Functional Analysis of Hypoxia-Inducible Factor-1-mediated Transactivation – Identification of Amino Acid Residues Critical for Transcriptional Activation and / or Interaction with CBP *

The hypoxia-inducible factor-1 alpha (HIF-1 alpha) is a key regulator of adaptive responses to hypoxia. HIF-1 alpha has two independent transactivation domains (TADs). Whereas the N-terminal TAD (N-TAD) also constitutes a degradation box, the C-terminal TAD (C-TAD) functions in a strictly hypoxia-inducible fashion. Oxygen-dependent hydroxylation of an asparagine residue has recently been reported to regulate C-TAD function by disrupting the interaction with the CH1 domain of the p300/CBP coactivator at normoxia. Here we have performed alanine-scanning mutagenesis of a predicted alpha-helix within the C-TAD of mouse HIF-1 alpha to identify residues important for transactivation and interaction of the C-TAD with transcriptional coactivators. We observed that several hydrophobic residues, Ile(802), Leu(808), Leu(814), Leu(815), and Leu(818), were critical for transactivation and binding to the CH1 domain of CBP in hypoxic cells. Moreover, E812A/E813A and D819A mutations impaired hypoxia-dependent transactivation without disrupting binding to CH1. In the context of full-length HIF-1 alpha, mutation of the leucine residues conferred conformational changes to the protein and significantly reduced the transactivation function as well as functional interaction with the transcriptional coactivators CBP and SRC-1. These mutations also affected intranuclear redistribution of HIF-1 alpha in the presence of CBP, indicating that the integrity of the C-TAD is critical for intracellular localization of mouse HIF-1 alpha.


INTRODUCTION
In mammalian organisms a reduction in cellular or systemic levels of O 2 (hypoxia) induces a coordinated cascade of biological responses.A critical regulator of these responses to hypoxia is the hypoxia-inducible factor-1α (HIF-1α 1 ).Together with a partner factor, Arnt, HIF-1α forms in the nucleus of hypoxic cells a heterodimeric complex of basic-helix-loop-helix (bHLH) / Per, Arnt and Sim (PAS) domain proteins that activates transcription of a network of genes encoding erythropoietin, vascular endothelial growth factor and glycolytic enzymes (1).Under normoxic conditions, HIF-1α is very rapidly degraded by the 26S proteasome upon ubiquitylation mediated by the von-Hippel Lindau tumor supressor protein (pVHL) (2)(3)(4)(5)(6).Activation of transcription by the HIF-1α/Arnt dimer requires the recruitment of transcriptional coactivators such as p300/CBP and factors belonging to the p160/SRC-1 family of proteins (7)(8)(9)(10)(11)(12).Proteins of the CBP/p300 and p160/SRC family of coactivators possess histone acetyltransferase activity and therefore have the potential to remodel chromatin structure (13,14).CBP and p300 are coactivator proteins that operate at the end points of a variety of signal transduction pathways, thereby modulating specific gene expression programs involved in cell growth, differentiation, homeostasis and viral pathogenesis (15).
Low availability of oxygen will inhibit this mechanism leading to the release of pVHL, thus stabilizing the N-TAD and facilitating the transactivation function.In contrast, C-TAD protein levels are not affected by changes in oxygen concentration but transactivation mediated by this domain is strongly induced by hypoxia ( 8,10,16,17).
Two mechanisms have been proposed for the hypoxia-inducible function of the C-TAD.Mahon et al. (27) have identified a protein, FIH-1, which binds to the Cterminus of HIF-1α repressing C-TAD mediated transactivation, and have proposed a model involving recruitment of pVHL and histone deacetylases to FIH-1 to form a repressor complex.In addition, a more recent report has demonstrated that an asparagine residue in the C-TAD is hydroxylated at normoxia, resulting in inhibition of the interaction between the C-TAD and the CH1 domain of p300 (28).This mechanism seems also to be mediated by Fe(II)-2-oxoglutarate-dependent dioxygenases since the use of specific inhibitors for these enzymes leads to constitutive activity of C-TAD (28,29).The structure of the HIF-1α C-TAD bound to the CH1 domain of p300 has recently been solved in two independent studies (30,31).Depending on the fragment used, the C-TAD is shown to form 2 or 3 α-helices upon binding to the CH1 domain of CBP.Both studies show that the last α-helix is in intimate contact with CH1.Previous observations have shown that deletion of the last thirteen amino acids renders the C-TAD transcriptionaly inactive, indicating that this region is critical for the transactivation function (8,17).
