Curvature of designed armadillo repeat proteins allows modular peptide binding

Designed armadillo repeat proteins (dArmRPs) were developed to create a modular peptide binding technology where each of the structural repeats binds two residues of the target peptide. An essential prerequisite for such a technology is a dArmRP geometry that matches the peptide bond length. To this end, we determined a large set (n = 27) of dArmRP X-ray structures, of which 12 were previously unpublished, to calculate curvature parameters that define their geometry. Our analysis shows that consensus dArmRPs exhibit curvatures close to the optimal range for modular peptide recognition. Binding of peptide ligands can induce a curvature within the desired range, as confirmed by single-molecule FRET experiments in solution. On the other hand, computationally designed ArmRPs, where side chains have been chosen with the intention to optimally fit into a geometrically optimized backbone, turned out to be more divergent in reality, and thus not suitable for continuous peptide binding. Furthermore, we show that the formation of a crystal lattice can induce small but significant deviations from the curvature adopted in solution, which can interfere with the evaluation of repeat protein scaffolds when high accuracy is required. This study corroborates the suitability of consensus dArmRPs as a scaffold for the development of modular peptide binders.     

Structures of designed armadillo repeat proteins binding to peptides fused to globular domains

Designed armadillo repeat proteins (dArmRP) are α‐helical solenoid repeat proteins with an extended peptide binding groove that were engineered to develop a generic modular technology for peptide recognition. In this context, the term “peptide” not only denotes a short unstructured chain of amino acids, but also an unstructured region of a protein, as they occur in termini, loops, or linkers between folded domains. Here we report two crystal structures of dArmRPs, in complex with peptides fused either to the N‐terminus of Green Fluorescent Protein or to the C‐terminus of a phage lambda protein D. These structures demonstrate that dArmRPs bind unfolded peptides in the intended conformation also when they constitute unstructured parts of folded proteins, which greatly expands possible applications of the dArmRP technology. Nonetheless, the structures do not fully reflect the binding behavior in solution, that is, some binding sites remain unoccupied in the crystal and even unexpected peptide residues appear to be bound. We show how these differences can be explained by restrictions of the crystal lattice or the composition of the crystallization solution. This illustrates that crystal structures have to be interpreted with caution when protein–peptide interactions are characterized, and should always be correlated with measurements in solution.     

Characterization and further stabilization of designed ankyrin repeat proteins by combining molecular dynamics simulations and experiments

Multiple molecular dynamics simulations with explicit solvent at room temperature and at 400 K were carried out to characterize designed ankyrin repeat (AR) proteins with full-consensus repeats. Using proteins with one to five repeats, the stability of the native structure was found to increase with the number of repeats. The C-terminal capping repeat, originating from the natural guanine-adenine-binding protein, was observed to denature first in almost all high-temperature simulations. Notably, a stable intermediate is found in experimental equilibrium unfolding studies of one of the simulated consensus proteins. On the basis of simulation results, this intermediate is interpreted to represent a conformation with a denatured C-terminal repeat. To validate this interpretation, constructs without C-terminal capping repeat were prepared and did not show this intermediate in equilibrium unfolding experiments. Conversely, the capping repeats were found to be essential for efficient folding in the cell and for avoiding aggregation, presumably because of their highly charged surface. To design a capping repeat conferring similar solubility properties yet even higher stability, eight point mutations adapting the C-cap to the consensus AR and adding a three-residue extension at the C-terminus were predicted in silico and validated experimentally. The in vitro full-consensus proteins were also compared with a previously published designed AR protein, E3_5, whose internal repeats show 80% identity in primary sequence. A detailed analysis of the simulations suggests that networks of salt bridges between β-hairpins, as well as additional interrepeat hydrogen bonds, contribute to the extraordinary stability of the full consensus.     

Crystal structures of designed armadillo repeat proteins: Implications of construct design and crystallization conditions on overall structure

Designed armadillo repeat proteins (dArmRP) are promising modular proteins for the engineering of binding molecules that recognize extended polypeptide chains. We determined the structure of a dArmRP containing five internal repeats and 3rd generation capping repeats in three different states by X‐ray crystallography: without N‐terminal His₆‐tag and in the presence of calcium (YM₅A/Ca²⁺), without N‐terminal His₆‐tag and in the absence of calcium (YM₅A), and with N‐terminal His₆‐tag and in the presence of calcium (His‐YM₅A/Ca²⁺). All structures show different quaternary structures and superhelical parameters. His‐YM₅A/Ca²⁺ forms a crystallographic dimer, which is bridged by the His₆‐tag, YM₅A/Ca²⁺ forms a domain‐swapped tetramer, and only in the absence of calcium and the His₆‐tag, YM₅A forms a monomer. The changes of superhelical parameters are a consequence of calcium binding, because calcium ions interact with negatively charged residues, which can also participate in the modulation of helix dipole moments between adjacent repeats. These observations are important for further optimizations of dArmRPs and provide a general illustration of how construct design and crystallization conditions can influence the exact structure of the investigated protein.     

