On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can...

32
1 On the Effect of Cyclization on Peptide Backbone Dynamics Conan K. Wang 1 , Joakim E. Swedberg 1 , Susan E. Northfield 1 , David J. Craik 1,* 1 Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, 4072, Australia *Corresponding Author: Professor David J. Craik Institute for Molecular Bioscience, The University of Queensland, Brisbane, QLD, 4072, Australia Tel: 61-7-3346 2019 Fax: 61-7-3346 2101 e-mail: [email protected]

Transcript of On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can...

Page 1: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

1

On the Effect of Cyclization on Peptide Backbone Dynamics

Conan K. Wang1, Joakim E. Swedberg1, Susan E. Northfield1, David J. Craik1,*

1Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, 4072, Australia

*Corresponding Author:

Professor David J. Craik

Institute for Molecular Bioscience,

The University of Queensland,

Brisbane, QLD, 4072, Australia

Tel: 61-7-3346 2019

Fax: 61-7-3346 2101

e-mail: [email protected]

Page 2: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

2

Abstract

Despite the widespread use of cyclization as a structure optimization tool in peptide chemistry, little is

known about the effect of cyclization on peptide internal dynamics. Here we used a combination of multi-

field NMR relaxation and molecular dynamics techniques to study monocyclic as well as polycyclic

peptides that have promising biopharmaceutical properties – VH, SFTI-1 and cVc1.1 – and their less

constrained analogues to study the effect of backbone cyclization (which forms a macrocycle) and disulfide

bond cyclization (which forms internal cycles). We confirmed that backbone cyclization contributes to the

rigidity of the monocyclic VH. Interestingly though, backbone cyclization of the bicyclic SFTI-1 had a

limited effect on rigidity, with changes in internal dynamics localized around the ligation site. This suggests

that the disulfide bond, which creates an internal cycle, has a insulating effect, protecting the internal cycle

from external motional effects. An insulating effect was also observed for the polycyclic cVc1.1 – the

rigidity of the core was not enhanced by macrocyclization. Additionally, we found that disulfide bonds had a

greater contribution to overall rigidity than macrocyclization. Overall, our results suggest that, although

backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

particularly for polycyclic systems.

Keywords: cyclic peptide, disulfide bond, model free, molecular dynamics, NMR relaxation

Page 3: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

3

Introduction

Cyclization is an important strategy that is widely used to improve the biopharmaceutical properties of

peptide leads in drug discovery.1-3 Cyclization is generally believed to impart improved metabolic stability

to a target peptide, enabling medicinal chemists to exploit the properties of high selectivity, high potency

and low toxicity that are inherent to peptides and underpin the renewed interest in them as drugs. Peptides

can be cyclized in many ways, including through head-to-tail (e.g. backbone cyclization) and side-chain-to-

side-chain (e.g. disulfide bond formation) connections.1 One of the earliest examples of cyclization in drug

design involved a potent cyclic hexapeptide analogue of somatostatin, also known as the Veber-Hirschmann

peptide (VH; Figure 1a and 1c).4 Backbone cyclization not only constrained the peptide into a bioactive

conformation but also improved its metabolic stability, resulting in a cyclic peptide that had both a long

duration of action and oral activity. Subsequent structural studies on other monocyclic peptides have further

shown how backbone cyclization can constrain peptides into bioactive conformations.5

We recently extended the concept of cyclization to disulfide-rich peptides by re-engineering a conotoxin

from the venom of Conus victoriae (cone snail) to make it backbone cyclic by adding a seven-residue linker

to connect the N- and C-termini (Figure 1a and 1c).6 The engineered cyclic peptide showed potent and

selective inhibition of voltage-activated calcium channels associated with pain responses. Whereas the

backbone acyclic peptide was inactive via the oral delivery route in an animal model of neuropathic pain, the

backbone cyclic peptide had oral activity and was more than two orders of magnitude more potent than

gabapentin, the current leading therapy for neuropathic pain.6 Along with this example demonstrating the

benefits of cyclization, there are now a growing number of cyclic peptides that have been reported to have

potent bioactivities and/or high absorption within the gastro-intestinal tract.7-10

The benefits of cyclization have been exploited not only by medicinal chemists but also by nature. To

understand the advantages of cyclization, we have conducted structural studies on a wide range of cyclic

peptides found in bacteria, fungi, plants and mammals.10 In one study, we examined the effect of backbone

Page 4: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

4

cyclization on the structure and activity of sunflower trypsin inhibitor-1 (SFTI-1; Figure 1a and 1b), a cyclic

peptide that comprises one cross-bracing disulfide bond and is the smallest and most potent known inhibitor

of trypsin, and found that cyclization was integral to its proteolytic stability and inhibitory activity.11,12

Another example of a plant-derived cyclic disulfide-rich peptide is kalata B1, originally discovered as the

active uterotonic agent in an African medicinal tea and later shown to be resistant to adverse enzymatic,

thermal and chemical conditions.13 The remarkable in vitro enzymatic, chemical and thermal stability of

these peptides, which have been attributed to their cyclic backbones and disulfide-rich nature, makes them

excellent scaffolds in drug development.10 Indeed, they have led to the design of several leads with potent

activities in animal models of disease, including pain, multiple sclerosis and cancer.10

Although the importance of cyclization as a design tool is well-established, the interplay between cyclization

and peptide backbone dynamics is poorly understood. For proteins, dynamics is intimately linked to the

folding free energy14 and other aspects of protein behavior in solution, including thermostability15 and

solubility.16 Dynamics can also regulate protein activity and, importantly, in ways that sometimes cannot be

predicted based on the protein's ground-state structure.17 For example, internal dynamics can have a

significant effect on binding to a target protein even if no conformational changes are observed upon

binding.17 Specifically, changes in internal dynamics can have a marked effect on the strength of the

enthalpy-entropy compensation that contributes to the free energy of binding. Therefore, a study of

dynamics can uncover cryptic structural information that is fundamental to understanding function.

Nuclear magnetic resonance (NMR) relaxation is a powerful technique that is well-suited for the

characterization of the internal and overall (diffusion controlled) motions of molecules (Figure 1b).18-20

Combined with advances in molecular dynamics simulation, NMR relaxation can provide a comprehensive

and informative picture of the dynamics of a biomolecular system.21 Here we compared the backbone

dynamics of a series of representative cyclic peptides with that of their acyclic derivatives (i.e. backbone and

disulfide variants). We focused on three sets of peptides of varying size and amino acid content: the Veber-

Hirschmann peptide (VH) and its acyclic analogue ([lin]VH); SFTI-1, its backbone acyclic analogue

Page 5: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

5

([lin]SFTI-1), and a disulfide bond-deleted mutant ([C3A,C11A]SFTI-1); and cVc1.1 and its linear analogue

Vc1.1, as well as a series of disulfide-bond deletion/substitution analogues ([C2A,C8A]cVc1.1,

[C2A,C8A]Vc1.1 and [C2H,C8F]cVc1.1). We obtained multi-field NMR relaxation measurements for these

peptides and derived parameters describing their backbone dynamics. We evaluated these results with

respect to molecular dynamics simulations in explicit solvent.

Page 6: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

6

Figure 1: Structures and sequences of the peptides in this study. a) Structures of cyclic peptides with

selected residues labeled. The site where the linear peptides are joined to create the cyclic analogue is

shown for VH and SFTI-1. The linker sequence is highlighted for cVc1.1 with a dotted line. A

schematic of each cyclic peptide is shown to the right of its structure. b) Schematic of some of the

types of motion that can be measured using NMR. For internal motion, the order parameter, S2, for the

Cα-Hα backbone vector can be used as an indicator of molecular flexibility. c) Sequences and disulfide

bond connectivity. Backbone cyclic peptides are indicated with the term 'cyclo'. Residues are shown as

single letter amino acid codes. Cysteine residues and substituted residues are highlighted in bold. The

cystine connectivity is shown. The asterisk indicates an amidated C-terminus.

