Skip Navigation


Human Molecular Genetics Advance Access originally published online on October 27, 2004
Human Molecular Genetics 2004 13(24):3139-3149; doi:10.1093/hmg/ddh338
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
13/24/3139    most recent
ddh338v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Barbacci, E.
Right arrow Articles by Cereghini, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Barbacci, E.
Right arrow Articles by Cereghini, S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Human Molecular Genetics, Vol. 13, No. 24 © Oxford University Press 2004; all rights reserved

HNF1ß/TCF2 mutations impair transactivation potential through altered co-regulator recruitment

Elena Barbacci1, Angeliki Chalkiadaki2, Christelle Masdeu1, Cécile Haumaitre1, Ludmilla Lokmane1, Chantal Loirat3, Sylvie Cloarec4, Iannis Talianidis2, Christine Bellanne-Chantelot5 and Silvia Cereghini1,*

1Biologie du Développement, UMR 7622, CNRS, Université Pierre et Marie Curie, 9 quai St Bernard, 75005 Paris, France, 2Institute of Molecular Biology and Biotechnology, FORTH 1527, Vassilika Vouton, 711 10 Herakleion, Crete, Greece, 3Hôpital Robert-Debré, Service de Néphrologie, 48 bd Serrurier, 75019 Paris, France, 4Hôpital de Tours, Centre Hospitalier Universitaire, 49 bd Beranger, 37044 Tours Cedex, France and 5Hôpital Saint-Antoine, 184 rue du Faubourg Saint-Antoine, 75012 Paris, France

Received August 2, 2004; Accepted October 13, 2004


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Mutations in the HNF1ß gene, encoding the dimeric POU-homeodomain transcription factor HNF1ß (TCF2 or vHNF1), cause various phenotypes including maturity onset diabetes of the young 5 (MODY5), and abnormalities in kidney, pancreas and genital tract development. To gain insight into the molecular mechanisms underlying these phenotypes and into the structure of HNF1ß, we functionally characterized eight disease-causing mutations predicted to produce protein truncations, amino acids substitutions or frameshift deletions in different domains of the protein. Truncated mutations, retaining the dimerization domain, displayed defective nuclear localization and weak dominant-negative activity when co-expressed with the wild-type protein. A frameshift mutation located within the C-terminal QSP-rich domain partially reduced transcriptional activity, whereas selective deletion of this domain abolished transactivation. All five missense mutations, which concern POU-specific and homeodomain residues, were correctly expressed and localized to the nucleus. Although having different effects on DNA-binding capacity, which ranged from complete loss to a mild reduction, these mutations exhibited a severe reduction in their transactivation capacity. The transcriptional impairment of those mutants, whose DNA-binding activity was weakly or not affected, correlated with the loss of association with one of the histone-acetyltransferases CBP or PCAF. In contrast to wild-type HNF1ß, whose transactivation potential depends on the synergistic action of CBP and PCAF, the activity of these mutants was not increased by the synergistic action of these two coactivators or by treatment with the specific histone-deacetylase inhibitor TSA. Our findings suggest that the complex syndrome associated with HNF1ß-MODY5 mutations arise from either defective DNA-binding or transactivation function through impaired coactivator recruitment.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Maturity onset diabetes of the young (MODY) is a genetically heterogeneous form of non-insulin dependent diabetes mellitus characterized by early onset, autosomal dominant inheritance and abnormal glucose stimulated insulin secretion. Heterozygous mutations in genes encoding five cell-specific transcription factors are associated with different MODY subtypes, HNF4{alpha} (MODY1), HNF1{alpha}, (MODY3), insulin promoter factor (IPF1/PDX1; MODY4), HNF1ß/vHNF1 (MODY5) and NeuroD1 (MODY6). These transcription factors have been proposed to form an integrated regulatory network in ß-cells, which is involved in glucose metabolism and insulin secretion (1) (reviewed in 2). Yet, the precise mechanism and the affected target genes leading to the MODY phenotype are largely unknown.

Two of these MODY genes, HNF1{alpha} and HNF1ß, belong to a distinct subclass of the homeodomain family. They share a highly conserved DNA-binding domain composed of an atypical POU-specific (POUS) and POU homeodomain (POUH) and a more divergent C-terminal transactivation domain. Members of this family are also characterized by a highly conserved dimerization domain located at the first 32 amino acids. Because of these structural properties, HNF1{alpha} and HNF1ß display similar DNA-binding sequence specificity and bind as homodimers (3,4) (reviewed in 5). The structural and functional properties of HNF1{alpha}, including interactions with coactivator proteins have been extensively studied (6–10). In contrast, the molecular properties of HNF1ß remain essentially uncharacterized and mainly inferred to its homology to HNF1{alpha}.

A distinctive feature of the HNF1ß mutations identified is their association with severe non-diabetic renal defects and genital malformations, indicating that this transcription factor plays a critical role in normal development of kidney and Müllerian/Wolffian duct derivatives (11–15) (reviewed in 16,17). However, the molecular mechanisms by which heterozygous mutations in this gene cause a broad spectrum of clinical symptoms are largely unknown. Targeted disruption of the HNF1ß gene in mice results in early embryonic death (18,19), precluding an analysis of its role at later stages of development. Remarkably, mice heterozygous for a HNF1ß null allele have apparent normal kidney structure and function and show no signs of diabetes or abnormal glucose tolerance (S. Cereghini, unpublished data). The majority of MODY5 mutations initially identified are predicted to cause either truncations at early codons or an internal deletion within the POUS domain and therefore lack part of the DNA-binding domain. In vitro studies on the functional properties of some of these mutations suggested that the clinical phenotypes may be related to either loss of function or dominant-negative mechanisms (11,20). In contrast, two mutations (P328L329delCCTCT and S36F) exhibited a gain-of-function phenotype (21,22). More recently, four mutations at the splice donor site of exon 2 (14,23–25) and several novel missense mutations, mainly located in the DNA-binding domain, have been reported (12,24,26,27).

To obtain a more comprehensive picture of HNF1ß function and of the impairment caused by the mutations of this gene, we have analyzed the functional role of selected MODY5-mutations located in different domains of the protein. These include five missense mutations (24), two non-sense mutations in the POUS domain and one frameshift mutation in the C-terminal transactivation domain. We investigated their intracellular localization, DNA-binding and transactivation activities, as well as the possibility that these mutations could interfere with the function of the protein encoded by the wild-type allele. These analyses uncover the involvement of diverse mechanisms: some mutations, which affect DNA-binding, possess a weak dominant-negative activity, whereas others, with intact or partially decreased DNA-binding activity, impair primarily the transactivation potential of the protein. In addition, we demonstrate that HNF1ß physically interacts with the histone-acetyltransferases CREB-binding protein (CBP) and the p300/CBP-associated factor PCAF and with the histone-de-acetylase HDAC-1. We provide evidence that the transactivation potential of HNF1ß depends on the synergistic action of CBP and PCAF and that the impaired transactivation function of some HNF1ß mutations could be related to a defective interaction with these coactivator proteins.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Selected mutations in the HNF1ß gene
Our analysis was focused on five missense mutations of HNF1ß (24), as well as two nonsense mutations (Q147X, R177X), and a frameshift deletion in the transactivation domain of the protein (Q454fsdelAG) (Fig. 1A). These mutations were identified during a screening of patients exhibiting renal clinical features characteristic of MODY5-HNF1ß patients (hypoplastic glomerulocystic kidney, cysts, dysplastic kidneys) associated or not with early onset of diabetes. Three of them have not yet been reported.