Molecular dynamics study of the stabilities of consensus designed ankyrin repeat proteins

Two designed ankyrin repeat (AR) proteins (E3_5 and E3_19) are high homologous (with about 87% sequence identity) and their crystal structures have a Calpha atom-positional root-mean-square difference of about 0.14 nm. However, it was found that E3_5 is considerably more stable than E3_19 in guanidinium hydrochloride and thermal denaturation experiments. With the goal of providing insights into the various factors contributing to the stabilities of the designed AR proteins and suggesting possible mutations to enhance their stabilities, homology modeling and molecular dynamics (MD) simulations with explicit solvent have been performed. Because the crystal structure of E3_19 was solved later than that of E3_5, a homology model of E3_19 based on the crystal structure of E3_5 was also used in the simulations. E3_5 shows a very stable trajectory in both crystal and solution simulations. In contrast, the C-terminal repeat of E3_19 unfolds in the simulations starting from either the modeled structure or the crystal structure, although it has a sequence identical to that of E3_5. A continuum electrostatic model was used to estimate the effect of single mutations on protein stability and to study the interaction between the internal ARs and the C-terminal capping AR. Mutations involving charged residues were found to have large effects on stability. Due to the difference in charge distribution in the internal ARs of E3_19 and E3_5, their interaction with the C-terminal capping AR is less favorable in E3_19. The simulation trajectories suggest that the stability of the designed AR proteins can be increased by optimizing the electrostatic interactions within and between the different repeats.     

Evaluation of a fast implicit solvent model for molecular dynamics simulations.

A solvation term based on the solvent accessible surface area (SASA) is combined with the CHARMM polar hydrogen force field for the efficient simulation of peptides and small proteins in aqueous solution. Only two atomic solvation parameters are used: one is negative for favoring the direct solvation of polar groups and the other positive for taking into account the hydrophobic effect on apolar groups. To approximate the water screening effects on the intrasolute electrostatic interactions, a distance-dependent dielectric function is used and ionic side chains are neutralized. The use of an analytical approximation of the SASA renders the model extremely efficient (i.e., only about 50% slower than in vacuo simulations). The limitations and range of applicability of the SASA model are assessed by simulations of proteins and structured peptides. For the latter, the present study and results reported elsewhere show that with the SASA model it is possible to sample a significant amount of folding/unfolding transitions, which permit the study of the thermodynamics and kinetics of folding at an atomic level of detail.     

Structure refinement of membrane proteins via molecular dynamics simulations

A refinement protocol based on physics‐based techniques established for water soluble proteins is tested for membrane protein structures. Initial structures were generated by homology modeling and sampled via molecular dynamics simulations in explicit lipid bilayer and aqueous solvent systems. Snapshots from the simulations were selected based on scoring with either knowledge‐based or implicit membrane‐based scoring functions and averaged to obtain refined models. The protocol resulted in consistent and significant refinement of the membrane protein structures similar to the performance of refinement methods for soluble proteins. Refinement success was similar between sampling in the presence of lipid bilayers and aqueous solvent but the presence of lipid bilayers may benefit the improvement of lipid‐facing residues. Scoring with knowledge‐based functions (DFIRE and RWplus) was found to be as good as scoring using implicit membrane‐based scoring functions suggesting that differences in internal packing is more important than orientations relative to the membrane during the refinement of membrane protein homology models.     

