Computational Modeling of Stapled Peptides toward a Treatment

Subscriber access provided by UNIV OF DURHAM

B: Biophysical Chemistry and Biomolecules

Computational Modeling of Stapled Peptides Towards a Treatment Strategy for CML and Broader Implications in the Design of Lengthy Peptide Therapeutics Sean P. Cornillie, Benjamin J Bruno, Carol S. Lim, and Thomas E. Cheatham J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.8b01014 • Publication Date (Web): 08 Mar 2018 Downloaded from http://pubs.acs.org on March 9, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Computational Modeling of Stapled Peptides Towards a Treatment Strategy for CML and Broader Implications in the Design of Lengthy Peptide Therapeutics Sean P. Cornillie1, Benjamin J. Bruno2, Carol S. Lim2, and Thomas E. Cheatham III1,*

1

2

*

Department of Medicinal Chemistry, University of Utah, Utah, USA

Department of Pharmaceutics and Pharmaceutical Chemistry, University of Utah, Utah, USA

To whom correspondence should be addressed. Email: [email protected]

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 37

1

Abstract The oncogenic gene product Bcr-Abl is the principal cause of chronic myeloid leukemia (CML) and though several therapies exist to curb the aberrant kinase activity of Bcr-Abl through targeting of the Abl kinase domain, these therapies are rendered ineffective by frequent mutations in the corresponding gene. It has been demonstrated that a designed protein known as CCmut3 is able to produce a dominant negative inactivating effect on Bcr-Abl kinase by preferentially oligomerizing with the N-terminal coiled-coil oligomerization domain of Bcr-Abl (Bcr-CC) to effectively reduce the oncogenic potential of Bcr-Abl. However, the sheer length of the CCmut3 peptide introduces a high degree of conformational variability and opportunity for targeting by intracellular proteolytic mechanisms. Here we have examined the effects on introducing one or two molecular staples, or cross-links, spanning i,i+7 backbone residues of the CCmut3 construct which have been suggested to reinforce α-helical conformation, enhance cellular internalization, and increase resistance to proteolytic degradation leading to enhanced pharmacokinetic properties. The importance of optimizing staple location along a highly-tuned biological construct such as CCmut3 has been widely emphasized and as such, we have employed in silico techniques to swiftly build, relax, and characterize a large number of candidates. This approach effectively allowed exploring each and every possible staple location along the peptide backbone so that every possible candidate is considered. While many of stapled candidate peptides displayed enhanced binding characteristics for Bcr-CC and improved conformational stability in the (Bcr-CC) bound form, simulations of the stapled peptides in the unbound form revealed widespread conformational variability among stapled candidates dependent on staple type and location, implicating the molecular replacement of helix-stabilizing residues with staple-containing residues in disrupting the native α-helical conformation of


Page 3 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


2 CCmut3, further highlighting a need for careful optimization of the CCmut3 construct. A candidate set has been assembled which retains the native backbone α-helical integrity in both the bound and unbound forms whilst providing enhanced binding affinity for the Bcr-CC target, as research disseminated in this manuscript is intended to guide the development of a nextgeneration CCmut3 inhibitor peptide in an experimental setting.

Introduction The classical treatment of chronic myeloid leukemia (CML) involves the lifelong disease management with tyrosine kinase inhibitors (TKIs) which target the oncogenic gene product BcrAbl, a constitutively active tyrosine kinase that interferes with key cell signaling pathways and drives the malignant CML phenotype.1,2 CML is characterized by dysregulated cellular proliferation through the suppression of apoptosis mechanisms, constitutively active mitogenic signaling, and altered adhesion to the bone marrow stroma.1–5 A number of TKIs have been developed (imantinib, nilotinib, ponatinib) which effectively target the Abl kinase domain to curb malignant kinase activity.6–8 However the majority of approved TKIs are heavily affected by resistance mutations within the Abl kinase region and beset with a notable risk of serious side-effects (arterial thrombosis, severe blood clotting, blood vessel narrowing), such that the exploration of novel and alternative therapies is still of relevant therapeutic interest.9–12 The Bcr domain of the fusion protein Bcr-Abl is composed of an alpha-helical coiled-coil motif (Bcr-CC), which serves as an oligomerization domain through which (inactive) Bcr-Abl monomers are assembled into active homo-tetramers.13–15 It has been established that Bcr-Abl constructs lacking the Bcr-CC domain have severely reduced transformation potential, and that the disruption of Bcr-mediated assembly through hetero-oligomerization between Bcr-CC and



Page 4 of 37

3 isolated coiled-coils depletes oncogenic Abl kinase activity.13,15–18 Our previous work combined both biomolecular simulation and in vitro experimentation techniques to develop the coiled-coil CCmut3, a Bcr-CC mimic which preferentially hetero-oligomerizes with native Bcr-CC to produce a dominant negative inactivating effect on Bcr-Abl kinase activity.16,17 A series of mutations (C38A, S41R, L45D, E48R, Q60E) characterize CCmut3, promoting preferential oligomerization with the wild type Bcr-CC domain and disfavoring CCmut3:CCmut3 selfassociation.16,17 The CCmut3 construct is an effective inhibitor of Bcr-Abl activity in vitro in human CML cell lines, including cell lines containing the T315I “gatekeeper” mutation as well as the compound mutant T315I/E255V.16,17,19,20 However the sheer size of CCmut3 results in relatively low cellular uptake when compared to ‘traditional’ small-molecule therapeutics, and introduces a high degree of conformational variability which allows for degradation by intracellular proteolytic mechanisms, each contributing to limit bioavailability in vivo – thus facilitating an interest to develop a next-generation Bcr-Abl inhibitor based on the CCmut3 construct.17,21–23 Here we utilized computational modelling and biomolecular simulation techniques to design a 3rd generation peptide inhibitor of Bcr-mediated Bcr-Abl assembly following the development of inhibitor CCmut3. We present a set of constructs that seek to remedy the shortcomings characteristic of peptide therapeutics: the peptide constructs have been truncated to eliminate the subdomain region CCmut3-α1 (residues 1-27) of innate high conformational variability (Figure S1) and incorporated molecular ‘staples’ (or cross-links) spanning i,i+7 residues along the backbone of the alpha-helical CCmut3-α2 subdomain (residues 28-67). Molecular stapling is rapidly emerging as an effective approach to stabilize α-helical peptide motifs, and the inclusion of staples have been by reported by numerous groups to limit


Page 5 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4 conformational variability to ‘lock in’ a stabilized helical state, enhance cellular internalization, and increase resistance to proteolytic degradation, leading to substantial increases in target affinity in a varied array of peptide constructs.22–27 Amongst a diverse assortment of staple types28,29, all-hydrocarbon24 cross-links have certainly garnered the most widespread attention; however the relatively expensive nature of synthesizing lengthy peptide constructs (e.g. CCmut3) incorporated by one or more hydrocarbon staples is enough to facilitate the exploration of linkers which are more economically viable to produce. A novel approach to linker design was recently realized upon the discovery of a novel peptide stapling technique involving the coupling of cysteine residues and α,ω-diene-containing linkers, enabling the ability to staple recombinantly-expressed CCmut3 with a variety of linker types via selective modification of unprotected cysteine residues.30,31 Evidence suggests that varying linker type may influence the bioactive properties of stapled peptides, such that we have chosen to explore two varied backbone staple types (Figure 1) in pursuit of an optimized stapled CCmut3 construct: first is that of the traditional all-hydrocarbon variant spanning i,i+7 residues described by Schafmeister et. al.24 and second is that of a decreasingly-hydrophobic urea-based linker spanning i,i+7 cysteine residues, described by Wang et. al.31 Protein-protein interactions are by design highly sequence-specific such that the molecular replacement of residues in a highly-tuned biological construct, in favor of staplecontaining residues, is posed to alter target affinity and/or binding mechanics. Moreover, the optimization of staple positioning along the peptide construct becomes key in order to maintain or enhance target affinity.22,25,32 Stapled peptides which are reported in literature have typically been designed via the thorough examination of high-resolution structures or comprehensive alanine scanning studies, but these methods are by no means exhaustive or all-inclusive and



Page 6 of 37

5 provide little insight into the behavior of a stapled construct in practice.22,32 While staple positioning would ideally be optimized via the synthesis and biophysical characterization of each and every possible stapled variant of a peptide construct, to characterize this in an experimental setting potentially of hundreds of candidates is neither economically viable nor feasible within a reasonable timeline, especially in a lengthy construct such as CCmut3. However, utilizing computational tools we are able to quickly and effectively build, relax, and characterize candidates in silico; thus we are able to explore each possible staple location along the peptide backbone so that all possible candidates are considered in the search – greatly cutting down on costs and time necessary to design and optimize lead compounds for detailed experimental studies.

