AIDA: ab initio domain assembly for automated multi-domain protein structure prediction and domain-domain interaction prediction

M-DeepAssembly: enhanced DeepAssembly based on multi-objective multi-domain protein conformation sampling

BACKGROUND

Association and cooperation among structural domains play an important role in protein function and drug design. Despite remarkable advancements in highly accurate single-domain protein structure prediction through the collaborative efforts of the community using deep learning, challenges still exist in predicting multi-domain protein structures when the evolutionary signal for a given domain pair is weak or the protein structure is large.

RESULTS

To alleviate the above challenges, we proposed M-DeepAssembly, a protocol based on multi-objective protein conformation sampling algorithm for multi-domain protein structure prediction. Firstly, the inter-domain interactions and full-length sequence distance features are extracted through DeepAssembly and AlphaFold2, respectively. Secondly, subject to these features, we constructed a multi-objective energy model and designed a sampling algorithm for exploring and exploiting conformational space to generate ensembles. Finally, the output protein structure was selected from the ensembles using our in-house developed model quality assessment algorithm. On the test set of 164 multi-domain proteins, the results show that the average TM-score of M-DeepAssembly is 15.4% and 2.0% higher than AlphaFold2 and DeepAssembly, respectively. It is worth noting that there are models with higher accuracy in ensembles, achieving an improvement of 20.3% and 6.4% relative to the two baseline methods, although these models were not selected. Furthermore, when compared to the prediction results of AlphaFold2 for CASP15 multi-domain targets, M-DeepAssembly demonstrates certain performance advantages.

CONCLUSIONS

M-DeepAssembly provides a distinctive multi-domain protein assembly algorithm, which can alleviate the current challenges of weak evolutionary signals and large structures to some extent by forming diverse ensembles using multi-objective protein conformation sampling algorithm. The proposed method contributes to exploring the functions of multi-domain proteins, especially providing new insights into targets with multiple conformational states.

Recent Progress of Protein Tertiary Structure Prediction

The prediction of three-dimensional (3D) protein structure from amino acid sequences has stood as a significant challenge in computational and structural bioinformatics for decades. Recently, the widespread integration of artificial intelligence (AI) algorithms has substantially expedited advancements in protein structure prediction, yielding numerous significant milestones. In particular, the end-to-end deep learning method AlphaFold2 has facilitated the rise of structure prediction performance to new heights, regularly competitive with experimental structures in the 14th Critical Assessment of Protein Structure Prediction (CASP14). To provide a comprehensive understanding and guide future research in the field of protein structure prediction for researchers, this review describes various methodologies, assessments, and databases in protein structure prediction, including traditionally used protein structure prediction methods, such as template-based modeling (TBM) and template-free modeling (FM) approaches; recently developed deep learning-based methods, such as contact/distance-guided methods, end-to-end folding methods, and protein language model (PLM)-based methods; multi-domain protein structure prediction methods; the CASP experiments and related assessments; and the recently released AlphaFold Protein Structure Database (AlphaFold DB). We discuss their advantages, disadvantages, and application scopes, aiming to provide researchers with insights through which to understand the limitations, contexts, and effective selections of protein structure prediction methods in protein-related fields.

Recent Advances and Challenges in Protein Structure Prediction

Artificial intelligence has made significant advances in the field of protein structure prediction in recent years. In particular, DeepMind's end-to-end model, AlphaFold2, has demonstrated the capability to predict three-dimensional structures of numerous unknown proteins with accuracy levels comparable to those of experimental methods. This breakthrough has opened up new possibilities for understanding protein structure and function as well as accelerating drug discovery and other applications in the field of biology and medicine. Despite the remarkable achievements of artificial intelligence in the field, there are still some challenges and limitations. In this Review, we discuss the recent progress and some of the challenges in protein structure prediction. These challenges include predicting multidomain protein structures, protein complex structures, multiple conformational states of proteins, and protein folding pathways. Furthermore, we highlight directions in which further improvements can be conducted.

Multi-domain and complex protein structure prediction using inter-domain interactions from deep learning

Accurately capturing domain-domain interactions is key to understanding protein function and designing structure-based drugs. Although AlphaFold2 has made a breakthrough on single domain, it should be noted that the structure modeling for multi-domain protein and complex remains a challenge. In this study, we developed a multi-domain and complex structure assembly protocol, named DeepAssembly, based on domain segmentation and single domain modeling algorithms. Firstly, DeepAssembly uses a population-based evolutionary algorithm to assemble multi-domain proteins by inter-domain interactions inferred from a developed deep learning network. Secondly, protein complexes are assembled by means of domains rather than chains using DeepAssembly. Experimental results show that on 219 multi-domain proteins, the average inter-domain distance precision by DeepAssembly is 22.7% higher than that of AlphaFold2. Moreover, DeepAssembly improves accuracy by 13.1% for 164 multi-domain structures with low confidence deposited in AlphaFold database. We apply DeepAssembly for the prediction of 247 heterodimers. We find that DeepAssembly successfully predicts the interface (DockQ ≥ 0.23) for 32.4% of the dimers, suggesting a lighter way to assemble complex structures by treating domains as assembly units and using inter-domain interactions learned from monomer structures.

