The Wilms’ tumour suppressor gene (WT1) encodes a protein(s) with 4 zinc fingers that is essential for the development of the genitourinary system. A considerable body of evidence exists to support the idea that WT1 binds DNA and functions as a transcription factor. However, we have shown recently by confocal microscopy and immunoprecipitation studies that a significant proportion of WT1 is associated with splice factors in kidney cell lines, fetal tissues and transfected Cos cells. Different isoforms of WT1 are produced by an alternative splice that leads to the presence or absence of a 3 amino acid insertion (KTS) between zinc fingers 3 and 4. We have shown that these different forms localise differently in the nucleus. The +KTS form mainly localises with splice factors, the −KTS form mainly with transcription factors. Here we propose a model to account for these different localisations. Also, we discuss the possible significance of these findings.

Wilms’ tumour (WT) is a relatively common paediatric malignancy of the kidney affecting in the order of one in 10,000 children (Hastie, 1994). WT is perhaps the most striking example of how cancer can arise through disruption of development. In this case metanephric stem cells which should normally differentiate into the epithelial components of the nephron, continue to divide in an uncontrolled manner, leading to tumours which may be in the order of 2 kg in size. These tumours attempt to recapitulate the normal stages of nephrogenesis but do so in an abortive fashion (Hastie, 1994).

From genetic analysis there is evidence for at least 3 different WT disposition genes, 2 of these mapping on the short arm of human chromosome 11. So far only one of these has been isolated, the Wilms’ tumour suppressor gene 1 or WT1, mapping to chromosome 1 Ipl3 (Call et al., 1990; Gessler et al., 1990). WT1 is a classical tumour suppressor gene in that complete loss of function in kidney stem cells is usually required for the initiation of tumorigenesis.

The WT1 gene encodes a protein with 4 zinc fingers, the last 3 of which have a high degree of similarity with the 3 zinc fingers of known transcription factors, Spl, EGR1 and EGR2. However, unlike these other zinc finger proteins, 4 slightly different isoforms of WT1 are produced by alternative splicing (Fig. 1). One of these alternative splices leads to the inclusion of an extra 17 amino acids upstream of the zinc fingers and another leads to the insertion of 3 amino acids: lysine, threonine and serine (KTS) between zinc fingers 3 and 4.

Fig. 1.

Structure of the 4 different WT1 isoforms produced through alternative splicing. The - and + refer to the absence or presence of the 17 amino acid exon 5 or the 3 amino acid insertion at exon 9 (+KTS). The rectangular box depicts the large proline/glutamine-rich regulatory domain.

Fig. 1.

Structure of the 4 different WT1 isoforms produced through alternative splicing. The - and + refer to the absence or presence of the 17 amino acid exon 5 or the 3 amino acid insertion at exon 9 (+KTS). The rectangular box depicts the large proline/glutamine-rich regulatory domain.

The highest levels of WT1 expression are detected during fetal development, particularly in the kidney, gonad and mesothelium, the latter being a tissue that surrounds the body cavity and the thoracic organs (Pritchard-Jones et al., 1990). These 3 tissues all arise from mesoderm and all undergo a mesenchyme to epithelial transition in the cells expressing WT1 at high levels.

The study of spontaneous mutations in humans and manufactured mutations in mice have shown that WT1 function is essential for the normal development of the kidney, gonad and mesothelium. Humans inheriting heterozygous deletions of chromosome lip 13 i.e. with one functional copy of WT1, often have mild developmental abnormalities of the kidneys and gonads as well as Wilms’ tumours. However, a particular class of inherited WT1 mutations can lead to much more severe abnormalities of the kidneys and gonads in children with the so-called Dennis Drash Syndrome (DDS) (Pelletier et al., 1991; Hastie, 1994). These WT1 mutations, either missense or nonsense mutations, always affect the zinc finger region of the proteins specifically so that nucleic acid binding is destroyed. In these children only one copy of the gene is mutated, the other being wild type, so it has been concluded that the severe phenotype is caused by dominant negative mutations in the gene. The most convincing proof that WT1 is required for normal genitourinary development comes from analysis of homozygous null mice created by gene targeting (Kreidberg et al., 1993). These mice die at 14-15 days gestation, completely lack kidneys and gonads and have an abnormal mesothelium. Consideration of all the different phenotypes arising through different WT1 mutations has led us to conclude that WT1 is required for at least 3 different stages of kidney development (Hastie, 1994). Hence WT1 is a tumour suppressor gene essential for normal genitourinary development.

