly better outcome than those without this aberration (P = 0.001). There was no association between gain of 5q or der(3)t(3;5) and any of the well-known variables for prognosis, including low versus high clinical stage and grade of malignancy. Among additional chromosomal aberrations, loss of chromosome 9/9p was associated with distant metastasis at diagnosis (P = 0.006). The data indicate that gain of 5q identifies a clinically favorable cytogenetic variant of clear cell RCC and demonstrate the impact of specific chromosome aberrations as additional prognostic indicators in clear cell RCC.

Identification and classification of differentially expressed genes in renal cell carcinoma by expression profiling on a global human 31,500-element cDNA array

We investigated the changes in gene expression accompanying the development and progression of kidney cancer by use of 31,500-element complementary DNA arrays. We measured expression profiles for paired neoplastic and noncancerous renal epithelium samples from 37 individuals. Using an experimental design optimized for factoring out technological and biological noise, and an adapted statistical test, we found 1738 differentially expressed cDNAs with an expected number of six false positives. Functional annotation of these genes provided views of the changes in the activities of specific biological pathways in renal cancer. Cell adhesion, signal transduction, and nucleotide metabolism were among the biological processes with a large proportion of genes overexpressed in renal cell carcinoma. Down-regulated pathways in the kidney tumor cells included small molecule transport, ion homeostasis, and oxygen and radical metabolism. Our expression profiling data uncovered gene expression changes shared with other epithelial tumors, as well as a unique signature for renal cell carcinoma. [Expression data for the differentially expressed cDNAs are available as a Web supplement at http://www.dkfz-heidelberg.de/abt0840/whuber/rcc.]

Allelic imbalance and fine mapping of the 17p13.3 subregion in sporadic breast carcinomas

Chromosome arm 17p is frequently altered in a variety of human cancers, especially in breast cancer, and allelic imbalances (AIs) in the region 17p13.1 do not always coincide with mutations in the TP53 gene. A second interval that frequently shows AIs at 17p is the chromosomal band 17p13.3. This region is suspected to harbor another tumor suppressor gene. In order to get more information concerning the pattern of AIs in 17p13.3, we performed analysis of AI of 49 breast carcinomas at 6 polymorphic loci in 17p13.3. Eighty-six percent of the tumors present AI at least at one marker in 17p13.3. Among all loci tested, the highest percentage of Al was observed at loci D17S5 (77%) and D17S1528 (72%). According to these results, a minimal region of deletion could be determined between the markers D17S28 and D17S5. Fine mapping of this region revealed that the size of the deleted region is about 100-150 kb. Furthermore, a subset of the patients shows two other areas with AI close to the markers D17S1574/D17S513 and D17S849, respectively.

Construction of a physical map of an autism susceptibility region in 7q32.3-q33

The fast evolving progress of the human genome mapping and sequencing efforts facilitate the detection of genes also for complex traits. We focus on the detection of susceptibility loci for autism, a prototypical pervasive developmental disorder. Five genome screens worldwide have identified several putative locations of susceptibility genes thus far, with the most common region on chromosome 7q. In order to identify new candidate genes for infantile autism we constructed a physical map of bacterial artificial chromosome, P1-derived artificial chromosome and yeast artificial chromosome clones of a 3 Mb region between D7S1575 and D7S500, including a complete contig of the approximately 1.2 Mb region around D7S2533, the marker with the most significant association result. We developed 16 novel sequence tag sites and mapped 23 genes/expressed sequence tags to the contigs. As this map contains a putative autistic disorder locus this integrated physical and transcript map provides a valuable resource for identification of candidate gene(s).

Toward a catalog of human genes and proteins: sequencing and analysis of 500 novel complete protein coding human cDNAs

With the complete human genomic sequence being unraveled, the focus will shift to gene identification and to the functional analysis of gene products. The generation of a set of cDNAs, both sequences and physical clones, which contains the complete and noninterrupted protein coding regions of all human genes will provide the indispensable tools for the systematic and comprehensive analysis of protein function to eventually understand the molecular basis of man. Here we report the sequencing and analysis of 500 novel human cDNAs containing the complete protein coding frame. Assignment to functional categories was possible for 52% (259) of the encoded proteins, the remaining fraction having no similarities with known proteins. By aligning the cDNA sequences with the sequences of the finished chromosomes 21 and 22 we identified a number of genes that either had been completely missed in the analysis of the genomic sequences or had been wrongly predicted. Three of these genes appear to be present in several copies. We conclude that full-length cDNA sequencing continues to be crucial also for the accurate identification of genes. The set of 500 novel cDNAs, and another 1000 full-coding cDNAs of known transcripts we have identified, adds up to cDNA representations covering 2%--5 % of all human genes. We thus substantially contribute to the generation of a gene catalog, consisting of both full-coding cDNA sequences and clones, which should be made freely available and will become an invaluable tool for detailed functional studies.

RAI1 is a novel polyglutamine encoding gene that is deleted in Smith-Magenis syndrome patients

The human chromosomal band 17p11.2 is a genetically unstable interval. It has been shown to be deleted in patients suffering from Smith-Magenis syndrome. Previous efforts of physical and transcriptional mapping in 17p11.2 and subsequent genomic sequencing of the candidate interval allowed the identification of new genes that might be responsible for the Smith-Magenis syndrome. In this report, one of these genes named RAI1, the human homologue of the mouse Rai1 gene, has been investigated for its contribution to the syndrome. Expression analysis on different human adult and fetal tissues has shown the existence of at least three splice variants. Moreover, the most interesting feature of the gene is the presence of a polymorphic CAG repeat coding for a polyglutamine stretch in the amino terminal domain of the protein.