E2EDA: Protein Domain Assembly Based on End-to-End Deep Learning

With the development of deep learning, almost all single-domain proteins can be predicted at experimental resolution. However, the structure prediction of multi-domain proteins remains a challenge. Achieving end-to-end protein domain assembly and further improving the accuracy of the full-chain modeling by accurately predicting inter-domain orientation while improving the assembly efficiency will provide significant insights into structure-based drug discovery. In this work, we propose an End-to-End Domain Assembly method based on deep learning, named E2EDA. We first develop RMNet, an EfficientNetV2-based deep learning model that fuses multiple features using an attention mechanism to predict inter-domain rigid motion. Then, the predicted rigid motions are transformed into inter-domain spatial transformations to directly assemble the full-chain model. Finally, the scoring strategy RMscore is designed to select the best model from multiple assembled models. The experimental results show that the average TM-score of the model assembled by E2EDA on the benchmark set (282) is 0.827, which is better than those of other domain assembly methods SADA (0.792) and DEMO (0.730). Meanwhile, on our constructed multi-domain data set from AlphaFold DB, the model reassembled by E2EDA is 7.0% higher in TM-score compared to the full-chain model predicted by AlphaFold2, indicating that E2EDA can capture more accurate inter-domain orientations to improve the quality of the model predicted by AlphaFold2. Furthermore, compared to SADA and AlphaFold2, E2EDA reduced the average runtime on the benchmark by 64.7% and 19.2%, respectively, indicating that E2EDA can significantly improve assembly efficiency through an end-to-end approach. The online server is available at http://zhanglab-bioinf.com/E2EDA.

DomBpred: Protein Domain Boundary Prediction Based on Domain-Residue Clustering Using Inter-Residue Distance

Domain boundary prediction is one of the most important problems in the study of protein structure and function, especially for large proteins. At present, most domain boundary prediction methods have low accuracy and limitations in dealing with multi-domain proteins. In this study, we develop a sequence-based protein domain boundary prediction, named DomBpred. In DomBpred, the input sequence is first classified as either a single-domain protein or a multi-domain protein through a designed effective sequence metric based on a constructed single-domain sequence library. For the multi-domain protein, a domain-residue clustering algorithm inspired by Ising model is proposed to cluster the spatially close residues according inter-residue distance. The unclassified residues and the residues at the edge of the cluster are then tuned by the secondary structure to form potential cut points. Finally, a domain boundary scoring function is proposed to recursively evaluate the potential cut points to generate the domain boundary. DomBpred is tested on a large-scale test set of FUpred comprising 2549 proteins. Experimental results show that DomBpred better performs than the state-of-the-art methods in classifying whether protein sequences are composed by single or multiple domains, and the Matthew's correlation coefficient is 0.882. Moreover, on 849 multi-domain proteins, the domain boundary distance and normalised domain overlap scores of DomBpred are 0.523 and 0.824, respectively, which are 5.0% and 4.2% higher than those of the best comparison method, respectively. Comparison with other methods on the given test set shows that DomBpred outperforms most state-of-the-art sequence-based methods and even achieves better results than the top-level template-based method. The executable program is freely available at https://github.com/iobio-zjut/DomBpred and the online server at http://zhanglab-bioinf.com/DomBpred/.

Fusion of two unrelated protein domains in a chimera protein and its 3D prediction: Justification of the x-ray reference structures as a prediction benchmark

Proteins are naturally formed by domains edging their functional and structural properties. A domain out of the context of an entire protein can retain its structure and to some extent also function on its own. These properties rationalize construction of artificial fusion multidomain proteins with unique combination of various functions. Information on the specific functional and structural characteristics of individual domains in the context of new artificial fusion proteins is inevitably encoded in sequential order of composing domains defining their mutual spatial positions. So the challenges in designing new proteins with new domain combinations lie dominantly in structure/function prediction and its context dependency. Despite the enormous body of publications on artificial fusion proteins, the task of their structure/function prediction is complex and nontrivial. The degree of spatial freedom facilitated by a linker between domains and their mutual orientation driven by noncovalent interactions is beyond a simple and straightforward methodology to predict their structure with reasonable accuracy. In the presented manuscript, we tested methodology using available modeling tools and computational methods. We show that the process and methodology of such prediction are not straightforward and must be done with care even when recently introduced AlphaFold II is used. We also addressed a question of benchmarking standards for prediction of multidomain protein structures-x-ray or Nuclear Magnetic Resonance experiments. On the study of six two-domain protein chimeras as well as their composing domains and their x-ray structures selected from PDB, we conclude that the major obstacle for justified prediction is inappropriate sampling of the conformational space by the explored methods. On the other hands, we can still address particular steps of the methodology and improve the process of chimera proteins prediction.

Structural analogue-based protein structure domain assembly assisted by deep learning

MOTIVATION

With the breakthrough of AlphaFold2, the protein structure prediction problem has made remarkable progress through deep learning end-to-end techniques, in which correct folds could be built for nearly all single-domain proteins. However, the full-chain modelling appears to be lower on average accuracy than that for the constituent domains and requires higher demand on computing hardware, indicating the performance of full-chain modelling still needs to be improved. In this study, we investigate whether the predicted accuracy of the full-chain model can be further improved by domain assembly assisted by deep learning.

RESULTS

In this article, we developed a structural analogue-based protein structure domain assembly method assisted by deep learning, named SADA. In SADA, a multi-domain protein structure database was constructed for the full-chain analogue detection using individual domain models. Starting from the initial model constructed from the analogue, the domain assembly simulation was performed to generate the full-chain model through a two-stage differential evolution algorithm guided by the energy function with an inter-residue distance potential predicted by deep learning. SADA was compared with the state-of-the-art domain assembly methods on 356 benchmark proteins, and the average TM-score of SADA models is 8.1% and 27.0% higher than that of DEMO and AIDA, respectively. We also assembled 293 human multi-domain proteins, where the average TM-score of the full-chain model after the assembly by SADA is 1.1% higher than that of the model by AlphaFold2. To conclude, we find that the domains often interact in the similar way in the quaternary orientations if the domains have similar tertiary structures. Furthermore, homologous templates and structural analogues are complementary for multi-domain protein full-chain modelling.