What is the role of WT1 in the developmental process -what is the biochemical function of the protein? Evidence has accu-mulated to suggest that WT1 is a transcription factor, in particular a transcriptional repressor. There are several lines of evidence to support this idea: (1) WT1 binds G-rich DNA motifs with moderately high affinity (Rauscher et al., 1990; Bickmore et al., 1992; Drummond et al., 1994). (2) Inherited missense mutations in children with DDS affect residues predicted to contact DNA directly from crystallographic studies (Pelletier et al., 1991; Pavletich and Pabo, 1991). (3) WT1 can repress the expression of reporters linked to promoters containing G-rich binding sequences. (4) The proline/glutamine-rich repression domain from WT1 can be transferred to other proteins such as EGR1, transforming these proteins from transcriptional activators to repressors (Madden et al., 1991). Several potential target genes for repression by WT1 during development have been identified -these include the insulin-like growth factor 2 gene (IGF2), the IGF1/2 receptor gene, the PDGFA gene and the PAX2 developmental gene (Hastie, 1994). Although these are all good candidates for physiological targets of WT1, there is no direct in vivo evidence to prove this association.

What is the significance of the 4 different isoforms of WT1 created by alternative splicing? It is important to stress that the predominant forms of WT1 mRNA are those containing the 2 inserts (+/+, Fig. 1), though it is not clear whether this reflects the level of the protein (Haber et al., 1991). It appears that the ratio of the different mRNAs is consistent throughout development and in different tissues. We have little understanding of the significance of the 17 amino acid insert, though it can influence the strength of transcriptional repression under certain circumstances. However, it is clear that the +KTS form of the molecule has different DNA binding properties from the −-KTS form (Bickmore et al., 1992; Drummond et al., 1994). The −-KTS form binds a nonomer of the form GCGGGGGCG with moderately high affinity but the +KTS form will not bind to this sequence. It appears that inserting KTS makes the protein one which will now bind through all 4 zinc fingers to a I2mer. Hence the +KTS form will bind to a more limited set of DNA targets than the −KTS form.

There is strong evidence to suggest that the + and −-KTS forms are essential for normal development and that their ratio is critical; this comes from 2 sources. Firstly, we have shown recently that the −- and +KTS forms are conserved throughout all vertebrate WTls, including those in fish, amphibians, reptiles, birds and mammals; however the 17 amino acid insert is only found in mammals (Kent et al., unpublished; Hastie, 1994). Secondly, several children with DDS have inherited remarkable mutations which affect splice sites such that only the -KTS form of the mRNA is synthesised from one allele (Bruening et al., 1992). Hence, there is a reduced ratio of +KTS form to −-KTS form in these children who develop severe abnormalities of the kidneys and gonads as well as Wilms’ tumour. From this it is tempting to conclude that +KTS and −-KTS forms have different functions and that their ratio is critical.

Over the past few years all the studies on transcriptional properties of WT1 have been carried out in artificial systems using transient transfections in cells not normally expressing the protein. Recently we have set up fetal kidney and gonad cell lines from mice in order to study WT1 protein in its natural environment. By studying WT1 in these cells we hope to identify the physiologically relevant target genes and the proteins with which WT1 may interact to perform its function. As we will now describe, these studies have led to some surprising conclusions about WT1 localisation in the nucleus (Larsson et al., 1995). Contrary to our expectations, most WT1 molecules appear to associate with splice factors, not with the transcriptional apparatus.

WT1 colocalises with snRNPs in the nucleus

In order to study the localisation of endogenous WT1 protein we studied cell lines from fetal kidney and gonad of transgenic mice carrying the polyoma large T immortalisation gene (Larsson et al., 1995). These new cell lines express high levels of WT1 mRNA and protein, comparable to the highest levels found during development. A battery of monoclonal and polyclonal WT1 antibodies was then used to investigate the localisation of WT1. A number of different controls were carried out to prove that any signals we observed were specific for WT1. The results we obtained were the same for several different WT1 antibodies. Only cell lines expressing WT1 at the mRNA level were positive with the antibody. A typical WT1 staining pattern using confocal microscopy is shown in Fig. 2. Contrary to expectations for a transcription factor, WT1 was found to be localised in 20-50 discrete spots as well as in a diffuse background signal. The pattern we obtained was reminiscent of that observed in many previous studies using Sm antibodies that recognise a number of snRNPs which function in splicing (Nyman et al., 1986; Spector et al., 1991). Using these antibodies it has been shown that snRNPs are concentrated in 2 types of structures, 20-50 speckles and 2-5 larger structures called coiled bodies. These latter structures also contain a protein called p8O coilin which is generally absent from the speckles. As can be seen in Fig. 2, WT1 antibodies light up the same pattern as the snRNP antibodies. This is clear when the 2 images are merged (yellow colour). There is, however, more diffuse background staining with the WT1 antibody than with the snRNP antibodies. We observe this colocalisation in several different WT1 expressing lines established from kidney, testis or mesothelium.

