The TSG101 Tumor Susceptibility Gene is Located in Chromosome 11 Band p15 and is Mutated in Human Breast Cancer

Limin Li, Xu Li, Uta Francke, and Stanley N. Cohen
Department of Genetics Howard Hughes Medical Institute
Stanford University School of Medicine Stanford, California 94305-5120

Reviewed by Coleman Dominiak ('97)

The mouse gene, tsg101, appears to suppress malignant cell growth, and the putative protein encoded by tsg 101 has the sequence characteristics of a transcription factor. The authors present the cloning and mapping of the human homologue to this gene, TSG101, and report on the presence of TSG101 mutations in human breast cancer.

First, the mouse tsg101 cDNA sequence was used to query the Expressed Sequences Tags (ESTs) database of the National Center for Biotechnology Information (NCBI) by the BLAST program in order to find homologous sequences in the human genome. The query yielded ten sequences of cDNA with 85%-95% identity to the mouse tsg101 cDNA (Figure 1Aa). Oligonucleotide primers were made based on a region of identity between two of the ESTs. With these primers and others corresponding to sequences bracketing the Lambda gt10 cloning site of a human cDNA library, the human TSG101 cDNA was amplified by PCR (Figure 1Ab). Then, the longest 5' and 3' fragments obtained from PCR amplification were joined into the pAMP1 cloning vector and sequenced.

The cDNA insert was a 1494 bp fragment with 86% identity to mouse tsg101. An open reading frame of 1140 nucleotides found in the cDNA sequence was predicted to encode a 380 amino acid protein with a mass of 42.841 kDa. Alignment of the putative human and mouse TSG101 proteins (Figure 1B) shows 94% identity between the protein sequences, with 20 amino acid mismatches and one gap. A coiled-coil domain and a proline-rich region, both typical features of the activation domain of transcription factors, are highly conserved (only one mismatch), and the leucine zipper motifs in that coiled-coil domain of each protein are identical. In addition, the human TSG101 amino acid sequence also contains seven possible protein kinase C phosphorylation sites, five possible casein kinase II phosphorylation sites, and three possible N-glycosylation sites that are found in the mouse protein. A query of the NCBI database returned no human gene sequences that had significant homology to the TSG101 gene, supporting the notion that the gene is a unique human homologue to the mouse tsg101.

A Northern blot (Figure 2) of mRNA from several human tissues probed with TSG101 cDNA showed that all tissues produced a 1.5 kb strand of mRNA that bound to the probe. The size of the mRNA was in good agreement with the size of the cloned cDNA fragment, which indicated that the cloned fragment was virtually full-length TSG101 cDNA. The figure includes a positive control showing the presence of the ubiquitous ß actin mRNA in all lanes, but the figure lacks a negative control.

The authors then report (without showing the data) that PCR amplification of the 3'-untranslated region of the TSG101 gene generated a 210 bp PCR product when each of the following sources of template DNA was used: human cells, human-Chinese hamster hybrid cells containing human chromosome 11, and human-Chinese hamster cells containing the p arm of chromosome 11. The PCR product was not generated when hamster cells or hybrid cells containing only the q arm of chromosome 11 were used as sources of template DNA. Given these results, the authors assigned the human TSG101 gene to chromosome 11p.

Fluorescence in situ hybridization (Figure 3A) was then used to specify the subbands of chromosome 11p to which TSG101 is localized. A human genomic library screened by PCR with TSG101-specific primers yielded two clones that contained the gene. Each clone was hybridized to arrays of human chromosomes and gave signals on chromosome 11 in 20 cells. The signals indicated that TSG101 is localized to subbands p15.1-p15.2 of chromosome 11. Radiation hybrid mapping confirmed this location of the TSG101 gene, and the localization of the gene by each of the three methods described is shown in an ideogram of chromosome 11 (Figure 3B).

The authors also mapped the mouse tsg101 gene. Mouse-rodent hybrid cells were used as a source of template DNA for PCR analysis, which localized the gene to mouse chromosome 7. This location lends further credence to the hypothesis that human TSG101 is a genuine homologue of the mouse tsg101 given the presence of conserved syntenic regions on human chromosome 11p and mouse chromosome 7. Efforts to map mouse tsg101 more specifically were unsuccessful, but the authors suggest that the gene is located on genetic map unit 23.5-24 based on the human mapping data.

Chromosome 11p15 is believed to contain at least one tumor suppressor gene, because deletion or loss of heterozygosity (LOH) at this region is found in several human cancers. But breast cancer is the malignancy with which mutation at 11p15 is primarily associated. The conclusion that 11p15 mutation plays a role in tumorigenesis is supported by findings that introduction of a normal chromosome 11 into breast cancer cells stifles their ability to form metastatic tumors. Because TSG101 had been mapped to chromosome 11p15, this gene may be involved in human breast cancer; the authors next examined that possibility.

RT-PCR was used to amplify TSG101 transcripts from breast cancer cells and normal breast tissue from the same patients. Figure 4A shows the primers which bracket the TSG101 open reading frame, and Figure 4B shows the gel electrophoresis of the PCR products (with primers Pr3 and Pr4 as shown in Figure 4A) from seven patients who showed TSG101 mutations out of the fifteen patients screened. Normal breast tissue from each patient contained only full-length TSG101 transcripts while cancerous tissue contained one or more TSG101 transcripts with significantly increased mobility in agarose gel indicating alleles with sizable deletions.

