Molecular Nanosprings for Protein-Based Nanorobotics

Mustapha Hamdi 1, Antoine Ferreira 1 and Constantinos Mavroidis 2 Laboratoire Vision et Robotique, ENSI Bourges-Universit dOrlans, 18000, Bourges, France. 2 Computational BioNano Robotics Laboratory, Northeastern University, Boston, MA-02120, USA antoine.ferreira@ensi-bourges.fr , mavro@coe.neu.edu
1

This paper presents a molecular mechanics study using a molecular dynamics software (NAMD2) for characterization of molecular elastic joints for bio nanorobotic prototyping. Simple protein-like elastic joints elements for structural links have be carefully studied and simulated to understand the functional limits of each one of them: fibronectin, titin, deca-alanine, fibrillin, topocollagen, tropomyosin, Rop and double-helical DNA molecules. We simulated the restoring forces involved under various external mechanical stress (stretching, contraction, shearing, bending) to predict the type of force spectra, reversibility, degrees of freedom and irreversible work that may be expected from single-molecule protein manipulation experiments. Standard models for the elasticity of unfolded proteins are based on entropic arguments in combination with enthalpic contributions. I. INTRODUCTION Bio-nanorobotics is an emerging area of scientific and technological opportunity bridging the fields of biology and nanotechnology. Developing nanomachines out of proteins elements requires the merging of two fields of research approaches: the inspiration by nature and biology (biomimetics) and the inspiration by large scale machines and the traditional mechanisms and machine theory (machine nanomimetics). Both DNA and protein molecules possess a number of intrinsic characteristics that make them excellent candidates for the assembly of dynamic nanostructures and nanodevices. In this way proteins and DNA could act as motors, mechanical joints, transmission elements, or sensors. Most research efforts are concentrated on enzymes that function as nanoscale biological motors such as kinesin [1], RNA polymerase [2], myosin [3], dynein [4], adenosine triphosphate (ATP) synthase [5], viral protein linear (VPL) motor [6] and DNA. These motors, which are called biomolecular motors, have attracted a great deal of attention recently because they have high efficiency, they could be self-replicating, and hence cheaper in mass usage, and they are readily available in nature. Biological mechanisms and joints with multi-degree of freedom are also currently investigated in the scientific community. As potential transmission components in bionanorobotics, we can note 2-dof molecular hinges based calix[4]arenes [7], based DNA [8], linear ratchets mechanisms [9], molecular nanosprings using fibronectin [10] and titin [11] which can act as compliant joints in biomolecular robotic systems. Spring-like proteins are particularly interesting from their elastic, resilient and stability characteristics: stretch and relax without any net energy dissipation, can store and release

energy and rectify motion in physical nanomachines in a full reversible way [12]. Protein flexibility refers to the capacity of the protein to experience dynamic changes in conformation under biological conditions. In this project we are studying the development of protein based joint mechanisms. In order to draw analogies between mechanical protein methodologies and robotics methodologies, several protein-like elastic joints for structural links have be carefully studied and simulated to understand their functional and elastic limits. Contribution from molecular dynamic (MD) simulations is important in order to be able to understand the bio-nanomechanics of proteins and develop dynamic and kinematic models to study their performances. Computational studies of different classes of proteins: helix bundle proteins class (rop, DNA), -sheet class (fibronectin, titin,) and compound class (collagen, silk, actin) have been carried out. It allows to predict the type of force spectra, reversibility, degrees of freedom and irreversible work that may be expected from molecular springs. Finally, quantitative models of biological spring mechanisms are developed and classified following their nanomechanical performances. II. DESIGN OF BIONANOMECHANISMS MIMICKING BIOLOGICAL STRUCTURAL PROTEINS New nanostructures (with dimensions of nanometers to micrometers) using synthetic strategies have recently emerged. Inspired by biological systems, concepts of selforganization and self-assembly of building blocks allows the self-association (in series or in parallel) of particular patterns to form higher order organized biological nanomechanisms. Elementary biological structural components composed of protein-based nanosprings are of prime importance in the elasticity and stability of main biological functions at macroscale, i.e., tendons, muscles, lung and skin. As illustration, Fig.1 shows some of these structural protein applications. Marine mussel produce byssal threads to attach to solid surfaces that ressembles tiny tendons (Fig.1a). Each thread is stiff at one end and extensible at the other. The mussel byssus thread protein found in tendon exhibits powerful elasticity characteristics, i.e., tendon is five times tougher and more extensible than Achilles tendon. Mussel thread is a microcomposite fiber with a graded distribution of tensile molecular elements: flexible near the mussel, strong and stiff

