Haruspex: A Neural Network for the Automatic Identification of Oligonucleotides and Protein Secondary Structure in Cryo-Electron Microscopy Maps

In recent years, three-dimensional density maps reconstructed from single particle images obtained by electron cryo-microscopy (cryo-EM) have reached unprecedented resolution. However, map interpretation can be challenging, in particular if the constituting structures require de-novo model building or are very mobile. Herein, we demonstrate the potential of convolutional neural networks for the annotation of cryo-EM maps: our network Haruspex has been trained on a carefully curated set of 293 experimentally derived reconstruction maps to automatically annotate RNA/DNA as well as protein secondary structure elements. It can be straightforwardly applied to newly reconstructed maps in order to support domain placement or as a starting point for main-chain placement. Due to its high recall and precision rates of 95.1 % and 80.3 %, respectively, on an independent test set of 122 maps, it can also be used for validation during model building. The trained network will be available as part of the CCP-EM suite.     

Electron Cryo‐microscopy as a Tool for Structure‐Based Drug Development

For decades, X‐ray crystallography and NMR have been the most important techniques for studying the atomic structure of macromolecules. However, as a result of size, instability, low yield, and other factors, many macromolecules are difficult to crystallize or unsuitable for NMR studies. Electron cryo‐microscopy (cryo‐EM) does not depend on crystals and has therefore been the method of choice for many macromolecular complexes that cannot be crystallized, but atomic resolution has mostly been beyond its reach. A new generation of detectors that are capable of sensing directly the incident electrons has recently revolutionized the field, with structures of macromolecules now routinely being solved to near‐atomic resolution. In this review, we summarize some of the most recent examples of high‐resolution cryo‐EM structures. We put particular emphasis on proteins with pharmacological relevance that have traditionally been inaccessible to crystallography. Furthermore, we discuss examples where interactions with small molecules have been fully characterized at atomic resolution. Finally, we stress the current limits of cryo‐EM, and methodological issues related to its usage as a tool for drug development.     

Polyoxometalate‐Engineered Building Blocks with Gold Cores for the Self‐Assembly of Responsive Water‐Soluble Nanostructures

The controlled assembly of gold nanoparticles (AuNPs) with the size of quantum dots into predictable structures is extremely challenging as it requires the quantitatively and topologically precise placement of anisotropic domains on their small, approximately spherical surfaces. We herein address this problem by using polyoxometalate leaving groups to transform 2 nm diameter gold cores into reactive building blocks with hydrophilic and hydrophobic surface domains whose relative sizes can be precisely tuned to give dimers, clusters, and larger micelle‐like organizations. Using cryo‐TEM imaging and ¹H DOSY NMR spectroscopy, we then provide an unprecedented “solution‐state” picture of how the micelle‐like structures respond to hydrophobic guests by encapsulating them within 250 nm diameter vesicles whose walls are comprised of amphiphilic AuNP membranes. These findings provide a versatile new option for transforming very small AuNPs into precisely tailored building blocks for the rational design of functional water‐soluble assemblies.     

Spherical and Worm‐Like Micelles from Fructose‐Functionalized Polyether Block Copolymers

This paper presents the synthesis and characterization of d ‐fructose modified poly(ethylene glycol) (Fru‐PEG) and fructose modified poly(ethylene glycol)‐block‐poly(ethyl hexyl glycidyl ether) (Fru‐PEG‐b‐PEHG) that are both prepared by initiation with isopropyliden protected fructose, followed by deprotection of the sugar. The block copolymers are self‐assembled into micelles, and are subsequently characterized by cryo‐TEM and dynamic light scattering. The fluorescent dye Nile red is encapsulated as a model hydrophobic compound and fluorescent marker to perform initial uptake tests with breast cancer cells. The uptake of sugar and nonsugar decorated micelles is compared.     

