Structure-aware machine learning strategies for antimicrobial peptide discovery.

Machine learning models are revolutionizing our approaches to discovering and designing bioactive peptides. These models often need protein structure awareness, as they heavily rely on sequential data. The models excel at identifying sequences of a particular biological nature or activity, but they frequently fail to comprehend their intricate mechanism(s) of action. To solve two problems at once, we studied the mechanisms of action and structural landscape of antimicrobial peptides as (i) membrane-disrupting peptides, (ii) membrane-penetrating peptides, and (iii) protein-binding peptides. By analyzing critical features such as dipeptides and physicochemical descriptors, we developed models with high accuracy (86-88%) in predicting these categories. However, our initial models (1.0 and 2.0) exhibited a bias towards α-helical and coiled structures, influencing predictions. To address this structural bias, we implemented subset selection and data reduction strategies. The former gave three structure-specific models for peptides likely to fold into α-helices (models 1.1 and 2.1), coils (1.3 and 2.3), or mixed structures (1.4 and 2.4). The latter depleted over-represented structures, leading to structure-agnostic predictors 1.5 and 2.5. Additionally, our research highlights the sensitivity of important features to different structure classes across models.

Helical Content Correlations and Hydration Structures of the Folding Ensemble of the B Domain of Protein A.

The B domain of protein A (BdpA), a small three-helix bundle, folds on a time scale of a few microseconds with heterogeneous native and unfolded states. It is widely used as a model for understanding protein folding mechanisms. In this work, we use structure-based models (SBMs) and atomistic simulations to comprehensively investigate how BdpA folding is associated with the formation of its secondary structure. The energy landscape visualization method (ELViM) was used to characterize the pathways that connect the folded and unfolded states of BdpA as well as the sets of structures displaying specific ellipticity patterns. We show that the native state conformational diversity is due mainly to the conformational variability of helix I. Helices I, II, and III occur in a weakly correlated manner, with Spearman's rank correlation coefficients of 0.1539 (I and II), 0.1259 (I and III), and 0.2561 (II and III). These results, therefore, suggest the highest cooperativity between helices II and III. Our results allow the clustering of partially folded structures of folding of the B domain of protein A on the basis of its secondary structure, paving the way to an understanding of environmental factors in the relative stability of the basins of the folding ensemble, which are illustrated by the structural dependency of the protein hydration structures, as computed with minimum-distance distribution functions.

The 1H, 15N and 13C resonance assignments of dengue virus capsid protein with the deletion of the intrinsically disordered N-terminal region.

Dengue virus belongs to the Flaviviridae family, being responsible for an endemic arboviral disease in humans. It is an enveloped virus, whose genome is a positive-stranded RNA packaged by the capsid protein. Dengue virus capsid protein (DENVC) forms homodimers in solution organized in 4 α-helices and an intrinsically disordered N-terminal region. The N-terminal region is involved in the binding of membranous structures in host cells and in the recognition of nucleotides. Here we report the 1H, 15N and 13C resonance assignments of the DENVC with the deletion of the first 19 intrinsically disordered residues. The backbone chemical shift perturbations suggest changes in the α1 and α2 helices between full length and the truncated proteins.

Antimicrobial Spectrum of Activity and Mechanism of Action of Linear Alpha-Helical Peptides Inspired by Shrimp Anti-Lipopolysaccharide Factors.

Shrimp antilipopolysaccharide factors (ALFs) form a multifunctional and diverse family of antimicrobial host defense peptides (AMPs) composed of seven members (groups A to G), which differ in terms of their primary structure and biochemical properties. They are amphipathic peptides with two conserved cysteine residues stabilizing a central ß-hairpin that is understood to be the core region for their biological activities. In this study, we synthetized three linear (cysteine-free) peptides based on the amino acid sequence of the central ß-hairpin of the newly identified shrimp (Litopenaeus vannamei) ALFs from groups E to G. Unlike whole mature ALFs, the ALF-derived peptides exhibited an α-helix secondary structure. In vitro assays revealed that the synthetic peptides display a broad spectrum of activity against both Gram-positive and Gram-negative bacteria and fungi but not against the protozoan parasites Trypanosoma cruzi and Leishmania (L.) infantum. Remarkably, they displayed synergistic effects and showed the ability to permeabilize bacterial membranes, a mechanism of action of classical AMPs. Having shown low cytotoxicity to THP-1 human cells and being active against clinical multiresistant bacterial isolates, these nature-inspired peptides represent an interesting class of bioactive molecules with biotechnological potential for the development of novel therapeutics in medical sciences.