AVAILABILITY AND IMPLEMENTATION

http://zhanglab-bioinf.com/SADA.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

I-TASSER-MTD: a deep-learning-based platform for multi-domain protein structure and function prediction

Most proteins in cells are composed of multiple folding units (or domains) to perform complex functions in a cooperative manner. Relative to the rapid progress in single-domain structure prediction, there are few effective tools available for multi-domain protein structure assembly, mainly due to the complexity of modeling multi-domain proteins, which involves higher degrees of freedom in domain-orientation space and various levels of continuous and discontinuous domain assembly and linker refinement. To meet the challenge and the high demand of the community, we developed I-TASSER-MTD to model the structures and functions of multi-domain proteins through a progressive protocol that combines sequence-based domain parsing, single-domain structure folding, inter-domain structure assembly and structure-based function annotation in a fully automated pipeline. Advanced deep-learning models have been incorporated into each of the steps to enhance both the domain modeling and inter-domain assembly accuracy. The protocol allows for the incorporation of experimental cross-linking data and cryo-electron microscopy density maps to guide the multi-domain structure assembly simulations. I-TASSER-MTD is built on I-TASSER but substantially extends its ability and accuracy in modeling large multi-domain protein structures and provides meaningful functional insights for the targets at both the domain- and full-chain levels from the amino acid sequence alone.

DEMO2: Assemble multi-domain protein structures by coupling analogous template alignments with deep-learning inter-domain restraint prediction

Most proteins in nature contain multiple folding units (or domains). The revolutionary success of AlphaFold2 in single-domain structure prediction showed potential to extend deep-learning techniques for multi-domain structure modeling. This work presents a significantly improved method, DEMO2, which integrates analogous template structural alignments with deep-learning techniques for high-accuracy domain structure assembly. Starting from individual domain models, inter-domain spatial restraints are first predicted with deep residual convolutional networks, where full-length structure models are assembled using L-BFGS simulations under the guidance of a hybrid energy function combining deep-learning restraints and analogous multi-domain template alignments searched from the PDB. The output of DEMO2 contains deep-learning inter-domain restraints, top-ranked multi-domain structure templates, and up to five full-length structure models. DEMO2 was tested on a large-scale benchmark and the blind CASP14 experiment, where DEMO2 was shown to significantly outperform its predecessor and the state-of-the-art protein structure prediction methods. By integrating with new deep-learning techniques, DEMO2 should help fill the rapidly increasing gap between the improved ability of tertiary structure determination and the high demand for the high-quality multi-domain protein structures. The DEMO2 server is available at https://zhanggroup.org/DEMO/.

Characterizing conformational ensembles of multi-domain proteins using anisotropic paramagnetic NMR restraints

It has been over two decades since paramagnetic NMR started to form part of the essential techniques for structural analysis of proteins under physiological conditions. Paramagnetic NMR has significantly expanded our understanding of the inherent flexibility of proteins, in particular, those that are formed by combinations of two or more domains. Here, we present a brief overview of techniques to characterize conformational ensembles of such multi-domain proteins using paramagnetic NMR restraints produced through anisotropic metals, with a focus on the basics of anisotropic paramagnetic effects, the general procedures of conformational ensemble reconstruction, and some representative reweighting approaches.

The structural coverage of the human proteome before and after AlphaFold

The protein structure field is experiencing a revolution. From the increased throughput of techniques to determine experimental structures, to developments such as cryo-EM that allow us to find the structures of large protein complexes or, more recently, the development of artificial intelligence tools, such as AlphaFold, that can predict with high accuracy the folding of proteins for which the availability of homology templates is limited. Here we quantify the effect of the recently released AlphaFold database of protein structural models in our knowledge on human proteins. Our results indicate that our current baseline for structural coverage of 48%, considering experimentally-derived or template-based homology models, elevates up to 76% when including AlphaFold predictions. At the same time the fraction of dark proteome is reduced from 26% to just 10% when AlphaFold models are considered. Furthermore, although the coverage of disease-associated genes and mutations was near complete before AlphaFold release (69% of Clinvar pathogenic mutations and 88% of oncogenic mutations), AlphaFold models still provide an additional coverage of 3% to 13% of these critically important sets of biomedical genes and mutations. Finally, we show how the contribution of AlphaFold models to the structural coverage of non-human organisms, including important pathogenic bacteria, is significantly larger than that of the human proteome. Overall, our results show that the sequence-structure gap of human proteins has almost disappeared, an outstanding success of direct consequences for the knowledge on the human genome and the derived medical applications.

Protein structures are key to understand many biological phenomena at the molecular scale: from the effects of genetic variation to how different proteins interact with each other to create molecular pathways that, together, have a biological function. Obtaining experimental structures, however, is extremely consuming in terms of both, time and resources.

For this and other reasons, scientists have long worked to develop computational approaches that predict the structure of a protein using only its sequence as input. Recently, a group of scientists at Deepmind have developed AlphaFold2, a computational tool that is extremely accurate at this task. Moreover, they have used this tool to predict the structures of all human proteins.

In this manuscript we provide an overview of the structural coverage of the human proteome before AlphaFold models were released and how much we have gained thanks to these models. We also show how the gain affects our understanding of human pathogenic variants, both germline and somatic. Finally, we provide evidence suggesting that the gain in non-human organisms is larger than for the human proteome, particularly in the case of bacteria.

Reconstruction of Coupled Intra- and Interdomain Protein Motion from Nuclear and Electron Magnetic Resonance

Proteins composed of multiple domains allow for structural heterogeneity and interdomain dynamics that may be vital for function. Intradomain structures and dynamics can influence interdomain conformations and vice versa. However, no established structure determination method is currently available that can probe the coupling of these motions. The protein Pin1 contains separate regulatory and catalytic domains that sample "extended" and "compact" states, and ligand binding changes this equilibrium. Ligand binding and interdomain distance have been shown to impact the activity of Pin1, suggesting interdomain allostery. In order to characterize the conformational equilibrium of Pin1, we describe a novel method to model the coupling between intra- and interdomain dynamics at atomic resolution using multistate ensembles. The method uses time-averaged nuclear magnetic resonance (NMR) restraints and double electron-electron resonance (DEER) data that resolve distance distributions. While the intradomain calculation is primarily driven by exact nuclear Overhauser enhancements (eNOEs), J couplings, and residual dipolar couplings (RDCs), the relative domain distribution is driven by paramagnetic relaxation enhancement (PREs), RDCs, interdomain NOEs, and DEER. Our data support a 70:30 population of the compact and extended states in apo Pin1. A multistate ensemble describes these conformations simultaneously, with distinct conformational differences located in the interdomain interface stabilizing the compact or extended states. We also describe correlated conformations between the catalytic site and interdomain interface that may explain allostery driven by interdomain contact.

