Dyslexia, or specific reading disability, is the unexpected failure in learning to read and write when intelligence and senses are normal. One of the susceptibility genes, DYX1C1, has been implicated in neuronal migration, but little is known about its interactions and functions. As DYX1C1 was suggested to interact with the U-box protein CHIP (carboxy terminus of Hsc70-interacting protein), which also participates in the degradation of estrogen receptors alpha (ERα) and beta (ERβ), we hypothesized that the effects of DYX1C1 might be at least in part mediated through the regulation of ERs. ERs have shown to be important in brain development and cognitive functions. Indeed, we show that DYX1C1 interacts with both ERs in the presence of 17β-estradiol, as determined by co-localization, co-immunoprecipitation and proximity ligation assays. Protein levels of endogenous ERα or exogenous ERβ were reduced upon over-expression of DYX1C1, resulting in decreased transcriptional responses to 17β-estradiol. Furthermore, we detected in vivo complexes of DYX1C1 with ERα or ERβ at endogenous levels along neurites of primary rat hippocampal neurons. Taken together, our data suggest that DYX1C1 is involved in the regulation of ERα and ERβ, and may thus affect the brain development and regulate cognitive functions. These findings provide novel insights into the function of DYX1C1 and link neuronal migration and developmental dyslexia to the estrogen-signaling effects in the brain.
DYX1C1 was the first susceptibility gene suggested for dyslexia (1), the most common childhood learning disorder. The neurobiology of dyslexia remains largely unknown but evidence suggests that neuronal migration and axon guidance pathways may be disrupted in dyslexia. It has been shown that individuals with dyslexia have neocortical neuronal migration abnormalities including molecular layer heterotopias, laminar dysplasias and periventricular nodular heterotopias (2). The knockdown of Dyx1c1 by in utero RNA interference (RNAi) in rats causes a deficit in neuronal migration in the developing neocortex, supporting the role of DYX1C1 in dyslexia (3,4). Further examination of rats subjected to Dyx1c1 RNAi showed behavioral impairments in auditory processing and spatial learning (5). Also other dyslexia candidate genes have shown to regulate neuronal migration (6,7) and axonal guidance (8).
The estrogen receptors (ERs) and their endogenous ligand 17β-estradiol (E2) have been recognized to be important in brain development (9–12) and in neuronal processes such as neuronal differentiation, survival and plasticity (13–18). ERs have also been shown to be involved in cognitive processes and memory (18–21). ERα and ERβ display different expression patterns during brain development, suggesting distinct developmental functions for these two related receptors (22–24). Interestingly, ERβ has a role in neuronal migration and neuronal survival in the developing cortex, and Erβ knock-out (KO) mice present very similar phenotypes as the post-mortem brains of dyslexic individuals (9,11).
The molecular and cellular functions of DYX1C1 have been poorly characterized so far. On the basis of its protein domain structure, DYX1C1 may be engaged in various protein–protein interactions and function in multiprotein complexes. Indeed, Hatakeyama et al. (25) showed that DYX1C1 (then called EKN-1) interacts with the U-box family ubiquitin protein ligase CHIP (carboxy terminus of Hsc70-interacting protein) encoded by the gene STUB1. CHIP ubiquitylates and promotes the degradation of a variety of proteins in a chaperone-assisted manner (26). Intriguingly, CHIP targets also the ERs for degradation by at least two different mechanisms: CHIP-mediated degradation of ERα is ligand-independent and blocked when estrogen is added, whereas ERβ degradation is strictly estrogen-dependent (27–29).
In the current study, we aimed to elucidate novel protein–protein interactions and cellular roles of DYX1C1. We hypothesized that DYX1C1 might interact with and be involved in the regulation of ERs, thereby joining two pathways implicated in brain development, higher cognitive functions and etiology of dyslexia.