In this study we have characterized the C-TAD by performing alanine-scanning mutagenesis based on the predicted structure of this transactivation domain.We have identified residues within the C-TAD that are critical for the hypoxia-inducible transactivation response and for potentiation of this response by CBP and SRC-1.We show that mutation of specific amino acids in the C-TAD results in loss of transactivation through disruption of the binding to the CH1 domain of CBP.Moreover, we demonstrate that in vivo colocalization of mHIF-1α and CBP in intranuclear foci is impaired by mutations in the C-TAD of HIF-1α which negatively affect the hypoxiainducible transactivation function of the protein.The precipitated protein complex was separated by SDS-PAGE, blotted onto a nitrocellulose filter and probed with specific antibodies as described below.

Cell Culture and Transient Transfection Experiments
Immunoblotting and Detection-Ten µg of total cell protein was blotted after separation by SDS-PAGE on a nitrocellulose filter and blocked overnight with 5% non-fat milk in TBS.Antibodies for FLAG epitope (Sigma) or GST (AmershamPharmacia) were diluted as recommended by the manufacturer in 5% nonfat milk in TBS and incubated at room temperature for 1h.After washes, 1:2000 dilution of anti-mouse or anti-goat IgG-horseradish peroxidase conjugate (Amersham) in TBS containing 1% non-fat milk, was used as secondary antibody.The filters were then extensively washed with TBS, 0.05% Tween-20 and complexes were visualized using enhanced chemiluminescence (Amersham) according to the manufacturer's recommendations.
Limited Proteolysis Assay-FLAG-tagged wild-type or mutant mHIF-1α proteins were expressed in the presence of [ 35 S]methionine in a coupled cell-free transcription-translation kit (Promega) according to the manufacturer's instructions.
After brief centrifugation the Sepharose pellets were washed two times with TBS buffer and in vitro translated proteins were added in a total volume of 50 µl in TBS and further incubated for 1h under the same conditions.After three washes with TBS, immunoprecipitated proteins were eluted with FLAG peptide (0.1 mg/ml) for 30 min at room temperature and an aliquote of the eluted material was analyzed by 7.5%  (33,34).
This analysis indicated two α-helices within the mHIF-1α C-TAD spanning amino acids 772-782 (helix 1) and 806-820 (helix 2; Fig. 1A).These two regions have recently been shown to partially overlap with the first and last helices of the C-TAD when bound to the CH1 domain of p300 (30,31).To investigate the functional importance of these two predicted α-helices, we constructed plasmids encoding the FLAG epitope fused in-frame with the GAL4 DNA-binding domain and the minimal C-TAD (FLAG-GAL4/mHIF-1α(772-822)), as well as two deletion fragments of this domain: FLAG-GAL4/mHIF-1α(779-822) corresponding to deletion of the first 6 amino acids, and FLAG-GAL4/mHIF-1α(772-809) corresponding to deletion of the last 13 amino acids of the C-TAD (8).These constructs were tested for their ability to activate transcription of a GAL4-driven luciferase reporter gene in transient transfection assays in HEK 293 cells.(Fig. 1B).All mutants showed equivalent levels of expression as assessed by immunoblotting using anti-FLAG antibodies (Fig. 1C).
In order to identify the amino acids responsible for the dramatic effect on transactivation caused by deletion of helix 2, an extensive analysis of this region of the Fig. 1A).The ability of each mutant to activate transcription was tested as before in transient transfection assays (Fig. 2A).These results showed that mutation to alanine of hydrophobic amino acids produced in the case of the I802A mutant either a dramatic decrease or, in the case of the mutants LL808AA, LL814AA, or L818A, complete abrogation of the transactivation function and the hypoxia-inducible response.In contrast, mutation of acidic amino acids in the EE812AA and D819A mutants led to lower levels of relative luciferase activity.Mutations such as R806A and Q810A had no significant effect on the transactivation function of C-TAD, whereas the mutants N807A and R816A showed an increased transactivation function at hypoxia when compared to the wild type C-TAD.The S805A mutation seemed to have no effect per se on the transactivation function of C-TAD either at normoxia or hypoxia, in agreement with a recent report where mutation of residue S857 in HLF (corresponding to S805 in mHIF-1α) does not affect transactivation mediated by the C-TAD (36).The expression levels of all the mutants tested in gene reporter assays were investigated by immunoblot analysis and proven to be very comparable (data not shown).2B).These results suggest that in the context of the GAL4/C-TAD these leucine residues within C-TAD were critical for functional interaction with CBP.In contrast, the EE812AA mutation that showed a decreased transactivation potency by itself was still responsive to CBP-or SRC-1-mediated enhancement, indicating that these acidic amino acids are not as critical as hydrophobic residues for C-TAD function.