Methods Model Building All models of the modified Bcr-Abl oligomerization domain were constructed based on a crystal structure of the N-terminal oligomerization domain of Bcr-Abl (Protein Data Bank (PDB) entry 1K1F), where residues 1-67 in each Chain A and Chain B were used to build the models. The swapaa tool within Chimera was used to convert selenomethionine residues to methionine and to convert residue 38 to cysteine, consistent with the wild-type protein.33–35 To build the modified CCmut3 constructs, the mutations C38A, S41R, L45D, E48R and Q60E were sequentially implemented (chain A, residues 1-67) using the swapaa tool within Chimera.34 Chain B (residues 68-134) was unmodified apart from the changes described to revert back to its wild-type form. The CCmut3-α1 domain of high conformational variability (residues 1-27) was deleted to attain the truncated CCmut3 construct (residues 28-67). To build the stapled CCmut3


Page 7 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


6 coils, non-standard amino acid residues were created and parameterized to contain the molecular staples. Two distinct staple types were built and parameterized for this work: an all-hydrocarbon staple designed to span i,i+7 peptide residues, and a urea-based staple also designed to span i,i+7 peptide residues. Each staple type is depicted in Figure 1. Equivalent parameterization methods were used to build each of the two unique staple types: Atom coordinates for the non-standard residues were developed using Gaussview 5.0.36 Geometry and partial charges of said residues were assigned using the RESP ESP charge Derive (RED) server, in which the geometries of the stapled residues were optimized at a HF 6-31G* level of theory and assigned partial charges based on restrained electrostatic potential (RESP)

Figure 1. Structures of the (A) hydrocarbon (colored by heteroatom, carbons colored orange) and (B) urea (colored by heteroatom, carbons colored cyan) staple types which have been incorporated into i,i+7 residues of the CCmut3 peptide. Atomistic representations of the (C) hydrocarbon staple incorporated by a cis-olefin linkage and (D) urea staple spanning two cysteine residues are additionally presented.



Page 8 of 37

7 methodology.37–39 Antechamber (AmberTools15) was used to assign AMBER atom types and generate library files (available in the Supplemental Information) to define the non-standard amino acid residues within the ff14SB force field.40 The modified, staple-containing residues were incorporated into CCmut3 coils by residue replacement within the PDB file. AMBER ff14SB force field parameters were used to build the models, which were explicitly solvated in a truncated octahedron with a 10 Å minimum surrounding buffer of TIP3P water.40–42 Joung and Cheatham ion parameters were used to incorporate net-neutralizing counter-ions (Na+ / Cl-), and 50 additional Na+ / Cl- atoms were incorporated to reach a biologically-representative ion concentration of ~200 mM.43 All solute hydrogen masses were repartitioned to a mass of 3.024 Da to allow for a doubling of the MD time integration step to 4 fs.44

Minimization and Equilibration An extensive minimization and equilibration protocol was applied to all models in order to relax and steer systems towards energetically favorable conformations prior to performing production molecular dynamics (MD). An initial minimization (500 steps of steepest descent, 500 steps of conjugate gradient) was performed prior to equilibrating the system to a biologically-representative temperature of 300 K. For the duration of the initial minimization and heating steps, A 25 kcal/(mol-Å2) restraint was placed upon backbone Cα atoms. Five cycles of minimization (500 steps of steepest descent each, 500 steps of conjugate gradient each) and equilibration were then performed (upon each system), where restraint weights were sequentially reduced from 5 kcal/(mol-Å2) to 1 kcal/(mol-Å2). A final equilibration was performed with a restraint weight of 0.5 kcal/(mol-Å2) for 500 ps. A Berendsen thermostat with a 0.2 ps coupling


Page 9 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


8 time was used to control both constant temperature (300 K) and pressure (1 bar) throughout the minimization and equilibration protocol.45

Production Molecular Dynamics Simulations All production MD simulations were performed using the GPU implementation of the Amber14 modelling suite with K20X GPUs on the Blue Waters Petascale Resource in explicit solvent using a 4 fs time step, a Langeven thermostat with a collision frequency of 5 ps-1 to regulate constant temperature and pressure, a 10 Å non-bonded cutoff, default particle mesh Ewald treatment of electrostatics, and SHAKE applied to bonds to hydrogen.46–52 Three copies of each system were run for statistical purposes, and combined account for between 2.2 µs and 4.9 µs of simulation time for each of 62 stapled peptide candidates and the unstapled CCmut3 control, using a different random seed in each simulation to prevent synchronization artifacts, generated every 24 hours of simulation time.53

Analysis Protocols Analyses of the candidate peptides were performed with the CPPTRAJ toolset included within the AmberTools15 package, and clustering analyses accumulating the most commonlysampled ligand and receptor conformations (described subsequently in this manuscript) were performed upon ligand and receptor peptide backbone atoms, using an average linkage hierarchical agglomerative (bottom up) approach where the minimum distance between clusters was varied in each simulation to generate between five and ten clusters within each unique simulation scenario.54 Energetic analyses were performed using the MM-PBSA (Molecular Mechanics Poisson



Page 10 of 37

9 Boltzmann Surface Area) methodology, where a single-trajectory approach was utilized to estimate the relative binding energetics between each stapled peptide (or unstapled CCmut3) and the Bcr-CC receptor.55 The MMPBSA.py program compiled within AmberTools15 was used to calculate the relative overall free energies, as well as to compute the energetic decompositions on a per-residue basis.56 In all cases, energetics between distinct stapled peptides (or unstapled CCmut3) and Bcr-CC were calculated of a trajectory snapshot defined by the final ten nanoseconds (final 1000 frames) in each of three independent simulation copies. Relative energetic values (averaged over the three distinct energetic calculations in each simulation scenario) and standard error of the mean are reported.

Results & Discussion All-atom MD simulations have been performed to investigate the structure and dynamics of 62 truncated and stapled CCmut3 variants, including 32 candidates which have been incorporated by fully hydrocarbon staples and 30 candidates which have been incorporated by urea-based linkers. Candidates were incorporated by either: one i,i+7 hydrocarbon staple, one i,i+7 urea staple, two i,i+7 hydrocarbon staples, or two i,i+7 urea staples. Staples were strategically placed along the CCmut3 backbone to avoid interference with the CCmut3:Bcr-CC binding interface and avoid the destruction of any favorable binding partners such as salt bridges or contacts identified as potentially critical for maintaining dimerization. MD simulation and analyses were used to investigate the structure and dynamics of stapled CCmut3 analogues in both the unbound and bound conditions: unbound CCmut3 analogues were simulated in the absence of Bcr-CC while simulations of the bound condition probed interactions between the stapled CCmut3 analogues and Bcr-CC (Figure 2).


Page 11 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


10

Figure 2. Double[hydrocarbon]-stapled candidate DHC-030 (grey) bound to native Bcr-CC (purple). Hydrocarbon staples bridging i,i+7 residues along the peptide backbone are represented atomistically (orange) and key mutations defining CCmut3 are represented in spherical atomistic notation.

