Targeted Identification of Protein Interactions in Eukaryotic mRNA Translation

To identify protein–protein interactions and phosphorylated amino acid sites in eukaryotic mRNA translation, replicate TAP‐MudPIT and control experiments are performed targeting Saccharomyces cerevisiae genes previously implicated in eukaryotic mRNA translation by their genetic and/or functional roles in translation initiation, elongation, termination, or interactions with ribosomal complexes. Replicate tandem affinity purifications of each targeted yeast TAP‐tagged mRNA translation protein coupled with multidimensional liquid chromatography and tandem mass spectrometry analysis are used to identify and quantify copurifying proteins. To improve sensitivity and minimize spurious, nonspecific interactions, a novel cross‐validation approach is employed to identify the most statistically significant protein–protein interactions. Using experimental and computational strategies discussed herein, the previously described protein composition of the canonical eukaryotic mRNA translation initiation, elongation, and termination complexes is calculated. In addition, statistically significant unpublished protein interactions and phosphorylation sites for S. cerevisiae’s mRNA translation proteins and complexes are identified.     

Protein Function Prediction: From Traditional Classifier to Deep Learning

Deep learning demonstrates greater competence over traditional machine learning techniques for many tasks. In last several years, deep learning has been applied to protein function prediction and a series of good achievements has been obtained. These findings extensively advanced our understanding of protein function. However, the accuracy of protein function prediction based upon deep learning still has yet to be improved. In article number 1900019, Issue 12, Zhang et al. construct DeepFunc, a deep learning framework using derived feature information of protein sequence and protein interactions network. They find that implementing DeepFunc for protein function prediction is more accurate than using DeepGO, a similar method reported previously. Meanwhile, they find that the method of combining multiple derived feature information in DeepFunc is much better than the method of using only single derived feature information. Due to its fully exploiting feature representation learning ability, deep learning with more derived feature information will enable it to be a promising method for solving more complicated protein function prediction problems and other bioinformatics challenges. Recent researches have provided some major insights into the value for using deep learning to protein function prediction problem.     

Subunit-specific backbone NMR assignments of a 64 kDa trp repressor/DNA complex: A role for N-terminal residues in tandem binding

Deuterium decoupled, triple resonance NMR spectroscopy was used to analyze complexes of ²H,¹⁵N,¹³C labelled intact and (des2–7) trp repressor (2–7 trpR) from E. coli bound in tandem to an idealized 22 basepair trp operator DNA fragment and the corepressor 5-methyltryptophan. The DNA sequence used here binds two trpR dimers in tandem resulting in chemically nonequivalent environments for the two subunits of each dimer. Sequence- and subunit-specific NMR resonance assignments were made for backbone ¹HN, ¹⁵N, ¹³C positions in both forms of the protein and for¹³ C in the intact repressor. The differences in backbone chemical shifts between the two subunits within each dimer of 2–7 trpR reflect dimer-dimer contacts involving the helix-turn-helix domains and N-terminal residues consistent with a previously determined crystal structure [Lawson and Carey (1993) Nature, 366, 178–182]. Comparison of the backbone chemical shifts of DNA-bound 2–7 trpR with those of DNA-bound intact trpR reveals significant changes for those residues involved in N-terminal-mediated interactions observed in the crystal structure. In addition, our solution NMR data contain three sets of resonances for residues 2–12 in intact trpR suggesting that the N-terminus has multiple conformations in the tandem complex. Analysis of C chemical shifts using a chemical shift index (CSI) modified for deuterium isotope effects has allowed a comparison of the secondary structure of intact and 2–7 tprR. Overall these data demonstrate that NMR backbone chemical shift data can be readily used to study specific structural details of large protein complexes.     