Interaction of the C-TAD with the CH1 Domain of CBP is Affected by Specific
Mutations in Helix 2-The interaction of the HIF-1α C-TAD with the transcriptional coactivator CBP has been shown to be mediated by the CH1 domain of CBP (20).
Based on the observation that some of the mutations in the predicted helix 2 of the C-TAD resulted in loss of transactivation even following overexpression of CBP, we investigated whether there was any correlation between the functional activity of the The di-leucine repeats present in predicted helix 2 of the C-TAD were further studied to establish the relevance of each individual leucine residue for the interaction with CH1.Fig. 3B shows the results of the protein-protein interaction assay using Coexpression of CBP or SRC-1 potentiated the transactivation function of wildtype mHIF-1α both at normoxia and hypoxia (Fig. 4B).A similar effect was observed for mHIF-1α(772∆822) and mHIF-1α(532∆583) (data not shown), indicating that the potentiating effect can be mediated by each transactivation domain individually (8,10).
Gel mobility shift analysis showed that in the presence of Arnt all mutants were able to bind DNA (data not shown).To examine if introduction of selected mutations into C-TAD induced conformational changes in the full length protein, the different in vitro translated 35 S-methionine labelled FLAG tagged proteins were first immunoprecipitated with anti-FLAG antibodies, followed by specific elution with the Accumulation foci with colocalization of both proteins were observed in 22% of cells at normoxia.Treatment with CoCl 2 led to a significant increase in the number of cells with these discrete foci (92% of cells).However, the number of foci per nucleus (40 -100) was not altered by CoCl 2 treatment.Redistribution of CFP-mHIF-1α and YFP-CBP upon coexpression of both proteins strongly suggests that these proteins interact in vivo and that this interaction is hypoxia-inducible.Coexpression of CFP-mHIF-1α(LL808AA) with YFP-CBP led to the formation of accumulation foci in only 11% of cells at normoxia and in 38% of cells after treatment with CoCl 2 (Fig. 5B), indicating that the C-TAD plays an important role in mHIF-1α-mediated redistribution of CBP.
Similar results were obtained with CFP-mHIF-1α(LL814AA) and CFP-mHIF-1α(L818A) (data not shown).These results were comparable with those obtained following coexpression of YFP-CBP and CFP-mHIF-1α(772∆822) (accumulation foci in 39% of cells in the presence of CoCl 2 ) suggesting that helix 2 of C-TAD may be involved in the recruitment process leading to colocalization of mHIF-1α with CBP in discrete accumulation foci.Consistent with this model, YFP-CBP was able to redistribute CFP-mHIF-1α(532∆583) from a diffuse intranuclear pattern to the characteristic accumulation foci although in a lower number of cells both at normoxia (11%) and hypoxia (38%).Taken together these results indicate that efficient interaction between mHIF-1α and CBP resulting in colocalization of both proteins in 92% of cells at hypoxia is dependent on the integrity of mHIF-1α transactivation domains and that the putative helix 2 of the C-TAD serves as the interaction interface between the transactivation domain and CBP.In excellent agreement with this notion, no maintained a diffuse intranuclear distribution and did not colocalize with YFP-CBP under any of the conditions described above (Fig. 5B).

DISCUSSION
In the present study we have identified amino acids within the predicted helix 2 of the HIF-1α C-TAD that are essential for the hypoxia-dependent transactivation response, interaction with CBP, and intranuclear colocalization with CBP.CBP/p300 has been shown to play a critical role in cellular responses to hypoxia.Functional assays have demonstrated that CBP is able to enhance HIF-1α-mediated transactivation function, indicating that CBP functions as a coactivator in hypoxia-induced transcriptional responses (7,8,10,37).The CH1 region of p300 has been shown to interact with HIF-1α (7) by directly targeting the C-TAD, as assessed by proteinprotein interaction assays (11,12,20).Based on this mechanism, overexpression of the C-TAD motif of HIF-1α was demonstrated to inhibit tumor growth in animal models, possibly by disrupting the interaction of HIF-1α with p300 (12).More recently the solution structure of the complex formed by the HIF-1α C-TAD and the CH1 domain of p300/CBP has been determined in two independent studies (30,31).In one of these studies the C-TAD forms three short helices upon binding to the CH1 domain (31).