Key Residue Interactions Realized Between CCmut3 and Bcr-CC A per-residue free energy decomposition of the CCmut3:Bcr-CC complex was performed to identify residues key in maintaining CCmut3 binding to Bcr-CC. Extensive hydrophobic contacts are observed between the dimerization interface of CCmut3 and Bcr-CC coils, largely maintaining the ‘knobs-in-holes’ fashion of hydrophobic residue packing along the heptad repeat



Page 12 of 37

11 found within the wild-type Bcr-CC homo-dimerization interface. Involved are CCmut3 residues Ile31, Leu35, Ile42, Val49, Met56, Leu59, and Leu63 and Bcr-CC residues Ile31, Leu35, Ile42, Leu45, Val49, Met56, Leu59, and Leu63 of Bcr-CC, each of which make moderate contributions to the overall binding affinity of CCmut3 for Bcr-CC (Table S1). Concurrent with experimental findings, residues Phe7, Trp11, and Phe15 of the Bcr-CC α1 domain (Residues 1-27) make contacts with CCmut3 residues Phe54 and Tyr58, forming an aromatic core opposite the dimerization interface to provide additional binding stability between the two coils.33

(CCmut3)

(Bcr-CC)

Simulation Occupancy (% of Frames)

Glu32

Lys67

11.4

-1.2 ± 0.3

Lys67

Glu32

16.6

-1.5 ± 0.1

Glu46

Arg53

74.0

-4.5 ± 0.2

Arg53

Glu46

77.7

-4.1 ± 0.3

Glu34

Arg55

78.8

-5.6 ± 0.3

Arg55

Glu34

89.6

-4.7 ± 0.4

Arg41

Glu52

65.3

-4.4 ± 0.9

Glu60

Lys39

28.2

-1.2 ± 0.1

Salt Bridge Residue

Energetic Contribution (kcal/mol)

Table 1. Identified potential CCmut3:Bcr-CC salt bridge interactions and the collective energetic contributions (in kcal/mol) of residues involved. Per-residue energy decompositions were performed to obtain the energetic contribution of each salt bridge interaction during the portion of time during the molecular dynamics trajectory that the salt bridge was formed. Five salt bridge interactions (Glu46-Arg53, Arg53-Glu46, Glu-34-Arg55, Arg55-Glu34, and Arg41-Glu52) were reported to significantly contribute to heterodimer affinity based on energetic contribution and salt bridge occupancy among three independent simulation replicas. Reference Figure S2 for salt bridge occupancies as a function of time.


Page 13 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


12 Eight prospective salt bridges were initially identified as possessing the potential to contribute to hetero-dimer stability, but only five salt bridges (Glu46-Arg53, Arg53-Glu46, Glu34-Arg55, Arg55-Glu34, and Arg41-Glu52) were ultimately observed to contribute significantly to the overall binding affinity between CCmut3 and Bcr-CC based on energetic contribution and occupancy of each salt bridge throughout each of three 1 µs independent replica simulations (Table 1, Figure S2). The original CCmut3 construct was engineered to feature a salt bridge between CCmut3 residue Arg41 (S41R mutation) and Bcr-CC residue Glu48; however the Arg41-Glu48 salt bridge interaction was observed only in ~9% of total simulation frames. The possibility of Arg41-Glu48 bridging via an inter-residue water bridge was ruled out using a distance metric which places the greatest population of atoms involved in each of six unique bridging opportunities ~10 Å apart. The distance metric reports minor populations of Arg41Glu48 bridging atoms positioned both ~3 Å apart and ~5-6 Å apart, suggesting at least temporary Arg41 and Glu48 interactions mediated via salt bridging or water bridging (Figure S3).17 Instead, the formation of salt bridge Arg41-Glu52 was observed throughout 65% of simulation frames and the residues involved contribute substantially (-4.4 kcal/mol) to the overall binding affinity of CCmut3 for Bcr-CC. A secondary designed salt bridge between CCmut3 residue Glu60 and Bcr-CC residue Lys39 was engineered via a Q60E mutation which was predicted not only to enhance CCmut3:Bcr-CC hetero-oligomerization, but to disfavor CCmut3:CCmut3 homooligomerization and aggregation through charge-charge repulsion. While the Q60E mutation has been observed to contribute to the disruption of the CCmut3:CCmut3 interaction in vitro, simulations of the hetero-dimer reveal only a modest contribution of the Glu60-Lys39 salt bridge the overall affinity of CCmut3 for Bcr-CC (-1.2 kcal/mol, 28% salt bridge occupancy).17



Page 14 of 37

13 Simulations of Stapled Constructs in Complex with Bcr-CC With the exception of a handful of constructs, stapled peptides bound to Bcr-CC are observed to exhibit minimal structural deviations from the native CCmut3:Bcr-CC heterodimer as reported by root-mean-square deviation (RMSD) analyses (Figure S4) and radius of gyration calculations (Figure S6). A slight RMS deviation towards the native structure is observed to be common among stapled peptides and reported in Figure S4, indicating that many stapled candidates retain the native conformation observed in the reference crystal structure (PDB: 1K1F, chains A & B) more closely than unstapled CCmut3 following ~2-5 µs of combined simulation per candidate. Examining closely the most noticeably shifted candidate peptide single[urea]-stapled SUR-010, we observe increased σ-helicity (Table S3) and fewer RMS fluctuations in SUR-010 (especially in N-terminal, staple-occluded residues 29-36) comparable to unstapled CCmut3 (Figure S7). Additionally, a principle component analysis (PCA) reports slight deviations in the principle modes of motion between SUR-010 and the CCmut3 control (Figure S8). Together, these analyses suggest that the i,i+7 staples placed along the backbone are (in the majority of stapled candidates) acting to stabilize the native α-helical conformation observed in the crystal structure (1K1F), whereas an unrestricted unstapled CCmut3 peptide is permitted to explore a greater conformational landscape. A clustering analysis was employed to accumulate and visualize the most highly-sampled conformations in of the stapled candidates, indicating a low degree of structural variability in the bound condition among the majority of stapled candidates. Five candidates, in particular, stick out as noticeably deviating in RMSD space (Figure S4). Visualization of these candidates reveals a partial translation of the peptide backbone at Cterminal residues of single[hydrocarbon]-stapled SHC-130, single[urea]-stapled SUR-070,


Page 15 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


14 double[urea]-stapled DUR-130. However, a significantly ‘bent’ peptide conformation is observed in stapled peptides SHC-060 and DHC-050 where the two peptides feature a loss in backbone helicity and significant ‘bending’ of the peptide backbone.

The ‘bent’ peptide

structure is manifested as the major conformation of SHC-060 in each of three independent replica simulations (720 ns per replica), and the major conformation of DHC-050 in one of three independent replica simulations (1.04 µs per replica). The SHC-060 and DHC-050 also appear to influence Bcr-CC to adopt a similar bent conformation (Figure S9) and while peptides SHC-060 and DHC-050 remain bound to Bcr-CC for the duration of each of three independent simulations (per system) and both retain an overall binding affinity for Bcr-CC comparable to that of the unstapled control (Table S2, Table S4), the breakdown in secondary structure in the bound condition is quite possibly an indicator of ligand conformational instability instigated by staple placement, highlighting the need for an optimized stapled control construct. Using MM-PBSA energetic analyses, we were able to identify seventeen stapled CCmut3 analogues conferring enhanced binding affinity for Bcr-CC, and seventeen additional constructs retaining equivalent binding affinity for Bcr-CC (Tables S2-S5). The remaining twenty-eight stapled CCmut3 variants remain bound to the Bcr-CC receptor despite suffering losses in overall binding affinity. Twelve of the hydrocarbon-stapled candidates (five single[hydrocarbon]-stapled and seven double[hydrocarbon]-stapled) account for the stapled variants conferring enhanced binding affinity for Bcr-CC, while only five of the urea-stapled candidates (three single[urea]stapled and double[urea]-stapled) were reported as having enhanced binding affinity over unstapled CCmut3. A trend observed amongst both singly-stapled and doubly-stapled candidates was the relatively greater binding affinity of hydrocarbon-stapled candidates for Bcr-CC compared to the equivalent urea-stapled candidates, which becomes apparent when comparing



Page 16 of 37

15 candidates featuring one or more backbone staple near the C-terminus of the peptide ligand, where a staple has the potential to interact with the Bcr-CC α1 flexible tail domain (Residues 127). Unstapled CCmut3 is posed to make several hydrophobic contacts with the α1 domain, such that it is not out of the realm of possibility that peptides incorporated by a staple near the Cterminus will interact with the α1 domain – and that staple type has a direct influence on the benefits or detriments of this interaction. Additional investigation into this interaction is discussed in subsequent sections.