Mussel
Mussel
Human muscle fiber

Titin strand formed by homologous domains

Proximal portion

Sarcomere 2m

SILK 5kD PreCol-D

Listal portion Attachment plaque

Substratum COLLAGEN 46kD

Titin Immunoglobulinlike (Ig) domain

40

(a)

(b)

Fig. 1: Biological elastic microfilaments and fibers : (a) Structure of a mussel byssus thread protein and (b) giant multidomain fibrillar protein titin found in striated muscle.

towards the substrate. Its elastic characteristics are composed of two types of macro proteins: proximal and distal proteins. Distal protein is composed of several domains such as collagen (47kD), silk fibroin (20kD) and crosslinking sites (HIS tag). Proximal protein is mainly composed of elastin domains. Giant multidomain fibrillar protein titin (connectin) composing the sarcomeres of human striated muscles (Fig.1b) behaves the strong passive elasticity of muscle. Titin, a 1-m-long protein found in striated muscle myofibrils, possesses unique elastic and extensibility properties due to the serial assembly of coiled titin molecules. Titin is composed of 300 repeats of two types of domains, fibronectin type III-like (Fn-III) domains and immunoglobulin-like (Ig) domains. Just as muscles magnify forces and movements by using geometrical hierarchy mechanisms (Fig.1b), bio-inspired nanomechanisms using similar principles are investigated: cumulative nanoscale changes in elementary protein subunits are amplified by their linear arrangement in filaments in order to perform micro-scale mechanical work. These nanoscale fibrillar structures have equivalencies with macro-scale robotic kinematic chains. The synthetic serial and/or parallel composition of protein-like springs allows to mimic the passive elasticity of biomechanisms found in nature for application in new design concepts of bionanorobotic links. Figure 2a shows a bio-nanorobotic Stewart platform composed of protein components. Four biological elastic filaments or fibers can be used as passive spring elements to join two platforms and form a single degree of freedom parallel platform that is actuated by three viral protein linear actuator (center) [13]. In Fig.2b, the VPL actuators has stretched out and this results in the upward linear motion of

the platform. The four protein fibers acting as mechanical springs are also stretched out. Their elastic behavior can be used as a passive control element or as the restitution force that will bring the platform back to its original position.
VPL actuators Molecular springs

(a)

(b)

Fig. 2: Bio-nanorobotic paralell platform using four biological elastic microfilaments and three viral protein linear actuators.

The main challenge in bio-nanorobotic design involves the synthetic formation of elastic fibrillar structures that are homogeneous, structurally and kinematically well defined, with unique mechanical properties combining a graded distribution of tensile strength with good elasticity. Towards this goal, different protein structures and assemblies have been simulated and modeled using steered molecular dynamics simulations (SMD). III. METHODS AND TECHNIQUES FOR STEERED MOLECULAR DYNAMICS SIMULATION We present in this section the simulated mechanical properties of different molecular nanosprings for passive

