Functional characterization of hepatocyte nuclear factor-4a dimerization interface mutants Eleni Aggelidou*, Panagiota Iordanidou*, Constantinos Demetriades, Olga Piltsi and Margarita Hadzopoulou-Cladaras Department of Genetics, Development and Molecular Biology, Laboratory of Developmental Biology, School of Biology, Aristotle University of Thessaloniki, Greece Hepatocyte nuclear factor-4a (HNF-4a) is a transcript- ion factor that belongs to the superfamily of nuclear receptors. Like other members of the family, HNF-4a is characterized by the presence of two well-conserved functional domains, the DNA-binding domain (DBD) consisting of two zinc-finger motifs that specifically bind to DR1 elements found in promoters of target genes, and the ligand-binding domain (LBD), consist- ing of a ligand-binding site, dimerization domain and hydrophobic AF-2 transactivation domain [1]. HNF-4a interacts with regulatory elements in pro- moters of genes whose products are involved in diverse functions such as cholesterol, fatty acid, amino acid and glucose metabolism, as well as liver development and differentiation [2–7]. Mutations in the HNF-4a coding region have been directly linked to the disease maturity-onset diabetes of the young (MODY)-1 [8] and one mutation in the HNF-4a binding site of the HNF-1a gene has been linked to MODY-3 [9]. Keywords coactivator; dimerization; HNF-4; nuclear receptor; transcription Correspondence M. Hadzopoulou-Cladaras, Department of Genetics, Development and Molecular Biology, Laboratory of Developmental Biology, School of Biology, Aristotle University of Thessaloniki, Thessaloniki, 54124, Greece Fax: +30 2310 998298 Tel: +30 2310 998303 E-mail: cladaras@bio.auth.gr *These authors contributed equally to this study. (Received 31 August 2005, revised 21 February 2006, accepted 2 March 2006) doi:10.1111/j.1742-4658.2006.05208.x Hepatocyte nuclear factor-4 (HNF-4a), a member of the nuclear receptor superfamily, binds DNA exclusively as a homodimer. Dimerization con- trols important aspects of receptor function, such as DNA binding, protein stability, ligand binding and interaction with coactivators. Crystallographic data of the HNF-4a ligand-binding domain (LBD) demonstrated that the homodimer interface is composed of residues in helices 7, 9 and 10 with intermolecular salt bridges, hydrogen bonds and hydrophobic interactions contributing to the stability of the interface. To investigate the importance of the proposed ionic interactions for HNF-4a dimerization, interactions critical for formation of the LBD homodimer interface were disrupted by introducing point mutations in residues D261N (H7), E269Q (H7), Q307L (H9), D312N (H9) and Q336L (H10). Mutants were analysed for transacti- vation, coactivator interaction, DNA binding and dimerization. EMSA analysis showed that the mutants are able to bind DNA as dimers and coimmunoprecipitation assays confirmed dimerization in solution. Further- more, the mutations do not compromise HNF-4a activity and are respon- sive to PPAR-gamma coactivator-1 (PGC-1). Finally, residue R324, located in the H9 ⁄ H10 loop, which was suspected to be involved in dimer stabilization via an ionic interaction with residue E276, was studied. In contrast to the conservative substitution R324H the mutation R324L abol- ishes HNF-4a transcriptional activity and coactivator recruitment, reveal- ing that the nature of substitution may play an important role in HNF-4a function. Abbreviations AF, activation function; b-gal, b-galactosidase; DBD, DNA-binding domain; DMEM, Dulbecco’s modified Eagle’s medium; HNF-4, hepatocyte nuclear factor-4; LBD, ligand-binding domain; MODY, maturity-onset diabetes of the young; NR, nuclear receptor; PGC-1, PPAR-gamma coactivator-1; PMSF, phenylmethylsulfonyl fluoride; RAR, retinoic acid receptor; RXR, retinoid X receptor. 1948 FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS HNF-4a tends to bind DNA exclusively as a homodimer. Two regions are responsible for dimeriza- tion: the DBD, responsible for dimerization on DNA; and the LBD, responsible for dimerization in solution [10]. Crystallographic data of the HNF-4a LBD showed that it crystallizes as a canonical homodimer with near perfect two fold symmetry about the inter- face and revealed that the LBD dimerization interface is made up of residues in helices 7, 9 and 10 [11]. More specifically, a coiled coil interaction between helix 10 from each molecule dominates the interface, with hydrophobic side-chain⁄ side-chain interactions (F325– F325, L329–L329, L330–I283, L332–L330, P333–P333 and W340–W340) along its length. Intermolecular salt bridges (E269–K300, R303–E327 and R322–D312) and hydrogen bonds (Q307–E327, E311–G323 and Q336– Q336) appear to contribute to the stability of the inter- face. In addition, acidic residues D261 and E262 create a bulge in helix 5 that may be important for dimeriza- tion [11]. In agreement, the HNF-4a LBD crystallo- graphy revealed a homodimer interface that includes specific side-chain ⁄ side-chain interactions, with hydro- gen bonds between Q267 and E287 (Q307 and E327 in HNF-4a) and between Q296 and Q296 (Q336 in HNF-4a), as well as salt bridges between E229 and K260 (E269 and K300 in HNF-4a) [12]. Similarly, the amino acid residue L419 in retinoid X receptor (RXRa; L329 in HNF-4a) was found to be important for the RXRa dimer interface [13] and the mutant V361G in retinoic acid receptor (RARa; D312 in HNF-4a) was found to diminish strongly homo- and heterodimerization of this molecule [14]. Protein dimerization affects receptor function by influencing DNA and ligand binding, protein stability and interaction with coregulatory molecules. Although many mutational studies have been conducted in the LBD of HNF-4a [1,10,15,16], all the critical determi- nants of receptor dimerization have not yet been char- acterized. In this study, we mutated amino acid residues in helices 7, 9 and 10 that were shown to com- prise the LBD dimerization interface. In particular, mutations D261N in H5, E269Q