Stapled Peptides Display a High Degree of Conformational Variability in Solution A number of the stapled CCmut3 analogues were observed to produce significant increases in binding affinity for Bcr-CC over that of unstapled CCmut3, and almost all candidates feature exceptional stabilization of the native α-helical conformation in the bound condition. However, it is important to consider the behavior of peptides in the unbound form in solution to elucidate any conformational instabilities which might exist to limit ligand activity in vivo. Because it has been demonstrated that unstapled CCmut3 maintains an almost entirely helical conformation when bound to Bcr-CC31 (Tables S2-S5), ideal stapled constructs will feature enhanced (or equivalent) α-helicity along the peptide backbone, as stapled CCmut3 analogues which do not maintain helicity in solution would theoretically suffer a conformational penalty in adopting the helical conformation upon binding and presumed to be less biologically active.26 Surprisingly, the majority of the stapled CCmut3 analogues tend to deviate from the native α-helical peptide conformation via a bending or folding of the peptide backbone to permanently form a ‘collapsed’ peptide conformation (Figure 3). Fifty-eight of the sixty-two


Page 17 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


16

Figure 3. (A) CCmut3 exhibits an extended peptide conformation, while the bulk of stapled candidates diverge from the native folded structure in solution, adopting one of various ‘collapsed’ peptide conformations (B, C).

stapled CCmut3 constructs are observed to permanently form a collapsed conformation, leaving four original candidates which preserve the native ‘extended’ peptide conformation. The folding events are observed to accompany a marked loss in α-helicity in the peptide backbone, where the strongest losses in α-helicity and ensuing peptide backbone collapse occurs between peptide residues Glu12-Gln24, independent of staple type. The collapsed conformations were realized following the observation of a major shift in both peptide RMSD (Figure S10) and radius of gyration (Figure S11) in affected candidates, indicating a significant re-ordering of peptide secondary structure. Clustering analyses were used



Page 18 of 37

17 to accumulate the most frequently-sampled conformations of the stapled constructs in solution, revealing high populations of collapsed peptide conformations across the vast majority of the stapled CCmut3 analogues. While the unstapled CCmut3 control was observed to repeatedly explore various ‘bent’ peptide conformations for ~20-30 ns per event, a collapsed conformation was not observed within the unstapled CCmut3 control throughout an accumulated 4.3 µs of sampling (between three independent simulations), further supporting the notion that an optimized staple location is critical for maintaining CCmut3 helical integrity.57,58 The collapsing events observed in the stapled CCmut3 constructs are additionally corroborated by research performed by Wang & Bruno et al., in which circular dichroism (CD) experiments confirm a marked loss in alpha-helicity in single[urea]-stapled analogue SUR-010 and double[urea]-stapled analogue DUR-190, and size-exclusion chromatography (SEC) results support a two-state model in the SUR-010 and DUR-190 constructs, and a single-state model in the unstapled CCmut3 control.31 While many of the stapled constructs observed to deviate significantly from the native αhelical ‘extended’ conformation retain affinity for Bcr-CC (reasonably explained by the notion that simulations of stapled peptides bound to Bcr-CC began in the ‘optimal’ binding conformation, where peptides were not required to pay an entropic penalty of binding), sharp increases in proteolytic sensitivity are reported in constructs SUR-010 and DUR-190 (which are observed to form the collapsed conformation in solution) in vitro relative to native CCmut3, suggesting that loss of the native α-helical secondary structure and adoption of a collapsed peptide conformation causes an increase proteolytic sensitivity.31 It is possible that candidates suffering loss of secondary structure will evade the reported proteolytic sensitivity and retain some activity in vivo, but the aim of this research is to enhance stabilization of the native


Page 19 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


18 CCmut3 α-helical conformation to decrease the entropic penalty for binding through peptide backbone stapling, such that candidates which feature secondary structure collapse and deviate from the native helical conformation are not ideal.

Examining Stapled Constructs which Retain the Native α-Helical Conformation in Solution In total, four of the stapled CCmut3 analogues retain conformational stability in solution and feature a resistance to formation of the collapsed peptide conformation. Clustering the most highly-populated conformations among these candidates yields nearly identical conformational sampling of the native extended conformation, save for double[hydrocarbon]-stapled peptide DHC-090, in which the most highly sampled conformation in solution is that of a similar ‘bent’ peptide conformation as observed in the unstapled CCmut3 control minor structure (Figure 4). DHC-090 is distinguished among the original 62 stapled CCmut3 constructs that is able to recover from the formation of a collapsed conformation, which is typically observed as a permanent conformational shift in affected candidates. In the case of DHC-090 however, the peptide is observed to repeatedly form a collapsed conformation (Figure S12) in solution for durations of ~20-30 ns per collapsing event before extending to and largely maintaining the bent conformation through the duration of each of three parallel simulations accumulating over 2 µs of sampling (Figure S13). The folding events exclusive to DHC-090 are additionally reflected in Table 2 where a significant loss in overall helicity in solution is observed relative to that of unstapled CCmut3, most notably in residues Glu19-Val22 (Figure 5). Interestingly enough, the loss in DHC-090 helicity occurs within the staple-occluded region Arg16-Asn23. A loss in helicity in this region is in direct contrast to what is typically observed in an experimental setting, where strong increases



Page 20 of 37

19

Figure 4. Clustering analyses reveal the most highly-sampled conformations in unbound stapled CCmut3 analogues (A) single[hydrocarbon]-stapled peptide SHC-120, (B) double[hydrocarbon]stapled peptide DHC-090, (C) double[urea]-stapled peptide DUR-040, and (D) double[urea]stapled peptide DUR-060.

in helicity are characteristically observed in helical constructs between residues incorporated with an i,i+7 hydrocarbon staple.22,25 Previous research in the development of the original CCmut3 construct cites that experimenting with multiple residue mutations within the region which is, in DHC-090 constrained by an i,i+7 hydrocarbon staple, introduced a kink into CCmut3, which could explain the formation of the bent formation observed in DHC-090.17 Despite the loss of helicity in solution, in the bound condition DHC-090 provides a noticeable increase in helicity over that of unstapled CCmut3, and additionally maintains helicity in


Page 21 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


20 residues affected by the loss of helicity in solution. A slight reduction in overall helicity within single[hydrocarbon]-stapled candidate SHC-120 relative to the unstapled control is observed in solution, but no glaring loss in helicity is observed on a per-residue basis and the candidate features enhanced helicity in the bound condition. The two double[urea]-stapled candidates DUR-040 and DUR-060 feature enhanced helicity over that of CCmut3 in both the bound condition and in solution, due especially to the stabilization of N-terminal peptide regions by staples spanning residues 29-36 (DUR-040) and residues 30-37 (DUR-060).

System

Helicity (%) [solution]

Helicity (%) [complex]

CCmut3

79.4 ± 0.7

87.4 ± 2.0

SHC-120

77.2 ± 0.8

89.3 ± 0.2

DHC-090

71.6 ± 2.0

91.3 ± 1.0

DUR-040

80.4 ± 1.1

93.0 ± 0.1

DUR-060

80.7 ± 0.8

88.8 ± 0.2

Table 2. Overall peptide backbone α-helicity of CCmut3 peptide analogues in solution and in complex with Bcr-CC. Hydrocarbon-stapled candidates SHC-120 and DHC-090 feature overall losses in α-helicity in solution however enhanced α-helicity whilst bound to Bcr-CC, and ureastapled candidates DUR-040 and DUR 060 feature enhanced α-helicity both in the bound and unbound conditions.



Page 22 of 37

21

Figure 5. α-helicity of peptide analogues in solution as a function of peptide residue. Candidates SHC-120, DUR-040 and DUR-060 enjoy enhanced α-helicity at the N-termini while candidate DHC-090 suffers a significant loss in α-helicity reflected by the ‘bent’ peptide conformation.

