When Alphafold2 predictions go wrong for protein-protein complexes, is there something to be learnt?

In this short communication, I analyze cases of failed predictions for protein-protein complexes with Alphafold2, and show that they either point to erroneous annotation in the PDB or correct binding site regions.

Values for the Digestibility of Pea Protein Isolate or Casein Amino Acids Determined using the Dual Isotope Method Are Not Similar to Those Derived with the Standard Ileal Balance Method in Healthy Volunteers.

BACKGROUND: The measurement of ileal amino acid (AA) digestibility is invasive and inappropriate when applied to vulnerable populations. The dual isotope method has been developed over the past 5 y as an alternative method. OBJECTIVE: The aim of this work was to compare the indispensable amino acid (IAA) digestibility values of 2 different proteins obtained using the dual isotope and the standard ileal balance methods in the same subjects. METHODS: Fifteen healthy adults completed the study. Over 4 h, they ingested 9 successive portions of mashed potatoes containing the test protein (pea protein or casein) labeled intrinsically with 15N and 2H, and a 13C-free AA mixture as a reference for the dual isotope method. Plasma was sampled regularly over the 8-h postprandial period, whereas the ileal digesta was collected continuously via a naso-ileal tube. Isotopic enrichments (15N and 13C) were measured in the digesta for the direct determination of ileal IAA digestibility, whereas plasma enrichments (2H and 13C) were measured to determine IAA digestibility using the dual isotope method. RESULTS: The 4-h repeated meal procedure enabled the almost complete digestion of test proteins at 8 h and the attainment of a plasma isotopic plateau between 2.5 and 4 h. These conditions were necessary to perform the ileal balance and dual isotope methods simultaneously. For pea protein, the mean IAA digestibility was similar between the 2 methods, but significant differences (from 10% to 20%) were observed for individual IAA values. For casein, IAA digestibility was significantly lower with the dual isotope method for all the IAA analyzed. CONCLUSIONS: Under our experimental conditions, the degree of agreement between the dual isotope and ileal balance methods varied among AAs and depended on the protein source. Further research is needed to validate the dual isotope method. This study was registered at clinicaltrials.gov as NCT04072770.

AlphaFold2 Predicts Whether Proteins Interact Amidst Confounding Structural Compatibility.

Predicting whether two proteins physically interact is one of the holy grails of computational biology, galvanized by rapid advancements in deep learning. AlphaFold2, although not developed with this goal, is promising in this respect. Here, I test the prediction capability of AlphaFold2 on a very challenging data set, where proteins are structurally compatible, even when they do not interact. AlphaFold2 achieves high discrimination between interacting and non-interacting proteins, and the cases of misclassifications can either be rescued by revisiting the input sequences or can suggest false positives and negatives in the data set. AlphaFold2 is thus not impaired by the compatibility between protein structures and has the potential to be applied on a large scale.

Discriminating physiological from non-physiological interfaces in structures of protein complexes: A community-wide study.

Reliably scoring and ranking candidate models of protein complexes and assigning their oligomeric state from the structure of the crystal lattice represent outstanding challenges. A community-wide effort was launched to tackle these challenges. The latest resources on protein complexes and interfaces were exploited to derive a benchmark dataset consisting of 1677 homodimer protein crystal structures, including a balanced mix of physiological and non-physiological complexes. The non-physiological complexes in the benchmark were selected to bury a similar or larger interface area than their physiological counterparts, making it more difficult for scoring functions to differentiate between them. Next, 252 functions for scoring protein-protein interfaces previously developed by 13 groups were collected and evaluated for their ability to discriminate between physiological and non-physiological complexes. A simple consensus score generated using the best performing score of each of the 13 groups, and a cross-validated Random Forest (RF) classifier were created. Both approaches showed excellent performance, with an area under the Receiver Operating Characteristic (ROC) curve of 0.93 and 0.94, respectively, outperforming individual scores developed by different groups. Additionally, AlphaFold2 engines recalled the physiological dimers with significantly higher accuracy than the non-physiological set, lending support to the reliability of our benchmark dataset annotations. Optimizing the combined power of interface scoring functions and evaluating it on challenging benchmark datasets appears to be a promising strategy.

Internal Normal Mode Analysis Applied to RNA Flexibility and Conformational Changes.

