A generalized approach to sampling backbone conformations with RosettaDock for CAPRI rounds 13-19

Bilirubin is a new ligand for nuclear receptor Liver Receptor Homolog-1

The nuclear receptor Liver Receptor Homolog-1 (LRH-1, NR5A2 ) binds to phospholipids that regulate important LRH-1 functions in the liver. A recent compound screen unexpectedly identified bilirubin, the product of liver heme metabolism, as a possible ligand for LRH-1. Here, we show unconjugated bilirubin directly binds LRH-1 with apparent K d =9.3uM, altering LRH-1 interaction with all transcriptional coregulator peptides tested. Bilirubin decreased LRH-1 protease sensitivity, consistent with MD simulations predicting bilirubin stably binds LRH-1 within the canonical ligand binding site. Bilirubin activated a luciferase reporter specific for LRH-1, dependent on co-expression with the bilirubin membrane transporter SLCO1B1 , but bilirubin failed to activate ligand-binding genetic mutants of LRH-1. Gene profiling in HepG2 cells shows bilirubin selectively regulated transcripts from endogenous LRH-1 ChIP-seq target genes, which was significantly attenuated by either genetic knockdown of LRH-1, or by a specific chemical competitor of LRH-1. Gene set enrichment suggests bilirubin and LRH-1 share roles in cholesterol metabolism and lipid efflux, thus we propose a new role for LRH-1 in directly sensing intracellular levels of bilirubin.

Harnessing the power of artificial intelligence to advance cell therapy

Cell therapies are powerful technologies in which human cells are reprogrammed for therapeutic applications such as killing cancer cells or replacing defective cells. The technologies underlying cell therapies are increasing in effectiveness and complexity, making rational engineering of cell therapies more difficult. Creating the next generation of cell therapies will require improved experimental approaches and predictive models. Artificial intelligence (AI) and machine learning (ML) methods have revolutionized several fields in biology including genome annotation, protein structure prediction, and enzyme design. In this review, we discuss the potential of combining experimental library screens and AI to build predictive models for the development of modular cell therapy technologies. Advances in DNA synthesis and high-throughput screening techniques enable the construction and screening of libraries of modular cell therapy constructs. AI and ML models trained on this screening data can accelerate the development of cell therapies by generating predictive models, design rules, and improved designs.

Elucidation of genome-wide understudied proteins targeted by PROTAC-induced degradation using interpretable machine learning

Proteolysis-targeting chimeras (PROTACs) are hetero-bifunctional molecules that induce the degradation of target proteins by recruiting an E3 ligase. PROTACs have the potential to inactivate disease-related genes that are considered undruggable by small molecules, making them a promising therapy for the treatment of incurable diseases. However, only a few hundred proteins have been experimentally tested for their amenability to PROTACs, and it remains unclear which other proteins in the entire human genome can be targeted by PROTACs. In this study, we have developed PrePROTAC, an interpretable machine learning model based on a transformer-based protein sequence descriptor and random forest classification. PrePROTAC predicts genome-wide targets that can be degraded by CRBN, one of the E3 ligases. In the benchmark studies, PrePROTAC achieved a ROC-AUC of 0.81, an average precision of 0.84, and over 40% sensitivity at a false positive rate of 0.05. When evaluated by an external test set which comprised proteins from different structural folds than those in the training set, the performance of PrePROTAC did not drop significantly, indicating its generalizability. Furthermore, we developed an embedding SHapley Additive exPlanations (eSHAP) method, which extends conventional SHAP analysis for original features to an embedding space through in silico mutagenesis. This method allowed us to identify key residues in the protein structure that play critical roles in PROTAC activity. The identified key residues were consistent with existing knowledge. Using PrePROTAC, we identified over 600 novel understudied proteins that are potentially degradable by CRBN and proposed PROTAC compounds for three novel drug targets associated with Alzheimer's disease.

Crystal structures of glycoprotein D of equine alphaherpesviruses reveal potential binding sites to the entry receptor MHC-I

Cell entry of most alphaherpesviruses is mediated by the binding of glycoprotein D (gD) to different cell surface receptors. Equine herpesvirus type 1 (EHV-1) and EHV-4 gDs interact with equine major histocompatibility complex I (MHC-I) to initiate entry into equine cells. We have characterized the gD-MHC-I interaction by solving the crystal structures of EHV-1 and EHV-4 gDs (gD1, gD4), performing protein-protein docking simulations, surface plasmon resonance (SPR) analysis, and biological assays. The structures of gD1 and gD4 revealed the existence of a common V-set immunoglobulin-like (IgV-like) core comparable to those of other gD homologs. Molecular modeling yielded plausible binding hypotheses and identified key residues (F213 and D261) that are important for virus binding. Altering the key residues resulted in impaired virus growth in cells, which highlights the important role of these residues in the gD-MHC-I interaction. Taken together, our results add to our understanding of the initial herpesvirus-cell interactions and will contribute to the targeted design of antiviral drugs and vaccine development.

Elucidation of Genome-wide Understudied Proteins targeted by PROTAC-induced degradation using Interpretable Machine Learning

Proteolysis-targeting chimeras (PROTACs) are hetero-bifunctional molecules. They induce the degradation of a target protein by recruiting an E3 ligase to the target. The PROTAC can inactivate disease-related genes that are considered as understudied, thus has a great potential to be a new type of therapy for the treatment of incurable diseases. However, only hundreds of proteins have been experimentally tested if they are amenable to the PROTACs. It remains elusive what other proteins can be targeted by the PROTAC in the entire human genome. For the first time, we have developed an interpretable machine learning model PrePROTAC, which is based on a transformer-based protein sequence descriptor and random forest classification to predict genome-wide PROTAC-induced targets degradable by CRBN, one of the E3 ligases. In the benchmark studies, PrePROTAC achieved ROC-AUC of 0.81, PR-AUC of 0.84, and over 40% sensitivity at a false positive rate of 0.05, respectively. Furthermore, we developed an embedding SHapley Additive exPlanations (eSHAP) method to identify positions in the protein structure, which play key roles in the PROTAC activity. The key residues identified were consistent with our existing knowledge. We applied PrePROTAC to identify more than 600 novel understudied proteins that are potentially degradable by CRBN, and proposed PROTAC compounds for three novel drug targets associated with Alzheimer's disease.