in the H7 ⁄ H8 loop, Q307L and D312N in H9 and Q336L in H10, were investigated for their behaviour concerning DNA-bind- ing properties, protein dimerization ability, transcrip- tional activation and recruitment of coactivators. In addition, another missense mutation, R324L in the H9 ⁄ H10 loop, was examined. The naturally occurring mutation R324H was found in a type-2 diabetic nephr- opathic patient [17]. Furthermore, R324 has been shown to form an ionic interaction with MODY-1- associated residue E276 [11]. These residues are not directly involved in the dimer interface but help tether the 9 ⁄ 10 loops, thus it is likely that loss of interactions between E276 and R324 would destabilize dimeriza- tion. In this respect, we found it interesting to study a more drastic mutation, R324L, in contrast to the con- servative mutation already studied. We found that point mutations of the residues located in helices H7, H9 and H10 do not affect either dimerization or DNA binding. Furthermore, transcriptional analysis revealed that the mutations D261N, E269Q, Q307L, D312N and Q336L do not impair HNF-4a activation or PGC- 1 coactivation potential, whereas mutagenesis of R324 affects both constitutive and coactivator-stimulated activity. Results In order to investigate the role of LBD residues pro- posed to be important for dimerization of HNF-4a we introduced mutations into amino acids present in helices 7, 9 and 10, namely D261N, E269Q, Q307L, D312N and Q336L. Based on the crystallographic data, those residues were predicted to participate in the homodimer interface of HNF-4a contributing to the stability of the interface [11]. The assignment of helices H3–H12 based on the crystal structure of HNF-4a, together with the point mutations of amino acids impli- cated in receptor dimerization are shown in Fig. 1. Mutants were first examined for their ability to bind DNA. An EMSA analysis was performed, using the CIIIB element of the apolipoprotein CIII promoter as a probe, which is a high-affinity binding site for HNF- 4a. The results depicted in Fig. 2A show that mutants D261N, E269Q, Q307L, D312N and Q336L bind strongly to the CIIIB site, suggesting that, similarly to the wild-type, all mutants bind DNA as homodimers. Western blot analysis revealed the equal expression of both wild-type and mutated proteins (Fig. 2B). Furthermore, it was interesting to investigate the hypothesis that the side chain from residue R324 would play a role in dimerization. This residue forms a hydrogen bond with residue E276 and it was previ- ously proposed that R324 substitution is likely to destabilize dimerization as a result of loss of interac- tions between E276 and R324 [11]. The results from the EMSA analysis demonstrate that mutation R324L does not affect DNA binding (Fig. 2A). To investigate the dimerization properties of the mutants, nuclear extracts from transfected cells expres- sing CD1b were included in the binding assays. CD1b is a truncated HNF-4a protein that lacks the F region and was previously shown to retain wild-type binding and dimerization properties [1]. The results shown in Fig. 3 indicate the formation of heterodimeric E. Aggelidou et al. HNF-4 dimerization FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS 1949 complexes, which are detected as shifted complexes of intermediate mobility between those of the point mutant and the CD1b homodimers, confirming the ability of mutants D261N, E269Q, Q307L, D312N, Q336L and R324L to heterodimerize with the wild-type. Although the mutants D261N, E269Q, Q307L, D312N, Q336L and R324L were shown to heterodi- merize efficiently, a subtle change was noticed in dime- rization of the mutants D261N and Q307L (Fig. 3). To evidence possibly impaired heterodimerization we performed EMSA experiments in less-favourable con- ditions, by increasing the ionic strength of the binding buffer from 300 to 600 mm KCl. The mutants formed an intermediate complex even at 600 mm KCl (Fig. 4), confirming that mutations D261N and Q307L do not cause any effect in dimerization. To further confirm the dimerization potential of the HNF-4a substitution mutations, we examined whether A Q336L D261N E269Q Q307L HEK-293 W T D312N R324L Free probe HNF-4α B R324L WB: anti-HNF-4α (C-terminal) D261N E269Q Q307L W T D312N Q336L 62 47.5 HNF-4α Fig. 2. DNA binding of wild-type and mutated in helices 7, 9 and 10, HNF-4a proteins. (A) EMSA analysis of DNA binding of nuclear extracts from HEK293 cells, which were transfected with wild-type HNF-4a and point mutants (upper), using the 32 P-labelled double- stranded oligonucleotide CIIIB as a probe. Protein–DNA complexes were analysed by electrophoresis in a 5% nondenaturing gel, fol- lowed by autoradiography. As control either the probe alone or nuclear extracts from untransfected HEK293 cells was used. (B) Detection of the expression of wild-type and mutated in helices 7, 9 and 10, HNF-4a proteins by western blot analysis. HEK293 cells were transfected with the indicated expression plasmids (18 lg). Nuclear extracts were analysed by 10% SDS ⁄ PAGE. The expres- sion of HNF-4a and mutant proteins was detected by using a goat anti-HNF-4a polyclonal serum. Numbers indicate molecular mass protein markers in kDa. D261N E269Q Q307L WT D312N R324L Q336L CD1b CD1b heterodimer HNF-4α CD1b ++++++++ Fig. 3. Dimerization between wild-type or mutated in helices 7, 9 and 10, HNF-4a proteins and its deletion mutant, CD1b. EMSA ana- lysis of heterodimers formed by CD1b, a truncated HNF-4a protein that retains DNA binding and dimerization properties and LBD point mutants (upper). Nuclear extracts from HEK293 cells were cotrans- fected with wild-type HNF-4a or LBD point mutants and CD1b, or CD1b alone, using the 32 P-labelled double-stranded oligonucleotide CIIIB as a probe. Heterodimers were analysed by electrophoresis in a 5% nondenaturing gel, followed by autoradiography. Upper and lower arrows indicate the formation of HNF-4a or CD1b homo- dimers, respectively. Middle arrow indicates heterodimer