DYX1C1 and ligand-activated ERs co-localize in mammalian cells
We first determined the intracellular co-localization of over-expressed DYX1C1, ERα and ERβ in COS-7 cells. DYX1C1 localized both in the nucleus and in the cytoplasm, whereas ERs were exclusively localized in nuclei (data not shown), consistent with previous findings (1,3,30). Upon co-expression, DYX1C1 and ERα had a diffuse localization in the nucleus in the absence of ER ligand (Fig. 1B). A distinct co-localization pattern of DYX1C1 with ERα was detected when cells were treated with the agonist 17β-estradiol (E2). This clear punctuate nuclear pattern of co-localization was not seen when the cells were treated with the antagonist tamoxifen (4-hydroxytamoxifen—4-OHT) (Fig. 1B). To confirm the specificity of this co-localization, we used different expression constructs of DYX1C1 and ERα (ERα-V5 and DYX1C1-myc) with similar results (data not shown). Moreover, identical co-localization patterns were obtained for DYX1C1 and ERβ (Supplementary Material, Fig. S1A).
To determine whether the co-localization of ERα and DYX1C1 depended on the helix 12 of ERα, which is part of the conserved estrogen-dependent co-regulator surface called AF-2 (Fig. 1A) (31), we transiently transfected the ERα H12-truncated mutant (ERαΔH12-VP16) or wild-type ERα-VP16 together with DYX1C1. We observed that ERαΔH12 co-localized only partially, compared with wild-type ERα, with DYX1C1 in the presence of E2 (Fig. 1C). Finally, to compare the DYX1C1/ER co-localization pattern, we expressed the established co-regulator RIP140 (32) with ERα in COS-7 cells (Fig. 1D). The detected E2-dependent co-localization between RIP140 and ERα showed a similar pattern as DYX1C1 and ERα. Taken together, these results suggest that DYX1C1 interacts with ligand-activated ERs and shares features with a variety of established ER co-regulators (31).
ER interactions are mediated by the p23 domain of DYX1C1
To further study the DYX1C1/ERα interaction, we performed co-immunoprecipitations (Co-IPs) using SH-SY5Y cell extracts expressing ERα and three different V5-tagged DYX1C1 constructs (Fig. 2A): one encoding the full-length protein (DYX1C1-V5), one lacking the three TPR domains (DYX1C1ΔTPR-V5), and one lacking the p23 domain (DYX1C1Δp23-V5). Co-IP using antibody against ERα revealed that full-length DYX1C1 and DYX1C1ΔTPR associated with ERα, although deletion of the p23 domain weakened the interaction (Fig. 2B). In addition, by Co-IP using an antibody against CHIP, we verified the interaction between CHIP and DYX1C1 and determined which domain of DYX1C1 was needed for binding. As expected, full-length DYX1C1 associated with CHIP (Fig. 2B). We detected the CHIP complex also with DYX1C1ΔTPR-V5, but no complex was observed with DYX1C1Δp23-V5 (Fig. 2B). The DYX1C1 domain requirement for ER interaction was further investigated by co-localization analysis in COS-7 cells. It has previously been shown that the N-terminus of DYX1C1 lacking the TPR domains localized predominantly in the nucleus and C-terminus of DYX1C1 lacking the p23 domain localized predominantly in the cytoplasm in COS-7 cells (3). In our experiments, we observed that both deletion constructs were found in the nucleus as well as in the cytoplasm (data not shown). Upon co-expression, DYX1C1ΔTPR-V5 co-localized with ERα similarly to the full-length DYX1C1 (Fig. 2C), although deletion of the p23 domain abolished co-localization (Fig. 2C). Similar results were obtained with ERβ (Supplementary Material, Fig. S1B). The Co-IP and co-localization results together suggest that the interactions between DYX1C1 with ERs are mediated mainly by the N-terminal p23 domain, which is also needed for CHIP interaction.
DYX1C1 over-expression reduces ER protein levels and transcriptional activity
As DYX1C1 co-localizes and interacts with both ERα and ERβ, we tested next whether DYX1C1 might regulate the protein levels of ERs. First, we transiently transfected increasing amounts of DYX1C1-V5 into human MCF-7 breast cancer cells that express endogenous ERα. We found that over-expression of DYX1C1 decreased steady-state protein levels of ERα in a dose-dependent manner (Fig. 3A). Consistent with the interaction studies, the decrease of ERα protein levels was greater in the E2-treated samples, in which over-expression of DYX1C1 reduced ERα levels to as low as 30% of the control levels. Furthermore, treatment with the ER antagonist 4-OHT partially blocked the down-regulatory effect of DYX1C1 (Supplementary Material, Fig. S2A). To determine whether proteasome activity is needed for the DYX1C1-mediated ERα down-regulation, we incubated the cells first with E2 and then for 6 h with the proteasome inhibitor MG132. The results indicate that MG132 blocked the down-regulation of ERα protein levels mediated by DYX1C1 (Supplementary Material, Fig. S2B).