Fig. 2.

Colocalisation of WT1 with snRNPs in the nuclei of kidney cell lines. (A) The pattern using an autoimmune antibody that recognises snRNPs. (B) The pattern using a polyclonal WT1 antibody. (C) The merged pattern -colocalisation gives a yellow or orange colour; green shows the WT1 pattern -note that there is some localisation in the cytoplasm in this cell.

Fig. 2.

Colocalisation of WT1 with snRNPs in the nuclei of kidney cell lines. (A) The pattern using an autoimmune antibody that recognises snRNPs. (B) The pattern using a polyclonal WT1 antibody. (C) The merged pattern -colocalisation gives a yellow or orange colour; green shows the WT1 pattern -note that there is some localisation in the cytoplasm in this cell.

Furthermore, the same pattern was seen in frozen sections of fetal kidney and gonad (Larsson et al., 1995). We were also able to show that WT1 proteins can be immunoprecipitated by several different splice factor antibodies from all these different cell lines suggesting a direct interaction between WT1 and components of splicing factor complexes (Larsson et al., 1995).

WT1 and splice factors relocalise in a similar fashion following heat shock

When cells are exposed to a variety of insults splice factors reorganise in the nucleus. For example when transcription is inhibited by actinomycin D most snRNPs localise in larger bodies (Carmo-Fonseca et al., 1991, 1992). However coilin, U170K and the splice factors U2AF now relocalise in a ring round the nucleolus. We have shown that WT1 dissociates from snRNPs following actinomycin treatment and now relocalises around the nucleolus where it can still be immunoprecipitated by antibodies to coilin (Larsson et al., 1995). Another treatment which causes relocalisation of snRNPs in the nucleus is heat shock; this leads to a diffuse staining pattern which reorganises back into speckles and coil bodies after 20 minutes at 37°C (Fig. 3). M15 kidney cells were incubated at 45°C for 15 and then stained with WT1 and Sm antibodies for confocal microscopy. As for the snRNPs (data not shown) the WT1 pattern became diffuse after heat shock. However after 15 minutes at 37°C both snRNPs and WTI reappeared together in speckles. Together with the other data we have accumulated these results support the idea that WTI and splice factors are intimately associated in the nucleus.

Fig. 3.

WTI relocalises following heat shock. The M15 kidney cell line was incubated at 45°C for 15 minutes and then allowed to recover for 15 minutes at 37°C. (A) The WTI staining pattern for heat shock, (B) after heat shock and (C) after recovery. Note that the pattern becomes more diffuse after heat shock but returns to a spotty pattern after recovery.

Fig. 3.

WTI relocalises following heat shock. The M15 kidney cell line was incubated at 45°C for 15 minutes and then allowed to recover for 15 minutes at 37°C. (A) The WTI staining pattern for heat shock, (B) after heat shock and (C) after recovery. Note that the pattern becomes more diffuse after heat shock but returns to a spotty pattern after recovery.

Different WT1 isoforms localise to different compartments in the nucleus

The pattern of WTI localisation we observed in the kidney and gonad cell lines is likely to be a composite of all the 4 isoforms produced through alternative splicing. We were interested in knowing whether the different WTI splice forms localise in a similar fashion in the nucleus, particularly in light of genetic evidence which suggests that the +KTS and -KTS forms may have different functions. To address this question we introduced each of the 4 WTI isoforms separately into Cos cells, admittedly an artificial system, with the hope that WTI localisation might reflect localisation of the endogenous proteins at least to some extent (Larsson et al., 1995).

When WTI was introduced on expression vectors containing SV4O origins of replication into Cos cells, approximately 5% of the cells now stained with WTI antibody. The pattern of staining observed could be subdivided into 3 types; speckles, diffuse or large coherent domains. The proportion of the types of patterns differed between the WTI splice forms.

