The sequence–structure–function relationship of intrinsic ERα disorder

Recombinant protein expression and purification

The NTD of human ER (ER-NTD, residues M1–Y184) was subcloned into the bacterial expression vector pMCSG7 and expressed as inclusion bodies in Escherichia coli BL21(DE3) cells using lysogeny broth medium. Protein expression was induced at optical density OD₆₀₀ = 0.8 by the addition of 0.1 mM isopropyl-ß-d-thiogalactopyranoside (IPTG), followed by overnight incubation at 25 °C. Mutants were generated using the Phusion site-directed mutagenesis kit (Thermo Fisher, catalogue no. F541) and verified by DNA sequencing.

For isotopic ¹⁵N and ¹³C labelling, cells were cultured in M9 minimal medium supplemented with 1 g l^{−1 15}NH₄Cl and 3 g l^{−1 13}C-glucose (Cambridge Isotope Laboratories, catalogue nos. CLM-1396-50 and NLM-467-50, respectively). Harvested cells were lysed by high-pressure homogenization (Avestin EmulsiFlex-C3) in 50 mM Tris pH 8.0 and 500 mM NaCl buffer with protease inhibitors (Millipore Sigma, catalogue no. 11836153001). Inclusion bodies were solubilized in 50 mM sodium phosphate pH 7.4, 5 mM imidazole and 1 mM phenylmethanesulfonyl fluoride (PMSF) with 6 M guanidine hydrochloride, and clarified by centrifugation at 4 °C.

Untagged and His-tagged ER-NTD proteins were purified from solubilized inclusion bodies. Following the purification protocol⁵⁷, untagged proteins were dialysed into 20 mM Tris pH 7.5, 10 mM NaCl and 0.1 mM PMSF, and purified by anion exchange using a HiTrap Q column (GE Healthcare, catalogue no. 29-0513-25) and gel filtration with an Enrich SEC650 column (Bio-Rad, catalogue no. 7801650). His-tagged proteins were purified on cobalt resin (Thermo Fisher, catalogue no. 89964), eluted with imidazole and dialysed into 20 mM Tris pH 7.4, 10 mM NaCl and 0.1 mM PMSF. The His-tag was removed by overnight TEV protease digestion at 4 °C, and further purified through the same anion-exchange and size-exclusion steps. Purified ERα-NTD proteins were stored at either 4 or −80 °C.

In vitro protein phosphorylation

Purified ER-NTD proteins were phosphorylated in vitro by incubation with recombinant, activated MAP kinase 1 (MAPK1/ERK2) possessing an N-terminal glutathione S-transferase (GST) tag (Millipore Sigma, catalogue no. 14-550). Phosphorylation followed a previously published protocol¹⁹, using 3 mg of ER-NTD proteins and 5 µg of activated MAPK1 in a kinase reaction buffer (25 mM Tris pH 7.5, 25 mM MgCl₂, 10 mM dithiothreitol and 100 µM ATP). Reactions proceeded for 4 h at 30 °C. Site-specific phosphorylation at S118 was verified by immunoblot using a phosphorylation-ER Ser118 antibody (Cell Signaling, catalogue no. 2511). Following the reaction, GST-MAPK1 enzyme was removed by incubation with glutathione-agarose resin (Macherey-Nagel, catalogue no. NC0389541).

SEC-SAXS

SEC-SAXS measurements were performed on protein samples at concentrations ranging from 2 to 4 mg ml⁻¹. The running buffer, consisting of 20 mM sodium phosphate, 150 mM NaCl, 0.5 mM EDTA and 0.1 mM PMSF pH 7.4, was precooled to 4 °C and maintained on ice during data acquisition. Samples were injected into either an Enrich SEC650 column (Bio-Rad, catalogue no. 7801650) or a Superdex 75 Increase 5/150 column (Cytiva, catalogue no. 29148722), at flow rates ranging from 0.35 to 0.70 ml min⁻¹. Scattering images were collected at either 0.5 or 2-s exposure per frame during column elution. Background subtraction was applied to the elution peak to obtain the final protein-scattering profile.

For WT, S118D and S118A samples, data were collected at the BioCAT-18ID beamline of the Advanced Photon Source (Argonne, IL, USA) using a Pilatus3-X-1M detector with X-rays at an energy of 13 keV. For pS118 and E56Q-pS118 samples, data collection was performed at the 16-ID (LiX) beamline of the National Synchrotron Light Source II (Upton, NY, USA) using a combination of Pilatus3 S 1 M and Pilatus3 900-K detectors with X-rays at an energy of 15 keV.

