Review Review The Effects of Shear Flow on Protein Structure and Function The Effects of Shear Flow on Protein

Structure and Function

Innocent B. Bekard,1 Peter Asimakis,1 Joseph Bertolini,2 Dave E. Dunstan1
1 2

Department of Chemical and Biomolecular Engineering, University of Melbourne, Parkville, VIC 3010, Australia Research and Development Department, CSL Biotherapies, 189-209 Camp Rd, Broadmeadows, VIC 3047, Australia

Received 14 February 2011; revised 20 April 2011; accepted 26 April 2011 Published online 4 May 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/bip.21646

ABSTRACT:
Protein molecules are subjected to potentially denaturing uid shear forces during processing and in circulation in the body. These complex molecules, involved in numerous biological functions and reactions, can be signicantly impaired by molecular damage. There have been many studies on the effects of hydrodynamic shear forces on protein structure and function. These studies are reviewed and the implications to bioprocessing and pathophysiology of certain diseases are discussed. # 2011 Wiley Periodicals, Inc. Biopolymers 95: 733745, 2011. Keywords: shear ow; protein; unfolding; aggregation; amyloid bril; physiology; bioprocessing

This article was originally published online as an accepted preprint. The Published Online date corresponds to the preprint version. You can request a copy of the preprint by emailing the Biopolymers editorial office at biopolymers@wiley. com

INTRODUCTION

roteins are linear polymer chains of combinations of the 20 amino acid residues. These complex macromolecules form an essential part of an array of biological processes in living systems, such as structural components of cells, biological catalysts (enzymes), and chemical messengers (hormones). The functional propCorrespondence to: Dave E. Dunstan; e-mail: davided@unimelb.edu.au Innocent B. Bekard and Peter Asimakis contributed equally to this work. C 2011 Wiley Periodicals, Inc. V

erties of proteins depend on their unique three-dimensional structure which is determined primarily by the amino acid sequence of individual chains. The energy threshold between the native and denatured state of a protein is low, thus the physicochemical environment is a critical determinant of protein structure and function.1 Perturbation of the native protein conformation results in loss of catalytic activity or biological function,2,3 and in many cases aggregation and amyloid bril formation.46 Structural perturbation of protein molecules occurs as a result of conditions of temperature, pH, and ionic strength as well as through the presence of denaturants (e.g., urea, guanidine), organic solvents, and surfactants.7 In addition, structural destabilization can result from exposure to hydrodynamic shear forces810 originating from shaking, sonication, mixing, vortexing, and ow through conduits.7,11,12 The effect of uid shear stress on protein structure is a subject of practical interest because it is a common phenomenon in bioprocessing.13 Proteins have been shown to exhibit conformational dynamics in shear ow.1417 Protein solutions are exposed to shear stresses during bioprocessing steps as a result of centrifugation, fractionation, pumping, and ultraltration.13,18,19 Similarly, the shipping and handling of biotherapeutics and protein based bioprocessing reagents such as monoclonal antibodies, hormones, cytokines, and enzymes can result in signicant agitation. For these reasons, an understanding of the effects of shear ow on protein stability, and a means of testing and measuring this effect, would allow the appropriate design and selection of manufacturing processes, conditions, and formulations which would ensure maximum yield and stability. It has been known for some time that considerable shear stress is generated by blood ow in the circulatory system.20 However the pathophysiological implications of this have not been explored. Thus, this review will also examine relevant papers in this area and highlight implications to the pathogenesis of some diseases.

Biopolymers Volume 95 / Number 11

733

734

Bekard et al.

FIGURE 1 Schematics of the common velocity proles in ow. (A) Homogeneous extensional ow typically associated with cross-slot geometries. The shear rate is constant. (B) Well developed ow in a conduit. The velocity prole is parabolic, typical of Poiseuille ow. In addition, the uid velocity is maximal at the center of the tube, approaching zero at the boundaries. The converse is true for the uid shear rate. (C) Simple shear ow, showing a velocity gradient perpendicular to the direction of ow. In all cases (AC), the arrows show both the direction of ow as well as the velocity prole.
vx 16 _@ strain rate c @ y . In the rotational component of the ow eld, protein molecules experience whole body rotation with no hydrodynamic shear strain, hence maintain their structural integrity. However, protein molecules are subjected to stretching events when oriented in the extensional ow eld. The hydrodynamic shear stress resulting from these events is thought to destabilize the native protein structure, leading to aggregate formation.17 For Newtonian uids, the mathematical relationship between shear stress _ ) in a given ow eld is expressed (s) and velocity gradient (c _ .23 A tangential stress of this type is called shearing, as s gc where innitesimally thin layers of uid slide over each other in laminar ow. Shear stress has units of force per unit area (N/m2), and is a measure of the actual hydrodynamic drag forces acting on the protein solution under ow. The velocity gradient, also referred to as the shear rate, has units of inverse time (s21). It is noteworthy that the magnitudes of rotational and elongational components of simple shear ow are equal (|x| _ |). This implies that the conformational dynamics of 5 |c protein molecules in the ow eld are dictated by the random occurrence of both events. In practice, the rotational and elongational components of ow differ in magnitude.