Folding of a designed simple ankyrin repeat protein

Ankyrin repeats (AR) are 33-residue motifs containing a beta-turn, followed by two alpha-helices connected by a loop. AR occur in tandem arrangements and stack side-by-side to form elongated domains involved in very different cellular tasks. Recently, consensus libraries of AR repeats were constructed. Protein E1_5 represents a member of the shortest library, and consists of only a single consensus repeat flanked by designed N- and C-terminal capping repeats. Here we present a biophysical characterization of this AR domain. The protein is compactly folded, as judged from the heat capacity of the native state and from the specific unfolding enthalpy and entropy. From spectroscopic data, thermal and urea-induced unfolding can be modeled by a two-state transition. However, scanning calorimetry experiments reveal a deviation from the two-state behavior at elevated temperatures. Folding and unfolding at 5 degrees C both follow monoexponential kinetics with k(folding) = 28 sec(-1) and k(unfolding) = 0.9 sec(-1). Kinetic and equilibrium unfolding parameters at 5 degrees C agree very well. We conclude that E1_5 folds in a simple two-state manner at low temperatures while equilibrium intermediates become populated at higher temperatures. A chevron-plot analysis indicates that the protein traverses a very compact transition state along the folding/unfolding pathway. This work demonstrates that a designed minimal ankyrin repeat protein has the thermodynamic and kinetic properties of a compactly folded protein, and explains the favorable properties of the consensus framework.     

Calorimetric study of a series of designed repeat proteins: modular structure and modular folding

Repeat proteins comprise tandem arrays of a small structural motif. Their structure is defined and stabilized by interactions between residues that are close in the primary sequence. Several studies have investigated whether their structural modularity translates into modular thermodynamic properties. Tetratricopeptide repeat proteins (TPRs) are a class in which the repeated unit is a 34 amino acid helix-turn-helix motif. In this work, we use differential scanning calorimetry (DSC) to study the equilibrium stability of a series of TPR proteins with different numbers of an identical consensus repeat, from 2 to 20, CTPRa2 to CTPRa20. The DSC data provides direct evidence that the folding/unfolding transition of CTPR proteins does not fit a two-state folding model. Our results confirm and expand earlier studies on TPR proteins, which showed that apparent two-state unfolding curves are better fit by linear statistical mechanics models: 1D Ising models in which each repeat is treated as an independent folding unit.     

Structural dynamics of pentapeptide repeat proteins

Pentapeptide repeat proteins (PRPs) represent a large superfamily with more than 38 000 sequences in nearly 3500 species, the majority belonging to cyanobacteria but represented among all branches of life. PRPs contain at least eight consecutive pentapeptide repeats with the consensus (A/C/S/V/T/L/I)(D/N/S/K/E/I/R)(L/F)(S/T/R/E/Q/K/V/D)(G/D/E/N/R/Q/K). PRPs fold into right-handed quadrilateral β helices, also known as repeat-five-residue (Rfr)-folds, with four consecutive pentapeptide repeats comprising a single coil, the ~90° change in polypeptide direction in square-shaped coils achieved by type I, II and IV β turns, and hydrogen bonds between coils establishing β ladders on each Rfr-fold face. PRPs are broadly categorized into group 1 and 2 involved in antibiotic resistance and group 3 currently having unknown functions. Motivated by their intriguing structures, we are investigating PRP biophysical characteristics, including Rfr-fold thermal stability, β turn and β ladder hydrogen bond amide exchange rates and backbone dynamics. Here, we present analysis of 20 ns molecular dynamics (MD) simulations and all atom normal mode analysis (aaNMA) calculations for four group 1 and group 2 and four group 3 PRPs whose structures have been determined by X-ray crystallography. The MD cross-correlation matrices and aaNMA indicated strong correlated motion between adjacent coils and weak coupled motion between coils separated by one or more intervening coils. Slow anticorrelated motions were detected between adjacent coils in aaNMA modes that we hypothesize are requisite to access exchange-competent states necessary to permit solvent exchange of amide hydrogens involved in β-ladder and β-turns hydrogen bonds, which can have lifetimes on the order of months.     

Refinement of comparative models of protein structure by using multicanonical molecular dynamics simulations

Comparative modelling is a powerful method that easily predicts a considerably accurate structure of a protein by using a template structure having a similar amino-acid sequence to the target protein. However, in the region where the amino-acid sequence is different between the target and the template, the predicted structure remains unreliable. In such a case, the model has to be refined. In the present study, we explored the possibility of a molecular dynamics-based method, using the human SAP Src Homology 2 (SH2) domain as the modelling target. The multicanonical method was used to alleviate the multiple-minima problem and the generalised Born/surface area model was used to reduce the computational cost. In addition, position restraints were imposed on the atoms in the reliable regions to avoid unnecessary conformational sampling. We analyzed the conformational distribution of the ligand-recognition loop of the domain and found that the most populated conformational clusters in the ensemble of the model agreed well with one of the two major clusters in the ensemble of the reference simulation starting from the crystal structure. This demonstrates that the current refinement method can significantly improve the accuracy of an unreliable region in a comparative model.     