Two of the C-TAD helices identified in the solution structure overlap partially with the predicted helices described in the present study while the central helix containing N803 (hHIF-1α) was not predicted by the program and was therefore was not considered in this study.
Alanine-scanning mutagenesis of the two predicted α-helices of the C-TAD demonstrated that mutations of hydrophobic residues within helix 2 completely abrogated C-TAD-mediated transactivation, physical interaction with the isolated CH1 domain of CBP, as well as CBP-mediated enhancement of transcription.The predicted helix 2 has been reported to be essential for the interaction of HIF-1α C-TAD with the CH1 region (20).In the solution structure of the C-TAD/CH1 complex the last helix of the C-TAD was docked in the groove of the CH1 domain, and many hydrophobic and polar interactions contribute to the stabilization of the complex (30,31).Two of the leucines within the predicted helix 2 that were identified in our mutation analysis have been previously shown to constitute part of the interface between HIF-1α C-TAD and the CH1 motif (L818 and L822 in hHIF-1α corresponding to L814 and L818 in mHIF-1α).In the present study we have identified three additional hydrophobic residues, I802, L808, and L815 in mHIF-1α C-TAD, that were critical for transcriptional activation by the C-TAD and interaction with CBP.In the solution structure of the complex between the C-TAD and the minimal CH1 all these amino acids were proposed to make hydrophobic and polar interactions important for stabilization of the complex.
Residues I802 and L808 are located in an intrahelix bridge while L814, L815, and L818 are present in the last helix of the C-TAD (30,31).Interestingly, in our study mutation of acidic residues to alanine (mutants EE812AA and D819A) decreased or abrogated hypoxia-dependent transactivation mediated by the C-TAD, but in contrast to any mutation of the hydrophobic amino acids, EE812AA and D819A were still able to in mHIF-1α) have been proposed to make electrostatic interactions with CH1 in the structural studies but do not seem to be critical for binding to CBP in our functional analysis.
It has recently been proposed that hypoxia-dependent transactivation mediated by the C-TAD is regulated by hydroxylation of N803 (N799 in mHIF-1α) by Fe(II)-2oxoglutarate-dependent dioxygenases (28).When mutated to alanine N803 has been shown to generate a constitutive C-TAD (28).In contrast, mutation of N807 present in the predicted helix 2 resulted in increased transactivation activity only under hypoxic conditions.N807 has been proposed to make polar interactions with the p300 CH1 domain in one of the structural studies while in the other study the N807 was in a disordered structure (30,31).Based on our results N807 appears not to be critical for either transactivation or interaction with CBP.
In the present study we have visualised the in vivo interaction between HIF-1α and CBP using CFP and YFP fusion proteins.Upon coexpression of both proteins the to recruit coactivators such as SRC-1 and TIF2 to accumulation foci in a liganddependent manner (42,45).This process of coregulator redistribution by a transcription factor, also known as intranuclear marshaling, correlates with the formation of complexes potentially involved in regulation of gene expression.Thus, the N-TAD, in addition to the previously described role in targeting HIF-1α for degradation at normoxia (2,5,21), also contributes to the hypoxia-inducible transcriptional activation in the hypoxic cell by mediating interaction in vivo between HIF-1α and CBP, resulting in potentiation of the transactivation function of HIF-1α.
Taken together, we have identified several residues in the C-TAD that are important for hypoxia-dependent activation of transcription by HIF-1α.We have also obtained evidence that mutation of several hydrophobic residues present in a predicted α-helix of the C-TAD not only disrupts the in vitro interaction with the CH1 domain of CBP but also impairs the interaction of HIF-1α with CBP in vivo.