The order of PDZ3 and TrpCage in fusion chimeras determines their properties-a biophysical characterization

Most of the structural proteins known today are composed of domains that carry their own functions while keeping their structural properties. It is supposed that such domains, when taken out of the context of the whole protein, can retain their original structure and function to a certain extent. Information on the specific functional and structural characteristics of individual domains in a new context of artificial fusion proteins may help to reveal the rules of internal and external domain communication. Moreover, this could also help explain the mechanism of such communication and address how the mutual allosteric effect plays a role in a such multi-domain protein system. The simple model system of the two-domain fusion protein investigated in this work consisted of a well-folded PDZ3 domain and an artificially designed small protein domain called Tryptophan Cage (TrpCage). Two fusion proteins with swapped domain order were designed to study their structural and functional features as well as their biophysical properties. The proteins composed of PDZ3 and TrpCage, both identical in amino acid sequence but different in composition (PDZ3-TrpCage, TrpCage-PDZ3), were studied using circualr dichroism (CD) spectrometry, analytical ultracentrifugation, and molecular dynamic simulations. The biophysical analysis uncovered different structural and denaturation properties of both studied proteins, revealing their different unfolding pathways and dynamics.

A novel predicted ADP-ribosyltransferase-like family conserved in eukaryotic evolution

The presence of many completely uncharacterized proteins, even in well-studied organisms such as humans, seriously hampers full understanding of the functioning of the living cells. ADP-ribosylation is a common post-translational modification of proteins; also nucleic acids and small molecules can be modified by the covalent attachment of ADP-ribose. This modification, important in cellular signalling and infection processes, is usually executed by enzymes from the large superfamily of ADP-ribosyltransferases (ARTs). Here, using bioinformatics approaches, we identify a novel putative ADP-ribosyltransferase family, conserved in eukaryotic evolution, with a divergent active site. The hallmark of these proteins is the ART domain nestled between flanking leucine-rich repeat (LRR) domains. LRRs are typically involved in innate immune surveillance. The novel family appears as putative novel ADP-ribosylation-related actors, most likely pseudoenzymes. Sequence divergence and lack of clearly detectable “classical” ART active site suggests the novel domains are pseudoARTs, yet atypical ART activity, or alternative enzymatic activity cannot be excluded. We propose that this family, including its human member LRRC9, may be involved in an ancient defense mechanism, with analogies to the innate immune system, and coupling pathogen detection to ADP-ribosyltransfer or other signalling mechanisms.

Revisiting Bap Multidomain Protein: More Than Sticking Bacteria Together

One of the major components of the staphylococcal biofilm is surface proteins that assemble as scaffold components of the biofilm matrix. Among the different surface proteins able to contribute to biofilm formation, this review is dedicated to the Biofilm Associated Protein (Bap). Bap is part of the accessory genome of Staphylococcus aureus but orthologs of Bap in other staphylococcal species belong to the core genome. When present, Bap promotes adhesion to abiotic surfaces and induces strong intercellular adhesion by self-assembling into amyloid like aggregates in response to the levels of calcium and the pH in the environment. During infection, Bap enhances the adhesion to epithelial cells where it binds directly to the host receptor Gp96 and inhibits the entry of the bacteria into the cells. To perform such diverse range of functions, Bap comprises several domains, and some of them include several motifs associated to distinct functions. Based on the knowledge accumulated with the Bap protein of S. aureus, this review aims to summarize the current knowledge of the structure and properties of each domain of Bap and their contribution to Bap functionality.

Homology modeling in the time of collective and artificial intelligence

Homology modeling is a method for building protein 3D structures using protein primary sequence and utilizing prior knowledge gained from structural similarities with other proteins. The homology modeling process is done in sequential steps where sequence/structure alignment is optimized, then a backbone is built and later, side-chains are added. Once the low-homology loops are modeled, the whole 3D structure is optimized and validated. In the past three decades, a few collective and collaborative initiatives allowed for continuous progress in both homology and ab initio modeling. Critical Assessment of protein Structure Prediction (CASP) is a worldwide community experiment that has historically recorded the progress in this field. Folding@Home and Rosetta@Home are examples of crowd-sourcing initiatives where the community is sharing computational resources, whereas RosettaCommons is an example of an initiative where a community is sharing a codebase for the development of computational algorithms. Foldit is another initiative where participants compete with each other in a protein folding video game to predict 3D structure. In the past few years, contact maps deep machine learning was introduced to the 3D structure prediction process, adding more information and increasing the accuracy of models significantly. In this review, we will take the reader in a journey of exploration from the beginnings to the most recent turnabouts, which have revolutionized the field of homology modeling. Moreover, we discuss the new trends emerging in this rapidly growing field.

A highly potential cleavable linker for tumor targeting antibody-chemokines

Chemokines are the large family of chemotactic cytokines that play an important role in leukocyte movement and migration stimulation. Until now, several antibody-cytokine (chemokine) fusion proteins have been investigated in clinical trials because of their ability to evoke the circulating leukocytes far from the tumor site. In this case, creating the concentration gradient regarding the chemokine is very important to recruit the circulating leukocytes with maximum performance to the tumor environment. To achieve a proper gradient, the chemokine separation from the tumor antigen-bounded antibody can be very crucial. Thus, we designed a novel linker that can be cleaved by enzymes presented around the tumor site including cathepsin B, urokinase-type plasminogen activator (uPA) and matrix metalloproteinases (MMPs). Also, it can inhibit tumor progression by competing with the native substrate of key proteases in the tumor microenvironment. The proposed linker was evaluated using some bioinformatics approaches. In silico results showed that the linker is structurally stable and could be detected and cleaved using the mentioned enzymes.Communicated by Ramaswamy H. Sarma.