Solid-phase Chemistry: A Useful Tool to Discover Modulators of Protein Interactions

The Solid phase synthesis (SPS) concept, first developed for biopolymers, has spread in every field where organic synthesis is involved. While the potential of the solid-phase method was obvious in 1959 to its discoverer, Prof. R. B. Merrifield, it was unpredictable its dominance in peptide synthesis and especially in combinatorial chemistry, an area not yet conceived. SPS paved the way for solid-phase combinatorial approaches (extensively reviewed in (Choong, I. C. and Ellman, J. A.: 1996, Annu. Rep. Med. Chem. 31, 309–318; Obrecht, D. and Villalgordo, J. M.: 1998, Solid-supported Combinatorial and Parallel Synthesis of Small-Molecular-Weight Compound Libraries. Pergamon Press Ltd., Oxford, UK; Chabala, J. C.: 1995, Curr. Opin. Biotechnol. 6, 632–639; Kamal, A., Reddy, K. L., Devaiah, V., Shankaraiah, N., Reddy, D. R.: 2006, Mini Rev. Med. Chem. 6, 53–69; Whitehead, D. M., McKeown, S. C., Routledge, A.: 2005, Comb. Chem. HTS 8, 361–371; Nefzi, A., Ostresh, J. M., Houghten, R. A.: 1997, Chem. Rev. 97, 449–472; Gordon, E. M., Gallop, M. A., Patel, D. V.: 1996, Acc. Chem. Res. 29, 144–154)) as many laboratories and companies focused on the development of technologies and chemistry suitable to this new methodology. This resulted in the spectacular outburst of combinatorial chemistry, which profoundly changed the approach for new drug discovery. Combinatorial chemistry is currently considered a valid approach for a wide range of biomedical applications, such as, target validation and drug discovery.     

Protein Interaction Network-based Deep Learning Framework for Identifying Disease-Associated Human Proteins

Infectious diseases in humans appear to be one of the most primary public health issues. Identification of novel disease-associated proteins will furnish an efficient recognition of the novel therapeutic targets. Here, we develop a Graph Convolutional Network (GCN)-based model called PINDeL to identify the disease-associated host proteins by integrating the human Protein Locality Graph and its corresponding topological features. Because of the amalgamation of GCN with the protein interaction network, PINDeL achieves the highest accuracy of $83.45\%$ while AUROC and AUPRC values are $0.90$ and $0.88$, respectively. With high accuracy, recall, F1-score, specificity, AUROC, and AUPRC, PINDeL outperforms other existing machine-learning and deep-learning techniques for disease gene/protein identification in humans. Application of PINDeL on an independent dataset of 24320 proteins, which are not used for training, validation, or testing purposes, predicts 6448 new disease-protein associations of which we verify 3196 disease-proteins through experimental evidence like disease ontology, Gene Ontology, and KEGG pathway enrichment analyses. Our investigation informs that experimentally-verified 748 proteins are indeed responsible for pathogen-host protein interactions of which 22 disease-proteins share their association with multiple diseases such as cancer, aging, chem-dependency, pharmacogenomics, normal variation, infection, and immune-related diseases. This unique Graph Convolution Network-based prediction model is of utmost use in large-scale disease-protein association prediction and hence, will provide crucial insights on disease pathogenesis and will further aid in developing novel therapeutics.     

深度学习在药物活性预测研究中的应用

药物从研发到临床应用需要耗费较长的时间，研发期间的投入成本可高达十几亿元。而随着医药研发与人工智能的结合以及生物信息学的飞速发展，药物活性相关数据急剧增加，传统的实验手段进行药物活性预测已经难以满足药物研发的需求。借助算法来辅助药物研发，解决药物研发中的各种问题能够大大推动药物研发进程。传统机器学习方法尤其是随机森林、支持向量机和人工神经网络在药物活性方面能够达到较高的预测精度。深度学习由于具有多层神经网络，模型可以接收高维的输入变量且不需要人工限定数据输入特征，可以拟合较为复杂的函数模型，应用于药物研发可以进一步提高各个环节的效率。在药物活性预测中应用较为广泛的深度学习模型主要是深度神经网络（deep neural networks，DNN）、循环神经网络（recurrent neural networks，RNN）和自编码器（auto encoder，AE），而生成对抗网络（generative adversarial networks，GAN）由于其生成数据的能力常常被用来和其他模型结合进行数据增强。近年来深度学习在药物分子活性预测方面的研究和应用综述表明，深度学习模型的准确度和效率均高于传统实验方法和传统机器学习方法。因此，深度学习模型有望成为药物研发领域未来十年最重要的辅助计算模型。     

Functional protein domains from the thermally driven motion of polypeptide chains: A proposal

It is proposed that the thermally driven motion of certain polypeptide chains, including those that are part of an otherwise stable folded protein, produces time-averaged three-dimensional domains that confer unique functions to a protein. These domains may be controlled by collapsing the polypeptide into an enthalpically favored structure, or extending it into an entropically dominated form. In the extended form, these domains occupy a relatively large space, which may be used to regulate protein–protein interactions and confer mechanical properties to proteins. This “entropic bristle” model makes several predictions about the structure and properties of these domains, and the predictions are used to reevaluate a range of biophysical studies on proteins. The outcome of the analysis suggests that the entropic bristle can be used to explain a wide range of disparate and apparently unrelated experimental observations. Proteins 32:223–228, 1998. © 1998 Wiley-Liss, Inc.     