NMR spectroscopy

Isotope-labelled proteins were prepared in 20 mM sodium phosphate buffer pH 7.4, 100 mM NaCl, 0.5 mM EDTA, 0.1 mM PMSF and 5% (v/v) D₂O. NMR data were acquired using either a Bruker Avance III HD 700 or 850-MHz spectrometer equipped with a TXO- or TCI triple-resonance cryogenic probe, respectively, at a temperature of 4 °C, to reduce the line broadening of solvent-exposed residues due to their amide hydrogen exchange with the solvent.

For ¹⁵N,¹³C ER-NTD samples at a concentration of 400 µM, a set of standard triple-resonance backbone assignment experiments⁵⁸ were recorded at 850 MHz, using sensitivity-enhanced gradient coherence selection^59,60, semiconstant time acquisition in the ¹⁵N dimension⁶¹ and non-uniform sampling following either Poisson-gap⁶² or exponentially weighted sampling schedules. Briefly, the typical direct ¹H dimension contained 1,024 complex points for a 100-ms acquisition time, whereas maximum data size of the indirect dimensions (¹⁵N × ¹³C), in terms of both complex points (acquisition times) and sparsity levels, was as follows: ¹⁵N HSQC with 100 ms (380; 47.4%), HNCO with 100 × 50 ms (380 × 150; 0.53%), HNCA with 100 × 14 ms (380 × 90; 1.75%), HNCOCA with 100 × 14 ms (380 × 90; 0.88%), HNCACO with 100 × 50 ms (380 × 150; 1.05%), HNCACB with 100 × 14 ms (380 × 240; 1.31%) and CBCACONH with 100 × 6.7 ms (380 × 115; 1.37%). Additional three-dimensional NNH-NOESY spectra were acquired with 380 complex points for 100-ms acquisition times in indirect ¹⁵N dimensions using a sparsity level of 0.83%. The typical experimental times for spectra were as follows: 22 min for HSQC, 10 h for HNCO, 37 h for HNCA, 28 h for HNCOCA, 38 h for HNCACO, 75 h for HNCACB, 38 h for CBCACONH and 84 h for NNH-NOESY.

In addition to proton-detected spectra, ¹³CO-detected hCACON and hCANCO spectra⁶³ were recorded at 700 MHz for resonance assignments of most proline residues. Data were collected using virtual decoupling, and without non-uniform sampling, at acquisition times (complex points) of 103 ms (512) × 30 ms (60) × 5.6 ms (30) for ¹³CO × ¹⁵N × ¹³Cα dimensions. All NMR spectra were recorded at 4 °C.

Data were acquired and processed using BRUKER Topspin (v.3.6) and NMRPipe (v.1.0)⁶⁴. Spectra were analysed using NMRfam-SPARKY (v.1.470) with Sparky 3.190 (ref. ⁶⁵).

Spectral referencing was performed directly for the ¹H dimension using the temperature-dependent water resonance frequency, whereas ¹³C and ¹⁵N were referenced indirectly. Combined chemical shift changes for ¹⁵N HSQC were calculated as \(\Delta \delta =\sqrt{(\Delta \delta {{\rm{H}}}^{2}+{(\Delta \delta {\rm{N}}\times 0.154)}^{2}}\), where ∆δH and ∆δN represent the changes in proton and nitrogen chemical shift, respectively, compared with the free or WT protein spectrum.

Paramagnetic relaxation enhancement

Single-cysteine mutants were labelled with the paramagnetic nitroxide spin label MTSL (Toronto Research Chemicals, catalogue no. O875000), following the established protocol⁶⁶. Protein samples (100–200 µM) were buffer exchanged into a labelling buffer (50 mM Tris-HCl pH 7.4 and 50 mM NaCl) supplemented with 2.5 mM freshly prepared dithiothreitol (DTT) to prevent cysteine oxidation. Immediately before labelling, DTT was removed by passing the samples through a Zeba spin-desalting column equilibrated with DTT-free labelling buffer. MTSL was dissolved in acetonitrile (0.1% v/v) to a minimum concentration of 2 mM, and over 20-fold molar excess was added to the protein solution. The reaction mixture was incubated overnight at 4 °C on a rotating device, protected from light to ensure efficient labelling. Excess MTSL was removed using a Zeba spin-desalting column, and samples were concentrated to the desired level using an Amicon Ultra-15 centrifugal filter (4 °C). Spin-labelling efficiency (over 95%) was confirmed by electrospray ionization mass spectrometry. Diamagnetic control samples were prepared by reducing the paramagnetic samples with a tenfold molar excess of l-ascorbic acid (Millipore Sigma, catalogue no. A7631), with overnight incubation at 4 °C.