Unambiguous Intracellular Localization and Quantification of a Potent Iridium Anticancer Compound by Correlative 3D Cryo X‐Ray Imaging

The iridium half‐sandwich complex [Ir(η⁵:κ¹‐C₅Me₄CH₂py)(2‐phenylpyridine)]PF₆ is highly cytotoxic: 15–250× more potent than clinically used cisplatin in several cancer cell lines. We have developed a correlative 3D cryo X‐ray imaging approach to specifically localize and quantify iridium within the whole hydrated cell at nanometer resolution. By means of cryo soft X‐ray tomography (cryo‐SXT), which provides the cellular ultrastructure at 50 nm resolution, and cryo hard X‐ray fluorescence tomography (cryo‐XRF), which provides the elemental sensitivity with a 70 nm step size, we have located the iridium anticancer agent exclusively in the mitochondria. Our methodology provides unique information on the intracellular fate of the metallodrug, without chemical fixation, labeling, or mechanical manipulation of the cells. This cryo‐3D correlative imaging method can be applied to a number of biochemical processes for specific elemental localization within the native cellular landscape.     

“Sandglass”‐Shaped Self‐Assembly of Coil–rod–coil Triblock Copolymer Containing Rigid Aniline‐Pentamer

The amphiphilic PEG_{1 500}‐b‐EM AP‐b‐PEG_{1 500} (EM PAP) triblock copolymer of poly(ethylene glycol) (PEG) and emeraldine aniline‐pentamer (EM AP) in its concentrated solution can self‐assemble into a special shape like “sandglass”, as observed by transmission electron microscopy (TEM), field emission scanning electron microscopy (ESEM) and atomic force microscopy (AFM). This “sandglass”‐shaped assembly is composed of several “rods” aggregated in the middle, with every “rod” being about 8 µm in length and 300 nm in diameter. We conclude that the special “sandglass”‐shaped assembly may come into being because of the inducement effect of the crystallization of EM AP segments, by studying electron diffraction (ED) results and wide‐angle X‐ray diffusion (WAXD) characterization of the EM PAP triblock copolymer.

     

Representation of molecular electrostatic potentials of biopolymer by self‐organizing feature map

The Kohonen self‐organizing map was introduced to map the protein molecular surface features. The protein or polypeptide properties, such as shape and molecular electrostatic potential, can be visualized by self‐organizing map, which was trained by the 3D surface coordinates. Such maps allow the visual comparison of molecular properties between proteins having common topological or chemical features.     

Induced Folding of Protein‐Sized Foldameric β‐Sandwich Models with Core β‐Amino Acid Residues

The mimicry of protein‐sized β‐sheet structures with unnatural peptidic sequences (foldamers) is a considerable challenge. In this work, the de novo designed betabellin‐14 β‐sheet has been used as a template, and α→β residue mutations were carried out in the hydrophobic core (positions 12 and 19). β‐Residues with diverse structural properties were utilized: Homologous β³‐amino acids, (1R,2S)‐2‐aminocyclopentanecarboxylic acid (ACPC), (1R,2S)‐2‐aminocyclohexanecarboxylic acid (ACHC), (1R,2S)‐2‐aminocyclohex‐3‐enecarboxylic acid (ACEC), and (1S,2S,3R,5S)‐2‐amino‐6,6‐dimethylbicyclo[3.1.1]heptane‐3‐carboxylic acid (ABHC). Six α/β‐peptidic chains were constructed in both monomeric and disulfide‐linked dimeric forms. Structural studies based on circular dichroism spectroscopy, the analysis of NMR chemical shifts, and molecular dynamics simulations revealed that dimerization induced β‐sheet formation in the 64‐residue foldameric systems. Core replacement with (1R,2S)‐ACHC was found to be unique among the β‐amino acid building blocks studied because it was simultaneously able to maintain the interstrand hydrogen‐bonding network and to fit sterically into the hydrophobic interior of the β‐sandwich. The novel β‐sandwich model containing 25 % unnatural building blocks afforded protein‐like thermal denaturation behavior.     

