Interspecies Variation between Fish and Human Transthyretins in Their Binding of Thyroid-Disrupting Chemicals.

Thyroid-disrupting chemicals (TDCs) are xenobiotics that can interfere with the endocrine system and cause adverse effects in organisms and their offspring. TDCs affect both the thyroid gland and regulatory enzymes associated with thyroid hormone homeostasis. Transthyretin (TTR) is found in the serum and cerebrospinal fluid of vertebrates, where it transports thyroid hormones. Here, we explored the interspecies variation in TDC binding to human and fish TTR (exemplified by Gilthead seabream ( Sparus aurata)). The in vitro binding experiments showed that TDCs bind with equal or weaker affinity to seabream TTR than to the human TTR, in particular, the polar TDCs (>500-fold lower affinity). Crystal structures of the seabream TTR-TDC complexes revealed that all TDCs bound at the thyroid binding sites. However, amino acid substitution of Ser117 in human TTR to Thr117 in seabream prevented polar TDCs from binding deep in the hormone binding cavity, which explains their low affinity to seabream TTR. Molecular dynamics and in silico alanine scanning simulation also suggested that the protein backbone of seabream TTR is more rigid than the human one and that Thr117 provides fewer electrostatic contributions than Ser117 to ligand binding. This provides an explanation for the weaker affinities of the ligands that rely on electrostatic interactions with Thr117. The lower affinities of TDCs to fish TTR, in particular the polar ones, could potentially lead to milder thyroid-related effects in fish.

Structure-Based Virtual Screening Protocol for in Silico Identification of Potential Thyroid Disrupting Chemicals Targeting Transthyretin.

Thyroid disruption by xenobiotics is associated with a broad spectrum of severe adverse outcomes. One possible molecular target of thyroid hormone disrupting chemicals (THDCs) is transthyretin (TTR), a thyroid hormone transporter in vertebrates. To better understand the interactions between TTR and THDCs, we determined the crystallographic structures of human TTR in complex with perfluorooctanesulfonic acid (PFOS), perfluorooctanoic acid (PFOA), and 2,2',4,4'-tetrahydroxybenzophenone (BP2). The molecular interactions between the ligands and TTR were further characterized using molecular dynamics simulations. A structure-based virtual screening (VS) protocol was developed with the intention of providing an efficient tool for the discovery of novel TTR-binders from the Tox21 inventory. Among the 192 predicted binders, 12 representatives were selected, and their TTR binding affinities were studied with isothermal titration calorimetry, of which seven compounds had binding affinities between 0.26 and 100 µM. To elucidate structural details in their binding to TTR, crystal structures were determined of TTR in complex with four of the identified compounds including 2,6-dinitro-p-cresol, bisphenol S, clonixin, and triclopyr. The compounds were found to bind in the TTR hormone binding sites as predicted. Our results show that the developed VS protocol is able to successfully identify potential THDCs, and we suggest that it can be used to propose THDCs for future toxicological evaluations.

Lipid interacting regions in phosphate stress glycosyltransferase atDGD2 from Arabidopsis thaliana.

Membrane lipid glycosyltransferases (GTs) in plants are enzymes that regulate the levels of the non-bilayer prone monogalactosyldiacylglycerol (GalDAG) and the bilayer-forming digalactosyldiacylglycerol (GalGalDAG). The relative amounts of these lipids affect membrane properties such as curvature and lateral stress. During phosphate shortage, phosphate is rescued by replacing phospholipids with GalGalDAG. The glycolsyltransferase enzyme in Arabidopsis thaliana responsible for this, atDGD2, senses the bilayer properties and interacts with the membrane in a monotopic manner. To understand the parameters that govern this interaction, we have identified several possible lipid-interacting sites in the protein and studied these by biophysical techniques. We have developed a multivariate discrimination algorithm that correctly predicts the regions in the protein that interact with lipids, and the interactions were confirmed by a variety of biophysical techniques. We show by bioinformatic methods and circular dichroism (CD), fluorescence, and NMR spectroscopic techniques that two regions are prone to interact with lipids in a surface-charge dependent way. Both of these regions contain Trp residues, but here charge appears to be the dominating feature governing the interaction. The sequence corresponding to residues 227-245 in the protein is seen to be able to adapt its structure according to the surface-charge density of a bilayer. All results indicate that this region interacts specifically with lipid molecules and that a second region in the protein, corresponding to residues 130-148, also interacts with the bilayer. On the basis of this, and sequence charge features in the immediate environment of S227-245, a response model for the interaction of atDGD2 with the membrane bilayer interface is proposed.