The Human DNA Mismatch Repair Protein MSH3 Contains Nuclear Localization and Export Signals That Enable Nuclear-Cytosolic Shuttling in Response to Inflammation

Inactivation of DNA mismatch repair propels colorectal cancer (CRC) tumorigenesis. CRCs exhibiting elevated microsatellite alterations at selected tetranucleotide repeats (EMAST) show reduced nuclear MutS homolog 3 (MSH3) expression with surrounding inflammation and portend poor patient outcomes. MSH3 reversibly exits from the nucleus to the cytosol in response to the proinflammatory cytokine interleukin-6 (IL-6), suggesting that MSH3 may be a shuttling protein. In this study, we manipulated three putative nuclear localization (NLS1 to -3) and two potential nuclear export signals (NES1 and -2) within MSH3. We found that both NLS1 and NLS2 possess nuclear import function, with NLS1 responsible for nuclear localization within full-length MSH3. We also found that NES1 and NES2 work synergistically to maximize nuclear export, with both being required for IL-6-induced MSH3 export. We examined a 27-bp deletion (Δ27bp) within the polymorphic exon 1 that occurs frequently in human CRC cells and neighbors NLS1. With oxidative stress, MSH3 with this deletion (Δ27bp MSH3) localizes to the cytoplasm, suggesting that NLS1 function in Δ27bp MSH3 is compromised. Overall, MSH3's shuttling in response to inflammation enables accumulation in the cytoplasm; reduced nuclear MSH3 increases EMAST and DNA damage. We suggest that polymorphic sequences adjacent to NLS1 may enhance cytosolic retention, which has clinical implications for inflammation-associated neoplastic processes.

Multidomain protein structure prediction using information about residues interacting on multimeric protein interfaces

Protein functions can be predicted based on their three-dimensional structures. However, many multidomain proteins have unstable structures, making it difficult to determine the whole structure in biological experiments. Additionally, multidomain proteins are often decomposed and identified based on their domains, with the structure of each domain often found in public databases. Recent studies have advanced structure prediction methods of multidomain proteins through computational analysis. In existing methods, proteins that serve as templates are used for three-dimensional structure prediction. However, when no protein template is available, the accuracy of the prediction is decreased. This study was conducted to predict the structures of multidomain proteins without the need for whole structure templates.

We improved structure prediction methods by performing rigid-body docking from the structure of each domain and reranking a structure closer to the correct structure to have a higher value. In the proposed method, the score for the domain-domain interaction obtained without a structural template of the multidomain protein and score for the three-dimensional structure obtained during docking calculation were newly incorporated into the score function. We successfully predicted the structures of 50 of 55 multidomain proteins examined in the test dataset.

Interaction residue pair information of the protein-protein complex interface contributes to domain reorganizations even when a structural template for a multidomain protein cannot be obtained. This approach may be useful for predicting the structures of multidomain proteins with important biochemical functions.

SAXSDom: Modeling multidomain protein structures using small-angle X-ray scattering data

Many proteins are composed of several domains that pack together into a complex tertiary structure. Multidomain proteins can be challenging for protein structure modeling, particularly those for which templates can be found for individual domains but not for the entire sequence. In such cases, homology modeling can generate high quality models of the domains but not for the orientations between domains. Small-angle X-ray scattering (SAXS) reports the structural properties of entire proteins and has the potential for guiding homology modeling of multidomain proteins. In this article, we describe a novel multidomain protein assembly modeling method, SAXSDom that integrates experimental knowledge from SAXS with probabilistic Input-Output Hidden Markov model to assemble the structures of individual domains together. Four SAXS-based scoring functions were developed and tested, and the method was evaluated on multidomain proteins from two public datasets. Incorporation of SAXS information improved the accuracy of domain assembly for 40 out of 46 critical assessment of protein structure prediction multidomain protein targets and 45 out of 73 multidomain protein targets from the ab initio domain assembly dataset. The results demonstrate that SAXS data can provide useful information to improve the accuracy of domain-domain assembly. The source code and tool packages are available at https://github.com/jianlin-cheng/SAXSDom.

A Legionella effector kinase is activated by host inositol hexakisphosphate

The transfer of a phosphate from ATP to a protein substrate, a modification known as protein phosphorylation, is catalyzed by protein kinases. Protein kinases play a crucial role in virtually every cellular activity. Recent studies of atypical protein kinases have highlighted the structural similarity of the kinase superfamily despite notable differences in primary amino acid sequence. Here, using a bioinformatics screen, we searched for putative protein kinases in the intracellular bacterial pathogen Legionella pneumophila and identified the type 4 secretion system effector Lpg2603 as a remote member of the protein kinase superfamily. Employing an array of biochemical and structural biology approaches, including in vitro kinase assays and isothermal titration calorimetry, we show that Lpg2603 is an active protein kinase with several atypical structural features. Importantly, we found that the eukaryote-specific host signaling molecule inositol hexakisphosphate (IP6) is required for Lpg2603 kinase activity. Crystal structures of Lpg2603 in the apo-form and when bound to IP6 revealed an active-site rearrangement that allows for ATP binding and catalysis. Our results on the structure and activity of Lpg2603 reveal a unique mode of regulation of a protein kinase, provide the first example of a bacterial kinase that requires IP6 for its activation, and may aid future work on the function of this effector during Legionella pathogenesis.

