Structural insights into the Switching Off of the Interaction between the Archaeal Ribosomal Stalk and aEF1A by Nucleotide Exchange Factor aEF1B

The ribosomal stalk protein plays a crucial role in functional interactions with translational GTPase factors. It has been shown that the archaeal stalk aP1 binds to both GDP- and GTP-bound conformations of aEF1A through its C-terminal region in two different modes. To obtain an insight into how the aP1•aEF1A binding mode changes during the process of nucleotide exchange from GDP to GTP on aEF1A, we have analyzed structural changes in aEF1A upon binding of the nucleotide exchange factor aEF1B. The isolated archaeal aEF1B has nucleotide exchange ability in the presence of aa-tRNA but not deacylated tRNA, and increases activity of polyphenylalanine synthesis 4-fold. The aEF1B mutation, R90A, results in loss of its original nucleotide exchange activity but retains a remarkable ability to enhance polyphenylalanine synthesis. These results suggest an additional functional role for aEF1B other than in nucleotide exchange. The crystal structure of the aEF1A•aEF1B complex, resolved at 2.0 Å resolution, shows marked rotational movement of domain 1 of aEF1A compared to the structure of aEF1A•GDP•aP1, and this conformational change results in disruption of the original aP1 binding site between domains 1 and 3 of aEF1A. The loss of aP1 binding to the aEF1A•aEF1B complex was confirmed by native gel analysis. The results suggest that aEF1B plays a role in switching off the interaction between aP1 and aEF1A•GDP, as well as in nucleotide exchange, and promote translation elongation.     

Conformational Dynamics Allows Sampling of an “Active-like” State by Oncogenic K-Ras-GDP

Mutations in K-Ras GTPase replacing Gly12 with either Asp or Val are common in cancer. These mutations decelerate intrinsic and catalyzed GTP hydrolysis, leading to accumulation of K-Ras-GTP in cells. Signaling cascades initiated by K-Ras-GTP promote cell proliferation, survival, and invasion. Despite functional differences between the most frequent G12D mutation and the most aggressive and chemotherapy resistant G12V mutation, their long-suspected distinct structural features remain elusive. Using NMR, X-ray structures, and computational methods, we found that oncogenic mutants of K-Ras4B, the predominant splice variant of K-Ras, exhibit distinct conformational dynamics when GDP-bound, visiting the “active-like” conformational state similar to the one observed in GTP-bound K-Ras. This behavior distinguishes G12V from wild type and G12D K-Ras4B-GDP. The likely reason is interactions between the aliphatic sidechain of V12 and the Switch II region of K-Ras4B^G12V-GDP, which are distinct in K-Ras4B^G12D-GDP. In the X-ray structures, crystal contacts reduce the dynamics of the sidechain at position 12 by stabilizing the Switch I region of the protein. This explains why structural differences between G12V and G12D K-Ras have yet not been reported. Together, our results suggest a previously unknown mechanism of K-Ras activation. This mechanism relies on conformational dynamics caused by specific oncogenic mutations in the GDP-bound state. Our findings also imply that the therapeutic strategies decreasing the level of K-Ras-GTP by interfering with nucleotide exchange or by expediting GTP hydrolysis may work differently in different oncogenic mutants.     

EGCG has Dual and Opposing Effects on the N-terminal Region of Self-associating α-synuclein Oligomers

Oligomers of the protein α-synuclein (α-syn) are thought to be a major toxic species in Parkinson’s disease, particularly through their ability to permeabilize cell membranes. The green tea polyphenol epigallocatechin gallate (EGCG) has been found to reduce this ability. We have analyzed α-syn oligomer dynamics and interconversion by H/D exchange monitored by mass spectrometry (HDX-MS). Our results show that the two oligomers OI and OII co-exist in equilibrium; OI is a multimer of OII and its dissociation can be followed by HDX-MS by virtue of the correlated exchange of the N-terminal region. Urea destabilizes the α-syn oligomers, dissociating OI to OII and monomers. Oligomers exposed to EGCG undergo Met oxidation. Intriguingly, EGCG induces an oxidation-dependent effect on the structure of the N-terminal region. For the non-oxidized N-terminal region, EGCG increases the stability of the folded structure as measured by a higher level of protection against H/D exchange. In contrast, protection is clearly abrogated in the Met oxidized N-terminal region. Having a non-oxidized and disordered N-terminal region is known to be essential for efficient membrane binding. Therefore, our results suggest that the combined effect of a structural stabilization of the non-oxidized N-terminal region and the presence of a disordered oxidized N-terminal region renders the oligomers less cytotoxic by decreasing the ability of the N-terminal region to bind to cell membranes and facilitate their permeabilization.     