Toward a catalog of human genes and proteins: sequencing and analysis of 500 novel complete protein coding human cDNAs

With the complete human genomic sequence being unraveled, the focus will shift to gene identification and to the functional analysis of gene products. The generation of a set of cDNAs, both sequences and physical clones, which contains the complete and noninterrupted protein coding regions of all human genes will provide the indispensable tools for the systematic and comprehensive analysis of protein function to eventually understand the molecular basis of man. Here we report the sequencing and analysis of 500 novel human cDNAs containing the complete protein coding frame. Assignment to functional categories was possible for 52% (259) of the encoded proteins, the remaining fraction having no similarities with known proteins. By aligning the cDNA sequences with the sequences of the finished chromosomes 21 and 22 we identified a number of genes that either had been completely missed in the analysis of the genomic sequences or had been wrongly predicted. Three of these genes appear to be present in several copies. We conclude that full-length cDNA sequencing continues to be crucial also for the accurate identification of genes. The set of 500 novel cDNAs, and another 1000 full-coding cDNAs of known transcripts we have identified, adds up to cDNA representations covering 2%--5 % of all human genes. We thus substantially contribute to the generation of a gene catalog, consisting of both full-coding cDNA sequences and clones, which should be made freely available and will become an invaluable tool for detailed functional studies.

A physical map of the human genome

The human genome is by far the largest genome to be sequenced, and its size and complexity present many challenges for sequence assembly. The International Human Genome Sequencing Consortium constructed a map of the whole genome to enable the selection of clones for sequencing and for the accurate assembly of the genome sequence. Here we report the construction of the whole-genome bacterial artificial chromosome (BAC) map and its integration with previous landmark maps and information from mapping efforts focused on specific chromosomal regions. We also describe the integration of sequence data with the map.

Physical and transcriptional mapping of the 17p13.3 region that is frequently deleted in human cancer

Studies of chromosomal losses at 17p13 have suggested the presence of at least two distinct regions for tumor suppressor genes, the TP53 region at 17p13.1 and a more distal region at 17p13.3. Within the latter region, Hypermethylated in Cancer 1 (HIC1) is located, a likely candidate for a tumor suppressor gene that has also been suggested to play a role in the pathogenesis of Miller-Diecker syndrome (MDS). However, single-gene isolation efforts have retrieved additional genes from 17p13.3 that could play a role in tumorigenesis. This indicates that the full potential of this chromosomal region with respect to disease-related genes has not yet been exhausted and that there may exist still unknown genes that contribute to tumorigenesis or to the complex MDS phenotype. To provide a basis for the systematic isolation and evaluation of such genes, we established a physical map over 1.5 Mb of 17p13.3 and assigned 29 transcriptional units within this region.

Mapping of X chromosome inversion breakpoints [inv(X)(q11q28)] associated with FG syndrome: a second FG locus [FGS2]?

FG syndrome is an X-linked condition comprising mental retardation, congenital hypotonia, macrocephaly, distinctive facial changes, and constipation or anal malformations. In a linkage analysis, we mapped a major FG syndrome locus [FGS1] to Xq13, between loci DXS135 and DXS1066. The same data, however, clearly demonstrated genetic heterogeneity. Recently, we studied a French family in which an inversion [inv(X)(q12q28)] segregates with clinical symptoms of FG syndrome. This suggests that one of the breakpoints corresponds to a second FG syndrome locus [FGS2]. We report the results of fluorescence in situ hybridization analysis performed in this family using YACs and cosmids encompassing the Xq11q12 and Xq28 regions. Two YACs, one positive for the DXS1 locus at Xq11.2 and one positive for the color vision pigment genes and G6PD loci at Xq28, were found to cross the breakpoints, respectively. We postulate that a gene might be disrupted by one of the breakpoints.

Processing and quality control of DNA array hybridization data

MOTIVATION

The technology of hybridization to DNA arrays is used to obtain the expression levels of many different genes simultaneously. It enables searching for genes that are expressed specifically under certain conditions. However, the technology produces large amounts of data demanding computational methods for their analysis. It is necessary to find ways to compare data from different experiments and to consider the quality and reproducibility of the data.

RESULTS

Data analyzed in this paper have been generated by hybridization of radioactively labeled targets to DNA arrays spotted on nylon membranes. We introduce methods to compare the intensity values of several hybridization experiments. This is essential to find differentially expressed genes or to do pattern analysis. We also discuss possibilities for quality control of the acquired data.

AVAILABILITY

http://www.dkfz.de/tbi

CONTACT

M.Vingron@dkfz-heidelberg.de

Systematic subcellular localization of novel proteins identified by large-scale cDNA sequencing

As a first step towards a more comprehensive functional characterization of cDNAs than bioinformatic analysis, which can only make functional predictions for about half of the cDNAs sequenced, we have developed and tested a strategy that allows their systematic and fast subcellular localization. We have used a novel cloning technology to rapidly generate N- and C-terminal green fluorescent protein fusions of cDNAs to examine the intracellular localizations of > 100 expressed fusion proteins in living cells. The entire analysis is suitable for automation, which will be important for scaling up throughput. For > 80% of these new proteins a clear intracellular localization to known structures or organelles could be determined. For the cDNAs where bioinformatic analyses were able to predict possible identities, the localization was able to support these predictions in 75% of cases. For those cDNAs where no homologies could be predicted, the localization data represent the first information.