Author Summary

Many human diseases remain incurable because disease-causing genes cannot by selectively and effectively targeted by small molecules. Proteolysis-targeting chimera (PROTAC), an organic compound that binds to both a target and a degradation-mediating E3 ligase, has emerged as a promising approach to selectively target disease-driving genes that are not druggable by small molecules. Nevertheless, not all of proteins can be accommodated by E3 ligases, and be effectively degraded. Knowledge on the degradability of a protein will be crucial for the design of PROTACs. However, only hundreds of proteins have been experimentally tested if they are amenable to the PROTACs. It remains elusive what other proteins can be targeted by the PROTAC in the entire human genome. In this paper, we propose an intepretable machine learning model PrePROTAC that takes advantage of powerful protein language modeling. PrePROTAC achieves high accuracy when evaluated by an external dataset which comes from different gene families from the proteins in the training data, suggesting the generalizability of PrePROTAC. We apply PrePROTAC to the human genome, and identify more than 600 understudied proteins that are potentially responsive to the PROTAC. Furthermore, we design three PROTAC compounds for novel drug targets associated with Alzheimer's disease.

Emerging degrader technologies engaging lysosomal pathways

Targeted protein degradation (TPD) provides unprecedented opportunities for drug discovery. While the proteolysis-targeting chimera (PROTAC) technology has already entered clinical trials and changed the landscape of small-molecule drugs, new degrader technologies harnessing alternative degradation machineries, especially lysosomal pathways, have emerged and broadened the spectrum of degradable targets. We have recently proposed the concept of autophagy-tethering compounds (ATTECs) that hijack the autophagy protein microtubule-associated protein 1A/1B light chain 3 (LC3) for targeted degradation. Other groups also reported degrader technologies engaging lysosomal pathways through different mechanisms including AUTACs, AUTOTACs, LYTACs and MoDE-As. In this review, we analyse and discuss ATTECs along with other lysosomal-relevant degrader technologies. Finally, we will briefly summarize the current status of these degrader technologies and envision possible future studies.

Domain Expansion and Functional Diversification in Vertebrate Reproductive Proteins

The rapid evolution of fertilization proteins has generated remarkable diversity in molecular structure and function. Glycoproteins of vertebrate egg coats contain multiple zona pellucida (ZP)-N domains (1-6 copies) that facilitate multiple reproductive functions, including species-specific sperm recognition. In this report, we integrate phylogenetics and machine learning to investigate how ZP-N domains diversify in structure and function. The most C-terminal ZP-N domain of each paralog is associated with another domain type (ZP-C), which together form a "ZP module." All modular ZP-N domains are phylogenetically distinct from nonmodular or free ZP-N domains. Machine learning-based classification identifies eight residues that form a stabilizing network in modular ZP-N domains that is absent in free domains. Positive selection is identified in some free ZP-N domains. Our findings support that strong purifying selection has conserved an essential structural core in modular ZP-N domains, with the relaxation of this structural constraint allowing free N-terminal domains to functionally diversify.

Unraveling the Role of Linker Design in Proteolysis Targeting Chimeras

A current bottleneck in the development of proteolysis targeting chimeras (PROTACs) is the empirical nature of linker length structure-activity relationships (SARs). A multidisciplinary approach to alleviate the bottleneck is detailed here. First, we examine four published synthetic approaches that have been developed to increase synthetic throughput. We then discuss advances in structural biology and computational chemistry that have led to successful rational PROTAC design efforts and give promise to de novo linker design in silico. Lastly, we present a model generated from a curated list of linker SARs studies normalized to reflect how linear linker length affects the observed degradation potency (DC50).

Discovery of an orally active VHL-recruiting PROTAC that achieves robust HMGCR degradation and potent hypolipidemic activity in vivo

HMG-CoA reductase (HMGCR) protein is usually upregulated after statin (HMGCR inhibitor) treatment, which inevitably diminishes its therapeutic efficacy, provoking the need for higher doses associated with adverse effects. The proteolysis targeting chimera (PROTAC) technology has recently emerged as a powerful approach for inducing protein degradation. Nonetheless, due to their bifunctional nature, developing orally bioavailable PROTACs remains a great challenge. Herein, we identified a powerful HMGCR-targeted PROTAC (21c) comprising a VHL ligand conjugated to lovastatin acid that potently degrades HMGCR in Insig-silenced HepG2 cells (DC₅₀ = 120 nmol/L) and forms a stable ternary complex, as predicated by a holistic modeling protocol. Most importantly, oral administration of the corresponding lactone 21b reveled favorable plasma exposures referring to both the parent 21b and the conversed acid 21c. Further in vivo studies of 21b demonstrated robust HMGCR degradation and potent cholesterol reduction in mice with diet-induced hypercholesterolemia, highlighting a promising strategy for treating hyperlipidemia and associated diseases.

Targeted degradation of the enhancer lysine acetyltransferases CBP and p300

The enhancer factors CREB-binding protein (CBP) and p300 (also known as KAT3A and KAT3B) maintain gene expression programs through lysine acetylation of chromatin and transcriptional regulators and by scaffolding functions mediated by several protein-protein interaction domains. Small molecule inhibitors that target some of these domains have been developed; however, they cannot completely ablate p300/CBP function in cells. Here we describe a chemical degrader of p300/CBP, dCBP-1. Leveraging structures of ligand-bound p300/CBP domains, we use in silico modeling of ternary complex formation with the E3 ubiquitin ligase cereblon to enable degrader design. dCBP-1 is exceptionally potent at killing multiple myeloma cells and can abolish the enhancer that drives MYC oncogene expression. As an efficient degrader of this unique class of acetyltransferases, dCBP-1 is a useful tool alongside domain inhibitors for dissecting the mechanism by which these factors coordinate enhancer activity in normal and diseased cells.