nanomechanisms when subjected to external forces. The mechanical characteristics (stretching, shearing or bending) are simulated through an Interactive Steered Molecular Dynamics (SMD) system. 3.1. Methodology Molecular dynamics (MD) was carried out using NAMD2 [14] with the CHARMM19 potential set. All simulations were carried out at 300K, with temperature rescaling performed every 10 timesteps. In more detail, the protocol to generate a single simulation comprises four steps: 1. The first step is the energy minimization of the biomolecular structure in order to remove any strong Van der Waals interactions that may exist which might otherwise lead to unstable simulations. At this point, all proteins were solvated by a periodic box of dimensions (measuring 2606565 3) covering the module by at least five layers of water molecules with the TIP3P model. The entire box of water is overlayed onto the protein and those water molecules that overlap the protein are removed. Energy re-minimization is realized in order to readjust the water molecules to the protein molecule. 2. Then, an heating phase is initiated. Initial velocities at a low temperature are assigned to each atom of the system and Newtons equations of motion are integrated to propagate the system in time. During the heating phase, initial velocities are assigned at a low temperature and the simulation is started periodically, new velocities are assigned at a slightly higher temperature and the simulation is allowed to continue. The entire system gradually was heated up to 300 K in increments of 30 K for time intervals of 1 ps, while leaving the box volume unchanged. 3. When the desired temperature is reached, the equilibration procedure consists to run the simulation until that the structure parameters, i.e., pressure, temperature and energy, become stable with respect to time. During equilibration at a temperature of 300K, the water molecules composing the box were harmonically restrained to their original positions to maintain the shape of the water bubble. All simulations were performed with a time step of 1 femtosecond and a uniform dielectric constant of 1. 4. The final simulation step is to run the simulation in a production phase for the desired time length. However, the biological processes that involve transitions from one equilibrium state (native state) to another (conformation state) in molecular springs are difficult to reproduce on MD time scales, which today are still confined to the order of tens of nanoseconds. To address this issue, the application of external forces in a variety of ways may be used to guide the system from one state to another, yielding new information about the mechanics of protein-based nanosprings. 3.2. Steered Molecular Dynamics

Steered molecular dynamics (SMD) simulations were carried out with either a constant force applied to an atom (or a set of atoms) or by attaching a harmonic (spring-like) restraint to one or more atoms in the system. SMD simulations were carried out by fixing one terminus of the domain, and applying external forces to the other terminus. The forces were applied by restraining the pulled end harmonically to a restraint point and moving the restraint point with a constant velocity in the desired direction. The procedure is equivalent to attaching one end of a harmonic spring to the end of the domain and pulling on the other end of the spring. Force-induced unfolding processes with several choices of pulling positions were simulated in order to characterize stretching, bending and shearing mechanical properties of the various -helix and -sheet biological springs.
VMD: Virtual Molecular Dynamics NAMD: Molecular Dynamics

AFM tip

++ ++ +++ +++ + ++ + +pH

Native
Coordinates (x,y,z)

Conformation
External constraints

Force Field

Communication Protocol

GHOST SDK
Force Position

Fig.3: Basic concept of virtual environment and haptics technology for bio-nanorobot simulation.

The developed simulation system presented in Fig.1 permits manipulation of bionanorobotic components in MD simulations with real-time force feedback and 3D graphical display [16]. It consists of three primary components: a haptic device controlled by a computer that generates the force environment, a MD simulation for determining the effects of force application, and a visualization program for display of the results. Communication is achieved through an efficient protocol between the visualization program Visual Molecular Dynamics (VMD) [15] and the molecular dynamics program (NAMD2) running on single machine (can be extended to multiple machines). A force-feedback PHANToM device measures a users hand position and exerts a precisely controlled force on the hand in order to apply different mechanical constraints, force and energy fields on the virtual model.

Coordinates of molecule

Four-bundle -helix ROP protein

Double stranded DNA protein

FJC Hookes law WLC

-sheet titin protein

F1 F2

WLC model
Force (pN)

kstret

I
Distance (nm)

II

III
Distance (nm)
35 30

Distance (nm)

Distance (nm)

Distance (nm)

relaxation stretching

Force (pN)

relaxation stretching
DNA
Time (pSec)

25 20 15 10 5 0 0 100

relaxation stretching
Titin I27
200 300 400 500

ROP
Time (pSec)

Time (pSec)

600

(a)

(b)

(c)

Fig. 4: Force curve and hysteresis curve done by forward stretching (with v=0.1 /sec) and relaxation of the structure when the stretching force is zero: (a) four -helix to coil stretching, (b) double-stranded DNA stretching and (c) titin Ig 27 stretching simulations.