We investigated the capability of internal normal modes to reproduce RNA flexibility and predict observed RNA conformational changes and, notably, those induced by the formation of RNA-protein and RNA-ligand complexes. Here, we extended our iNMA approach developed for proteins to study RNA molecules using a simplified representation of the RNA structure and its potential energy. Three data sets were also created to investigate different aspects. Despite all the approximations, our study shows that iNMA is a suitable method to take into account RNA flexibility and describe its conformational changes opening the route to its applicability in any integrative approach where these properties are crucial.

Rational Prediction of PROTAC-Compatible Protein-Protein Interfaces by Molecular Docking.

Proteolysis targeting chimeras (PROTACs) are heterobifunctional ligands that mediate the interaction between a protein target and an E3 ligase, resulting in a ternary complex, whose interaction with the ubiquitination machinery leads to target degradation. This technology is emerging as an exciting new avenue for therapeutic development, with several PROTACs currently undergoing clinical trials targeting cancer. Here, we describe a general and computationally efficient methodology combining restraint-based docking, energy-based rescoring, and a filter based on the minimal solvent-accessible surface distance to produce PROTAC-compatible PPIs suitable for when there is no a priori known PROTAC ligand. In a benchmark employing a manually curated data set of 13 ternary complex crystals, we achieved an accuracy of 92% when starting from bound structures and 77% when starting from unbound structures, respectively. Our method only requires that the ligand-bound structures of the monomeric forms of the E3 ligase and target proteins be given to run, making it general, accurate, and highly efficient, with the ability to impact early-stage PROTAC-based drug design campaigns where no structural information about the ternary complex structure is available.

Transformative adaptation through nature-based solutions: a comparative case study analysis in China, Italy, and Germany.

This paper explores how claims for transformative adaptation toward more equitable and sustainable societies can be assessed. We build on a theoretical framework describing transformative adaptation as it manifests across four core elements of the public-sector adaptation lifecycle: vision, planning, institutional frameworks, and interventions. For each element, we identify characteristics that can help track adaptation as transformative. Our purpose is to identify how governance systems can constrain or support transformative choices and thus enable targeted interventions. We demonstrate and test the usefulness of the framework with reference to three government-led adaptation projects of nature-based solutions (NBS): river restoration (Germany), forest conservation (China), and landslide risk reduction (Italy). Building on a desktop study and open-ended interviews, our analysis adds evidence to the view that transformation is not an abrupt system change, but a dynamic complex process that evolves over time. While each of the NBS cases fails to fulfill all the transformation characteristics, there are important transformative elements in their visions, planning, and interventions. There is a deficit, however, in the transformation of institutional frameworks. The cases show institutional commonalities in multi-scale and cross-sectoral (polycentric) collaboration as well as innovative processes for inclusive stakeholder engagement; yet, these arrangements are ad hoc, short-term, dependent on local champions, and lacking the permanency needed for upscaling. For the public sector, this result highlights the potential for establishing cross-competing priorities among agencies, cross-sectoral formal mechanisms, new dedicated institutions, and programmatic and regulatory mainstreaming. Supplementary Information: The online version contains supplementary material available at 10.1007/s10113-023-02066-7.

The Candida glabrata glycogen branching enzyme structure reveals unique features of branching enzymes of the Saccharomycetaceae phylum.

Branching enzymes (BE) are responsible for the formation of branching points at the 1,6 position in glycogen and starch, by catalyzing the cleavage of α-1,4-linkages and the subsequent transfer by introducing α-1,6-linked glucose branched points. BEs are found in the large GH13 family, eukaryotic BEs being mainly classified in the GH13_8 subfamily, GH13_9 grouping almost exclusively prokaryotic enzymes. With the aim of contributing to the understanding of the mode of recognition and action of the enzymes belonging to GH13_8, and to the understanding of features distinguishing these enzymes from those belonging to subfamily 13_9, we solved the crystal structure of the glycogen branching enzyme (GBE) from the yeast Candida glabrata, CgGBE, in ligand-free forms and in complex with a maltotriose. The structures revealed the presence of a domain already observed in Homo sapiens and Oryza sativa BEs that we named α-helical N-terminal domain, in addition to the three conserved domains found in BE. We confirmed by phylogenetic analysis that this α-helical N-terminal domain is always present in the GH13_8 enzymes suggesting that it could actually present a signature for this subfamily. We identified two binding sites in the α-helical N-terminal domain and in the carbohydrate binding module 48 (CBM48), respectively, which show a unique structural organization only present in the Saccharomycotina phylum. Our structural and phylogenetic investigation provides new insight into the structural characterization of GH13_8 GBE revealing that unique structural features only present in the Saccharomycotina phylum thereby conferring original properties to this group of enzymes.