Two-dimensional ¹H-¹⁵N HSQC spectra were acquired for each MTSL-label, single-cysteine mutant in both paramagnetic and diamagnetic states at 4 °C. Comparison of chemical shifts between labelled and unlabelled samples confirmed that MTSL labelling had not significantly perturbed the protein’s structural ensemble. PRE effects were quantified by calculating the peak intensity ratio (I_para/I_dia) between paramagnetic and diamagnetic spectra for each resolved cross-peak. Errors in intensity ratios were propagated using root-mean-square deviation between observed and Gaussian-fit peak heights.

To investigate the role of hydrophobicity in long-range interactions, PRE measurements were conducted for ER-NTD with S46C spin labelling (30 μM) with and without DDM (120 μM; Anatrace, catalogue no. D310). ¹H-¹⁵N HSQC spectra were recorded under identical conditions for DDM-containing and DDM-free samples.

The effect of paramagnetic relaxation on cross-peak intensity^26,67 is expressed as

where t_o = 10.87 ms (combined INEPT time) and R₂(¹H^N) is the intrinsic relaxation rate of the amide proton. Γ₂, paramagnetic rate enhancement, is related to the distance between the amino proton and spin label by

where K = 1.2 × 10⁻³² cm⁶ s⁻² for the MTSL spin label²⁶ and ω_h is the Larmor frequency of proton spin at 850 MHz. The apparent correlation time (τ_c) for electron-amide proton interactions is 5.9 ± 1.4 ns (mean ± s.d.).

¹⁵N relaxation measurement

Protein samples were used at a concentration of 60 µM for R₁ and R_1ρ (spin-lattice relaxation rate in the rotating frame) measurements, at 850 MHz and 4 °C, using the HSQC-based sequences of Lakomek et al.⁶⁸ but with modifications for non-deuterated proteins⁶⁹. Experiments were performed with ¹H and ¹⁵N sweep width (acquisition times) of 12 ppm (100 ms) and 21 ppm (88 ms), respectively, using eight signal-averaging scans and 2.5-s recovery delay. Delays for R₁ (R_1ρ) relaxation were 0, 280, 560 × 2 and 840 × 2 ms (0, 55, 110 × 2 and 165 × 2) with repeats for statistics. Magnetization was aligned with the effective field in R_1ρ experiments using 2-ms adiabatic half passages⁷⁰ at a calibrated B₁ field of 1.08 kHz centred at 121 ppm. R₁ and R_1ρ rates were obtained by fitting mono-exponential decays using in-house Python scripts. Transverse R₂ rates were derived from R₁ and R_1ρ using the equation

where \(\theta =\,\tan ({\omega }_{1}/\varOmega )\) is the effective tilt angle of B1 field, ω₁ is B1 field strength and Ω is ¹⁵N signal offset from spinlock carrier frequency⁷¹. Errors were propagated from signal amplitude variations.

In vitro GST pulldown

The GST-tagged proteins were cloned into pGEX-4T-1 vector and expressed in bacterial Rosetta (DE3) pLysS-competent cells. Expression was induced with 0.2 mM IPTG overnight at 20 °C. Purification was performed using a glutathione spin column (Thermo Fisher, catalogue no. 16105). Untagged NTD proteins (10 µg) were incubated with GST-tagged proteins (100 µg) and immobilized overnight at 4 °C on glutathione-Sepharose beads in 20 mM Tris-HCl pH 8.0, 150 mM NaCl, 10% glycerol, 1% Nonidet P-40 (Sigma-Aldrich, catalogue no. N-3516) and 2 mM EDTA. Subsequently, the resin was washed three times with the same binding buffer to remove unbound proteins. Pulled-down complexes were denatured in tricine sample buffer, separated by 4–20% SDS–polyacrylamide gel electrophoresis and visualized using GelCode blue staining (Thermo Fisher, catalogue no. 24594).

Transient transfection reporter assay

To assay Gal4-NTD reporter activity, HEK293 or MCF7 cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS). Cells were seeded at 2 × 10⁶ per well on 6-cm plates and transfected with 1 µg of Gal4-NTD, 3 µg of 5× Gal4 BS-TATA-luciferase plasmid (Addgene, catalogue no. 46756) and 0.3 µg of pRL-Renilla-luciferase plasmid (Promega, catalogue no. E2231), using Lipofectamine 2000 (Invitrogen, catalogue no. 11668027). Twenty-four hours after transfection, cells were washed with 1× PBS and grown in DMEM with 10% FBS for a further 24 h, and lysates prepared using 1× Passive Lysis Buffer (Promega, catalogue no. E1941).

