The subcellular localization of bHLH transcription factor TCF4 is mediated by multiple nuclear localization and nuclear export signals

Transcription factor 4 (TCF4) is a class I basic helix-loop-helix (bHLH) transcription factor which regulates the neurogenesis and specialization of cells. TCF4 also plays an important role in the development and functioning of the immune system. Additionally, TCF4 regulates the development of Sertoli cells and pontine nucleus neurons, myogenesis, melanogenesis and epithelial-mesenchymal transition. The ability of transcription factors to fulfil their function often depends on their intracellular trafficking between the nucleus and cytoplasm of the cell. The trafficking is regulated by specific sequences, i.e. the nuclear localization signal (NLS) and the nuclear export signal (NES). We performed research on the TCF4 trafficking regulating sequences by mapping and detailed characterization of motifs potentially acting as the NLS or NES. We demonstrate that the bHLH domain of TCF4 contains an NLS that overlaps two NESs. The results of in silico analyses show high conservation of the sequences, especially in the area of the NLS and NESs. This high conservation is not only between mouse and human TCF4, but also between TCF4 and other mammalian E proteins, indicating the importance of these sequences for the functioning of bHLH class I transcription factors.

N-terminally YFP-tagged full-length TCF4 using fluorescent microscopy. The expression of YFP-TCF4, and in further experiments YFP-tagged TCF4 mutants in the COS-7 cells, was confirmed by western blot analysis using an anti-GFP antibody (Fig. S1). As the expression of TCF4 was documented in various tissues 7 , we decided to use two different cell lines: COS-7 cells used previously for bHLH protein localization studies 37,38 and Neuro2A mouse neuroblastoma (N2a) cells with a documented expression of TCF4 5 . The distribution of TCF4 fused with YFP in more than 95% of COS-7 cells was strictly nuclear (Fig. 1a,a'), however there were some cells (less than 5%) presenting fluorescence not only in the nucleus, but also in the cytoplasm in the puncta that were non-uniformly distributed in the cytoplasm of cells (Fig. S2). In the case of the N2a cells, the expression of YFP-TCF4 was exclusively nuclear (Fig. 1b,b'). As proof that YFP does not influence TCF4 localization, we expressed the YFP protein alone in the COS-7 (Fig. 1 c) and N2a cells (Fig. 1d), which, as expected, resulted in ubiquitous localization. Differences in localization of TCF4, depending on the cell type and literature data presented here, led us to the hypothesis about the putative presence of the composite system of NLSs and NESs in TCF4. This hypothesis is substantiated by a previous report indicating that related proteins, E2a and HEB, can be observed in the nucleus and cytoplasm of human embryonic stem cells (hESCs) 30 .
The N-terminal fragment of the canonical TCF4-B isoform do not possess an active NES. It was previously suggested that the N-terminal region of TCF4 could contain an active NES, however experimental verification of prediction was not performed 39 . We repeated in silico analysis using currently available NES predictors and obtained the same positive results for 10-19 aa (Fig. 2Aa). To experimentally test the activity of putative NES, we prepared deletion fragments comprising amino acid residues 1-30, 31-146, and 1-146 fused C-terminally to YFP (Fig. 2B). The expression of YFP-TCF4/1-30 (Fig. 2Ca,a'), YFP-TCF4/31-146 (Fig. 2Cb,b') and YFP-TCF4/1-146 (Fig. 2Cc,c') in the COS-7 and N2a cells resulted in the ubiquitous distribution of fluorescence throughout the whole cell. These results show that the sequence predicted by the NetNES server (Fig. 2Aa) is not an active NES in this region. The lack of activity can be explained by results of in silico analysis by NetSurfP, showing that leucine residues important for classical NES are predicted to be not exposed to the surface of the protein (Fig. 2Ab).