Page 7: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

7

Methods and Materials

Materials

N-terminal protected amino acids, resins and coupling agents were purchased from ChemImpex

International, and all other reagents were purchased from Auspep, Merck and Sigma, and used without

further purification. D2O (D, 99.96%) and acetonitrile-d3 (D, 99.8%) were purchased from Cambridge

Isotope Laboratories, Inc.

Peptide synthesis and purification

VH, [lin]VH, [lin]SFTI-1, [C3A,C11A]SFTI-1, Vc1.1, [C2A,C8A]cVc1.1, and [C2A,C8A]Vc1.1 were

synthesized on 2-chlorotrityl chloride (2CTC) resin on either 0.125 or 0.25 mmol scale using

fluorenylmethyloxycarbonyl chloride (Fmoc)-chemistry and cleaved from the resin, as described

previously.6,22 Head-to-tail cyclization of [lin]VH was performed in DMF (2 mM) with 3 equiv. HATU and

6 equiv. DIPEA for 6–16 h, and the solvent was subsequently removed. Side-chain protecting groups of VH

and the linear peptides were removed in 94:3:3 TFA/TIPS (triisopropylsilane)/water and purification was

achieved using reverse-phase (RP) preparative high performance liquid chromatography (HPLC). Purity of

fractions was assessed using electron spray ionization-mass spectrometry (ESI-MS) and analytical HPLC.

[lin]SFTI-1, Vc1.1, [C2A,C8A]cVc1.1, and [C2A,C8A]Vc1.1 were oxidized at room temperature in acidic

conditions in the presence of iodine for 20 min. ESI-MS was used to determine reaction completion and

reactions were quenched with L-ascorbic acid. Oxidized peptides were purified by RP-HPLC and purity of

fractions assessed by ESI-MS and analytical HPLC. Synthesis and purification of [C2H,C8F]cVc1.1 has

been described previously.23

SFTI-1 and cVc1.1 were synthesized on Boc-Gly-PAM resin with a thioester linker on 0.5 mmol scale using

tert-butyloxycarbonyl (Boc)-chemistry, as described previously.22 Following peptide cleavage from the resin

using hydrofluoric acid (HF), the crude peptides were purified by RP-HPLC. Cyclization and oxidation were

subsequently carried out in a one-pot reaction in NH4HCO3 buffer (0.1 M, pH 8.5) overnight. The peptides

Page 8: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

8

were purified by reverse-phase (RP) preparative high performance liquid chromatography (HPLC). Purity of

fractions was assessed using ESI-MS and analytical HPLC.

NMR relaxation theory

The NMR parameters T1 (= 1/R1), T2 (=1/R2), and NOE are related by equations (1), (2), and (3) to the

spectral density function, J(ω), which describes the energy available for transitions at different frequencies

and is dependent on the amplitudes and frequencies of overall and internal motions.

𝑅𝑅1 = 14�𝜇𝜇0

2𝛾𝛾𝐶𝐶2𝛾𝛾𝐻𝐻

2ℎ2

16𝜋𝜋2𝑟𝑟𝐶𝐶𝐻𝐻2 � [𝐽𝐽(𝜔𝜔𝐻𝐻 − 𝜔𝜔𝐶𝐶) + 3𝐽𝐽(𝜔𝜔𝐶𝐶) + 6𝐽𝐽(𝜔𝜔𝐻𝐻 + 𝜔𝜔𝐶𝐶)] + 𝑐𝑐2𝐽𝐽(𝜔𝜔𝐶𝐶) (1)

𝑅𝑅2 = 18�𝜇𝜇0

2𝛾𝛾𝐶𝐶2𝛾𝛾𝐻𝐻

2ℎ2

16𝜋𝜋2𝑟𝑟𝐶𝐶𝐻𝐻2 � [𝐽𝐽(𝜔𝜔𝐻𝐻 − 𝜔𝜔𝐶𝐶) + 3𝐽𝐽(𝜔𝜔𝐶𝐶) + 6𝐽𝐽(𝜔𝜔𝐻𝐻 +𝜔𝜔𝐶𝐶) + 4𝐽𝐽(0) + 6𝐽𝐽(𝜔𝜔𝐻𝐻)] + 𝑐𝑐2

6[4𝐽𝐽(0) + 3𝐽𝐽(𝜔𝜔𝐶𝐶)] + 𝑅𝑅𝑒𝑒𝑒𝑒

(2)

𝑁𝑁𝑁𝑁𝑁𝑁 = 1 + 14𝛾𝛾𝐻𝐻𝛾𝛾𝐶𝐶�𝜇𝜇0

2𝛾𝛾𝐶𝐶2𝛾𝛾𝐻𝐻

2ℎ2

16𝜋𝜋2𝑟𝑟𝐶𝐶𝐻𝐻2 � �6𝐽𝐽(𝜔𝜔𝐻𝐻+𝜔𝜔𝐶𝐶)−𝐽𝐽(𝜔𝜔𝐻𝐻−𝜔𝜔𝐶𝐶)

𝑅𝑅1� (3)

Here μ0 is the permeability of free space, h is Planck’s constant, γC and γH are the gyromagnetic ratios for

13C and 1H, respectively, ωC and ωH are the Larmour frequencies of 15C and 1H, respectively, rCH is the C−H

internuclear distance (1.09 Å), Rex is the chemical exchange contribution, and c = ωCΔσ/√3, where Δσ is the

chemical shift anisotropy for 13C nuclei (25 ppm). In the model-free formalism, developed by Lipari and

Szabo,24 J(ω) is defined in terms of motional parameters according to the following equation:

𝐽𝐽(ω) = 25� 𝑆𝑆2τ𝑜𝑜1+(ωτ𝑚𝑚)

+ (1−𝑆𝑆2)τ1+(ωτ)2� (4)

where the overall correlation time, τo, describes the overall molecular tumbling; the generalized order

parameter, S2, describes the amplitude of internal motion; and the internal correlation time, τi, describes the

rate of internal motion on the ps – ns timescale. In addition, τ = τoτi / (τo + τi).

Page 9: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

9

Under certain limiting assumptions, different motional parameters dominate in terms of their contribution to

the spectral density function, giving rise to different motional models. Four standard motional models were

investigated in this study: model 1 (τo, S2), model 2 (τo, S2 and τi), model 3 (τo, S2, and Rex) and model 4 (τo,

S2, τi and Rex).