by guest on September 1, 2017 http://www.jbc.org/Downloaded from Ruas et al.: 10 SDS-polyacrylamide gel electrophoresis (SDS-PAGE) followed by autoradiography.Limited proteolysis was performed as previously described (32) with the following modifications: equal amounts of the different proteins were incubated with 5, 50, 150 or 300 ng of Staphylococcus aureus endoproteinase Glu-C (V8 protease; Boehringer Mannheim) in a reaction buffer containing 40mM HEPES pH 7.5, 120 mM KCl, 5 mM MgCl 2 and 0.5 mM EDTA in a total volume of 50 µl for 30 min at 30ºC.The reaction products were separated by 20% SDS-PAGE, fixed in 10% (vol/vol) acetic acid, and dried.Radioactive peptides were detected by autoradiography.Nuclear Translocation and Colocalization Studies-HEK 293 cells were cultured on coverslips in 3-cm diameter dishes for 24h and transfected as before with 500 ng of each plasmid.36h after transfection medium was changed for treatment in the presence or absence of 100 µM CoCl 2 for 2h.Coverslips were then removed from the culture dishes, and cells were fixed for 30 min in 4% paraformaldehyde in phosphate-buffered saline (PBS).The coverslips were then mounted on glass slides and observed using a Axiovert 35 Zeiss microscope (Carl Zeiss Jena GmbH, Jena, Germany) with epifluorescence and illumination from a Gixenon burner, and equipped with selective filter sets for CFP (exciter: D436/20, emitter D480/40) and YFP (exciter: HQ500/20, emitter: HQ535/30) (Chroma Tech).Images were captured by a CCD-CH250 camera (Photometrics, Arizona, USA), and signal superimposition obtained with the camera software (QED imaging).Subcellular and intranuclear distribution of CFP-and YFP-fused proteins was determined by counting around 300 cells.Ruas et al.: 11 RESULTS Two Predicted α-Helices Within the C-TAD Are Critical for the Hypoxiainducible Transactivation Function-When fused to the GAL4 DNA-binding domain both N-and C-TAD can independently activate transcription of a GAL4-driven luciferase reporter gene in a hypoxia-dependent manner (8).To investigate the structure and function of the C-TAD in closer detail we performed a computer-based analysis of the amino acid sequence of mHIF-1α using the secondary structure prediction program PREDATOR (EMBL [http://www.embl-heidelberg.de/cgi/predator_serv.pl])

Fig. 1B shows
Fig.1Bshows the levels of relative luciferase activity obtained at normoxia or Helix 2 of the C-TAD Affect Potentiation by CBP and SRC-It has been previously demonstrated that CBP and SRC-1 are able to stimulate transcriptional activation mediated by the HIF-1α C-TAD (8, 10), and that this potentiation effect is lost by the deletion of helix 2 which disrupts the binding to CBP (8, 20).Against this background we next examined if the loss of transactivation function observed by mutation of the leucine residues LL808-809, LL814-815 and L818 could be rescued by overexpression of CBP and/or SRC-1 in transient transfection assays.As shown in Fig. 2B, under these conditions expression of CBP or SRC-1 had no detectable by guest on September 1, 2017 http://www.jbc.org/Downloaded from Ruas et al.: 14 effect on the transcriptional activity of GAL4-DBD alone.However, as expected (8) C-TAD-mediated transactivation was potentiated about 2.5 -3.0-fold both at normoxia and hypoxia by coexpression of either CBP or SRC-1.The transactivation function of the mutants LL808AA, LL814AA, and L818A was not significantly altered by coexpression of CBP or SRC-1 (Fig. various C-TAD mutants and their ability to interact with the isolated CH1 domain of CBP.To this end a GST-CH1 fusion protein was used to precipitate C-TAD from whole cell extracts prepared from transiently transfected HEK 293 cells grown in the presence or absence of 100µM CoCl 2 .Fig.3Ashows the results of the GST-pull down assay analysed by immunoblotting.In control experiments, GST-CH1 did not interact with the FLAG epitope or with the GAL4 DNA binding domain under any of the tested conditions (Fig.3A, lanes 1 and 2).Moreover, no interaction was observed by guest on September 1, 2017 http://www.jbc.org/Downloaded from Ruas et al.: 15 between GST and wild-type or mutant C-TAD under any of these conditions (data not shown).No significant interaction was observed between GST-CH1 and the C-TAD at normoxia (Fig. 3A, lane 3) but treatment with CoCl 2 led to efficient precipitation of the C-TAD (lane 4).In excellent agreement with the functional studies, mutants that failed to activate transcription were unable to interact with the CH1 domain of CBP.Mutation of C-TAD residues I802 (lanes 5 and 6), LL808-809 (lanes 9 and 10), LL814-815 (lanes 13 and 14) or L818 (lanes 15 and 16) resulted in total abrogation of interaction with CH1 both at normoxia or upon treatment with CoCl 2. Furthermore, in good correlation with the funcional assays the N807A mutant that showed increased hypoxia-dependent transactivation when compared to wild type C-TAD, interacted more strongly than the wild-type protein with CH1 under hypoxic conditions (lane 8) although no interaction was observed at normoxia (lane 7).Mutation of the acidic amino acid residues EE812-813 that showed decreased levels of transactivation, resulted in no interaction with CH1 at normoxia (lane 11) but interaction of CH1 when cells were treated with CoCl 2 (lane 12).Consistent with this result, EE812AA-mediated activation of transcription could still be enhanced by CBP overexpression.Interestingly, the D819A mutant that produced basal transactivation activity, showed an increased afinity for the CH1 domain when compared to the C-TAD both at normoxia (lane 17) and upon treatment with CoCl 2 (lane 18).