Hybrid methods for combined experimental and computational determination of protein structure

Knowledge of protein structure is paramount to the understanding of biological function, developing new therapeutics, and making detailed mechanistic hypotheses. Therefore, methods to accurately elucidate three-dimensional structures of proteins are in high demand. While there are a few experimental techniques that can routinely provide high-resolution structures, such as x-ray crystallography, nuclear magnetic resonance (NMR), and cryo-EM, which have been developed to determine the structures of proteins, these techniques each have shortcomings and thus cannot be used in all cases. However, additionally, a large number of experimental techniques that provide some structural information, but not enough to assign atomic positions with high certainty have been developed. These methods offer sparse experimental data, which can also be noisy and inaccurate in some instances. In cases where it is not possible to determine the structure of a protein experimentally, computational structure prediction methods can be used as an alternative. Although computational methods can be performed without any experimental data in a large number of studies, inclusion of sparse experimental data into these prediction methods has yielded significant improvement. In this Perspective, we cover many of the successes of integrative modeling, computational modeling with experimental data, specifically for protein folding, protein-protein docking, and molecular dynamics simulations. We describe methods that incorporate sparse data from cryo-EM, NMR, mass spectrometry, electron paramagnetic resonance, small-angle x-ray scattering, Förster resonance energy transfer, and genetic sequence covariation. Finally, we highlight some of the major challenges in the field as well as possible future directions.

Enhancing fragment-based protein structure prediction by customising fragment cardinality according to local secondary structure

BACKGROUND

Whenever suitable template structures are not available, usage of fragment-based protein structure prediction becomes the only practical alternative as pure ab initio techniques require massive computational resources even for very small proteins. However, inaccuracy of their energy functions and their stochastic nature imposes generation of a large number of decoys to explore adequately the solution space, limiting their usage to small proteins. Taking advantage of the uneven complexity of the sequence-structure relationship of short fragments, we adjusted the fragment insertion process by customising the number of available fragment templates according to the expected complexity of the predicted local secondary structure. Whereas the number of fragments is kept to its default value for coil regions, important and dramatic reductions are proposed for beta sheet and alpha helical regions, respectively.

RESULTS

The evaluation of our fragment selection approach was conducted using an enhanced version of the popular Rosetta fragment-based protein structure prediction tool. It was modified so that the number of fragment candidates used in Rosetta could be adjusted based on the local secondary structure. Compared to Rosetta's standard predictions, our strategy delivered improved first models, + 24% and + 6% in terms of GDT, when using 2000 and 20,000 decoys, respectively, while reducing significantly the number of fragment candidates. Furthermore, our enhanced version of Rosetta is able to deliver with 2000 decoys a performance equivalent to that produced by standard Rosetta while using 20,000 decoys. We hypothesise that, as the fragment insertion process focuses on the most challenging regions, such as coils, fewer decoys are needed to explore satisfactorily conformation spaces.

CONCLUSIONS

Taking advantage of the high accuracy of sequence-based secondary structure predictions, we showed the value of that information to customise the number of candidates used during the fragment insertion process of fragment-based protein structure prediction. Experimentations conducted using standard Rosetta showed that, when using the recommended number of decoys, i.e. 20,000, our strategy produces better results. Alternatively, similar results can be achieved using only 2000 decoys. Consequently, we recommend the adoption of this strategy to either improve significantly model quality or reduce processing times by a factor 10.

Covalent-Fragment Screening of BRD4 Identifies a Ligandable Site Orthogonal to the Acetyl-Lysine Binding Sites

BRD4, a member of the bromodomain and extraterminal domain (BET) family, has emerged as a promising epigenetic target in cancer and inflammatory disorders. All reported BET family ligands bind within the bromodomain acetyl-lysine binding sites and competitively inhibit BET protein interaction with acetylated chromatin. Alternative chemical probes that act orthogonally to the highly conserved acetyl-lysine binding sites may exhibit selectivity within the BET family and avoid recently reported toxicity in clinical trials of BET bromodomain inhibitors. Here, we report the first identification of a ligandable site on a bromodomain outside the acetyl-lysine binding site. Inspired by our computational prediction of hotspots adjacent to nonhomologous cysteine residues within the C-terminal BRD4 bromodomain (BRD4-BD2), we performed a midthroughput mass spectrometry screen to identify cysteine-reactive fragments that covalently and selectively modify BRD4. Subsequent mass spectrometry, NMR, and computational docking analyses of electrophilic fragment hits revealed a novel ligandable site near Cys356 that is unique to BRD4 among human bromodomains. This site is orthogonal to the BRD4-BD2 acetyl-lysine binding site as Cys356 modification did not impact binding of the pan-BET bromodomain inhibitor JQ1 in fluorescence polarization assays nor an acetylated histone peptide in AlphaScreen assays. Finally, we tethered our top-performing covalent fragment to JQ1 and performed NanoBRET assays to provide proof of principle that this orthogonal site can be covalently targeted in intact human cells. Overall, we demonstrate the potential of targeting sites orthogonal to bromodomain acetyl-lysine binding sites to develop bivalent and covalent inhibitors that displace BRD4 from chromatin.

Efficient flexible backbone protein-protein docking for challenging targets

Motivation

Binding-induced conformational changes challenge current computational docking algorithms by exponentially increasing the conformational space to be explored. To restrict this search to relevant space, some computational docking algorithms exploit the inherent flexibility of the protein monomers to simulate conformational selection from pre-generated ensembles. As the ensemble size expands with increased flexibility, these methods struggle with efficiency and high false positive rates.

Results

Here, we develop and benchmark RosettaDock 4.0, which efficiently samples large conformational ensembles of flexible proteins and docks them using a novel, six-dimensional, coarse-grained score function. A strong discriminative ability allows an eight-fold higher enrichment of near-native candidate structures in the coarse-grained phase compared to RosettaDock 3.2. It adaptively samples 100 conformations each of the ligand and the receptor backbone while increasing computational time by only 20-80%. In local docking of a benchmark set of 88 proteins of varying degrees of flexibility, the expected success rate (defined as cases with ≥50% chance of achieving 3 near-native structures in the 5 top-ranked ones) for blind predictions after resampling is 77% for rigid complexes, 49% for moderately flexible complexes and 31% for highly flexible complexes. These success rates on flexible complexes are a substantial step forward from all existing methods. Additionally, for highly flexible proteins, we demonstrate that when a suitable conformer generation method exists, the method successfully docks the complex.

Availability and implementation

As a part of the Rosetta software suite, RosettaDock 4.0 is available at https://www.rosettacommons.org to all non-commercial users for free and to commercial users for a fee.

Supplementary information

Supplementary data are available at Bioinformatics online.