Non-transactivational, dual pathways for LPA-induced Erk1/2 activation in primary cultures of brown pre-adipocytes.

In many cell types, G-protein-coupled receptor (GPCR)-induced Erk1/2 MAP kinase activation is mediated via receptor tyrosine kinase (RTK) transactivation, in particular via the epidermal growth factor (EGF) receptor. Lysophosphatidic acid (LPA), acting via GPCRs, is a mitogen and MAP kinase activator in many systems, and LPA can regulate adipocyte proliferation. The mechanism by which LPA activates the Erk1/2 MAP kinase is generally accepted to be via EGF receptor transactivation. In primary cultures of brown pre-adipocytes, EGF can induce Erk1/2 activation, which is obligatory and determinant for EGF-induced proliferation of these cells. Therefore, we have here examined whether LPA, via EGF transactivation, can activate Erk1/2 in brown pre-adipocytes. We found that LPA could induce Erk1/2 activation. However, the LPA-induced Erk1/2 activation was independent of transactivation of EGF receptors (or PDGF receptors) in these cells (whereas in transformed HIB-1B brown adipocytes, the LPA-induced Erk1/2 activation indeed proceeded via EGF receptor transactivation). In the brown pre-adipocytes, LPA instead induced Erk1/2 activation via two distinct non-transactivational pathways, one G(i)-protein dependent, involving PKC and Src activation, the other, a PTX-insensitive pathway, involving PI3K (but not Akt) activation. Earlier studies showing LPA-induced Erk1/2 activation being fully dependent on RTK transactivation have all been performed in cell lines and transfected cells. The present study implies that in non-transformed systems, RTK transactivation may not be involved in the mediation of GPCR-induced Erk1/2 MAP kinase activation.

Structural basis for transthyretin amyloid formation in vitreous body of the eye.

Amyloid transthyretin (ATTR) amyloidosis is characterized by the abnormal accumulation of ATTR fibrils in multiple organs. However, the structure of ATTR fibrils from the eye is poorly understood. Here, we used cryo-EM to structurally characterize vitreous body ATTR fibrils. These structures were distinct from previously characterized heart fibrils, even though both have the same mutation and type A pathology. Differences were observed at several structural levels: in both the number and arrangement of protofilaments, and the conformation of the protein fibril in each layer of protofilaments. Thus, our results show that ATTR protein structure and its assembly into protofilaments in the type A fibrils can vary between patients carrying the same mutation. By analyzing and matching the interfaces between the amino acids in the ATTR fibril with those in the natively folded TTR, we are able to propose a mechanism for the structural conversion of TTR into a fibrillar form.

Scanning electron microscopy as a tool for evaluating morphology of amyloid structures formed on surface plasmon resonance chips.

We demonstrate the use of Scanning Electron microscopy (SEM) in combination with Surface Plasmon Resonance (SPR) to probe and verify the formation of amyloid and its morphology on an SPR chip. SPR is a technique that measures changes in the immobilized weight on the chip surface and is frequently used to probe the formation and biophysical properties of amyloid structures. In this context it is of interest to also monitor the morphology of the formed structures. The SPR chip surface is made of a layer of gold, which represent a suitable material for direct analysis of the surface using SEM. The standard SPR chip used here (CM5-chip, GE Healthcare, Uppsala, Sweden) can easily be disassembled and directly analyzed by SEM. In order to verify the formation of amyloid fibrils in our experimental conditions we analyzed also in-solution produced structures by using Transmission Electron Microscopy (TEM). For further details and experimental findings, please refer to the article published in Journal of Molecular Biology, (Brännström K. et al., 2018) [1].

Transthyretin Interferes with Aß Amyloid Formation by Redirecting Oligomeric Nuclei into Non-Amyloid Aggregates.

The pathological Aß aggregates associated with Alzheimer's disease follow a nucleation-dependent path of formation. A nucleus represents an oligomeric assembly of Aß peptides that acts as a template for subsequent incorporation of monomers to form a fibrillar structure. Nuclei can form de novo or via surface-catalyzed secondary nucleation, and the combined rates of elongation and nucleation control the overall rate of fibril formation. Transthyretin (TTR) obstructs Aß fibril formation in favor of alternative non-fibrillar assemblies, but the mechanism behind this activity is not fully understood. This study shows that TTR does not significantly disturb fibril elongation; rather, it effectively interferes with the formation of oligomeric nuclei. We demonstrate that this interference can be modulated by altering the relative contribution of elongation and nucleation, and we show how TTR's effects can range from being essentially ineffective to almost complete inhibition of fibril formation without changing the concentration of TTR or monomeric Aß.