To investigate the effect of DYX1C1 on ERβ stability, we used the SH-SY5Y neuroblastoma cells to over-express ERβ together with DYX1C1. We transiently co-transfected a stable amount of ERβ (200 ng) and increasing amounts of DYX1C1. The cells were treated with E2 for 18 h and DMSO was used as a control. We detected a clear decrease in the protein levels of ERβ in the presence of E2 (Fig. 3B, lanes 5–8) and the absence of ligand (Fig. 3B, lanes 1–4). Thus, over-expression of DYX1C1 had relatively similar down-regulatory effects on both ER subtypes, consistent with similar interaction features.
To test whether down-regulation of ERs mediated by DYX1C1 had functional consequences on the transcriptional responses of ER upon ligand activation, we employed transient transfection-based reporter assays. We co-transfected ERα-positive MCF-7 cells with estrogen response element (ERE2) containing luciferase reporter, DYX1C1 and pRL-TK (containing Renilla luciferase for normalization of transfection efficiency). To test the ERβ transcriptional activity, the SH-SY5Y cells were co-transfected with ERβ, ERE2 luciferase reporter, DYX1C1 and pRL-TK. We treated cells with E2 for 18 h and measured the luciferase activity. We found that DYX1C1-mediated ER down-regulation significantly diminished ERE2-mediated reporter gene expression in MCF-7 cells corresponding to the ERα signaling (Fig. 3C) and in SH-SY5Y cells corresponding to the ERβ signaling (Fig. 3D). These results indicate that DYX1C1 can negatively modulate transcriptional ER-signaling pathways.
Co-localization experiments were designed to address the effect of over-expressed DYX1C1 wild-type and domain-deleted derivatives (Fig. 2A) on the stability of endogenous ERα in MCF-7 cells (Fig. 3E). Confocal fluorescence images demonstrate that ERα staining was abolished in cells expressing DYX1C1 wild-type or slightly diminished when expressing DYX1C1ΔTPR, whereas it was unaffected in cells expressing DYX1C1Δp23. These results are consistent with the ER interaction requirement of the p23 domain (Fig. 2) and suggest that p23 domain-mediated interactions are a prerequisite for proteasomal down-regulation of ER.
In vivo association of endogenous rat Dyx1c1 and ERs in the neurites of primary hippocampal neurons
We finally studied whether the interactions shown between the ERs and DYXC1 could be verified under endogenous conditions, i.e. cell expressing native protein levels. For this, we used primary hippocampal neurons derived from E17 rat brain. To visualize native protein complexes, we used the in situ proximity ligation assay (in situ PLA) (33–35) and were able to detect protein complexes that included Dyx1c1 with either Erα or Erβ in the presence of E2 (Fig. 4A and D). Interestingly, the interactions occurred along the neurite extensions of the hippocampal neurons when cells were treated with E2. When no ligand was added, less or no interactions were seen (Fig. 4B and E), consistent with the interaction requirements observed with over-expressed proteins as shown in Figure 1. Negative controls omitting DYX1C1 primary antibody demonstrated the specificity of the assay (Fig. 4C and F). As a further control of specificity of the proximity ligation assay and the ERα antibody, we performed in situ PLA in MCF-7 cells and detected in vivo interactions of ERα with the established steroid receptor co-activator SRC-3 (also NCoA3, AIB-1, ACTR) and CHIP (Supplementary Material, Fig. S3).
In this study, we report an interaction between the two ER subtypes and DYX1C1 by several parallel methods and demonstrate that over-expression of DYX1C1 regulates the levels of ERα and ERβ in a dose-dependent manner, accompanied by a decrease in the transcriptional activity of the ERs. Finally, we detected native complexes of Dyx1c1/Erα and Dyx1c1/Erβ along extensions of primary rat hippocampal neurons.