NMR spectroscopy

NMR samples were prepared by dissolving dried peptide in 100% (v/v) D2O or 30% (v/v) acetonitrile-d3,

70% (v/v) D2O. Acetonitrile was included in the solvent to improve peptide solubility and prevent

aggregation. VH was prepared at 3.4 mM and pH* 2.6; [lin]VH at 3.4 mM and pH* 2.7; SFTI-1 at 3.3 mM

and pH* 2.6; [lin]SFTI-1 at 3.5 mM and pH* 2.9; [C3A,C11A]SFTI-1 at 3.2 mM and pH* 2.9; cVc1.1 at

3.4 mM and pH* 2.5; Vc1.1 at 3.4 mM and pH* 2.5; [C2A,C8A]cVc1.1 at 3.8 mM and pH* 2.7;

[C2A,C8A]Vc1.1 at 3.9 mM and pH* 2.9; and [C2H,C8F]cVc1.1 at 3.4 mM and pH* 2.8. Concentrations of

SFTI-1 and its analogues were determined based on weight. For all other peptides, concentrations were

determined based on their absorbance at 280 nm. The pH of samples was measured using a micro pH

combination electrode (Sigma-Aldrich) and adjusted using small amounts of concentrated DCl and NaOD.

pH* is the direct reading of the pH in a D2O solution of the H2O-calibrated pH-meter. The conversion of

pH* into pD was accomplished by adding a constant value of 0.4. NMR spectra were recorded on Bruker

Avance 500, 600, or 900 MHz spectrometers at 298 K. Spectra were processed with TopSpin (Bruker) using

a zero scaling factor, phased, and calibrated and then assigned with CCPNMR software. Chemical shifts in

the 1H dimension were referenced to internal 4,4-dimethyl-4-silapentane-1-sulfate (DSS) and those in the

13C dimension were indirectly referenced to DSS. Assignment of the 13C chemical shifts was achieved using

a 13C HSQC-TOCSY spectrum.

Diffusion measurements

Translational diffusion coefficients were measured as described previously.25 Briefly, PFG-NMR spectra

were acquired at 25ºC on a Bruker Avance 500 MHz spectrometer. Diffusion coefficients were measured by

incrementing either the duration of the field gradient pulses or the amplitude while the separations of the

Page 10: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

10

field gradients and the total echo time remained constant. Translational diffusion coefficients can be used to

identify the oligomeric state of the peptide sample; knowledge of the oligomeric state is important for

interpreting NMR relaxation data.26,27

Amide temperature coefficients

Amide temperature coefficients were measured as described previously.8 Briefly, amide proton chemical

shifts were derived from 1D and TOCSY spectra recorded on a Bruker Avance 500 MHz spectrometer at

temperatures between 288 K and 308 K, in 5 K increments. Assignment of the spectra was performed using

the program CCPNMR.

Relaxation measurements

Relaxation measurements were carried out as described previously.28 T1 (spin-lattice) and T2 (spin-spin) 13C

relaxation times were measured using the Bruker pulse programs hsqct1etgpsi3d.2 and hsqct2etgpsi3d.2,

respectively. For VH and [lin]VH, NMR spectra were acquired with a spectral width of 10 ppm over 2048

complex points in the 1H dimension and 10 ppm over 40 complex points in the 13C dimension. For all other

peptides, NMR spectra were acquired with a spectral width of 10 ppm over 2048 complex points in the 1H

dimension and 22 ppm over 60 complex points in the 13C dimension. To determine T1 values, 12 relaxation

delays in the range 0.01 to 2.5 s were used, whereas, to determine T2 values, 9 relaxation delays in the range

0.02 to 0.04 s were used. Recycle delays (> 5 T1) of 2 s (500 MHz), 3 s (600 MHz), and 3 s (900 MHz) was

used. Although T1 measurements were conducted at 500, 600 and 900 MHz, we only attempted T2

measurements at 500 MHz due to technical limitations on the other instruments. Peak heights were

measured in CCPNMR and fitted using a two-parameter fit exponential equation to determine T1 or T2

values. Experiments were carried out in triplicate, with a different order of relaxation delays used in each

experiment, and the mean and standard deviation for the T1 or T2 values were derived.

13C NOEs were determined from the ratio of peak intensities from 2D 1H–13C correlation spectra measured

via double inept transfer using sensitivity improvement with decoupling during acquisition, using Bruker

Page 11: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

11

pulse program hsqcnoegpsi.2. For VH and [lin]VH, recycle delays of 4.5 s (> 10 T1 at 500 MHz), 4 s (> 10

T1 at 600 MHz), and 5 s (> 10 T1 at 900 MHz). For SFTI-1, [lin]SFTI-1, and [C3A,C11A]SFTI-1 recycle

delays of 3 s (> 10 T1 at 500 MHz), 3 s (> 10 T1 at 600 MHz) and 5 s (> 10 T1 at 900 MHz) were used. For

cVc1.1 and Vc1.1, recycle delays of 3 s (> 10 T1 at 500 MHz), 3 s (> 10 T1 at 600 MHz) and 3 s (> 6 T1 at

900 MHz) were used. For [C2A,C8A]cVc1.1, [C2A,C8A]Vc1.1 and [C2H,C8F]cVc1.1, recycle delays of 3

s (> 10 T1 at 500 MHz), 4 s (> 10 T1 at 600 MHz) and 5.5 s (> 10 T1 at 900 MHz) were used. NOE on and

NOE off spectra were processed with a zero scaling factor, phased identically and the intensities of the

respective peaks were measured using CCPNMR software. Errors were calculated as the standard deviation

over triplicate measurements.

Relaxation analysis

The relaxation data were fitted to equations described in the ‘model-free’ formalism using the ‘Solver’

function in Microsoft Excel (data presented in Supplementary Tables S3, S6, and S9) and the program

Modelfree 4.1529,30 (data presented in Supplementary Tables S4, S7, and S10). Errors were estimated by

fitting to simulated data generated using a Monte Carlo algorithm. In the Excel spreadsheet, the fits of the

NOE values were given a 50% weighting compared to the T1 and T2 values to take into account the higher

uncertainty in the experimental NOE measurements. The target function is as follows:

𝜒𝜒2 = ∑ ∑ ��𝑇𝑇1,𝑖𝑖,𝑗𝑗

𝑜𝑜𝑜𝑜𝑜𝑜−𝑇𝑇1,𝑖𝑖,𝑗𝑗𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐�

2

𝜎𝜎𝑇𝑇1,𝑖𝑖,𝑗𝑗2 +

�𝑇𝑇2,𝑖𝑖,𝑗𝑗𝑜𝑜𝑜𝑜𝑜𝑜−𝑇𝑇2,𝑖𝑖,𝑗𝑗

𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐�2

𝜎𝜎𝑇𝑇2,𝑖𝑖,𝑗𝑗2 + 0.5

�𝑁𝑁𝑁𝑁𝑁𝑁𝑖𝑖,𝑗𝑗𝑜𝑜𝑜𝑜𝑜𝑜−𝑁𝑁𝑁𝑁𝑁𝑁𝑖𝑖,𝑗𝑗

𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐�2

𝜎𝜎𝑁𝑁𝑁𝑁𝑁𝑁𝑖𝑖,𝑗𝑗2 �𝑀𝑀

𝑗𝑗𝑁𝑁𝑖𝑖 (5)

Calculation of biophysical properties of cyclic peptides

Diffusion coefficients of the peptides, based on their average conformations from the molecular dynamics

simulations, were predicted using the program HYDROPRO31 with Mode 2 (i.e. shell-model from residue-

level) and the solvent parameters for water (i.e. solvent density of 1 g cm-3 and viscosity of 0.89 cP).