by guest on September 1, 2017 http://www.jbc.org/Downloaded from Ruas et al.: 16GST-CH1 and point mutations of L808, L809, L814 and L815, respectively.As observed above the interaction between C-TAD and CH1 was induced upon treatment of the cells with CoCl 2 (lanes 1 and 2).Mutation of L808 resulted in total loss of interaction under normoxic (lane 3) or hypoxic conditions (lane 4).The C-TAD mutants L813A (lanes 7 and 8) and L814A (lanes 9 and 10) showed week affinity for CH1, comparable to that observed for the C-TAD at normoxia (lane 1).L809 seemed to be less critical for the interaction between the HIF-1α C-TAD and CH1.Mutation of this amino acid resulted in efficient binding both at normoxia (lane 5) and upon treatment with CoCl 2 (lane 6).Mutations in the Predicted Helix 2 Mimic the Effects of Deleting the C-TAD withRegard to HIF-1α-mediated Transactivation and Interaction with Coactivators-Toestablish if the leucine to alanine mutations that inactivated C-TAD function would affect mHIF-1α-mediated transactivation function we introduced the same mutations into the full length protein, thus generating mHIF-1α(LL808AA), mHIF-1α(LL814AA), and mHIF-1α(L818A), respectively.These mutants were tested for their ability to activate transcription of an HRE-driven reporter gene in HEK 293 cells (Fig.4A).When compared to the transactivation function of wild-type mHIF-1α at normoxia and hypoxia mHIF-1α(LL808AA), mHIF-1α(LL814AA) and mHIF-1α(L818A) all showed a significant (about 2.0-fold) decrease in functional activity at hypoxia.The activity of these mutants is comparable to that of mHIF-1α (772∆822) where the C-TAD is completely deleted, reflecting the transactivation function mediated by the N-TAD alone.To further corroborate the critical role of helix 2 in C-TAD-mediated transactivation the same mutations were introduced into mHIF-1α(532∆583) lacking the by guest on September 1, 2017 http://www.jbc.org/Downloaded from Ruas et al.: 17 N-TAD.In the absence of the N-TAD the leucine mutations completely abolished the transactivation function, further ilustrating the critical role of these residues for C-TAD function.The expression levels of the HIF-1α mutants were analysed by immunoblotting using anti-FLAG antibodies and found not to vary sgnificantly (data not shown).