The DYX1C1 protein has two recognizable motifs, a p23 domain in the N-terminus and three TPR domains in the C-terminus. The functional importance of the DYX1C1 TPR domains was previously demonstrated when impaired neuronal migration caused by siRNA-mediated down-regulation of DYX1C1 was rescued with the expression of a DYX1C1 TPR domain derivative (3). Although the TPR domains appear dispensable for ER interactions, it remains to be seen whether they participate in ER regulation in in vivo conditions. Indeed, TPR domains are functional in diverse chaperone, cell cycle, transcription and protein transport complexes (36). P23 domains are generally implicated in stabilizing intramolecular protein folding and/or in mediating protein–protein interactions (37). Our results specifically suggest that the DYX1C1 p23 domain is necessary for interactions with ERs. Although it is possible that DYX1C1 interacts indirectly with the ERs via CHIP bound to the p23 domain, robustness of the co-localization patterns upon over-expression indicates that the interaction is not mediated by limiting endogenous bridging factors. Interestingly, three independent studies reported opposite effects of CHIP on the degradation of over-expressed ERα versus ERβ (27–29). In contrast, we demonstrate that DYX1C1 interacts with and down-regulates levels of both ERs in a similar manner, suggesting that DYX1C1 might function without CHIP. In support of this assumption, our direct experimental comparison suggests DYX1C1 to share features with two most relevant ER co-regulators, namely the co-repressor RIP140 and the co-activator SRC-3. Common features include (i) recognition of the activated ER conformation, i.e. induced by agonistic ligands, (ii) requirement for helix 12, part of the ligand-dependent activation domain AF-2 within the ligand-binding domain of ERs, (iii) abolishment of interactions upon treatment with ER antagonists, i.e. ligands such as tamoxifen that induce a distinct ‘non-productive’ ER conformation, including re-arrangements of helix 12 (reviewed in 31). Whether DYX1C1 directly binds to the conserved AF-2 surface remains to be seen, but leucine-rich peptide motifs in the DYX1C1 p23 domain might present candidate ER-binding domains and will be subjected to further investigations. Irrespective of the binding mode, interaction with DYX1C1 has negative consequences on ER stability and transcriptional activation. Thus, DYX1C1 represents a candidate co-regulator of ER that integrates two mechanistically distinct, yet functionally linked, mechanisms.
It is tempting to speculate that DYX1C1 modulates cellular ER functions within at least two signaling pathways referred to as ‘genomic’ and ‘non-genomic’ (reviewed in 31) (Fig. 5). In brief, genomic (also nuclear) actions of ERs describe the direct regulation (i.e. activation or repression) of transcription in conjunction with numerous co-regulators. Non-genomic (also extranuclear, non-transcriptional) actions involve cytoplasmic ERs that are activated, for example, by rapid phosphorylation-signaling cascades (38). Indicative of both genomic and non-genomic actions, we detected native complexes of ERs and DYX1C1 in primary hippocampal neurons with a distinct pattern along the neurite extensions (Fig. 4). Indeed, both ERs have been found at extranuclear sites in different brain regions such as hippocampus and cortex (39–41), which suggest ERs to be important in the rapid activation of different signaling pathways such as p42/p44 MAPK and MAPK ERK1/2 (38). Furthermore, Xu et al. (42) also showed that human ERα is expressed in punctuated areas within neurites and suggested that ERα regulates compartmentalized hormone-mediated modulation of signaling systems within neurons. Also CHIP has been detected along neurites of NT2-N cells in glucocorticoid receptor transport complexes together with Hsp90 (43).
Early observations by Galaburda and colleagues show that dyslectic individuals, both males and females, have abnormalities in the cerebral cortex such as molecular layer ectopias and microgyria (44–46). Three dyslexia candidate genes including DYX1C1 have been shown to be involved in neuronal migration leading to neocortical malformations in rodents (3,4,6,7). In cortical development, neurons proliferate in the ventricular (VZ) and subventricular zones and subsequently migrate along radial and tangential pathways to the cortical plate, which will later form the six-layered neocortex. Both ERs are expressed in the developing brain, but distribution pattern of their expression is still unclear (23,47,48). The importance of the expression of ERβ during brain development has been shown with studies on Erβ KO mice that demonstrate deficits at the time when post-mitotic neurons migrate to the cortex. The deficits were caused by abnormal neuronal migration and an increased level of apoptotic neuronal death (9,11). Since no brain abnormalities have been detected in the Erα KO mouse, the importance of ERα in cortex development remains to be seen. One suggested mechanism is that ERα is important for the function of Cajal–Retzius cells, which produce the extracellular matrix protein Reelin, important for the correct positioning of cortical neurons (23). Also localization of ERα in the embryonic VZ cells has been demonstrated with the majority of the VZ cells expressing ERα in the cytoplasm and membrane (49).