MM-PBSA calculations report significant increases in overall binding affinity for Bcr-CC in hydrocarbon-stapled candidates DHC-090 and SHC-120 relative to unstapled CCmut3 (Table 3). Both urea-stapled candidates DUR-040 and DUR-060 suffer losses in overall affinity for BcrCC despite featuring enhanced helicity in the bound and unbound conditions - suggesting that absolute peptide helicity may not be the number one indicator of binding affinity. Furthermore, regression analyses performed across each of the 62 stapled candidates reports no significant correlation between binding affinity and absolute helicity of peptides in the bound condition, among each of the 62 candidates. (R2HC-Di7 = 0.18, R2UR-Di7 = 0.06, R2HC-Si7 = 0.01, R2UR-Si7 = 0.01).


Page 23 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


22

System

∆∆Gbind [kcal/mol]

CCmut3

0.0 ± 4.8

SHC-120

-5.3 ± 2.1

DHC-090

-9.3 ± 5.6

DUR-040

+2.7 ± 5.3

DUR-060

+9.0 ± 6.5

Table 3. Relative binding energies of stapled CCmut3 analogues bound to Bcr-CC (relative to unstapled CCmut3). While hydrocarbon-stapled candidate peptides SHC-120 and DHC-090 feature a notable increase in affinity for the Bcr-CC receptor relative to the unstapled CCmut3, both urea-stapled candidates DUR-040 and DUR-060 suffer losses in overall binding affinity.

Per-residue free-energy decompositions were performed to further probe the observed differences in overall binding energies between each of the four candidates which retain the native helical conformation and Bcr-CC, allowing us to examine the energetic contributions of individual residues. Each of the stapled CCmut3 variants are observed to retain the hydrophobic ‘knobs-in-holes’ contacts defining the CCmut3:Bcr-CC binding interface, and no significant deviations in individual energetic contributions (relative to unstapled CCmut3) of the involved hydrophobic contacts along the binding interface exist to explain the differences in overall binding affinity. The energetic contributions of residues involved in potential salt bridging interactions were not remarkably affected by staple placement. However, tracking the occupancies of salt bridge interactions throughout each of the unique simulations revealed that staple placement along the backbone does affect salt bridge formation between CCmut3 and BcrCC, as the eight identified potential CCmut3:Bcr-CC salt bridges (Table 1) were variably occupied among each uniquely-stapled candidates. No correlation could be made however



Page 24 of 37

23 between staple placement (relative to salt bridge location) and salt bridge occupancy such that while the underlying mechanism remains unclear, it is obvious that salt bridge formation is influenced by staple placement in CCmut3. Though a direct correlation could not be made between salt bridge energetic contributions and relative binding affinities, a single charge-charge interaction was able to explain the significant loss in affinity of urea-stapled peptide DUR-060 for Bcr-CC. The original CCmut3 peptide is defined by six point mutations which differentiate the ligand from Bcr-CC, each of which exist to either stabilize the heterodimer (CCmut3:Bcr-CC) or destabilize the homodimer (CCmut3:CCmut3) and prevent self-oligomerization. Mutation L45D in particular was designed to destabilize the homodimer via a D45:D45 charge-charge repulsion at the cost of potentially destabilizing the heterodimer through the destruction of several hydrophobic contacts along the ‘knobs-in-holes’ interface (Leu45 is positioned at position ‘d’ along the heptad repeat). However, the per-residue energetic breakdown reveals that the L45D mutation may have greater implications than just the loss of the wild-type hydrophobic contacts, as residue Asp45 of CCmut3 charges the CCmut3:Bcr-CC heterodimer with an energetic penalty of +3.7 kcal/mol to overall binding affinity. While this was largely the extent of the penalty in the unstapled control and candidates SHC-120, DHC-090, and DUR-040 are penalized comparably, in candidate DUR-060 the energetic penalty of residue Asp45 jumps to +16.0 kcal/mol, and a charge-charge repulsion with Bcr-CC residues Glu48 and Glu52 becomes very apparent, where the two glutamic acid residues contribute energetic penalties of +7.6 kcal/mol and +4.1 kcal/mol respectively to the overall binding of DUR-060 and Bcr-CC. Because the DUR-060 C-terminal i,i+7 urea staple lies directly adjacent to Asp45, we can posit that the proximity of the backbone staple directly influences the interaction of Asp45 with both Glu48 and Glu52, perhaps through


Page 25 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


24 limiting Asp45 conformational selection to produce one or more unfavorable electrostatic interactions between Asp45 and Glu48 and/or Glu52 – especially since the energetic penalty observed in the presence of the adjacent backbone staple is much more extensive than that observed between unstapled CCmut3 and Bcr-CC. These results suggest that the presence of a backbone molecular staple immediately adjacent or near bulky side-chains or groups which require some degree of conformational freedom may be detrimental to ligand:receptor binding through interference of inherent ligand or receptor flexibility.

Staple interaction with the Bcr-CC α1 Flexible Tail Domain Shared among each of the four candidates is the presence of a backbone staple which interacts with the Bcr-CC α1 flexible tail domain (Bcr-CC residues 1-27). It has been demonstrated that the α1 domain provides additional stability to the CCmut3:Bcr-CC heterodimer, and that disrupting contacts between CCmut3 and the α1 domain would be detrimental to heterodimer stability and as such, one question we set out to answer is how the insertion of a staple near α1 domain would influence binding affinity. Among a number of candidates, we noticed two distinct modes of interaction between the staples and the α1 domain: backbone staples were positioned to either ‘drape’ over the α1 domain (and the observed aromatic patch) or insert themselves between the α1 domain and CCmut3 (bridging the aromatic patch) (Figure 6). Energy decomposition analyses reveal that while the insertion of a staple between the aromatic patch is detrimental to heterodimer stability, a staple which is positioned to drape over the Bcr-CC α1 domain serves to enhance overall heterodimer stability. The Bcr-CC α1 domain contributes -16.6 kcal/mol to CCmut3:Bcr-CC stability, but in candidate DUR-040 which



Page 26 of 37

25

Figure 6. Profile and top-down representations of stapled heterodimers are truncated to highlight the (A) draping mechanism of the SHC-120 i,i+7 hydrocarbon staple (orange) interaction with Bcr-CC (purple) and (B) inserting mechanism of DUR-040 i,i+7 urea staple (cyan) interaction with Bcr-CC. Relevant Bcr-CC aromatic residues Phe7, Trp11, and Phe15 are displayed atomistically.


Page 27 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


26 features a urea staple interjecting the Bcr-CC α1 domain, the overall contribution of the α1 domain is decreased, providing -13.4 kcal/mol to heterodimer stability. In the case of SHC-120 where a hydrocarbon staple is draped over the Bcr-CC α1 domain, the α1 domain interacts with the backbone staple much more favorably, providing -20.8 kcal/mol to the overall binding affinity of the SHC-120:Bcr-CC heterodimer. To rule out possible discrepancies due to staple type (rather than staple location), the energy decompositions were additionally performed with DHC-040 and SUR-120. While DHC040 (inserting mechanism) was additionally plagued by a reduction in α1 binding contribution, SUR-120 (draping mechanism) garnered no additional affinity from the α1 domain over the control (Table 4) – revealing that staple type additionally plays a role in the interaction with the α1 domain. The energetic analyses were very telling in this regard: the SHC-120 ‘draped’ hydrocarbon staple-containing residues provide -6.3 kcal/mol to overall SHC-120 affinity for Bcr-CC, whereas the equivalent SUR-120 urea staple-containing residues provide merely -2.9 kcal/mol in binding energy towards heterodimer formation. Together, these results suggest that the draped fully hydrocarbon staple makes several stabilizing contacts with the largely hydrophobic Bcr-CC α1 domain, while the relatively hydrophilic urea staple does not benefit from draping the α1 domain as its equivalent hydrocarbon variant does, remaining relatively inert. Such that the design of a stapled peptide utilizing a fully hydrocarbon staple positioned to drape over the Bcr-CC α1 domain is suggested to provide a noticeable increase in overall binding affinity of the heterodimer, reasonably explaining the net increase in affinity of SHC-120 and DHC-090 for the native Bcr-CC relative to unstapled CCmut3.