previously determined nLS can act as noLS. When testing the localization of different isoforms of TCF4, Sepp with co-workers 5 identified bipartite NLS in the region of 8-9 exons of TCF4 comprising amino acid residues 156-178 in the TCF4-B isoform. This NLS was shown to be conserved among E-proteins 5,40 . We prepared YFP-tagged truncation mutants of TCF4 containing this area (Fig. 3A). The expression of YFP-TCF4/147-187 resulted in a strictly nuclear fluorescence in the COS-7 cells (Fig. 3Ba). Importantly, nucleolar localization was observed, which was proven by the staining with Draq5 dye that was used for visualization of chromatin and nucleoli positions (Fig. 3Ba'). Moreover, in the case of YFP-TCF4/147-187 expression in the N2a cells, we observed fluorescence in both the nuclei and nucleoli (Fig. 3Bb). Interestingly, the expression  Regions of the TCF4 (NLS and SMART predicted bHLH) are depicted using different patterns and upper numbering of aa residues. The length of each domain in the diagram is arbitrary. Bottom aa residues refer to first aa residues of all TCF4 fragments used in this study. Region of actual studied area of TCF4 is shown by the red rectangle. Expressed deletion mutants of TCF4 are depicted as arrows and particular deletion mutants are depicted as arrows. The same fashion is used for Figs 2-4 and 6-10 (C) Subcellular distribution of deletion mutants of TCF4. Subcellular localizations of the expressed proteins were analysed by confocal microscopy 20-24 h after transfecting COS-7 and N2a cells. Representative images (single confocal plane for confocal microscopy) of subcellular distribution of the derivatives of the TCF4/1-146 area. Bar, 10 µm. Ratios between mean fluorescence intensity in cytoplasmic, nuclear and nucleolar compartments are presented as an accumulated bar graph (no-nucleolus; nu-nucleus, cyt-cytoplasm).   41 detected the nucleolar presence of TCF4 in primary cultures of forebrain neurons and astrocytes. We performed in silico prediction of the nucleolar localization signal (NoLS) in TCF4 with a Nucleolar localization sequence Detector (NoD) and got negative results. However, prediction of the NoLS is characterized by low accuracy and needs experimental verification 42 . A typical characteristic of the NoLS is the high presence of basic amino acid residues that are located in the alpha-helical and easily accessible region on the surface of the protein. There are three known classes of NoLSs and NLSs: NoLS-only signals, NLS-only signals and the joint NoLS-NLS region containing an overlapping NLS and NoLS 42 . As all characteristics necessary for acting as a NoLS possess a NLS (156-178aa) in the TCF4, we propose that this NLS may present additional NoLS activity that is regulated by an unknown mechanism (potentially posttranslational modification or interaction with a masking NoLS activity partner) that maybe uses some of the amino acid residues present in the 188-300 area of the TCF4.
No active localization signal exists in the central part of TCF4. NLS predictors recognize classical signals rich in basic amino acid residues, while the last few years has also seen the discovery of proline-tyrosine-NLSs 43 . Moreover, NES appears to be a complex and diverse signal, not always being a classical leucine-rich motif 44 . To systematically test the presence of potential sequences that influence subcellular localization in the central region of TCF4, we analysed fragments of this region fused C-terminally to YFP. These TCF4 fragments were defined according to the secondary structure prediction, which was performed using PSIPRED (data not shown). Eventually, four fragments, encompassing residues: 188-300, 301-367, 368-410 and 411-549 were selected (Fig. 4A). The expressed proteins: YFP-TCF4/188-300 (Fig. 4Ba,a'), YFP-TCF4/301-367 ( Fig. 4Bb,b'), YFP-TCF4/368-410 (Fig. 4Bc,c') and YFP-TCF4/411-549 (Fig. 4Bd,d') presented uniform distribution between the nucleus and cytoplasm in the COS-7 and N2a cells. To guarantee that we did not disrupt any structural motif in the constructs presented above, we also prepared fused versions of the previously selected fragments (Fig. 4A). Similarly to the results described above, YFP-tagged TCF4/188-367 (Fig. 4Be,e'), YFP-TCF4/301-410 (Fig. 4Bf,f ') and YFP-TCF4/368-549 (Fig. 4Bg,g') were ubiquitously distributed in the COS-7 and N2a cells. As no NLS or NES motif was predicted in this area, this result proved the lack of active subcellular localization signals in this region under the used conditions. The bHLH domain of TCF4 possesses an overlapping NLS and two NES motifs. It was previously shown that the full length GFP tagged bHLH domain of TCF4 was present in the nucleus and cytoplasm of the HEK293 cells, while the expression of the separated N-and C-terminal parts of this domain resulted in cytoplasmic localization of both truncations 5 . As previously published data did not unambiguously define the molecular features that affect subcellular distribution of the bHLH domain, we decided to perform a detailed study of this area. We asked the question about the presence of putative NESs and NLSs in fragments containing bHLH using different predictors. According to CDD (conserved domain database), a resource for the annotation of functional units in proteins, homology area of bHLH of TCF4 was shown for aa residues 553-634 in contrast to predicted by Smart (568-626) (see Fig. 5A) and Prosite (567-620) (Fig. S3). Due to this, we decided to add additional aa residues to both the N-and C-terminus of the predicted bHLH in order to be sure that we did not destroy any structural motif. In this paper we refer to the bHLH domain as area 550-670 aa, additionally divided for the N-terminal (550-601) and C-terminal (602-670) part (Fig. 4A).