Engineering defensin α-helix to produce high-affinity SARS-CoV-2 spike protein binding ligands.

The binding of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike protein to the angiotensin-converting enzyme 2 (ACE2) receptor expressed on the host cells is a critical initial step for viral infection. This interaction is blocked through competitive inhibition by soluble ACE2 protein. Therefore, developing high-affinity and cost-effective ACE2 mimetic ligands that disrupt this protein-protein interaction is a promising strategy for viral diagnostics and therapy. We employed human and plant defensins, a class of small (2-5 kDa) and highly stable proteins containing solvent-exposed alpha-helix, conformationally constrained by two disulfide bonds. Therefore, we engineered the amino acid residues on the constrained alpha-helix of defensins to mimic the critical residues on the ACE2 helix 1 that interact with the SARS-CoV-2 spike protein. The engineered proteins (h-deface2, p-deface2, and p-deface2-MUT) were soluble and purified to homogeneity with a high yield from a bacterial expression system. The proteins demonstrated exceptional thermostability (Tm 70.7°C), high-affinity binding to the spike protein with apparent Kd values of 54.4 ± 11.3, 33.5 ± 8.2, and 14.4 ± 3.5 nM for h-deface2, p-deface2, and p-deface2-MUT, respectively, and were used in a diagnostic assay that detected SARS-CoV-2 neutralizing antibodies. This work addresses the challenge of developing helical ACE2 mimetics by demonstrating that defensins provide promising scaffolds to engineer alpha-helices in a constrained form for designing of high-affinity ligands.

Computational study, synthesis and evaluation of active peptides derived from Parasporin-2 and spike protein from Alphacoronavirus against colorectal cancer cells.

Parasporin-2Aa1 (PS2Aa1) is a toxic protein of 37 KDa (30 kDa, activated form produced by proteolysis) that was shown to be cytotoxic against specific human cancer cells, although its mechanism of action has not been elucidated yet. In order to study the role of some native peptide fragments of proteins on anticancer activity, here we investigated the cytotoxic effect of peptide fragments from domain-1 of PS2Aa1 and one of the loops present in the binding region of the virus spike protein from Alphacoronavirus (HCoV-229E), the latter according to scientific reports, who showed interaction with the human APN (h-APN) receptor, evidence corroborated through computational simulations, and thus being possible active against colon cancer cells. Peptides namely P264-G274, Loop1-PS2Aa, and Loop2-PS2Aa were synthesized using the Fmoc solid-phase synthesis and characterized by mass spectrometry (MS). Additionally, one region from loop 1 of HCoV-229E, Loop1-HCoV-229E, was also synthesized and characterized. The A4W-GGN5 anticancer peptide and 5-fluorouracil (5-FU) were taken as a control in all experiments. Circular dichroism revealed an α-helix structure for the peptides derived from PS2Aa1 (P264-G274, Loop1-PS2Aa, and Loop2-PS2Aa) and ß-laminar structure for the peptide derived from Alphacoronavirus spike protein Loop1-HCoV-229E. Peptides showed a hemolysis percentage of less than 20% at 100 µM concentration. Besides, peptides exhibited stronger anticancer activity against SW480 and SW620 cells after exposure for 48 h. Likewise, these compounds showed significantly lower toxicity against normal cells CHO-K1. The results suggest that native peptide fragments from Ps2Aa1 may be optimized as a novel potential cancer-therapeutic agents.

Interaction of Spike protein and lipid membrane of SARS-CoV-2 with Ursodeoxycholic acid, an in-silico analysis.