MotifAnalyzer‐PDZ: A computational program to investigate the evolution of PDZ‐binding target specificity

Recognition of short linear motifs (SLiMs) or peptides by proteins is an important component of many cellular processes. However, due to limited and degenerate binding motifs, prediction of cellular targets is challenging. In addition, many of these interactions are transient and of relatively low affinity. Here, we focus on one of the largest families of SLiM‐binding domains in the human proteome, the PDZ domain. These domains bind the extreme C‐terminus of target proteins, and are involved in many signaling and trafficking pathways. To predict endogenous targets of PDZ domains, we developed MotifAnalyzer‐PDZ, a program that filters and compares all motif‐satisfying sequences in any publicly available proteome. This approach enables us to determine possible PDZ binding targets in humans and other organisms. Using this program, we predicted and biochemically tested novel human PDZ targets by looking for strong sequence conservation in evolution. We also identified three C‐terminal sequences in choanoflagellates that bind a choanoflagellate PDZ domain, the Monsiga brevicollis SHANK1 PDZ domain (mbSHANK1), with endogenously‐relevant affinities, despite a lack of conservation with the targets of a homologous human PDZ domain, SHANK1. All three are predicted to be signaling proteins, with strong sequence homology to cytosolic and receptor tyrosine kinases. Finally, we analyzed and compared the positional amino acid enrichments in PDZ motif‐satisfying sequences from over a dozen organisms. Overall, MotifAnalyzer‐PDZ is a versatile program to investigate potential PDZ interactions. This proof‐of‐concept work is poised to enable similar types of analyses for other SLiM‐binding domains (e.g., MotifAnalyzer‐Kinase). MotifAnalyzer‐PDZ is available at http://motifAnalyzerPDZ.cs.wwu.edu .     

Three-dimensional domain swapping in homooligomeric proteins and its functional significance

In the process of oligomeric structure formation through a mechanism of three-dimensional domain swapping, one domain of a monomeric protein is replaced by the same domain from an identical monomer. The swapped domain can represent an entire tertiary globular domain or an element of secondary protein structure, such as an -helix or a -strand. Different examples of three-dimensional domain swapping are reviewed; the functional importance of this phenomenon and its role in the development of new properties by some proteins in the process of evolution are considered. The contribution of three-dimensional domain swapping to the formation of linear protein polymers and amyloids is discussed.     

A simple contact mapping algorithm for identifying potential peptide mimetics in protein–protein interaction partners

A simple, static contact mapping algorithm has been developed as a first step at identifying potential peptide biomimetics from protein interaction partner structure files. This rapid and simple mapping algorithm, “OpenContact” provides screened or parsed protein interaction files based on specified criteria for interatomic separation distances and interatomic potential interactions. The algorithm, which uses all‐atom Amber03 force field models, was blindly tested on several unrelated cases from the literature where potential peptide mimetics have been experimentally developed to varying degrees of success. In all cases, the screening algorithm efficiently predicted proposed or potential peptide biomimetics, or close variations thereof, and provided complete atom‐atom interaction data necessary for further detailed analysis and drug development. In addition, we used the static parsing/mapping method to develop a peptide mimetic to the cancer protein target, epidermal growth factor receptor. In this case, secondary, loop structure for the peptide was indicated from the intra‐protein mapping, and the peptide was subsequently synthesized and shown to exhibit successful binding to the target protein. The case studies, which all involved experimental peptide drug advancement, illustrate many of the challenges associated with the development of peptide biomimetics, in general. Proteins 2014; 82:2253–2262. © 2014 The Authors. Proteins: Structure, Function, and Bioinformatics Published by Wiley Periodicals, Inc.     