The Structural Basis of the Farnesylated and Methylated KRas4B Interaction with Calmodulin

Ca²⁺-calmodulin (CaM) extracts KRas4B from the plasma membrane, suggesting that KRas4B/CaM interaction plays a role in regulating Ras signaling. To gain mechanistic insight, we provide a computational model, supported by experimental structural data, of farnesylated/methylated KRas4B_1-185 interacting with CaM in solution and at anionic membranes including signaling lipids. Due to multiple interaction modes, we observe diverse conformational ensembles of the KRas4B-CaM complex. A highly populated conformation reveals the catalytic domain interacting with the N-lobe and the hypervariable region (HVR) wrapping around the linker with the farnesyl docking to the extended CaM's C-lobe pocket. Alternatively, KRas4B can interact with collapsed CaM with the farnesyl penetrating CaM's center. At anionic membranes, CaM interacts with the catalytic domain with large fluctuations, drawing the HVR. Signaling lipids establishing strong salt bridges with CaM prevent membrane departure. Membrane-interacting KRas4B-CaM complex can productively recruit phosphatidylinositol 3-kinase α (PI3Kα) to the plasma membrane, serving as a coagent in activating PI3Kα/Akt signaling.

Measuring Intrinsic Disorder and Tracking Conformational Transitions Using Rosetta ResidueDisorder

Many proteins contain regions of intrinsic disorder, not folding into unique, stable conformations. Numerous experimental methods have been developed to measure the disorder of all or select residues. In the absence of experimental data, computational methods are often utilized to identify these disordered regions and thus gain a better understanding of both structure and function. Many freely available computational methods have been developed to predict regions of intrinsic disorder from the primary sequence of a protein, including our recently developed Rosetta ResidueDisorder. While these methods are very useful, they are only designed to predict intrinsic disorder from the sequence. However, it would be useful to have a method that could also measure intrinsic disorder directly from structure. Such a method might also be used to identify changes in the structure of systems that can transition from folded to unfolded or vice versa, even systems that are not intrinsically disordered. Here we extended the capabilities of Rosetta ResidueDisorder to measure the intrinsic disorder from the coordinates of a single conformation of a protein. As a proof of principle, we show that ResidueDisorder can measure the intrinsic disorder from the coordinates with a higher accuracy (69.2%) than when predicted from sequence (65.4%) using a benchmark set of 229 proteins, showing that intrinsic disorder can be measured accurately from single structures over a large range of intrinsic disorder (0-100%). Additionally, we used ResidueDisorder to analyze unfolding trajectories of 12 fast-folding, nonintrinsically disordered proteins generated using molecular dynamics (MD), specifically steered MD (SMD), high-temperature MD, and accelerated MD (aMD) as well as long-time scale folding/unfolding trajectories. Using ResidueDisorder, a clear correlation between RMSD with respect to the native structure and measured fraction of denatured residues was observed. Finally, we introduced methods to predict folding/unfolding transitions as well as a native-like structure in the absence of a crystal structure from folding/unfolding MD trajectories. Rosetta ResidueDisorder is available as an application in the Rosetta software suite with the addition of new capabilities for directly identifying denatured regions and predicting events.

Oxidation-induced destabilization of the fibrinogen αC-domain dimer investigated by molecular dynamics simulations

Upon activation, fibrinogen is converted to insoluble fibrin, which assembles into long strings called protofibrils. These aggregate laterally to form a fibrin matrix that stabilizes a blood clot. Lateral aggregation of protofibrils is mediated by the αC domain, a partially structured fragment located in a disordered region of fibrinogen. Polymerization of αC domains links multiple fibrin molecules with each other enabling the formation of thick fibrin fibers and a fibrin matrix that is stable but can also be digested by enzymes. However, oxidizing agents produced during the inflammatory response have been shown to cause thinner fibrin fibers resulting in denser clots, which are harder to proteolyze and pose the risk of deep vein thrombosis and lung embolism. Oxidation of Met⁴⁷⁶ located within the αC domain is thought to hinder its ability to polymerize disrupting the lateral aggregation of protofibrils and leading to the observed thinner fibers. How αC domains assemble into polymers is still unclear and yet this knowledge would shed light on the mechanism through which oxidation weakens the lateral aggregation of protofibrils. This study used temperature replica exchange molecular dynamics simulations to investigate the αC-domain dimer and how this is affected by oxidation of Met⁴⁷⁶ . Analysis of the trajectories revealed that multiple stable binding modes were sampled between two αC domains while oxidation decreased the likelihood of dimer formation. Furthermore, the side chain of Met⁴⁷⁶ was observed to act as a docking spot for the binding and this function was impaired by its conversion to methionine sulfoxide.

The selective disruption of presynaptic JNK2/STX1a interaction reduces NMDA receptor-dependent glutamate release

The neuronal loss caused by excessive glutamate release, or ‘excitotoxicity’, leads to several pathological conditions, including cerebral ischemia, epilepsy, and neurodegenerative diseases. Over-stimulation of presynaptic N-methyl-D-aspartate (NMDA) receptors is known to trigger and support glutamate spillover, while postsynaptic NMDA receptors are responsible for the subsequent apoptotic cascade. Almost all molecules developed so far are unable to selectively block presynaptic or postsynaptic NMDA receptors, therefore a deeper knowledge about intracellular NMDA pathways is required to design more specific inhibitors. Our previous work showed that presynaptic c-Jun N-terminal kinase 2 (JNK2) specifically regulates NMDA-evoked glutamate release and here we demonstrate that an interaction between Syntaxin-1a and JNK2 is fundamental to this mechanism. Based on this evidence, a new cell permeable peptide (CPP), “JGRi1”, has been developed to disrupt the JNK2/STX1a interaction to indirectly, but specifically, inhibit presynaptic NMDA receptor signaling. JGRi1 reduces the NMDA-evoked release of glutamate both in in-vitro and ex-vivo experiments while also being able to widely diffuse throughout brain tissue via intraperitoneal administration. In conclusion, the JNK2/STX1 interaction is involved in presynaptic NMDA-evoked glutamate release and the novel CPP, JGRi1, acts as a pharmacological tool that promotes neuroprotection.