For full-length ER assays, MCF7 cells in phenol red-free DMEM with 10% charcoal-stripped FBS were transfected with 2.25 µg of HA-tagged, full-length ER expression vector, 6.75 µg of 3× ERE-TATA-luciferase plasmid (Addgene, catalogue no. 11354) and 0.67 µg of pRL-Renilla-luciferase plasmid. Cells were seeded at 7 × 10⁶ per well on 10-cm plates. Twenty-four hours after transfection, cells were washed with 1× PBS, split and cultured in phenol red-free DMEM with 10% charcoal-stripped FBS. Cells were treated with either vehicle (0.1% ethanol), 100 nM E2 or 1 μM 4-hydroxytamoxifen (Cayman, catalogue no. 17308) before lysate preparation.

Firefly and Renilla luciferase activities were measured using the Dual-Luciferase Reporter Assay kit (Promega, catalogue no. E1910), and firefly activity was normalized to Renilla luciferase. Mean ± s.e.m. was calculated from four technical repeats with three biological replicates.

Nucleocytoplasmic fractionation and whole-cell lysate preparation

Cells were lysed with NP-40 buffer (10 mM Tris pH 7.4, 10 mM NaCl, 3 mM MgCl₂, 0.5% NP-40 and 5% glycerol) and centrifuged to obtain cytoplasmic fractions. Nuclear fractions were isolated by lysing the resulting pellets in RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS and 50 mM Tris pH 7.4). Both lysis buffers included protease (Millipore Sigma, catalogue no. 04693116001) and phosphatase inhibitors (Millipore Sigma, catalogue no. 4906845001). MCF7 and HEK293T cells were lysed in RIPA buffer with protease (Millipore Sigma, catalogue no. 04693116001) and phosphatase (Millipore Sigma, catalogue no. 4906845001) inhibitors to obtain whole-cell lysates.

Immunoblotting

Proteins were separated by 10% SDS–polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes (Bio-Rad, catalogue no. 1620177). Membranes were immunoblotted overnight at 4°C with primary antibodies against ER (abcam, catalogue no. ab16660), phospho-ER-Ser118 (Cell Signaling, catalogue no. 2511), β-actin (Cell Signaling, catalogue no. 4967), Lamin B1 (Santa Cruz, catalogue no. sc-374015), β-tubulin (Santa Cruz, catalogue no. sc-166729), Gal4 (Santa Cruz, catalogue no. sc-510), HA-tag (Santa Cruz, catalogue no. sc-7392), His-tag (Millipore Sigma, catalogue no. SAB4301134) and GST-tag (Millipore Sigma, catalogue no. SAB4301139). Secondary anti-IgG antibodies (Bio-Rad, catalogue no. 1706515 or 1706516) were applied at room temperature before imaging on a ChemiDoc MP system (Bio-Rad, catalogue no. 12003154).

RNA extraction and qPCR with reverse transcription

To determine the effects of WT and mutant ER on the expression of ER target genes, cells on 6-cm plates were transfected with 2 μg of either HA-tagged ER expression vector or empty vector control (HA-vector) using Lipofectamine 3000 (Thermo Fisher, catalogue no. L3000001) and 400 μl of Opti-MEM (Gibco, catalogue no. 31985062). Twelve hours after transfection, culture medium was replaced with regular DMEM supplemented with 10% FBS, and cells were grown for an additional 24 h before harvest. Cells were collected by scraping in cold 1× PBS and pelleted. Cell pellets were collected by scraping in cold 1× PBS, centrifuged, resuspended and evenly split for RNA and protein analysis.

Total RNA was isolated using the RNeasy Mini Kit (Qiagen, catalogue no. 74104) and reverse transcribed using iScript supermix (Bio-Rad, catalogue no. 1708841). Quantitative PCR was performed on a Bio-Rad CFX Connect using iQ SYBR Green supermix (Bio-Rad, catalogue no. 1708880).

The following primers were used:

TFF1 forward 5′-ATACCATCGACGTCCCTCCA-3′, reverse 5′-AAGCGTGTCTGAGGTGTCCG-3′;

MYC forward 5′-GGTGCTCCATGAGGAGACA-3′, reverse 5′-CCTGCCTCTTTTCCACAGAA-3′;

CCND1 forward 5′- GGATGCTGGAGGTCTGCGA-3′, reverse 5′-AGAGGCCACGAACATGCAAG-3′ and GAPDH forward 5′-CCACAGTCCATGCCATCA-3′, reverse 5′-GGATGACCTTGCCCACAG-3′.

Calculations were based on mean ± s.e.m. from three technical repeats with three biological replicates.