View larger version (28K):
[in this window]
[in a new window]
  		Figure 1. (A) Schematic representation of HNF1ß domain organization. Domains are represented on the basis of three-dimensional studies reported for the structurally related HNF1{alpha} protein (10). The position of the nine mutations analyzed is indicated. The position of an artificially truncated HNF1ß mutant (L329X) is also indicated by a dotted line. (B) Pedigree of a French family with the Q454fs mutation in the HNF1ß gene. (C) Pedigree of a French family with the Q147X mutation in the HNF1ß gene. Subjects affected are indicated by filled symbols, non-affected by open symbols and probands by an arrow, with a summary of clinical findings in the text. Present subject's age (years), genotype and age at diagnosis of diabetes are indicated. NLS, nuclear localization signal; GCKD, glomerulo-cystic kidney disease.

 
The frameshift mutation Q454fsdelAG, further designed as Q454fs, is predicted to lead to a premature stop codon at position 549, instead of 557, with the insertion of novel amino acids. It was identified in a three-generation French family with variable expressivity of both renal cysts and diabetes (Fig. 1B). The Q147X was found in a three-generation French family, which co-segregated with renal dysfunction, hypoplastic glomerulocystic kidney, and diabetes. The clinical history of this family was initially described in 1982 (28). Pointing to variability of phenotype within the same family and similar to other families with HNF1ß mutations (11), one carrier had a severe pancreas atrophy of both body and tail, whereas another carrier had a rather mild phenotype and relatively normal kidney structure (Fig. 1C). The mutation R177X, which was first reported in a Japanese family (29) and more recently in an Italian family (30), was found in a 13-year-old French girl who presented hyperechogenic kidneys.

DNA-binding and heterodimerization properties of HNF1ß mutants
To examine the effect of different mutations on DNA-binding activity and heterodimerization potential, we expressed them in HEK293 or C33 cells, either alone or in combination with the wild-type protein. Both cell lines lack endogenous HNF1{alpha} and ß expression, but their epithelial nature and the embryonic kidney origin of HEK293 suggested that these cell lines probably express the cofactors required for dimerization and/or transcriptional activation. Total cell extracts from the transfected cells were assayed by gel mobility assay using as probe a double-stranded oligonucleotide [proximal element (PE)], encompassing the high affinity HNF1 binding site of the albumin proximal promoter (33). The wild-type protein bound strongly to this probe and the protein–DNA complexes were specifically supershifted by the addition of a polyclonal HNF1ß antibody (Fig. 2). Because all HNF1ß mutations identified so far retained an intact dimerization domain, we also investigate the possibility that mutations defective in DNA-binding may interfere with the function of the wild-type protein through the formation of inactive heterodimers.



View larger version (58K):
[in this window]
[in a new window]
  		Figure 2. Co-expression of various HNF1ß truncated mutants and wild-type protein results in decreased binding activity of wild-type HNF1ß. Cell extracts were prepared from HEK293 cells transiently transfected with the pCB6 vector containing the wild-type HNF1ß-cDNA and the indicated mutant cDNAs, or both. (A) Extracts producing either HNF1ß alone or co-expressing HNF1ß and the indicated mutant proteins were used in gel-shift assays, with the labelled PE oligonucleotide (Proximal Element of the rat albumin promoter encompassing the HNF1{alpha}/ß-binding site) as double-strand probe. The free-probe is not shown. The specific HNF1ß–DNA complexes super-shifted by addition of a HNF1ß-antibody ({alpha}), are indicated by head-arrows. Arrows show either HNF1ß homodimers or heterodimers between HNF1ß and the indicated mutation. Since similar levels of wild-type HNF1ß are present in all lanes shown in the western blot (C), the attenuation of HNF1ß–DNA complex is probably due to the formation of either non-functional heterodimers between wild-type and mutant proteins, or heterodimeric-complexes of intermediate migration, which are indicated by arrows. (B) Efficient heterodimerization and binding of wild-type HNF1ß and L329X proteins. Note that all wild-type protein present in the extracts heterodimerize with L329X, without evidence of attenuation in DNA-binding. (C) Western blot analysis showing comparable expression levels of wild-type protein in all samples, whereas truncated proteins are expressed at higher levels than the wild-type protein. Extracts from transfected cells were resolved on a 10% SDS–polyacrylamide gel and immunoblotted using with a HNF1ß antibody directed against residues 30–65. The wild-type (60 kDa) and truncated proteins are labelled with the HNF1ß antibody. Molecular weight markers (kDa) are shown on the left. NS, non-specific complexes; S, supershifted complexes.

 
As predicted from the structure, the Q147X and R177X mutations were unable to bind DNA (data not shown). We therefore examined the potential effects of these nonsense mutations on the DNA-binding activity of the wild-type protein. The A263fsinsGG mutant (stop codon at residue 265), which has a defective DNA-binding potential and behaves as a dominant-negative mutant both in vivo and ex vivo (20,31) was used as control. As an additional control, an artificially truncated HNF1ß protein containing a stop codon at the L329 residue (L329X) was used. L329X binds efficiently DNA as homo- or heterodimers, but does not exhibit transcriptional activity, and possess dominant-negative activity in ex vivo reporter assays (see later; Figs 5 and 6).



View larger version (49K):
[in this window]
[in a new window]
  		Figure 5. Effect of non-sense mutations Q147X, R177X and A263fsinsGG on the transcriptional activity of wild-type HNF1ß. HEK293 cells were transiently transfected with a constant amount of HNF1ß expression vector and increasing molar ratios (1x, 2x, 4x) of the indicated expression vector, together with the mouse HNF4{alpha}1-Cat (A) or DTK-HNF1-Cat reporters (B). CAT activity was normalized to transfection efficiency by measuring the ß-galactosidase activity obtained with a co-transfected RSV-ß-galactosidase construct. Normalized CAT activities and standard errors from at least three independent experiments are expressed in percentage of that of the wild-type HNF1ß alone shown as black bars (100%). A decrease of ~25–30% in the transactivation potential of HNF1ß is reproducibly observed when a 4-molar ratio of the indicated expression vector is co-expressed with the wild-type protein.

 


View larger version (23K):
[in this window]
[in a new window]
  		Figure 6. Differential transactivation of CAT reporter promoter constructs by missense and frameshift HNF1ß mutants. HEK 293 cells were transiently transfected with the indicated amounts of mutant and wild-type expression vectors, together with 2 µg of a reporter construct containing the indicated promoter. The reporter constructs used contain the mouse HNF4{alpha}1 proximal promoter (–572/+25) (A), the mouse Afp promoter (–1023/+33) (B) or the rat albumin promoter (–386/+4) (C). CAT activity was normalized to transfection efficiency by measuring the ß-galactosidase activity obtained with a co-transfected RSV-ß-galactosidase construct. It is expressed in fold-activation above the activity obtained by the CAT-reporter alone and the values and standard deviations plotted, are the mean of at least four independent experiments. In (B), for visibility reasons, only the standard deviations of the HNF1ß and Q454fs transactivation values are shown. The standard deviations of R295H, R165H, K164Q, R112P and HNF1ß-Del Ex7 transactivation values ranged from ±0.05 to ±0.3.