Molecular dynamics

Page 12: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

12

Molecular dynamics simulations were done in 30% (v/v) acetonitrile with the structures of SFTI-1 (PDB ID:

4TTK)22 and cVc1.1 (PDB ID: 4TTL)22 used as starting structures. The structures of [lin]SFTI-1 and

[C3A,C11A]SFTI-1 were modeled using YASARA 14.7.17 based on the structure of SFTI-1. The structure

of Vc1.1 was modeled using YASARA by removing the linker sequence from cVc1.1. The structures of the

disulfide mutants of cVc1.1 were also modeled from the structure of cVc1.1. VH was modeled in CYANA

by using restraints reported previously. The structure of [lin]VH was modeled in CYANA by generating

random structures; then, the effect of adding different number of hydrogen bond restraints based on the

cyclic analogue on the final conformations was examined. Force field parameters for acetonitrile were

obtained from SwissParam (http://www.swissparam.ch/). Protonation states of carboxylic acid groups were

calculated in YASARA for each peptide at the pH determined for NMR measurements. Force field

parameters for protonated C-termini were derived from protonated Asp and Glu residues in the standard

CHARMM27 force field. Each peptide was equilibrated using a stepwise relaxation procedure. All heavy-

atoms were harmonically restrained with a force constant of 2 kcal mol-1 Å-2 before a conjugate gradient

minimization of 500 steps was applied using NAMD v2.10 and CHARMM27 force field parameters. The

simulation was heated to 298 K before gradual release over 2.5 ns under NPT conditions and periodic

boundary conditions. A Langevin thermostat with a damping coefficient of 0.5 ps-1 was used to maintain the

system temperature. The system pressure was maintained at 1 atm using a Langevin piston barostat. The

particle mesh Ewald algorithm was used to compute long-range electrostatic interactions at every time step

and non-bonded interactions were truncated smoothly between 10.5 Å and 12 Å. All covalent hydrogen

bonds were constrained by the SHAKE algorithm (or the SETTLE algorithm for water), permitting an

integration time step of 2 fs. Production runs of 20 ns were carried out using ACEMD. These simulations

were performed under NVT with otherwise identical force field and simulation parameters as above.

Coordinates were saved every 1000 simulation steps producing 10000 frames per run.

Page 13: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

13

Results

Effect of backbone cyclization on dynamics

Using solid phase peptide chemistry, we initially synthesized three sets of peptides, which ranged in size

from six to 22 amino acids to examine the effect of backbone cyclization on backbone dynamics (Figure 1).

VH differs from [lin]VH by a single peptide bond between Thr-6 and Phe-1. Similarly, SFTI-1 differs from

[lin]SFTI-1 by a peptide bond between Asp-14 and Gly-1. Thus, both VH and SFTI-1 contain the same

amino acid sequence as their respective linear counterparts, differing in mass by only that of a single water

molecule. Unlike these two sets of peptides, cVc1.1 and Vc1.1 differ in sequence by seven amino acids,

which were introduced to bridge the gap that separates the N- and C-termini of Vc1.1.

To obtain a preliminary assessment of the effect of cyclization/linearization on the backbone dynamics of

the peptides, we began by analyzing their NMR chemical shifts, which provide information on both

structure and dynamics on the ms timescale.32 For the cyclic analogues, the Hα–Cα cross-peaks as well as

the backbone amide resonances were well-defined and corresponded with a single conformation, with the

exception of VH, where a second minor conformer (undergoing slow exchange with the major conformer

and existing at an abundance of approximately 10% according to peak height) was observed, as shown in

Supplementary Figures S2 and S3). Notably, the presence of conformational heterogeneity was more

pronounced for the linear analogues. This was most evident for [lin]VH, where a second conformer was

observed with an abundance of approximately 30%. For [lin]SFTI-1, a second minor conformer (~12%;

Supplementary Figure S2b) was also observed, whereas none was observed for Vc1.1 (Supplementary

Figure S2c). In addition to increased conformational heterogeneity, the linear analogues had some

differences in their Hα and Cα chemical shifts compared to those of the cyclic peptides, indicating changes

in structure upon cyclization. For [lin]VH, these differences in chemical shifts occurred for all residues. The

most significant differences between [lin]SFTI-1 and SFTI-1 were for residues near or at the termini (i.e.

Gly-1, Arg-2, Cys3, Phe-12 and Pro13). Similarly, for Vc1.1, the most significant differences were also near

or at the termini (i.e. Gly-1, Cys-2, Cys-3, Ser-4 and Cys-16).

Page 14: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

14

To determine the effect of cyclization/linearization on the hydrogen bond network, we measured

temperature coefficients for the backbone amides of the peptides (Figure 2). For VH/[lin]VH and SFTI-

1/[lin]SFTI-1, the linear peptides have weaker hydrogen bond networks; for example, Phe-3 and Thr-6 of

[lin]VH have amide temperature coefficients more negative than that of VH. For cVc1.1/Vc1.1,

linearization/cyclization has a mixed effect on the hydrogen bond network; for example, Cys-3 has an amide

temperature coefficient that is less negative in the cyclic peptide, whereas Asn-9 has an amide temperature

coefficient that is more negative in the cyclic peptide.

Figure 2: Amide temperature coefficients (Δδ/ΔT) of cyclic peptides and their backbone acyclic

analogues: VH and [lin]VH (a), SFTI-1 and [lin]SFTI-1 (b), and cVc1.1 and Vc1.1 (c). The sequence

of each peptide is shown at the top of each panel. Amide temperature coefficients for the N-terminal

residue or for proline residues (indicated by asterisks) cannot be measured. Schematic illustrations of

the backbone acyclic peptides showing their disulfide bond content are shown at the top-right of the

figure.

Page 15: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

15

Next, we considered relaxation measurements to characterize ps-ns dynamics. To begin, we confirmed that

the linear and cyclic peptides were predominantly monomeric in the conditions tested by carrying out NMR

translational diffusion coefficient measurements (Supplementary Table S1), which can be used to identify

the oligomeric state of peptides, including those studied herein, via the Stokes-Einstein relation.25 Next, we

measured relaxation parameters for their backbone Cα–Hα bonds at multiple field strengths (i.e. 500 MHz,

600 MHz and 900 MHz). For molecules of similar size and shape, a comparison of their relaxation

parameters can give an indication of similarities/differences in dynamics. More specifically, larger T1, T2

and NOE values are generally indicative of a higher degree of flexibility. As shown in Supplementary

Figure S4 and Supplementary Table S2, relaxation parameters are higher for [lin]VH compared to VH, with

the differences increasing for residues that are closer to the termini of [lin]VH (e.g. Phe-1 and Thr-6). The

residues near the termini of [lin]SFTI-1 also show larger relaxation parameters than SFTI-1 (Supplementary

Figure S5 and Supplementary Table S5). Aside from the terminal residues, the other residues show similar

values for both [lin]SFTI-1 and SFTI-1: as expected, the relaxation parameters for the loop between Ser-6

and Pro-9 are larger than for the residues within the β-strands. Supplementary Figure S6 and Supplementary

Table S8 show that the relaxation parameters are similar for cVc1.1 and Vc1.1 and that the linker region is

more flexible than the main part of cVc1.1. A similar trend was also observed for the backbone N–H bond

(Supplementary Figure S7).

To determine parameters quantifying the backbone dynamics, we used the model-free formalism.24 The

fitted data and their deviations from the observed relaxation data are shown in Supplementary Tables S3, S6,

and S9. To illustrate the effect of cyclization/linearization on peptide backbone dynamics, we mapped the

differences in S2 (which corresponds to differences in the magnitude of motion) between each cyclic peptide

and its linear counterpart onto the structure of the cyclic peptide, as shown in Figure 3. For both [lin]VH and

[lin]SFTI-1, cyclization leads to an increase in rigidity. Whereas for [lin]VH the effect is observed for most

of the molecule, for [lin]SFTI-1 changes in rigidity are more localized near the ligation point. It was initially

surprising to us that cyclization of Vc1.1 did not result in a substantial increase in rigidity but instead led to

Page 16: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

16

a slight decrease in rigidity (i.e. at Cys-3, Cys-8 and Asn-9); though we rationalize this observation later in

the discussion using concepts relevant to the thermodynamics of protein folding.