Relating nucleotide‐dependent conformational changes in free tubulin dimers to tubulin assembly

The complex dynamic behavior of microtubules (MTs) is believed to be primarily due to the αβ‐tubulin dimer architecture and its intrinsic GTPase activity. Hence, a detailed knowledge of the conformational variations of isolated α‐GTP‐β‐GTP‐ and α‐GTP‐β‐GDP‐tubulin dimers in solution and their implications to interdimer interactions and stability is directly relevant to understand the MT dynamics. An attempt has been made here by combining molecular dynamics (MD) simulations and protein–protein docking studies that unravels key structural features of tubulin dimer in different nucleotide states and correlates their association to tubulin assembly. Results from simulations suggest that tubulin dimers and oligomers attain curved conformations in both GTP and GDP states. Results also indicate that the tubulin C‐terminal domain and the nucleotide state are closely linked. Protein–protein docking in combination with MD simulations suggest that the GTP‐tubulin dimers engage in relatively stronger interdimer interactions even though the interdimer interfaces are bent in both GTP and GDP tubulin complexes, providing valuable insights on in vitro finding that GTP‐tubulin is a better assembly candidate than GDP‐tubulin during the MT nucleation and elongation processes. © 2012 Wiley Periodicals, Inc. Biopolymers 99: 282–291, 2013.     

Competitive Microtubule Binding of PEX14 Coordinates Peroxisomal Protein Import and Motility

Human PEX14 plays a dual role as docking protein in peroxisomal protein import and as peroxisomal anchor for microtubules (MT), which relates to peroxisome motility. For docking, the conserved N-terminal domain of PEX14 (PEX14-NTD) binds amphipathic alpha-helical ligands, typically comprising one or two aromatic residues, of which human PEX5 possesses eight. Here, we show that the PEX14-NTD also binds to microtubular filaments in vitro with a dissociation constant in nanomolar range. PEX14 interacts with two motifs in the C-terminal region of human ß-tubulin. At least one of the binding motifs is in spatial proximity to the binding site of microtubules (MT) for kinesin. Both PEX14 and kinesin can bind to MT simultaneously. Notably, binding of PEX14 to tubulin can be prevented by its association with PEX5. The data suggest that PEX5 competes peroxisome anchoring to MT by occupying the ß-tubulin-binding site of PEX14. The competitive correlation of matrix protein import and motility may facilitate the homogeneous dispersion of peroxisomes in mammalian cells.     

Evolutionarily Conserved Proline Residues Impede the Misfolding of the Mouse Prion Protein by Destabilizing an Aggregation-competent Partially Unfolded Form

The misfolding of the prion protein has been linked to several neurodegenerative diseases. Despite extensive studies, the mechanism of the misfolding process remains poorly understood. The present study structurally delineates the role of the conserved proline residues present in the structured C-terminal domain of the mouse prion protein (moPrP) in the misfolding process. It is shown that mutation of these Pro residues to Ala leads to destabilization of the native (N) state, and also to rapid misfolding. Using hydrogen–deuterium exchange (HDX) studies coupled with mass spectrometry (MS), it has been shown that the N state of moPrP is in rapid equilibrium with a partially unfolded form (PUF2*) at pH 4. It has been shown that the Pro to Ala mutations make PUF2* energetically more accessible from the N state by stabilizing it relative to the unfolded (U) state. The apparent rate constant of misfolding is found to be linearly proportional to the extent to which PUF2* is populated in equilibrium with the N state, strongly indicating that misfolding commences from PUF2*. It has also been shown that the Pro residues restrict the boundary of the structural core of the misfolded oligomers. Overall, this study highlights how the conserved proline residues control misfolding of the prion protein by modulating the stability of the partially unfolded form from which misfolding commences.     