Computational affinity maturation of camelid single-domain intrabodies against the nonamyloid component of alpha-synuclein

Improving the affinity of protein-protein interactions is a challenging problem that is particularly important in the development of antibodies for diagnostic and clinical use. Here, we used structure-based computational methods to optimize the binding affinity of V_HNAC1, a single-domain intracellular antibody (intrabody) from the camelid family that was selected for its specific binding to the nonamyloid component (NAC) of human α-synuclein (α-syn), a natively disordered protein, implicated in the pathogenesis of Parkinson's disease (PD) and related neurological disorders. Specifically, we performed ab initio modeling that revealed several possible modes of V_HNAC1 binding to the NAC region of α-syn as well as mutations that potentially enhance the affinity between these interacting proteins. While our initial design strategy did not lead to improved affinity, it ultimately guided us towards a model that aligned more closely with experimental observations, revealing a key residue on the paratope and the participation of H4 loop residues in binding, as well as confirming the importance of electrostatic interactions. The binding activity of the best intrabody mutant, which involved just a single amino acid mutation compared to parental V_HNAC1, was significantly enhanced primarily through a large increase in association rate. Our results indicate that structure-based computational design can be used to successfully improve the affinity of antibodies against natively disordered and weakly immunogenic antigens such as α-syn, even in cases such as ours where crystal structures are unavailable.

Plasticity in binding confers selectivity in ligand-induced protein degradation

Heterobifunctional small-molecule degraders that induce protein degradation through ligase-mediated ubiquitination have shown considerable promise as a new pharmacological modality. However, we currently lack a detailed understanding of the molecular basis for target recruitment and selectivity, which is critically required to enable rational design of degraders. Here we utilize a comprehensive characterization of the ligand-dependent CRBN-BRD4 interaction to demonstrate that binding between proteins that have not evolved to interact is plastic. Multiple X-ray crystal structures show that plasticity results in several distinct low-energy binding conformations that are selectively bound by ligands. We demonstrate that computational protein-protein docking can reveal the underlying interprotein contacts and inform the design of a BRD4 selective degrader that can discriminate between highly homologous BET bromodomains. Our findings that plastic interprotein contacts confer selectivity for ligand-induced protein dimerization provide a conceptual framework for the development of heterobifunctional ligands.

Human DHEA sulfation requires direct interaction between PAPS synthase 2 and DHEA sulfotransferase SULT2A1

The high-energy sulfate donor 3′-phosphoadenosine-5′-phosphosulfate (PAPS), generated by human PAPS synthase isoforms PAPSS1 and PAPSS2, is required for all human sulfation pathways. Sulfotransferase SULT2A1 uses PAPS for sulfation of the androgen precursor dehydroepiandrosterone (DHEA), thereby reducing downstream activation of DHEA to active androgens. Human PAPSS2 mutations manifest with undetectable DHEA sulfate, androgen excess, and metabolic disease, suggesting that ubiquitous PAPSS1 cannot compensate for deficient PAPSS2 in supporting DHEA sulfation. In knockdown studies in human adrenocortical NCI-H295R1 cells, we found that PAPSS2, but not PAPSS1, is required for efficient DHEA sulfation. Specific APS kinase activity, the rate-limiting step in PAPS biosynthesis, did not differ between PAPSS1 and PAPSS2. Co-expression of cytoplasmic SULT2A1 with a cytoplasmic PAPSS2 variant supported DHEA sulfation more efficiently than co-expression with nuclear PAPSS2 or nuclear/cytosolic PAPSS1. Proximity ligation assays revealed protein–protein interactions between SULT2A1 and PAPSS2 and, to a lesser extent, PAPSS1. Molecular docking studies showed a putative binding site for SULT2A1 within the PAPSS2 APS kinase domain. Energy-dependent scoring of docking solutions identified the interaction as specific for the PAPSS2 and SULT2A1 isoforms. These findings elucidate the mechanistic basis for the selective requirement for PAPSS2 in human DHEA sulfation.

Modeling oblong proteins and water-mediated interfaces with RosettaDock in CAPRI rounds 28-35

The 28th-35th rounds of the Critical Assessment of PRotein Interactions (CAPRI) served as a practical benchmark for our RosettaDock protein-protein docking protocols, highlighting strengths and weaknesses of the approach. We achieved acceptable or better quality models in three out of 11 targets. For the two α-repeat protein-green fluorescent protein (αrep-GFP) complexes, we used a novel ellipsoidal partial-global docking method (Ellipsoidal Dock) to generate models with 2.2 Å/1.5 Å interface RMSD, capturing 49%/42% of the native contacts, for the 7-/5-repeat αrep complexes. For the DNase-immunity protein complex, we used a new predictor of hydrogen-bonding networks, HBNet with Bridging Waters, to place individual water models at the complex interface; models were generated with 1.8 Å interface RMSD and 12% native water contacts recovered. The targets for which RosettaDock failed to create an acceptable model were typically difficult in general, as six had no acceptable models submitted by any CAPRI predictor. The UCH-L5-RPN13 and UCH-L5-INO80G de-ubiquitinating enzyme-inhibitor complexes comprised inhibitors undergoing significant structural changes upon binding, with the partners being highly interwoven in the docked complexes. Our failure to predict the nucleosome-enzyme complex in Target 95 was largely due to tight constraints we placed on our model based on sparse biochemical data suggesting two specific cross-interface interactions, preventing the correct structure from being sampled. While RosettaDock's three successes show that it is a state-of-the-art docking method, the difficulties with highly flexible and multi-domain complexes highlight the need for better flexible docking and domain-assembly methods. Proteins 2017; 85:479-486. © 2016 Wiley Periodicals, Inc.

Modeling and docking of antibody structures with Rosetta

We describe Rosetta-based computational protocols for predicting the 3D structure of an antibody from sequence (RosettaAntibody) and then docking the antibody to protein antigens (SnugDock). Antibody modeling leverages canonical loop conformations to graft large segments from experimentally determined structures, as well as offering (i) energetic calculations to minimize loops, (ii) docking methodology to refine the V_L-V_H relative orientation and (iii) de novo prediction of the elusive complementarity determining region (CDR) H3 loop. To alleviate model uncertainty, antibody-antigen docking resamples CDR loop conformations and can use multiple models to represent an ensemble of conformations for the antibody, the antigen or both. These protocols can be run fully automated via the ROSIE web server (http://rosie.rosettacommons.org/) or manually on a computer with user control of individual steps. For best results, the protocol requires roughly 1,000 CPU-hours for antibody modeling and 250 CPU-hours for antibody-antigen docking. Tasks can be completed in under a day by using public supercomputers.

