Protein Structure Prediction Program Research Articles

Putative purine nucleoside interacting residues in the malaria parasite purine uptake transporter PfENT1 are critical for transporter function.

Malaria remains a major public health threat for billions of people worldwide. Infection with obligate intracellular, unicellular parasites from the genus Plasmodium causes malaria. Plasmodium falciparum causes the deadliest form of human malaria. Plasmodium parasites are purine auxotrophic. They rely on purine import from the host red blood cell cytoplasm via equilibrative nucleoside transporters to supply substrates to the purine salvage pathway. We previously developed a high throughput screening assay to identify inhibitors of the P. falciparum Equilibrative Nucleoside Transporter Type 1 (PfENT1). Screening a small molecule library identified PfENT1 inhibitors that blocked proliferation of P. falciparum parasites in in vitro culture. The goal of the current work was to validate a high-resolution model of PfENT1 predicted by the AlphaFold protein structure prediction program. We superimposed the predicted PfENT1 structure on the human homologue structure, hENT1, and developed a structure-based sequence alignment. We mutated the residues in PfENT1 aligned with and flanking the residues in hENT1 that interact with the purine analog, nitrobenzylthioinosine (NBMPR). Mutation of the PfENT1 residues Q135, D287, and R291 that are predicted to form hydrogen bonds to purine nucleosides eliminated purine and pyrimidine transport function in various yeast-based growth and radiolabeled substrate uptake assays. Mutation of two flanking residues, W53 and S290, also resulted in inactive protein. Mutation of L50 that forms hydrophobic interactions with the purine nucleobase reduced transport function. Based on our results the AlphaFold predicted structure for PfENT1 may be useful in guiding medicinal chemistry efforts to improve the potency of our PfENT1 inhibitors.

Partial Atomic Model of the Tailed Lactococcal Phage TP901-1 as Predicted by AlphaFold2: Revelations and Limitations.

Bacteria are engaged in a constant battle against preying viruses, called bacteriophages (or phages). These remarkable nano-machines pack and store their genomes in a capsid and inject it into the cytoplasm of their bacterial prey following specific adhesion to the host cell surface. Tailed phages possessing dsDNA genomes are the most abundant phages in the bacterial virosphere, particularly those with long, non-contractile tails. All tailed phages possess a nano-device at their tail tip that specifically recognizes and adheres to a suitable host cell surface receptor, being proteinaceous and/or saccharidic. Adhesion devices of tailed phages infecting Gram-positive bacteria are highly diverse and, for the majority, remain poorly understood. Their long, flexible, multi-domain-encompassing tail limits experimental approaches to determine their complete structure. We have previously shown that the recently developed protein structure prediction program AlphaFold2 can overcome this limitation by predicting the structures of phage adhesion devices with confidence. Here, we extend this approach and employ AlphaFold2 to determine the structure of a complete phage, the lactococcal P335 phage TP901-1. Herein we report the structures of its capsid and neck, its extended tail, and the complete adhesion device, the baseplate, which was previously partially determined using X-ray crystallography.

How AlphaFold2 Predicts Conditionally Folding Regions Annotated in an Intrinsically Disordered Protein Database, IDEAL.

AlphaFold2 (AF2) is a protein structure prediction program which provides accurate models. In addition to predicting structural domains, AF2 assigns intrinsically disordered regions (IDRs) by identifying regions with low prediction reliability (pLDDT). Some regions in IDRs undergo disorder-to-order transition upon binding the interaction partner. Here we assessed model structures of AF2 based on the annotations in IDEAL, in which segments with disorder-to-order transition have been collected as Protean Segments (ProSs). We non-redundantly selected ProSs from IDEAL and classified them based on the root mean square deviation to the corresponding region of AF2 models. Statistical analysis identified 11 structural and sequential features, possibly contributing toward the prediction of ProS structures. These features were categorized into two groups: one that contained pLDDT and the other that contained normalized radius of gyration. The typical ProS structures in the former group comprise a long α helix or a whole or part of the structural domain and those in the latter group comprise a short α helix with terminal loops.

The Active Site of the Enzyme 10-Formyl-THFDH in the Honey Bee Apis mellifera-A Key Player in Formic Acid Detoxification.