Figure 3: Differences in order parameter (ΔS2) between cyclic and backbone linear peptides. The

differences in order parameter upon cyclization are mapped onto the structure of the cyclic peptides.

The color map is shown to the left. Gly residues are colored grey because relaxation parameters were

not measured for these residues. The linker sequence of cVc1.1 is also colored grey because Vc1.1

does not have the linker sequence and a comparison between order parameters could not be performed.

The disulfide bonds are colored in yellow. The residues that experienced significant change in their

order parameters are labeled and the percentage change is indicated.

Effect of disulfide bonds on dynamics

Aside from having cyclic backbones, another distinguishing feature of SFTI-1 and cVc1.1 is the presence of

intramolecular disulfide bonds, which form internal cycles within the peptides. Interested in the effect of

these internal connections on dynamics, we synthesized a series of disulfide bond mutants, including

[C3A,C11A]SFTI-1, which contains Ala in place of the Cys residues of the parent peptide and consequently

contains no disulfide bonds, as well as [C2A,C8A]cVc1.1 and [C2A,C8A]Vc1.1, both of which contain only

one disulfide bond compared to the two of cVc1.1. We also synthesized [C2H,C8F]cVc1.1, which is also a

disulfide-deleted mutant but was designed to retain the in vitro serum stability and activity of cVc1.1 by

increasing the hydrophobic core of the peptide.23 The sequences of these peptides are shown in Figure 1.

Page 17: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

17

We compared the dynamics of the disulfide bond mutants to the parent peptides using the same approaches

that we used for the backbone acyclic analogues as described above. Analysis of the chemical shift

deviations indicate that removal of the disulfide bonds leads to changes in the structure and conformational

rigidity compared to the parent peptide. For example, deletion of the disulfide-bond within SFTI-1 reduced

the dispersion of the amide resonances and increased conformational heterogeneity (Supplementary Figures

S8 and S9). As expected and in agreement with our previous report on [C2H,C8F]cVc1.1,23 replacement of

the Cys-2 to Cys-8 disulfide bond with a pair of His and Phe residues is less disruptive to the conformation

of cVc1.1 than replacement with a pair of Ala residues (Supplementary Figures S8 and S9).

Figure 4 shows a comparison of the amide temperature coefficients of the disulfide bond mutants to that of

their parent cyclic peptides. For SFTI-1, removal of the disulfide bond results in an overall weakening of its

hydrogen bond network, with most amides of [C3A,C11A]SFTI-1 (e.g. Thr-4, Ile-10, and Phe-12)

displaying temperature coefficients that are more negative compared to their respective values of SFTI-1.

For cVc1.1 and its disulfide bond analogues, replacement of the Cys-2 to Cys-8 disulfide bond has a mixed

effect on the hydrogen bond network, with some backbone amides of the disulfide bond mutants having

more negative temperature coefficients, while others have less negative temperature coefficients compared

to cVc1.1.

Page 18: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

18

Figure 4: Amide temperature coefficients (Δδ/ΔT) of cyclic peptides and their disulfide-bond

analogues: SFTI-1 and [C3A,C11A]SFTI-1 (a), cVc1.1 and [C2A,C8A]cVc1.1, cVc1.1 and

[C2A,C8A]Vc1.1, and cVc1.1 and [C2H,C8F]cVc1.1 (b). The sequence of SFTI-1 and cVc1.1 is

shown at the top of each panel. Amide temperature coefficients for the N-terminal residue or for

proline residues (indicated by asterisks) cannot be measured. Schematic illustrations of the disulfide-

bond mutants are shown at the right of the figure.

We measured NMR relaxation data for the disulfide bond mutants at 500 MHz, 600 MHz and 900 MHz and

derived their relaxation parameters, as presented in Supplementary Figures S10 and S11 and Supplementary

Tables S6 and S9. The relaxation data suggest that replacement of the disulfide bonds reduces rigidity,

particularly near the substituted residues (Supplementary Figures S10 and S11). For example, Ala-3 of

[C3A,C11A]SFTI-1 has larger T1 values compared to that of Cys-3 of SFTI-1 at all measured field strengths

(Supplementary Figure S10). In agreement with this qualitative analysis of the relaxation data, the fitted

Page 19: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

19

order parameters also indicate that the disulfide bond mutants experience increased internal motion

compared to their respective parent cyclic peptide, as shown in Figure 5. For example, formation of the

single disulfide bond in SFTI-1 results in increases in the order parameters at residues 3 and 11 by 18%.

Formation of the disulfide bond between residues 2 and 9 of cVc1.1 results in increases in rigidity for

several residues in addition to residues 2 and 9 (as indicated from a comparison between [C2A,C8A]cVc1.1

with cVc1.1 and [C2A,C8A]Vc1.1 with cVc1.1). The peptide cVc1.1 is also more rigid than

[C2H,C8F]cVc1.1, though the presence of the His-Phe interaction makes [C2H,C8F]cVc1.1 more rigid than

[C2A,C8A]cVc1.1.

In silico dynamics

We used molecular dynamics simulation to obtain complementary information on the differences in

molecular flexibility between the cyclic peptides and their backbone acyclic and disulfide-bond substituted

forms. Figure 6 shows a superposition of several snapshots taken throughout the simulation run for each

peptide, providing a qualitative picture of peptide dynamics. Order parameters can be calculated from

molecular dynamics simulations, although it should be noted that discrepancies between order parameters

extracted from simulations and experiments have been reported.33 Nevertheless, differences in the calculated

order parameter for each cyclic and linear peptide pair based on the molecular dynamics simulations were

calculated (Supplementary Figure S12), and were compared to those derived from NMR as shown in

Supplementary Figure S13, and showed a general agreement between the NMR and simulated data. For the

backbone linear peptides, there was a high degree of flexibility at the terminal residues; for example, Phe-1

and Thr-6 of [lin]VH, Gly-1 and Asp-14 of [lin]SFTI-1, and Gly-1 of Vc1.1. The cyclic peptides VH and

SFTI-1 showed greater rigidity than their backbone linear counterparts. For cVc1.1 and Vc1.1, the degree of

backbone motion was similar across the core part of the molecule (i.e. Cys-2 to Ile-15). The linker region of

cVc1.1 was very flexible relative to the rest of the molecule. Consistent with the relaxation analysis, the

disulfide bond mutants showed greater flexibility than their parent peptides.

Page 20: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

20

Figure 5: Differences in order parameter (ΔS2) between cyclic and disulfide-bond mutant peptides.

The differences in order parameter upon disulfide bond replacement are mapped onto the structure of

the cyclic peptides. The color map is shown to the top-left. Gly residues are colored grey because

relaxation parameters were not measured for these residues. The disulfide bonds are colored in yellow.

The residues that experienced significant change in their order parameters are labeled and the

percentage change is indicated.

Page 21: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

21

Figure 6: Molecular dynamics simulation of VH and [lin]VH (a), SFTI-1, [lin]SFTI-1, and

[C3A,C11A]SFTI-1 (b), and cVc1.1, Vc1.1, and [C2A,C8A]cVc1.1 (c). Backbone structures of the

cyclic peptides and their analogues at selected time points of the simulation are overlaid onto their

respective initial structures. Selected residues are labeled, and disulfide bonds if present are colored

yellow. Regions of flexibility are highlighted with dotted arrow lines.

Page 22: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

22

Discussion

Cyclization is widely believed to impart improved thermodynamic stability to cyclic peptides relative to

their linear counterparts. Additionally, there have been a large number of studies demonstrating the

beneficial effects of peptide cyclization on activity, metabolic stability and bioavailability.2,6,34-37 Although

cyclization is frequently used for enhancing stability of peptide drug leads, little is known about its effect on

internal dynamics. To address this gap in understanding, we analyzed the structure and dynamics of cyclic

peptides and their linear and less constrained counterparts.