Studying the Effect of Shear Flow on Proteins

Extensional Flow vs. Simple Shear Flow. The common ow elds applied in shear studies are extensional ow and simple shear ow. A homogeneous extensional ow, also called elongational ow or stretching ow, is characterized by a lin_ y along the direction ear velocity gradient of the form vy c _ of ow (Figure 1).21 The strain rate, c
@ xy @y ,

is constant. An

example is the ow of a polymer solution through a converging channel with hyperbolically-shaped walls. This should not be confused with well developed ow of an incompressible uid through a circular pipe of innite length, which is characterized by inhomogeneous shear. Here, the ow pattern is of the Poiseulle type, showing a parabolic velocity prole where both the shear rate and uid velocity varies as a function of the distance from the vessel boundary (Figure 1). The strain rate is given as _ 2y0 a where y0 is the distance between the center of the c polymer and the ow axis and a is a measure of the curvature of the ow prole.22 Simple shear ow on the other hand describes uid ow characterized by a velocity gradient perpendicular to the ow eld (Figure 1). It is modeled as a linear superposition of rotational ow with vorticity x, and elongational ow with

Biopolymers

Effects of Shear Flow on Protein Structure and Function Table I Studies of the Shear-Stability of Protein Systems Using a Variety of Shear Devices Protein Insulin Poly-L-lysine b-lactoglobulin von Willebrand factor Amyloid-b Lysozyme Cytochrome-c a-amylase Fibrinogen Catalase Rennet Glycoprotein Ib and IIb-IIIa von Willebrand factor Urease Catalase Alcohol dehydrogenase Deoxyribonuclease Growth hormone Urease Catalase Catalase Cytochrome-c von Willebrand factor Shear Rate/Rotational Speed 200 s21 [333 s21 25 s21, 150 s21 \3.0 3 103 s21 50 s21 750 rpm, [14 s21 750 rpm 120 s21 290 s21 91.5 s21 9.15 s21 8.23 103 s21 6.7 3 103 s21 683 s21 683 s21 683 s21 1.5 3 103 rpm 1.5 3 103 rpm 48 s21 67 s21 4.6 3 104 s21 2.0 3 105 s21 [13 103 s21 Result Unfolding and aggregation Unfolding Fibril formation Unfolding Fibril formation Unfolding No change Loss of activity Degradation Loss of activity Loss of activity Unfolding Aggregation Particle formation Particle formation Particle formation No change Unfolding and fragmentation Loss of activity Loss of activity Minor loss of activity No change Unfolding

735

Device Geometry Couette

References 14 24 25, 26 27 28 29, 30 29 31 32 33 33 34 35 36 36 37 38 38 39 40 36 41 42

Taylor-Couette Cone and plate

Concentric cylinder

Capillary

Microuidic cell

Devices for Generating Shear Flow

In many protein studies, shear ow is generated by subjecting protein solutions to ow patterns characterized by uniform or heterogeneous velocity gradients.18 Shear devices that create a near uniform velocity gradient through protein solutions provide a well controlled and quantiable shear stress. Hence, such devices are ideal for studying the effects of shear forces on protein molecules. Conversely, heterogeneous velocity gradients, for example, resulting from stirring or shaking, provide poorly controlled shear conditions that are difcult to quantify. It is worth noting that the differences in experimental devices applied in shear studies, and the associated data interpretation, could be a major contributory factor to the inconsistent experimental results reported throughout the literature. Several studies that show the effect of shear ow on a number of proteins, along with the shear devices employed, are summarized in Table I. By virtue of the differences in experimental design, the shear (strain history) as well as the shear stress values, which are not given here, varies signicantly among the different experiments. The experimental ow devices used for protein studies fall into two main categories: capillary and rotational devices.

between the inlet and outlet of the conduit.43 Here, the uid velocity prole is of the Poiseuille type (non-homogeneous shear ow) with the maximum shear rate at the uid-vessel boundary, reducing gradually towards the vessel center (Figure 1B). Very high shear rates, up to 105 s21, can be achieved but over short residence times on the order of milliseconds.41,44 A schematic of a typical capillary ow tube is shown in Figure 2. It has been reported that the short residence time spent in the ow eld could contribute to incomplete protein stretching (unfolding).45 However, recirculation of protein solutions through conduits have been successfully applied to increase residence times,32 offering a good model system for uid shear stress in ultraltration units and membranes.