Honey bees are important managed pollinators that fulfill important ecological and economic functions. In recent decades, the obligate ectoparasite Varroa destructor severely affected the survival of honey bees, as it weakened them by different means. A common treatment against V. destructor is formic acid fumigation, which has been used for decades by beekeepers across the world. This treatment is known to be effective, but many beekeepers report adverse effects of formic acid on bees, which include damage to the brood, worker bee mortality, and queen loss. Little is known about the molecular mechanisms of formic acid detoxification in honey bees. Recently, we reported upregulation of the bee enzyme, 10-formyl-THFDH, under formic acid fumigation. Here, the active site of this enzyme is characterized by an interdisciplinary approach combining homology modeling and protein mutagenesis. In addition, the limitations of the 3D protein structure prediction program AlphaFold2 are shown in regard to docking studies. This study provides a more thorough understanding of the molecular detoxification mechanisms of formic acid in Apis mellifera.

Exploring Structural Diversity among Adhesion Devices Encoded by Lactococcal P335 Phages with AlphaFold2.

Bacteriophages, or phages, are the most abundant biological entities on Earth. They possess molecular nanodevices to package and store their genome, as well as to introduce it into the cytoplasm of their bacterial prey. Successful phage infection commences with specific recognition of, and adhesion to, a suitable host cell surface. Adhesion devices of siphophages infecting Gram-positive bacteria are very diverse and remain, for the majority, poorly understood. These assemblies often comprise long, flexible, and multi-domain proteins, which limit their structural analyses by experimental approaches. The protein structure prediction program AlphaFold2 is exquisitely adapted to unveil structural and functional details of such molecular machineries. Here, we present structure predictions of adhesion devices from siphophages belonging to the P335 group infecting Lactococcus spp., one of the most extensively applied lactic acid bacteria in dairy fermentations. The predictions of representative adhesion devices from types I-IV P335 phages illustrate their very diverse topology. Adhesion devices from types III and IV phages share a common topology with that of Skunavirus p2, with a receptor binding protein anchored to the virion by a distal tail protein loop. This suggests that they exhibit an activation mechanism similar to that of phage p2 prior to host binding.

Computation-Aided Design of Albumin Affibody-Inserted Antibody Fragment for the Prolonged Serum Half-Life.

Single-chain variable fragments (scFvs) have been recognized as promising agents in cancer therapy. However, short serum half-life of scFvs often limits clinical application. Fusion to albumin affibody (ABD) is an effective and convenient half-life extension strategy. Although one terminus of scFv is available for fusion of ABD, it is also frequently used for fusion of useful moieties such as small functional proteins, cytokines, or antibodies. Herein, we investigated the internal linker region for ABD fusion instead of terminal region, which was rarely explored before. We constructed two internally ABD-inserted anti-HER2 4D5scFv (4D5-ABD) variants, which have short (4D5-S-ABD) and long (4D5-L-ABD) linker length respectively. The model structures of these 4D5scFv and 4D5-ABD variants predicted using the deep learning-based protein structure prediction program (AlphaFold2) revealed high similarity to either the original 4D5scFv or the ABD structure, implying that the functionality would be retained. Designed 4D5-ABD variants were expressed in the bacterial expression system and characterized. Both 4D5-ABD variants showed anti-HER2 binding affinity comparable with 4D5scFv. Binding affinity of both 4D5-ABD variants against albumin was also comparable. In a pharmacokinetic study in mice, the 4D5-ABD variants showed a significantly prolonged half-life of 34 h, 114 times longer than that of 4D5scFv. In conclusion, we have developed a versatile scFv platform with enhanced pharmacokinetic profiles with an aid of deep learning-based structure prediction.

A structural discovery journey of streptococcal phages adhesion devices by AlphaFold2.

Successful bacteriophage infection starts with specific recognition and adhesion to the host cell surface. Adhesion devices of siphophages infecting Gram-positive bacteria are very diverse and remain, for the majority, poorly understood. These assemblies often comprise long, flexible, and multi-domain proteins, which limits their structural analyses by experimental approaches such as X-ray crystallography and electron microscopy. However, the protein structure prediction program AlphaFold2 is exquisitely adapted to unveil structural and functional details of such molecular machineries. Here, we present structure predictions of whole adhesion devices of five representative siphophages infecting Streptococcus thermophilus, one of the main lactic acid bacteria used in dairy fermentations. The predictions highlight the mosaic nature of these devices that share functional domains for which active sites and residues could be unambiguously identified. Such AlphaFold2 analyses of phage-encoded host adhesion devices should become a standard method to characterize phage-host interaction machineries and to reliably annotate phage genomes.