IV. MECHANICAL SIMULATION OF BIOLOGICAL SPRINGS FOR FILAMENT BIOMECHANISMS We present in this section the simulated mechanical properties of different molecular nanosprings for filaments and fibrils nanomechanisms when subjected to external forces. A quantitative understanding of flexibility, extensibility, reversibility characteristics of passive biomechanisms are presented. 4.1. Biomechanisms using helical structures. 4.1.1. Four helix-bundle ROP protein: The repressor of primer protein (ROP) is a small, dimeric molecule consisting of two or four identical chains of 63 amino acids (Fig.4a). Each monomer consists of two -helices connected by a short turn and a seven-residue C-terminal tail. The two monomers pack together as a fully antiparallel four helixbundle. The bend region of Rop has attracted considerable interest as a parallel molecular spring due to its stability and elasticity properties [17]. In the simulation, we constrain both ends of two -helices of the molecule to move only along the z axis for stretching simulations and fix the short turn. Fig.4a shows two distinct conformational transitions

that provoke the conversion of double -helix to an extended form approaching the coil conformation. Both -helix extensions are very similar and can not be distinguished in Fig.4a. A useful approximation for spring constant kstret is given by the inextensible worm-like chain (WLC) model [18]. The WLC model of entropic elasticity predicts the relationship between the relative extension of a polymer (z/L) and the entropic restoring force (f) through
f = (k BT / A) z / L + (1 / 4(1 z / L)2 ) 1 / 4 ,

(1)

where kB is the Boltzmann constant, T is absolute temperature, A is the persistence length, z is the end-to-end length, and L is the length. The four -helix structure is naturally fully reversible in a wide domain of extension without the need of an external pulling force. Its behavior during extension might be modeled as four entropic parallel springs (see Fig.6a). 4.1.2. Double stranded DNA protein: Chemically, DNA is a long polymer made up of a linear series of subunits known as nucleotides. Structurally, DNA is usually found as a double helix, with two strands wrapped around one another (Fig.4b).

/ sheet fibrillin protein

Triple-stranded helix tropocollagen t i

Double-helix tropomyosin protein

3500 3000 2500

x 10

6000

k2stret
Force (pN)
force in pN

2.5

WLC
5000

Force (pN)

Force (pN)

FJC

Force (pN)
3 3.5 4 4.5 5 distance in nm 5.5 6 6.5 7

2000 1500 1000 500 0 -500 -1000 5

k1stret

Force (pN)

4000

1.5

3000

0.5

2000

lk
10 15 20 25 30

1000

-0.5 2.5

Distance (nm) Distance

Distance (nm)

0 15

20

25

30

35 40 Distance in nm

45

50

55

60

Distance (nm)
50 45

Distance (nm)

Distance (nm)

Distance (nm)

relaxation stretching

relaxation

Distance (nm)

40 35 30 25 20 15 10 5 -50

Relaxation

Stretching

Fibrillin Time (pSec) Time (pSec)

(a)

stretching g Tropocollagen
Time (pSec)

Tropomyosin
0 50 100 150 200 250 300 350 400

Time (pSec)

(b)

(c)

Fig. 5: Force curve and hysteresis curve done by forward stretching (with v=0.1 /sec) and relaxation of the structure when the stretching force is zero: (a) fibrillin protein, (b) three helical tropocollagen protein and (c) double helical tropomyosin protein.

Assemblies of double-stranded DNA (ds-DNA) proteins can provide strong elasticity and full reversibility. The waterDNA system was gradually heated over 7ps to 300K, and then equilibrated with a thermal bath at 300K for another 7ps. The dynamics of the DNA were performed with the double-stranded terminations stretched and the other end fixed. This elastic behavior is thus purely entropic [19]. For very low tension f 1 pN , the restoring force is provided by "entropic elasticity". In the absence of any force applied to its ends, the DNA's RMS end-to-end distance (chain length, L) is small compared to its contour length defined as the maximum end-to-end distance (maximum length, L0) and the chain enjoys a large degree of conformational disorder. Stretching DNA reduces its entropy and increases the free energy. The corresponding force f increases linearly as a Hookes law with the extension L:
f 3k B T L , ADNA L0 L << L0 (2)

forces ( f 10 pN ) , the end-to-end distance L is close to L0 and the elastic restoring force is due to distortion of the internal structure of DNA. In this regime, the force extension curve can be approximated by two models that are often used to describe the entropic elasticity of DNA. In the freely jointed chain (FJC) model, the molecule is made up of rigid, orientationally independent Kuhn segments whose length, b, is a measure of chain stiffness. The resulting entropic elastic behavior can be summarized in the force-extension relation:
z Ltot fb = coth k T B k BT f b

(3)

defining the well-known Langevin funtion. Expanding Eq.(3) gives the effective spring constant for the FJC as k FJC = 3k B T / b . The terms Ltot represent the total length of the protein and f the stretching force. The alignment of segments by tension is described by Boltzmann distribution.