Predicting protein conformational changes for unbound and homology docking: learning from intrinsic and induced flexibility

Predicting protein conformational changes from unbound structures or even homology models to bound structures remains a critical challenge for protein docking. Here we present a study directly addressing the challenge by reducing the dimensionality and narrowing the range of the corresponding conformational space. The study builds on cNMA-our new framework of partner- and contact-specific normal mode analysis that exploits encounter complexes and considers both intrinsic and induced flexibility. First, we established over a CAPRI (Critical Assessment of PRedicted Interactions) target set that the direction of conformational changes from unbound structures and homology models can be reproduced to a great extent by a small set of cNMA modes. In particular, homology-to-bound interface root-mean-square deviation (iRMSD) can be reduced by 40% on average with the slowest 30 modes. Second, we developed novel and interpretable features from cNMA and used various machine learning approaches to predict the extent of conformational changes. The models learned from a set of unbound-to-bound conformational changes could predict the actual extent of iRMSD with errors around 0.6 Å for unbound proteins in a held-out benchmark subset, around 0.8 Å for unbound proteins in the CAPRI set, and around 1 Å even for homology models in the CAPRI set. Our results shed new insights into origins of conformational differences between homology models and bound structures and provide new support for the low-dimensionality of conformational adjustment during protein associations. The results also provide new tools for ensemble generation and conformational sampling in unbound and homology docking. Proteins 2017; 85:544-556. © 2016 Wiley Periodicals, Inc.

3DIANA: 3D Domain Interaction Analysis: A Toolbox for Quaternary Structure Modeling

Electron microscopy (EM) is experiencing a revolution with the advent of a new generation of Direct Electron Detectors, enabling a broad range of large and flexible structures to be resolved well below 1 nm resolution. Although EM techniques are evolving to the point of directly obtaining structural data at near-atomic resolution, for many molecules the attainable resolution might not be enough to propose high-resolution structural models. However, accessing information on atomic coordinates is a necessary step toward a deeper understanding of the molecular mechanisms that allow proteins to perform specific tasks. For that reason, methods for the integration of EM three-dimensional maps with x-ray and NMR structural data are being developed, a modeling task that is normally referred to as fitting, resulting in the so called hybrid models. In this work, we present a novel application—3DIANA—specially targeted to those cases in which the EM map resolution is medium or low and additional experimental structural information is scarce or even lacking. In this way, 3DIANA statistically evaluates proposed/potential contacts between protein domains, presents a complete catalog of both structurally resolved and predicted interacting regions involving these domains and, finally, suggests structural templates to model the interaction between them. The evaluation of the proposed interactions is computed with DIMERO, a new method that scores physical binding sites based on the topology of protein interaction networks, which has recently shown the capability to increase by 200% the number of domain-domain interactions predicted in interactomes as compared to previous approaches. The new application displays the information at a sequence and structural level and is accessible through a web browser or as a Chimera plugin at http://3diana.cnb.csic.es.

A flexible docking approach for prediction of T cell receptor-peptide-MHC complexes

T cell receptors (TCRs) are immune proteins that specifically bind to antigenic molecules, which are often foreign peptides presented by major histocompatibility complex proteins (pMHCs), playing a key role in the cellular immune response. To advance our understanding and modeling of this dynamic immunological event, we assembled a protein-protein docking benchmark consisting of 20 structures of crystallized TCR/pMHC complexes for which unbound structures exist for both TCR and pMHC. We used our benchmark to compare predictive performance using several flexible and rigid backbone TCR/pMHC docking protocols. Our flexible TCR docking algorithm, TCRFlexDock, improved predictive success over the fixed backbone protocol, leading to near-native predictions for 80% of the TCR/pMHC cases among the top 10 models, and 100% of the cases in the top 30 models. We then applied TCRFlexDock to predict the two distinct docking modes recently described for a single TCR bound to two different antigens, and tested several protein modeling scoring functions for prediction of TCR/pMHC binding affinities. This algorithm and benchmark should enable future efforts to predict, and design of uncharacterized TCR/pMHC complexes.

Extending RosettaDock with water, sugar, and pH for prediction of complex structures and affinities for CAPRI rounds 20-27

Rounds 20-27 of the Critical Assessment of PRotein Interactions (CAPRI) provided a testing platform for computational methods designed to address a wide range of challenges. The diverse targets drove the creation of and new combinations of computational tools. In this study, RosettaDock and other novel Rosetta protocols were used to successfully predict four of the 10 blind targets. For example, for DNase domain of Colicin E2-Im2 immunity protein, RosettaDock and RosettaLigand were used to predict the positions of water molecules at the interface, recovering 46% of the native water-mediated contacts. For α-repeat Rep4-Rep2 and g-type lysozyme-PliG inhibitor complexes, homology models were built and standard and pH-sensitive docking algorithms were used to generate structures with interface RMSD values of 3.3 Å and 2.0 Å, respectively. A novel flexible sugar-protein docking protocol was also developed and used for structure prediction of the BT4661-heparin-like saccharide complex, recovering 71% of the native contacts. Challenges remain in the generation of accurate homology models for protein mutants and sampling during global docking. On proteins designed to bind influenza hemagglutinin, only about half of the mutations were identified that affect binding (T55: 54%; T56: 48%). The prediction of the structure of the xylanase complex involving homology modeling and multidomain docking pushed the limits of global conformational sampling and did not result in any successful prediction. The diversity of problems at hand requires computational algorithms to be versatile; the recent additions to the Rosetta suite expand the capabilities to encompass more biologically realistic docking problems.

Rapid calculation of protein pKa values using Rosetta

We developed a Rosetta-based Monte Carlo method to calculate the pK(a) values of protein residues that commonly exhibit variable protonation states (Asp, Glu, Lys, His, and Tyr). We tested the technique by calculating pK(a) values for 264 residues from 34 proteins. The standard Rosetta score function, which is independent of any environmental conditions, failed to capture pK(a) shifts. After incorporating a Coulomb electrostatic potential and optimizing the solvation reference energies for pK(a) calculations, we employed a method that allowed side-chain flexibility and achieved a root mean-square deviation (RMSD) of 0.83 from experimental values (0.68 after discounting 11 predictions with an error over 2 pH units). Additional degrees of side-chain conformational freedom for the proximal residues facilitated the capture of charge-charge interactions in a few cases, resulting in an overall RMSD of 0.85 pH units. The addition of backbone flexibility increased the overall RMSD to 0.93 pH units but improved relative pK(a) predictions for proximal catalytic residues. The method also captures large pK(a) shifts of lysine and some glutamate point mutations in staphylococcal nuclease. Thus, a simple and fast method based on the Rosetta score function and limited conformational sampling produces pK(a) values that will be useful when rapid estimation is essential, such as in docking, design, and folding.