Designing repeat proteins: well-expressed,soluble and stable proteins from combinatorial libraries of consensus ankyrin repeat proteins

We describe an efficient way to generate combinatorial libraries of stable, soluble and well-expressed ankyrin repeat (AR) proteins. Using a combination of sequence and structure consensus analyses, we designed a 33 amino acid residue AR module with seven randomized positions having a theoretical diversity of 7.2x10(7). Different numbers of this module were cloned between N and C-terminal capping repeats, i.e. ARs designed to shield the hydrophobic core of stacked AR modules. In this manner, combinatorial libraries of designed AR proteins consisting of four to six repeats were generated, thereby potentiating the theoretical diversity. All randomly chosen library members were expressed in soluble form in the cytoplasm of Escherichia coli in amounts up to 200 mg per 1 l of shake-flask culture. Virtually pure proteins were obtained in a single purification step. The designed AR proteins are monomeric and display CD spectra identical with those of natural AR proteins. At the same time, our AR proteins are highly thermostable, with T(m) values ranging from 66 degrees C to well above 85 degrees C. Thus, our combinatorial library members possess the properties required for biotechnological applications. Moreover, the favorable biophysical properties and the modularity of the AR fold may account, partly, for the abundance of natural AR proteins.     

Structure-based optimization of designed Armadillo-repeat proteins

The armadillo domain is a right‐handed super‐helix of repeating units composed of three α‐helices each. Armadillo repeat proteins (ArmRPs) are frequently involved in protein–protein interactions, and because of their modular recognition of extended peptide regions they can serve as templates for the design of artificial peptide binding scaffolds. On the basis of sequential and structural analyses, different consensus‐designed ArmRPs were synthesized and show high thermodynamic stabilities, compared to naturally occurring ArmRPs. We determined the crystal structures of four full‐consensus ArmRPs with three or four identical internal repeats and two different designs for the N‐ and C‐caps. The crystal structures were refined at resolutions ranging from 1.80 to 2.50 Å for the above mentioned designs. A redesign of our initial caps was required to obtain well diffracting crystals. However, the structures with the redesigned caps caused domain swapping events between the N‐caps. To prevent this domain swap, 9 and 6 point mutations were introduced in the N‐ and C‐caps, respectively. Structural and biophysical analysis showed that this subsequent redesign of the N‐cap prevented domain swapping and improved the thermodynamic stability of the proteins. We systematically investigated the best cap combinations. We conclude that designed ArmRPs with optimized caps are intrinsically stable and well‐expressed monomeric proteins and that the high‐resolution structures provide excellent structural templates for the continuation of the design of sequence‐specific modular peptide recognition units based on armadillo repeats.     

Modular peptide binding: From a comparison of natural binders to designed armadillo repeat proteins

Several binding scaffolds that are not based on immunoglobulins have been designed as alternatives to traditional monoclonal antibodies. Many of them have been developed to bind to folded proteins, yet cellular networks for signaling and protein trafficking often depend on binding to unfolded regions of proteins. This type of binding can thus be well described as a peptide–protein interaction. In this review, we compare different peptide-binding scaffolds, highlighting that armadillo repeat proteins (ArmRP) offer an attractive modular system, as they bind a stretch of extended peptide in a repeat-wise manner. Instead of generating each new binding molecule by an independent selection, preselected repeats – each complementary to a piece of the target peptide – could be designed and assembled on demand into a new protein, which then binds the prescribed complete peptide. Stacked armadillo repeats (ArmR), each typically consisting of 42 amino acids arranged in three α-helices, build an elongated superhelical structure which enables binding of peptides in extended conformation. A consensus-based design approach, complemented with molecular dynamics simulations and rational engineering, resulted in well-expressed monomeric proteins with high stability. Peptide binders were selected and several structures were determined, forming the basis for the future development of modular peptide-binding scaffolds.     