In a nutshell, the 2 forms of the protein containing the +KTS inserts (−/+,+/+) were more likely to give a speckled pattern, whereas the forms lacking the +KTS motif (−/ −,+/ −) localised to large domains or gave a diffuse pattern in most cells. To determine whether the speckles corresponded to snRNP-con-taining bodies the cells were also stained with Sm antibody. This confirmed that the +KTS form colocalised with snRNP in the majority of cells whereas −KTS forms only showed colocalisation with snRNPs in a very small minority (Larsson et al., 1995).

What is the nature of the large domains of staining observed with both −KTS forms of WT1?

When vectors containing SV4O origins are introduced into Cos cells this induces a lytic SV4O infection. Viral transcription takes place in large domains in the nucleus. The transcription factor Spl is known to be induced in this situation and localised to these large domains. To determine whether the large domains observed with the -KTS form of WT1 correspond to these large domains containing Spl and other transcription factors, co-immunohistochemistry experiments were carried out. The results were very clear cut; the -KTS forms of WT1 colocalised with Spl and the basal transcription factor TFIIB in large domains in the Cos cells (Larsson et al., 1995).

What factors determine the differential localisation of WT1 isoforms in the nucleus?

Although the +KTS forms of WT1 mainly colocalise with snRNPs in Cos cells and the -KTS forms mainly with transcription factors, the difference is not absolute. In other words -KTS forms are found to colocalise with snRNPs in a small percentage of cells and the +KTS forms are found in the transcriptional factor domains in a small proportion of cells. Hence as all 4 forms can associate with snRNPs, it is unlikely that the KTS motif itself plays a role in binding of the protein to the splice factors. One model we have considered is that the relative DNA binding affinities of the + and −KTS forms may dictate the nuclear localisation of the proteins. Hence the −KTS form is known to bind a wider set of DNA sequences with a higher affinity than the +KTS forms (Bickmore et al., 1992; Drummond et al., 1994).

Does the +KTS form tend to colocalise with snRNPs because of this reduced affinity with DNA?

We then asked what would happen if we treated the Cos cells with DNase -would the −KTS form now localise with the snRNPs in a high proportion of cells? This is exactly what we observed (Fig. 4).

Fig. 4.

DNase treatment changes the distribution of WT 1 (−KTS) in transfected Cos cells. Percentage of transfected cells refers to percentage of transfected cells showing colocalisation of WT1 with snRNPs. Two forms of WT1 (−17aa/-KTS) and (+17aa/+17KTS) were studied.

Fig. 4.

DNase treatment changes the distribution of WT 1 (−KTS) in transfected Cos cells. Percentage of transfected cells refers to percentage of transfected cells showing colocalisation of WT1 with snRNPs. Two forms of WT1 (−17aa/-KTS) and (+17aa/+17KTS) were studied.

We also asked what would happen to localisation if we introduced into Cos cells WT1 forms incapable of binding to DNA because they had mutations in the zinc finger domains. Two such mutant forms (both mimicking forms observed in patients with DDS) were introduced into Cos cells. One of these was a missense mutation in zinc finger 3 converting an Arg394 to Trp; the other form lacked the last 2 zinc fingers

We have shown that both these forms are incapable of binding to a range of WT1 binding sites (Little et al., 1995). Remarkably both these mutant forms localise in speckles in 8095% of cells, showing better colocalisation with snRNPs than the wild-type WT1 forms. Thus both these experiments support the notion that DNA affinity is a strong determinant of nuclear localisation of WT1.

We have found that WT 1 associates with snRNPs in the developing genitourinary system, several fetal cell lines and transfected Cos cells. What is more, forms of the protein containing the KTS alternative splice insert are more likely to associate with splicing factors in Cos cells. What does all this mean for WT1 function?

Firstly, it is important to say that splicing in the main is not thought to occur in speckles (interchromatin granules or IGs) or coiled bodies but co-transcriptionally in the perichromatin fibrils (Mattaj, 1994). This is still somewhat controversial as the splicing of some specific mRNA precursors may well take place in the IGs. Given the fact that WT 1 immunoprecipitates with different splice factor antibodies we feel it is reasonable to assume that WT1 is associated with snRNPs in active splice complexes, where ever they may be. However, we still have to prove this and are in the process of setting up experiments to do so.