Rotational Flow Devices

Cone and plate, parallel plate and concentric cylinder viscometers are three types of rotational ow devices commonly

Capillary/Microuidic Flow Devices

Capillary ow is where a uid is forced through a conduit of known dimensions, by applying a given pressure difference
Biopolymers FIGURE 2 Schematic of a simple extrusion glass capillary viscometer. The arrows show the direction of ow.

736

Bekard et al.

FIGURE 3 Schematics of the common rotational ow devices: (A) concentric cylinder, (B) cone and plate, and (C) parallel plate rheometers.

used to impart shear stress on protein systems across a narrow gap (Figure 3). All models consist of two parts where one rotates with the other being stationary. In the case of the concentric cylinder, also known as Couette or coaxial cylinder ow-cell, there are several sorts that differ in basic design but operate on the same principles.46 However, a throughcell Couette arrangement is ideal for generating a simple shear ow, as the design minimizes any end effects.47 This type of device also curtails particle settling or migration away from the solid boundaries, thereby limiting variations of uid properties across the gap.48 In addition, a nominal shear rate can be calculated for each rotational speed. At high rotational speeds, large deviations from the ideal ow may occur due to the formation of vortices. It is recognized that vortices occur in extensional ow regions in uids, and should therefore be avoided if a simple shear ow is desired.49 The main advantage of rotational shear devices over that of the capillary design is that the entire test solution is sheared at a near constant rate. This permits the study of any time-dependent behavior of protein samples over an extended period of time. In the through-cell Couette arrangement, the shear rate can be regarded as constant only if the gap width is much less than the radius of the inner cylinder.50 In the cone-and-plate design, the solution is contained in the space between a cone of large apex angle and a plate with a at surface normal to the axis. For small angles between the cone and plate, the shear rate is considered constant for all radial positions.

The Effects of Shear Flow on Protein Solutions

Protein Function. Preliminary studies on the relationship between shear ow and protein structure exploited the catalytic activity of enzyme solutions as a means of measuring protein integrity. Thus, a decrease in enzyme activity implied a structural deformation of the native enzyme molecules.

These early studies were critical in dening the appropriate, and operation of, devices to allow the study of the behavior of proteins in shear ow, allowing the estimation of shear forces required to perturb protein function with insight into possible resulting structural changes. Studies on catalase, rennet, and carboxypeptidase solutions sheared in a narrow gap coaxial viscometer40 at shear rates between 9.15 and 1155 s21 for 90 min, showed a strong correlation between decrease in enzyme activity and shear (shear rate 3 time). A loss of enzyme activity occurred for shear greater than 104 although the rennet solutions partially regained activity upon the cessation of shearing. The observed enzyme inactivation was attributed to the breaking of the molecules tertiary structure when oriented appropriately in the shear eld. The authors discounted any signicant contribution from the airliquid interface in the viscometer, where protein molecules, by virtue of their amphipathic nature, may aggregate (causing inactivity). It is also possible that the loss in enzyme activity was due to a ow-induced limited residence time for enzyme-substrate interaction. A similar shear-effect was observed upon exposure of catalase and rennet solutions to shear rates in the region of 104 s21 in model ultraltration systems.33 Whereas the catalase solutions lost some activity during ultraltration, the rennet solutions did not suffer inactivation in the circulation lter systems. Further studies on plasma brinogen, sheared in a Weissenberg rheogoniometer (shear rates 290 1155 s21) with a Couette attachment, revealed a temperature-independent loss in its clottability.32 Studies on the shear-stability of urease, using a coaxial cylinder viscometer, which allowed direct monitoring of urease catalyzed hydrolysis of urea during shear exposure, showed continuous decrease in the rate of urease activity with increasing shear rate (48, 288, 741, and 1717 s21).39 The partial deactivation of urease revealed both reversible and irreversible elements. The reversible inactivation was ascribed to hydrodynamic distortion of the enzymes structure in ow, whereas permanent inactivation was attributed to the breaking of the enzymes tertiary structure. The permanent loss in urease activity correlated with shear strain, with a critical shear value of 105, similar to that previously reported.40 A comparison of catalase degradation in streamline and turbulent ow regimes, upon mixing in a cylindrical container, showed a signicant inactivation in the latter where a vortex and air bubbles had formed in the solution.51 This suggested that the airliquid interface may have contributed signicantly to catalase degradation. It was also shown that shear stress in selected microltration membranes caused a signicant decrease in the catalytic activity of yeast alcohol dehydrogenase (ADH), under solution conditions where the
Biopolymers