Coarse-grained molecular dynamics simulations of membrane proteins and peptides

Molecular dynamics (MD) simulations provide a valuable approach to the dynamics, structure, and stability of membrane-protein systems. Coarse-grained (CG) models, in which small groups of atoms are treated as single particles, enable extended (>100 ns) timescales to be addressed. In this study, we explore how CG-MD methods that have been developed for detergents and lipids may be extended to membrane proteins. In particular, CG-MD simulations of a number of membrane peptides and proteins are used to characterize their interactions with lipid bilayers. CG-MD is used to simulate the insertion of synthetic model membrane peptides (WALPs and LS3) into a lipid (PC) bilayer. WALP peptides insert in a transmembrane orientation, whilst the LS3 peptide adopts an interfacial location, both in agreement with experimental biophysical data. This approach is extended to a transmembrane fragment of the Vpu protein from HIV-1, and to the coat protein from fd phage. Again, simulated protein/membrane interactions are in good agreement with solid state NMR data for these proteins. CG-MD has also been applied to an M3-M4 fragment from the CFTR protein. Simulations of CFTR M3-M4 in a detergent micelle reveal formation of an alpha-helical hairpin, consistent with a variety of biophysical data. In an I231D mutant, the M3-M4 hairpin is additionally stabilized via an inter-helix Q207/D231 interaction. Finally, CG-MD simulations are extended to a more complex membrane protein, the bacterial sugar transporter LacY. Comparison of a 200 ns CG-MD simulation of LacY in a DPPC bilayer with a 50 ns atomistic simulation of the same protein in a DMPC bilayer shows that the two methods yield comparable predictions of lipid-protein interactions. Taken together, these results demonstrate the utility of CG-MD simulations for studies of membrane/protein interactions.     

Molecular dynamics simulations of peptides and proteins with a continuum electrostatic model based on screened Coulomb potentials

A continuum electrostatics approach for molecular dynamics (MD) simulations of macromolecules is presented and analyzed for its performance on a peptide and a globular protein. The approach incorporates the screened Coulomb potential (SCP) continuum model of electrostatics, which was reported earlier. The model was validated in a broad set of tests some of which were based on Monte Carlo simulations that included single amino acids, peptides, and proteins. The implementation for large-scale MD simulations presented in this article is based on a pairwise potential that makes the electrostatic model suitable for fast analytical calculation of forces. To assess the suitability of the approach, a preliminary validation is conducted, which consists of (i) a 3-ns MD simulation of the immunoglobulin-binding domain of streptococcal protein G, a 56-residue globular protein and (ii) a 3-ns simulation of Dynorphin, a biological peptide of 17 amino acids. In both cases, the results are compared with those obtained from MD simulations using explicit water (EW) molecules in an all-atom representation. The initial structure of Dynorphin was assumed to be an alpha-helix between residues 1 and 9 as suggested from NMR measurements in micelles. The results obtained in the MD simulations show that the helical structure collapses early in the simulation, a behavior observed in the EW simulation and consistent with spectroscopic data that suggest that the peptide may adopt mainly an extended conformation in water. The dynamics of protein G calculated with the SCP implicit solvent model (SCP-ISM) reveals a stable structure that conserves all the elements of secondary structure throughout the entire simulation time. The average structures calculated from the trajectories with the implicit and explicit solvent models had a cRMSD of 1.1 A, whereas each average structure had a cRMSD of about 0.8A with respect to the X-ray structure. The main conformational differences of the average structures with respect to the crystal structure occur in the loop involving residues 8-14. Despite the overall similarity of the simulated dynamics with EW and SCP models, fluctuations of side-chains are larger when the implicit solvent is used, especially in solvent exposed side-chains. The MD simulation of Dynorphin was extended to 40 ns to study its behavior in an aqueous environment. This long simulation showed that the peptide has a tendency to form an alpha-helical structure in water, but the stabilization free energy is too weak, resulting in frequent interconversions between random and helical conformations during the simulation time. The results reported here suggest that the SCP implicit solvent model is adequate to describe electrostatic effects in MD simulation of both peptides and proteins using the same set of parameters. It is suggested that the present approach could form the basis for the development of a reliable and general continuum approach for use in molecular biology, and directions are outlined for attaining this long-term goal.     

Implicit solvent models for flexible protein-protein docking by molecular dynamics simulation

The suitability of three implicit solvent models for flexible protein-protein docking by procedures using molecular dynamics simulation is investigated. The three models are (i) the generalized Born (GB) model implemented in the program AMBER6.0; (ii) a distance-dependent dielectric (DDD) model; and (iii) a surface area-dependent model that we have parameterized and call the NPSA model. This is a distance-dependent dielectric model modified by neutralizing the ionizable side-chains and adding a surface area-dependent solvation term. These solvent models were first tested in molecular dynamics simulations at 300 K of the native structures of barnase, barstar, segment B1 of protein G, and three WW domains. These protein structures display a range of secondary structure contents and stabilities. Then, to investigate the performance of the implicit solvent models in protein docking, molecular dynamics simulations of barnase/barstar complexation, as well as PIN1 WW domain/peptide complexation, were conducted, starting from separated unbound structures. The simulations show that the NPSA model has significant advantages over the DDD and GB models in maintaining the native structures of the proteins and providing more accurate docked complexes.     