We started with a cyclic hexapeptide analogue of somatostatin, VH, and its linear analogue, [lin]VH. Before

performing the experiments, we expected that the linear peptide would be highly flexible compared to the

cyclic peptide, with all residues undergoing a large degree of molecular motion when the peptide was in its

linear form (Figure 7a). In initial agreement with this prediction, there was increased motion on the ms

timescale as indicated by the higher abundance of a second conformer upon linearization. However, the

global increase in motion over all residues was not reflected in the relaxation analysis. Although the terminal

residues displayed increased motion, we observed that the central residues (i.e. D-Trp-4 and Lys-5) did not

experience any significant change in motion on the ps-ns timescale. This meta-stable structure might be due

to the conformational preference of D-Trp-4 and Lys-5 to participate in a β-turn.38 Indeed, we could re-create

the presence of a meta-stable structure in [lin]VH using molecular dynamics and by using starting structures

that contained the β-turn. The meta-stable β-turn might underpin the reported activity of linear analogues of

VH,39 as the formation of a βII'-turn involving the residues D-Trp-4 and Lys-5 has been shown to be

essential for the biological activity of the VH peptide.36 Overall, this study of VH and [lin]VH demonstrates

that linear peptides can have regions of order due to conformational preferences of constituent residues

(Figure 7b). It is also possible that meta-stable structures are stabilized by hydrogen bonds and/or

electrostatic attractions.37

Page 23: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

23

Similarly, backbone cyclization of [lin]SFTI-1 improved its rigidity. In this case, there was some

improvement in rigidity on the ms timescale and the main effect of cyclization on the ps-ns timescale was on

the rigidity of the terminal residues, i.e. Gly-1, Arg-2, Pro-13 and Asp-14 (schematically represented in

Figure 7c). The difference in rigidity between cyclic and linear forms may explain the order of magnitude

difference in their trypsin-inhibitory activities.11 Aside from differences in motion at the terminal residues,

the core region of [lin]SFTI-1 was largely unaffected by backbone cyclization, probably because it was

already stabilized by a network of hydrogen bonds, as confirmed by amide temperature coefficient

measurements, and a disulfide bond between Cys-3 and Cys-11, which forms an embedded cycle involving

the intervening residues. Comparison between the dynamics of a disulfide-deleted mutant of SFTI-1 and that

of SFTI-1 confirmed that the disulfide bond indeed contributes to the rigidity of SFTI-1 (Figure 7d), and

may have a larger effect on rigidity than backbone cyclization. Indeed, whereas linearized SFTI-1 has

weaker inhibitory activity than SFTI-1, disulfide-bond deleted mutants are inactive.40 It appears that SFTI-1

is an example where nature uses two types of covalent cyclization to enhance activity and confer multiple

levels of rigidity. In support of this notion, we have previously shown that the remarkable chemical and

enzymatic stability of another naturally-occurring peptide, kalata B1, is ascribed to the combination of its

cyclic backbone and disulfide bond connectivity.13

We previously explored the potential of combining a cyclic backbone with disulfide bonds in enhancing

peptide serum stability by connecting the N- and C-termini of a disulfide-rich peptide from the venom of C.

victoriae.6 Encouragingly, the study found that the backbone cyclic form showed increased stability in

human serum and also displayed potent oral activity in an animal model of neuropathic pain. Based on our

earlier promising results, it was thus surprising to find in this study that addition of a linker to cyclize the

backbone does not enhance the rigidity of the core region of Vc1.1 (schematically depicted in Figure 7c).

Instead, we observed the opposite effect, although the effect was marginal, as supported by molecular

dynamics simulations. It would seem then that in this case the biggest benefit of backbone cyclization is to

prevent degradation by exopeptidases, leading to improved proteolytic stability as previously observed.6

Another benefit of backbone cyclization in this case might be the improvement of the energetic properties of

Page 24: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

24

the active structure with the Cys-2 to Cys-16 and Cys-3 to Cys-6 disulfide connectivity, preventing

formation of inactive conformations with alternative disulfide connectivities; the susceptibility of Vc1.16

and other similar peptides to disulfide bond re-arrangements is well-known.41

For both SFTI-1 and cVc1.1, we observed that disulfide bond formation (which creates smaller internal

cycles) provides a significant contribution to structural rigidity, and in fact provides a greater contribution

than from backbone cyclization (which conceptually creates one large macrocycle). Though both strategies

are effective at increasing structural rigidity and could be used in concert for that purpose, our result

suggests that directed segmentation of a peptide into rigid sub-structures is a very effective strategy for

increasing rigidity. One could liken disulfide bonds to reinforcement techniques in construction and

engineering, such as cross-bracing, which re-distribute compression and tension forces. As the formation of

disulfide bonds can significantly improve the structural rigidity of peptides, it is thus not surprising that

disulfide bonds are abundant in nature, occurring in many peptides and proteins,42,43 and also attracting

significant interest in peptide drug design.44 For therapeutic peptides, a disadvantage of disulfide bonds is

their susceptibility to reducing environments in the human body. Therefore, we support the view that non-

reducible disulfide-bond surrogates, which retain the rigidifying effect of the disulfide bond but are

chemically more stable in vivo, will have increasingly wider applications in cyclic peptide drug design.

From a thermodynamics perspective, there are parallels between the effects of cyclization on peptide

stability and on protein stability. It is thought that backbone cyclization and disulfide bond cyclization

improves folding stability of proteins by reducing the entropy of the unfolded state,45,46 and the same

principle could apply to peptides. In a study of the folding stability of a 122 residue analogue of the DnaB

helicase protein from Eschericia coli,47 it was calculated that backbone cyclization stabilizes the native

protein with a free energy of the order of 8 kJ mol-1, which although modest, resulted in an increase in

melting temperature of 15°C. The change in free energy upon cyclization for peptides would depend on both

entropic and enthalpic changes.

Page 25: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

25

Unlike for [lin]VH and [lin]SFTI-1, we did not observe an increase in rigidity of Vc1.1 upon cyclization. In

principle, cyclization does not have to rigidify the peptide, and could have no effect on rigidity or even lead

to increased flexibility. It has been shown that the effect of cyclization on the PIN1 WW domain, a 34

residue structure, was dependent on the length of the linker used to join the N- and C-termini.48 Linkers that

were too short or too long were unfavorable for stability because they introduced strain, which probably

disrupted native interactions and distorted the enthalpy-entropy compensation phenomenon of the native

structure. It is worth noting that the introduction of a linker to Vc1.1 for cyclization also resulted in the

introduction of an entropic element to Vc1.1 as the linker region showed a high degree of flexibility in the

cyclic analogue. Interestingly, increased entropy is not necessarily detrimental to stability, as long as

changes in entropy exceed the loss of enthalpic contributions, as explained in a recent study.49

In conclusion, backbone cyclization and internal cyclization (via disulfide bonds, for example) can improve

the rigidity of a candidate peptide, leading to changes in entropy and enthalpy that might be favorable for

thermodynamic stability. For some peptides, cyclization may act to reinforce existing embedded cycles

within the starting peptide that involve pre-organized structures stabilized by non-covalent (e.g. hydrogen

bonds) or covalent (e.g. disulfide bonds) interactions. In other cases, cyclization may have no effect or a

destabilizing effect on the dynamics of the linear peptides. Whether an effect on rigidity is caused by

cyclization ultimately depends on the interplay between enthalpy and entropy of the cyclic and linear forms.