During adult life in humans, the protein levels of ERα are lower than those of ERβ in the cortex and hippocampus. Studies on Erα and Erβ KO mice suggest that both are needed for learning and higher cognitive functions (50). In the absence of the ERs, spatial learning is impaired (21,51). In a recent study, the authors show that the effects of estrogen on hippocampal synaptic plasticity and hippocampus-dependent memory functions are mediated through ERβ (18). Interestingly, spatial learning is impaired also in the rats that have received in utero siRNA against Dyx1c1 (5).
In less than 5 years, six susceptibility genes for dyslexia have been identified. Four of these genes DYX1C1, DCDC2, KIAA0319 and ROBO1 have been shown to be involved with neuronal migration and axon guidance, thus implicating common pathways or interaction between these genes. For this reason, it is very important to understand the molecular function of the genes and their interacting partners. Genetic association studies have been ambiguous in replicating the effects of DYX1C1 on dyslexia. However, recently, new insights support the role of DYX1C1 as a candidate gene for dyslexia. Two new SNPs in association with dyslexia have been identified in the DYX1C1 promoter region, and a meta-analysis of existing association studies also supports the association of DYX1C1 to dyslexia. DYX1C1 seems to have a sex-specific risk effect on dyslexia, which is interesting in the light of dyslexia being more common in boys than in girls (52,53). We also recently showed that a complex of TFII-I, SFPQ and PARP1 regulates the levels of DYX1C1 in an allele-specific manner (54).
Our data presented in this study links DYX1C1 to the regulation of the estrogen pathway, which is known to be important in brain development. On the basis of previous results on the rapid signaling of ERs in the extra-nuclear compartments and of our findings, we speculate that DYX1C1 could affect the rapid non-genomic signaling in brain through the ERs (Fig. 5). These findings provide insights into the function of DYX1C1 that potentially links neuronal migration and developmental dyslexia to the effects of ERs in the brain.
MATERIALS AND METHODS
Cell culture and transient transfections
The African green monkey kidney fibroblast cell line COS-7 was grown in Dulbecco's modified Eagle's medium (DMEM) containing GlutaMAX-I and supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 µg/ml streptomycin. For the localization experiments, the cells were grown on glass coverslips on 24-well plates for 24 h in phenol red-free DMEM/F12 medium containing 2.5 mml-glutamine, supplemented with 10% dextran charcoal (DCC)-treated FBS and with 100 U/ml penicillin and 100 µg/ml streptomycin when followed by transfections with Fugene 6 (Roche).
The human breast cancer cell line MCF-7 was cultured in DMEM with l g/l glucose, 1% l-glutamine, 10% FBS and 100 U/ml penicillin and 100 µg/ml streptomycin. For all the experiments, cells were cultured 24 h in phenol red-free DMEM, supplemented with 4% of dextran–charcoal-treated FBS.
The human neuroblastoma SH-SY5Y cell line was cultured in modified essential media (MEM) with Earle's salts and GLUTAMAX-I supplemented with 10% FBS, 100 U/ml penicillin and 100 µg/ml streptomycin. For the experiments, cells were cultured 24 h in phenol red-free MEM, supplemented with l-glutamine and 4% of DCC-treated FBS. Transfections were performed using either Lipofectamine 2000 (Invitrogen) or FuGene 6 (Roche), following the recommendations from the manufacturers. The cells were treated with 17β-estradiol (E2) (Sigma) (10 nm), 4-hydroxytamoxifen (4-OHT) (Sigma) (100 nm), MG-132 (Sigma) (10 nm) and DMSO as control, as indicated for each experiment.