Using normal mode analysis on protein structural models. How far can we go on our predictions?

Normal mode analysis (NMA) is a fast and inexpensive approach that is largely used to gain insight into functional protein motions, and more recently to create conformations for further computational studies. However, when the protein structure is unknown, the use of computational models is necessary. Here, we analyze the capacity of NMA in internal coordinate space to predict protein motion, its intrinsic flexibility, and atomic displacements, using protein models instead of native structures, and the possibility to use it for model refinement. Our results show that NMA is quite insensitive to modeling errors, but that calculations are strictly reliable only for very accurate models. Our study also suggests that internal NMA is a more suitable tool for the improvement of structural models, and for integrating them with experimental data or in other computational techniques, such as protein docking or more refined molecular dynamics simulations.

Identification and visualization of protein binding regions with the ArDock server.

ArDock (ardock.ibcp.fr) is a structural bioinformatics web server for the prediction and the visualization of potential interaction regions at protein surfaces. ArDock ranks the surface residues of a protein according to their tendency to form interfaces in a set of predefined docking experiments between the query protein and a set of arbitrary protein probes. The ArDock methodology is derived from large scale cross-docking studies where it was observed that randomly chosen proteins tend to dock in a non-random way at protein surfaces. The method predicts interaction site of the protein, or alternate interfaces in the case of proteins with multiple interaction modes. The server takes a protein structure as input and computes a score for each surface residue. Its output focuses on the interactive visualization of results and on interoperability with other services.

Decomposing protein-DNA binding and recognition using simplified protein models.

We analyze the role of different physicochemical factors in protein/DNA binding and recognition by comparing the results from all-atom molecular dynamics simulations with simulations using simplified protein models. These models enable us to separate the role of specific amino acid side chains, formal amino acid charges and hydrogen bonding from the effects of the low-dielectric volume occupied by the protein. Comparisons are made on the basis of the conformation of DNA after protein binding, the ionic distribution around the complex and the sequence specificity. The results for four transcription factors, binding in either the minor or major grooves of DNA, show that the protein volume and formal charges, with one exception, play a predominant role in binding. Adding hydrogen bonding and a very small number of key amino acid side chains at the all-atom level yields results in DNA conformations and sequence recognition close to those seen in the reference all-atom simulations.

Dynamics and recognition within a protein-DNA complex: a molecular dynamics study of the SKN-1/DNA interaction.

Molecular dynamics simulations of the Caenorhabditis elegans transcription factor SKN-1 bound to its cognate DNA site show that the protein-DNA interface undergoes significant dynamics on the microsecond timescale. A detailed analysis of the simulation shows that movements of two key arginine side chains between the major groove and the backbone of DNA generate distinct conformational sub-states that each recognize only part of the consensus binding sequence of SKN-1, while the experimentally observed binding specificity results from a time-averaged view of the dynamic recognition occurring within this complex.

Protein-DNA interfaces: a molecular dynamics analysis of time-dependent recognition processes for three transcription factors.

We have studied the dynamics of three transcription factor-DNA complexes using all-atom, microsecond-scale MD simulations. In each case, the salt bridges and hydrogen bond interactions formed at the protein-DNA interface are found to be dynamic, with lifetimes typically in the range of tens to hundreds of picoseconds, although some interactions, notably those involving specific binding to DNA bases, can be a hundred times longer lived. Depending on the complex studied, this dynamics may or may not lead to the existence of distinct conformational substates. Using a sequence threading technique, it has been possible to determine whether DNA sequence recognition is sensitive or not to such conformational changes, and, in one case, to show that recognition appears to be locally dependent on protein-mediated cation distributions.

The C-terminal Domains of Apoptotic BH3-only Proteins Mediate Their Insertion into Distinct Biological Membranes.

Changes in the equilibrium of pro- and anti-apoptotic members of the B-cell lymphoma-2 (Bcl-2) protein family in the mitochondrial outer membrane (MOM) induce structural changes that commit cells to apoptosis. Bcl-2 homology-3 (BH3)-only proteins participate in this process by either activating pro-apoptotic effectors or inhibiting anti-apoptotic components and by promoting MOM permeabilization. The association of BH3-only proteins with MOMs is necessary for the activation and amplification of death signals; however, the nature of this association remains controversial, as these proteins lack a canonical transmembrane sequence. Here we used an in vitro expression system to study the insertion capacity of hydrophobic C-terminal regions of the BH3-only proteins Bik, Bim, Noxa, Bmf, and Puma into microsomal membranes. An Escherichia coli complementation assay was used to validate the results in a cellular context, and peptide insertions were modeled using molecular dynamics simulations. We also found that some of the C-terminal domains were sufficient to direct green fluorescent protein fusion proteins to specific membranes in human cells, but the domains did not activate apoptosis. Thus, the hydrophobic regions in the C termini of BH3-only members associated in distinct ways with various biological membranes, suggesting that a detailed investigation of the entire process of apoptosis should include studying the membranes as a setting for protein-protein and protein-membrane interactions.