Page 28 of 37

27

System

Bcr-CC α1 Contribution (∆G, kcal/mol)

Interaction Type

CCmut3

-16.6 ± 0.9

Unstapled

SHC-120

-20.8 ± 1.0

Draping

SUR-120

-16.1 ± 1.2

Draping

DHC-040

-14.8 ± 2.0

Inserting

DUR-040

-13.4 ± 4.9

Inserting

Table 4. Energetic contribution of the Bcr-CC α1 domain (residues 1-27) to overall heterodimer affinity in candidates featuring a staple which is ‘draped’ over the α1 domain or a staple which is ‘inserted’ into the α1 domain. Draped candidates confer equivalent (SUR-120) or enhanced (SHC-120) in the Bcr-CC α1 domain, whilst inserted candidates suffer an energetic penalty in both hydrocarbon-stapled and urea-stapled variants.

Conclusions We have investigated extensively 62 stapled CCmut3 analogues incorporated by two varied linker types to define a set of constructs which feature both enhanced affinity for native Bcr-CC and stabilization of the native α-helical conformation in solution. Though we had hypothesized that replacement of residues along the CCmut3 peptide backbone with staplecontaining residues would influence Bcr-CC binding, we were surprised to observe such a breakdown of the secondary structure among stapled candidates in solution. Together, our in silico data paired with the SEC and CD experimental data suggest careful optimization of staple type and location along CCmut3 is required in order to retain or enhance the natively α-helical secondary structure, echoing the common sentiment among stapled peptide design that staple


Page 29 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


28 inclusion does not necessarily guarantee structural enhancement.25,31,59 Though peptide helicity was not found to correlate with Bcr-CC affinity in silico, loss in CCmut3 secondary structure has been directly linked with sharp increases in proteolytic sensitivity (which can reasonably justified by the contention that folded proteins are more often resistant to proteolytic mechanisms than equivalent unfolded proteins, which are readily digested by proteases), hence the optimization of a fully helical stapled CCmut3 analogue which retains the native folded α-helical conformation in solution is ideal.31,60–62 It has become abundantly clear, through observations derived of both in silico modeling and physical experiment, that the molecular stapling of the lengthy CCmut3 construct (in its current state) may be wholly detrimental to overall efficacy of the peptide. The inclusion of molecular staples along the backbone of the peptide effectively promotes alternative peptide monomer geometries that are rapidly digested by intracellular proteolytic mechanisms, rendering the stapled peptides wholly inert.31 Four of the 62 stapled peptide monomers that were modeled computationally were observed to retain the native helical conformation in simulations performed out to 1 µs (100,000 simulation frames saved at 10 ps intervals) However, the data suggest that, given enough simulation time, each of the stapled peptide constructs will adopt an alternative geometry featuring a disruption of the native helical geometry. Because stapled peptides tend to ‘fold’ inward to bury the staples (bringing together the N-terminal and C-terminal peptide regions), we can hypothesize that perhaps by decreasing the overall length (trimming off N-terminal or C-terminal residues) of the stapled peptide constructs, we can reduce (or eliminate) the potential for native helical disruption – where an incorporated staple no longer has the ability to interact with residues at the opposite termini (which no longer



Page 30 of 37

29 exist). Utilizing in silico techniques, it would be possible to trim down peptides sequentially until a balance between conformational stability and target affinity can be reached. Binding energetics measured between the diverse set of stapled constructs and the BcrCC target were observed to vary widely dependent on staple location and type, though exploring the effects of stapling on energetic contributions of individual residues and residues sets defined as critical for heterodimer formation (hydrophobic contacts, salt bridges) yielded similar patterns across the stapled constructs. While critical binding partners were typically unaffected by staple inclusion, the addition of a staple occluding Asp45 was observed to dramatically exacerbate an unfavorable charge-charge interaction to the detriment of a collective +27.6 kcal/mol, demonstrating the need to consider the effects of obstructing bulky or charged residue sidechains in optimizing staple placement. Perhaps the greatest indicator of affinity for Bcr-CC was the presence of a hydrocarbon staple neighboring the Bcr-CC α1 domain; where significant increases in heterodimer affinity were observed in candidates featuring a hydrocarbon staple positioned to drape over the largely hydrophobic α1 helix to form additional hydrocarbon contacts along the ‘secondary’ binding interface. The increasingly-hydrophilic urea-stapled peptides were not observed to benefit from such an interaction, implicating hydrocarbon-stapled candidates as collectively more favorable binders of Bcr-CC overall. Peptide therapeutics are becoming increasingly relevant in the treatment of a myriad of disease scenarios. However, the design and optimization processes are still being refined; often hampered by the inherent complexities of peptide-protein interaction and conformational sensitivity of peptides to even minute alterations. Computational methods have the potential to play a large role in the design process of peptide therapeutics in the ability to consider an exceptionally large and diverse set of candidates in a timeframe oftentimes inaccessible in an


Page 31 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


30 experimental setting. By guiding experimental research in investigating only the most promising candidates or candidate sets, computational methods can greatly cut down on the costs and time required to design and optimize lead peptides. While the primary goal of this research is to define a set of next-generation CCmut3 constructs to target the oncogenic activity of Bcr-Abl, literature reporting on the subject of molecular stapling in lengthy biological peptides is rather limited62–64, such that it is our hope that lessons gleaned from this research can be applied in a broader sense to stapled peptide design and optimization.

Supporting Information Available General schema of the CCmut3 peptide. Energetic contributions of hydrophobic and salt bridging interactions. Root-mean-square deviations, radii of gyration, and α-helicical content of monomer peptides and heterodimer complexes. Relative binding affinities of heterodimer complexes. Principle component histograms and root-mean-square fluctuations of various simulations. Stapled residue parameter sets.

Acknowledgements We would like to acknowledge funding and allocations on the Blue Waters Petascale Resource provided by NSF ACI-1521728 and OAC PRAC-151557, as well as time on XSEDE resources through XSEDE allocation MCA017027. We would like to thank Rodrigo Galindo for the valuable insights and scientific discussion.



Page 32 of 37

31

References (1)

Bedi, A.; Zehnbauer, B. A.; Barber, J. P.; Sharkis, S. J.; Jones, R. J. Inhibition of Apoptosis by BCR-ABL in Chronic Myeloid Leukemia. Blood 1994, 83 (8), 2038–2044.

(2)

Puil, L.; Liu, J.; Gish, G.; Mbamalu, G.; Bowtell, D.; Pelicci, P. G.; Arlinghaus, R.; Pawson, T. Bcr-Abl Oncoproteins Bind Directly to Activators of the Ras Signalling Pathway. EMBO J. 1994, 13 (4), 764–773.

(3)

Gordon, M. Y.; Dowding, C. R.; Riley, G. P.; Goldman, J. M.; Greaves, M. F. Altered Adhesive Interactions with Marrow Stroma of Haematopoietic Progenitor Cells in Chronic Myeloid Leukaemia. Nature 1987, 328 (6128), 342–344.

(4)

Goldman, J. M. Chronic Myeloid Leukemia: A Historical Perspective. Semin. Hematol. 2010, 47 (4), 302–311.

(5)

Okimoto, R. A.; Van Etten, R. A. Navigating the Road toward Optimal Initial Therapy for Chronic Myeloid Leukemia. Curr. Opin. Hematol. 2011, 18 (2), 89–97.

(6)

Buchdunger, E.; Zimmermann, J.; Mett, H.; Meyer, T.; Müller, M.; Druker, B. J.; Lydon, N. B. Inhibition of the Abl Protein-Tyrosine Kinase in Vitro and in Vivo by a 2Phenylaminopyrimidine Derivative. Cancer Res. 1996, 56 (1), 100–104.

(7)

Weisberg, E.; Manley, P. W.; Breitenstein, W.; Brüggen, J.; Cowan-Jacob, S. W.; Ray, A.; Huntly, B.; Fabbro, D.; Fendrich, G.; Hall-Meyers, E.; et al. Characterization of AMN107, a Selective Inhibitor of Native and Mutant Bcr-Abl. Cancer Cell 2005, 7 (2), 129–141.