formation. ITDVCESMKEQLLVLVEWAKYIPAFCELLLDDQVALLRAHAGEHLLLGATKRSMVFKDVLLLGNDYIVPR HCPELAEMSRVSIRILDELVLPFQELQIDDNEYACLKAIIF FDPDAKGLS DPGKIKRLRSQVQVSLEDYI NDRQYDSRGRFGELLLLLPTLQSITWQMIEQIQFIKLFGMAKIDNLLQEMLLGGSP H3 S1 S2 H6 H7 H8 H9 H10 H12 β β -turn H11 R 3 2 4 L D 2 6 1 N E 2 6 9 Q Q 3 0 7 L D 3 1 2 N Q 3 3 6 L H4 H5 Fig. 1. Schematic representation of the sec- ondary structure of the LBD of HNF-4a (heli- ces H3–H12). The position of the a helices and the corresponding point mutations at the dimer interface are indicated. HNF-4 dimerization E. Aggelidou et al. 1950 FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS the above mutants can dimerize in solution. Thus, coimmunoprecipitation assays were performed in HEK293 cells that were transiently transfected with equimolar amounts of expression vectors for full- length HNF-4a LBD point mutants and the HNF-4a C-terminal deletion mutant, CD1b. The extracts were subjected to western blot analysis and coexpression of full-length and truncated HNF-4a proteins was verified (Fig. 5A). Protein detection was performed using a specific antibody for the N-terminus of HNF-4a [anti- (N-terminal HNF-4a)], which recognizes both the full- length and truncated CD1b protein. Next, we sought to determine whether the two proteins are able to interact and heterodimerize in solution, forming HNF-4a ⁄ CD1b complexes in the nuclear extracts. All HNF-4a point mutant proteins expressed in HEK293 cells were immunoprecipitated by an antibody specific for the C-terminus of HNF-4a [anti-(C-terminal HNF-4a)] that recognizes the LBD point mutants, but not CD1b, and were subsequently immunoblotted with the anti-(N-terminal HNF-4a) serum. The rationale was that, if both LBD point mutants and CD1b could be detected in the precipitates by western blot analysis using the anti-(N-terminal HNF-4a) serum that would be because of a specific interaction in solution between the two proteins. The results, shown in Fig. 5B, indi- cate that all point mutants are able to form hetero- dimers with wild-type HNF-4a in solution, because CD1b was sufficiently coimmunoprecipitated with all HNF-4a LBD point mutants. Furthermore, the anti- (C-terminal HNF-4a) serum failed to immunoprecipi- tate the CD1b protein confirming the specificity of the protein–protein interactions (Fig. 5C). Based on the fact that the LBD is a multifunctional region, important not only for dimerization, but also for ligand binding and transactivation, we investigated WT D261N Q307L 300 600 300 600 300 600 KCl (m M) CD1b heterodimer HNF-4α Fig. 4. Dimerization properties of wild-type and HNF-4a point mutants D261N and Q307L performed in increasing ionic strength conditions. EMSA analysis was performed in 300 or 600 m M KCl in order to examine dimerization of mutants. D 2 6 1 N E 2 6 9 Q Q 30 7 L W T D 31 2 N R 32 4 L Q 33 6 L Input HNF-4α HNF-4α CD1b WB: anti-HNF-4α (N-terminal) CD1b ++++++ + A 47.5 32.5 D261N E 269Q Q307L W T D312N R324L Q336L 47.5 32.5 HNF-4α CD1b IP: anti-HNF-4α (C-terminal) WB: anti-HNF-4α (N-terminal) IP: anti-HNF-4α (C-terminal) WB: anti-HNF-4α (N-terminal) CD1b ++++ ++ + B D261N CD1b D261N CD1b PS -+-+ HNF-4α CD1b C Fig. 5. Dimerization in solution, by coimmunoprecipitation assays, of WT HNF-4a or its point mutants in helices 7, 9 and 10, with a deletion HNF-4a mutant. (A) HNF-4a protein input was taken into account for coimmunoprecipitation assays. HEK293 cells were cotransfected with the indicated pcDNA3.1-LBD expression plas- mids (4 lg) and the deletion mutant CD1b (4 lg). Nuclear extracts were analysed by 8% SDS ⁄ PAGE. The coexpression of HNF-4a point and deletion mutant proteins was detected by using an N-terminal rabbit anti-HNF-4a polyclonal serum. Numbers indicate molecular mass protein markers in kDa. (B) HEK293 cells were cotransfected with the indicated plasmids expressing point mutants and the C-terminal HNF-4a deletion mutant, CD1b, in equimolar quantities (4 lg). The antibody used for coimmunoprecipitation was the C-terminal goat HNF-4a, which recognizes the HNF-4a point mutants but not the C-terminal deletion mutant, CD1b. Immunopre- cipitated proteins were analysed by 8% SDS ⁄ PAGE. The coexpres- sion of HNF-4a point mutants and CD1b was detected by using the N-terminal rabbit HNF-4a antibody, which is able to detect both full- length and truncated C-terminal HNF-4a constructs. Numbers indi- cate molecular mass protein markers in kDa. (C) Control protein incubated with agarose-protein G beads but without antibody both in supernatant (S) and pellet (P) and cells transfected with CD1b alone, in supernatant (S) and pellet (P). E. Aggelidou et al. HNF-4 dimerization FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS 1951 the effect of point mutations on HNF-4a transcrip- tional activity. Transient transfection experiments were performed in order to test whether mutants D261N, E269Q, Q307L, D312N, Q336L and R324L are able to activate transcription from the apoCIII promoter. As shown in Fig. 6, HNF-4a transcriptional activity is unaffected by mutants D261N, E269Q, Q307L, D312N and Q336L, whereas it is abolished in the case of mutant R324L. The loss of activation was not due to lower protein expression, as monitored by western blot analysis, suggesting that this mutation causes a partial loss of HNF-4a function. Furthermore, it was interest- ing to examine whether the mutations influence the ability of the coactivator PGC-1 to enhance transcrip- tion. It is known that PGC-1 forms a complex with HNF-4a enhancing transcriptional activation [18,19] and we sought to determine whether mutagenesis in the particular amino acid residues has an effect on HNF-4a transactivation potential in the presence of PGC-1. In agreement with the abovementioned results, mutants D261N, E269Q, Q307L, D312N and Q336L that were not shown to affect HNF-4a activity retained the ability to further increase transcriptional activity in the presence of PGC-1 [Fig. 6]. In the case of mutation of the MODY-1-implicated residue, R324L, it was observed that the presence of PGC-1 