Present Impact of AlphaFold2 Revolution on Structural Biology, and an Illustration With the Structure Prediction of the Bacteriophage J-1 Host Adhesion Device.

In 2021, the release of AlphaFold2 - the DeepMind’s machine-learning protein structure prediction program - revolutionized structural biology. Results of the CASP14 contest were an immense surprise as AlphaFold2 successfully predicted 3D structures of nearly all submitted protein sequences. The AlphaFold2 craze has rapidly spread the life science community since structural biologists as well as untrained biologists have now the possibility to obtain high-confidence protein structures. This revolution is opening new avenues to address challenging biological questions. Moreover, AlphaFold2 is imposing itself as an essential step of any structural biology project, and requires us to revisit our structural biology workflows. On one hand, AlphaFold2 synergizes with experimental methods including X-ray crystallography and cryo-electron microscopy. On the other hand, it is, to date, the only method enabling structural analyses of large and flexible assemblies resistant to experimental approaches. We illustrate this valuable application of AlphaFold2 with the structure prediction of the whole host adhesion device from the Lactobacillus casei bacteriophage J-1. With the ongoing improvement of AlphaFold2 algorithms and notebooks, there is no doubt that AlphaFold2-driven biological stories will increasingly be reported, which questions the future directions of experimental structural biology.

GhoT of the GhoT/GhoS toxin/antitoxin system damages lipid membranes by forming transient pores

GhoT is a bacterial toxin of the type V toxin/antitoxin system that allows Escherichia coli to reduce its metabolism in response to oxidative and bile stress. GhoT functions by increasing membrane permeability and reducing both ATP levels and the proton motive force. However, how GhoT damages the inner membrane has not been elucidated. Here we investigated how GhoT damages membranes by studying its interaction with lipid bilayers and determined that GhoT does not cause macroscopic disruption of the lipid bilayer to increase membrane permeability to the dye carboxyfluorescein. Using circular dichroism, we found that GhoT forms an alpha helical structure in lipid bilayers that agrees with the structure predicted by the I-TASSER protein structure prediction program. The structure generated using I-TASSER was used to conduct coarse-grained molecular dynamics simulations, which indicate that GhoT damages the cell membrane, as a multimer, by forming transient transmembrane pores.

Suppression of RNAi by dsRNA-degrading RNaseIII enzymes of viruses in animals and plants.

Certain RNA and DNA viruses that infect plants, insects, fish or poikilothermic animals encode Class 1 RNaseIII endoribonuclease-like proteins. dsRNA-specific endoribonuclease activity of the RNaseIII of rock bream iridovirus infecting fish and Sweet potato chlorotic stunt crinivirus (SPCSV) infecting plants has been shown. Suppression of the host antiviral RNA interference (RNAi) pathway has been documented with the RNaseIII of SPCSV and Heliothis virescens ascovirus infecting insects. Suppression of RNAi by the viral RNaseIIIs in non-host organisms of different kingdoms is not known. Here we expressed PPR3, the RNaseIII of Pike-perch iridovirus, in the non-hosts Nicotiana benthamiana (plant) and Caenorhabditis elegans (nematode) and found that it cleaves double-stranded small interfering RNA (ds-siRNA) molecules that are pivotal in the host RNA interference (RNAi) pathway and thereby suppresses RNAi in non-host tissues. In N. benthamiana, PPR3 enhanced accumulation of Tobacco rattle tobravirus RNA1 replicon lacking the 16K RNAi suppressor. Furthermore, PPR3 suppressed single-stranded RNA (ssRNA)—mediated RNAi and rescued replication of Flock House virus RNA1 replicon lacking the B2 RNAi suppressor in C. elegans. Suppression of RNAi was debilitated with the catalytically compromised mutant PPR3-Ala. However, the RNaseIII (CSR3) produced by SPCSV, which cleaves ds-siRNA and counteracts antiviral RNAi in plants, failed to suppress ssRNA-mediated RNAi in C. elegans. In leaves of N. benthamiana, PPR3 suppressed RNAi induced by ssRNA and dsRNA and reversed silencing; CSR3, however, suppressed only RNAi induced by ssRNA and was unable to reverse silencing. Neither PPR3 nor CSR3 suppressed antisense-mediated RNAi in Drosophila melanogaster. These results show that the RNaseIII enzymes of RNA and DNA viruses suppress RNAi, which requires catalytic activities of RNaseIII. In contrast to other viral silencing suppression proteins, the RNaseIII enzymes are homologous in unrelated RNA and DNA viruses and can be detected in viral genomes using gene modeling and protein structure prediction programs.

