Wilms' tumour is a paediatric malignancy of the kidneys and is the most common solid tumour found in children. The Wilms' tumour suppressor protein WT1 is mutated in approx. 15% of Wilms' tumours, and is aberrantly expressed in many others. WT1 can manifest both tumour suppressor and oncogenic activities, but the reasons for this are not yet clear. The Wilms' tumour suppressor protein WT1 is a transcriptional activator, the function of which is under cell-context-specific control. We have previously described a small region at the N-terminus of WT1 (suppression domain) that inhibits the transcriptional activation domain by contacting a co-suppressor protein. We recently identified BASP1 as one of the components of the co-suppressor. Here, we analyse the mechanism of action of the WT1 suppression domain, and discuss its function in the context of the role of WT1 as a regulator of development.
WT1 (Wilms' tumour suppressor protein) and tumorigenesis
Wilms' tumour is a malignancy of kidneys that affects 1 in 10000 children, making it the most common solid paediatric tumour (reviewed in [1–5]). This condition is also associated with hereditary genitourinary disorders such as Denys Drash syndrome and Frasier syndrome. Isolation of genes associated with Wilms' tumour led to the identification of a zinc finger protein, WT1. Subsequently, WT1 was shown to be a transcriptional regulator, with putative target genes including growth factors and regulators of cell division. Although only 15% of sporadic Wilms' tumours have been found to contain mutations in WT1, several others show aberrant WT1 expression. Indeed, gene dosage appears to have major effects on WT1 function .
WT1 knockout mice do not survive gestation, displaying absence/incorrect development of the kidney, gonads, spleen, heart and diaphragm [6,7]. More recent studies have shown that WT1 is also required for formation of the retinal ganglia, and it is likely that further analyses will reveal other developmental roles for WT1 . Thus the knockout mice confirm a major role for WT1 in the formation of the genitourinary system, and also a wider role in the development of other tissues. Not surprisingly, therefore, WT1 has been implicated not only in Wilms' tumour, but also in other malignancies, such as leukaemia, desmoplastic small round cell tumour and breast cancer (reviewed in ).
Since its discovery, WT1 has proved to be a fascinating protein. Alternative splicing, RNA editing and an alternative translation start codon combine to produce a plethora of isoforms of the protein. One alternative splice gives rise to the insertion of a three amino acid sequence (KTS) between zinc fingers 3 and 4 of WT1. This insertion switches the affinity of WT1 from a specific DNA sequence to a specific RNA sequence. Moreover, the +KTS form of WT1 associates with RNA processing factors such as U2AF, and localizes to regions of RNA processing in the nucleus . Thus WT1 has been proposed to function in both transcription and RNA processing in a splice-isoform-dependent manner.
Recently, mice have been generated that express exclusively either the −KTS or +KTS isoform . The birth of both mice demonstrated that the two WT1 isoforms perform overlapping functions. However, the distinct phenotypes of each isoform-specific knockout mice suggested that they are also required to perform distinct functions. Specifically, the +KTS isoform of WT1 is critical in the determination of the male phenotype. On the other hand, the −KTS form of WT1 plays a greater role in kidney development. Thus the transcriptional regulatory function of WT1 is critical in nephrogenesis.
The other alternative splice inserts 17 amino acids (17AA) N-terminal to the WT1 zinc fingers. This WT1 17AA isoform is frequently over-represented in Wilms' tumours and other malignancies, and has been shown to have effects both on cell division and cell survival [11–14]. Specific ablation of the +17AA isoform of WT1 in mice does not result in any obvious defects in genitourinary development, suggesting that it may be required specifically for a tumour suppressor role or that it performs a subtle function .
WT1 and nephrogenesis
As mentioned above, WT1 plays a critical role in kidney development and is required for the condensation of the metanephric mesenchyme on to the ureteric bud (reviewed in [1–5]). WT1 is expressed in the mesenchymal cells, and in the WT1 knockout mice these cells undergo apoptosis. Under normal conditions, the metanephric mesenchyme mediate reciprocal signalling with the ureteric bud, and following condensation proliferate and undergo differentiation into the epithelial cells of the nephron. WT1 expression levels decrease in the latter stages of kidney development, and are very low in the adult kidney, where WT1 expression is confined to a very specialized cell type, the podocytes . The podocyte cells extend processes around the capillaries of the glomerulus and provide a filtration barrier that prevents the entry of blood proteins into the nephron tubule network. The expression of WT1 is required for the maintenance of podocyte function .
As mentioned above, the transcriptionally active form of WT1 plays a critical role in nephrogenesis and perhaps the development of other organs. Thus an analysis of WT1 transcription function will not only enhance our understanding of kidney development, but also help us to understand how this process is altered in the course of Wilms' tumorigenesis.