(8)

O’Hare, T.; Shakespeare, W. C.; Zhu, X.; Eide, C. A.; Rivera, V. M.; Wang, F.; Adrian, L. T.; Zhou, T.; Huang, W.-S.; Xu, Q.; et al. AP24534, a Pan-BCR-ABL Inhibitor for Chronic Myeloid Leukemia, Potently Inhibits the T315I Mutant and Overcomes MutationBased Resistance. Cancer Cell 2009, 16 (5), 401–412.

(9)

Wei, G.; Rafiyath, S.; Liu, D. First-Line Treatment for Chronic Myeloid Leukemia: Dasatinib, Nilotinib, or Imatinib. J. Hematol. Oncol. 2010, 3, 47.

(10)

Fausel, C. Targeted Chronic Myeloid Leukemia Therapy: Seeking a Cure. Am. J. Health. Syst. Pharm. 2007, 64 (24 Suppl 15), S9-15.

(11)

Pavlovsky, C.; Kantarjian, H.; Cortes, J. E. First-Line Therapy for Chronic Myeloid Leukemia: Past, Present, and Future. Am. J. Hematol. 2009, 84 (5), 287–293.

(12)

Radich, J. P. Chronic Myeloid Leukemia 2010: Where Are We Now and Where Can We Go? Hematology Am. Soc. Hematol. Educ. Program 2010, 2010, 122–128.

(13)

McWhirter, J. R.; Galasso, D. L.; Wang, J. Y. A Coiled-Coil Oligomerization Domain of Bcr Is Essential for the Transforming Function of Bcr-Abl Oncoproteins. Mol. Cell. Biol. 1993, 13 (12), 7587–7595.

(14)

McWhirter, J. R.; Wang, J. Y. Effect of Bcr Sequences on the Cellular Function of the Bcr-Abl Oncoprotein. Oncogene 1997, 15 (14), 1625–1634.

(15)

Beissert, T.; Hundertmark, A.; Kaburova, V.; Travaglini, L.; Mian, A. a; Nervi, C.; Ruthardt, M. Targeting of the N-Terminal Coiled Coil Oligomerization Interface by a Helix-2 Peptide Inhibits Unmutated and Imatinib-Resistant BCR/ABL. Int. J. Cancer 2008, 122 (12), 2744–2752.


Page 33 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


32 (16)

Dixon, A. S.; Pendley, S. S.; Bruno, B. J.; Woessner, D. W.; Shimpi, A. A.; Cheatham, T. E.; Lim, C. S. Disruption of Bcr-Abl Coiled Coil Oligomerization by Design. J. Biol. Chem. 2011, 286 (31), 27751–27760.

(17)

Dixon, A. S.; Miller, G. D.; Bruno, B. J.; Constance, J. E.; Woessner, D. W.; Fidler, T. P.; Robertson, J. C.; Cheatham, T. E.; Lim, C. S. Improved Coiled-Coil Design Enhances Interaction with Bcr-Abl and Induces Apoptosis. Mol. Pharm. 2012, 9 (1), 187–195.

(18)

Mian, A. A.; Oancea, C.; Zhao, Z.; Ottmann, O. G.; Ruthardt, M. Oligomerization Inhibition, Combined with Allosteric Inhibition, Abrogates the Transformation Potential of T315I-Positive BCR/ABL. Leukemia 2009, 23 (12), 2242–2247.

(19)

Bruno, B. J.; Lim, C. S. Inhibition of Bcr-Abl in Human Leukemic Cells with a CoiledCoil Protein Delivered by a Leukemia-Specific Cell-Penetrating Peptide. Mol. Pharm. 2015, 12 (5), 1412–1421.

(20)

Woessner, D. W.; Eiring, A. M.; Bruno, B. J.; Zabriskie, M. S.; Reynolds, K. R.; Miller, G. D.; O’Hare, T.; Deininger, M. W.; Lim, C. S. A Coiled-Coil Mimetic Intercepts BCRABL1 Dimerization in Native and Kinase-Mutant Chronic Myeloid Leukemia. Leukemia 2015, 29 (8), 1668–1675.

(21)

Jenssen, H.; Aspmo, S. I. Serum Stability of Peptides. Methods Mol. Biol. 2008, 494, 177– 186.

(22)

Verdine, G. L.; Hilinski, G. J. Stapled Peptides for Intracellular Drug Targets. Methods Enzymol. 2012, 503, 3–33.

(23)

Bird, G. H.; Bernal, F.; Pitter, K.; Walensky, L. D. Synthesis and Biophysical Characterization of Stabilized Alpha-Helices of BCL-2 Domains. Methods Enzymol. 2008, 446, 369–386.

(24)

Schafmeister, C. E.; Po, J.; Verdine, G. L. An All-Hydrocarbon Cross-Linking System for Enhancing the Helicity and Metabolic Stability of Peptides. J. Am. Chem. Soc. 2000, 122 (24), 5891–5892.

(25)

Walensky, L. D.; Bird, G. H. Hydrocarbon-Stapled Peptides: Principles, Practice, and Progress. J. Med. Chem. 2014, 57 (15), 6275–6288.

(26)

Guo, Z.; Mohanty, U.; Noehre, J.; Sawyer, T. K.; Sherman, W.; Krilov, G. Probing the Alpha-Helical Structural Stability of Stapled p53 Peptides: Molecular Dynamics Simulations and Analysis. Chem. Biol. Drug Des. 2010, 75 (4), 348–359.

(27)

Bird, G. H.; Irimia, A.; Ofek, G.; Kwong, P. D.; Wilson, I. A.; Walensky, L. D. Stapled HIV-1 Peptides Recapitulate Antigenic Structures and Engage Broadly Neutralizing Antibodies. Nat. Struct. Mol. Biol. 2014, 21 (12), 1058–1067.

(28)

Henchey, L. K.; Jochim, A. L.; Arora, P. S. Contemporary Strategies for the Stabilization of Peptides in the α-Helical Conformation. Curr. Opin. Chem. Biol. 2008, 12 (6), 692– 697.

(29)

Lau, Y. H.; de Andrade, P.; Wu, Y.; Spring, D. R. Peptide Stapling Techniques Based on Different Macrocyclisation Chemistries. Chem. Soc. Rev. 2015, 44 (1), 91–102.

(30)

Wang, Y.; Chou, D. H.-C. A Thiol-Ene Coupling Approach to Native Peptide Stapling and Macrocyclization. Angew. Chemie Int. Ed. 2015, 54 (37), 10931–10934.



Page 34 of 37

33 (31)

Wang, Y.; Bruno, B. J.; Cornillie, S.; Nogieira, J. M.; Chen, D.; Cheatham, T. E.; Lim, C. S.; Chou, D. H.-C. Application of Thiol-yne/Thiol-Ene Reactions for Peptide and Protein Macrocyclizations. Chem. - A Eur. J. 2017, 23 (29), 7087–7092.

(32)

Tan, Y. S.; Lane, D. P.; Verma, C. S. Stapled Peptide Design: Principles and Roles of Computation. Drug Discov. Today 2016, 21 (10), 1642–1653.

(33)

Zhao, X.; Ghaffari, S.; Lodish, H.; Malashkevich, V. N.; Kim, P. S. Structure of the BcrAbl Oncoprotein Oligomerization Domain. Nat. Struct. Biol. 2002, 9 (2), 117–120.

(34)

Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. UCSF Chimera--a Visualization System for Exploratory Research and Analysis. J. Comput. Chem. 2004, 25 (13), 1605–1612.

(35)

Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.; Weissig, H.; Shindyalov, I. N.; Bourne, P. E. The Protein Data Bank. Nucleic Acids Res. 2000, 28 (1), 235–242.

(36)

Dennington, R.; Keith, T.; Millam, J. GaussView, Version 5. Semichem Inc. 2009, Shawnee Mission, KS.

(37)

Vanquelef, E.; Simon, S.; Marquant, G.; Garcia, E.; Klimerak, G.; Delepine, J. C.; Cieplak, P.; Dupradeau, F.-Y. R.E.D. Server: A Web Service for Deriving RESP and ESP Charges and Building Force Field Libraries for New Molecules and Molecular Fragments. Nucleic Acids Res. 2011, 39 (Web Server issue), W511-7.