Is it that excess WT1 protein is made in the cell and that the speckles are dumping grounds for the excess protein? This is of course a possibility but one we think is incorrect. Firstly, we see the snRNP association in a variety of cell lines expressing WT1 at different levels. Secondly, the transcription factor Spl is expressed at very high levels (probably higher than transfected WT1) in transfected Cos cells but is never seen associated with snRNPs. Thirdly, another transcription factor, PAX6, only localises in transcription factor domains when transfected into Cos cells (Larsson et al., 1995). Furthermore, when Cos cells are treated with DNasel all the WT1 moves to the speckles whereas Spl exits the nucleus. Moreover, we should point out that the genetic studies in humans suggest that both +KTS and − KTS forms of the protein are essential and that they have different functions. Finally, in cells treated with actinomycin D, WT1 separates from most of the snRNP and relocates to the nucleus with a few specific proteins including the UI 7OK protein, U2AF, an auxiliary splice factor involved in alternative splicing, and P8O coilin. This again supports the idea that WT1 is involved in specific interactions with a subset of splice factors.

Hence we conclude that the interaction of WT1 with splice factors is likely to be physiologically meaningful, though we have a great deal to do to unravel the functional significance. Thus we propose that WT1, in its different forms, plays roles in regulating mRNA processing as well as transcription. Work from our collaborator, Dr Andrew Ward′s laboratory, supports this idea. They have shown that WT1 can bind RNA with high affinity through the zinc fingers whereas Egrl cannot. They have also preliminary evidence to support the idea that WT1 can regulate the expression of genes post-transcriptionally. If WT1 does play a role in regulating RNA processing this suggests that such post-transcriptional control is essential for normal kidney and gonad development and that misregulation can lead to tumours and other developmental abnormalities.

Our experiments have shown that the +KTS form is more likely to associate with snRNPs than the -KTS form though this distinction is not absolute. It is clear that all 4 forms are capable of interaction and that the domains created by alternative splicing are not themselves essential for the interaction. Why then do the 2 forms localise differently in the nucleus? The scheme we favour is that the −KTS form binds to a wide range of DNA targets and with higher affinity than the +KTS form. Hence, more of the +KTS form is free to associate with splice factors. This is supported by the fact that when we removed DNA binding sites by treating cells with DNasel or by mutating the zinc fingers, the protein associates with splice factors in a high proportion of cells (Fig. 4).

Finally, it is worth considering the components of the splicing complex with which WT1 directly interacts. In the presence of actinomycinD, WT1 relocates with a subset of splice factors including U2AF, U170K and P8O coilin. In this situation WT1 can still be immunoprecipitated by coilin antibody but not by antibody to the snRNPs. It is tempting to speculate that WT1 binds directly to one of these proteins. Studies are in progress to test this.