Numerous repositioned drugs have been sought to decrease the severity of SARS-CoV-2 infection. It is known that among its physicochemical properties, Ursodeoxycholic Acid (UDCA) has a reduction in surface tension and cholesterol solubilization, it has also been used to treat cholesterol gallstones and viral hepatitis. In this study, molecular docking was performed with the SARS-CoV-2 Spike protein and UDCA. In order to confirm this interaction, we used Molecular Dynamics (MD) in "SARS-CoV-2 Spike protein-UDCA". Using another system, we also simulated MD with six UDCA residues around the Spike protein at random, naming this "SARS-CoV-2 Spike protein-6UDCA". Finally, we evaluated the possible interaction between UDCA and different types of membranes, considering the possible membrane conformation of SARS-CoV-2, this was named "SARS-CoV-2 membrane-UDCA". In the "SARS-CoV-2 Spike protein-UDCA", we found that UDCA exhibits affinity towards the central region of the Spike protein structure of - 386.35 kcal/mol, in a region with 3 alpha helices, which comprises residues from K986 to C1032 of each monomer. MD confirmed that UDCA remains attached and occasionally forms hydrogen bonds with residues R995 and T998. In the presence of UDCA, we observed that the distances between residues atoms OG1 and CG2 of T998 in the monomers A, B, and C in the prefusion state do not change and remain at 5.93 ± 0.62 and 7.78 ± 0.51 Å, respectively, compared to the post-fusion state. Next, in "SARS-CoV-2 Spike protein-6UDCA", the three UDCA showed affinity towards different regions of the Spike protein, but only one of them remained bound to the region between the region's heptad repeat 1 and heptad repeat 2 (HR1 and HR2) for 375 ps of the trajectory. The RMSD of monomer C was the smallest of the three monomers with a value of 2.89 ± 0.32, likewise, the smallest RMSF was also of the monomer C (2.25 ± 056). In addition, in the simulation of "SARS-CoV-2 membrane-UDCA", UDCA had a higher affinity toward the virion-like membrane; where three of the four residues remained attached once they were close (5 Å, to the centre of mass) to the membrane by 30 ns. However, only one of them remained attached to the plasma-like membrane and this was in a cluster of cholesterol molecules. We have shown that UDCA interacts in two distinct regions of Spike protein sequences. In addition, UDCA tends to stay bound to the membrane, which could potentially reduce the internalization of SARS-CoV-2 in the host cell.

Protease Inhibitors as Promising Weapons against COVID-19: Focus on Repurposing of Drugs used to Treat HIV and HCV Infections.

As a part of the efforts to quickly develop pharmaceutical treatments for COVID-19 through repurposing existing drugs, some researchers around the world have combined the recently released crystal structure of SARS-CoV-2 Mpro in complex with a covalently bonded inhibitor with virtual screening procedures employing molecular docking approaches. In this context, protease inhibitors (PIs) clinically available and currently used to treat infectious diseases, particularly viral ones, are relevant sources of promising drug candidates to inhibit the SARS-CoV-2 Mpro, a key viral enzyme involved in crucial events during its life cycle. In the present perspective, we summarized the published studies showing the promising use of HIV and HCV PIs as potential repurposing drugs against the SARS-CoV-2 Mpro.

DFT calculations to investigate silver ions as a virucide from SARS-CoV-2.

The world has face the COVID-19 pandemic which has already caused millions of death. Due to the urgency in fighting the virus, we study five residues of free amino acids present in the structure of the SARS-CoV-2 spike protein (S). We investigated the spontaneous interaction between amino acids and silver ions (Ag+), considering these ions as a virucide chemical agent for SARS-CoV-2. The amino acid-Ag+ systems were investigated in a gaseous medium and a simulated water environment was described with a continuum model (PCM) the calculations were performed within the framework of density functional theory (DFT). Calculations related to the occupied orbitals of higher energy showed that Ag+ has a tendency to interact with the nitrile groups (-NH). The negative values of the Gibbs free energies show that the interaction process between amino acids-Ag+ in both media occurs spontaneously. There is a decrease in Gibbs free energy from the amino acid-Ag+ interactions immersed in a water solvation simulator.

Crystallographic approach to fragment-based hit discovery against Schistosoma mansoni purine nucleoside phosphorylase.

Several Schistosoma species cause Schistosomiasis, an endemic disease in 78 countries that is ranked second amongst the parasitic diseases in terms of its socioeconomic impact and human health importance. The drug recommended for treatment by the WHO is praziquantel (PZQ), but there are concerns associated with PZQ, such as the lack of information about its exact mechanism of action, its high price, its effectiveness - which is limited to the parasite's adult form - and reports of resistance. The parasites lack the de novo purine pathway, rendering them dependent on the purine salvage pathway or host purine bases for nucleotide synthesis. Thus, the Schistosoma purine salvage pathway is an attractive target for the development of necessary and selective new drugs. In this study, the purine nucleotide phosphorylase II (PNP2), a new isoform of PNP1, was submitted to a high-throughput fragment-based hit discovery using a crystallographic screening strategy. PNP2 was crystallized and crystals were soaked with 827 fragments, a subset of the Maybridge 1000 library. X-ray diffraction data was collected and structures were solved. Out of 827-screened fragments we have obtained a total of 19 fragments that show binding to PNP2. Fourteen of these fragments bind to the active site of PNP2, while five were observed in three other sites. Here we present the first fragment screening against PNP2.