Enhanced GROMACS: toward a better numerical simulation framework

The GROMACS software package represented a promising direction toward the molecular dynamic simulation and there is ongoing interest to extend it. In this study, we introduce a new component into the conventional package with the goal being to facilitate the process of finding the native structure of proteins with minimal free-energy value. We achieved this through incorporating a wide range of metaheuristic optimization algorithms and force fields, leading up to the EGROMACS molecular simulation toolkit. Compared with other programs, the EGROMACS supports all standard force fields as well as new minimization algorithms and Hybrid MPI/OpenMP parallelization. We applied the proposed EGROMACS framework to minimize the structure of several target sequences. The obtained results showed comparative performance of the introduced framework to current well-known molecular simulation algorithms. This extension to the GROMACS, however, uses metaheuristic algorithms to address the problem.

Assembling multidomain protein structures through analogous global structural alignments

Most proteins exist with multiple domains in cells for cooperative functionality. However, structural biology and protein folding methods are often optimized for single-domain structures, resulting in a rapidly growing gap between the improved capability for tertiary structure determination and high demand for multidomain structure models. We have developed a pipeline, termed DEMO, for constructing multidomain protein structures by docking-based domain assembly simulations, with interdomain orientations determined by the distance profiles from analogous templates as detected through domain-level structure alignments. The pipeline was tested on a comprehensive benchmark set of 356 proteins consisting of 2-7 continuous and discontinuous domains, for which DEMO generated models with correct global fold (TM-score > 0.5) for 86% of cases with continuous domains and for 100% of cases with discontinuous domain structures, starting from randomly oriented target-domain structures. DEMO was also applied to reassemble multidomain targets in the CASP12 and CASP13 experiments using domain structures excised from the top server predictions, where the full-length DEMO models showed a significantly improved quality over the original server models. Finally, sparse restraints of mass spectrometry-generated cross-linking data and cryo-EM density maps are incorporated into DEMO, resulting in improvements in the average TM-score by 6.3% and 12.5%, respectively. The results demonstrate an efficient approach to assembling multidomain structures, which can be easily used for automated, genome-scale multidomain protein structure assembly.

Modeling the Tertiary Structure of the Rift Valley Fever Virus L Protein

A tertiary structure governs, to a great extent, the biological activity of a protein in the living cell and is consequently a central focus of numerous studies aiming to shed light on cellular processes central to human health. Here, we aim to elucidate the structure of the Rift Valley fever virus (RVFV) L protein using a combination of in silico techniques. Due to its large size and multiple domains, elucidation of the tertiary structure of the L protein has so far challenged both dry and wet laboratories. In this work, we leverage complementary perspectives and tools from the computational-molecular-biology and bioinformatics domains for constructing, refining, and evaluating several atomistic structural models of the L protein that are physically realistic. All computed models have very flexible termini of about 200 amino acids each, and a high proportion of helical regions. Properties such as potential energy, radius of gyration, hydrodynamics radius, flexibility coefficient, and solvent-accessible surface are reported. Structural characterization of the L protein enables our laboratories to better understand viral replication and transcription via further studies of L protein-mediated protein-protein interactions. While results presented a focus on the RVFV L protein, the following workflow is a more general modeling protocol for discovering the tertiary structure of multidomain proteins consisting of thousands of amino acids.

Application of molecular dynamics simulations to design a dual-purpose oligopeptide linker sequence for fusion proteins

Proteins are often monitored by combining a fluorescent polypeptide tag with the target protein. However, due to the high molecular weight and immunogenicity of such tags, they are not suitable choices for combining with fusion proteins such as immunotoxins. In this study, we designed a polypeptide sequence with a dual role (it acts as both a linker and a fluorescent probe) to use with fusion proteins. Two common fluorescent tag sequences based on tetracysteine were compared to a commonly used rigid linker as well as our proposed dual-purpose sequence. Computational investigations showed that the dual-purpose sequence was structurally stable and may be a good choice to use as both a linker and a fluorescence marker between two moieties in a fusion protein.

Identifying weak interdomain interactions that stabilize the supertertiary structure of the N-terminal tandem PDZ domains of PSD-95

Previous studies of the N-terminal PDZ tandem from PSD-95 produced divergent models and failed to identify interdomain contacts stabilizing the structure. We used ensemble and single-molecule FRET along with replica-exchange molecular dynamics to fully characterize the energy landscape. Simulations and experiments identified two conformations: an open-like conformation with a small contact interface stabilized by salt bridges, and a closed-like conformation with a larger contact interface stabilized by surface-exposed hydrophobic residues. Both interfaces were confirmed experimentally. Proximity of interdomain contacts to the binding pockets may explain the observed coupling between conformation and binding. The low-energy barrier between conformations allows submillisecond dynamics, which were time-averaged in previous NMR and FRET studies. Moreover, the small contact interfaces were likely overridden by lattice contacts as crystal structures were rarely sampled in simulations. Our hybrid approach can identify transient interdomain interactions, which are abundant in multidomain proteins yet often obscured by dynamic averaging.

Recent advances in sequence-based protein structure prediction

The most accurate characterizations of the structure of proteins are provided by structural biology experiments. However, because of the high cost and labor-intensive nature of the structural experiments, the gap between the number of protein sequences and solved structures is widening rapidly. Development of computational methods to accurately model protein structures from sequences is becoming increasingly important to the biological community. In this article, we highlight some important progress in the field of protein structure prediction, especially those related to free modeling (FM) methods that generate structure models without using homologous templates. We also provide a short synopsis of some of the recent advances in FM approaches as demonstrated in the recent Computational Assessment of Structure Prediction competition as well as recent trends and outlook for FM approaches in protein structure prediction.

Missense mutation in SLIT2 associated with congenital myopia, anisometropia, connective tissue abnormalities, and obesity

Background

SLIT2 is a protein ligand for the Roundabout (ROBO) receptor and was found to play a major role in repulsive midline axon guidance in central nervous system development. Based on studies utilizing knockout models, it has been postulated that SLIT2 is important for preventing inappropriate axonal routing during mammalian optic chiasm development.