Using the RosettaSurface algorithm to predict protein structure at mineral surfaces

Determination of protein structure on mineral surfaces is necessary to understand biomineralization processes toward better treatment of biomineralization diseases and design of novel protein-synthesized materials. To date, limited atomic-resolution data have hindered experimental structure determination for proteins on mineral surfaces. Molecular simulation represents a complementary approach. In this chapter, we review RosettaSurface, a computational structure prediction-based algorithm designed to broadly sample conformational space to identify low-energy structures. We summarize the computational approaches, the published applications, and the new releases of the code in the Rosetta 3 framework. In addition, we provide a protocol capture to demonstrate the practical steps to employ RosettaSurface. As an example, we provide input files and output data analysis for a previously unstudied mineralization protein, osteocalcin. Finally, we summarize ongoing challenges in energy function optimization and conformational searching and suggest that the fusion between experiment and calculation is the best route forward.

Computer-aided antibody design

Recent clinical trials using antibodies with low toxicity and high efficiency have raised expectations for the development of next-generation protein therapeutics. However, the process of obtaining therapeutic antibodies remains time consuming and empirical. This review summarizes recent progresses in the field of computer-aided antibody development mainly focusing on antibody modeling, which is divided essentially into two parts: (i) modeling the antigen-binding site, also called the complementarity determining regions (CDRs), and (ii) predicting the relative orientations of the variable heavy (V(H)) and light (V(L)) chains. Among the six CDR loops, the greatest challenge is predicting the conformation of CDR-H3, which is the most important in antigen recognition. Further computational methods could be used in drug development based on crystal structures or homology models, including antibody-antigen dockings and energy calculations with approximate potential functions. These methods should guide experimental studies to improve the affinities and physicochemical properties of antibodies. Finally, several successful examples of in silico structure-based antibody designs are reviewed. We also briefly review structure-based antigen or immunogen design, with application to rational vaccine development.

IBC's 22nd Annual Antibody Engineering and 9th Annual Antibody Therapeutics International Conferences and the 2011 Annual Meeting of The Antibody Society, December 5-8, 2011, San Diego, CA

The 22nd Annual Antibody Engineering and 9th Annual Antibody Therapeutics international conferences, and the 2011 Annual Meeting of The Antibody Society, organized by IBC Life Sciences with contributions from The Antibody Society and two Scientific Advisory Boards, were held December 5-8, 2011 in San Diego, CA. The meeting drew ~800 participants who attended sessions on a wide variety of topics relevant to antibody research and development. As a preview to the main events, a pre-conference workshop held on December 4, 2011 focused on antibodies as probes of structure. The Antibody Engineering Conference comprised eight sessions: (1) structure and dynamics of antibodies and their membrane receptor targets; (2) model-guided generation of binding sites; (3) novel selection strategies; (4) antibodies in a complex environment: targeting intracellular and misfolded proteins; (5) rational vaccine design; (6) viral retargeting with engineered binding molecules; (7) the biology behind potential blockbuster antibodies and (8) antibodies as signaling modifiers: where did we go right, and can we learn from success? The Antibody Therapeutics session comprised five sessions: (1)Twenty-five years of therapeutic antibodies: lessons learned and future challenges; (2) preclinical and early stage development of antibody therapeutics; (3) next generation anti-angiogenics; (4) updates of clinical stage antibody therapeutics and (5) antibody drug conjugates and bispecific antibodies.

Protein-protein complex structure predictions by multimeric threading and template recombination

The total number of protein-protein complex structures currently available in the Protein Data Bank (PDB) is six times smaller than the total number of tertiary structures in the PDB, which limits the power of homology-based approaches to complex structure modeling. We present a threading-recombination approach, COTH, to boost the protein complex structure library by combining tertiary structure templates with complex alignments. The query sequences are first aligned to complex templates using a modified dynamic programming algorithm, guided by ab initio binding-site predictions. The monomer alignments are then shifted to the multimeric template framework by structural alignments. COTH was tested on 500 nonhomologous dimeric proteins, which can successfully detect correct templates for 50% of the cases after homologous templates are excluded, which significantly outperforms conventional homology modeling algorithms. It also shows a higher accuracy in interface modeling than rigid-body docking of unbound structures from ZDOCK although with lower coverage. These data demonstrate new avenues to model complex structures from nonhomologous templates.

Benchmarking and analysis of protein docking performance in Rosetta v3.2

RosettaDock has been increasingly used in protein docking and design strategies in order to predict the structure of protein-protein interfaces. Here we test capabilities of RosettaDock 3.2, part of the newly developed Rosetta v3.2 modeling suite, against Docking Benchmark 3.0, and compare it with RosettaDock v2.3, the latest version of the previous Rosetta software package. The benchmark contains a diverse set of 116 docking targets including 22 antibody-antigen complexes, 33 enzyme-inhibitor complexes, and 60 ‘other’ complexes. These targets were further classified by expected docking difficulty into 84 rigid-body targets, 17 medium targets, and 14 difficult targets. We carried out local docking perturbations for each target, using the unbound structures when available, in both RosettaDock v2.3 and v3.2. Overall the performances of RosettaDock v2.3 and v3.2 were similar. RosettaDock v3.2 achieved 56 docking funnels, compared to 49 in v2.3. A breakdown of docking performance by protein complex type shows that RosettaDock v3.2 achieved docking funnels for 63% of antibody-antigen targets, 62% of enzyme-inhibitor targets, and 35% of ‘other’ targets. In terms of docking difficulty, RosettaDock v3.2 achieved funnels for 58% of rigid-body targets, 30% of medium targets, and 14% of difficult targets. For targets that failed, we carry out additional analyses to identify the cause of failure, which showed that binding-induced backbone conformation changes account for a majority of failures. We also present a bootstrap statistical analysis that quantifies the reliability of the stochastic docking results. Finally, we demonstrate the additional functionality available in RosettaDock v3.2 by incorporating small-molecules and non-protein co-factors in docking of a smaller target set. This study marks the most extensive benchmarking of the RosettaDock module to date and establishes a baseline for future research in protein interface modeling and structure prediction.