Results suggested the presence of NESs in two fragments -one comprising the N-terminal part of the bHLH, and one comprising the C-terminal part of the bHLH (Fig. 5A). Additionally, we performed a prediction of the putative NLS with a cNLS Mapper and got a positive result for aa 569-603 ( Fig. 5B) overlapping predicted NESs. We decided to experimentally test the activity of all the predicted signals.
First, we prepared the YFP-TCF4/550-601 construct (Fig. 6A) comprising the N-terminal fragment of the bHLH. As expected and according to the prediction of the NES in this area, localization of this fragment was dominantly localised in cytoplasm of both the COS-7 and N2a cells (Fig. 6Ba,a'). Leptomycin B (LMB) is known as an inhibitor of protein transport from the nucleus to the cytoplasm, which depends on a leucine-enriched sequence known as classical NES by interaction with exportin-1, which is responsible for protein export from the nucleus [27][28][29] . As predicted, the NESs were rich in leucine residues (see Fig. 5A). We used LMB to verify our hypothesis regarding the presence of an NES in the N-terminus of the bHLH. The presence of LMB resulted in a fluorescence signal that was present throughout the whole cell for both the COS-7 (Fig. 6Bb) and N2a (Fig. 6Bb') cells. To exclude the influence of LMB solvent (70% methanol), we added methanol as a control and observed cytoplasmic localization of YFP-TCF/550-601 (Fig. 6Bc,c') similarly to results without LMB addition. These results substantiated our hypothesis about the presence of active NES. By analysing different predictor results, we hypothesized that a sequence comprising aa 589-598 might be responsible for the NES activity in this fragment. To prove this hypothesis, we prepared a mutant where L596 and L598 were substituted by A (YFP-TCF4/550-601/ L596 A/L598A; Fig. 6A). Localization of this mutant changed from being dominantly cytoplasmic for the unmodified fragment to being distributed in both the nucleus and cytoplasm of the cell. Interestingly, the COS-7 and N2a cells presented a slightly higher fluorescence signal in the cytoplasm (Fig. 6Bd To verify the activity of the predicted NES in the C-terminal part of the bHLH domain, we prepared a YFP-TCF4/602-670 construct (Fig. 7A). The expression of this construct resulted in fully cytoplasmic fluorescence in the COS-7 cells and cytoplasm prevailing in the N2a cells (Fig. 7Ba). Next, we added LMB and observed a shift of localization from the cytoplasm to being equal in the cytoplasm and nucleus (Fig. 7Bb), presenting the inhibitory effect of LMB according to the prediction of the NES rich in L residues. Again, addition of methanol had no effect (Fig. 7Bc). After detailed analysis of the predictor results (see Fig. 5A), we chose the sequence comprising aa 604-618 as a hypothetic NES. To investigate this assumption, we prepared a mutant where L607 and L608 were substituted by A (YFP-TCF4/602-670/L607 A/L608 A; Fig. 7A). Localization of this mutant changed from being exclusively cytoplasmic, observed for the unmodified fragment, to being distributed equally in both compartments of the cell for bothCOS-7 and N2a cells (Fig. 7Bd).   Fig. 8Ba). However, for some cells we also observed equal distribution in both compartments of the cell (<10% cells; Fig. S2), and dominantly nuclear localisation for others (approximately 30% cells) (Fig. 8Bb). Repeated imaging 48 h after transfection showed a highly dominant nuclear localisation of the bHLH in more than 90% of cells (Fig. 8Bc), while in less than 10% of cells we could still observe fluorescence in the cytoplasm (Fig. 8Bd). In the case of the N2a cells, 24 h after transfection, dominant nuclear localisation was presented by more than 95% of cells (Fig. 8Be), and we observed cytoplasmic fluorescence in only some cells (<5%) (Fig. S2). All the above results suggest the presence of an active NLS in the bHLH-containing fragment. The cNLS Mapper predicted the region aa 569-603 as a putative bipartite NLS motif rich in basic amino acid residues (see Fig. 5B). Knowing that the N-terminal part of the bHLH (aa 550-601) presented strictly cytoplasmic localization (see Fig. 6Ba,a'), we deduced that basic amino acid residues K602 (predicted as part of NLS) and K606 (not predicted as part of the NLS, however situated in close proximity