The Properties of Amyloid-ß Fibrils Are Determined by their Path of Formation.

Fibril formation of the amyloid-ß peptide (Aß) follows a nucleation-dependent polymerization process and is associated with Alzheimer's disease. Several different lengths of Aß are observed in vivo, but Aß1-40 and Aß1-42 are the dominant forms. The fibril architectures of Aß1-40 and Aß1-42 differ and Aß1-42 assemblies are generally considered more pathogenic. We show here that monomeric Aß1-42 can be cross-templated and incorporated into the ends of Aß1-40 fibrils, while incorporation of Aß1-40 monomers into Aß1-42 fibrils is very poor. We also show that via cross-templating incorporated Aß monomers acquire the properties of the parental fibrils. The suppressed ability of Aß1-40 to incorporate into the ends of Aß1-42 fibrils and the capacity of Aß1-42 monomers to adopt the properties of Aß1-40 fibrils may thus represent two mechanisms reducing the total load of fibrils having the intrinsic, and possibly pathogenic, features of Aß1-42 fibrils in vivo. We also show that the transfer of fibrillar properties is restricted to fibril-end templating and does not apply to cross-nucleation via the recently described path of surface-catalyzed secondary nucleation, which instead generates similar structures to those acquired via de novo primary nucleation in the absence of catalyzing seeds. Taken together these results uncover an intrinsic barrier that prevents Aß1-40 from adopting the fibrillar properties of Aß1-42 and exposes that the transfer of properties between amyloid-ß fibrils are determined by their path of formation.

Modifications of the 7-Hydroxyl Group of the Transthyretin Ligand Luteolin Provide Mechanistic Insights into Its Binding Properties and High Plasma Specificity.

Amyloid formation of the plasma protein transthyretin (TTR) has been linked to familial amyloid polyneuropathy and senile systemic amyloidosis. Binding of ligands within its natural hormone binding site can stabilize the tetrameric structure and impair amyloid formation. We have recently shown that the flavonoid luteolin stabilizes TTR in human plasma with a very high selectivity. Luteolin, however, is inactivated in vivo via glucuronidation for which the preferred site is the hydroxy group at position 7 on its aromatic A-ring. We have evaluated the properties of two luteolin variants in which the 7-hydroxy group has been exchanged for a chlorine (7-Cl-Lut) or a methoxy group (7-MeO-Lut). Using an in vitro model, based on human liver microsomes, we verified that these modifications increase the persistence of the drug. Crystal structure determinations show that 7-Cl-Lut binds similarly to luteolin. The larger MeO substituent cannot be accommodated within the same space as the chlorine or hydroxy group and as a result 7-MeO-Lut binds in the opposite direction with the methoxy group in position 7 facing the solvent. Both 7-Cl-Lut and 7-MeO-Lut qualify as high-affinity binders, but in contrast to luteolin, they display a highly non-specific binding to other plasma components. The binding of the two conformations and the key-interactions to TTR are discussed in detail. Taken together, these results show a proof-of-concept that the persistence of luteolin towards enzymatic modification can be increased. We reveal two alternative high-affinity binding modes of luteolin to TTR and that modification in position 7 is restricted only to small substituents if the original orientation of luteolin should be preserved. In addition, the present work provides a general and convenient method to evaluate the efficacy of TTR-stabilizing drugs under conditions similar to an in vivo environment.

Tetrabromobisphenol A Is an Efficient Stabilizer of the Transthyretin Tetramer.

Amyloid formation of the human plasma protein transthyretin (TTR) is associated with several human disorders, including familial amyloidotic polyneuropathy (FAP) and senile systemic amyloidosis. Dissociation of TTR's native tetrameric assembly is the rate-limiting step in the conversion into amyloid, and this feature presents an avenue for intervention because binding of an appropriate ligand to the thyroxin hormone binding sites of TTR stabilizes the native tetrameric assembly and impairs conversion into amyloid. The desired features for an effective TTR stabilizer include high affinity for TTR, high selectivity in the presence of other proteins, no adverse side effects at the effective concentrations, and a long half-life in the body. In this study we show that the commonly used flame retardant tetrabromobisphenol A (TBBPA) efficiently stabilizes the tetrameric structure of TTR. The X-ray crystal structure shows TBBPA binding in the thyroxine binding pocket with bromines occupying two of the three halogen binding sites. Interestingly, TBBPA binds TTR with an extremely high selectivity in human plasma, and the effect is equal to the recently approved drug tafamidis and better than diflunisal, both of which have shown therapeutic effects against FAP. TBBPA consequently present an interesting scaffold for drug design. Its absorption, metabolism, and potential side-effects are discussed.