An improved method to detect correct protein folds using partial clustering

BackgroundStructure-based clustering is commonly used to identify correct protein folds among candidate folds (also called decoys) generated by protein structure prediction programs. However, traditional clustering methods exhibit a poor runtime performance on large decoy sets. We hypothesized that a more efficient “partial“ clustering approach in combination with an improved scoring scheme could significantly improve both the speed and performance of existing candidate selection methods.ResultsWe propose a new scheme that performs rapid but incomplete clustering on protein decoys. Our method detects structurally similar decoys (measured using either Cα RMSD or GDT-TS score) and extracts representatives from them without assigning every decoy to a cluster. We integrated our new clustering strategy with several different scoring functions to assess both the performance and speed in identifying correct or near-correct folds. Experimental results on 35 Rosetta decoy sets and 40 I-TASSER decoy sets show that our method can improve the correct fold detection rate as assessed by two different quality criteria. This improvement is significantly better than two recently published clustering methods, Durandal and Calibur-lite. Speed and efficiency testing shows that our method can handle much larger decoy sets and is up to 22 times faster than Durandal and Calibur-lite.ConclusionsThe new method, named HS-Forest, avoids the computationally expensive task of clustering every decoy, yet still allows superior correct-fold selection. Its improved speed, efficiency and decoy-selection performance should enable structure prediction researchers to work with larger decoy sets and significantly improve their ab initio structure prediction performance.

Cotranslational protein folding and terminus hydrophobicity.

Peptides fold on a time scale that is much smaller than the time required for synthesis, whence all proteins potentially fold cotranslationally to some degree (followed by additional folding events after release from the ribosome). In this paper, in three different ways, we find that cotranslational folding success is associated with higher hydrophobicity at the N-terminus than at the C-terminus. First, we fold simple HP models on a square lattice and observe that HP sequences that fold better cotranslationally than from a fully extended state exhibit a positive difference (N−C) in terminus hydrophobicity. Second, we examine real proteins using a previously established measure of potential cotranslationality known as ALR (Average Logarithmic Ratio of the extent of previous contacts) and again find a correlation with the difference in terminus hydrophobicity. Finally, we use the cotranslational protein structure prediction program SAINT and again find that such an approach to folding is more successful for proteins with higher N-terminus than C-terminus hydrophobicity. All results indicate that cotranslational folding is promoted in part by a hydrophobic start and a less hydrophobic finish to the sequence.

Designing succinct structural alphabets

Motivation: The 3D structure of a protein sequence can be assembled from the substructures corresponding to small segments of this sequence. For each small sequence segment, there are only a few more likely substructures. We call them the ‘structural alphabet’ for this segment. Classical approaches such as ROSETTA used sequence profile and secondary structure information, to predict structural fragments. In contrast, we utilize more structural information, such as solvent accessibility and contact capacity, for finding structural fragments.Results: Integer linear programming technique is applied to derive the best combination of these sequence and structural information items. This approach generates significantly more accurate and succinct structural alphabets with more than 50% improvement over the previous accuracies. With these novel structural alphabets, we are able to construct more accurate protein structures than the state-of-art ab initio protein structure prediction programs such as ROSETTA. We are also able to reduce the Kolodny's library size by a factor of 8, at the same accuracy.Availability: The online FRazor server is under constructionContact:scli@uwaterloo.ca,mli@uwaterloo.ca, j3xu@tti-c.org