Hippocampal neuron cultures were prepared from the brains of E17 rat fetuses as described previously (55). Briefly, the hippocampi were dissected, the meninges removed and the neurons dissociated in single-cell suspension with papain (0.5 mg/ml) digestion and mechanical trituration. The cells were centrifuged, suspended in DMEM containing Glutamax I and supplemented with 10% FBS, 100 U/ml penicillin and 100 µg/ml streptomycin (DMEM medium; Gibco BRL). In total, 100 000 cells/well on 12-well plates were plated onto glass cover slips coated with 0.001% poly-l-ornithine (Sigma) and 10 mg/ml laminin (Invitrogen). The cells were cultured in phenol red-free neurobasal medium (lutGibco) supplemented with B27 (Gibco), penicillin–streptomycin (Euroclone) and l-gamine (Euroclone) at 37°C in 5% CO2 for 12 h. Subsequently, the neuronal cells were treated with 1 nm E2 (Sigma) for 48 h before fixation.
Preparation of expression constructs
The full-length DYX1C1-V5 expression construct was cloned in pcDNA3.1/V5-His-TOPO vector as described earlier (1). The DYX1C1Δp23-V5 lacks the p23 domain (residues 7–103 in NP_570722.2, nucleotides 387–677 in NM_130810.2) and has a V5-His tag at the C-terminus. A PCR fragment was amplified using the full-length DYX1C1-V5 plasmid as template. The forward primer (5′-ACCGGAATGCCTCTTCAGGTTAGCATTTTACAAGCACAAGAGAG-3′) had been designed by combining the sequences flanking on both sides of the p23 deletion breakpoint and includes the initiation codon. The reverse primer (5′-AGATTTTAGTTCTGTTCCTTGAAT-3′) had been designed to anneal at the 5′ side of the stop codon. The PCR fragment containing the p23 deletion was directly cloned into the pcDNA-V5-His vector using the TOPO TA cloning kit (Invitrogen) according to the manufacturer's instructions. The DYX1C1ΔTPR-V5 construct lacks all of the three TPR domains (residues 290–399 in NP_570722.2, nucleotides 1236–1565 in NM_130810.2) and has a V5-His tag in the C-terminus. Two PCR fragments were first amplified from each side of the deletion. The 5′ portion of the insert was amplified using a forward primer that anneals upstream of the DYX1C1 initiation codon (5′-CAAGAATCGGCATCACTCT-3′) and deletion-specific reverse primer (5′-AATTTGTACAATTTTGTTCTTTTCTTCTTC-3′). The 3′ portion was amplified using deletion-specific forward primer (5′-GAAGAAGAAAAGAACAAAATTGTACAAATT-3′) with the reverse primer that anneals at the 5′ side of the stop codon of DYX1C1 (5′-AGATTTTAGTTCTGTTCCTTGAAT-3′). The primary PCR products were mixed and used as templates in a secondary PCR, with the forward and reverse primers used in the first PCR reactions that were not deletion-specific. The PCR fragments containing deletions were directly cloned into the pcDNA-V5-His vector using the TOPO TA cloning kit (Invitrogen) according to the manufacturer's instructions. Epitope-tagged expression constructs for human ERα, ERβ and RIP140 were described previously (30,32). psG5-ERα, psG5-ERβ and pcDNA3.1-CHIP were also previously described (56–58). The identity of each construct was confirmed by DNA sequencing.
The cells were grown 24 h after transfecting them with expression constructs and subsequently treated with E2, 4-OHT or EtOH for 1–2 h. The cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 5 min. After the fixation, the cells were made permeable with 0.1% Triton X-100 in PBS. The cells were incubated in 3% bovine serum albumin (BSA) in PBS for 30 min to block unspecific binding of the antibodies. Subsequently, the cells were incubated with primary antibodies for 45 min followed by three washing steps with PBS and staining with secondary antibodies for 45 min. Finally, the nuclei were stained with 4′,6-diamino-2-phenylindole (DAPI, Sigma-Aldrich), and the cells were mounted on glass slides with ProLong Gold antifade reagent (Invitrogen).