PPICurator: A Tool for Extracting Comprehensive Protein–Protein Interaction Information

Protein–protein interaction extraction through biological literature curation is widely employed for proteome analysis. There is a strong need for a tool that can assist researchers in extracting comprehensive PPI information through literature curation, which is critical in research on protein, for example, construction of protein interaction network, identification of protein signaling pathway, and discovery of meaningful protein interaction. However, most of current tools can only extract PPI relations. None of them are capable of extracting other important PPI information, such as interaction directions, effects, and functional annotations. To address these issues, this paper proposes PPICurator, a novel tool for extracting comprehensive PPI information with a variety of logic and syntax features based on a new support vector machine classifier. PPICurator provides a friendly web‐based user interface. It is a platform that automates the extraction of comprehensive PPI information through literature, including PPI relations, as well as their confidential scores, interaction directions, effects, and functional annotations. Thus, PPICurator is more comprehensive than state‐of‐the‐art tools. Moreover, it outperforms state‐of‐the‐art tools in the accuracy of PPI relation extraction measured by F‐score and recall on the widely used open datasets. PPICurator is available at https://ppicurator.hupo.org.cn .     

CRK: An evolutionary approach for distinguishing biologically relevant interfaces from crystal contacts

Protein crystals contain two different types of interfaces: biologically relevant ones, observed in protein–protein complexes and oligomeric proteins, and nonspecific ones, corresponding to crystal lattice contacts. Because of the increasing complexity of the objects being tackled in structural biology, distinguishing biological contacts from crystal contacts is not always a trivial task and can lead to wrong interpretation of macromolecular structures. We devised an approach (CRK, core‐rim K_a/K_s ratio) for distinguishing biologically relevant interfaces from nonspecific ones. Given a protein–protein interface, CRK finds a set of homologs to the sequences of the proteins involved in the interface, retrieves and aligns the corresponding coding sequences, on which it carries out a residue‐by‐residue K_a/K_s ratio (ω) calculation. It divides interface residues into a “rim” and a “core” set and analyzes the selection pressure on the residues belonging to the two sets. We developed and tested CRK on different datasets and test cases, consisting of biologically relevant contacts, nonspecific ones or of both types. The method proves very effective in distinguishing the two categories of interfaces, with an overall accuracy rate of 84%. As it relies on different principles when compared with existing tools, CRK is optimally suited to be used in combination with them. In addition, CRK has potential applications in the validation of structures of oligomeric proteins and protein complexes. Proteins 2010. © 2010 Wiley‐Liss, Inc.     

Coupling Molecular Dynamics and Deep Learning to Mine Protein Conformational Space

1. Download : Download high-res image (302KB)
2. Download : Download full-size image

     

Importance of the interaction protein–protein of the CaM–PDE1A and CaM–MLCK complexes in the development of new anti‐CaM drugs

Protein–protein interactions play central roles in physiological and pathological processes. The bases of the mechanisms of drug action are relevant to the discovery of new therapeutic targets. This work focuses on understanding the interactions in protein–protein–ligands complexes, using proteins calmodulin (CaM), human calcium/calmodulin‐dependent 3′,5′‐cyclic nucleotide phosphodiesterase 1A active human (PDE1A), and myosin light chain kinase (MLCK) and ligands αII–spectrin peptide (αII–spec), and two inhibitors of CaM (chlorpromazine (CPZ) and malbrancheamide (MBC)). The interaction was monitored with a fluorescent biosensor of CaM (hCaM M124C–mBBr). The results showed changes in the affinity of CPZ and MBC depending on the CaM–protein complex under analysis. For the Ca²⁺–CaM, Ca²⁺–CaM–PDE1A, and Ca²⁺–CaM–MLCK complexes, CPZ apparent dissociation constants (K_ds) were 1.11, 0.28, and 0.55 μM, respectively; and for MBC K_ds were 1.43, 1.10, and 0.61 μM, respectively. In competition experiments the addition of calmodulin binding peptide 1 (αII–spec) to Ca²⁺–hCaM M124C–mBBr quenched the fluorescence (K_d = 2.55 ± 1.75 pM) and the later addition of MBC (up to 16 μM) did not affect the fluorescent signal. Instead, the additions of αII–spec to a preformed Ca²⁺–hCaM M124C–mBBr–MBC complex modified the fluorescent signal. However, MBC was able to displace the PDE1A and MLCK from its complex with Ca²⁺–CaM. In addition, docking studies were performed for all complexes with both ligands showing an excellent correlation with experimental data. These experiments may help to explain why in vivo many CaM drugs target prefer only a subset of the Ca²⁺–CaM regulated proteins and adds to the understanding of molecular interactions between protein complexes and small ligands. Copyright © 2013 John Wiley & Sons, Ltd.     