Cold depolymerization of microtubules to double rings: geometric stabilization of assemblies

The kinetic pathway of microtubule depolymerization at 0 degrees C has been examined. Microtubules made of MAP-containing and MAP-free tubulins were depolymerized at 0 degree C in the presence of [3H]GDP or [3H]GTP or of trace amounts of 125I dimeric tubulin. The products of depolymerization were separated on a column, their structures were identified by electron microscopy, and the time course of incorporation of 3H or 125I labels in the different components of the system was determined. Two predominant assembly states of tubulin found in the nonmicrotubule state were alpha-beta dimers and double rings. Kinetic data indicate that ring formation from disassembling microtubules does not occur by direct coiling of protofilaments as previously thought, but disassembling GDP subunits are in very rapid equilibrium with curved oligomers that are kinetic intermediates in the isodesmic assembly of GDP-tubulin. The formation of oligomers and rings from dimers, at concentrations as low as 10 microM, is much faster than nucleotide exchange on alpha-beta-tubulin. Disassembly of double rings, in contrast, is slower than nucleotide exchange on alpha-beta-tubulin, by 1 order of magnitude in the absence of MAPs and 2 orders of magnitude in the presence of MAPs. These results support the model proposed previously to explain spontaneous oscillations in microtubule assembly. They are consistent with the existence of an equilibrium between two conformations of tubulin, "straight", i.e., microtubule forming, and "curved", i.e., ring forming, under the allosteric control of bound nucleotide. The straight conformation requires the presence of two ionizable hydroxyls on the gamma-phosphate in GTP or GDP-Pi.     

Dynamic Solution Structures of Whole Human NAP1 Dimer Bound to One and Two Histone H2A-H2B Heterodimers Obtained by Integrative Methods

Nucleosome assembly protein 1 (NAP1) binds to histone H2A-H2B heterodimers, mediating their deposition on and eviction from the nucleosome. Human NAP1 (hNAP1) consists of a dimerization core domain and intrinsically disordered C-terminal acidic domain (CTAD), both of which are essential for H2A-H2B binding. Several structures of NAP1 proteins bound to H2A-H2B exhibit binding polymorphisms of the core domain, but the distinct structural roles of the core and CTAD domains remain elusive. Here, we have examined dynamic structures of the full-length hNAP1 dimer bound to one and two H2A-H2B heterodimers by integrative methods. Nuclear magnetic resonance (NMR) spectroscopy of full-length hNAP1 showed CTAD binding to H2A-H2B. Atomic force microscopy revealed that hNAP1 forms oligomers of tandem repeated dimers; therefore, we generated a stable dimeric hNAP1 mutant exhibiting the same H2A-H2B binding affinity as wild-type hNAP1. Size exclusion chromatography (SEC), multi-angle light scattering (MALS) and small angle X-ray scattering (SAXS), followed by modelling and molecular dynamics simulations, have been used to reveal the stepwise dynamic complex structures of hNAP1 binding to one and two H2A-H2B heterodimers. The first H2A-H2B dimer binds mainly to the core domain of hNAP1, while the second H2A-H2B binds dynamically to both CTADs. Based on our findings, we present a model of the eviction of H2A-H2B from nucleosomes by NAP1.     

Inhibition of microtubule assembly competent tubulin synthesis leads to accumulation of phosphorylated tau in neuronal cell bodies

Neurofibrillary tangles, a pathological hallmark of Alzheimer’s disease (AD), are somatodendritic filamentous inclusions composed of hyperphosphorylated tau. Microtubule loss is also a common feature of affected neurons in AD. However, whether and how the disruptions of microtubules and the microtubule-associated proteins occur in the pathogenesis of AD remain unclear. Recent evidence indicates that reduced expression of tubulin by knocking down a tubulin chaperon can cause tau neurotoxicity. Thus, the disruption of tubulin homeostasis may result in the acquisition of tau pathogenesis and ultimately cause tauopathy. To investigate whether the disruption of tubulin maintenance induces tau abnormalities in mammalian neurons, we developed a miRNA-mediated knockdown system of tubulin-specific chaperon E (Tbce), which is a factor required for the de novo synthesis of tubulin. Tbce knockdown in mouse primary cultured neurons induced an increase in tubulin in the cell body at 14 days in vitro. Accumulated tubulin was not acetylated or incorporated in microtubules, indicating that they were functionally inert. Concomitantly, tau also accumulated in neuronal cell bodies. The mis-localized tau was phosphorylated at Ser202/Thr205 and Ser396/Ser404. These results indicate that Tbce knockdown in mammalian neurons induces not only a reduction in properly folded tubulins, which are microtubule assembly competent, but also an accumulation of phosphorylated tau in the cell body of mammalian neurons. These findings suggest that disruption of the homeostatic mechanism for maintaining tubulin biosynthesis and/or microtubules can cause tau accumulation in the cell body, which is commonly observed in tauopathies.     