(38)

Bayly, C. I.; Cieplak, P.; Cornell, W.; Kollman, P. A. A Well-Behaved Electrostatic Potential Based Method Using Charge Restraints for Deriving Atomic Charges: The RESP Model. J. Phys. Chem. 1993, 97 (40), 10269–10280.

(39)

Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Kollmann, P. A. Application of RESP Charges to Calculate Conformational Energies, Hydrogen Bond Energies, and Free Energies of Solvation. J. Am. Chem. Soc. 1993, 115 (21), 9620–9631.

(40)

Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.; Kollman, P. A. A Second Generation Force Field for the Simulation of Proteins, Nucleic Acids, and Organic Molecules J . Am . Chem . Soc . 1995 , 117 , 5179−5197. J. Am. Chem. Soc. 1996, 118 (9), 2309–2309.

(41)

Maier, J. A.; Martinez, C.; Kasavajhala, K.; Wickstrom, L.; Hauser, K. E.; Simmerling, C. ff14SB: Improving the Accuracy of Protein Side Chain and Backbone Parameters from ff99SB. J. Chem. Theory Comput. 2015, 11 (8), 3696–3713.

(42)

Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79 (2), 926.

(43)

Joung, I. S.; Cheatham, T. E. Determination of Alkali and Halide Monovalent Ion Parameters for Use in Explicitly Solvated Biomolecular Simulations. J. Phys. Chem. B 2008, 112 (30), 9020–9041.

(44)

Hopkins, C. W.; Le Grand, S.; Walker, R. C.; Roitberg, A. E. Long-Time-Step Molecular Dynamics through Hydrogen Mass Repartitioning. J. Chem. Theory Comput. 2015, 11 (4), 1864–1874.


Page 35 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


34 (45)

Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; DiNola, A.; Haak, J. R. Molecular Dynamics with Coupling to an External Bath. J. Chem. Phys. 1984, 81 (8), 3684.

(46)

Salomon-Ferrer, R.; Götz, A. W.; Poole, D.; Le Grand, S.; Walker, R. C. Routine Microsecond Molecular Dynamics Simulations with AMBER on GPUs. 2. Explicit Solvent Particle Mesh Ewald. J. Chem. Theory Comput. 2013, 9 (9), 3878–3888.

(47)

Le Grand, S.; Götz, A. W.; Walker, R. C. SPFP: Speed without compromise—A Mixed Precision Model for GPU Accelerated Molecular Dynamics Simulations. Comput. Phys. Commun. 2013, 184 (2), 374–380.

(48)

Pearlman, D. A.; Case, D. A.; Caldwell, J. W.; Ross, W. S.; Cheatham, T. E.; DeBolt, S.; Ferguson, D.; Seibel, G.; Kollman, P. AMBER, a Package of Computer Programs for Applying Molecular Mechanics, Normal Mode Analysis, Molecular Dynamics and Free Energy Calculations to Simulate the Structural and Energetic Properties of Molecules. Comput. Phys. Commun. 1995, 91 (1–3), 1–41.

(49)

Case, D. A.; Cheatham, T. E.; Darden, T.; Gohlke, H.; Luo, R.; Merz, K. M.; Onufriev, A.; Simmerling, C.; Wang, B.; Woods, R. J. The Amber Biomolecular Simulation Programs. J. Comput. Chem. 2005, 26 (16), 1668–1688.

(50)

Loncharich, R. J.; Brooks, B. R.; Pastor, R. W. Langevin Dynamics of Peptides: The Frictional Dependence of Isomerization Rates of N-Acetylalanyl-N’-methylamide. Biopolymers 1992, 32 (5), 523–535.

(51)

Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. A Smooth Particle Mesh Ewald Method. J. Chem. Phys. 1995, 103 (19), 8577.

(52)

Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. . Numerical Integration of the Cartesian Equations of Motion of a System with Constraints: Molecular Dynamics of N-Alkanes. J. Comput. Phys. 1977, 23 (3), 327–341.

(53)

Sindhikara, D. J.; Kim, S.; Voter, A. F.; Roitberg, A. E. Bad Seeds Sprout Perilous Dynamics: Stochastic Thermostat Induced Trajectory Synchronization in Biomolecules. J. Chem. Theory Comput. 2009, 5 (6), 1624–1631.

(54)

Roe, D. R.; Cheatham, T. E. PTRAJ and CPPTRAJ: Software for Processing and Analysis of Molecular Dynamics Trajectory Data. J. Chem. Theory Comput. 2013, 9 (7), 3084– 3095.

(55)

Kollman, P. A.; Massova, I.; Reyes, C.; Kuhn, B.; Huo, S.; Chong, L.; Lee, M.; Lee, T.; Duan, Y.; Wang, W.; et al. Calculating Structures and Free Energies of Complex Molecules: Combining Molecular Mechanics and Continuum Models. Acc. Chem. Res. 2000, 33 (12), 889–897.

(56)

Miller, B. R.; McGee, T. D.; Swails, J. M.; Homeyer, N.; Gohlke, H.; Roitberg, A. E. MMPBSA.py : An Efficient Program for End-State Free Energy Calculations. J. Chem. Theory Comput. 2012, 8 (9), 3314–3321.

(57)

Sim, A. Y. L.; Joseph, T.; Lane, D. P.; Verma, C. Mechanism of Stapled Peptide Binding to MDM2: Possible Consequences for Peptide Design. J. Chem. Theory Comput. 2014, 10 (4), 1753–1761.

(58)

Czabotar, P. E.; Westphal, D.; Dewson, G.; Ma, S.; Hockings, C.; Fairlie, W. D.; Lee, E.



Page 36 of 37

35 F.; Yao, S.; Robin, A. Y.; Smith, B. J.; et al. Bax Crystal Structures Reveal How BH3 Domains Activate Bax and Nucleate Its Oligomerization to Induce Apoptosis. Cell 2013, 152 (3), 519–531. (59)

Bernal, F.; Tyler, A. F.; Korsmeyer, S. J.; Walensky, L. D.; Verdine, G. L. Reactivation of the p53 Tumor Suppressor Pathway by a Stapled p53 Peptide. J. Am. Chem. Soc. 2007, 129 (9), 2456–2457.

(60)

Hubbard, S. J.; Eisenmenger, F.; Thornton, J. M. Modeling Studies of the Change in Conformation Required for Cleavage of Limited Proteolytic Sites. Protein Sci. 1994, 3 (5), 757–768.

(61)

Fontana, A.; Polverino de Laureto, P.; De Filippis, V.; Scaramella, E.; Zambonin, M. Probing the Partly Folded States of Proteins by Limited Proteolysis. Fold. Des. 1997, 2 (2), R17–R26.

(62)

Bird, G. H.; Madani, N.; Perry, A. F.; Princiotto, A. M.; Supko, J. G.; He, X.; Gavathiotis, E.; Sodroski, J. G.; Walensky, L. D. Hydrocarbon Double-Stapling Remedies the Proteolytic Instability of a Lengthy Peptide Therapeutic. Proc. Natl. Acad. Sci. U. S. A. 2010, 107 (32), 14093–14098.

(63)

Giordanetto, F.; Revell, J. D.; Knerr, L.; Hostettler, M.; Paunovic, A.; Priest, C.; Janefeldt, A.; Gill, A. Stapled Vasoactive Intestinal Peptide (VIP) Derivatives Improve VPAC 2 Agonism and Glucose-Dependent Insulin Secretion. ACS Med. Chem. Lett. 2013, 4 (12), 1163–1168.

(64)

Bird, G. H.; Boyapalle, S.; Wong, T.; Opoku-Nsiah, K.; Bedi, R.; Crannell, W. C.; Perry, A. F.; Nguyen, H.; Sampayo, V.; Devareddy, A.; et al. Mucosal Delivery of a DoubleStapled RSV Peptide Prevents Nasopulmonary Infection. J. Clin. Invest. 2014, 124 (5), 2113–2124.


Page 37 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


36

TOC Graphic


Computational Modeling of Stapled Peptides toward a Treatment

Recommend Documents