Effects of Shear Flow on Protein Structure and Function

737

enzyme was slightly unstable.52 This observation was attributed to membrane-enzyme interactions resulting from a shear-induced deformation of the enzyme structure. In studies of the shear-stability of ADH solutions using a high shear concentric viscometer, a rotating disk reactor and several pumps, it was found that fully active tetrameric molecules of ADH showed negligible change in activity even for strain rates as high as 26, 000 s21 over 1 h in the high shear device.53 Turbulent regimes in the rotating-disk reactor (3600 rpm for 5 h) and up to about 1500 passes in the various pumps likewise had little effect on ADH activity. However, the introduction of a gas-liquid interface in the rotating-disk reactor resulted in a 40% decrease in the initial ADH activity after 5 h. Similarly, a signicant decrease in ADH activity was observed in both a gear and a Jabsco pump after 2000 passes.53 This report suggested that interfacial denaturation, and not shear per se, accounts for the deformation of proteins in solutions exposed to shear ow, and that the native conformation of globular proteins, when in a chemically stable environment, are stable in shear ows usually encountered in practice. The importance of an airliquid interface in promoting shear-induced protein denaturation was shown in a study with progesterone 11a-hydroxylase complex, using an open concentric cylinder viscometer and a horizontal viscometer closed at both ends.54 It was observed that the gas-liquid interface, in concert with high shear regimes, which replenished the protein at the interface, promoted rapid protein denaturation as previously reported.53 In addition, a wall surface effect on the enzyme complex was observed as the gap width of the viscometer was reduced from 0.475 to 0.15 mm. Furthermore, it was observed that viscous enzyme solutions showed a marked decrease in activity as a function of shear stress.54 These observations lead the authors to propose that the applied shear stress resulted in enzyme denaturation via either the rupture of the enzyme membrane or the enhanced solubilization of component(s) of the enzyme complex by the velocity gradient in the agitated solution. The above studies clearly showed that shear could have a profound effect on the activity of enzymes, and by inference on protein structure. The shear-effect was more pronounced in the presence of an airliquid interface. A key limitation of the above studies was that any deduction of structural change was based on indirect evidence derived from enzyme activity. Indeed, whereas shear may not have caused signicant structural damage, a decrease in enzyme activity could result from minor modications to an enzymes active site. In many of the experiments, the enzymes were subjected to intermittent shearing force to allow sampling. Therefore, signicant time
Biopolymers

would have elapsed between sampling of solutions and testing for enzyme activity. This would have allowed reversible structural changes to revert and thus go undetected. The advent of new experimental techniques, allowing real-time measurements of the solution conformation of proteins, is providing new insight into the shear-stability of proteins with respect to variations in the sensitivity of different proteins to shear, and the specic molecular changes occurring in response to shear exposure.14,55

Protein Structure
Human von Willebrand Factor (vWF), a blood plasma protein with important functions in coagulation, was amongst the rst of protein systems, besides enzymes, to be studied for shear induced structural changes. vWF molecules, which are initially secreted by the endothelial cells as large, hyperactive multimers, with relative molecular mass up to 50 3 106, circulate as a series of smaller, less active proteolytic fragments in the plasma as a result of the action of the metalloprotease ADAMTS-13.56,57 Interestingly, the proteolysis of large vWF multimers is not observed when normal plasma is incubated in vitro.58 This lead to the conclusion that the generation of proteolytic fragments of vWF in circulation depends on haemodynamic shear stress.57 The size distribution of vWF multimers in microcirculation is critical to its haemostatic potential. Tsai et al. investigated this theory by perfusing normal plasma through long capillary tubes (190 or 354 cm in length), at shear rates between 1508 and 4761 s21, followed by an examination of the size distribution of vWF using a combination of SDS agarose gel electrophoresis and densitometric scanning analysis of autoradiographs.59 Relative to unsheared plasma, an increase in smaller vWF multimers was observed in the sheared samples at the expense of the larger multimers, especially in high shear regimes. These observations lead to the conclusion that shear stress might be involved in the modulation of the size distribution of vWF in circulation by enhancing the proteolytic cleavage of large vWF multimers.59 In a subsequent study by Siedlecki et al., using a combination of atomic force microscopy and a rotating disk system, vWF molecules were observed (visually) to undergo a shearinduced conformational transition from a globular state to an extended chain conformation at a critical shear stress _  4618.8 s21).60 The result was the value [31.5 dyn/cm2 (c exposure of intra-molecular domains of vWF. It was suggested that shear-induced structural changes in vWF may have an important role in platelet adhesion and thrombus formation in regions of high shear stress, as would develop at the site of a bleeding injury.60

738

Bekard et al.