 
When Q147X, R177X or A263fsinsGG were co-expressed with the wild-type protein, the DNA-binding activity of HNF1ß was significantly attenuated, although equivalent levels of wild-type protein were present in all samples (Fig. 2, compare gel shift in A with western blot in C). The three mutations decreased HNF1ß binding activity to a similar degree. As expected, all wild-type HNF1ß protein appeared in HNF1ß/L329X heterodimeric complexes, with no apparent decrease in the binding activity (Fig. 2B). Notably, truncated proteins were expressed at higher levels than the full-length wild-type protein (Fig. 2C). Unexpectedly, we could detect the formation of very weak faster migrating complexes, corresponding to heterodimers between either the R177X or A263fsinsGG and the wild-type HNF1ß, as indicated by the displacement of these complexes with the HNF1ß specific antibody. In contrast, we did not detect any complexes corresponding to the binding of Q147X and HNF1ß heterodimers, indicating that such heterodimers do not bind DNA (Fig. 2A).

We next investigated whether missense mutations would affect the DNA binding of HNF1ß. DNA binding by the Q136E mutant was not detectable at any protein concentration, indicating that the Q136 residue is essential for binding (Fig. 3A). The K164Q mutant had a barely detectable DNA-binding activity, which could be visualized only after a longer autoradiographic exposure, arguing that this mutation had also a drastic effect on DNA-binding. In contrast, R112P, R165H and R295H mutations interacted with DNA with ~40–60% efficiency compared with wild-type protein (Fig. 3A). DNA-binding and competition analysis using either two high affinity HNF1 binding sites present in the albumin (PE site, Fig. 3A) and in the HNF4{alpha}1 promoters or two weak sites present in the Alphafetoprotein promoter gave similar results (data not shown). Thus, despite their location within the POUS and the POUH domains, the R112P, R165H and R295H mutations retain significant DNA-binding activity and appear to exhibit similar sequence specificity. These results further suggest that either these residues are not essential for binding or these mutations do not drastically disrupt the structure of the POUS and POUH domains of HNF1ß. In addition, as equivalent amounts of either mutant or wild-type protein were detected in all extracts (Fig. 3B), the missense mutations examined do not appear to affect protein stability.



View larger version (49K):
[in this window]
[in a new window]
  		Figure 3. DNA-binding activity and heterodimerization properties of missense and C-terminal frameshift mutations. (A) HEK293 cell-extracts overproducing single mutant proteins or the wild-type HNF1ß, were used in gel shift assays using the labelled PE-oligonucleotide. Arrows indicate wild-type or mutant homodimeric DNA-complexes. S, Specific HNF1ß–DNA complexes supershifted by the addition of a HNF1ß antibody ({alpha}) are shown by head-arrows. (B) Western blot of the cell-extracts of the indicated proteins used in (A) and revealed using a HNF1ß antibody, show that all missense mutations are produced at similar levels. (C) Gel-shift using cell-extracts from cells co-expressing the indicated proteins and the artificially truncated HNF1ß protein L329X, showing the formation of heterodimers of intermediate mobility. (D) DNA-binding and heterodimerization of Q454fs mutant. In (C) and (D), arrows show the position of HNF1ß, L329X or Q454fs homodimeric complexes as well as the heterodimers between the indicated mutant proteins and L329X, which display an intermediate mobility relative to the corresponding homodimers.

 
To detect heterodimer formation in gel mobility shift assays, full-length missense mutations were co-expressed with the truncated L329X protein into HEK293 cells. Equivalent amounts of either mutant or wild-type proteins were present in all extracts, except for L329X that was expressed at roughly 5–6 times higher amounts compared with the other HNF1ß proteins, to facilitate complete heterodimerization. The presence of DNA–protein complexes with intermediate mobility relative to that corresponding to homodimers, indicated that similarly to wild-type HNF1ß protein all mutants did heterodimerize with L329X. Compared with the HNF1ß/L329X heterodimers, the binding capacity of these heterodimeric complexes was strongly attenuated in K164Q/L329X and Q136/L329X heterodimers, whereas it was either slightly decreased in R112P/L329X and R165H/L329 heterodimers or no affected in the case of R295H/L329X heterodimers (Fig. 3C). These results further confirm that the R112P, R165H and R295H mutations do not perturb significantly the structure of the POUS and POUH bound to DNA.

The frameshift mutation Q454fs, which has intact DNA-binding and dimerization domains, bound efficiently to the HNF1 probe and heterodimerized with L329X protein (Fig. 3D). This mutant protein, as the other truncated forms of HNF1ß, was also expressed at higher levels than wild-type protein (data not shown), suggesting that the stability of HNF1ß may be modulated by its most C-terminal amino acids.

Thus, in agreement with the presence of an intact N-terminal dimerization, all mutant proteins form heterodimers and are able to interfere with the DNA-binding of wild-type HNF1ß. Moreover, the DNA-binding capacity of the heterodimeric complexes formed appears to depend on the integrity of the POUS domain.

Intracellular localization of HNF1ß mutants
The sub-cellular localization of mutant proteins was examined by indirect immunofluorescence in transfected C33 cells. As shown in Figure 4, the mutations Q147X and R177X were preferentially located in the cytoplasm, whereas the wild-type protein and the other mutants examined, including A263fsinsGG, L329X, Q136E, R112P, K164Q, R165IH, Q454fs, were all localized in the nucleus (data not shown). These results are in agreement with the predicted nuclear localization signal between residues K229 and K237, encompassing a cluster of basic amino acids (KKMRRNRFK), at the N-terminus of the homeodomain (10). These results also indicate that the missense or frameshift mutations do not perturb folding of the protein and its subsequent correct trans-location to the nucleus.



View larger version (45K):
[in this window]
[in a new window]
  		Figure 4. Sub-cellular localization of the mutant and wild-type proteins. Immunocytochemistry analyses of C33 cells transiently expressing Q147X and R177X, using an anti-HNF1ß antibody (green) and propidium iodide (PI, red) as nuclear staining, reveal that while HNF1ß is nuclear, the truncated mutant proteins are predominantly localized in the cytoplasm. To discriminate between wild-type and mutant proteins when co-expressed in C33 cells, myc tagged-HNF1ß proteins in the nucleus were visualized with the myc-epitope tag antibody, whereas anti-HNF1ß antibody reveals both wild-type and mutant proteins in the nucleus and cytoplasm. The restricted localization of myc-tagged HNF1ß proteins in the nucleus indicates that Q147X is unable to sequestrate wild-type protein to detectable levels in the cytoplasm through heterodimerization.

 
We next examined whether heterodimerization between Q147X or R177X and wild-type HNF1ß could, at least in part, relocate the wild-type protein into the cytoplasm and therefore interfere with its function. Untagged and myc-tagged versions of mutant and wild-type proteins were co-expressed in the cells and their intracellular localization was subsequently examined by indirect immunofluorescence. As shown in Figure 4, co-expression of Q147X or R177X with the myc-tagged HNF1ß protein did not modify the nuclear localization of the wild-type protein, even though significantly higher levels of truncated proteins were present in the cytoplasm (Fig. 4B; data not shown). Conversely, when mutant myc-tagged Q147X or R177X were co-expressed with the wild-type protein, we could detect immunofluorescence signals predominantly in the cytoplasm in most of transfected cells. In the latter case, however, the high expression of truncated proteins and their distribution in the nucleus and cytoplasm, did not allow us to exclude definitely the possibility that a fraction of these proteins may translocate into the nuclei through heterodimerization with wild-type HNF1ß.