The eisosome core is composed of BAR domain proteins

Eisosomes define sites of plasma membrane organization. In Saccharomyces cerevisiae, eisosomes delimit furrow-like plasma membrane invaginations that concentrate sterols, transporters, and signaling molecules. Eisosomes are static macromolecular assemblies composed of cytoplasmic proteins, most of which have no known function. In this study, we used a bioinformatics approach to analyze a set of 20 eisosome proteins. We found that the core components of eisosomes, paralogue proteins Pil1 and Lsp1, are distant homologues of membrane-sculpting Bin/amphiphysin/Rvs (BAR) proteins. Consistent with this finding, purified recombinant Pil1 and Lsp1 tubulated liposomes and formed tubules when the proteins were overexpressed in mammalian cells. Structural homology modeling and site-directed mutagenesis indicate that Pil1 positively charged surface patches are needed for membrane binding and liposome tubulation. Pil1 BAR domain mutants were defective in both eisosome assembly and plasma membrane domain organization. In addition, we found that eisosome-associated proteins Slm1 and Slm2 have F-BAR domains and that these domains are needed for targeting to furrow-like plasma membrane invaginations. Our results support a model in which BAR domain protein-mediated membrane bending leads to clustering of lipids and proteins within the plasma membrane.

Analysis and modeling of the variable region of camelid single-domain antibodies

Camelids have a special type of Ab, known as heavy chain Abs, which are devoid of classical Ab light chains. Relative to classical Abs, camelid heavy chain Abs (cAbs) have comparable immunogenicity, Ag recognition diversity and binding affinities, higher stability and solubility, and better manufacturability, making them promising candidates for alternate therapeutic scaffolds. Rational engineering of cAbs to improve therapeutic function requires knowledge of the differences of sequence and structural features between cAbs and classical Abs. In this study, amino acid sequences of 27 cAb variable regions (V(H)H) were aligned with the respective regions of 54 classical Abs to detect amino acid differences, enabling automatic identification of cAb V(H)H CDRs. CDR analysis revealed that the H1 often (and sometimes the H2) adopts diverse conformations not classifiable by established canonical rules. Also, although the cAb H3 is much longer than classical H3 loops, it often contains common structural motifs and sometimes a disulfide bond to the H1. Leveraging these observations, we created a Monte Carlo-based cAb V(H)H structural modeling tool, where the CDR H1 and H2 loops exhibited a median root-mean-square deviation to natives of 3.1 and 1.5 Å, respectively. The protocol generated 8-12, 14-16, and 16-24 residue H3 loops with a median root-mean-square deviation to natives of 5.7, 4.5, and 6.8 Å, respectively. The large deviation of the predicted loops underscores the challenge in modeling such long loops. cAb V(H)H homology models can provide structural insights into interaction mechanisms to enable development of novel Abs for therapeutic and biotechnological use.

Computational docking of antibody-antigen complexes, opportunities and pitfalls illustrated by influenza hemagglutinin

Antibodies play an increasingly important role in both basic research and the pharmaceutical industry. Since their efficiency depends, in ultimate analysis, on their atomic interactions with an antigen, studying such interactions is important to understand how they function and, in the long run, to design new molecules with desired properties. Computational docking, the process of predicting the conformation of a complex from its separated components, is emerging as a fast and affordable technique for the structural characterization of antibody-antigen complexes. In this manuscript, we first describe the different computational strategies for the modeling of antibodies and docking of their complexes, and then predict the binding of two antibodies to the stalk region of influenza hemagglutinin, an important pharmaceutical target. The purpose is two-fold: on a general note, we want to illustrate the advantages and pitfalls of computational docking with a practical example, using different approaches and comparing the results to known experimental structures. On a more specific note, we want to assess if docking can be successful in characterizing the binding to the same influenza epitope of other antibodies with unknown structure, which has practical relevance for pharmaceutical and biological research. The paper clearly shows that some of the computational docking predictions can be very accurate, but the algorithm often fails to discriminate them from inaccurate solutions. It is of paramount importance, therefore, to use rapidly obtained experimental data to validate the computational results.

Methods for the homology modeling of antibody variable regions

Antibodies are one of the critical molecules of our immune system and are unique in their enormous diversity required for recognizing various antigens. Antibodies are protein molecules and their antigen interacting region, the fragment variable (F (V)), is typically composed of a light (V (L)) and heavy (V (H)) chain. In particular, three loops each at the tip of the V (L) and the V (H), known as the complementarity determining region (CDR) loops, are responsible for binding to the antigen. While the framework regions of the V (L) and V (H) are relatively constant across the entire repertoire of antibodies, the conformation of the CDR loops varies extensively to enable the antibody to recognize different antigens. Three-dimensional structures of antibodies illustrating the V (L)-V (H) relative orientation and the CDR conformations are needed to gain insight into antibody stability, immunogenicity, and antibody-antigen interactions. Computational modeling provides a fast and inexpensive route for generating antibody structural models. This chapter highlights the various features crucial for creating a successful antibody homology model.

Mechanism of polyubiquitin chain recognition by the human ubiquitin conjugating enzyme Ube2g2

Ube2g2 is a human ubiquitin conjugating (E2) enzyme involved in the endoplasmic reticulum-associated degradation pathway, which is responsible for the identification and degradation of unfolded and misfolded proteins in the endoplasmic reticulum compartment. The Ube2g2-specific role is the assembly of Lys-48-linked polyubiquitin chains, which constitutes a signal for proteasomal degradation when attached to a substrate protein. NMR chemical shift perturbation and paramagnetic relaxation enhancement approaches were employed to characterize the binding interaction between Ube2g2 and ubiquitin, Lys-48-linked diubiquitin, and Lys-63-linked diubiquitin. Results demonstrate that ubiquitin binds to Ube2g2 with an affinity of 90 μM in two different orientations that are rotated by 180° in models generated by the RosettaDock modeling suite. The binding of Ube2g2 to Lys-48- and Lys-63-linked diubiquitin is primarily driven by interactions with individual ubiquitin subunits, with a clear preference for the subunit containing the free Lys-48 or Lys-63 side chain (i.e. the distal subunit). This preference is particularly striking in the case of Lys-48-linked diubiquitin, which exhibits an ∼3-fold difference in affinities between the two ubiquitin subunits. This difference can be attributed to the partial steric occlusion of the subunit whose Lys-48 side chain is involved in the isopeptide linkage. As such, these results suggest that Lys-48-linked polyubiquitin chains may be designed to bind certain proteins like Ube2g2 such that the terminal ubiquitin subunit carrying the reactive Lys-48 side chain can be positioned properly for chain elongation regardless of chain length.