coactivator failed to restore transactivation potential, in contrast to the other LBD residues studied. The drastic reduction in transcriptional activity observed in the presence of point mutant R324L is in contrast to the results obtained in a previous study in which the mutation R324H was examined [20]. In this context, we found it challenging to investigate in more detail the effect of mutagenesis of the particular resi- due under the same conditions. It was found that replacement of arginine at position 324 with leucine eliminates both transcriptional activity and PGC-1 coactivation potential, whereas replacement with histi- dine does not affect HNF-4a activity (Fig. 7), as shown previously [20]. The differential effect in HNF-4a transcriptional activity caused by substitution mutations of R324 resi- due prompted us to determine whether these mutants behave differently in physical interactions with the coactivator PGC-1. As seen in Fig. 8, mutant R324L and mutant R324H retain the ability to physically interact with coactivator PGC-1, much like the wild- type. Thus, despite the fact that mutation R324L 0 100 200 300 400 500 600 700 800 900 Relative CAT activity (%) to wild-type - PGC-1 + PGC-1 D261N E269Q Q307L WT D312N R324L Q336L PGC-1 - +-+-+ -+- +-+-+ -890 HNF-4 Enhancer Promoter +24 ApoCIII (-890/+24) CAT HNF-4 CAT COS-7 Fig. 6. Effects of HNF-4a point mutants in helices 7, 9 and 10 on the enhancement of wild-type transcriptional activity by PGC-1. The reporter construct apoCIII()890 ⁄ +24)chloramphenicol acetyltrans- ferase is shown at the top of the figure. COS-7 cells were trans- iently transfected with 0.2 lg wild-type or mutated HNF-4a expression vector and 1.25 lg of empty vector or PGC-1. The relat- ive chloramphenicol acetyltransferase activity (± SEM) of three independent experiments is shown as the percentage of the activ- ity obtained with the apoCIII-890chloramphenicol acetyltransferase reporter construct cotransfected with pcDNA3.1-LBD WT HNF-4a. CAT -890 HNF-4 Enhancer Promoter +24 ApoCIII (-890/+24) CAT HNF-4 W T PGC-1 -+ -+ +- 0 200 400 600 800 1000 1200 1400 1600 - PGC-1 + PGC-1 R324H R324L Relative CAT activity (%) COS-7 Fig. 7. Transcriptional activity of HNF-4a point mutants R324L and R324H and their effect on the enhancement of wild-type transcript- ional activity by PGC-1. The reporter construct apoCIII()890 ⁄ +24)chloramphenicol acetyltransferase is shown at the top of the figure. COS-7 cells were transiently transfected with 0.2 lg wild- type or mutated HNF-4a expression vector and 1.25 lg empty vector or PGC-1. The relative chloramphenicol acetyltransferase activity (± SEM) of three independent experiments is shown as the percentage of the activity obtained with the apoCIII-890chloram- phenicol acetyltransferase reporter construct cotransfected with pcDNA3.1-LBD WT HNF-4a. HNF-4 dimerization E. Aggelidou et al. 1952 FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS eliminates transcriptional enhancement by PGC-1 a protein–protein interaction between the mutant and the coactivator is observed. Discussion In contrast to other nuclear receptors that tend to bind DNA either as heterodimers or homodimers, HNF-4a is known to be an obligate homodimer [21]. In agree- ment with the interface seen in other nuclear receptors, crystal structures of the HNF-4a and c LBD regions revealed that the LBD dimerization interface is made up of residues in helices 7, 9 and 10 [11,12]. Although protein dimerization controls many important aspects of receptor function, all the critical residues responsible for HNF-4a dimerization have not been identified. We performed site-directed mutagenesis studies in amino acid residues that were shown to participate in a num- ber of side-chain interactions important for the stabil- ity of the dimer interface in order to examine their potential impact on protein dimerization. We observed that mutations in residues that parti- cipate either in hydrogen bonds or salt bridges, in particular, amino acids E269, Q307, D312 and Q336, do not affect DNA binding and dimerization proper- ties. Indeed, crystallographic data reveal that salt bridges and hydrogen bonds are important for part- ner selectivity, whereas the key role in stabilization of the HNF)4 dimer interface is played by the hydro- phobic side-chain interactions between helices 10 from each molecule [11,12]. In more detail, in HNF-4a, salt bridges are formed between E269–K300 and D312– R322, with the equivalent salt bridges in RXRa being D359–E390 and A402–P412 [22]. Formation of an HNF-4 ⁄ RXR heterodimer is prevented by a charge incompatibility that would create an unfavourable pairing in the case of E269 in HNF-4a and E390 in RXRa or D312 and P412, respectively [10]. The above observations, in combination with our experi- mental data, lead to the conclusion that the particular salt bridge may have a sole role in conferring partner specificity without affecting other properties, such as DNA binding. In addition, in the case of the hydrogen bond between Q307 and E327, the importance of residue E327 has already been studied [10]. In detail, when the charge of residues K300 and E327 was inverted, according to the corresponding residues in RXRa (E390 and K417, respectively), the double mutant did not heterodimerize with wild-type HNF-4a [10]. In our study, mutation Q307L did not have an analogous effect, highlighting the importance of E327 rather than Q307 in dimer formation, via the salt bridge that E327 forms with K300. The above observations suggest that specific residues may participate simultaneously in dif- ferent interactions, with a gradient of importance in dimer formation. It has been reported that in the crystal structure of the RXRa LBD helix 7 adopts an unusual geometry that gives rise to the formation of a series of inter- and intramolecular hydrogen bonds, which improve RXRa input HNF-4α α WB: anti-HNF-4α R 3 2 4 H W T R 3 24 L 62 47.5 HNF-4α - B irA bound Bio-PGC-1 WB: streptavidin-HRP R3 2 4 H W T R3 24 L Bio-PGC-1 83 