These results do not support the concept of a dominant-negative action of Q147X or R177X mutations, which could be mediated by an abnormal sub-cellular localization of wild-type HNF1ß via heterodimerization.

Transactivation potential of HNF1ß mutations
To assess the functional consequences of HNF1ß mutations, vectors expressing mutant and wild-type HNF1ß proteins were transfected into HEK293 cells, along with CAT-reporter vectors, containing either native target promoters (HNF4{alpha}1, Alphafetoprotein and albumin) or a chimeric promoter encompassing two HNF1 binding sites upstream of the thymidine kinase promoter (DTK-HNF1-CAT). The proximal HNF4{alpha}1 promoter (–572/+25) and the albumin promoter (–151/+25), contain both a high affinity HNF1 binding site at positions –98 and –60, respectively (32,33). The Alphafetoprotein (Afp) promoter contains two HNF1 sites at –131 and –67 nucleotide positions (34). These promoters are transactivated by HNF1ß, albeit at lower levels than HNF1{alpha} (35).

The nonsense mutations Q147X and R177X, which lack part of the DNA-binding domain and the entire transactivation domain, were transcriptionally inactive (data not shown). We assessed their potential interference with wild-type HNF1ß-mediated transactivation by co-expressing increasing amounts of mutant proteins with a constant amount of wild-type HNF1ß. A 30% decrease in the HNF1ß transcriptional activity was observed with CAT-reporter vectors containing either the HNF4{alpha}1 promoter (Fig. 5A), the Afp promoter (–1023/+33) (data not shown), or the chimeric DTK-HNF1-Cat reporter construct (Fig. 5B).

Analysis of individual missense mutations showed that at saturating concentrations the Q136E, R112P, KI64Q and R165H mutants were transcriptionally inactive on the three promoters examined, despite the fact that only Q136E and K164Q mutations abolished DNA binding (Fig. 6; data not shown). Notably, the R295H mutant exhibited a target promoter-specific effect. At lower concentrations of expression vector, the transactivation potential of R295H on the HNF4{alpha}1 promoter varies between 75 and 78% of the activation obtained by the wild-type protein, whereas at saturating concentrations it reached wild-type transactivation values (Fig. 6A). In contrast, R295H was not able to transactivate the Afp promoter at all tested concentrations (Fig. 6B). Transactivation of the albumin promoter by this mutant was barely detectable in cells transfected with lower amounts of expression vector and reached only 25% of the wild-type activity at saturating concentrations (Fig. 6C). These results indicate that there is no a strict correlation between DNA-binding activity and transactivation function of these substitution mutants. The observation that R295H, R165H and R112P mutations retain significant DNA-binding activity, whereas their transactivation capacity is severely reduced, suggests that these mutations most likely interfere with protein–protein interactions that are necessary for transactivation function.

The frameshift Q454fs mutant, at non-saturating amounts of expression vector, exhibited 40–50% decrease in its transcriptional activity compared with that obtained with the wild-type protein, using either the Afp or the albumin promoter. When assayed on the HNF4{alpha}1 promoter, this mutation reduced transcriptional activity by 30% at non-saturating concentrations and by 10% at saturating concentrations. The Q454 residue is located within a QSP-rich domain. Although the C-terminal regions of HNF1{alpha} and HNF1ß do not exhibit extensive similarity, this domain is partially conserved and corresponds to one of the two activation domains identified in HNF1{alpha} (6). This prompted us to examine the importance of this domain in the transactivation properties of HNF1ß. As shown in Figure 6B and C, a HNF1ß-mutant construct encompassing a selective in phase deletion of exon 7 (HNF1ß-DelEx7) failed to transactivate, suggesting that important determinants for HNF1ß transactivation are conferred by the QSP-rich domain.

In contrast to Q147X, R177X and A263fsinsGG mutations, co-expression of missense mutations or Q454fs with either HNF1ß or HNF1{alpha} did not decrease their respective transcriptional activity (data not shown). Although these mutant proteins did not interfere with the activity of HNF1ß or HNF1{alpha} in co-transfection assays, it is uncertain whether these results reflect the in vivo situation. It remains possible that at physiological concentrations, these mutant proteins, in particular those that bind DNA, but show strongly decreased transactivation potential, may partially interfere with the activity of wild-type HNF1ß or HNF1{alpha}.

Interaction of HNF1ß with coactivator and co-repressor proteins
Coactivators possessing an intrinsic histone-acetyltransferase activity have been shown to stimulate transcription through the interaction with specific DNA-binding factors and the simultaneous acquisition of an active configuration [(36) and references therein]. We reasoned that the impaired transactivation function of those HNF1ß mutations that retained DNA-binding activity and, can therefore be recruited to target promoters, could be related to a defective interaction with coactivator proteins or an increased interaction with co-repressors.

The potential direct interaction of wild-type HNF1ß with coactivator and co-repressor proteins was first investigated by in vitro GST pull-down experiments. As shown in Figure 7A, full length HNF1ß interacted efficiently with CBP and PCAF histone-acetyltransferases and the histone-deacetylase, HDAC-1. PCAF interaction was observed by GST fusion proteins, containing either the N- or C-terminal part of HNF1ß. CBP and HDAC-1 selectively interacted with either the N-terminal region, or the C-terminal region of the protein, respectively.



View larger version (35K):
[in this window]
[in a new window]
  		Figure 7. In vitro and in vivo interaction of HNF1ß mutants with CBP, PCAF and HDAC-1. (A) In vitro GST pull-down experiments were performed with 35S-labeled CBP, PCAF, or HDAC-1 probes and bacterially expressed GST fusion proteins containing the full-length (FL), N-terminal part (1–313), or C-terminal part (314–557) of HNF1ß. (B) Cos-1 cells were transfected with myc epitope-tagged HNF1ß vectors together with pCMV-CBP, pCMV-Flag-PCAF, pCMV-Flag-HDAC-1 vectors as indicated. Nuclear extracts from the transfected cells were prepared and subjected to immunoprecipitation with non-immune (Pre), rabbit polyclonal CBP ({alpha}CBP) or mouse monoclonal Flag epitope ({alpha}Flag) antibodies. Co-immunoprecipitated HNF1ß proteins were detected in western blots using myc epitope antibody ({alpha}-myc). (C) Schematic representation of HNF1ß protein domains and their interactions with CBP, PCAF and HDAC-1. (D) Summary of the interaction pattern of the indicated HNF1ß mutations with CBP, PCAF and HDAC-1.