FLAG
peptide, and then submitted to limited proteolysis by S. aureus Glu-C endoproteinase (V8 protease).As shown in Fig.4C, analysis of the generated protein fragments by 20% SDS-PAGE and autoradiography showed qualitative and quantitative differences in degradation products produced by digestion of either the wild-type or the by guest on September 1, 2017 http://www.jbc.org/Downloaded from Ruas et al.: 18 mutant mHIF-1α proteins, as indicated by the appearance of different patterns of discrete bands at the lower part of the gel (indicated by arrowheads).Thus, these experiments indicate that introduction of mutations into helix 2 of the C-TAD may have induced a conformational change in the full length protein.CBP Induces Recruitment of HIF-1α into Discrete Nuclear Accumulation Foci -Effects of Mutations in the Predicted Helix 2 of the C-TAD-Hypoxia and hypoxiamimicking agents such as CoCl 2 are known to profoundly affect the subcellular localization of HIF-1α by inducing rapid nuclear accumulation of the protein(37).This effect has been shown to be mediated by a C-terminal NLS motif(37).It has been demonstrated that intranuclear distribution of several proteins (including HIF-1α)(8) can be affected by different factors such as addition of ligands (in the case of some nuclear receptors)(38)(39)(40)(41)(42) or coexpression of interacting proteins (PML, CBP, SRC-1, TIF2)(8,(42)(43)(44)(45).Under these conditions a diffuse intranuclear distribution pattern will change to either a punctuated one or to discrete accumulation foci.The nature and function of these discrete foci is not yet completely understood but several reports suggest that they may dependent on the existence of a nuclear matrix targeting signal (NMTS)(46,47), or on the recruitment to these foci by proteins that have an NMTS motif.Based on the results indicating an important role of the putative helix 2 of the C-TAD for interaction between mHIF-1α and CBP, we investigated the effect of mutations within helix 2 on intranuclear distribution of mHIF-1α in the absence and presence of this coactivator.To this end we constructed vectors carrying in-frame CFP fusions of wild-type mHIF-1α or HIF-1α mutants [LL808AA, LL814AA, and L818A, mHIF-1α(532∆583), mHIF-1α(532∆583)(LL808AA) and mHIF-1α(772∆822)].The by guest on September 1, 2017 http://www.jbc.org/Downloaded from Ruas et al.: 19abilities of these CFP fusion proteins to activate transcription of an HRE-driven reporter gene corresponded to those exhibited by the FLAG-tagged chimeras and they showed similar expression levels, as assessed by immunoblot analysis using anti-GFP antibodies (data not shown).For colocalization studies we also constructed a YFP-CBP fusion protein.The intracellular localization of the CFP and YFP fusion proteins was investigated in the absence of presence of CoCl 2 .CoCl 2 was chosen as hypoxiamimicking agent to maintain the hypoxic signal throughout the procedure, thus avoiding potential problems that may occur by reoxygenation of the samples.In these assays CFP and YFP alone were distributed throughout the cell, and CoCl 2 treatment had no effect either on the intensity or the subcellular distribution of fluorescence (data not shown).In agreement with a previous report using GFP/hHIF-1α fusion proteins(37), CFP-mHIF-1α was distributed throughout the cell at normoxia and accumulated in the nucleus upon treatment with the hypoxia-mimicking agent CoCl 2 (Fig.5A).The same effect was observed with CFP-mHIF-1α(532∆583) and CFP-mHIF-1α(772∆822), and introduction of helix 2 mutations into the full length protein or into CFP-mHIF-1α(532∆583) had no effect on the kinetics or quantitative aspects of CoCl 2 -inducible nuclear accumulation of any of these proteins (data not shown).Expression of a YFP-CBP chimera in HEK 293 cells showed an exclusively nuclear protein localization exhibiting a punctuated pattern (Fig. 5A) that was not altered by treatment with CoCl 2 .Coexpression of CFP-mHIF-1α with YFP-CBP led to a redistribution of both proteins with the formation of accumulation foci were both fluorescence signals could be observed with the help of selective excitation/emission by guest on September 1, 2017 http://www.jbc.org/Downloaded from Ruas et al.: 20 filters, thus indicating colocalization of CFP-mHIF-1α and YFP-CBP (Fig. 5B).
alterations observed with regard to intranuclear distribution patterns indicate that an in vivo interaction occurs in a hypoxia-inducible manner.These interactions are dependent on the integrity of the HIF-1α transactivation domains since recruitment of the two proteins to accumulation foci was lost when a mutant, transcriptionally inert form of HIF-1α was used in the assay.Intranuclear redistribution of CBP by a transcription factor has been previously shown with C/EBPα and this event is dependent on the C/EBPα transactivation domain(44).Also steroid hormone receptors have been shown by guest on September 1, 2017 http://www.jbc.org/Downloaded from Ruas et al.: 24 Our results demonstrate the importance of the helix 2 motif of the C-TAD in the context of HIF-1α full-length protein since mutations of helix 2 leucines affected the conformation of the protein and the hypoxia-dependent transactivation function of HIF-1α, as well as interaction with CBP in vivo.However, in contrast to the report by Gu et al.,(11) our data showed that mutations targeting the C-TAD predicted helix 2 are not enough to render full length HIF-1α transcriptionally inert, thus indicating an important role of the N-TAD for the transactivation function of the full-length protein.Complete inactivation of the transactivation function of HIF-1α was only achieved by deletion of the N-TAD concomitant with introduction of critical mutations within the C-TAD.