On‐the‐fly reconstruction of free‐energy profiles using logarithmic mean‐force dynamics

Mean‐force dynamics (MFD), which is a fictitious dynamics for a set of collective variables on a potential of mean‐force, is a powerful algorithm to efficiently explore free‐energy landscapes. Recently, we have introduced logarithmic MFD (LogMFD) (Morishita et al., Phys. Rev. E 2012, 85, 066702) which overcomes difficulties encounterd in free‐energy calculations using standard approaches such as thermodynamic integration. Here, we present a guide to implementing LogMFD calculations paying attention to the practical issues in choosing the parameters in LogMFD. A primary focus is given to the effect of the parameters on the accuracy of the reconstructed free‐energy profiles. A recipe for reducing the errors due to energy dissipation is presented. We also demonstrate that multidimensional free‐energy landscapes can be reconstructed on‐the‐fly using LogMFD, which cannot be accomplished using any other free‐energy calculation techniques. © 2013 Wiley Periodicals, Inc.     

Hydrophobic Monomers Recognize Microenvironments in Hydrogel Microspheres during Free‐Radical‐Seeded Emulsion Polymerization

The three‐dimensional structure of nanocomposite microgels was precisely determined by cryo‐electron micrography. Several nanocomposite microgels that differ with respect to their nanocomposite structure, which were obtained from seeded emulsion polymerization in the presence of microgels, were used as model nanocomposite materials for cryo‐electron micrography. The obtained three‐dimensional segmentation images of these nanocomposite microgels provide important insights into the interactions between the hydrophobic monomers and the microgels, that is, hydrophobic styrene monomers recognize molecular‐scale differences in polarity within the microgels during the emulsion polymerization. This result led to the formation of unprecedented multi‐layered nanocomposite microgels, which promise substantial potential in colloidal applications.     

Hyaluronic acid‐poly(vinyl alcohol) composite cryo‐gel for biofunctional material application

Recently, applications of hyaluronic acid (HA) as a biomaterial were investigated. However, the weak structure of HA gel and the effects of using cross‐linker raised concerns during in vivo resolution. In this study, we investigated the method to solve these two problems using physical cross‐linking and compositing with poly(vinyl alcohol) (PVA). Various weight ratios of HA and PVA solutions were mixed, adjustment of pH to 1.8 using HCl then used to fabricate HA‐PVA cryo‐gel by freezing‐thawing. Young's modulus of the prepared gel rose with the increase of both HA and PVA concentrations or either one of them. We estimated that HA and PVA have exhibited these mechanical properties due to forming a double network. HA‐PVA gel showed kinetic friction force of approximately 10 times of PVA gel, while water contact angle and protein adsorption of HA‐PVA gels were remarkably decreased. The properties of the prepared gel suggest that it can be used for postoperative adhesion prevention applications.     

3,6,7‐Triamino‐[1,2,4]triazolo[4,3‐b][1,2,4]triazole: A Non‐toxic,High‐Performance Energetic Building Block with Excellent Stability

A novel strategy for the design of energetic materials that uses fused amino‐substituted triazoles as energetic building blocks is presented. The 3,6,7‐triamino‐7H‐[1,2,4]triazolo[4,3‐b][1,2,4]triazolium (TATOT) motif can be incorporated into many ionic, nitrogen‐rich materials to form salts with advantages such as remarkably high stability towards physical or mechanical stimuli, excellent calculated detonation velocity, and toxicity low enough to qualify them as “green explosives”. Neutral TATOT can be synthesized in a convenient and inexpensive two‐step protocol in high yield. To demonstrate the superior properties of TATOT, 13 ionic derivatives were synthesized and their chemical‐ and physicochemical properties (e.g., sensitivities towards impact, friction and electrostatic discharge) were investigated extensively. Low toxicity was demonstrated for neutral TATOT and its nitrate salt. Both are insensitive towards impact and friction and the nitrate salt combines outstanding thermal stability (decomposition temperature=280 °C) with promising calculated energetic values.     

Understanding the Signatures of Secondary‐Structure Elements in Proteins with Raman Optical Activity Spectroscopy