Target of rapamycin FATC domain as a general membrane anchor: The FKBP‐12 like domain of FKBP38 as a case study

Increased efforts have been undertaken to better understand the formation of signaling complexes at cellular membranes. Since the preparation of proteins containing a transmembrane domain or a prenylation motif is generally challenging an alternative membrane anchoring unit that is easy to attach, water‐soluble and binds to different membrane mimetics would find broad application. The 33‐residue long FATC domain of yeast TOR1 (y1fatc) fulfills these criteria and binds to neutral and negatively charged micelles, bicelles, and liposomes. As a case study, we fused it to the FKBP506‐binding region of the protein FKBP38 (FKBP38‐BD) and used ¹H–¹⁵N NMR spectroscopy to characterize localization of the chimeric protein to micelles, bicelles, and liposomes. Based on these and published data for y1fatc, its use as a C‐terminally attachable membrane anchor for other proteins is compatible with a wide range of buffer conditions (pH circa 6–8.5, NaCl 0 to >150 mM, presence of reducing agents, different salts such as MgCl₂ and CaCl₂). The high water‐solubility of y1fatc enables its use for titration experiments against a membrane‐localized interaction partner of the fused target protein. Results from studies with peptides corresponding to the C‐terminal 17–11 residues of the 33‐residue long domain by 1D ¹H NMR and CD spectroscopy indicate that they still can interact with membrane mimetics. Thus, they may be used as membrane anchors if the full y1fatc sequence is disturbing or if a chemically synthesized y1fatc peptide shall be attached by native chemical ligation, for example, unlabeled peptide to ¹⁵N‐labeled target protein for NMR studies.     

DockTrina: Docking triangular protein trimers

In spite of the abundance of oligomeric proteins within a cell, the structural characterization of protein–protein interactions is still a challenging task. In particular, many of these interactions involve heteromeric complexes, which are relatively difficult to determine experimentally. Hence there is growing interest in using computational techniques to model such complexes. However, assembling large heteromeric complexes computationally is a highly combinatorial problem. Nonetheless the problem can be simplified greatly by considering interactions between protein trimers. After dimers and monomers, triangular trimers (i.e. trimers with pair‐wise contacts between all three pairs of proteins) are the most frequently observed quaternary structural motifs according to the three‐dimensional (3D) complex database. This article presents DockTrina, a novel protein docking method for modeling the 3D structures of nonsymmetrical triangular trimers. The method takes as input pair‐wise contact predictions from a rigid body docking program. It then scans and scores all possible combinations of pairs of monomers using a very fast root mean square deviation test. Finally, it ranks the predictions using a scoring function which combines triples of pair‐wise contact terms and a geometric clash penalty term. The overall approach takes less than 2 min per complex on a modern desktop computer. The method is tested and validated using a benchmark set of 220 bound and seven unbound protein trimer structures. DockTrina will be made available at http://nano‐d.inrialpes.fr/software/docktrina . Proteins 2014; 82:34–44. © 2013 Wiley Periodicals, Inc.     

Tubulin rings: Curved filaments with limited flexibility and two modes of association

Tubulin rings have been previously identified as composed of linear polymers of tubulin subunits, equivalent to a protofilament in the microtubule wall but in a curved rather than a straight conformation. We have examined and measured a number of different ring structures obtained under different conditions. The preferred curvature is indicated by a single ring of 380 Å outside diameter. Radially double rings consist of two coplanar rings of 460 Å and 350 Å outside diameter, held together by a pattern of eight identical contacts between the 40 Å subunits in the inner and outer rings. In some circumstances a larger ring, 570 Å diameter, can be added to the outside, or a smaller ring, 240 Å diameter, may be added to the inside of the radially double ring, in both cases repeating the pattern of eight radial contacts. The distortion of the filament from its relaxed 380 Å diameter curvature apparently can be made without disrupting the longitudinal bond between subunits in the filament, but must be stabilized by the energy of the radial contacts. All of these rings (single and radially double and triple) are observed to associate axially to form pairs or in some cases larger stacks. The radially double rings or an axially associated pair of these (quadruple ring) may also associate to form crystals. These are thin plates, up to 100 μm in extent and several μm thick which have been of limited use so far in diffraction studies because of irregularities in the packing of adjacent rings.     