 
We next compared the in vivo interaction of the wild-type and four HNF1ß mutations with CBP, PCAF and HDAC-1 in co-immunoprecipitation experiments. Nuclear extracts from Cos-1 cells co-transfected with wild-type and mutant pCMV-myc-HNF1ß vectors in combination with pCMV-CBP, pCMV-Flag-PCAF or pCMV-Flag-HDAC-1 expression vectors, were immunoprecipitated with either the {alpha}CBP, or {alpha}Flag antibody. The presence of HNF1ß proteins in the precipitates was assessed in western blots using {alpha}-myc antibody. Efficient in vivo interaction of full-length HNF1ß was observed with all three factors (Fig. 7B). All three N-terminal missense HNF1ß mutants (R295H, R165H and R112P) were found in association with PCAF and HDAC-1, but not with CBP. In contrast, the C-terminal Q454fs mutant interacted only with CBP, but not with PCAF or with HDAC-1 (Fig. 7B and D). The differential effects of the N- and C-terminal mutations on interactions with CBP and HDAC-1 could be due to the selective association of these two proteins with the two domains of HNF1ß as evidenced by the in vitro GST pull-down experiments. PCAF, which associates with both regions of HNF1ß in vitro, did not interact with the Q454fs mutant, suggesting that the conformation adopted by this protein prevented such association.

Functional consequences of defective interactions of coactivator and co-repressor proteins with HNF1ß mutations
To investigate the functional significance of the identified interactions, we performed transient expression assays in NIH3T3 cells which lack endogenous HNF1ß and contain low levels of CBP and PCAF compared with other cell lines. We used a cell line, which contains genome-integrated 3xAlbPE-TK-CAT (9), because the chromatin structure adopted by the transiently transfected template may differ from the typical eukaryotic chromatin organization. In this assay, over-expression of CBP only marginally affected wild-type HNF1ß-induced transcription (Fig. 8). In contrast, substantial (4.5-fold) activation was observed by PCAF over-expression. This activation, however, was still far below the level of induction observed by co-expression of both CBP and PCAF proteins, which was at the range of ~24-fold (Fig. 8). This suggests that high-level transcriptional activation by wild-type HNF1ß requires synergism between the two coactivator proteins. Furthermore, treatment of the cells with TSA, a specific histone-deacetylase inhibitor, resulted in a similarly dramatic (17-fold) activation, indicating that the activity obtained by wild-type HNF1ß alone corresponds to a repressed state, possibly through its association with HDAC-1.



View larger version (13K):
[in this window]
[in a new window]
  		Figure 8. Activation of a genome-integrated HNF1-dependent reporter by HNF1ß mutants. NIH3T3 cells with stably integrated 3x AlbPE TK-CAT reporter were transfected with 100 ng of wild-type and mutant pCMV-HNF1ß vectors together with 3 µg of pCMV-CBP or pCMV-PCAF expression vectors, as indicated. Where indicated, the cells were treated with 1 µM trichostatin A (TSA) for 12 h before harvest. The bars represent normalized CAT activities and standard errors from three experiments, expressed as fold-activation above the activity obtained with transfection of wild-type HNF1ß alone.

 
We then examined the effect of expressing CBP and PCAF on those mutations whose DNA-binding activity was only weakly or not affected. The transactivation activities obtained by the R112P, R165H, R295H and Q454fs mutants were 14, 9, 21 and 18% of that detected with wild-type HNF1ß, respectively. Notably, over-expression of CBP and PCAF only marginally affected the transactivation observed by these mutants. Because all these mutants could interact in vivo with either CBP or PCAF, but not with both, this finding further corroborates the idea of the requirement of simultaneous recruitment of both coactivators to achieve high level transcription. In addition, with the exception of R112P mutant, which was induced 4.5-fold, the activity of the other mutants, even those that exhibited efficient in vivo interactions with HDAC-1, was not increased by TSA treatment.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 REFERENCES
 
In the present study, we examined the molecular consequences of HNF1ß mutations identified in patients presenting a wide clinical spectrum, including renal dysfunction and diverse renal abnormalities, diabetes, genital tract malformations and pancreas atrophy.

We investigated two major classes of mutations, substitutions in the POUS and POUH domains and truncations and/or frameshifts in different functional domains of the protein. Intragenic loss-of-function mutations caused either by nonsense mutations, deletions or insertions, account for the majority of functionally characterized MODY5 mutations. The mutations Q147X and R177X belong to this class. Our in vitro studies show that truncated HNF1ß proteins are expressed at reproducibly higher levels compared with the wild-type and can heterodimerize with wild-type protein. Interestingly, heterodimers between R177X or A263fsinsGG and wild-type HNF1ß, but not Q147X/HNF1ß heterodimers, are able to bind DNA, albeit with a significantly lower affinity than the wild-type homodimers. This suggests that the presence of the two invariant glutamines, Q136 and Q147, which create hydrogen bonds with DNA, allow the truncated POUS domain of either R177X or A263fsinsGG to contact DNA, which are subsequently stabilized by the wild-type partner in a heterodimeric complex.

In spite of their ability to heterodimerize with wild-type HNF1ß, the Q147X and R177X proteins did not sequester the wild-type protein into the cytoplasm, but exhibited a weak dominant-negative effect in transfection experiments. Nevertheless, it is possible that these mutations may be subject to nonsense mediated mRNA decay (NMD), a surveillance mechanism that remove transcripts with termination signals located at least 50 nucleotides upstream of the terminal exon–exon junction (37). This pathway, however, cannot be verified in human tissues, as expressing tissues are so far not available. On the basis of NMD, haploinsufficiency is the most favoured in vivo mechanism underlying the phenotypes associated with heterozygous nonsense alleles (38).

In contrast to premature nonsense mutations, the missense mutations analyzed here, as well as the frameshift Q454fs, are unlikely to be influenced by NMD (37). The five missense mutations affect residues that are conserved in HNF1ß proteins from zebra fish to humans and, with the exception of R112, the corresponding residues are also conserved in the structurally related HNF1{alpha} protein. Four out of five of these mutations are clustered within the POUS domain of the protein. This domain interacts with the atypical insertion of the homeodomain to create a stable interface and provides sequence specificity to the DNA interaction (10). The recently solved three-dimensional structures of the DNA-bound POUS and POUH domains of HNF1{alpha} (10) strongly suggested that the missense mutations would impair DNA binding. Substitutions in Q136 are predicted to disrupt base-specific hydrogen bonds with DNA, those in K164 and R295 are predicted to perturb cationic side-chain/phosphate backbone interactions, whereas substitutions in R165 may disrupt DNA recognition indirectly through perturbations in local environment (10). In agreement with these structural predictions, the mutations Q136E and K164Q abolished DNA binding. However, neither R165H nor R295H significantly affect DNA-binding activity.

Notably, we observed that the R112P, R165H and R295H mutations decrease severely their transcriptional potential, although retain significant DNA-binding capacity and an apparently intact transactivation domain. These data suggest that these substitutions may induce structural changes within the POUS and POUH domains with effects beyond the DNA-binding domain that influence the transactivation potential of the protein. Specifically, we present evidence that these mutations interfere with protein–protein interactions required for transactivation function. First, we show that different functional domains of HNF1ß interact in vivo with the histone-acetyltransferases CBP and PCAF and with the histone-deacetylase HDAC1. We also show that transcriptional activity of HNF1ß in a genome-integrated promoter relies on the synergistic action of CBP and PCAF.

Second, we demonstrate that the loss of interaction with either CBP or PCAF in the different HNF1ß mutants correlated with their transcriptional defects. Moreover, the activity of these mutants was not increased by treatment with the specific histone-deacetylase inhibitor TSA. These observations indicate that within the POUS and the POUH domains, DNA recognition and transcriptional properties are distinct and separable functions.