Biochemical characterization of LysVpKK5 endolysin from a marine vibriophage.

Endolysins have been proposed as a potential antibacterial alternative for aquaculture, especially against Vibrio; the bacterial-agents that most frequently cause disease. Although multiple marine vibriophages have been characterized to date, research on vibriophage endolysins is recent. In this study, biochemical characterization of LysVpKK5 endolysin encoded by Vibrio parahaemolyticus-infecting VpKK5 phage was performed. In silico analysis revealed that LysVpKK5 possesses a conserved amidase_2 domain with a zinc-binding motif of high structural similarity to T7 lysozyme (RMSD = 0.107 Å). Contrary to expectations, the activity was inhibited with Zn2+ and was improved with other divalent cations, especially Ca2+. It showed optimal muralytic activity at pH 10, and curiously, no lytic activity at pH ≤ 7 was recorded. As for the thermal stability test, the optimal activity was recorded at 30 °C; the higher residual activity was recorded at 4 °C, and was lost at ≥ 50 °C. On the other hand, increasing NaCl concentrations reduced the activity gradually; the optimal activity was recorded at 50 mM NaCl. On the other hand, the enzymatic activity at 0.5 M NaCl was approx 30% and of approx 50% in seawater. LysVpKK5 endolysin exhibited a higher activity on V. parahaemolyticus ATCC-17802 strain, in comparison with AHPND + strains.

An atomic model for the human septin hexamer by cryo-EM.

In order to form functional filaments, human septins must assemble into hetero-oligomeric rod-like particles which polymerize end-to-end. The rules governing the assembly of these particles and the subsequent filaments are incompletely understood. Although crystallographic approaches have been successful in studying the separate components of the system, there has been difficulty in obtaining high resolution structures of the full particle. Here we report a first cryo-EM structure for a hexameric rod composed of human septins 2, 6 and 7 with a global resolution of ~3.6 Å and a local resolution of between ~3.0 Å and ~5.0 Å. By fitting the previously determined high-resolution crystal structures of the component subunits into the cryo-EM map, we are able to provide an essentially complete model for the particle. This exposes SEPT2 NC-interfaces at the termini of the hexamer and leaves internal cavities between the SEPT6-SEPT7 pairs. The floor of the cavity is formed by the two α0 helices including their polybasic regions. These are locked into place between the two subunits by interactions made with the α5 and α6 helices of the neighbouring monomer together with its polyacidic region. The cavity may serve to provide space allowing the subunits to move with respect to one another. The elongated particle shows a tendency to bend at its centre where two copies of SEPT7 form a homodimeric G-interface. Such bending is almost certainly related to the ability of septin filaments to recognize and even induce membrane curvature.

New tyrosinases with putative action against contaminants of emerging concern.

Tyrosinases (EC 1.14.18.1) are type-3 copper metalloenzymes with strong oxidative capacities and low allosteric selectivity to phenolic and non-phenolic aromatic compounds, which have been used as biosensors and biocatalysts to mitigate the impacts of environmental contaminants over aquatic ecosystems. However, the widespread use of these polyphenol oxidases is limited by elevated production costs and restricted knowledge on their spectrum of action. Here, six tyrosinase homologs were identified and characterized from the genomes of four widespread freshwater ciliates using bioinformatics. Next, we performed a virtual screening to calculate binding energies between 3D models of these homologs and ~ 1000 contaminants of emerging concern (CECs), as an indirect approach to identify likely and unlikely targets for tyrosinases. Many fine chemicals, pharmaceuticals, personal care products, illicit drugs, natural toxins, and pesticides exhibited strong binding energies to these new tyrosinases, suggesting the spectrum of targets of these enzymes might be considerably broader than previously thought. Many ciliates, including those carrying tyrosinase genes, are fast-growing unicellular microeukaryotes that can be efficiently cultured, at large scales, under in vitro conditions, suggesting these organisms should be regarded as potential low-cost sources of new environmental biotechnological molecules.