to the predicted NLS) in the C-terminal part of the bHLH are indispensable for this NLS activity. For this reason we decided to substitute K606 for A (Fig. 8A). The expression of this point mutant (YFP-TCF4/55-670/K606A) in the COS-7 cells mainly resulted (>70%) in equal or slightly more intensive fluorescence in the cytoplasm (Fig. 8Bf). However, for some cells (<30%) we could observe localization at the same time in the cytoplasm, but could only observe this in some subnuclear punctate aggregates (Fig. S2). In the case of the N2a cells, localization was slightly more cytoplasmic (>80%) (Fig. 8Bg). For less than 20% of the N2a cells, the fluorescence signal was distributed equally through the cell (Fig. S2). Change of localization after substitution of K606 by A proved the importance of this amino acid residue for the detected NLS activity and substantiates the thesis that the motif 569-RMANNARERLRVRDINEAFKELGRMVQLHLKSDKPQTK-606 (bold are basic residues, bold and underlined is lysine substituted residue) is the second active NLS (NLS-2) in TCF4. Out results are consistent with the results of Lingbeck et al. 40 , who documented the presence of the second NLS in the basic region of other proteins from the I class bHLH family: E12 and E47.  (Fig. 8Bh) and cytoplasm (Fig. 8Bi). This presented a pattern similar to the expression without LMB, with a delicate shift www.nature.com/scientificreports www.nature.com/scientificreports/ in the direction of nuclear localization (from about 30% to 50% of the analysed cells). We are not able to clearly explain why the NESs in the bHLH of TCF4 are LMB sensitive when separated, while the whole bHLH domain is less sensitive. However, we hypothesise that complex multi-element regulation might be the reason for the differentiated response to LMB in the case of the whole bHLH domain (containing two NESs and an NLS) in comparison to the single separated element (NES). LMB addition to the N2a cells expressing YFP-TCF4/550-670 had no visible effect (Fig. 8Bj), while without LMB this fragment was already predominantly nuclear in more than 95% cells, as discussed above. Controlled addition of methanol as a solvent of LMB had no effect for both the COS-7 (Fig. 8Bk,l) and N2a (Fig. 8Bm) cells.  www.nature.com/scientificreports www.nature.com/scientificreports/ We intended to further investigate how particular signals in the bHLH domain impact the localization of TCF4. First, we prepared a C-terminally truncated mutant without the bHLH domain, instead containing only NLS-1 (YFP-TCF4/1-549; Fig. 9A). The expression of this YFP-tagged mutant in the COS-7 cells was strictly nuclear, as expected, and presented no fluorescence in the nucleoli (Fig. 9Ba,a'). Some of the COS-7 cells (<10%) also presented highly fluorescent puncta in the nucleus (Fig. S2). The YFP-TCF4/1-549 expression in the neuronal N2a cells resulted in a pattern analogical to the COS-7 cells (Fig. 9Bb,b'). Next, we created YFP-TCF4/1-601 (Fig. 10A) containing both NLS-1 and NES-1. The expressed protein distribution was extremely diversified in both the used cell lines (Fig. S4). The COS-7 cells were classified as cells presenting a predominantly nuclear (>75% of cells) and predominantly cytoplasmic (<20%) localization (Fig. 10Ba-c), and also presenting an equal distribution in both compartments (<5%) (Fig. S4). Additionally, in most cases of nuclear localization, some puncta with highly intensive fluorescence in the nucleus were also visible. Similar results were observed for the N2a cells, with >70% of the population presenting a nuclear (Fig. 10Bd,e) and <30% cytoplasmic signal of fluorescence (Fig. 10Bf). The addition of methanol did not change distribution (Fig. 10Bg-j), while LMB shifted the localization to be exclusively nuclear (Fig. 10Bk,l), proving the active role of NES-1 in directing protein to the final destination. We hypothesise that the intensive shift from the cytoplasm to the nucleus in the case of TCF4/1-601 in comparison to the TCF4/550-670 fragment comes from the presence of an NLS in the N-terminal part of the protein. Previously, it was shown that cells expressing TCF4 with a disrupted N-terminal NLS (158-172aa) presented a slightly dominant nuclear localization. This suggests the synergistic effect of NLSs from the N-terminal and C-terminal parts of the TCF4.