Artifactual FA dimers mimic FAHFA signals in untargeted metabolomics pipelines

FA esters of hydroxy FAs (FAHFAs) are lipokines with extensive structural and regional isomeric diversity that impact multiple physiological functions, including insulin sensitivity and glucose homeostasis. Because of their low molar abundance, FAHFAs are typically quantified using highly sensitive LC-MS/MS methods. Numerous relevant MS databases house in silico-spectra that allow identification and speciation of FAHFAs. These provisional chemical feature assignments provide a useful starting point but could lead to misidentification. To address this possibility, we analyzed human serum with a commonly applied high-resolution LC-MS untargeted metabolomics platform. We found that many chemical features are putatively assigned to the FAHFA lipid class based on exact mass and fragmentation patterns matching spectral databases. Careful validation using authentic standards revealed that many investigated signals provisionally assigned as FAHFAs are in fact FA dimers formed in the LC-MS pipeline. These isobaric FA dimers differ structurally only by the presence of an olefinic bond. Furthermore, stable isotope-labeled oleic acid spiked into human serum at subphysiological concentrations showed concentration-dependent formation of a diverse repertoire of FA dimers that analytically mimicked FAHFAs. Conversely, validated FAHFA species did not form spontaneously in the LC-MS pipeline. Together, these findings underscore that FAHFAs are endogenous lipid species.  However, nonbiological FA dimers forming in the setting of high concentrations of FFAs can be misidentified as FAHFAs. Based on these results, we assembled a FA dimer database to identify nonbiological FA dimers in untargeted metabolomics datasets.     

Direct incorporation of microtubule oligomers at high GTP concentrations

Chick brain microtubule protein consists primarily of a mixture of MAP2:tubulin oligomers and dimeric tubulin. The assembly of this protein is described by a single pseudofirst-order reaction at 20 microM GTP, but by the summation of two pseudofirst-order reactions at 1 mM GTP. The protein contains two GTP-binding species, corresponding to the tubulin dimers and the oligomers, and conditions which alter the dimer: oligomer equilibrium, affect the kinetics of microtubule assembly. The results indicate that the oligomers are only direct assembly intermediates at high GTP concentrations.     

Early vs. late rate of torque development: Relation with maximal strength and influencing factors

We re-examined the relationship between rate of torque development (RTD) and maximal voluntary contractions (MVC) torque, and investigated some possible neuromuscular determinants of early (≤100 ms) and late (≥200 ms) RTD. Seventeen healthy men performed maximal explosive isometric knee extensions at five joint angles, from which MVC torque, RTD at different time intervals (50–250 ms), and early quadriceps EMG activity (EMG₅₀) were evaluated. Quadriceps muscle thickness (MT) was quantified by longitudinal ultrasonography. The relationship between MVC torque, EMG₅₀ and MT against RTD was assessed with Pearson’s and repeated measures correlation coefficients. Moderate-to-strong correlation coefficients were observed between MVC torque and RTD (r = 0.50–0.88, p < 0.001), with stronger relationships for late RTD than for early RTD. Weak-to-strong correlation coefficients were observed amongst RTD and EMG₅₀ (r = 0.37–0.83, p < 0.001), with stronger relationships for early RTD than for late RTD. Only late RTD was significantly correlated with MT, though only moderately (r = 0.50–0.52, p < 0.05). These findings suggest that early and late knee extension RTD are potentially governed by different neuromuscular factors. Neuromuscular activation seems to have a greater influence on early RTD than on late RTD, and vice versa for muscle mass.     