175 - HNF-4 α - Bi rA HNF-4α R3 24 H W T R 3 2 4 L 62 47.5 bound HNF-4α WB:anti-HNF-4α - H N F-4 α - Bir A Fig. 8. Physical interactions of wild-type HNF-4a and its point mutants. The expression of HNF-4 WT and its point mutants trans- fected in HEK293 cells was monitored by western blot analysis using anti-HNF-4a serum. The nuclear extracts expressing HNF-4a, Bio-PGC-1 and pEV-BirA ligase were incubated with streptavidin beads. The bound proteins were eluted and examined by SDS ⁄ PAGE and western blot analysis using streptavidin–horserad- ish peroxidase to detect bound biotinylated PGC-1. The same mem- brane was stripped and reprobed with anti-HNF-4a serum and physical interactions of wild-type HNF-4a or its point mutants with Bio-PGC-1 were detected. E. Aggelidou et al. HNF-4 dimerization FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS 1953 LBD homodimerization [23]. The crystal structure of HNF-4a confirmed the presence of a bulge in H7, with the acidic residues D261 and E262 possibly having an important part in dimerization [11]. To investigate this hypothesis further, the amino acid residue D261 was mutated, because E262 has already been studied [16]. Our study revealed that mutation D261N did not influ- ence either DNA binding or dimerization suggesting that in helix 7 residue E262 is the key residue for the function of nuclear receptors in accordance with RXRa, where the homologous residue E352 is involved in a charge driven interaction with R348, which is crucial for the dimerization interface [23]. As mentioned above, the mutants were able to bind DNA either as homodimers or as heterodimers with wild-type HNF-4a. In order to investigate whether the ability to dimerize is consistently observed both ex vivo and in vitro we performed coimmunoprecipitat- ion assays, where it was verified that all mutants dimerize efficiently with the truncated HNF-4a mutant CD1b. Failure to detect loss of DNA binding or dimeriza- tion in the above point mutants does not exclude that these mutations have some consequence in the tran- scriptional activity of HNF-4a. This suggestion is based on the fact that the LBD domain is functionally complex because it is important not only for dimeriza- tion, but also for ligand binding and ligand-dependent transactivation. We have previously shown that muta- tions in amino acid residues, which line the LBD pocket and come in contact with the ligand, impair HNF-4a transcriptional activity [15]. Furthermore, we demonstrated that distinct residues in the LBD pocket might be involved in coactivator and ligand interac- tions, although some residues are critical for both functions [19]. In this context, we found it interesting to examine in detail the effect of the point mutations under study in transcriptional activity both in the absence or presence of the coactivator PGC-1. The transient transfection experiments revealed that the mutations do not affect HNF-4a activity or tran- scriptional enhancement by coactivator. Finally, we wanted to investigate the hypothesis that mutation in residue R324, which forms an ionic inter- action with E276, could affect dimerization, in light of the finding that this interaction helps tether the H9 ⁄ H10 loops of the two molecules in the homodimer crystal [11]. In order to explore the effect of a muta- tion in the above residue on the dimerization and transactivation properties of the HNF-4a protein, we chose to substitute arginine for leucine, bearing in mind that in a previous study, a naturally occurring mutation in the same residue, R324H, which was found in a type-2 diabetic patient, was shown to have no effect on HNF-4a-mediated activation of the ApoCIII promoter [20]. Although R324L appeared to form homodimers and heterodimerize with the wild- type efficiently, it eliminated transcription from the ApoCIII promoter in transfection experiments and the addition of PGC-1 coactivator failed to restore the mutation’s effect. In contrast, the wild-type transcrip- tional activity was unaffected when mutant R324H was examined and this activity was further stimulated by the presence of PGC-1. The observed difference in activity of R324L may be due to the drastic nature of this amino acid substitution compared with the conser- vative R324H mutation. In order to explain the functional data, we investi- gated the effects of point mutants R324L and R324H on the physical interaction with the coactivator PGC- 1. As seen in Fig. 7, the mutant R324L severely affec- ted transactivation and failed to be rescued in the presence of coactivator. This effect would not have been surprising, because reduced transcriptional activ- ity observed in certain mutated LBD residues can be rescued by coactivators, provided that these muta- tions do not disturb the coactivator interaction inter- face [19]. Interestingly, the protein–protein interaction experiments showed that mutant R324L retains a sub- stantial degree of interaction with PGC-1, in a similar manner as mutant R324H. Failure to detect a subtle loss of physical interaction between mutant R342L and PGC-1 does not exclude that this mutation may influence coactivator stimulated activity via another mechanism. In summary, we investigated the importance of ionic interactions to the formation of the HNF)4a homodimer interface, which was suggested by the LBD crystal structure. Our data indicate that the par- ticular single point mutations in residues participating in hydrogen bonds and salt-bridge interactions were not critical for stabilization of the interface. It is possible to speculate that simultaneous mutation of more than one residue of the one that were studied and were involved in ionic interactions may be required to disturb formation of the homodimer inter- face. By contrast, our results in combination with previous studies suggest that specific residues of one monomer may interact with more than one residue of the other monomer, with each interaction contribu- ting to the specificity of partner selection to a varying degree. Finally, it