Discussion
The subcellular distribution of TCF4 was proposed as one of the mechanisms regulating the function of this protein [63]. We showed that the previously documented NLS (NLS-1, Fig. 11A) 5 , located in the N-terminal part of the TCF4 protein (156-178aa), also possesses the putative activity of NoLS. We also showed that additional sequences may inhibit nucleolar localization. Our results are substantiated by a previous report stating that in cultured cortical-and hippocampus neurons and astrocytes, TCF4 in addition to nucleoplasm and pericarya was www.nature.com/scientificreports www.nature.com/scientificreports/ also present in the nucleoli. The authors also stated that PTHS-associated dominant-negative mutants R582P and R580Trp moderately reduced ribosomal biogenesis and general protein synthesis, pointing out the role of TCF4 in these processes 41 . www.nature.com/scientificreports www.nature.com/scientificreports/ Additionally, we identified the activity of two NESs (NES-1 and NES-2) and an overlapping NLS (NLS-2) in the bHLH domain of TCF4 (Fig. 11A). We performed alignment of TCF4-B+ isoforms of mouse and human TCF4, which present a high sequence identity (Fig. 11B), pointing out the conservation of the mosaic pattern of www.nature.com/scientificreports www.nature.com/scientificreports/ overlaying localization signals with opposing activities between these two organisms. The presence of the NLS and NES in close proximity, or overlapping in the bHLH domain, was reported previously for the bHLH-PAS transcription factor aryl hydrocarbon receptor (AhR) 45 and NPAS4 46 , while two NESs were shown in the bHLH of Sima (Drosophila homolog of HIF-1α) 47 . This suggests a more general function of the bHLH domain in the regulation of bHLH proteins shuttling. Laccase and Lefebvre proposed that overlaying of DNA binding motifs with localization signals allows for co-ordinate regulation 48 . We hypothesise that a system that regulates nucleocytoplasmic shuttling of these TF families is very complex and relies on many mutually dependent factors, both exportin-1 dependent and independent, like signal masking/unmasking, posttranslational modification and interaction with partner proteins such as calmodulin. This enables both precise and flexible signal transduction. Such complex multi-element regulation might also be the reason for the differentiated response to the LMB presence in the COS-7 cells during our experiments: high sensitivity in the case of separated NES-1 and NES-2 in the bHLH of TCF4, and low sensitivity in the case of the whole bHLH domain.
Localization studies of PTHS-connected point mutants performed previously, did not give consistent results, which was probably due to the different conditions of the performed experiments. Sepp et al. 52 expressed wt and R576Q, R578H, R580W, R582P and A614 PTHS-associated mutants of TCF4-B − proteins (isoform lacking RSRS) in HEK 293 cells, getting an analogical nuclear localization for both the wild type (wt) and all of the mutants. The expression of the bHLH tagged with GFP resulted in ubiquitous localization for wt and all the tested mutants. Forrest et al. 38 performed the expression of G358V, D535G, R578P, R580W and A614V mutants of the TCF4-B + isoform tagged with GFP in the COS-7 cells. Interestingly, while fluorescence for all the expressed proteins was visible only in the nucleus, the mutants R578P, R580W, and to a lesser extent A614V formed a small, spherical punctate, whose localization differed markedly from the distribution of the wild-type TCF4, G538V and D535G mutants. Similar results were obtained in SH-SY5Y neuroblastoma cells. We hypothesize that point mutations in the bHLH domain of TCF4, known for PTHS patients, influence pattern of distribution in nucleus and function of TCF4 probably by changes in DNA binding specificity. However, these single substitutions are not able to inhibit NLS-2 activity. In contrast, our experiments have led to the identification of amino acid residue K606 being indispensable for this NLS activity. Interestingly, this residue has not to date been connected to the PTHS disorder. Importantly, human aa residues located in the DNA binding region: R565, R576, E576, R578, R580, R587, K603 and K607 refer to murine aa residues R564, R575, E576, R577, R579, R586, K602 and K606 (Fig. 11B), which were pointed out by InterPro as being responsible for DNA binding Localization signals can be structurally exposed, accessible to transport machinery or masked by interacting partners. To visualise the arrangement of the detected NLSs and NESs in the TCF4, we generated a 3D model of the TCF4 using Phyre2 (Fig. 11C). We also performed in silico analysis of TCF4 structure disorder, and 89% of the TCF4 sequence was predicted to be intrinsically disordered. The only area with a tendency to possess a more rigid structure was the bHLH domain. Interestingly, for this area we obtained a prediction of the highest surface accessibility (Fig. S3). Results of repeated in silico analyses for the selected C-terminal part of the protein (550-670aa) containing the bHLH domain showed that short intrinsic disorder regions could exist in the external parts of this domain (Fig. S5). The preferential location of NLSs and NESs in the intrinsically disordered regions (IDRs) of proteins was found to enable flexible and easily accessible interactions with their binding partners 54 .