Cloning, purification, and characterization of the 6-phosphogluconate dehydrogenase (6 PGDH) from Giardia lamblia.

Giardia lamblia, due to the habitat in which it develops, requires a continuous supply of intermediate compounds that allow it to survive in the host. The pentose phosphate pathway (PPP) provides essential molecules such as NADPH and ribulose-5-phosphate during the oxidative phase of the pathway. One of the key enzymes during this stage is 6-phosphogluconate dehydrogenase (6 PGDH) for generating NADPH. Given the relevance of the enzyme, in the present work, the 6pgdh gene from G. lamblia was amplified and cloned to produce the recombinant protein (Gl-6 PGDH) and characterize it functionally and structurally after the purification of Gl-6 PGDH by affinity chromatography. The results of the characterization showed that the protein has a molecular mass of 54 kDa, with an optimal pH of 7.0 and a temperature of 36-42 °C. The kinetic parameters of Gl-6 PGDH were Km = 49.2 and 139.9 µM (for NADP+ and 6-PG, respectively), Vmax =26.27 µmol*min-1*mg-1, and Kcat = 24.0 s-1. Finally, computational modeling studies were performed to obtain a structural visualization of the Gl-6 PGDH protein. The generation of the model and the characterization assays will allow us to expand our knowledge for future studies of the function of the protein in the metabolism of the parasite.

Structure of an ancestral ADP-dependent kinase with fructose-6P reveals key residues for binding, catalysis, and ligand-induced conformational changes.

ADP-dependent kinases were first described in archaea, although their presence has also been reported in bacteria and eukaryotes (human and mouse). This enzyme family comprises three substrate specificities; specific phosphofructokinases (ADP-PFKs), specific glucokinases (ADP-GKs), and bifunctional enzymes (ADP-PFK/GK). Although many structures are available for members of this family, none exhibits fructose-6-phosphate (F6P) at the active site. Using an ancestral enzyme, we obtain the first structure of an ADP-dependent kinase (AncMsPFK) with F6P at its active site. Key residues for sugar binding and catalysis were identified by alanine scanning, D36 being a critical residue for F6P binding and catalysis. However, this residue hinders glucose binding because its mutation to alanine converts the AncMsPFK enzyme into a specific ADP-GK. Residue K179 is critical for F6P binding, while residues N181 and R212 are also important for this sugar binding, but to a lesser extent. This structure also provides evidence for the requirement of both substrates (sugar and nucleotide) to accomplish the conformational change leading to a closed conformation. This suggests that AncMsPFK mainly populates two states (open and closed) during the catalytic cycle, as reported for specific ADP-PFK. This situation differs from that described for specific ADP-GK enzymes, where each substrate independently causes a sequential domain closure, resulting in three conformational states (open, semiclosed, and closed).

Naphthalimide-Containing BP100 Leads to Higher Model Membranes Interactions and Antimicrobial Activity.

In a large variety of organisms, antimicrobial peptides (AMPs) are primary defenses against pathogens. BP100 (KKLFKKILKYL-NH2), a short, synthetic, cationic AMP, is active against bacteria and displays low toxicity towards eukaryotic cells. BP100 acquires a α-helical conformation upon interaction with membranes and increases membrane permeability. Despite the volume of information available, the action mechanism of BP100, the selectivity of its biological effects, and possible applications are far from consensual. Our group synthesized a fluorescent BP100 analogue containing naphthalimide linked to its N-terminal end, NAPHT-BP100 (Naphthalimide-AAKKLFKKILKYL-NH2). The fluorescence properties of naphthalimides, especially their spectral sensitivity to microenvironment changes, are well established, and their biological activities against transformed cells and bacteria are known. Naphthalimide derived compounds are known to interact with DNA disturbing related processes as replication and transcription, and used as anticancer agents due to this property. A wide variety of techniques were used to demonstrate that NAPHT-BP100 bound to and permeabilized zwitterionic POPC and negatively charged POPC:POPG liposomes and, upon interaction, acquired a α-helical structure. Membrane surface high peptide/lipid ratios triggered complete permeabilization of the liposomes in a detergent-like manner. Membrane disruption was driven by charge neutralization, lipid aggregation, and bilayer destabilization. NAPHT-BP100 also interacted with double-stranded DNA, indicating that this peptide could also affect other cellular processes besides causing membrane destabilization. NAPHT-BP100 showed increased antibacterial and hemolytic activities, compared to BP100, and may constitute an efficient antimicrobial agent for dermatological use. By conjugating BP100 and naphthalimide DNA binding properties, NAPHT-BP100 bound to a large extent to the bacterial membrane and could more efficiently destabilize it. We also speculate that peptide could enter the bacteria cell and interact with its DNA in the cytoplasm.