Previous work established that HNF1{alpha}-mediated transcription requires the synergistic action of CBP and PCAF and that HNF1{alpha} is subject to negative regulation by association with the co-repressors NCor and HDAC1 (36). Here, we showed a similar mechanism for HNF1ß. Thus, despite the sequence divergence of the C-terminal transactivation domains, HNF1{alpha} and HNF1ß exhibit a strikingly similar pattern of interactions with co-regulators and co-repressors, which function in part via chromatin remodelling. These findings further suggest that the two proteins use similar strategies to mediate transcriptional activation in a chromatin context.

The dramatic activation of HNF1ß upon treatment of the cells with TSA or by the simultaneous over-expression of CBP and PCAF, indicates that the repressed activity of wild-type HNF1ß can be overcome by either allosteric inhibition of histone-deacetylases or by the recruitment of the histone-acetyltransferases CBP and PCAF. The TSA-mediated induction of the transcriptional activity of the wild-type HNF1ß raises the possibility that the use of specific HDAC inhibitors may prove to be a potentially effective therapeutic strategy, as MODY5 patients possess a normal allele.

Our study on the frameshift mutation Q454fs uncovered the potential implication of a QSP-rich domain for HNF1ß transcriptional activity. An HNF1ß artificial mutant-construct encompassing a selective in phase deletion of exon 7, which primarily encodes this domain, fails to transactivate (Fig. 6). Thus, important determinants for HNF1ß transactivation appear to be restricted to a unique C-terminal domain, instead of two activation domains as it was described for HNF1{alpha} (6). The recent identification of two novel HNF1ß mutations located within this domain: a S465R substitution (26) and a frameshift mutation at residue 472 (S. Cereghini, unpublished data), further highlights the importance of this domain.

The clinical presentation of the mutations examined here reflects the large spectrum of phenotypes associated with HNF1ß mutations, suggesting a broad role for this transcription factor throughout development. However, the wide inter- and intra-familial phenotypic variability precludes a clear phenotype–genotype correlation. Although the mechanisms underlying this phenotypic variability are unknown, it is possible that environmental factors, modifier genes and stochastic developmental events are involved.

In conclusion, this comprehensive study on naturally occurring mutations within HNF1ß extends our understanding of the different functional domains of the protein, provides novel insights into structure–function relationships and demonstrates that they impair function by multiple mechanisms. Although the majority of the mutations are clustered within the DNA-binding domain, our results show that the complex syndrome associated with HNF1ß mutations may arise not only from defective DNA binding but also from decreased transactivation capacity through impaired recruitment of coactivator proteins. These observations also highlight the potential importance of a precise dosage of HNF1ß for normal function during early organogenesis and in the adult.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
REFERENCES
 
Subjects and mutation screening
Genomic DNA was prepared from EDTA blood samples and mutation analyses were performed by direct sequencing of polymerase chain reaction (PCR)-amplified HNF1ß proximal promoter and the nine coding exons, including intron–exon boundaries. The Q147X family came from France and has been followed by C. Loirat and S. Cloarec, since 1980 (28). The R177X mutation was identified in a 13-year-old girl, whose mother has kidney dysfunction. DNA from the mother was not available. The Ethics Committee of Debré and Necker Hospital approved the study and participants provided a written consent. The mutation screening and detailed clinical features of patients carrying missense mutations have been reported elsewhere (24).

Expression vectors and reporter plasmids
The wild-type human HNF1ß cDNA (kindly provided by I. Bach) was cloned between the HindIII and XbaI sites of the pCB6 vector. Mutant HNF1ß protein-constructs were generated by PCR based site-directed mutagenesis (Quick changeTM, Stratagene) on the wild-type vector. The mouse HNF4{alpha}1 promoter (–1.3 kb/+25) was generated by PCR using mouse 129sv genomic DNA. The PCR product was subcloned into pGEMT-easy vector, verified by sequence analysis, and subsequently cloned into CAT expression vector pBLCAT6 (35). The plasmid DTK-HNF1 was generated by cloning a double-stranded oligonucleotide containing two HNF1 sites of the human albumin promoter upstream to the thymidine kinase (TK) promoter (–51/+10) in pBLCAT6. To generate wild-type and mutant pCMV-myc-HNF1ß expression vectors the open reading frames of the cDNAs were amplified from the pCB6 constructs by PCR using primers containing EcoRI and XbaI sites and inserted to the same sites of pcDNA-myc vector (Invitrogen). GST fusion protein expression vectors were constructed by subcloning PCR amplified full length and fragments corresponding to 1–313, or 314–557 amino acid regions of HNF1ß to pGEX4T1 vectors. The constructions of pCMV-CBP, pCMV-Flag-PCAF and pCMV-Flag-HDAC-1 have been described previously (9,36).

Cell culture, transient transfection and immunofluorescence conditions
The NIH3T3 cell line containing genome-integrated copies of the 3x AlbPE-TK-CAT reporter (39) has been described (9). The human embryonic kidney HEK 293 and human epithelial carcinoma C33 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum as described (40). Cells were plated on 5 cm dishes 1 day before transfection at a density of 2x105/dish and transfected overnight using the standard calcium phosphate co-precipitation procedure. An aliquot of 0.5 µg of the ß-galactosidase expression vector pMJM20 was also included as a control for transfection and recovery efficiency. In co-transfection experiments, 2 µg of a promoter construct and the indicated amount of co-transfected expression vector were used. The total amount of expression vector was kept identical using the pCB6 empty vector. Transfected cells were harvested 48 h after washing and subsequently ß-gal and CAT assays were performed as described (40). To over-express wild-type and mutated proteins, HEK293 or C33 cells were transfected with the corresponding expression vectors as described earlier. Transfections were performed in 10 cm dishes using either 4 µg of expression plasmid and 4 µg of pCB6 vector; when co-expressed the wild-type and mutant protein, 4 µg of each expression were used.

Protein extracts and gel mobility-shift assays
To assess the specific DNA-binding affinity of each mutation, cells were transiently transfected with the corresponding expression vectors and the relative amount of proteins in the total cell extract was evaluated by western blotting. Nuclear extracts from transfected cell lines, gel mobility shift assays and western blots were performed as reported (41). Protein concentrations were estimated using the Bradford assay. Double stranded (ds) oligonucleotides used in gel retardation experiments correspond to the HNF1 binding site of the albumin promoter (33).

Co-immunoprecipitation assays
Nuclear extracts from transfected Cos-1 cells were prepared as described (39), and the lysates were adjusted to 25 mM HEPES pH 7.9, 150 mM KCl, 10% glycerol, 0.1% NP-40, 0.2 mM EDTA, 1 mM DTT, 10 mM NaF, 0.5 mM PMSF and 10 mg/ml approtinin. After a pre-clearing step with protein G-Sepharose (APB), the extracts were incubated with 4 µg/ml rabbit polyclonal CBP, (Santa-Cruz Biotechnology) or mouse monoclonal Flag antibody (Sigma) at 4°C, followed by adsorption to protein G-Sepharose. After extensive washing with the earlier mentioned buffer, the complexes were resuspended in SDS sample buffer, separated by 10% SDS–polyacrylamide gel electrophoresis, transferred to nitrocellulose membranes and immunostained with mouse monoclonal myc-epitope antiserum (Santa Cruz, Biotechnology). In vitro protein–protein interaction assays, 35S-labeled recombinant proteins were synthesized in vitro using the TNT coupled reticulocyte lysate system (Promega), in accordance with the manufacturer's instructions. An aliquot of 2 µg of GST-fusion proteins coupled to glutathione Sepharose column (APB) were incubated with the in vitro translated proteins in a buffer containing 20 mM HEPES pH 7.9, 200 mM NaCl, 5 mM MgCl2, 0.1% NP-40, 0.2% BSA, 10% glycerol, 1 mM PMSF and 10 µg/ml aprotinin, at 4°C with constant agitation. After excessive washing with the same buffer lacking BSA, the beads were resuspended in SDS-sample buffer and the proteins were separated by electrophoresis in SDS–polyacrylamide gels.