To this point, it had been considered that clot formation was initiated through direct shear activation of platelets.58 This was substantiated in a later study,61 which investigated the mechanics involved in the formation of reversible tether bonds between the platelet glycoprotein receptor Iba and the A1 domain of surface-immobilized vWF under hydrodynamic shear ow. Puried platelets were perfused over plasma vWF-A1 coated surfaces in a parallel-plate ow chamber at shear stresses between 0.2 and 8 dyn/cm2. Platelet tethering was visualized using a Nikon microscope equipped with a high speed video camera, and the motion of the tethered platelets calculated using an analytical two-dimensional model. The results showed that hydrodynamic compressive forces inuence tether bond formation between platelets and vWF.61 Recently, uorescence microscopy has been used to directly visualize individual uorescently labeled vWF multimers under hydrodynamic stress in a microuidic device.42 vWF was uorescently labeled with Alexa uor 488, which was attached to the primary amines of the glycoprotein using tetrauorophenyl ester. It was found that shear rates greater than 1000 s21 triggered a reversible stretching of vWF in solution. Furthermore, real-time measurements of the conformational dynamics of vWF in response to laminar shear ow in a Couette ow-cell, using a combination of small angle neutron scattering (SANS) spectroscopy and quantitative modeling, showed an irreversible conformational change in vWF, under hydrodynamic shear stress, at shear rates 3000 s21.27 The observed conformational change was found to involve the exposure of hydrophobic domains at shear rates [2300 s21, as indicated by an increase in the binding afnity of the hydrophobic dye 1,10-bis(anilino)-4-,40-bis(naphtalene)-8,80-disulfonate (bis-ANS).27 The shear-effect was more pronounced with increasing shear rate. A more recent single-molecule investigation of the shearstability of vWF has provided more insight into the relevance of shear stress in the homeostatic regulation of the size of circulating vWF multimers as well as the potentiation of vWFmediated platelet aggregation.62 Specically, the study examined the folded and unfolded states of a single A2 domain of vWF, which contains the site of cleavage for ADAMST-13, in the presence and absence of the metalloprotease. The A2 domain of vWF was unfolded by directly applying force with laser tweezers. The results showed that only the unfolded A2 domain is cleaved by ADAMST-13.62 These observations corroborated previous studies which found that shear-induced stretching of large vWF multimers results in the exposure of the scissile Tyr1605-Met1606 bond within the A2 domain, facilitating cleavage by ADAMST-13 in normal plasma.56,57 Smaller vWF multimers in microcirculation equally undergo

stretching in regions of high shear, such as a bleeding arteriole, which expose bonding sites in the A1 domain of vWF for complexation with platelets during haemostasis.57,60,63 While studies of vWF provided clear evidence that shear ow can alter protein structure, it may be argued that different proteins experience different shear since the strain rates required to trigger structural alterations decreases with increasing molecular weight and solvent viscosity. For example, studies on the effect of high shear rates on ferric equine cytochrome c, a small globular protein of molecular weight 12,384 Da, showed no signicant structural alterations when exposed to shear rates as high as 105 s21 in a capillary ow device.41 By applying a simple model, the authors suggested that a shear rate of 107 s21 would be required to denature a small globular protein in water.41 However, the transient residence time associated with capillary ow devices may explain the hysteresis observed in such systems.45 More importantly, the results41 contradict the observed abrupt stretching of vWF ([20,000 kDa) in a microuidic device at a threshold shear rate of 103 s21.42 Clearly, the larger, more complex vWF would be expected to be more susceptible to shear stress relative to cytochrome c. However, it is noteworthy that specic conformational characteristics such as a-helix and b-sheet composition, and intra-molecular interactions (e.g., hydrophobic, electrostatic) and covalent bonds (e.g., disulde), would be an important determinant of protein stability. Thus, protein systems with comparable molecular weights may differ in their shear-stability due to differences in their solution properties. For example, previous studies on the effects of both high shear ([107) and high shear rate ([105 s21), using concentric cylinder-based systems, on recombinant human growth hormone (rhGH) (22 kDa) and recombinant human deoxyribonuclease (rhDNase) (31 kDa) showed that rhDNase was more resistant to shear-effects.38 For instance, whereas scanning microcalorimetry and SDS-PAGE analysis of extensively sheared solutions showed changes in the melting temperature of rhGH and the presence of low molecular weight fragments, respectively, no such changes were observed for rhDNase. Interestingly, the observed structural changes in rhGH were not detected in both the near and the far-UV circular dichroism spectral prole of the sample.38 Further studies on both samples showed that rhGH denatured upon exposure to high shear in the presence of an airliquid interface whereas rhDNase remained relatively stable.64 The stability of rhDNase was attributed to its comparatively high surface tension and low foaming tendency in solution. Intriguingly, it has been shown that human and bovine albumin samples, which differ in only two amino acids,
Biopolymers