would be interesting to investigate the role of hydrophobic residues in the dimerization of HNF-4a, because the crystal structure revealed that the dimer interface is dominated by hydrophobic interactions. HNF-4 dimerization E. Aggelidou et al. 1954 FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS Experimental procedures Plasmid constructions All HNF-4a LBD mutants were generated by PCR-medi- ated site-directed mutagenesis as described previously [15], using appropriate primers and the rat HNF-4a cDNA as template. To construct the desirable mutants, the internal primers presented in Table 1 were used along with the N- and C-terminal primers, HNF-N and HNF-C. The PCR-amplified fragments were cloned into the vector pcDNA3.1[+] (Invitrogen, Carlsbad, CA) at the HindIII and BamHI sites. All mutations were verified by DNA sequencing analysis. The apoCIII[-890 ⁄ +24] chlorampheni- col acetyltransferase plasmid, containing the natural apo- lipoprotein CIII promoter [24], was used as reporter to assay the degree of transactivation. The pEV-BirA and pcDNA3.1-Bio plasmids, encoding the bacterial biotin ligase BirA and the 23 amino acid bioti- nylation epitope, were generous gifts of J. Strouboulis (Erasmus University Medical Center, the Netherlands) and D. Kardassis (University of Crete), respectively. The Bio- PGC-1 plasmid was created as described previously [19]. The resulting Bio-PGC-1-expressing plasmid was used in cell transfections for in vivo copurification assays. Cell transfections and chloramphenicol acetyltransferase assays COS-7 and HEK293 cells were maintained as stocks in Dulbecco’s modified Eagle’s medium (DMEM) supplemen- ted with 10% fetal bovine serum. 50–60% confluent 30 mm dishes were transfected using the calcium phosphate copre- cipitation method [25]. Plasmids were transfected into COS- 7 cells and assayed for their ability to promote transcription of the chloramphenicol acetyltransferase (CAT) gene. The transfection mixture contained 3 lg of the chlorampheni- col acetyltransferase reporter plasmid, 200 ng of the pcDNA3.1-HNF-4a wild-type or mutant plasmids, 1 lgof CMV b-gal plasmid and, in coactivator interaction experi- ments, 1.25 lg of PGC-1 plasmid. In each case, vector DNA was added as necessary to achieve a constant amount of transfected DNA (5.45 lg). Forty hours post transfec- tion, cells were washed with NaCl ⁄ P i and collected in TEN solution (0.04 m Tris ⁄ HCl, pH 7.8, 1 mm EDTA, pH 8.0, 0.15 m NaCl). Whole-cell extracts were prepared in 0.25 m Tris ⁄ HCl, pH 7.8, by three sequential freeze–thaw cycles. The b-galactosidase activity of cell lysates was determined as described previously [25], and the values obtained were used to normalize variability in the efficiency of transfec- tion. Chloramphenicol acetyltransferase activity was deter- mined using 14 C-chloramphenicol and acetyl-CoA as previously described [25]. Chloramphenicol acetyltrans- ferase enzyme levels that exhibited > 60% conversion of acetylated product were diluted and reassayed for chloram- phenicol acetyltransferase activity in the linear range. The results represent the mean of at least three independent transfection experiments, each carried out in duplicate. Coactivator plasmid pcDNA3.1-HA-PGC1 was a kind gift from A. Kralli (University of Basel, Switzerland). Electrophoretic mobility shift assay The various pcDNA3.1–LBD mutants were transfected into HEK293 cells, and nuclear extracts were isolated and used to evaluate the DNA binding and dimerization properties of all mutants. Cells were resuspended in hypotonic buffer A (10 mm Hepes pH 7.9, 10 mm KCl, 1.5 mm MgCl 2 , 0.5 mm dithiothreitol), supplemented with protease inhibi- tors (300 lm phenylmethylsulfonyl fluoride (PMSF), 200 lm leupeptin), and incubated on ice for 10 min, fol- lowed by homogenization and centrifugation (10 000 g, 5 min, 4 °C). The pellets were resuspended in nuclear extraction buffer B + (20 mm Hepes pH 7.9, 1.5 mm MgCl 2 , 20% v ⁄ v glycerol, 0.5 mm dithriothreitol, 0.3 m KCl), sup- plemented with protease inhibitors (300 lm PMSF, 200 lm leupeptin), and nuclei were further extracted on ice for 30 min with gentle mixing. Following centrifugation (10 000 g , 5 min, 4 °C), the supernatant was collected and 2 vol. of buffer B – (20 mm Hepes pH 7.9, 1.5 mm MgCl 2 , 20% v ⁄ v glycerol, 0.5 mm dithriothreitol, 0.1 m KCl) were added. A double-stranded oligonucleotide corresponding to the B regulatory element of the apoCIII promoter (CIIIB), which is a high-affinity binding site for HNF-4a [24], was used as a probe. The double-stranded oligonucleotide is com- posed of CIIIBfor, 5¢-GGTCAGCAGGTGACCTTTGCCC Table 1. Sequences of mutated oligonucleotides used to generate the point mutations in the pcDNA3.1-LBD constructs. The under- lined triplets show the mutated codons. Italic and underlined nucle- otides indicate the cloning sites HindIII and BamHI, as well as the initiator ATG and terminator TAG (CTA in reverse orientation) codons. The bold letters indicate the Kozak sequence placed adja- cent to the initiator ATG for optimal translation. Primer name 5¢-to3¢ primer oligonucleotide sequence D261N for 5¢-CGCATCCTC AATGAGCTGGTCTTG-3¢ D261N rev 5¢-CAAGACCAGCTC ATTGAGGATGCG-3¢ E269Q for 5¢-TTGCCCTTCCAA CAGCTGCAGATC-3¢ E269Q rev 5¢-GATCTGCAG CTGTTGGAAGGGCAA-3¢ Q307L for 5¢-GGTCACAGGTG CTGGTGAGCCTG-3¢ Q307L rev 5¢-CAGGCTCAC CAGCACCTGTGACC-3¢ D312N for 5¢-AGCCTGGAG AATTACATCAACGAC-3¢ D312N rev 5¢-GTCGTTGATGTA ATTCTCCAGGCT-3¢ R324L for 5¢-CTCTCGGGGT CTTTTTGGAGAGCT-3¢ R324L rev 5¢-AGCTCTCCAAA AAGACCCCGAGAG-3¢ Q336L for 5¢-CCCACTCTG CTGAGCATTACCTG-3¢ Q336L rev 5¢-CAGGTAATGCT CAGCAGAGTGGG-3¢ HNF-N 5¢-GATATC AAGCTTGCCGCCGCCATGGAC ATGGCTGACTACAGTGCT-3¢ HNF-C 5¢-TCTAGA GGATCCCTAGATGGCTTCC TGCTTGGTGAT-3¢ E. Aggelidou et al. HNF-4 dimerization FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS 1955 AGCG-3¢, and the complementary CIIIBrev, 5¢-CGCTGG GCAAAGGTCACCTGCTGACC-3¢. Nuclear extracts were incubated with the 32 P-labelled oligonucleotide probe for 30 min at 4 °C in the presence