WT1 as a transcription factor
The role of WT1 in transcription has been a subject of intense investigation, and attempts to identify relevant target genes have been difficult. As stated above, WT1 was initially characterized as a tumour suppressor and a repressor of transcription . Thus early studies looked to Wilms' tumours to identify potential target genes of WT1. It was reasoned that genes that are overexpressed in Wilms' tumours, such as IGFII (insulin-like growth factor II), might be targets of WT1 . Reporter assays with potential target genes were consistent with this idea. However, it soon became clear that WT1 could either activate or repress transcription. Moreover, differential transcription effects were observed with WT1 that depended on a variety of factors including the WT1 isoform and cell line under study (reviewed in ). The emergence of DNA microarrays allowed a more ‘unbiased’ analysis of potential WT1 target genes. The first surprise was that microarray analyses failed to identify target genes that are repressed by WT1, but did reveal genes that might be activated by WT1 . Thus more recent studies have provided the strongest evidence yet of biologically relevant target genes of WT1. In the first, WT1 was shown to co-operate with another transcription factor, steroidogenic factor 1, to activate the Mullerian inhibitory substance gene . This is very interesting, because the WT1-associated pathologies and the knockout mouse have shown that WT1 plays a central role in sex determination. In more recent studies, DNA microarrays were used to show that WT1 transcriptionally activates the amphiregulin promoter . Amphiregulin is expressed in the developing kidney, and can induce tubule branching in kidney organ explants. In addition, using the same approach, podocalyxin was found to be a target gene of WT1 . Podocalyxin is expressed in the differentiating and mature podocyte cells of the kidney.
WT1 transcriptional regulatory domains
A schematic diagram of WT1 is shown in Figure 1. WT1 contains four zinc fingers at the C-terminus, with the KTS insertion between zinc fingers 3 and 4 (shown by the solid bar in the Figure). Early domain-mapping experiments using deletion mutants of WT1 and GAL4-fusion proteins identified the transcriptional regulatory domains within the N-terminus of WT1 [23,24]. Residues 71–180 were characterized as a repression domain (‘R’ in Figure 1), and residues 180–250 were characterized as the transcriptional activation domain (‘A’). The 17-amino-acid insertion lies C-terminal to the transcriptional activation domain (shown by the solid bar). The N-terminal 70 amino acids of WT1 have not been well characterized, but they form part of the N-terminal region of WT1 that mediates self-association (see ). A physiological role for WT1 dimers (or multimers) is not yet known, but does provide a mechanism whereby WT1 mutations may exert a dominant-negative role over a remaining wild-type allele. More recent deletion mutagenesis has delineated the suppression domain (‘SD’ in Figure 1) that lies at the N-terminus of the repression domain . The suppression domain does not mediate transcriptional repression in itself, but inhibits the transcriptional activation domain of WT1. The WT1 suppression domain can act when linked to a heterologous transcriptional activation domain. Moreover, the suppression domain shows some degree of position independence, in that its spacing from the activation domain is not critical and it can act when located either N- or C-terminal to the activation domain.
The domain structure of WT1
The WT1 transcriptional suppression domain
The minimal WT1 suppression domain has been delineated to just 10 amino acids. In Figure 2(A), a transfection assay is performed using the −/− isoform of WT1 (which lacks both of the alternative splice insertions). The reporter (the IGFII promoter linked to luciferase) was transfected into human embryonic kidney 293 cells along with either wild-type WT1 (−/−) or a deletion mutant that lacks the 10 amino acids of the suppression domain (WT1 Δ92–101). The histogram shows luciferase activity of the promoter compared with a control transfection. Whereas wild-type WT1 fails to activate the promoter, the mutant that does not contain the minimal 10-amino-acid suppression domain stimulates transcription. The suppression domain is highly conserved between species, showing 90% identity (compared with 52% overall identity between puffer fish and mouse WT1). The suppression domain has so far not been identified in any other transcription factors. However, until a thorough mutagenesis study has been performed and the critical amino acids are identified, the presence of a similar motif in other proteins cannot be ruled out.
Definition of the minimal WT1 suppression domain and evidence that it contacts a co-suppressor
Experiments using an in vitro transcription system and nuclear extracts from different cell lines confirmed that the suppression domain functions by inhibiting the activation of transcription initiation . Moreover, the suppressive effect was specific for the activation of transcription: the suppression domain does not appear to inhibit activator-independent basal transcription. As this system used ‘naked’ DNA templates as a source of promoter, a histone deacetylase activity (at least for a chromatin-remodelling function) does not appear to be required for the inhibition of transcriptional activation by the suppression domain. However, it cannot be ruled out that the suppression domain utilizes chromatin-remodelling activities in vivo.