Methods

Case report.

Results

Here, we report a case of congenital myopia, anisometropia, and obesity in a patient with a SLIT2 point mutation. Examination of the patient’s skin biopsy revealed abnormalities in elastin and collagen fibrils that suggest an underlying connective tissue disorder. Structural modeling placed the novel mutation (p.D1407G) in the EGF-like domain 8 and was predicted to affect interactions with SLIT2 binding partners.

Conclusions

To the authors’ knowledge, this is the first report of a SLIT2 variant in the context of these ocular findings.

Electronic supplementary material

The online version of this article (10.1186/s13023-018-0885-4) contains supplementary material, which is available to authorized users.

A novel Vip3Aa16-Cry1Ac chimera toxin: Enhancement of toxicity against Ephestia kuehniella, structural study and molecular docking

Bacillus thuringiensis Vip3A protein has been widely used for crop protection and for delay resistance to existing insecticidal Cry toxins. During current study, a fusion between vip3Aa16 and the toxic core sequence of cry1Ac was constructed in pHT Blue plasmid. Vip3Aa16-Cry1Ac protein was expressed in the supernatant of B. thuringiensis with a size of about 150 kDa. Bioassays tested on Ephestia kuehniella showed that the use of the chimera toxin as biopesticide improved the toxicity to reach 90% ± 2 with an enhancement of 20% compared to the single Vip3Aa16 protein. The findings indicated that the fusion protein design opens new ways to enhance Vip3A toxicity against lepidopteran species and could avoiding insect tolerance of B. thuringiensis delta-endotoxins. Through computational study, we have predicted for the first time the whole 3D structure of a Vip3A toxin. We showed that Vip3Aa16 structure is composed by three domains like Cry toxins: an N-terminal domain containing hemolysin like fold as well as two others Carbohydrate Binding Module (CBM)-like domains. Molecular docking analysis of the chimera toxin and the single Vip3Aa16 protein against specific insect receptors revealed that residues of CBM like domains are clearly involved in the binding of the toxin to receptors.

Anti-ROR1 scFv-EndoG as a Novel Anti-Cancer Therapeutic Drug

Aim: Immunotoxins are proteins that consist of an antibody fragment linked to a toxin, used as agents for targeted therapy of cancers. Although the most potent immunotoxins are made from bacterial and plant toxins, obstacles which contribute to poor responses are immunogenicity in patients and rapid development of neutralizing antibodies. In the present study we proposed a new therapeutic immunotoxin for targeted cancer therapy of ROR1 expressing cancers: an anti ROR1 single chain fragment variable antibody (scFv)-endonuclease G (anti ROR1 scFv-EndoG). Methods: The three-dimensional structure of anti ROR1 scFv-EndoG protein was modeled and structure validation tools were employed to confirm the accuracy and reliability of the developed model. In addition, stability and integrity of the model were assessed by molecular dynamic (MD) simulation. Results: All results suggested the protein model to be acceptable and of good quality. Conclusions: Anti-ROR1 scFv-EndoG would be expected to bind to the ROR1 extracellular domain by its scFv portion and selectively deliver non-immunogenic human endonuclease G enzyme as an end-stage apoptosis molecule into ROR1-expressing cancer cells and lead rapidly to apoptosis. We believe that anti ROR1 and other anti-tumor antigen scFv-EndoG forms may be helpful for cancer therapy.

Structure of focal adhesion kinase in healthy heart versus pathological cardiac hypertrophy: A modeling and simulation study

Focal adhesion kinase (FAK) is required for signaling in the heart. S910 phosphorylated FAK is known to cause pathological cardiac hypertrophy. The switching of FAK between its inactive (-i), activated (-a) and hyperactive (-h) state is controlled by phosphorylation. FAK consists of three domains, namely: FERM, Kinase, and FAT joined by linkers L1 and L2. The structural basis of FAK phosphorylation and signaling to the downstream pathways is not understood. In this work, we carried out homology modeling and domain assembly of full length human iFAK and aFAK. 100 ns classical molecular dynamic simulations were performed using AMBER14 and effect of S910 phosphorylation on FAK was investigated. The iFAK model superposed on a small angel X-ray scattering (SAXS) derived model with RMSD of 1.18 Å for 590 Cα atoms. aFAK showed S910 phosphorylation site in L2 shielded by FERM. S910 phosphorylation in hFAK led to its exposure accompanied by a large conformational change and exposing the previously buried Grb2 interaction site responsible for causing cardiac hypertrophy. The models of FAK are in agreement with diverse experimental data and observed differences in biological action. Understanding the structure activity relationships of FAK in response to phosphorylation is important for its future therapeutic modulation.

A Novel Domain Assembly Routine for Creating Full-Length Models of Membrane Proteins from Known Domain Structures

Membrane proteins composed of soluble and membrane domains are often studied one domain at a time. However, to understand the biological function of entire protein systems and their interactions with each other and drugs, knowledge of full-length structures or models is required. Although few computational methods exist that could potentially be used to model full-length constructs of membrane proteins, none of these methods are perfectly suited for the problem at hand. Existing methods require an interface or knowledge of the relative orientations of the domains or are not designed for domain assembly, and none of them are developed for membrane proteins. Here we describe the first domain assembly protocol specifically designed for membrane proteins that assembles intra- and extracellular soluble domains and the transmembrane domain into models of the full-length membrane protein. Our protocol does not require an interface between the domains and samples possible domain orientations based on backbone dihedrals in the flexible linker regions, created via fragment insertion, while keeping the transmembrane domain fixed in the membrane. For five examples tested, our method mp_domain_assembly, implemented in RosettaMP, samples domain orientations close to the known structure and is best used in conjunction with experimental data to reduce the conformational search space.