Effects of Shear Flow on Protein Structure and Function

739

show different aggregation kinetics upon shaking at high shear rates ([8000 rpm).65 Multi-angle laser light-scattering measurements showed that in addition to the monomer-dimer transition observed in both albumin samples upon shearing, human serum albumin also formed trimers especially at high shear rates. Since the protein concentration and solventair interfaces were considered as xed factors, the observed differences in the degree of albumin aggregation was attributed to the minor variation in their primary structure.

Protein Aggregation
As has been discussed above, shear can result in protein denaturation and compromised function. It is also well established that denaturation of proteins can lead to aggregation.66 Hence, the use of agitation (uncontrolled shear stress) to both induce and accelerate protein aggregation and amyloid bril formation is common practice throughout the literature.66 For example, under favorable solution conditions, various forms of agitation including shaking, sonication and stirring have been successfully applied to accelerate the aggregation of insulin,6771 amyloid-b,28,72 b2-microglobulin,73,74 b-lactoglobulin,25 albumin,65 and the PDZ domain from murine Protein Tyrosine Phosphatase Bas-like (PDZ2)75 among others. In one particular study, a range of structurally diverse protein systems including BSA, myoglobin, lysozyme, Tm0979, SOD and hisactophin were shown to generate amyloid like structures following sonication.76 Shear effects on protein aggregation have also been studied in uniform ow elds. For example, a shear-induced nucleation, without ow enhanced aggregation, of preheated b-lactoglobulin solutions (0.5 wt%) at low pH has been reported.77 b-lactoglobulin solutions were exposed to Couette ow in a rheometer, at a moderate shear rate of 200 s21 over 5 h, while measuring the ow-induced birefringence. Thermal treatment of the b-lactoglobulin samples triggered the formation of metastable pre-brillar aggregates, without which bril formation was absent in the sheared samples. Shearing of b-lactoglobulin samples was done either continuously or in short pulses, both producing a similar degree of bril formation. These observations lead to the conclusion that brief mechanical perturbation of an aggregation prone protein solution was enough to enhance bril formation.77 That is, shear ow only triggered the nucleation of bril formation but did not inuence the polymerization of preaggregates into mature brils, discounting orthokinetic coagulation under the experimental conditions. However, the brils formed from the continuous shear treatment showed a smaller variance in Gaussian length distribution relative to those from the pulse shear treatments, indicating a homogeBiopolymers

nizing effect of the continuous shear treatment. Evidently, a synergy of heat and agitation triggered rapid bril formation. Another study showed that the aggregation of preheated b-lactoglobulin solutions in Couette ow was shear rate dependent.25 In addition, if shearing was performed prior to thermal treatment, a rapid enhancement of brillar species still occurred, suggesting a shear-induced formation of the precursor nuclei required for bril development as previously reported.77 Interestingly, at shear rates [500 s21, bril degradation, possibly arising from the persistent stretching (extensional ow) events in the ow eld, was observed. A further study78 revealed that a sufciently high b-lactoglobulin concentration ([ 3 wt%) was required to initiate bril formation when samples were heated and sheared simultaneously. In addition, the amount of brils formed increased as a function of shear rate up to 337 s21 beyond which an opposing shear effect, similar to that reported earlier,25 was observed. The paradoxical shear-effect was explained by a ow enhanced polymerization of the bril nucleus in low shear regimes, whereas in high shear regimes, the extensional component of the ow eld overwhelmed the nascent inter-b-strand hydrogen bonds stabilizing the aggregating brils. Furthermore, the authors observed that the shear applied produced shorter brils and also enhanced, below a critical bril concentration, the viscosity of the resulting bril solutions.78 More recently, a similar ow enhanced bril formation has been observed in aqueous solutions of bovine insulin,14 thermally treated amyloid-b28 and b-lactoglobulin.26 The shear-stability of bovine insulin in Couette ow was probed directly via novel circular dichroism and tyrosine autouorescence measurements, in situ and in real time. For a given shear rate, helical segments of native insulin were observed to unfold as a function of time. Atomic force microscopy images of the sheared samples revealed the presence of brillar forms in relatively strong shear regimes (600 s21).14 The shear-effect on preheated amyloid-b samples28 exposed to Couette ow showed similar trends to that reported earlier.25 For the b-lactoglobulin studies,26 the authors compared the morphology and mechanical properties of brils formed under heterogeneous (stirred) and uniform (Couette) ow conditions using atomic force microscopy. They found that b-lactoglobulin brils resulting from heterogeneous ow had twisted ribbon-like morphologies and higher mechanical strength relative to the beaded brils formed in Couette ow. Evidently, the intertwining of brils produced in the variable shear regime imparted additional mechanical stability to the overall bril structure. It was hypothesized that polymerization of bril seeds occurs in the ow eld, and in the case of stirred samples, mature brils anneal into twisted ribbons possibly via orthokinetic coagulation. Taking together, even