A prerequisite for the understanding of functional molecules like proteins is the elucidation of their structure under reaction conditions. Chiral vibrational spectroscopy is one option for this purpose, but provides only indirect access to this structural information. By first‐principles calculations, we investigate how Raman optical activity (ROA) signals in proteins are generated and how signatures of specific secondary‐structure elements arise. As a first target we focus on helical motifs and consider polypeptides consisting of twenty alanine residues to represent α‐helical and 3₁₀‐helical secondary‐structure elements. Although ROA calculations on such large molecules have not been carried out before, our main goal is the stepwise reconstruction of the ROA signals. By analyzing the calculated ROA spectra in terms of rigorously defined localized vibrations, we investigate in detail how total band intensities and band shapes emerge. We find that the total band intensities can be understood in terms of the reconstructed localized vibrations on individual amino acid residues. Two different basic mechanisms determining the total band intensities can be established, and it is explained how structural changes affect the total band intensities. The band shapes can be rationalized in terms of the coupling between the localized vibrations on different residues, and we show how different band shapes arise as a consequence of different coupling patterns. As a result, it is demonstrated for the chiral variant of Raman spectroscopy how collective vibrations in proteins can be understood in terms of well‐defined localized vibrations. Based on our calculations, we extract characteristic ROA signatures of α helices and of 3₁₀‐helices, which our analysis directly relates to differences in secondary structure.     

Methods for refining interactively established models of ribosomal RNA towards a physico‐chemically plausible structure

A computational multiscale modeling approach that was used for the refinement of the structure of the bacterial ribosome is presented. Though designed for the ribosome, the methods are applicable to other large biomolecules as well. Algorithms were developed, allowing defined groups of atoms to be clustered into rigid objects, which greatly reduces the number of parameters in the molecular dynamics approach and thus speeds up the computational process considerably (clustered molecular dynamics). The energy potential function, which is used in molecular dynamics to describe structural details of a particle, was extended to include terms that describe high‐level biochemical constraints resulting from crosslinking techniques and cryo‐electron microscopy. High‐ and low‐level features of the potential function were specified, and the clustered molecular dynamics technique was integrated into the interactive model‐building process, to establish a physico‐chemically plausible structure of the bacterial ribosomal RNA. © 2001 John Wiley & Sons, Inc. J Comput Chem 22: 407–417, 2001     

Relaxed energetic maps of κ‐carrabiose: A DFT study

The B3LYP density function was used with the 6‐31G(d) basis set to perform relaxed energetic contour maps of the charged form of κ‐carrabiose in the gas phase and for the neutral form first in the gas phase and then by simulating the presence of water as solvent using the Onsager model. Only one starting conformation has been considered to perform all the calculations. Rigid energetic maps have been then constructed either by addition of diffuse or polarization functions to the basis set obtaining in that way 6‐31+G(d)//6‐31G(d), 6‐31+G(d,p)//6‐31G(d), and 6‐311++G(d,p)//6‐31G(d) energetic maps that have been carefully examined. The obtained structures corresponding to the lower energy conformers have been then fully optimized using different basis sets with the B3LYP method, a reversion in term of energy has been observed for the two first minima in the case of the charged disaccharide in the gas phase, this was attributed to the large grid of 30° that could lead to the exclusion of an intermediate value corresponding to the real minimum of energy. We thus suggest that after establishing potential energy maps it is essential to proceed to full optimizations of the lower energy conformers. Calculations using the more accurate correlated method MP2 with the 6‐31G(d) basis set have also been performed for conformers of the two disaccharides in the gas phase. © 2010 Wiley Periodicals, Inc. J Comput Chem 2010     

ProteinUnet—An efficient alternative to SPIDER3‐single for sequence‐based prediction of protein secondary structures

Predicting protein function and structure from sequence remains an unsolved problem in bioinformatics. The best performing methods rely heavily on evolutionary information from multiple sequence alignments, which means their accuracy deteriorates for sequences with a few homologs, and given the increasing sequence database sizes requires long computation times. Here, a single‐sequence‐based prediction method is presented, called ProteinUnet, leveraging an U‐Net convolutional network architecture. It is compared to SPIDER3‐Single model, based on long short‐term memory‐bidirectional recurrent neural networks architecture. Both methods achieve similar results for prediction of secondary structures (both three‐ and eight‐state), half‐sphere exposure, and contact number, but ProteinUnet has two times fewer parameters, 17 times shorter inference time, and can be trained 11 times faster. Moreover, ProteinUnet tends to be better for short sequences and residues with a low number of local contacts. Additionally, the method of loss weighting is presented as an effective way of increasing accuracy for rare secondary structures.     