The length ADNA is known as the "thermal persistence length" of DNA and is of the order 50nm. For higher


B1 B2

Stretching force
Stretching force k1
A2

k4 k2 k3

A1

k1

k2

k3

k4

(a)

Stretching force

(b)

k1

k2

k n-1

kn F

Ig segment PEVK domain

Sequential unfolding

l1

l2 L

l n -1

ln

(c)

(d) k3 k2

k1

k2

k3

k4

Hydrogen bonds

k1

k4

Stretching force (e)

k3 k2 k4 Stretching forces

(f)

Fig. 6: Protein based nanomechanisms generating restoring force based on the mechanism of entropic elasticity. In the absence of external force, their entropic springs take a coiled form to take a coiled conformation while extension or compression of the chain generates a restoring force due to reduction of entropy. Different reversible spring nanomechanisms have been studied and modeled: (a) four helix Rop protein as a parallel spring nanomechanism and (b) a tandem of four double-strand DNA proteins,. In (c) and (d), we present superelastic structures based on (c) three-titin -sheet proteins and (d) three-fibrillin -sheet proteins as serial link mechanisms. In (e) and (f), we present serial structures with high tensile strength: (e) a pair of double helix tropomyosin filaments and (f) assembly of sandwhich -sheet silk proteins.

In the inextensible worm-like chain model (WLC) model, the molecule is treated as a flexible rod of length L that curves smoothly as a result of thermal fluctuation (see Equ.1). Results, shown in Fig.4b, indicate that, even though the FJC model can describe the behavior of dsDNA in the limit of

low and intermediate forces, it fails at high forces. The WLC model, on the otherhand, provides an excellent description of molecular elasticity at intermediate and high forces. Both models behave as a Hookes law for low stretching forces. The dsDNA is fully reversible as shown in Fig.4b. As shown

in Fig.6b, whenconsidering small extentions, a tandem of dsDNA proteins can be modeled as a series of hookian springs.

4.2. Biomechanisms using bundle structures. 4.2.1.Immunoglobulin-like (Ig) domains. The behavior of titin as a serial link entropic spring depends on the reversible unfolding of individual Ig domains. Currently, I27 is the only I-band Ig with an experimentally solved structure [20] and hence has been selected for investigation. The Ig domain was placed in the center of the water box and equilibrated with a thermal bath of 300C. Stretching simulations were carried out by fixing one terminus of the domain and applying external forces to the other terminus. The simulation began with an equilibrated folded structure and was stopped when a fully extended polypeptide was obtained. The extension of I27 was performed with a pulling speed v=0.5/ns and was stopped when the extension reached 33nm.The extension domain (Fig.4c) is divided into four sections: I. preburst at extension of 4nm during which the protein maintains -sheet structure and the external force remains smaller than 1500 pN; II. major burst immediately after the preburst burst at extension of 8nm; III. post-burst at extension of 27 nm during which the protein unravels; IV. pulling of fully extended chain up-to an extension of 33nm. Other simulations showed similar features of the unfolding process and force profiles with only small variations in force peak value and degree of extension at the force peak. Results of stretching-relaxation curve (Fig.4c) show a good reversibility of the protein motion when completely relaxed. Fig.6c suggests a mechanical portrait model. Its behavior during extension might be modeled as series of elastic spring with a viscous element corresponding to the unfolding of the individual I27 domain. Stretch would result first in straightening of the Ig domain chain (corresponding to the preburst-part I) as an entropic spring. The tightly folded Ig domain might function as a shock absorber (parts II: major burst and III: post-burst) by reversible unfolding only in the case of extremely high stretching forces. This structure allows avoiding the complete rupture of the protein due to overstretching. 4.2.2. Fibrillin proteins. Fibrillins form the structural framework of an essential class of extracellular microfibrils that endow dynamic connective tissues with long-range elasticity (skin, lung, eye) [21]. Mechanical extension induces a conformational change with an apparent decrease in randomly coiled regions of the protein and a relative increase in -helical regions. As we can see in Fig.5a, a series of three unfolding events that results ultimately in the generation of a stabilized tensioned polymer. Upon relaxation it appears that the protein spontaneously refolds to the original state, indicated by a return to the native state. WLC model fails in this case for small amount of stretching forces. It can be represented by a spring-stiff pair. For lengths below the length (lk), spring constant of the protein is k1stret , whereas for lengths greater than (lk), spring constant is