The primary antibodies used in the immunocytochemical stainings were anti-V5 (R960, Invitrogen and NB-600-381, Novus Biologicals), anti-myc (9E10, Berkeley Antibody Company), anti-VP-16 (sc-7545, Santa Cruz Biotechnology), anti-HA (MMS-101R Berkeley Antibody Company) and anti-ERα (sc-8002, Santa Cruz Biotechnology). The secondary antibodies used were Alexa Fluor 488 goat anti-rabbit IgG (H+L), Alexa Fluor 555 goat anti-mouse IgG (H+L) and Alexa Fluor 647 donkey anti-rabbit IgG (H+L) (Invitrogen).
The confocal immunofluorescence images were acquired with Meta LSM 510 confocal imaging system and Axiovert 200M microscope (Carl Zeiss) using 63× NA 1.4 or 40× NA 1.2 objective and multichannel scanning in frame mode. Filter settings for the blue, green, red and far-red channels for excitation and emission were 405/BP 420–480, 488/BP 505–530, 543/BP 585–615 or LP575, 633/LP 650, respectively. Zeiss LSM Image Browser (Carl Zeiss) was used to edit the confocal images. The brightness and contrast were enhanced for presentation. In the triple transfections, the far-red channel is pseudo-colored as blue and the DAPI channel as white.
For total extracts, cells were harvested in RIPA buffer [50 mm Tris–HCl, pH 7.4, 1% Triton X-100, 150 mm NaCl, 5 mm EDTA, 1 mm PMSF, 1 mm Na3VO4, 1 mm DTT and protease inhibitors (Roche)], followed by incubation on ice for 30 min (59). Cell lysates were centrifuged at 14 000g for 15 min, and the resulting supernatants were collected. Protein concentrations of the extracts were measured with the protein assay (Bio-Rad Laboratories), with BSA as standard.
SH-SY5Y cells were transiently transfected with the corresponding plasmids in six-well plates; after 24 h, cells were harvested with RIPA or NP-40 buffers with protease inhibitors (Roche). Protein extracts were incubated for 2 h at 4°C with 40 µl of protein G-sepharose slurry (GE Healthcare) and 1 µg of the respective antibody, either ERα (sc-8002) or CHIP (sc-33264, Santa Cruz Biotechnology) in IP-T150 buffer (50 mm Tris–HCl, pH 7.4, 150 mm NaCl, 0.2% Nonidet P-40, 1 mm EDTA and 10% glycerol). After incubation, the beads were washed three times using the IP-T150 buffer. For protein elution, beads were incubated with 1× SDS sample buffer, denatured by boiling, and protein extracts were resolved in NuPage polyacrilamide 10% gels (Invitrogen). Western blot analysis was performed using a V5-HRP antibody (R961-25, Invitrogen, 1:8000).
Western blot analysis
Protein extracts derived from the SH-SY5Y and MCF-7 cells were separated on 10% NuPAGE Bis–Tris gels (Invitrogen) in NuPAGE MOPS SDS running buffer and electroblotted to PVDF Hybond-P transfer membranes (GE Healthcare Bio-Sciences, Little Chalfont, UK). After transfer of proteins, the filters were blocked for unspecific protein binding by incubation 1 h at RT of 5% non-fat dry milk in 1% Tween-PBS. Subsequently, filters were incubated for 1 h with the primary antibody, washed in Tween-PBS for 15 min and three times 5 min followed by incubation with the secondary antibody 1 h. Filters were washed again at the same conditions as above, and detection was performed using the ECL advance western blotting detection kit (GE Healthcare).
The primary antibodies used in the western blots were anti-V5- HRP (R961-25, Invitrogen, 1:8000), anti-ERα (sc-8002, Santa Cruz Biotechnology, 1:200), anti-β-actin (ab6276, Abcam 1:5000) and anti-ERβ (ab14021, Abcam, 1:2000). The secondary antibodies used were ECL anti-mouse-HRP (NA931V, GE Healthcare) and anti-chicken-HRP (Ab8924, Abcam).
Relative protein levels of ER and β-actin proteins were measured using ImageJ software from the scanned western blots film. ER protein levels were adjusted against β-actin protein levels.