Page 26: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

26

Figure 7: The effect of cyclization/linearization on backbone dynamics. a) In principle, cyclization can

impart enhanced rigidity to a linear peptide that would otherwise be highly flexible (on the ns-ps and ms

timescale) throughout the peptide chain. b) In some cases, other structural features apart from the cyclic

backbone can stabilize the structure, and these structural features might be responsible for the presence of

transient meta-stable structures/cycles within the linear peptide. c) Disulfide bonds can form an embedded

cycle within the cyclic peptide and this embedded cycle by itself is also structurally rigid. d) Removal of the

disulfide bond increases flexibility.

Page 27: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

27

Associated Content

Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org.

Author Information

Corresponding Author

* Professor David J. Craik. Institute for Molecular Bioscience, The University of Queensland, Brisbane,

Qld, 4072, Australia. Tel: 61-7-3346 2019. Fax: 61-7-3346 2101. e-mail: [email protected]

Funding Sources

CKW was supported by a National Health and Medical Research Council (NHMRC) Early Career Research

Fellowship (546578). JES is an NHMRC Early Career Fellow (APP1069819). DJC is an NHMRC

Professorial Fellow (APP1026501). This work was supported by a grant from the NHMRC (APP1076136).

Acknowledgements

We thank Olivier Cheneval and Phillip Walsh for help with peptide synthesis and Peta Harvey and Greg

Pierens for help with NMR. The authors acknowledge the facilities, and the scientific and technical

assistance, of the Queensland NMR Network Research Facility at the University of Queensland.

Page 28: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

28

References

(1) White, C. J.; Yudin, A. K. Contemporary Strategies for Peptide Macrocyclization. Nat Chem 2011, 3, 509-524.

(2) Heinis, C. Drug Discovery: Tools and Rules for Macrocycles. Nat Chem Biol 2014, 10, 696-698.

(3) Marsault, E.; Peterson, M. L. Macrocycles Are Great Cycles: Applications, Opportunities, and Challenges of Synthetic Macrocycles in Drug Discovery. J Med Chem 2011, 54, 1961-2004.

(4) Veber, D. F.; Freidlinger, R. M.; Perlow, D. S.; Paleveda, W. J., Jr.; Holly, F. W.; Strachan, R. G.; Nutt, R. F.; Arison, B. H.; Homnick, C.; Randall, W. C.et al. A Potent Cyclic Hexapeptide Analogue of Somatostatin. Nature 1981, 292, 55-58.

(5) Kessler, H. Conformation and Biological Activity of Cyclic Peptides. Angew Chem Int Ed Engl 1982, 21, 512-523.

(6) Clark, R. J.; Jensen, J.; Nevin, S. T.; Callaghan, B. P.; Adams, D. J.; Craik, D. J. The Engineering of an Orally Active Conotoxin for the Treatment of Neuropathic Pain. Angew Chem Int Ed Engl 2010, 49, 6545-6548.

(7) Rezai, T.; Yu, B.; Millhauser, G. L.; Jacobson, M. P.; Lokey, R. S. Testing the Conformational Hypothesis of Passive Membrane Permeability Using Synthetic Cyclic Peptide Diastereomers. J Am Chem Soc 2006, 128, 2510-2511.

(8) Wang, C. K.; Northfield, S. E.; Colless, B.; Chaousis, S.; Hamernig, I.; Lohman, R. J.; Nielsen, D. S.; Schroeder, C. I.; Liras, S.; Price, D. A.et al. Rational Design and Synthesis of an Orally Bioavailable Peptide Guided by Nmr Amide Temperature Coefficients. Proc Natl Acad Sci U S A 2014, 111, 17504-17509.

(9) Wang, C. K.; Gruber, C. W.; Cemazar, M.; Siatskas, C.; Tagore, P.; Payne, N.; Sun, G.; Wang, S.; Bernard, C. C.; Craik, D. J. Molecular Grafting onto a Stable Framework Yields Novel Cyclic Peptides for the Treatment of Multiple Sclerosis. ACS Chem Biol 2014, 9, 156-163.

(10) Northfield, S. E.; Wang, C. K.; Schroeder, C. I.; Durek, T.; Kan, M. W.; Swedberg, J. E.; Craik, D. J. Disulfide-Rich Macrocyclic Peptides as Templates in Drug Design. Eur J Med Chem 2014, 77, 248-257.

(11) Korsinczky, M. L.; Schirra, H. J.; Rosengren, K. J.; West, J.; Condie, B. A.; Otvos, L.; Anderson, M. A.; Craik, D. J. Solution Structures by 1h Nmr of the Novel Cyclic Trypsin Inhibitor Sfti-1 from Sunflower Seeds and an Acyclic Permutant. J Mol Biol 2001, 311, 579-591.

(12) Colgrave, M. L.; Korsinczky, M. J.; Clark, R. J.; Foley, F.; Craik, D. J. Sunflower Trypsin Inhibitor-1, Proteolytic Studies on a Trypsin Inhibitor Peptide and Its Analogs. Biopolymers 2010, 94, 665-672.

(13) Colgrave, M. L.; Craik, D. J. Thermal, Chemical, and Enzymatic Stability of the Cyclotide Kalata B1: The Importance of the Cyclic Cystine Knot. Biochemistry 2004, 43, 5965-5975.

(14) Li, D. W.; Showalter, S. A.; Bruschweiler, R. Entropy Localization in Proteins. J Phys Chem B 2010, 114, 16036-16044.

Page 29: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

29

(15) Vinther, J. M.; Kristensen, S. M.; Led, J. J. Enhanced Stability of a Protein with Increasing Temperature. J Am Chem Soc 2011, 133, 271-278.

(16) De Simone, A.; Dhulesia, A.; Soldi, G.; Vendruscolo, M.; Hsu, S. T.; Chiti, F.; Dobson, C. M. Experimental Free Energy Surfaces Reveal the Mechanisms of Maintenance of Protein Solubility. Proc Natl Acad Sci U S A 2011, 108, 21057-21062.

(17) Tzeng, S. R.; Kalodimos, C. G. Protein Activity Regulation by Conformational Entropy. Nature 2012, 488, 236-240.

(18) Palmer, A. G., 3rd; Williams, J.; McDermott, A. Nuclear Magnetic Resonance Studies of Biopolymer Dynamics. J. Phys. Chem. 1996, 100, 13293-13310.

(19) Kay, L. E. Nmr Studies of Protein Structure and Dynamics - a Look Backwards and Forwards. J Magn Reson 2011, 213, 492-494.

(20) McElheny, D.; Schnell, J. R.; Lansing, J. C.; Dyson, H. J.; Wright, P. E. Defining the Role of Active-Site Loop Fluctuations in Dihydrofolate Reductase Catalysis. Proc Natl Acad Sci U S A 2005, 102, 5032-5037.

(21) Kasinath, V.; Sharp, K. A.; Wand, A. J. Microscopic Insights into the Nmr Relaxation-Based Protein Conformational Entropy Meter. J Am Chem Soc 2013, 135, 15092-15100.

(22) Wang, C. K.; King, G. J.; Northfield, S. E.; Ojeda, P. G.; Craik, D. J. Racemic and Quasi-Racemic X-Ray Structures of Cyclic Disulfide-Rich Peptide Drug Scaffolds. Angew Chem Int Ed Engl 2014, 53, 11236-11241.

(23) Yu, R.; Seymour, V. A.; Berecki, G.; Jia, X.; Akcan, M.; Adams, D. J.; Kaas, Q.; Craik, D. J. Less Is More: Design of a Highly Stable Disulfide-Deleted Mutant of Analgesic Cyclic Alpha-Conotoxin Vc1.1. Sci Rep 2015, 5, 13264.