k2stret. As the tension increases, the protein extends according to the unfolding events until to reach the length lk (as a spring component), and then the folded protein extends according to its stiffness (as a stiff component). When subjected to external mechanical bending force, the fibrillin protein behaves as a linearly elastic stiff spring. Such proteins are very interesting in compliant molecular links constituted by fibrillin-rich microfibrills. In this structure, the individual microfibrills acts as relatively stiff elastic polymers. Based on simulation studies (with v=1 /ns and k=10 pN/), it was concluded that individual isolated microfibrills were reversibly extensible (see Fig.5a) although the mechanism of this elasticity is unknown. The mechanical elastic model can be represented as a series model of spring-string components (Fig.6d). 4.3. Biomechanisms using compounds helix-bundle structures. 4.3.1. Triple-stranded helix tropocollagen protein. Collagens are a family of stuctural proteins of the extracellular matrix with one or more triple-helical domains. We studied here the triple-stranded helix tropocollagen (TP) macromolecule. The TC is a rigid rod of definite length composed of three helical peptide strands thatare equal in length. The collagen structural stability is provided by an orderly intramolecular hydrogen bond network [22]. To determine the elasticity of the tropocollagen molecule at the nanoscale, Fig.5b gives the force-extension characteristic. We verify the linear response, on average, when the molecule was extended 2.6-3.3 beyond its initial length, representing 2.0-2.5% of the molecular strain. For higher extensions (>3.3), high values of forces are reached (3104 pN) for small extension values 3.5-6.5 nm leading to high stiffness constants (>106 pN/nm). Figure 5 b shows clearly that the reversibility of the molecular spring is strongly perturbed by the destruction (during stretching) and reorganization (during relaxation) of the internal hydrogen bonds. The TP can be modeled as three strings in parallel (stiff spring) developing its stress when completely unfolded. When a stretching force F is applied, the system is considered homogeneous, each line bears a third of force and each spring in each line carries the same force. Then the strain is equally distributed within the lines and the springs. Since the TP protein is almost inelastic, it works as a stoplength fibre, preventing further strain when reaching its overstretched limits. This redistribution of the hydrogen bonds guarantees a constant mean of interactions needed to keep a reversible pathway when the stretching effect is interrupted. 4.3.2. Double-stranded helix myosin The nanomechanical properties of the coiled-coils of myosin are fundamentally important in understanding muscle assembly and contraction [23]. The molecular mechanics simulations of coiled-coil extension (shown in Fig.5c) demonstrates a soft hookian spring (10 pN/nm) for small

extensions (40-45 nm) preceding the final exponential phase leading to stiff springs modeled by WLC model of elasticity. The length Amyosin known as the "thermal persistence length" of myosin has been chosen on the order 40nm. In this case, the FJC model of elasticity fails. An important consideration in analyzing the force spectra of stretched tropomyosin is that serial assembly using carbon atom attachments are modeled by simple WLC springs connected in parallel. However, serial configuration using hydrogen bond attachments (see Fig.6e) behaves as a symmetric or asymmetric molecular network of interconnected soft and breakable springs (hydrogen bonds). For some compounds of proteins, the exponential fit of the data is necessary but not sufficient to claim non-entropic nature of the elastic response. Alternate models have been also developed composed of molecular network composed of interconnected springs.