of 25 mm Hepes, pH 7.6, 40 mm KCl, 1 mm dithriothreitol, 5 mm MgCl 2 , and 0.6 lg of poly(dI-dC). Protein–DNA complexes were subsequently analysed by electrophoresis on a 5% nondenaturing gel, followed by autoradiography. In dimerization experiments, extracts from cells transfected with the point mutants were incubated with extracts from CD1b transfectants for 15 min prior to the addition of the probe. CD1b is a truncated form of HNF-4a that has been previously shown to retain the wild-type DNA binding and dimerization properties [1]. Coimmunoprecipitation Western blotting assays For coimmunoprecipitation experiments, HEK293 cells were cotransfected with 4 lg of pcDNA3.1-LBD HNF-4a wild-type or mutant plasmids and 4 lg of CD1b. After 48 h, the cells were washed with ice-cold NaCl ⁄ P i and lysed in 100 lL of ice-cold RIPA buffer (1· NaCl ⁄ P i , 1% Noni- det P-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mm PMSF, 13 lgÆmL )1 aprotinin, 6.7 lgÆmL )1 leupeptin) for 30 min on ice. Cell debris was removed by centrifugation at 10 000 r.p.m. for 10 min at 4 °C and the supernatant was collected. Cell lysates were subjected to 2 h incubation with a preformed complex of the antibody coupled to pro- tein G–agarose on a shaking platform at 4 °C. This com- plex had been formed by incubating 1 lg of goat polyclonal anti-(HNF-4a C-terminal) serum (Santa Cruz Biotechnology, Santa Cruz, CA) together with 30 lLof 50% protein G PLUS-agarose beads (Santa Cruz, Biotech- nology) in 500 lL lysis (RIPA) buffer for 2 h at 4 °Cby rotation. The precise amounts of extracts used in each interaction were normalized to mutant expression, as assessed by western blot analysis. The beads were washed three times with lysis buffer, pelleted by a spin of 30 s and resuspended in 2· loading buffer (120 mm Tris ⁄ HCl, pH 6.8, 4% SDS, 20% glycerol, 0.02% bromophenol blue). Samples were boiled at 100 °C for 10 min, the proteins were separated on 8% SDS ⁄ PAGE and western blot analy- sis followed. In detail, proteins were transferred to an Immobilon P membrane (Millipore, Bedford, MA) by elec- troblotting and the membranes were preincubated in NaCl ⁄ P i , containing 0.1% Tween 20 (PBST), 5% nonfat dry milk and 0.5% bovine serum albumin, for 1 h at 25 °C. Subsequently, they were incubated with a primary rabbit anti-(N-terminal HNF-4a) serum at a dilution of 1 : 5000 in PBST, for 1 h at 25 °C. Membranes were washed three times in NaCl ⁄ P i containing 0.1% Tween 20 and incubated with the secondary antibody, goat anti-(rabbit IgG), conju- gated to horseradish peroxidase (Santa Cruz Biotechnology) at a dilution of 1 : 10 000 in PBST, for 1 h at 25 °C. Membranes were washed three times in NaCl ⁄ P i containing 0.1% Tween 20 and once in NaCl ⁄ P i , and proteins were visualized by exposure to ECL Plus reagent (Amersham Pharmacia Biotech, Piscataway, NJ) according to the manufacturer’s specifications. The rabbit anti-(N-terminal HNF-4a) serum was a generous gift from I. Talianidis (Institute of Molecular Biology and Biotechnology, Hera- kleion, Greece). Copurification of biotin-tagged PGC-1 and HNF-4a from cell extracts Fifty and 60% confluent 60 mm dishes of HEK293 cells were cotransfected with 5 lg each, of the following plasmid vectors: pEV-BirA, Bio-PGC-1 and the pcDNA3.1-LBD wild-type or mutant plasmids. Empty vector DNA was added as appropriate to achieve a constant amount of trans- fected DNA. Cells were harvested 48 h post transfection in 1 mL lysis buffer (20 mm Tris ⁄ HCl pH 7.5, 150 mm NaCl, 10% glycerol, 1% Triton X-100, plus protease inhibitors) by gentle rocking for 20 min at 4 °C. Depending on protein expression, 100–200 lL of the cell extracts were added to 50 lL of 50% streptavidin beads (Sigma, St. Louis, MO), in lysis buffer, up to a final volume of 500 lL per reaction, and reactions were incubated by rotation at 4 °C overnight. Following extensive washing of the beads, the bound pro- teins were eluted by boiling for 5 min in 20 lL2· loading buffer, as described previously [26], and they were electro- phoretically separated on 8% SDS–polyacrylamide gel. The proteins were then transferred to an Immobilon P mem- brane (Millipore) as described above, and membranes were incubated with a horseradish peroxidase-conjugated strept- avidin polymer (Sigma) at a 1 : 7500 dilution for 1 h at room temperature, following incubation in TBS-T 0.1% containing 5% nonfat dry milk and 0.5% bovine serum albumin for 1 h at room temperature. Biotinylated PGC-1 was visualized by exposure to ECL Plus reagent (Amersham Pharmacia Biotech), according to the manufacturer’s instructions. Following stripping of the membranes in 25 mm glycine, 1% Triton X-100, 1% SDS stripping solu- tion, pH 2.5, membranes were re-incubated with the anti- (HNF-4a C19) primary serum (Santa Cruz Biotechnology) and horseradishperoxidase-conjugated anti-goat secondary serum (Santa Cruz Biotechnology) at the dilutions described above. Proteins were again visualized with the aid of the ECL Plus reagent (Amersham Pharmacia Biotech). Acknowledgements This work was supported by funds from the Greek General Secreteriat for Science and Technology (Hera- cleitos) to MHC. We thank Drs A. Kralli, D. Kardassis., J. Strouboulis and B. Laine for plasmids pcDNA3- HA-PGC-1, pcDNA3.1-Bio, pEV-BirA and pcDNA3.1- HNF4a-R324H, respectively, and Dr I. Talianidis for HNF-4 dimerization E. Aggelidou et al. 1956 FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS the generous gift of anti-(N-terminal HNF-4a) serum. We would like to thank Drs D. Kardassis and C. Panag- iotidis for their stimulating discussions. References 1 Hadzopoulou-Cladaras M, Kistanova E, Evagelopoulou C, Zeng S, Cladaras C & Ladias J (1997) Functional domains of the nuclear receptor hepatocyte nuclear fac- tor 4. J Biol Chem 272, 539–550. 2 Hall RK, Scott DK, Noisin EL, Lucas PC & Granner DK (1992) Activation of the phosphoenolpyruvate car- boxykinase gene retinoic acid response element is depen- dent on a retinoic acid receptor ⁄ coregulator complex. Mol Cell Biol 12, 5527–5535. 