Nutritional stress exacerbates hepatic steatosis induced by deletion of the histidine nucleotide-binding (Hint2) mitochondrial protein.

The histidine nucleotide-binding protein, Hint2, is a mitochondrial phosphoramidase expressed in liver, brown fat, pancreas, and muscle. The livers of Hint2 knockout (Hint2(-/-)) mice accumulate triglycerides and show a pattern of mitochondrial protein lysine hyperacetylation. The extent and nature of the lysine acetylation changes and the response of Hint2(-/-) mice to nutritional challenges that elicit a modification of protein acetylation have not been investigated. To compare the adaptation of Hint2(-/-) and control (Hint2(+/+)) mice with episodes of fasting and high-fat diet (HFD), we subjected animals to either feeding ad libitum or fasting for 24 h, and to either a HFD or control diet for 8 wk. Triglyceride content was higher in Hint2(-/-) than in Hint2(+/+) livers, whereas plasma triglycerides were fourfold lower. Malonyl-CoA levels were increased twofold in Hint2(-/-) livers. After 24 h fasting, Hint2(-/-) displayed a decrease in body temperature, commensurate with a decrease in mass of brown fat and downregulation of uncoupling protein 1. HFD-treated Hint2(-/-) livers showed more steatosis, and plasma insulin and cholesterol were higher than in Hint(+/+) mice. Several proteins identified as substrates of sirtuin 3 and 5 and active in intermediary and ketone metabolism were hyperacetylated in liver and brown fat mitochondria after both HFD and fasting regimens. Glutamate dehydrogenase activity was downregulated in fed and fasted livers, and this was attributed to an increase in acetylation and ADP-ribosylation. The absence of Hint2 deregulates the posttranslational modification of several mitochondrial proteins, which impedes the adaptation to episodes of nutritional stress.

CLAP: a web-server for automatic classification of proteins with special reference to multi-domain proteins.

BACKGROUND: The function of a protein can be deciphered with higher accuracy from its structure than from its amino acid sequence. Due to the huge gap in the available protein sequence and structural space, tools that can generate functionally homogeneous clusters using only the sequence information, hold great importance. For this, traditional alignment-based tools work well in most cases and clustering is performed on the basis of sequence similarity. But, in the case of multi-domain proteins, the alignment quality might be poor due to varied lengths of the proteins, domain shuffling or circular permutations. Multi-domain proteins are ubiquitous in nature, hence alignment-free tools, which overcome the shortcomings of alignment-based protein comparison methods, are required. Further, existing tools classify proteins using only domain-level information and hence miss out on the information encoded in the tethered regions or accessory domains. Our method, on the other hand, takes into account the full-length sequence of a protein, consolidating the complete sequence information to understand a given protein better. RESULTS: Our web-server, CLAP (Classification of Proteins), is one such alignment-free software for automatic classification of protein sequences. It utilizes a pattern-matching algorithm that assigns local matching scores (LMS) to residues that are a part of the matched patterns between two sequences being compared. CLAP works on full-length sequences and does not require prior domain definitions.Pilot studies undertaken previously on protein kinases and immunoglobulins have shown that CLAP yields clusters, which have high functional and domain architectural similarity. Moreover, parsing at a statistically determined cut-off resulted in clusters that corroborated with the sub-family level classification of that particular domain family. CONCLUSIONS: CLAP is a useful protein-clustering tool, independent of domain assignment, domain order, sequence length and domain diversity. Our method can be used for any set of protein sequences, yielding functionally relevant clusters with high domain architectural homogeneity. The CLAP web server is freely available for academic use at http://nslab.mbu.iisc.ernet.in/clap/.

Disruption of tumor suppressor gene Hint1 leads to remodeling of the lipid metabolic phenotype of mouse liver.