The cDNA fragments were then cut from the gel, purified, and sequenced. Figure 5 shows the sequencing information for both wild-type and mutant TSG101 from each of the seven patients and reveals the positions of deletions found in the various disease associated alleles. Table 1 summarizes the cDNA deletions for each patient along with the corresponding genomic DNA deletions. The authors noted that all patients but one (P5) exhibited deletion of part or all of the portion of the transcript that encodes the coiled-coil domain of the TSG101 protein, which, as mentioned above, is typically found in the activation domain of a transcription factor. P5 contains a frameshift mutation that leads to a truncated polypeptide. The presence of two mutant transcripts in three patients (P2, P4, and P7) indicates that both copies of the TSG101 gene in these patients contain deletions. The wild-type alleles are also found in tumors from these patients because some nontumor cells were present in the malignant tissue specimens, which can be expected. The authors also remark that patients 4 and 7 show evidence of multiple mutation events in one allele (both a deletion and an insertion) and suggest that TSG101 may sustain multiple mutations throughout the development of a malignancy. Patients 2 and 5 contain a fragment with mobility that is slightly greater than that of the full length transcript, but the authors fail to discuss this band.

The available genomic DNA (P1, P2, P3, P4, P5, and P7) was then amplified by PCR to be analyzed by Southern blot (Figure 6). The primers used for this PCR are shown as Pr5 and Pr6 in Figure 7A, which also shows the hybridization probes H1 (used to probe the DNA from P2) and H2 (used to probe the DNA from the remaining five sources). Each specimen of tumor cells contained deleted copies of the TSG101 gene, which were not found in normal tissue from the same patients. Consistent with the cDNA analysis above, tumor cells from P2 and P4 were shown to contain multiple alleles which again suggests the occurrence of multiple mutation events during tumorigenesis. Unlike the cDNA analysis, however, the Southern blot did not indicate the presence of two alleles in P7 tissue, which the authors attribute to the similar size of those two alleles. The authors do not discuss the absence of a signal representing full-length TSG101, but since such a signal is not found in lanes containing normal or tumor cells, this point does not seem overly problematic. The probes (H1 and H2) are apparently unable to bind to the full-length gene. Figure 6 also shows the Southern blot of a PCR-amplified 375 bp segment of TSG101 containing exon 5 as a control demonstrating that similar amounts of DNA were loaded into each lane being compared. The primers used for this PCR appear as Pr7 and Pr8 in Figure 7A.

The authors then looked for point mutations in the nondeleted TSG101 alleles from tumors in patients P1, P3, and P5 in order to ascertain whether these patients also carried two mutant copies of the gene in their tumor cells. Exons 3, 4, and 5 of TSG101 were then amplified by PCR using genomic DNA as a template, and these fragments were sequenced to determine the presence or absence of point mutations. The nondeleted allele of P1 contained a T -> C transition at base pair 1162 in exon 5 which is predicted to result in an amino acid change from Valine to Alanine at a possible protein kinase C phosphorylation site. The presence of this mutation was confirmed by restriction endonuclease analysis of the fragment (Figure 8). The point mutation causes the loss of a BbsI restriction site, which is indicated by a 126 bp fragment that is present only in the restriction digest of tumor cell DNA.

In summary, seven of fifteen patients with breast cancer contained mutations in the TSG101 gene, and four of these seven were shown to possess mutations in both copies of the allele. No mutant TSG101 genes were found in normal tissue from any of these patients.

From this paper, future research regarding TSG101 could go in several possible directions. The TSG101 gene product appears to belong to the leucine zipper family of transcription factors, so research could be directed at characterizing that transcription factor according to the specific DNA sequence to which it binds. The DNA sequence recognized by the TSG101 protein could be found in the same way that the sequence bound by the myc protein was found (Figure 9-13 in Watson, Gilman, Witkowski, and Zoller's Recombinant DNA ). The TSG101 protein could be incubated with a pool of synthetic, degenerate oligonucleotides. The oligonucleotides would consist of fixed sequences bracketing a "window" of random nucleotides except for some nucleotides at specific positions in the sequences known to be bound by leucine zipper transcription factors. The TSG101 would bind to specific oligonucleotides in the mixture, and these could be distinguished by a band-shift assay. Successive band shift assays with the DNA taken from shifted bands would enrich the mixture for the preferred binding site of the TSG101 gene and that site could be sequenced.

The transcription factor could also be characterized according to the proteins with which it interacts in regulating transcription (e.g. stathmin). Such proteins could be found by using the yeast two-hybrid system described by Chien et al. Furthermore, the region of the TSG101 protein necessary for interaction with a specific protein could also be determined with the two-hybrid system by using different segments of the TSG101 gene.

Research could also be directed toward functional experiments. Previously, it was shown that inactivation of mouse tsg101 led to tumorigenesis in mouse cells, and that recovery of tsg101 function reversed the development of a malignancy. The human TSG101 gene could be tested for the same characteristics by replacing the TSG101 in a human cell line with a gene encoding a fusion protein consisting of the TSG101 protein and the estrogen binding domain of the estrogen receptor. The TSG101 protein would be inactive in these cells and could be activated by the addition of estrogen to the growth medium; thus, the effects of TSG101 inactivation and restoration could be easily observed. If cells containing inactive TSG101 were not transformed, these cells could be transfected with expression vectors containing different oncogenes to determine if TSG101 inactivation could lead to cell transformation when various oncoproteins were present. Again, the effect of TSG101 restoration could be observed by adding estrogen to the medium.

Given above are a few experiments that seem to follow from the paper reviewed here. However, relatively little is known about the TSG101 protein, and the limit to possibly useful experiments is nowhere in sight.

*This article appears in Cell , Vol.88, 143-154.

Return to Davidson College Biology Department Home Page