3 Kardassis D, Sacharidou E & Zannis VI (1998) Trans- activation of the human apolipoprotein CII promoter by orphan and ligand-dependent nuclear receptors. J Biol Chem 273, 17810–17816. 4 Ktistaki E, Lacorte JM, Katrakili N, Zannis VI & Talianidis I (1994) Transcriptional regulation of the apolipoprotein A-IV gene involves synergism between a proximal orphan receptor response element and a dis- tant enhancer located in the upstream promoter region of the apolipoprotein C-III gene. Nucleic Acids Res 22, 4689–4696. 5 Chen WS, Manova K, Weinstein DC, Duncan SA, Plump AS, Prezioso VR, Bachvarova RF & Darnell JE (1994) Disruption of the HNF-4 gene, expressed in visceral endoderm, leads to cell death in embryonic ectoderm and impaired gastrulation of mouse embryos. Genes Dev 8, 2466–2477. 6 Li J, Ning G & Duncan SA (2000) Mammalian hepato- cyte differentiation requires the transcription factor HNF-4alpha. Genes Dev 14, 464–474. 7 Hayhurst GP, Lee YH, Lambert G, Ward JM & Gonzalez FJ (2001) Hepatocyte nuclear factor 4alpha (nuclear receptor 2A1) is essential for maintenance of hepatic gene expression and lipid homeostasis. Mol Cell Biol 21, 1393–1403. 8 Yamagata K, Furuta H, Oda N, Kaisaki PJ, Menzel S, Cox NJ, Fajans SS, Signorini S, Stoffel M & Bell GI (1996) Mutations in the hepatocyte nuclear factor-4a gene in maturity-onset diabetes of the young. Nature (London) 384, 458–460. 9 Gragnoli C, Lindner T, Cockburn BN, Kaisaki PJ, Gragnoli F, Marozzi G & Bell GI (1997) Maturity-onset diabetes of the young due to a mutation in the hepato- cyte nuclear factor-4 alpha binding site in the promoter of the hepatocyte nuclear factor-1 alpha gene. Diabetes 46, 1648–1651. 10 Bogan A, Dallas-Young Q, Ruse MD Jr, Maeda Y, Jiang G, Nepomuceno L, Scanlan TS, Cohen FE & Sladek FM (2000) Analysis of protein dimerization and ligand binding of orphan receptor HNF4alpha. J Mol Biol 302, 831–851. 11 Dhe-Paganon S, Duda K, Iwamoto M, Chi Y-I & Shoe- lson SE (2002) Crystal structure of the HNF4a ligand binding domain in complex with endogenous fatty acid ligand. J Biol Chem 277, 37973–37976. 12 Wisely BG, Miller AB, Davis RG, Thornquest AD, Johnson R, Spitzer T, Sefler A, Shearer B, Moore JT, Miller AB et al. (2002) Hepatocyte nuclear factor 4 is a transcription factor that constitutively binds fatty acids. Structure 10, 1225–1234. 13 Zhang XK, Salbert G, Lee MO & Pfahl M (1994) Mutations that alter ligand-induced switches and dimer- ization activities in the retinoid X receptor. Mol Cell Biol 14, 4311–4323. 14 Rachez C, Sautiere P, Formstecher P & Lefebvre P (1996) Identification of amino acids critical for the DNA binding and dimerization properties of the human retinoic acid receptor a. J Biol Chem 271, 17996–18006. 15 Aggelidou E, Iordanidou P, Tsantili P, Papadopoulos G & Hadzopoulou-Cladaras M (2004) Critical role of resi- dues defining the ligand binding pocket in Hepatocyte Nuclear Factor-4a. J Biol Chem 279, 30680–30688. 16 Eeckhoute J, Oxombre B, Formstecher P, Lefebvre P & Laine B (2003) Critical role of charged residues in helix 7 of the ligand binding domain in Hepatocyte Nuclear Factor 4alpha dimerisation and transcriptional activity. Nucleic Acids Res 31, 6640–6650. 17 Price JA, Fossey SC, Sale MM, Brewer CS, Freedman BI, Wuerth JP & Bowden DW (2000) Analysis of the HNF4 alpha gene in Caucasian type II diabetic nephro- pathic patients. Diabetologia 43, 364–372. 18 Yoon JC, Puigserver P, Chen G, Donovan J, Wu Z, Rhee J, Adelmant G, Stafford J, Kahn CR, Granner DK et al. (2001) Control of hepatic gluconeogenesis through the transcriptional coactivator PGC-1. Nature 413, 131–138. 19 Iordanidou P, Aggelidou E, Demetriades C & Hadzo- poulou-Cladaras M (2005) Distinct amino acid residues may be involved in coactivator and ligand interactions in Hepatocyte Nuclear Factor-4a. J Biol Chem 280, 21810–21819. 20 Oxombre B, Moerman E, Eeckhoute J, Formstecher P & Laine B (2002) Mutations in hepatocyte nuclear fac- tor 4alpha (HNF4alpha) gene associated with diabetes result in greater loss of HNF4alpha function in pancrea- tic beta-cells than in nonpancreatic beta-cells and in reduced activation of the apolipoprotein CIII promoter in hepatic cells. J Mol Med 80, 423–430. 21 Jiang G, Nepomuceno L, Hopkins K & Sladek FM (1995) Exclusive homodimerization of the orphan recep- tor hepatocyte nuclear factor 4 defines a new subclass of nuclear receptors. Mol Cell Biol 15, 5131–5143. E. Aggelidou et al. HNF-4 dimerization FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS 1957 [...]...HNF-4 dimerization E Aggelidou et al 22 Jiang G & Sladek FM (1997) The DNA binding domain of hepatocyte nuclear factor 4 mediates cooperative, specific binding to DNA and heterodimerization with the retinoid X receptor alpha J Biol Chem 272, 1218– 1225 23 Gampe RT, Montana VG, Lambert MH, Wisely GB, Milburn MV & Xu HE (2000) Structural basis for autorepression of retinoid X receptor... Promoter elements and factors required for hepatic and intestinal transcription of the human ApoCIII gene J Biol Chem 265, 9808–9815 1958 25 Ladias JA, Hadzopoulou-Cladaras M, Kardassis D, Cardot P, Cheng J, Zannis V & Cladaras C (1992) Transcriptional regulation of human apolipoprotein genes ApoB, ApoCIII and ApoAII by members of the steroid hormone receptor superfamily HNF-4, ARP-1, EAR-2 and EAR-3 J... EAR-2 and EAR-3 J Biol Chem 267, 15849– 15860 26 De Boer E, Rodriguez P, Bonte E, Krijgsveld J, Katsantoni E, Heck A, Grosveld F & Strouboulis J (2003) Efficient biotinylation and single-step purification of tagged transcription factors in mammalian cells and transgenic mice Proc Natl Acad Sci USA 100, 7480– 7485 FEBS Journal 273 (2006) 1948–1958 ª 2006 The Authors Journal compilation ª 2006 FEBS . Functional characterization of hepatocyte nuclear factor-4a dimerization interface mutants Eleni Aggelidou*, Panagiota. the homodimer interface of HNF-4a contributing to the stability of the interface [11]. The assignment of helices H3–H12 based on the crystal structure of HNF-4a, together