Cofactor interactions and chemical shift perturbation

The Q-rich domain of transcriptional mediators/intermediary factor 2 (TIF2-QRD, residues G1180–N1288) was subcloned into pMCSG7 bacterial expression vector with an N-terminal 6xHis-tag, and expressed in bacterial BL21(DE3) cells as inclusion bodies using TB medium. TIF2-QRD was purified similarly to ER-NTD as described above, including tag cleavage.

For chemical shift perturbation measurements, ¹⁵N-labelled ER-NTD and unlabelled TIF2-QRD were exchanged into 20 mM sodium phosphate buffer pH 7.4, containing 100 mM NaCl. TIF2-QRD was titrated into 30 μM ER-NTD at ER-NTD/TIF2-QRD molar ratios of 1:0, 1:1, 1:2 and 1:4. Two-dimensional ¹H-¹⁵N HSQC spectra were recorded at each titration point. Combined chemical shift changes (∆δ) were calculated as \(\Delta \delta =\sqrt{(\Delta \delta {{\rm{H}}}^{2}+{(\Delta \delta {\rm{N}}\times 0.154)}^{2}}\), where ΔδH and ΔδN are the changes in proton and nitrogen chemical shift, respectively, relative to the free unbound protein spectrum.

CUT&RUN assay followed by qPCR

CUT&RUN assays were performed according to the kit protocol (Cell Signaling, catalogue no. 86652). Briefly, MCF7 cells were transfected with WT or mutant HA-tagged ER expression plasmids overnight and cultured for 48 h in phenol red-free medium supplemented with 5% charcoal-stripped FBS. Cells were then treated with 10 nM E2 for 150 min. For each immunoprecipitation reaction, 1 × 10⁵ cells were immobilized on Concanavalin A beads and incubated with either 2 μl of primary antibody anti-HA (Cell Signaling, catalogue no. 3724) or anti-TIF2 (Cell Signaling, catalogue no. 96687) for 2 h at 4 °C. Chromatin was fragmented by further incubation with pAG-MNase digestion. Fragmented DNA was purified and analysed by qPCR (normalized to input) using PowerUp SYBR Green Master Mix (Thermo Fisher, catalogue no. A25742) on a QuantStudio-3 real-time PCR system with TFF1 promoter primers (forward 5′-CTAGACGGAATGGGCTTCAT-3′, reverse 5′-TTGGCCGTGACAACAGTG-3′). calculations are based on mean ± s.e.m. from four technical repeats.

Cell growth

MCF7 cells were transfected with vector alone, WT or mutant ER expression plasmids. After transfection, 2 × 10³ MCF7 cells per well were seeded in 96-well plates. Following 12 h allowed for attachment, 10 μl of Cell Counting Kit-8 (CCK-8) reagent (Dojindo, catalogue no. CK04-05) was added and plates incubated for 2 h at 37 °C. Absorbance at 450 nm was measured as baseline (day 0). Subsequently, media were changed to regular DMEM with 10% FBS, 0.1% ethanol (vehicle) or 1 μM 4-hydroxytamoxifen, and refreshed every 2 days. On day 5, CCK-8 reagent was added and absorbance at 450 nm determined. Calculations are based on mean ± s.e.m. from three technical repeats with three biological replicates.

Colony formation

At 24 h after transfection with 2 μg of the respective plasmids using Lipofectamine 3000 (Thermo Fisher, catalogue no. L3000001), MCF7 cells were trypsinized. Viable cells were counted with a haemocytometer, and 500 per well seeded into six-well plates. After 10 days, culture supernatant was removed and cells were washed with PBS, fixed with 4% paraformaldehyde for 15 min then washed again with PBS. Colonies were stained with 0.2% crystal violet solution (Millipore Sigma, catalogue no.C0775) for 15 min, followed by additional PBS washes to remove background staining. Stained colonies were quantified by digital image analysis using NIH ImageJ software, and defined as entities with more than 50 cells. Calculations are based on mean ± s.e.m. from two technical repeats with two biological replicates.

Cell line information

Cell lines MCF7 (ATCC HTB-22), T47D (ATCC HTB-133) and HEK293T (ATCC CRL-3216) were obtained directly from the American Type Culture Collection (ATCC). MCF7 and T47D are human breast adenocarcinoma lines, and HEK293T is a human embryonic kidney epithelial line. Cell lines were authenticated by ATCC and validated in-house by comparison of morphology and growth characteristics against supplier specifications. All lines were regularly tested for mycoplasma contamination using a mycoplasma PCR detection kit (MP Biomedicals, catalogue no. 3050301).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.