   ACKNOWLEDGEMENTS
 
We thank clinicians and fetopathologists from Debré and Saint-Antoine Hospital, particularly C. Baumann, and AL Delezoide, for their interest. We also thank Professor Grundfeld, D. Chauveau and M.-C. Gubler, for discussions. This work was supported by Association de la Recherche sur le Cancer (ARC) contracts no. 5824 and 3231, by CNRS, and INSERM (Institut National de la Santé et de la Recherche Medicale) to S.C., and by GSRT (01ED509) to I.T.


   FOOTNOTES
 
* To whom correspondence should be addressed at: Biologie du Développement, UMR 7622, CNRS, Université Pierre et Marie Curie, 9 quai St Bernard, Bâtiment C, case 24, 75252 Paris Cedex 05, France. Tel: +33 144272155; Fax: +33 144273497; Email: silvia.cereghini{at}snv.jussieu.fr


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 

  1. Odom, D.T., Zizlsperger, N., Gordon, D.B., Bell, G.W., Rinaldi, N.J., Murray, H.L., Volkert, T.L., Schreiber, J., Rolfe, P.A., Gifford, D.K. et al. (2004) Control of pancreas and liver gene expression by HNF transcription factors. Science, 303, 1311–1312.[Abstract/Free Full Text]

  2. Servitja, J.M. and Ferrer, J. (2004) Transcriptional networks controlling pancreatic development and beta cell function. Diabetologia, 47, 597–613.[CrossRef][ISI][Medline]

  3. Rey-Campos, J., Chouard, T., Yaniv, M. and Cereghini, S. (1991) vHNF1 is a homeoprotein that activates transcription and forms heterodimers with HNF1. EMBO J., 10, 1445–1457.[ISI][Medline]

  4. De Simone, V., De Magistris, L., Lazzaro, D., Gerstner, J., Monaci, P., Nicosia, A. and Cortese, R. (1991) LFB3, a heterodimer-forming homeoprotein of the LFB1 family, is expressed in specialized epithelia. EMBO J., 10, 1435–1443.[ISI][Medline]

  5. Cereghini, S. (1996) Liver-enriched transcription factors and hepatocyte differentiation. FASEB J., 10, 267–282.[Abstract]

  6. Toniatti, C., Monaci, P., Nicosia, A., Cortese, R. and Ciliberto, G. (1993) A bipartite activation domain is responsible for the activity of transcription factor HNF1/LFB1 in cells of hepatic and nonhepatic origin. DNA Cell Biol., 12, 199–208.[ISI][Medline]

  7. Sourdive, D.J., Chouard, T. and Yaniv, M. (1993) The HNF1 C-terminal domain contributes to transcriptional activity and modulates nuclear localisation. C. R. Acad. Sci. III, 316, 385–394.[Medline]

  8. Vaxillaire, M., Abderrahmani, A., Boutin, P., Bailleul, B., Froguel, P., Yaniv, M. and Pontoglio, M. (1999) Anatomy of a homeoprotein revealed by the analysis of human MODY3 mutations. J. Biol. Chem., 274, 35639–35646.[Abstract/Free Full Text]

  9. Soutoglou, E., Katrakili, N. and Talianidis, I. (2000) Acetylation regulates transcription factor activity at multiple levels. Mol. Cell, 5, 745–751.[CrossRef][ISI][Medline]

  10. Chi, Y.I., Frantz, J.D., Oh, B.C., Hansen, L., Dhe-Paganon, S. and Shoelson, S.E. (2002) Diabetes mutations delineate an atypical POU domain in HNF-1alpha. Mol. Cell, 10, 1129–1137.[CrossRef][ISI][Medline]

  11. Lindner, T.H., Njolstad, P.R., Horikawa, Y., Bostad, L., Bell, G.I. and Sovik, O. (1999) A novel syndrome of diabetes mellitus, renal dysfunction and genital malformation associated with a partial deletion of the pseudo-POU domain of hepatocyte nuclear factor-1beta. Hum. Mol. Genet., 8, 2001–2008.[Abstract/Free Full Text]

  12. Bingham, C., Ellard, S., Allen, L., Bulman, M., Shepherd, M., Frayling, T., Berry, P.J., Clark, P.M., Lindner, T., Bell, G.I. et al. (2000) Abnormal nephron development associated with a frameshift mutation in the transcription factor hepatocyte nuclear factor-1 beta. Kidney Int., 57, 898–907.[CrossRef][ISI][Medline]

  13. Kolatsi-Joannou, M., Bingham, C., Ellard, S., Bulman, M.P., Allen, L.I., Hattersley, A.T. and Woolf, A.S. (2001) Hepatocyte nuclear factor-1beta: a new kindred with renal cysts and diabetes and gene expression in normal human development. J. Am. Soc. Nephrol., 12, 2175–2180.[Abstract/Free Full Text]

  14. Iwasaki, N., Okabe, I., Momoi, M.Y., Ohashi, H., Ogata, M. and Iwamoto, Y. (2001) Splice site mutation in the hepatocyte nuclear factor-1 beta gene, IVS2nt+1G>A, associated with maturity-onset diabetes of the young, renal dysplasia and bicornuate uterus. Diabetologia, 44, 387–388.[CrossRef][ISI][Medline]

  15. Bohn, S., Thomas, H., Turan, G., Ellard, S., Bingham, C., Hattersley, A.T. and Ryffel, G.U. (2003) Distinct molecular and morphogenetic properties of mutations in the human HNF1beta gene that lead to defective kidney development. J. Am. Soc. Nephrol., 14, 2033–2041.[Abstract/Free Full Text]

  16. Ryffel, G. (2001) Mutations in the human genes encoding the transcription factors of the hepatocyte nuclear factor (HNF)1 and HNF4 families: functional and pathological consequences. Review. J. Mol. Endocrinol., 27, 11–29.[Abstract]

  17. Yamagata, K. (2003) Regulation of pancreatic beta-cell function by the HNF transcription network: lessons from maturity-onset diabetes of the young. Review. Endocr. J., 50, 491–499.[CrossRef][ISI][Medline]

  18. Barbacci, E., Reber, M., Ott, M., Breillat, C., Huetz, F. and Cereghini, S. (1999) Variant Hepatocyte Nuclear Factor 1 is required for visceral endoderm specification. Development, 126, 4795–4805.[Abstract]

  19. Coffinier, C., Thepot, D., Babinet, C., Yaniv, M. and Barra, J. (1999) Essential role for the homeoprotein vHNF1/HNF1beta in visceral endoderm differentiation. Development, 126, 4785–4794.[Abstract]