The flavonoid luteolin, but not luteolin-7-O-glucoside, prevents a transthyretin mediated toxic response.

Transthyretin (TTR) is a homotetrameric plasma protein with amyloidogenic properties that has been linked to the development of familial amyloidotic polyneuropathy (FAP), familial amyloidotic cardiomyopathy, and senile systemic amyloidosis. The in vivo role of TTR is associated with transport of thyroxine hormone T4 and retinol-binding protein. Loss of the tetrameric integrity of TTR is a rate-limiting step in the process of TTR amyloid formation, and ligands with the ability to bind within the thyroxin binding site (TBS) can stabilize the tetramer, a feature that is currently used as a therapeutic approach for FAP. Several different flavonoids have recently been identified that impair amyloid formation. The flavonoid luteolin shows therapeutic potential with low incidence of unwanted side effects. In this work, we show that luteolin effectively attenuates the cytotoxic response to TTR in cultured neuronal cells and rescues the phenotype of a Drosophila melanogaster model of FAP. The plant-derived luteolin analogue cynaroside has a glucoside group in position 7 of the flavone A-ring and as opposed to luteolin is unable to stabilize TTR tetramers and thus prevents a cytotoxic effect. We generated high-resolution crystal-structures of both TTR wild type and the amyloidogenic mutant V30M in complex with luteolin. The results show that the A-ring of luteolin, in contrast to what was previously suggested, is buried within the TBS, consequently explaining the lack of activity from cynaroside. The flavonoids represent an interesting group of drug candidates for TTR amyloidosis. The present investigation shows the potential of luteolin as a stabilizer of TTR in vivo. We also show an alternative orientation of luteolin within the TBS which could represent a general mode of binding of flavonoids to TTR and is of importance concerning the future design of tetramer stabilizing drugs.

Enthalpic Forces Correlate with the Selectivity of Transthyretin-Stabilizing Ligands in Human Plasma.

The plasma protein transthyretin (TTR) is linked to human amyloidosis. Dissociation of its native tetrameric assembly is a rate-limiting step in the conversion from a native structure into a pathological amyloidogenic fold. Binding of small molecule ligands within the thyroxine binding site of TTR can stabilize the tetrameric integrity and is a potential therapeutic approach. However, through the characterization of nine different tetramer-stabilizing ligands we found that unspecific binding to plasma components might significantly compromise ligand efficacy. Surprisingly the binding strength between a particular ligand and TTR does not correlate well with its selectivity in plasma. However, through analysis of the thermodynamic signature using isothermal titration calorimetry we discovered a better correlation between selectivity and the enthalpic component of the interaction. This is of specific interest in the quest for more efficient TTR stabilizers, but a high selectivity is an almost universally desired feature within drug design and the finding might have wide-ranging implications for drug design.

A generic method for design of oligomer-specific antibodies.

Antibodies that preferentially and specifically target pathological oligomeric protein and peptide assemblies, as opposed to their monomeric and amyloid counterparts, provide therapeutic and diagnostic opportunities for protein misfolding diseases. Unfortunately, the molecular properties associated with oligomer-specific antibodies are not well understood, and this limits targeted design and development. We present here a generic method that enables the design and optimisation of oligomer-specific antibodies. The method takes a two-step approach where discrimination between oligomers and fibrils is first accomplished through identification of cryptic epitopes exclusively buried within the structure of the fibrillar form. The second step discriminates between monomers and oligomers based on differences in avidity. We show here that a simple divalent mode of interaction, as within e.g. the IgG isotype, can increase the binding strength of the antibody up to 1500 times compared to its monovalent counterpart. We expose how the ability to bind oligomers is affected by the monovalent affinity and the turnover rate of the binding and, importantly, also how oligomer specificity is only valid within a specific concentration range. We provide an example of the method by creating and characterising a spectrum of different monoclonal antibodies against both the Aß peptide and α-synuclein that are associated with Alzheimer's and Parkinson's diseases, respectively. The approach is however generic, does not require identification of oligomer-specific architectures, and is, in essence, applicable to all polypeptides that form oligomeric and fibrillar assemblies.