740

Bekard et al.

if in some cases the initial structural destabilization required for protein aggregation is not triggered by shear, its ability to accelerate the polymerization of metastable prebrillar aggregates into mature brils still has grave implications in bioprocessing and physiology. Regarding the mechanism of shear-induced protein aggregation, there is a general consensus that mechanical perturbation of a protein molecule often results in a structural destabilization of the native conformation, leading to the exposure of sequestered hydrophobic residues to the surrounding medium. Solvent-exposed hydrophobic groups nucleate via hydrophobic interactions and subsequently aggregate. Several studies show that the initial destabilization of the protein structure occurs at the solventair or solventsolid interface, where surface tension forces unfold the native protein.36,37,64,79 The airwater interfacial force is estimated to be 140 pN,80 taking the airwater interface to be 2 nm in depth with a 0.07 N/m surface tension.81 This value is comparable to the 150 pN observed to unfold globular proteins in atomic force microscopy studies.82 In addition, as agitation increases the turnover of the airliquid interface, and thus the number of protein molecules interacting with this interface, the nucleation of solvent-exposed hydrophobic species via hydrophobic interactions, and hence the aggregation of the bulk protein solution, increases dramatically.72 In fact, the inuence of agitation can be so profound that, in the case of insulin, it was found to attenuate the effects of other parameters, such as protein concentration and ionic strength, that inuence the kinetics of amyloid bril assembly.68

Molecular Models and Theoretical Aspects

The complexity of protein molecules with respect to conformational heterogeneity as well as intra- and intermolecular interactions between amino acid residues, with the consequent differences in the exibility of secondary conformations and tertiary congurations between proteins, results in variations in their solution properties in shear ow. Indeed, differences in the levels of association in diverse protein systems dictate the different degrees of shielding, especially of hydrophobic residues concealed in the protein matrix, from hydrodynamic drag during shear ow.83,84 Hence, it is difcult to obtain uniform data to allow the formulation of a theoretical model relating protein conformation to shear. For these reasons, dilute solutions of homopolymers, especially unbranched polymer chains, have been used as model systems for shear studies because of their inherent structural uniformity. By virtue of its conformational plasticity in response to temperature, pH, salt concentration and alcohol content in solution,24 poly-L-lysine is commonly used as a model system

in protein studies.85 The homopolypeptide exists as a random coil at neutral pH, a-helix at high alkaline pH and b-sheet at temperatures [308C.86 For these reasons, poly-Llysine is the ideal choice in probing the initial conformational transitions, under denaturing conditions, between the three basic protein secondary conformations namely: random coil, a-helix, and b-sheet. Experimental studies on the shear-stability of poly-L-lysine solutions (437 kDa), in an initial a-helical conformation, showed an increase in solution viscosity and turbidity with shearing time.87 A combination of Raman spectroscopy (collected in situ) and circular dichroism were employed for data acquisition in a Couette ow cell at 150 s21. In addition, a ow induced-gelation and an increase in the b-sheet content of the samples was observed in solution concentrations [0.3 g/dL. However, the circular dichroism data further revealed a reversal of the a-helix to b-sheet conformational transition after prolonged shearing ([50 min). The observed conformational change and subsequent aggregation of the sheared samples was thought to involve ow-enhanced hydrophobic interactions between individual poly-L-lysine molecules in solution.87 Further studies, via real-time circular birefringence measurements, showed a reversible, shear-induced, helix-to-coil transition of poly-L-lysine in simple shear ow.24 The conformational transition was observed at a critical strain rate of 300 s21, and the change was attributed to a shear-induced breakage of intramolecular hydrogen bonds in a-helical poly-L-lysine. More recently, the molecular-weight-dependence of the shear-induced unfolding of a-helical poly-L-lysine in Couette ow has been measured in real-time using circular dichroism spectroscopy.55 The authors observed that the shear-stability of the helices increased as a function of molecular weight. The hysteresis observed in the heavy chains was attributed to the large network of hydrogen bonds (cohesive force) stabilizing the helix structure, and the associated hydrodynamic screening of helical segments from the drag in the ow eld. Interestingly, irrespective of the molecular weight of poly-L-lysine, unfolding of the _ tc ) helical segments was found to occur at a critical strain (c 5 value of 10 (Figure 4), similar to that observed in globular proteins.39,40 For this reason, the authors proposed that the shear rate is not as critical as the duration of its application, which makes the idea of a critical shear rate arbitrary. In addition, the helix content, a, was found to show a power _ t 1= 2 . law dependence with strain: ac The inherent monodispersity associated with lamda bacteriophage DNA made it another ideal model for studies of polymer dynamics in hydrodynamic ows. For example, a novel real-time uorescence videomicrocopy of uorescently labeled single DNA molecules was developed to visualize the
Biopolymers