NMR analysis of the interaction of picornaviral proteinases Lb and 2A with their substrate eukaryotic initiation factor 4GII

Messenger RNA is recruited to the eukaryotic ribosome by a complex including the eukaryotic initiation factor (eIF) 4E (the cap‐binding protein), the scaffold protein eIF4G and the RNA helicase eIF4A. To shut‐off host–cell protein synthesis, eIF4G is cleaved during picornaviral infection by a virally encoded proteinase; the structural basis of this reaction and its stimulation by eIF4E is unclear. We have structurally and biochemically investigated the interaction of purified foot‐and‐mouth disease virus (FMDV) leader proteinase (Lb^pro), human rhinovirus 2 (HRV2) 2A proteinase (2A^pro) and coxsackievirus B4 (CVB4) 2A^pro with purified eIF4GII, eIF4E and the eIF4GII/eIF4E complex. Using nuclear magnetic resonance (NMR), we completed ¹³C/¹⁵N sequential backbone assignment of human eIF4GII residues 551–745 and examined their binding to murine eIF4E. eIF4GII_551–745 is intrinsically unstructured and remains so when bound to eIF4E. NMR and biophysical techniques for determining stoichiometry and binding constants revealed that the papain‐like Lb^pro only forms a stable complex with eIF4GII_551–745 in the presence of eIF4E, with K_D values in the low nanomolar range; Lb^pro contacts both eIF4GII and eIF4E. Furthermore, the unrelated chymotrypsin‐like 2A^pro from HRV2 and CVB4 also build a stable complex with eIF4GII/eIF4E, but with K_D values in the low micromolar range. The HRV2 enzyme also forms a stable complex with eIF4E; however, none of the proteinases tested complex stably with eIF4GII alone. Thus, these three picornaviral proteinases have independently evolved to establish distinct triangular heterotrimeric protein complexes that may actively target ribosomes involved in mRNA recruitment to ensure efficient host cell shut‐off.     

Structural characterization of the tunnels of Mycobacterium tuberculosis truncated hemoglobin N from molecular dynamics simulations

The structure of oxygenated trHbN from Mycobacterium tuberculosis shows an extended heme distal hydrogen‐bond network that includes Tyr33(B10), Gln58(E11), and the bound O₂. In addition, trHbN structure shows a network of hydrophobic cavities organized in two orthogonal branches. In the present work, the structure and the dynamics of oxygenated and deoxygenated trHbN in explicit water was investigated from 100 ns molecular dynamics (MD) simulations. Results show that, depending on the presence or the absence of a coordinated O₂, the Tyr33(B10) and Gln58(E11) side chains adopt two different configurations in concert with hydrogen bond network rearrangement. In addition, our data indicate that Tyr33(B10) and Gln58(E11) control the dynamics of Phe62(E15). In deoxy‐trHbN, Phe62(E15) is restricted to one conformation. Upon O₂ binding, the conformation of Gln58(E11) changes and residue Phe62(E15) fluctuates between two conformations. We also conducted a systematic study of trHbN tunnels by analyzing thousands of MD snapshots with CAVER. The results show that tunnel formation is the result of the dynamic reshaping of short‐lived hydrophobic cavities. The analyses indicate that the presence of these cavities is likely linked to the rigid structure of trHbN and also reveal two tunnels, EH and BE, that link the protein surface to the buried distal heme pocket and not present in the crystallographic structure. The cavities are sufficiently large to accomodate and store ligands. Tunnel dynamics in trHbN was found to be controlled by the side‐chain conformation of the Tyr33(B10), Gln58(E11), and Phe62(E15) residues. Importantly, in contrast to recently published works, our extensive systematic studies show that the presence or absence of a coordinated dioxygen does not control the opening of the long tunnel but rather the opening of the EH tunnel. In addition, the data lead to new and distinctly different conclusion on the impact of the Phe62(E15) residue on trHbN tunnels. We propose that the EH and the long tunnels are used for apolar ligands storage. The trajectories bring important new structural insights related to trHbN function and to ligand diffusion in proteins. Proteins 2009. © 2008 Wiley‐Liss, Inc.