(24) Lipari, G.; Szabo, A. Model-Free Approach to the Interpretation of Nuclear Magnetic Resonance Relaxation in Macromolecules. 1. Theory and Range of Validity. J Am Chem Soc 1982, 104, 4546-4559.

(25) Wang, C. K.; Northfield, S. E.; Swedberg, J. E.; Harvey, P. J.; Mathiowetz, A. M.; Price, D. A.; Liras, S.; Craik, D. J. Translational Diffusion of Cyclic Peptides Measured Using Pulsed-Field Gradient Nmr. J Phys Chem B 2014, 118, 11129-11136.

(26) Fushman, D.; Cahill, S.; Cowburn, D. The Main-Chain Dynamics of the Dynamin Pleckstrin Homology (Ph) Domain in Solution: Analysis of 15n Relaxation with Monomer/Dimer Equilibration. J Mol Biol 1997, 266, 173-194.

(27) Bernado, P.; Akerud, T.; Garcia de la Torre, J.; Akke, M.; Pons, M. Combined Use of Nmr Relaxation Measurements and Hydrodynamic Calculations to Study Protein Association. Evidence for Tetramers of Low Molecular Weight Protein Tyrosine Phosphatase in Solution. J Am Chem Soc 2003, 125, 916-923.

(28) Conibear, A. C.; Wang, C. K.; Bi, T.; Rosengren, K. J.; Camarero, J. A.; Craik, D. J. Insights into the Molecular Flexibility of Theta-Defensins by Nmr Relaxation Analysis. J Phys Chem B 2014, 118, 14257-14266.

Page 30: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

30

(29) Mandel, A. M.; Akke, M.; Palmer, A. G., 3rd. Backbone Dynamics of Escherichia Coli Ribonuclease Hi: Correlations with Structure and Function in an Active Enzyme. J Mol Biol 1995, 246, 144-163.

(30) Palmer, A. G., 3rd; Rance, M.; Wright, P. E. Intramolecular Motions of a Zinc Finger DNA-Binding Domain from Xfin Characterized by Proton-Detected Natural Abundance Carbon-13 Heteronuclear Nmr Spectroscopy. J Am Chem Soc 1991, 113, 4371-4380.

(31) Ortega, A.; Amoros, D.; Garcia de la Torre, J. Prediction of Hydrodynamic and Other Solution Properties of Rigid Proteins from Atomic- and Residue-Level Models. Biophys J 2011, 101, 892-898.

(32) Palmer, A. G., 3rd; Kroenke, C. D.; Loria, J. P. Nuclear Magnetic Resonance Methods for Quantifying Microsecond-to-Millisecond Motions in Biological Macromolecules. Methods Enzymol 2001, 339, 204-238.

(33) Maragakis, P.; Lindorff-Larsen, K.; Eastwood, M. P.; Dror, R. O.; Klepeis, J. L.; Arkin, I. T.; Jensen, M. O.; Xu, H.; Trbovic, N.; Friesner, R. A.et al. Microsecond Molecular Dynamics Simulation Shows Effect of Slow Loop Dynamics on Backbone Amide Order Parameters of Proteins. J Phys Chem B 2008, 112, 6155-6158.

(34) Bagnacani, V.; Franceschi, V.; Bassi, M.; Lomazzi, M.; Donofrio, G.; Sansone, F.; Casnati, A.; Ungaro, R. Arginine Clustering on Calix[4]Arene Macrocycles for Improved Cell Penetration and DNA Delivery. Nat Commun 2013, 4, 1721.

(35) Gunzburg, M. J.; Ambaye, N. D.; Del Borgo, M. P.; Perlmutter, P.; Wilce, J. A. Design and Testing of Bicyclic Inhibitors of Grb7-Are Two Cycles Better Than One? Biopolymers 2013, 100, 543-549.

(36) Ovadia, O.; Greenberg, S.; Laufer, B.; Gilon, C.; Hoffman, A.; Kessler, H. Improvement of Drug-Like Properties of Peptides: The Somatostatin Paradigm. Expert Opin Drug Discov 2010, 5, 655-671.

(37) Yudin, A. K. Macrocycles: Lessons from the Distant Past, Recent Developments, and Future Directions. Chem Sci 2015, 6, 30-49.

(38) Kessler, H. Peptoids-a New Approach to the Development of Pharmaceuticals. Angew Chem Int Ed Engl 1993, 32, 543-544.

(39) Raynor, K.; Murphy, W. A.; Coy, D. H.; Taylor, J. E.; Moreau, J. P.; Yasuda, K.; Bell, G. I.; Reisine, T. Cloned Somatostatin Receptors: Identification of Subtype-Selective Peptides and Demonstration of High Affinity Binding of Linear Peptides. Mol Pharmacol 1993, 43, 838-844.

(40) Korsinczky, M. L.; Clark, R. J.; Craik, D. J. Disulfide Bond Mutagenesis and the Structure and Function of the Head-to-Tail Macrocyclic Trypsin Inhibitor Sfti-1. Biochemistry 2005, 44, 1145-1153.

(41) Lovelace, E. S.; Gunasekera, S.; Alvarmo, C.; Clark, R. J.; Nevin, S. T.; Grishin, A. A.; Adams, D. J.; Craik, D. J.; Daly, N. L. Stabilization of Alpha-Conotoxin Auib: Influences of Disulfide Connectivity and Backbone Cyclization. Antioxid Redox Signal 2011, 14, 87-95.

(42) Cheek, S.; Krishna, S. S.; Grishin, N. V. Structural Classification of Small, Disulfide-Rich Protein Domains. J Mol Biol 2006, 359, 215-237.

Page 31: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

31

(43) van Vlijmen, H. W.; Gupta, A.; Narasimhan, L. S.; Singh, J. A Novel Database of Disulfide Patterns and Its Application to the Discovery of Distantly Related Homologs. J Mol Biol 2004, 335, 1083-1092.

(44) Gongora-Benitez, M.; Tulla-Puche, J.; Albericio, F. Multifaceted Roles of Disulfide Bonds. Peptides as Therapeutics. Chem Rev 2014, 114, 901-926.

(45) Zhou, H. X. Loops, Linkages, Rings, Catenanes, Cages, and Crowders: Entropy-Based Strategies for Stabilizing Proteins. Acc Chem Res 2004, 37, 123-130.

(46) Fass, D. Disulfide Bonding in Protein Biophysics. Annu Rev Biophys 2012, 41, 63-79.

(47) Williams, N. K.; Liepinsh, E.; Watt, S. J.; Prosselkov, P.; Matthews, J. M.; Attard, P.; Beck, J. L.; Dixon, N. E.; Otting, G. Stabilization of Native Protein Fold by Intein-Mediated Covalent Cyclization. J Mol Biol 2005, 346, 1095-1108.

(48) Deechongkit, S.; Kelly, J. W. The Effect of Backbone Cyclization on the Thermodynamics of Beta-Sheet Unfolding: Stability Optimization of the Pin Ww Domain. J Am Chem Soc 2002, 124, 4980-4986.

(49) Dagan, S.; Hagai, T.; Gavrilov, Y.; Kapon, R.; Levy, Y.; Reich, Z. Stabilization of a Protein Conferred by an Increase in Folded State Entropy. Proc Natl Acad Sci U S A 2013, 110, 10628-10633.

Page 32: On the Effect of Cyclization on Peptide Backbone Dynamics377355/UQ... · backbone cyclization can improve rigidity, there is a complex interplay between dynamics and cyclization,

32