Can molecular mimicry explain the cytokine storm of SARS-CoV-2?: An in silico approach.

PARP14 and PARP9 play a key role in macrophage immune regulation. SARS-CoV-2 is an emerging viral disease that triggers hyper-inflammation known as a cytokine storm. In this study, using in silico tools, we hypothesize about the immunological phenomena of molecular mimicry between SARS-CoV-2 Nsp3 and the human PARP14 and PARP9. The results showed an epitope of SARS-CoV-2 Nsp3 protein that contains consensus sequences for both human PARP14 and PARP9 that are antigens for MHC Classes 1 and 2, which can potentially induce an immune response against human PARP14 and PARP9; while its depletion causes a hyper-inflammatory state in SARS-CoV-2 patients.

Structural features of the plant N-recognin ClpS1 and sequence determinants in its targets that govern substrate selection.

In the N-degron pathway of protein degradation of Escherichia coli, the N-recognin ClpS identifies substrates bearing N-terminal phenylalanine, tyrosine, tryptophan, or leucine and delivers them to the caseinolytic protease (Clp). Chloroplasts contain the Clp system, but whether chloroplastic ClpS1 adheres to the same constraints is unknown. Moreover, the structural underpinnings of substrate recognition are not completely defined. We show that ClpS1 recognizes canonical residues of the E. coli N-degron pathway. The residue in second position influences recognition (especially in N-terminal ends starting with leucine). N-terminal acetylation abrogates recognition. ClpF, a ClpS1-interacting partner, does not alter its specificity. Substrate binding provokes local remodeling of residues in the substrate-binding cavity of ClpS1. Our work strongly supports the existence of a chloroplastic N-degron pathway.

Cassava (Manihot esculenta) defensins: Prospection, structural analysis and tissue-specific expression under biotic/abiotic stresses.

Defensins are a prominent family of antimicrobial peptides. They play sophisticated roles in the defense against pathogens in all living organisms, but few works address their expression under different conditions and plant tissues. The present work prospected defensins of Manihot esculenta Crantz, popularly known as cassava. Five defensin candidates (MeDefs) were retrieved from the genome sequences and characterized. Considering chromosome distribution, only MeDef1 and 2 occupy adjacent positions in the same chromosome arm. All 3D structures had antiparallel ß-sheets, an α-helix, and amphipathic residues distributed throughout the peptides with a predominance of cationic surface charge. MeDefs expression was validated by RT-qPCR, including two stress types (biotic: fungus Macrophomina pseudophaseolina, and abiotic: mechanical injury) and a combination of both stresses (fungus+injury) in three different tissues (root, stem, and leaf). For this purpose, ten reference genes (RGs) were tested, and three were chosen to characterize MeDef expression. MeDef3 was up-regulated at roots in all stress situations tested. MeDef1 and MeDef5 were induced in leaves under biotic and abiotic stresses, but not in both stress types simultaneously. Only MeDef2 was down-regulated in the stem tissue also with biotic/abiotic combined stresses. These results indicate that although defensins are known to be responsive to pathogen infection, they may act as preformed defense or, still, have tissue or stress specificities. Aspects of their structure, stability and evolution are also discussed.

Unique structural features of flaviviruses' capsid proteins: new insights on structure-function relationship.

The Flaviviridae family comprises important human pathogens, including Dengue, Zika, West Nile, Yellow Fever and Japanese Encephalitis viruses. The viral genome, a positive-sense single-stranded RNA, is packaged by a single protein, the capsid protein, which is a small and highly basic protein that form intertwined homodimers in solution. Atomic-resolution structures of four flaviviruses capsid proteins were solved either in solution by nuclear magnetic resonance spectroscopy, or after protein crystallization by X-ray diffraction. Analyses of these structures revealed very particular properties, namely (i) the predominance of quaternary contacts maintaining the structure; (ii) a highly electropositive surface throughout the protein; and (iii) a flexible helix (α1). The goal of this review is to discuss the role of these features in protein structure-function relationship.