Dynamic and Reversible Aggregation of the Human CAP Superfamily Member GAPR-1 in Protein Inclusions in Saccharomyces cerevisiae

Many proteins that can assemble into higher order structures termed amyloids can also concentrate into cytoplasmic inclusions via liquid–liquid phase separation. Here, we study the assembly of human Golgi-Associated plant Pathogenesis Related protein 1 (GAPR-1), an amyloidogenic protein of the Cysteine-rich secretory proteins, Antigen 5, and Pathogenesis-related 1 proteins (CAP) protein superfamily, into cytosolic inclusions in Saccharomyces cerevisiae. Overexpression of GAPR-1-GFP results in the formation GAPR-1 oligomers and fluorescent inclusions in yeast cytosol. These cytosolic inclusions are dynamic and reversible organelles that gradually increase during time of overexpression and decrease after promoter shut-off. Inclusion formation is, however, a regulated process that is influenced by factors other than protein expression levels. We identified N-myristoylation of GAPR-1 as an important determinant at early stages of inclusion formation. In addition, mutations in the conserved metal-binding site (His54 and His103) enhanced inclusion formation, suggesting that these residues prevent uncontrolled protein sequestration. In agreement with this, we find that addition of Zn²⁺ metal ions enhances inclusion formation. Furthermore, Zn²⁺ reduces GAPR-1 protein degradation, which indicates stabilization of GAPR-1 in inclusions. We propose that the properties underlying both the amyloidogenic properties and the reversible sequestration of GAPR-1 into inclusions play a role in the biological function of GAPR-1 and other CAP family members.     

Techniques for studying membrane pores

Pore-forming proteins (PFPs) are of special interest because of the association of their activity with the disruption of the membrane impermeability barrier and cell death. They generally convert from a monomeric, soluble form into transmembrane oligomers that induce the opening of membrane pores. The study of pore formation in membranes with molecular detail remains a challenging endeavor because of its highly dynamic and complex nature, usually involving diverse oligomeric structures with different functionalities. Here we discuss current methods applied for the structural and functional characterization of PFPs at the individual vesicle and cell level. We highlight how the development of high-resolution and single-molecule imaging techniques allows the analysis of the structural organization of protein oligomers and pore entities in lipid membranes.     

Description of Aestuariivivens marinum sp. nov., Aestuariivivens sediminis sp. nov. and Aestuariivivens sediminicola sp. nov., three new species isolated from the upper layer of tidal flat sediment

Nine bacterial strains designated MT3-5-12^T, MT3-5-27, MTV1-9, S-DT1-15^T, S-DT1-34, MTV5-3^T, MTV4-17, MTV5-12 and MTV5-13 were isolated from the upper layer (1–5 cm in depth) of tidal flat sediment in Quanzhou Bay, China. The 16S rRNA gene of these strains shared maximum sequence similarities with Aestuariivivens insulae KCTC 42350^T of 94.9–97.1%. Phylogenetic analyses based on 16S rRNA gene sequences and 120 conserved concatenated proteins placed these strains in three novel phylogenetic clades affiliated to the genus Aestuariivivens of the family Flavobacteriaceae. Strains MT3-5-12^T, MT3-5-27 and MTV1-9 were phylogenetically close to A. insulae KCTC 42350^T. The average nucleotide identity (ANI) and digital DNA-DNA hybridization (dDDH) values between strains MT3-5-12^Tand MTV1-9 and A. insulae KCTC 42350^T were estimated to be 78.5-78.7% and 22.5%, respectively. Strains S-DT1-15^T and S-DT1-34 formed a distinctly separated clade from A. insulae KCTC 42350^T. The ANI and dDDH values between strains S-DT1-15^T and S-DT1-34 and A. insulae KCTC 42350^T were 76.3–76.4% and 20.4–20.5%, respectively. The other four strains MTV5-3^T, MTV4-17, MTV5-12 and MTV5-13, formed a third novel clade, distinctly separated from A. insulae KCTC 42350^T. The ANI and dDDH values between strains MTV5-3^T and MTV4-17 and A. insulae KCTC 42350^T were 74.7% and 19.1–19.2%, respectively. The phylogenetic analyses and whole genomic comparisons, combined with phenotypic and chemotaxonomic features, strongly supported the nine strains could be classified as three novel species within the genus Aestuariivivens, for which the names Aestuariivivens marinum sp. nov. MT3-5-12^T, Aestuariivivens sediminis sp. nov. S-DT1-15^T, and Aestuariivivens sediminicola sp. nov. MTV5-3^T are proposed.     