Molecular and structural characteristics of multidrug resistance-associated protein 7 in Chinese liver fluke Clonorchis sinensis

Multidrug resistance-associated protein 7 (MRP7, ABCC10) is a C subfamily member of the ATP-binding cassette (ABC) superfamily. MRP7 is a lipophilic anion transporter that pumps endogenous and xenobiotic substrates from the cytoplasm to the extracellular milieu. Here, we cloned and characterized CsMRP7 as a novel ABC transporter from the Chinese liver fluke, Clonorchis sinensis. Full-length cDNA of CsMRP7 was 5174 nt, encoded 1636 amino acids (aa), and harbored a 147-bp 5'-untranslated region (5'-UTR) and 116-bp 3'-UTR. Phylogenetic analysis confirmed that CsMRP7 was closer to the ABCC subfamily than the ABCB subfamily. Tertiary structures of the N-terminal region (1-322 aa) and core region (323-1621 aa) of CsMRP7 were generated by homology modeling using glucagon receptor (PDB ID: 5ee7_A) and P-glycoprotein (PDB ID: 4f4c_A) as templates, respectively. CsMRP7 nucleotide-binding domain 2 (NBD2) was conserved more than NBD1, which was the sites of ATP binding and hydrolysis. Like typical long MRPs, CsMRP7 has an additional membrane-spanning domain 0 (MSD0) and cytoplasmic loop, along with a common structural fold consisting of MSD1-NBD1-MSD2-NBD2 as a single polypeptide assembly. MSD0, MSD1, and MSD2 consisted of TM1-7, TM8-13, and TM14-19, respectively. The CsMRP7 transcript was more abundant in the metacercariae than in the adult worms. Truncated NBD1 (39 kDa) and NBD2 (44 kDa) were produced in bacteria and mouse immune sera were raised. CsMRP7 was localized in the apical side of the intestinal epithelium, sperm in the testes and seminal receptacle, receptacle membrane, and mesenchymal tissue around intestine in the adult worm. These results provide molecular information and insights into structural and functional characteristics of CsMRP7 and homologs of flukes.

Towards novel Cry toxins with enhanced toxicity/broader: a new chimeric Cry4Ba / Cry1Ac toxin

Attempts have been made to express or to merge different Cry proteins in order to enhance toxic effects against various insects. Cry1A proteins of Bacillus thuringiensis form a typical bipyramidal parasporal crystal and their protoxins contain a highly conserved C-terminal region. A chimerical gene, called cry(4Ba-1Ac), formed by a fusion of the N-terminus part of cry4Ba and the C-terminus part of cry1Ac, was constructed. Its transformation to an acrystalliferous B. thuringiensis strain showed that it was expressed as a chimerical protein of 116 kDa, assembled in spherical to amorphous parasporal crystals. The chimerical gene cry(4Ba-1Ac) was introduced in a B. thuringiensis kurstaki strain. In the generated crystals of the recombinant strain, the presence of Cry(4Ba-1Ac) was evidenced by MALDI-TOF. The recombinant strain showed an important increase of the toxicity against Culex pipiens larvae (LC₅₀ = 0.84 mg l^-1 ± 0.08) compared to the wild type strain through the synergistic activity of Cry2Aa with Cry(4Ba-1Ac). The enhancement of toxicity of B. thuringiensis kurstaki expressing Cry(4Ba-1Ac) compared to that expressing the native toxin Cry4Ba, might be related to its a typical crystallization properties. The developed fusion protein could serve as a potent toxin against different pests of mosquitoes and major crop plants.

Mutation Drivers of Immunological Responses to Cancer

In cancer immunology, somatic missense mutations have been mostly studied with regard to their role in the generation of neoantigens. However, growing evidence suggests that mutations in certain genes, such as CASP8 or TP53, influence the immune response against a tumor by other mechanisms. Identifying these genes and mechanisms is important because, just as the identification of cancer driver genes led to the development of personalized cancer therapies, a comprehensive catalog of such cancer immunity drivers will aid in the development of therapies aimed at restoring antitumor immunity. Here, we present an algorithm, domainXplorer, that can be used to identify potential cancer immunity drivers. To demonstrate its potential, we used it to analyze a dataset of 5,164 tumor samples from The Cancer Genome Atlas (TCGA) and to identify protein domains in which mutation status correlates with the presence of immune cells in cancer tissue (immune infiltrate). We identified 122 such protein regions, including several that belong to proteins with known roles in immune response, such as C2, CD163L1, or FCγR2A. In several cases, we show that mutations within the same protein can be associated with more or less immune cell infiltration, depending on the specific domain mutated. These results expand the catalog of potential cancer immunity drivers and highlight the importance of taking into account the structural context of somatic mutations when analyzing their potential association with immune phenotypes. Cancer Immunol Res; 4(9); 789-98. ©2016 AACR.

AIDA: ab initio domain assembly server

AIDA: ab initio domain assembly server, available at http://ffas.burnham.org/AIDA/ is a tool that can identify domains in multi-domain proteins and then predict their 3D structures and relative spatial arrangements. The server is free and open to all users, and there is an option for a user to provide an e-mail to get the link to result page. Domains are evolutionary conserved and often functionally independent units in proteins. Most proteins, especially eukaryotic ones, consist of multiple domains while at the same time, most experimentally determined protein structures contain only one or two domains. As a result, often structures of individual domains in multi-domain proteins can be accurately predicted, but the mutual arrangement of different domains remains unknown. To address this issue we have developed AIDA program, which combines steps of identifying individual domains, predicting (separately) their structures and assembling them into multiple domain complexes using an ab initio folding potential to describe domain-domain interactions. AIDA server not only supports the assembly of a large number of continuous domains, but also allows the assembly of domains inserted into other domains. Users can also provide distance restraints to guide the AIDA energy minimization.