How does the suppression domain inhibit transcriptional activation? One possibility was that the suppression domain acts through a direct intramolecular mechanism to inhibit the function of the transcriptional activation domain. Alternatively, the suppression domain may act by recruiting a co-suppressor protein that then blocks the transcriptional activation process. Three lines of evidence suggested that the WT1 suppression domain acts through the latter mechanism. Firstly, the position-independent function of the suppression domain and the observation that it could act on a heterologous activation domain made an intramolecular mechanism unlikely . A second observation is shown in Figure 2(B). In this experiment, the IGFII promoter-luciferase construct was used as a reporter and transfected into HEK-293 cells along with WT1 (−/−). An expression vector containing either GAL4 or a GAL4-fusion protein linked to the WT1 suppression domain (GAL4-SD) was co-transfected, and the luciferase activity was measured. As described above, WT1 alone failed to activate the reporter. The same reporter response was also observed when GAL4 was overexpressed along with intact WT1. However, when the GAL4-suppression domain fusion was overexpressed, the intact WT1 elicited transcriptional activation of the reporter. One interpretation of these data is that overexpression of the suppression domain titrates a co-suppressor away from the DNA-bound intact WT1, and that this then allows the WT1 activation domain to stimulate transcription. In vitro transcription assays confirmed this observation.
The third line of evidence describes affinity chromatography to deplete the WT1 co-suppressor from a HeLa cell nuclear extract . In this experiment, HeLa nuclear extract was fractionated over a column containing the WT1 suppression domain. The resulting flow-through was depleted of co-suppressor activity, as evidenced by the fact that the WT1 suppression domain was rendered non-functional in transcription assays using this extract. This initiated the search for a co-suppressor that specifically inhibited the transcriptional activation function of WT1.
Identification of a WT1 co-suppressor
We used an in vitro transcription system coupled with protein affinity chromatography to produce an enriched fraction from HeLa nuclear extract that contained transcriptional co-suppressor activity . A component of the co-suppressor fraction was identified as BASP1 (brain acid soluble protein 1) . BASP1 had previously been characterized as a protein present in the cytoplasm of brain cells, and is also known as NAP-22 (neuron-enriched acidic peptide 22) . BASP1 can be N-terminally myristoylated, and has been found to associate with components of the membrane rafts of neuronal cells . Thus BASP1 seemed an unusual candidate for a transcriptional co-suppressor.
Sequence analysis of BASP1 shows that it contains a nuclear localization sequence and, indeed, BASP1 is exclusively nuclear in some cell lines (see below). BASP1 also contains several PEST sequences, which are normally associated with proteins that have a high turnover. Several recognition sites for casein kinase II (also known as protein kinase CK2) and protein kinase C are present in BASP1, although they have not been shown to be physiological targets for phosphorylation.
BASP1 associates with WT1 in cells that naturally express both proteins . Significantly, a specific form of BASP1 that arises from myristoylation does not associate with WT1. Thus BASP1 is likely to be involved in different cellular processes, one of which may be as a transcriptional co-suppressor for WT1. BASP1 is by no means a protein found exclusively in neuronal cells. Analysis of mouse embryos and human adult tissues shows that BASP1 is very widely expressed, including in the developing and adult kidney . Thus BASP1 is present in the nephrogenic intermediates of the embryonic kidney at the same time as WT1. Moreover, like WT1, BASP1 expression is restricted to the podocyte cells of the adult kidney. Thus WT1 and BASP1 are temporally and spatially co-expressed in the kidney.
Does BASP exhibit WT1 transcriptional co-suppressor activity? Transfection assays suggest that BASP1 can elicit transcriptional repression via WT1, and that this is dependent upon the suppression domain . Importantly, ablation of BASP1 expression using siRNA (small interfering RNA) leads to transcriptional activation by WT1, consistent with a role as the WT1 transcriptional co-suppressor. Whether BASP1 alone acts as the co-suppressor is not yet clear. It is likely that BASP1 acts as part of a co-suppressor complex, but the other components have yet to be identified.
As explained above, BASP1 can be found in either the cytoplasm or the nucleus. In many cases, it is almost exclusively within one compartment. The molecular basis for the cell-type-specific localization of BASP1 is not clear. The presence of WT1 does not appear to affect the cellular localization of BASP1. Future experiments will shed light on the mechanism by which BASP1 is distributed within the cell. However, it is tempting to speculate that BASP1 re-localization might play a role in modulating the transcriptional activation function of WT1.
Genes: Regulation, Processing and Interference: A Focus Topic at BioScience2004, held at SECC Glasgow, U.K., 18–22 July 2004. Edited by I. McEwan (Aberdeen, U.K.), B. White (Glasgow, U.K.), S. Graham (Glasgow, U.K.), S. Roberts (Manchester, U.K.), A. Sharrocks (Manchester, U.K.), D. Black (Organon, U.K.), S. Newbury (Oxford, U.K.), J. Sayers (Sheffield, U.K.) and A. Lloyd (University College London, U.K.).
This work was supported by the AICR, Cancer Research UK and the Wellcome Trust. S.G.E.R. is a Wellcome Trust Senior Research Fellow.