  20. Tomura, H., Nishigori, H., Sho, K., Yamagata, K., Inoue, I. and Takeda, J. (1999) Loss-of-function and dominant-negative mechanisms associated with hepatocyte nuclear factor-1beta mutations in familial type 2 diabetes mellitus. J. Biol. Chem., 274, 12975–12978.[Abstract/Free Full Text]

  21. Wild, W., Pogge von Strandmann, E., Nastos, A., Senkel, S., Lingott-Frieg, A., Bulman, M., Bingham, C., Ellard, S., Hattersley, A.T. and Ryffel, G.U. (2000) The mutated human gene encoding hepatocyte nuclear factor 1beta inhibits kidney formation in developing Xenopus embryos. Proc. Natl Acad. Sci. USA, 97, 4695–4700.[Abstract/Free Full Text]

  22. Yoshiuchi, I., Yamagata, K., Zhu, Q., Tamada, I., Takahashi, Y., Onigata, K., Takeda, J., Miyagawa, J. and Matsuzawa, Y. (2002) Identification of a gain-of-function mutation in the HNF-1beta gene in a Japanese family with MODY. Diabetologia, 45, 154–155.[CrossRef][ISI][Medline]

  23. Carbone, I., Cotellessa, M., Barella, C., Minetti, C., Ghiggeri, G.M., Caridi, G., Perfumo, F. and Lorini, R. (2002) A novel hepatocyte nuclear factor-1beta (MODY-5) gene mutation in an Italian family with renal dysfunctions and early-onset diabetes. Diabetologia, 45, 153–154.[CrossRef][ISI][Medline]

  24. Bellanne-Chantelot, C., Chauveau, D., Gautier, J., Dubois-Laforgue, D., Clauin, S., Beaufils, S., Wilhelm, J.M., Boitard, C., Noel, L.H., Velho, G. et al. (2004) Clinical spectrum associated with hepatocyte nuclear factor-1beta mutations. Ann. Intern. Med., 140, 510–517.[Abstract/Free Full Text]

  25. Harries, L.W., Ellard, S., Jones, R.W., Hattersley, A.T. and Bingham, C. (2004) Abnormal splicing of hepatocyte nuclear factor-1 beta in the renal cysts and diabetes syndrome. Diabetologia, 45, 937–943.

  26. Furuta, H., Furuta, M., Sanke, T., Ekawa, K., Hanabusa, T., Nishi, M., Sasaki, H. and Nanjo, K. (2002) Nonsense and missense mutations in the human hepatocyte nuclear factor-1 beta gene (TCF2) and their relation to type 2 diabetes in Japanese. J. Clin. Endocrinol. Metab., 87, 3859–3863.[Abstract/Free Full Text]

  27. Kitanaka, S., Miki, Y., Hayashi, Y. and Igarashi, T. (2004) Promoter-specific repression of hepatocyte nuclear factor (HNF)-1 beta and HNF-1 alpha transcriptional activity by an HNF-1 beta missense mutant associated with Type 5 maturity-onset diabetes of the young with hepatic and biliary manifestations. J. Clin. Endocrinol. Metab., 89, 1369–1378.[Abstract/Free Full Text]

  28. Rizzoni, G., Loirat, C., Levy, M., Milanesi, C., Zachello, G. and Mathieu, H. (1982) Familial hypoplastic glomerulocystic kidney. A new entity? Clin. Nephrol., 18, 263–268.[ISI][Medline]

  29. Horikawa, Y., Iwasaki, N., Hara, M., Furuta, H., Hinokio, Y., Cockburn, B.N., Lindner, T., Yamagata, K., Ogata, M., Tomonaga, O. et al. (1997) Mutation in hepatocyte nuclear factor-1 beta gene (TCF2) associated with MODY. Nat. Genet., 17, 384–385.[CrossRef][ISI][Medline]

  30. Montoli, A., Colussi, G., Massa, O., Caccia, R., Rizzoni, G., Civati, G. and Barbetti, F. (2002) Renal cysts and diabetes syndrome linked to mutations of the hepatocyte nuclear factor-1 beta gene: description of a new family with associated liver involvement. Am. J. Kidney Dis., 40, 397–402.[CrossRef][ISI]

  31. Hiesberger, T., Bai, Y., Shao, X., McNally, B.T., Sinclair, A.M., Tian, X., Somlo, S. and Igarashi, P. (2004) Mutation of hepatocyte nuclear factor-1beta inhibits Pkhd1 gene expression and produces renal cysts in mice. J. Clin. Invest., 113, 814–825.[CrossRef][ISI][Medline]

  32. Zhong, W., Mirkovitch, J. and Darnell, J.E.J. (1994) Tissue-specific regulation of mouse hepatocyte nuclear factor 4 expression. Mol. Cell. Biol., 14, 7276–84.[Abstract/Free Full Text]

  33. Cereghini, S., Blumenfeld, M. and Yaniv, M. (1988) A liver-specific factor essential for albumin transcription differs between differentiated and dedifferentiated rat hepatoma cells. Genes Dev., 2, 957–974.[Abstract/Free Full Text]

  34. Feuerman, M.H., Godbout, R., Ingram, R.S. and Tilghman, S.M. (1989) Tissue-specific transcription of the mouse alpha-fetoprotein gene promoter is dependent on HNF-1. Mol. Cell. Biol., 9, 4204–4212.[Abstract/Free Full Text]

  35. Haumaitre, C., Reber, M. and Cereghini, S. (2003) Functions of HNF1 family members in differentiation of the visceral endoderm cell lineage. J. Biol. Chem., 278, 40933–40942.[Abstract/Free Full Text]

  36. Soutoglou, E., Violet, B., Vaxillaire, M., Yaniv, M., Pontoglio, M. and Talianidis, I. (2001) Transcription factor-dependent regulation of CBP and P/CAF histone acetyltransferase activity. EMBO J., 20, 1984–1992.[CrossRef][ISI][Medline]

  37. Le Hir, H., Nott, A. and Moore, M.J. (2003) How introns influence and enhance eukaryotic gene expression. Trends Biochem. Sci., 28, 215–220.[CrossRef][ISI][Medline]

  38. Inoue, K., Khajavi, M., Ohyama, T., Hirabayashi, S.I., Wilson, J., Reggin, J.D., Mancias, P., Butler, I.J., Wilkinson, M.F., Wegner, M. et al. (2004) Molecular mechanism for distinct neurological phenotypes conveyed by allelic truncating mutations. Nat. Genet., 36, 361–369.[CrossRef][ISI][Medline]

  39. Ktistaki, E. and Talianidis, I. (1997) Modulation of hepatic gene expression by hepatocyte nuclear factor 1. Science, 277, 109–112.[Abstract/Free Full Text]

  40. Power, S.C. and Cereghini, S. (1996) Positive regulation of the vHNF1 promoter by the orphan receptors COUP-TF1/Ear3 and COUP-TFII/Arp1. Mol. Cell. Biol., 16, 778–791.[Abstract]

  41. Cereghini, S., Ott, M.O., Power, S. and Maury, M. (1992) Expression patterns of vHNF1 and HNF1 homeoproteins in early postimplantation embryos suggest distinct and sequential developmental roles. Development, 116, 783–797.[Abstract]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Endocr. Rev.Home page
L. Aguilar-Bryan and J. Bryan
Neonatal Diabetes Mellitus
Endocr. Rev., May 1, 2008; 29(3): 265 - 291.
[Abstract] [Full Text]