IDRs are known to be the targets of intensive posttranslational modifications. Modifications such as phosphorylation, especially close to the NLS or NES, have been shown to regulate the intracellular distribution of proteins by activating or deactivating localization motifs 55 . Interestingly, phosphorylation of S448, located very close to the bHLH domain, by protein kinase A (PKA) was documented by Sepp and co-workers as being necessary for TCF4 transcriptional activity in cultured neurons and in the developing brain in vivo 33 . We performed predictions of TCF4 phosphorylation sites and found many putative sites of phosphorylation, especially in the N-terminal part of protein with probability higher then 0.5, located in NLS-1 area (aa 156-178): S145, S152, S162, T169, T182. We observed much far less predicted sites located in the C-terminus of protein. However, some of the S and T residues of TCF4, predicted as putative phosphorylation sites with a probability higher or equal to 0.8, are located in the bHLH domain (S600) or in close proximity to this domain (S546, S548, S549, T557, S645, S646, S652). The location of putative phosphorylation sites in the C-terminal part of protein is presented in Fig. S5.
Importantly, using three programs, the motif 542-SITRSRSSNND-552 was predicted with a high probability as the phosphorylated (S548) binding site for 14-3-3 proteins, which were shown to influence the localization of partner proteins with intrinsically disordered regions 56 . The ubiquitous family of 14-3-3 proteins is involved in the regulation of signal transduction 57 . Binding by 14-3-3 modulates the enzyme activity, subcellular localization, structure, stability and molecular interactions of partner proteins 58 . In this aspect, very interesting is the observation by Liu et al. 59 that TCF4-B − isoform, which encodes full-length protein without motif 545-RSRS-548, significantly reduced the promoter containing E-box activity, while the splice variant TCF4-B + with RSRS did not affect this promoter activity in U1240MG cells. Taking all this into account, we hypothesize the importance of 14-3-3 interaction for the regulation of TCF4 localization, and in consequence, also this protein activity. On the other hand, we cannot exclude that a lack of phosphorylated S548 is the reason for the unbalanced and highly differentiated subcellular localization of the expressed TCF4/550-670.
The other mechanism for controlling TCF4 activity by differential subcellular localization could be Ca 2+ -dependent interaction with calmodulin, which was proposed as an important factor in the regulation of Scientific RepoRtS | (2019) 9:15629 | https://doi.org/10.1038/s41598-019-52239-w www.nature.com/scientificreports www.nature.com/scientificreports/ various target genes and cellular functions 60,61 . Though direct experimental data of how Ca 2+ could modulate TCF4 activity in neurons are still lacking, there are some hypothesis based on studies with class I and II bHLH factors about Ca 2+ -regulated shuttling of TCF4 from the cytoplasm to the nucleus 62 . Interestingly, it was shown that calmodulin binds to the basic sequence of E12, E47 and TCF4 63 , raising the possibility that calmodulin binding interferes with the activity of localization signals existing in this area.