4.3.2. -sheet silk fibroin protein. Silk is a protein produced by the silkworm Bombyx mori for the construction of the cocoon and also a large number of spider families for the manufacture of web. The silk fibres are made up of a protein called fibroin [24]. The protein is constructed from layers of antiparallel beta pleated sheets which run parallel to the silk fibre axis (see Fig.6f). Relaxed and extended stretching simulations show the large change in length that can occur with extension. Single molecules of fibroin silk are composed of tandem repeats of the entire ensemble, consisting of repetitive -sheet sequences plus non-repetitive spacer sequences. Even if the elasticity of fibroin silk is due to -strands acting as molecular springs, these springs are not necessary Hookian but function primarily as entropic springs. We are carrying out simulated pulling and stretching, using steered molecular dynamics, to investigate the molecular origins of the unique characteristics of the -sheet sequences in fibroin silk filalements. In such a case, filament can be modeled as a collection of chains joined by crosslinks (Fig.6f).
V. DISCUSSION AND CONCLUSION To summarize the performances of the different molecular springs, Table 1 makes a performance comparison and classification of protein-based elastic joints. These biosprings are classified following their mechanical characteristics: number of dof, elasticity, reversibility, displacement, linearity, relaxation time (see Table1). It is clear that we have treated only idealized cases in this modeling study and that we are using a very simple environmental representation. However, the present study shows clearly some important properties of molecular springs. All characterized molecular springs exhibit a conformational bistability of latched and unlatched states with different elastic features. The assembly of molecular springs features interesting mechanical properties. (i) Magnification of movements by a serial geometrical

hierarchy of proteins depends strongly of the nature of the spring. For structural smartness, assemblies built from helical units tend to be more readily reversible than structures dominated by -strand units. This difference arises principally from the relative importance of hydrogenbonding and hydrophobic contributions to structural stability. -Sheet structures have extensive interpeptide or interchain hydrogen bonds, which add great rigidity and irreversibility to such structures (i.e. the rigidity of silk fibroin is derived in large part from hydrogen bonding). The use of stretchy proteins in serial links, such as elastin or spiders silk, stretch and relax without any net energy dissipation. These proteins are highly resilient with smooth rebound when the elastic protein is relaxed. The fibtoin silk has tensile strength of ~1 Gpa, comparable in strength to Kevlar or steel. The other proteins presented in Table 1 dissipate energy as heat in the process of stretching and relaxing making them less resilient. (ii) Magnification of force by parallel structure are well adapted to large force-small displacement spring-like elastic joint, i.e., -helix bundles proteins. Through the database of Table 1, different kinematic and dynamic equivalences between macro and nanomechanisms have been drawn in order to allow prototyping of open-loop or closed-loop kinematic chains with adaptable stiffness and large displacements. I. REFERENCES
[1] S.M. Block, Kinesin what gives?, Cell, 93, pp.5-8 (1998). [2] M.D. Wang, M.J. Schnitzer, H.Yin, R. Landrick, J. Gelles, S.M. Block, Force and velocity measurement for single molecules of RNA polymerase, Science, 282, pp.902-907 (1998). [3] K. Kitamura, M. Tokunaga, A.H. Iwane, Y. Yanagida, A single myosin head moves along an actin filament with rectangular steps of 5.3nm, Nature, 397, PP.129 (1999). [4] C. Shingyoji, H. Higuchi, M. Yoshimura, E. Katayama, T. Yanagida, Dynein arms are oscillating force generators, Nature, 393, pp.711-714. [5] C.D. Montegano, G.D. Bachand, Constructing nanomechanical devices powered by biomolecular motors, Nanotechnology, 10, pp.225-331 (1999). [6] G. Sharma, M. Badescu, A. Dubey, C. Mavroidis, S.M. Tomassone., M.L. Yarmush, Kinematics and Workspace Analysis of Proetein Based Nano-Actuators, ASME Transactions on Mechanical Design, Vol.127, July 2005, pp.718-727. [7] H-H Yu, T.M. Swager, Molecular Actuators-Designing Actuating Materials at the Molecular Level, IEEE Journal of Oceanic Engineering, Vol.29, pp.692-695 (2004). [8] N.C. Sheman, From Genes to Machines: DNA Nanomechanical devices, Trends Biochemical Sciences, Vol.30, pp.119-125 (2005) [9] L. Mahadevan, P. Matsudaira, Mobility powered by supramolecular springs and ratchets, Science, 288, pp.95-99 (2000) [9] M. Knoblauch, G.A. Naull, T. Muller, D. Prufer, I. Schneider-Huther, et al. ATP-independent contractile proteins from plants, Nature Materials, 2, pp.600-603 (2003). [10] T.Yamamoto, O.Kurosawa, H.Kabata, N. Shimamoto, M. Washizu, Molecular surgery of DNA Based on Electrostatic micromanipulation, IEEE Trans. on Industry Applications, Vol.36, No.4, pp.1010-1017 (2000). [11] H.P. Erickson, Stretching single protein molecules: Titin of a wired spring, Science, 276, pp.1090-1092. [12] M. Hamdi, G. Sharma, A. Ferreira, C. Mavroidis, Characterization of Protein based Spring-like Elastic Joints for Biorobotic Applications , IEEE Int. Conference on Robotics and Automation, May, Florida, pp.1794-1799 (2006).