Size,Charge, and Stability of Fully Serine‐Based Catanionic Vesicles: Towards Versatile Biocompatible Nanocarriers

Vesicles based on mixed cationic and anionic surfactants (catanionic vesicles) offer a number of advantageous colloidal features over conventional lipid‐based vesicles, namely spontaneity in formation, long‐term stability, and easy modulation of size and charge. If biocompatibility is added through rational design of the chemical components, the potential for biorelated applications further emerges. Here, we report for the first time on two catanionic vesicle systems in which both ionic amphiphiles are derivatized from the same amino acid—serine—with the goal of enhancing aggregate biocompatibility. Phase behavior maps for a mixture with chain length symmetry, 12Ser/12‐12Ser, and another with asymmetry, 16Ser/8‐8Ser, are presented, for which regions of vesicles, micelles, and coexisting aggregates are identified. For the asymmetric mixture, detailed phase behavior and microstructure characterization have been carried out based on surface tension, light microscopy, cryo‐SEM, cryo‐TEM, and dynamic light scattering analysis. Vesicles are found with tunable mean size, pH, and zeta potential. Changes in aggregate shape with varying composition and the effect of preparation methods and aging on vesicle features and stability have been investigated in detail. The results are discussed in the light of self‐assembly models and related catanionic systems reported before. A versatile system of robust vesicles is thus presented for potential applications.     

Macroscopic Free‐Standing Hierarchical 3D Architectures Assembled from Silver Nanowires by Ice Templating

As macroscopic three dimensional (3D) architectures show increasing significance, much effort has been devoted to the hierarchical organization of 1D nanomaterials into serviceable macroscopic 3D assemblies. How to assemble 1D nanoscale building blocks into 3D hierarchical architectures is still a challenge. Herein we report a general strategy based on the use of ice as a template for assembling 1D nanostructures with high efficiency and good controllability. Free‐standing macroscopic 3D Ag nanowire (AgNW) assemblies with hierarchical binary‐network architectures are then fabricated from a 1D AgNW suspension for the first time. The microstructure of this 3D AgNW network endows it with electrical conductivity and allows it to be made into stretchable and foldable conductors with high electromechanical stability. These properties should make this kind of macroscopic 3D AgNW architecture and it composites suitable for electronic applications.     

Coacervation of Elastin‐Like Recombinamer Microgels

The investigation of the coacervation (self‐aggregation) behavior of biomicrogels which can potentially be used as drug carriers is an important topic, because self‐aggregation can not only cause loss of activity, but also toxicity and immunogenicity. To study this effect microgels from elastin‐like recombinamer are synthesized using miniemulsion technique. The existence of coacervation for such microgels, at different concentrations and temperatures, is studied and proved by cryo‐field emission scanning clectron microscopy (cryo‐FESEM), cryo‐transmission electron microscopy (cryo‐TEM), and by a novel ¹H high‐resolution magic angle sample spinning (HRMAS), nuclear magnetic resonance (NMR) spectroscopy, and relaxometry methods. The findings by ¹H HRMAS NMR spectroscopy and relaxometry show simultaneous processes of volume phase temperature transition and coacervation with different sensitivity for hydrophobic and hydrophilic amino acid side‐chains in the microgel. The coacervation process is more evidential by the behavior of glycine α‐CH₂, ¹H NMR peak as compared to the proline β‐CH₂.

     

Fluorescence Quenching‐based Assay for Measuring Golgi endo‐α‐Mannosidase

Golgi endo‐α‐mannosidase (G‐EM) catalyzes an alternative deglucosylation process for N‐glycans and plays important roles in the post‐endoplasmic reticulum (ER) quality control pathway. To understand the post‐ER quality control mechanism, we synthesized a tetrasaccharide probe for the detection of the hydrolytic activity of G‐EM based on a fluorescence quenching assay. The probe was labeled with an N‐methylanthraniloyl group as a reporter dye at the non‐reducing end and a 2,4‐dinitrophenyl group as a quencher at the reducing end. This probe is hydrolyzed to disaccharide derivatives by G‐EM, resulting in increased fluorescence intensity. Thus, the fluorescence signal is directly proportional to the amount of disaccharide derivative present, allowing the G‐EM activity to be evaluated easily and quantitatively.