A lipidomic and metabolomic investigation of serum and liver from mice was performed to gain insight into the tumor suppressor gene Hint1. A major reprogramming of lipid homeostasis was found in both serum and liver of Hint1-null (Hint(-/-)) mice, with significant changes in the levels of many lipid molecules, as compared with gender-, age-, and strain-matched WT mice. In the Hint1(-/-) mice, serum total and esterified cholesterol were reduced 2.5-fold, and lysophosphatidylcholines (LPCs) and lysophosphatidic acids were 10-fold elevated in serum, with a corresponding fall in phosphatidylcholines (PCs). In the liver, MUFAs and PUFAs, including arachidonic acid (AA) and its metabolic precursors, were also raised, as was mRNA encoding enzymes involved in AA de novo synthesis. There was also a significant 50% increase in hepatic macrophages in the Hint1(-/-) mice. Several hepatic ceramides and acylcarnitines were decreased in the livers of Hint1(-/-) mice. The changes in serum LPCs and PCs were neither related to hepatic phospholipase A2 activity nor to mRNAs encoding lysophosphatidylcholine acetyltransferases 1-4. The lipidomic phenotype of the Hint1(-/-) mouse revealed decreased inflammatory eicosanoids with elevated proliferative mediators that, combined with decreased ceramide apoptosis signaling molecules, may contribute to the tumor suppressor activity of Hint1.

Benchmarking protein-protein interface predictions: why you should care about protein size.

A number of predictive methods have been developed to predict protein-protein binding sites. Each new method is traditionally benchmarked using sets of protein structures of various sizes, and global statistics are used to assess the quality of the prediction. Little attention has been paid to the potential bias due to protein size on these statistics. Indeed, small proteins involve proportionally more residues at interfaces than large ones. If a predictive method is biased toward small proteins, this can lead to an over-estimation of its performance. Here, we investigate the bias due to the size effect when benchmarking protein-protein interface prediction on the widely used docking benchmark 4.0. First, we simulate random scores that favor small proteins over large ones. Instead of the 0.5 AUC (Area Under the Curve) value expected by chance, these biased scores result in an AUC equal to 0.6 using hypergeometric distributions, and up to 0.65 using constant scores. We then use real prediction results to illustrate how to detect the size bias by shuffling, and subsequently correct it using a simple conversion of the scores into normalized ranks. In addition, we investigate the scores produced by eight published methods and show that they are all affected by the size effect, which can change their relative ranking. The size effect also has an impact on linear combination scores by modifying the relative contributions of each method. In the future, systematic corrections should be applied when benchmarking predictive methods using data sets with mixed protein sizes.

Disruption of the histidine triad nucleotide-binding hint2 gene in mice affects glycemic control and mitochondrial function.

UNLABELLED: The histidine triad nucleotide-binding (HINT2) protein is a mitochondrial adenosine phosphoramidase expressed in the liver and pancreas. Its physiological function is unknown. To elucidate the role of HINT2 in liver physiology, the mouse Hint2 gene was deleted. Hint2(-/-) and Hint2(+/+) mice were generated in a mixed C57Bl6/J × 129Sv background. At 20 weeks, the phenotypic changes in Hint2(-/-) relative to Hint2(+/+) mice were an accumulation of hepatic triglycerides, decreased tolerance to glucose, a defective counter-regulatory response to insulin-provoked hypoglycemia, and an increase in plasma interprandial insulin but a decrease in glucose-stimulated insulin secretion and defective thermoregulation upon fasting. Leptin messenger RNA (mRNA) in adipose tissue and plasma leptin were elevated. In mitochondria from Hint2(-/-) hepatocytes, state 3 respiration was decreased, a finding confirmed in HepG2 cells where HINT2 mRNA was silenced. The linked complex II-III electron transfer was decreased in Hint2(-/-) mitochondria, which was accompanied by a lower content of coenzyme Q. Hypoxia-inducible factor-2α expression and the generation of reactive oxygen species were increased. Electron microscopy of mitochondria in Hint2(-/-) mice aged 12 months revealed clustered, fused organelles. The hepatic activities of 3-hydroxyacyl-coenzyme A dehydrogenase short chain and glutamate dehydrogenase (GDH) were decreased by 68% and 60%, respectively, without a change in protein expression. GDH activity was similarly decreased in HINT2-silenced HepG2 cells. When measured in the presence of purified sirtuin 3, latent GDH activity was recovered (126% in Hint2(-/-) versus 83% in Hint2(+/+) ). This suggests a greater extent of acetylation in Hint2(-/-) than in Hint2(+/+) . CONCLUSION: Hint2/HINT2 positively regulates mitochondrial lipid metabolism and respiration and glucose homeostasis. The absence of Hint2 provokes mitochondrial deformities and a change in the pattern of acetylation of selected proteins.