[12] Dubey A., Mavroidis C., Thornton A., Nikitczuk K.P., Yarmush M.L., Viral Protein Linear (VPL) Nano-Actuators, IEEE Int. Conference on Nanotechnology, San Francisco, CA, August 12-14, 2003. [14] M. Nelson, NAMD-A Parallel Object-Oriented Molecular Dynamics Program, Int. J. Supercomput. Appl. High. Perf. Comput. 10, p.251, 1996. [15] W.Humphrey, A.Dalke, K.Schulten, VMD:Visual Molecular Dynamics, Journal of Molecular Graphics, vol.14, pp.33-38, 1996. [16] M. Hamdi, G. Sharma, A. Ferreira, C. Mavroidis, Molecular Mechanics Simulation of Bio-Nanorobotic Components using Force Feedback, IEEE Int. Conference on Robotics and Biomimetics, June 29, Macau and Hong-Kong, pp.105-110 (2005). [17] H.P. Kresse et al., Four-helix bundle topology re-engineered: monomeric Rop protein variants with different loop arrangements, Protein Engineering, Vol.14, N11, pp.897-901, 2001. [18] S. Park, F.Khalili-Araghi, E.Tajkhorshid, K.Schulten, Free energy calculation from steered molecular dynamics simulations using Jarzynskis equality, Journal of Chemical Physics, vol.119, n6, pp.3559-3566, 2003.

[19] C. Bustamante, S.B. Smith, J.Liphardt, D.Smith, Single-molecule studies of DNA mechanics, Structural Biology, vol.10, pp.279-285, 2000. [20] K. Trombitas et al.,Titin extensibility in situ: Entropic elasticity of permanently folded and permanently unfolded molecular segments, Journal of Cell Biology, Vol.140, N.4, pp.853-859, 1998. [21] M.J. Sherratt et al., Fibrillin microfibrils are stiff reinforcing fibres in compliant tissues, Journal of Molecular Biology, 332, pp.183-193, 2003. [22] A. C. Lorenzo, E. R. Cafferana, Elastic properties, Youngs modulus determination and structural stability of the tropocollagen molecule: a computational study by steered molecular dynamics, Journal of Biomechanics, 38, pp.1527-1533, 2005. [23] D. D. Root, V. K. Yadavali, J. G. Forbes, K. Wang, Coiled-coil nanomechanics and uncoiling and unfolding of the superhelix and -helix of myosin, Journal of Biophysics, 2006. [24] N. Becker et al. Molecular nanosprings in spider capture-silks threads, Nature Materials, Vol.2, pp.278-283, 2003.

Table1: Comparison of elasticity characteristics of some proteinlike nanosprings. Proteins type Length Change Spring Stiffness Stretching kstret (pN/nm) 440 40 850 875 10000 40 200 1000 745 Bending kbend (pN/nm) 32 250 360 925 N.A. 190 450 N.A. N.A. Shearing kshear (pN/nm) 710 100 60 280 N.A. 240 220 N.A. N.A. Relaxation Time (sec) 9.04 0.163 0.1 0.1 0.365 0.48 0.192 0.1 0.1 Energy variation
(Kcal/mol)

Reversibility error (%) 30 10 5 5 24 5 <5 5 5

Elasticity model

l/l
Deca-alanine Fibrillin Fibronectin Fn-III10 Immunoglobulin I27 Topocollagen Rop DNA Silk Tropomyosin 1.305 4.17 8.483 5.769 1.707 1.2 1.476 N.A. N.A.

975 1835 6355 4778 1211 705 3018 1790 N.A.

Hookian Hookian WLC WLC WLC Hookian FJC WLC WLC and FJC

N.A: Not Available