Protein contact order prediction from primary sequences

BackgroundContact order is a topological descriptor that has been shown to be correlated with several interesting protein properties such as protein folding rates and protein transition state placements. Contact order has also been used to select for viable protein folds from ab initio protein structure prediction programs. For proteins of known three-dimensional structure, their contact order can be calculated directly. However, for proteins with unknown three-dimensional structure, there is no effective prediction method currently available.ResultsIn this paper, we propose several simple yet very effective methods to predict contact order from the amino acid sequence only. One set of methods is based on a weighted linear combination of predicted secondary structure content and amino acid composition. Depending on the number of components used in these equations it is possible to achieve a correlation coefficient of 0.857–0.870 between the observed and predicted contact order. A second method, based on sequence similarity to known three-dimensional structures, is able to achieve a correlation coefficient of 0.977. We have also developed a much more robust implementation for calculating contact order directly from PDB coordinates that works for > 99% PDB files. All of these contact order predictors and calculators have been implemented as a web server (see Availability and requirements section for URL).ConclusionProtein contact order can be effectively predicted from the primary sequence, at the absence of three-dimensional structure. Three factors, percentage of residues in alpha helices, percentage of residues in beta strands, and sequence length, appear to be strongly correlated with the absolute contact order.

VIRTUALLY CREATED ENZYMES

IN A DRAMATIC EXAMPLE of translation from in silico to in vitro, scientists have computationally designed a series of new enzymes that catalyze a multistep reaction—including the breaking of a carbon-carbon bond—in laboratory experiments. The designed enzymes, which catalyze the reverse of an aldol condensation reaction, aren’t nearly as active as naturally occurring enzymes that perform similar tasks ( Science 2008, 319 , 1387). But the fact that they work on a substrate not found in nature suggests that computer-aided design might yield enzymes that can perform nearly any reaction a chemist desires. Biochemistry professor David Baker, graduate student Lin Jiang, and postdoc Eric A. Althoff of the University of Washington, Seattle, and their colleagues used the Rosetta protein structure prediction program developed in the Baker lab to design a library of enzymes with active sites likely to catalyze the so-called retro-aldol reaction. They whittled down a virtual list of more than 180,000 candidates to a ...

PROSPECT II: protein structure prediction program for genome-scale applications.

A new method for fold recognition is developed and added to the general protein structure prediction package PROSPECT (http://compbio.ornl.gov/PROSPECT/). The new method (PROSPECT II) has four key features. (i) We have developed an efficient way to utilize the evolutionary information for evaluating the threading potentials including singleton and pairwise energies. (ii) We have developed a two-stage threading strategy: (a) threading using dynamic programming without considering the pairwise energy and (b) fold recognition considering all the energy terms, including the pairwise energy calculated from the dynamic programming threading alignments. (iii) We have developed a combined z-score scheme for fold recognition, which takes into consideration the z-scores of each energy term. (iv) Based on the z-scores, we have developed a confidence index, which measures the reliability of a prediction and a possible structure-function relationship based on a statistical analysis of a large data set consisting of threadings of 600 query proteins against the entire FSSP templates. Tests on several benchmark sets indicate that the evolutionary information and other new features of PROSPECT II greatly improve the alignment accuracy. We also demonstrate that the performance of PROSPECT II on fold recognition is significantly better than any other method available at all levels of similarity. Improvement in the sensitivity of the fold recognition, especially at the superfamily and fold levels, makes PROSPECT II a reliable and fully automated protein structure and function prediction program for genome-scale applications.

A distance-dependent atomic knowledge-based potential for improved protein structure selection.

A heavy atom distance-dependent knowledge-based pairwise potential has been developed. This statistical potential is first evaluated and optimized with the native structure z-scores from gapless threading. The potential is then used to recognize the native and near-native structures from both published decoy test sets, as well as decoys obtained from our group's protein structure prediction program. In the gapless threading test, there is an average z-score improvement of 4 units in the optimized atomic potential over the residue-based quasichemical potential. Examination of the z-scores for individual pairwise distance shells indicates that the specificity for the native protein structure is greatest at pairwise distances of 3.5-6.5 A, i.e., in the first solvation shell. On applying the current atomic potential to test sets obtained from the web, composed of native protein and decoy structures, the current generation of the potential performs better than residue-based potentials as well as the other published atomic potentials in the task of selecting native and near-native structures. This newly developed potential is also applied to structures of varying quality generated by our group's protein structure prediction program. The current atomic potential tends to pick lower RMSD structures than do residue-based contact potentials. In particular, this atomic pairwise interaction potential has better selectivity especially for near-native structures. As such, it can be used to select near-native folds generated by structure prediction algorithms as well as for protein structure refinement.