Luciferase reporter assay
The MCF-7 cells were seeded in 24-well plates 1 day prior to the transfections. The cells were co-transfected with 100 ng of ERE2-TK-luciferase plasmid (60), 100 ng DYX1C1-V5 plasmid and 10 ng of the pRLTK vector containing the herpes simplex thymidine kinase promoter linked to a constitutively expressing Renilla luciferase reporter gene (Promega) using Lipofectamine 2000 reagent according to the manufacturer's protocol. Six hours after transfections, the media were changed and the cells were treated with E2 or DMSO as control. After 18 h of treatment, proteins were extracted using passive lysis, and luciferase activity was measured using the dual-luciferase reporter assay system (Promega). All data were normalized against Renilla luciferase. The statistical significance was tested using Student's t-test with a threshold of P < 0.05.
Proximity ligation in situ assays
MCF-7 cells were cultured in Lab-Tek™ II chamber glasses (Nunc, Thermo Fisher Scientific) for 24 h. The cells were treated with 10 nm E2 for 6 h, washed in cold PBS and fixed in ice-cold 70% ethanol for 1 h. The hippocampal neurons were plated onto glass cover slips coated with 0.001% poly-l-ornithine (Sigma) and 10 mg/ml laminin (Invitrogen). Subsequently, the neuronal cells were treated with 1 nm 17-β-estradiol (Sigma) or ethanol for 48 h before fixation. The chamber wells were separated from the glass, and a hydrophobic barrier between individual wells was made using an ImmEdge pen (H-4000, Vector Laboratories). Blocking, antibody hybridizations, proximity ligations and detections were performed according to recommendations (Duolink IQ, OLINK Bioscience).
Antibodies used for the detection of protein–protein interactions were anti-NCoA-3 (sc-25742, Santa Cruz Biotechnology, 2 µg/ml), anti-ERα (sc-8002, Santa Cruz Biotechnology, 2 µg/ml, GTX22746, GeneTex, 3 µg/ml), anti-ERβ (GTX14021, GeneTex, 1:1000), anti-CHIP (GeneTex, GTX22917 2 µg/ml and Santa Cruz Biotechnology, sc-33264, 1 µg/ml), anti-β-actin (FITC) (ab6277, Abcam, 1:100) anti-DYX1C1 sera (Sigma-Genosys, 1:1000) made using synthetic peptide with the N-terminal epitope KAKIGNDTIVFTLY-C.
The proximity ligations were performed according to protocol with minor modifications. Briefly, after incubation with primary antibodies, we applied combinations of corresponding PLA probes (i.e. anti-rabbit PLUS, anti-mouse MINUS and anti-goat MINUS PLA probes) for 1 h at 37°C. Subsequent hybridizations, ligations and detections using DuoLink™ 100 Detection Kit 563 (OLINK Bioscience) were performed. DuoLink 100 Detection Kit 563 includes a Tye 563 fluorophore with excitation at 557 nm and emission at 563 nm and Hoechst 33432 nuclear dye. PBS with 0.05% or 0.1% Tween-20 (PBS-T) was used for washing. Preparations were mounted in DuoLink mounting medium (OLINK Bioscience). Assays was photographed using an Olympus IX71 inverted fluorescence microscope with oil immersion lenses at ×60 or ×100 magnifications. Filter settings for the Hoechst 33432, FITC and red signals for excitation and emission were 350/50, 480/40, 565/30 nm and 460/50, 535/50, 620/60 nm, respectively. Brightness and contrast of photos from the microscope were enhanced for presentation using the ImageJ 1.38× software, and figures were assembled in Adobe Illustrator CS2 and Adobe Photoshop CS3 software.
The study has been supported by Swedish Research Council, Swedish Royal Bank Tercentennial Foundation, Swedish Brain Foundation (Hjärnfonden), Knut and Alice Wallenberg Foundation, Swedish Cancer Foundation, Sigrid Jusélius Foundation, Päivikki and Sakari Sohlberg Foundation, Osk. Huttunen Foundation, Academy of Finland and European Union (Enlight and CRESCENDO).
We thank the members of the Unit for Receptor Biology for providing materials, Nina Gustafsson for excellent technical support and Myriam Peyrard Janvid for her valuable comments on the manuscript.
Conflict of Interest statement. None declared.
- signal transduction
- transcription, genetic
- hippocampus (brain)
- mental processes
- receptors, estrogen
- reading disorder, developmental
- brain development
- neuron migration
- protein microarrays
- carboxyl group