Tau Fibrillation Induced by Heparin or a Lysophospholipid Show Different Initial Oligomer Formation

The protein tau is involved in several neurogenerative diseases such as Alzheimer’s Disease, where tau content and fibrillation have been linked to disease progression. Tau colocalizes with phospholipids and glycosaminoglycans in vivo. We investigated if and how tau fibrillation can be induced by two lysophospholipids, namely the zwitterionic 1-myristoyl-2-hydroxy-sn-glycero-3-phosphocholine (LPC) and the anionic 1-myristoyl-2-hydroxy-sn-glycero-3-phospho-(1′-rac-glycerol) (LPG) as well as the glycosaminoglycan heparin. We used a range of biophysical techniques including small-angle X-ray scattering, Thioflavin T fluorescence, and SDS-PAGE, collecting data at various time points to obtain structural information on each phase of the fibrillation. We find that LPC does not induce fibrillation or interact with tau. Low concentrations of LPG induce fibrillation by formation of small hydrophobic clusters with monomeric tau. At higher LPG concentrations, a core–shell complex is formed where tau wraps around LPG micelles with regions extending away from the micelles. For heparin, loosely associated oligomers are formed rapidly with around ten tau molecules. Fibrils formed with either LPG or heparin show similar final cross-section dimensions. Furthermore, SDS-resistant oligomers are observed for both LPG and heparin. Our study demonstrates that tau fibrillation can be induced by two different biologically relevant cofactors leading to structurally different initial states but similar cross-sectional dimensions for the fibrils. Structural information about initial states prior to fibril formation is important both to gain a better understanding of the onset of fibrillation in vivo, and for the development of targeted drugs that can reduce or abolish tau fibrillation.     

Structural Insights into the Mechanism of Base Excision by MBD4

DNA glycosylases remove damaged or modified nucleobases by cleaving the N-glycosyl bond and the correct nucleotide is restored through subsequent base excision repair. In addition to excising threatening lesions, DNA glycosylases contribute to epigenetic regulation by mediating DNA demethylation and perform other important functions. However, the catalytic mechanism remains poorly defined for many glycosylases, including MBD4 (methyl-CpG binding domain IV), a member of the helix-hairpin-helix (HhH) superfamily. MBD4 excises thymine from G·T mispairs, suppressing mutations caused by deamination of 5-methylcytosine, and it removes uracil and modified uracils (e.g., 5-hydroxymethyluracil) mispaired with guanine. To investigate the mechanism of MBD4 we solved high-resolution structures of enzyme-DNA complexes at three stages of catalysis. Using a non-cleavable substrate analog, 2′-deoxy-pseudouridine, we determined the first structure of an enzyme-substrate complex for wild-type MBD4, which confirms interactions that mediate lesion recognition and suggests that a catalytic Asp, highly conserved in HhH enzymes, binds the putative nucleophilic water molecule and stabilizes the transition state. Observation that mutating the Asp (to Gly) reduces activity by 2700-fold indicates an important role in catalysis, but probably not one as the nucleophile in a double-displacement reaction, as previously suggested. Consistent with direct-displacement hydrolysis, a structure of the enzyme-product complex indicates a reaction leading to inversion of configuration. A structure with DNA containing 1-azadeoxyribose models a potential oxacarbenium-ion intermediate and suggests the Asp could facilitate migration of the electrophile towards the nucleophilic water. Finally, the structures provide detailed snapshots of the HhH motif, informing how these ubiquitous metal-binding elements mediate DNA binding.     

Optimisation of estrogen receptor subtype-selectivity of a 4-Aryl-4H-chromene scaffold previously identified by virtual screening

4-Aryl-4H-Chromene derivatives have been previously shown to exhibit anti-proliferative, apoptotic and anti-angiogenic activity in a variety of tumor models in vitro and in vivo generally via activation of caspases through inhibition of tubulin polymerisation. We have previously identified by Virtual Screening (VS) a 4-aryl-4H-chromene scaffold, of which two examples were shown to bind Estrogen Receptor α and β with low nanomolar affinity and <20-fold selectivity for α over β and low micromolar anti-proliferative activity in the MCF-7 cell line. Thus, using the 4-aryl-4H-chromene scaffold as a starting point, a series of compounds with a range of basic arylethers at C-4 and modifications at the C3-ester substituent of the benzopyran ring were synthesised, producing some potent ER antagonists in the MCF-7 cell line which were highly selective for ERα (compound 35; 350-fold selectivity) or ERβ (compound 42; 170-fold selectivity).