Bickmore
,
W. A.
,
Oghene
,
K.
,
Little
,
M. H.
,
Seawright
,
A.
,
van Heyningen
,
V.
and
Hastie
,
N. D.
(
1992
).
Modulation of DNA binding specificity by alternative splicing of the Wilms tumor wtl gene transcript
.
Science
257
,
235
237
.
Bruening
,
W.
,
Bardeesy
,
N.
,
Silverman
,
B. L.
,
Cohn
,
R. A.
,
Machin
,
G. A.
et al.  (
1992
).
Germline intronic and exonic mutations in the Wilms′ tumour gene (WT1) affecting urogenital development
.
Nature Genet
.
1
,
144
148
.
Call
,
K. M.
,
Glaser
,
T.
,
Ito
,
Cy
,
Buckler
,
A. J.
,
Pelletier
,
J.
,
Haber
,
D. A.
,
Rose
,
E. A.
,
Kral
,
A.
,
Yeger
,
H.
,
Lewis
,
W. H.
,
Jones
,
C.
and
Housman
,
D. E.
(
1990
).
Isolation and characterization of a zinc finger polypeptide gene at the human chromosome 11 Wilms′ tumor locus
.
Cell
60
,
509
520
.
Carmo-Fonseca
,
M.
,
Pepperkok
,
R.
,
Sproat
,
B. S.
,
Ansorge
,
W.
,
Swanson
,
M. S.
and
Lamond
,
A. I.
(
1991
).
In vivo detection of snRNP-rich organelles in the nuclei of mammalian cells
.
EMBO J
.
10
,
1863
1873
.
Carmo-Fonseca
,
M.
,
Pepperkok
,
R.
,
Carvalho
,
M. T.
and
Lamond
,
A. I.
(
1992
).
Transcription-dependent colocalization of the UI, U2, U4/U6 and U5 snRNPs in coiled bodies
.
J. Cell Biol
.
117
,
1
14
.
Drummond
,
I. A.
,
Rupprecht
,
H. D.
,
Rohwer-Nutter
,
P.
,
Lopez-Guisa
,
J. M.
,
Madden
,
S. L.
,
Rauscher
,
F. J.
III
and
Sukhatme
,
V. P.
(
1994
).
DNA recognition by splicing variants of the Wilms′ tumor suppressor, WT1
.
Mol. Cell Biol
.
14
,
3800
3809
.
Gessler
,
M.
,
Poustka
,
A.
,
Cavenee
,
W.
,
Neve
,
R. L.
,
Orkin
,
S. H.
and
Bruns
,
G. A. P.
(
1990
).
Homozygous deletion in Wilms′ tumours of a zinc-finger gene identified by chromosome jumping
.
Nature
343
,
774
778
.
Haber
,
D. A.
,
Sohn
,
R. L.
,
Buckler
,
A. J.
,
Pelletier
,
J.
,
Call
,
K. M.
and
Housman
,
D. E.
(
1991
).
Alternative splicing and genomic structure of the Wilms′ tumor gene WT1
.
Proc. Nat. Acad. Sci. USA
88
,
9618
9622
.
Hastie
,
N. D.
(
1994
).
The genetics of Wilms′ tumor - a case of disrupted development
.
Annu. Rev. Genet
.
28
,
523
58
.
Kreidberg
,
J. A.
,
Sariola
,
H.
,
Loring
,
J. M.
,
Maeda
,
M.
,
Pelletier
,
J.
,
Housman
,
D.
and
Jaenisch
,
R.
(
1993
).
WT-1 is required for early kidney development
.
Cell
74
,
679
691
.
Larsson
,
S. H.
,
Charlieu
,
J.-P.
,
Miyagawa
,
K.
,
Engelkamp
,
D.
,
Ross
,
A.
,
van Heyningen
,
V.
and
Hastie
,
N. D.
(
1995
).
Subnuclear localization of WT1 in splicing or transcription factor domains is regulated by alternative splicing
.
Cell
81
,
391
401
.
Little
,
M.
,
Holmes
,
G.
,
Bickmore
,
W.
,
van Heyningen
,
V.
,
Hastie
,
N.
and
Wainwright
,
B.
(
1995
).
DNA binding capacity of the WT1 protein is abolished by Denys-Drash syndrome WT1 point mutations
.
Hum. Mol. Genet
.
4
,
351
358
.
Madden
,
S. L.
,
Cook
,
D. M.
,
Morris
,
J. F.
,
Gashler
,
A.
,
Sukhatme
,
V. P.
and
Rauscher
,
F. J.
Ill
(
1991
).
Transcriptional repression mediated by the WT1 Wilms′ tumor gene product
.
Science
253
,
1550
1553
.
Mattaj
,
I. W.
(
1994
).
Splicing in space
.
Nature
372
,
727
728
.
Nyman
,
U.
,
Hallman
,
H.
,
Hadlaczky
,
G.
,
Pettersson
,
I.
,
Sharp
,
G.
and
Ringertz
,
N. R.
(
1986
).
Intranuclear localization of snRNP antigens
.
J. Cell Biol
.
102
,
137
144
.
Pavletich
,
N. P.
and
Pabo
,
C. O.
(
1991
).
Zinc finger-DNA recognition: crystal structure of a Zif268-DNA complex at2.lÅ
.
Science
252
,
809
817
.
Pelletier
,
J.
,
Schalling
,
M.
,
Buckler
,
A. J.
,
Rogers
,
A.
,
Haber
,
D. A.
and
Housman
,
D.
(
1991
).
Expression of the Wilms’ tumor gene WT1 in the murine urogenital system
.
Genes Dev
.
5
,
1345
1356
.
Pritchard-Jones
,
K.
,
Flemming
,
S.
,
Davidson
,
D.
,
Bickmore
,
W. A.
,
Porteous
,
D.
,
Gosden
,
C.
,
Bard
,
J.
,
Buckler
,
A.
,
Pelletier
,
J.
,
Housman
,
D.
,
van Heyningen
,
V.
and
Hastie
,
N. D.
(
1990
).
The candidate Wilms′ tumor gene is involved in genitourinary development
.
Nature
346
,
194197
.
Rauscher
,
F. J. HI
,
Morris
,
J. F.
,
Tournay
,
O. E.
,
Cook
,
D. M.
and
Curran
,
T.
(
1990
).
Binding of the Wilms′ tumor locus zinc finger protein to the EGR-1 consensus sequence
.
Science
250
,
1259
1262
.
Spector
,
D. I.
,
Fu
,
X.
and
Maniatis
,
T.
(
1991
).
Associations between distinct pre-mRNA splicing components and the cell nucleus
.
EMBO J
.
10
,
34673481
.