In this paper, we documented the presence of multiple localization signals with opposing activities, which suggests the complex and precisely balanced regulation of TCF4 subcellular shuttling by masking and unmasking of different localization signals in the bHLH domain by interacting partners and posttranslational modifications. We performed alignment of the TCF4/568-645 region, encompassing NLS-2, NES-1 and NES-2 in the bHLH domain with corresponding sequences of E12 (aa 547-624), E47 (aa 544-621) and HEB (aa577-654). Sequence homology in these areas is very high (Fig. 11D), and we therefore propose that not only NLS-2, confirmed by Lingbeck et al. [69], but also NES-1 and NES-2 activities might be present in all mammalian E proteins. We therefore believe that these signals are important elements that regulate dynamic exchange between the subcellular compartments of not only TCF4, but generally bHLH class I transcription factors. The complex interplay of regulatory mechanisms of the intracellular localization of TCF4 and other I class bHLH transcription factors requires further studies in vivo. cell culture and DnA transfection. As described previously 65 African green monkey kidney fibroblasts COS-7 (ATCC CRL-1651) and Mouse Albino neuroblastoma Neuro 2a (ECACC; Sigma-Aldrich) were cultured in Dulbecco's modified Eagle medium (DMEM) with 4.5 g/l glucose and L-glutamine (Lonza). The culture medium was supplemented with 10% foetal calf serum (FCS). Cells were cultured at 37 °C in a 95% air/5% CO 2 atmosphere. The cells were transfected with 3 μg of DNA/150,000 cells using Xfect Transfection Reagent (Takara Clontech) according to the manufacturer's instructions. LMB (Sigma) was dissolved in 70% methanol at a concentration of 10 −3 M and added to the medium during transfection at a final concentration of 10 −6 M. electrophoresis and Western blot analysis. To confirm the expression of TCF4 and TCF4 mutants in cultured COS-7 cells, additionally to the microscopy experiments, SDS-PAGE and Western blot analysis were performed. Samples of the total protein extract were prepared by replacing the cell medium 24 h after transfection (after washing cells with PBS) with a 2x SDS gel-loading buffer 66 , transferring extracts to Eppendorf tubes, boiling the samples for 5 min and then centrifuging the samples for 5 min (13,000 g). Proteins were separated by 10% SDS-PAGE and transferred to a Whatman PROTRAN Nitrocellulose Transfer Membrane (PROTRAN BA85, Schleicher and Schuell Pure, Sigma-Aldrich) with mini Trans-Blot apparatus (Bio-Rad). The membrane was blocked at room temperature (RT) for 1 h in 3% milk blocking buffer (milk powder; Milchpulver, blotting grade, Roth) that was prepared in PBS supplemented with 0.2% Tween 20 (Sigma). Next, the membrane was incubated overnight at 4 °C with anti-GFP polyclonal antibodies (Clontech) (diluted 1:300 with milk buffer), which cross-reacted with YFP. Secondary goat anti-rabbit antibodies coupled to horseradish peroxidase (Vector Laboratories) were added (1:10,000 with milk buffer) and incubated at RT for 2 h. Blots were developed using Pierce ECL Plus Western Blotting Substrate (Thermo Scientific) according to the manufacturer's manual. Confocal and fluorescence microscopy. As described previously 46,67 , prior to experiments of confocal microscopy imaging, cells were seeded onto 0.17-mm-thick round glass coverslips (Menzel) and submerged in a culture medium in 2-cm diameter Petri dishes. Next, 24 h, or in some mentioned cases 48 h after transfection, the coverslips with cell cultures were transferred onto a steel holder and mounted on a microscope stage. The standard culture medium was replaced with 1 ml of DMEM/F12 without phenol red, buffered with 15 mM HEPES (Sigma) and supplemented with 2% FBS (Sigma). Next, 0.1 μl of Draq5 DNA Dye (BioStatus) was added to the cells for DNA visualization. During microscopy observation, the cell culture temperature was stabilized at 37 °C using a microincubator (Life Imaging Services Box & Cube). Images of the fluorescently labelled proteins were acquired using a Leica TCS SP5 II confocal system equipped with argon and helium neon lasers and a 63x oil objective lens (NA: 1.4). YFP was excited using 514 nm light and the emitted fluorescence was observed at a range of 525-600 nm. The excitation of Draq5 was 633 nm and emission was 650-800 nm. Gamma correction of 0.6 was set for the Draq-5 channel. At least 50 cells were observed for each cDNA construct in one experiment and we calculated the cell percentages by dividing number of cells presenting exact pattern of localization to the total number of observed cells multiplying it by 100%. At least three independent experiments were performed for each cDNA construct. Images are presented for representative cells with a typical phenotype, which is characteristic for more than 95% of cases. In the case of cells presenting many phenotypes, representative images of all the cell phenotypes are presented. Ratios between mean fluorescence intensity in cytoplasmic, nuclear and nucleolar compartments were estimated by measuring mean fluorescence intensity inside each compartment of chosen cell