Relationship between energy distribution and fold stability: Insights from molecular dynamics simulations of native and mutant proteins

Most proteins must fold to a well-defined structure with a minimal stability to perform their function. Here we use a simple, molecular dynamics-based, energy decomposition approach to map the principal energetic interactions in a set of proteins representative of different folds. This work involves the all-atom simulation and analysis of the native structures and mutants of five different proteins representative of an all-alpha (yACPB, Protein A), all-beta (SH3), and a mixed alpha/beta fold (Proteins G and L). Given a certain structure, a native sequence and a set of mutants, we show that our model discriminates the ability of a mutation to yield a more or less stable protein, in agreement with experimental data, catching the principal energetic determinants of protein stabilization. Our approach identifies the interaction determinants responsible to define a fold and shows that mutations can either modulate the strength of pair-wise coupling between residues important for folding, or modify the profile of the principal interactions. Furthermore, we address the question of how to evaluate the fitness of a sequence to a given structure by comparing the information contained in the energy map, which recapitulates the chemistry of the sequence, to that contained in the contact map, which recapitulates the fold topology. The results show that the better fit between the energetic properties of the sequence and the fold topology corresponds to a higher stabilization of the protein. We discuss the relevance of these observations to the analysis of protein designability and to the rational evolution of new sequences.     

Thermostable designed ankyrin repeat proteins (DARPins) as building blocks for innovative drugs

Designed ankyrin repeat proteins (DARPins) are antibody mimetics with high and mostly unexplored potential in drug development. By using in silico analysis and a rationally guided Ala scanning, we identified position 17 of the N-terminal capping repeat to play a key role in overall protein thermostability. The melting temperature of a DARPin domain with a single full-consensus internal repeat was increased by 8 °C to 10 °C when Asp17 was replaced by Leu, Val, Ile, Met, Ala, or Thr. We then transferred the Asp17Leu mutation to various backgrounds, including clinically validated DARPin domains, such as the vascular endothelial growth factor-binding domain of the DARPin abicipar pegol. In all cases, these proteins showed improvements in the thermostability on the order of 8 °C to 16 °C, suggesting the replacement of Asp17 could be generically applicable to this drug class. Molecular dynamics simulations showed that the Asp17Leu mutation reduces electrostatic repulsion and improves van-der-Waals packing, rendering the DARPin domain less flexible and more stable. Interestingly, this beneficial Asp17Leu mutation is present in the N-terminal caps of three of the five DARPin domains of ensovibep, a SARS-CoV-2 entry inhibitor currently in clinical development, indicating this mutation could be partly responsible for the very high melting temperature (>90 °C) of this promising anti-COVID-19 drug. Overall, such N-terminal capping repeats with increased thermostability seem to be beneficial for the development of innovative drugs based on DARPins.     

Characterization of de novo four-helix bundles by molecular dynamics simulations

We have investigated the structure and dynamics of three cavitand-based four-helix bundles (caviteins) by computer simulation. In these systems, designed de novo, each of the four helices contain the identical basis sequence EELLKKLEELLKKG (N1). Each cavitein consists of a rigid macrocycle (cavitand) with four aryl linkages, to each of which is connected an N1 peptide by means of a linker peptide. The three caviteins studied here differ only in the linker peptide, which consist of one, two, or three glycine residues. Previous experimental work has shown that these systems exhibit very different behavior in terms of stability and oligomerization states despite the small differences in the linker peptide. Given that to date no three-dimensional structure is available for these caviteins, we have undertaken a series of molecular dynamics (MD) simulations in explicit water to try to rationalize the large differences in the experimentally observed behavior of these systems. Our results provide insight, for the first time, into why and how the cavitein with a single glycine linker forms dimers. In addition, our results indicate why although the two- and three-glycine-linked caviteins have similar stabilities, they have different native-like characteristics: the cavitein with three glycines can form a supercoiled helix, whereas the one with two glycines cannot. These findings may provide a useful guide in the rational de novo design of novel proteins with finely tunable structures and functions in the future.