Effects of Shear Flow on Protein Structure and Function

741

FIGURE 4 Change in the helix content of PLL as a function of shear strain. The gure shows that the a-helix-PLL structure (68.3 kDa) is stable below a strain value of 105. For clarity, the shear rates plotted are: (&) 74 s21, (^) 302 s21, (*) 518 s21, and (l) 715 s21. For clarity, error bars representing 6 standard deviation are shown only for the sample sheared at 518 s21.55

conformational dynamics of DNA chains in both elongation21 and shear16 ows. The elongation and shear ow experiments were performed in a microfabricated ow cell and between two parallel glass plates (50 lm gap), respectively, at very low shear rates ranging from 0.05 to 4.0 s21. It was found that, at relatively high shear rates, the sharp coil-stretch transitions of DNA chains observed in pure elongation ow was absent in shear ow.16,21 Individual chains in steady shear ow rather showed large, aperiodic temporal uctuations in conformation, consistent with what had been predicted in theory.84 In addition, stretching of DNA chains in shear ow occurred at a strain value \50 (101),16 which is several orders of magnitude lower than the 105 critical strain value observed for poly-L-lysine55 as well as folded proteins.39,40 Jaspe and Hagen41 suggests that coil-stretch transitions readily occur in homopolymer systems because the free energy cost of unfolding is negligible compared to that of a native protein, and this was purported to explain why an extraordinarily high shear rate (107 s21) is required to destabilize the latter. Another study88 estimated a destabilizing strain rate in the region of 108 s21 for collapsed polymers in planar shear ow. Classical theories on the shear-stability of polymer chains predict that a polymer chain would remain in a coil state in shear ow until a critical velocity gradient, which triggers an abrupt unwinding of the coil, is reached.86 In addition, abrupt conformational transitions are not expected in simple shear ow since the occurrence of extensional strain, hence
Biopolymers

polymer stretching, is random. It is hypothesized that at very high strain rates, where polymer chains spend a relative short time (residence time) in an extensional ow eld, no unfolding or incomplete chain extensions may occur.45 These predictions were substantiated in later studies that simulated polymer dynamics in shear ow. For example, in molecular dynamics (MD) simulations of a grafted collapsed macromolecule in strong uniform ow, it was observed that a exible macromolecule remains compact below a critical shear value but shows a globule-stretch transition above this critical shear value.89 The transition was reported to be more pronounced in strong extensional ow. Lemak et al.90 later compared the conformational instabilities of a minimalist b-barrel protein in both uniform and elongational ow elds using MD simulations. It was reported that whereas unfolding was abrupt in elongational ow, the uniform ow elds presented a multi-step unfolding process, with several intermediates, depending on the tethered terminus of the protein. In explaining this observation, it was proposed that whereas every monomer (i.e., amino acid residues) experienced a destabilizing hydrodynamic drag in elongational ow, only the terminal monomer experiences this force in uniform ow. Hence, in uniform ow, deformation of the protein follows an unzipping mechanism where the b-strands are unfolded one at a time, leading to the evolution of intermediate states. More importantly, it was evident that the nature of the applied external force dictated the preferred mechanism for protein unfolding. Szymczak and others observed similar ow effects on employing Brownian dynamics simulations to examine owinduced stretching of integrin and ubiquitin.91,92 They noted that the hydrodynamic drag on a grafted protein chain in extensional ow increases along the chain length as the polymer unfolds. Therefore, the initiation of unfolding triggers a positive feedback mechanism which leads to the rapid unraveling of the entire chain to the exclusion of intermediate states. Indeed, it had previously been predicted that the uid drag acting on an initially coiled polymer increases as the molecule unfolds due to the decreased hydrodynamic interactions between its constituent monomers.84,93 In addition, the authors observed that whereas the grafted protein chains unfold through a series of intermediate states in uniform ow, consistent with that reported earlier,90 synthetic helices unraveled smoothly with no detectable intermediate states.91 This highlights the fact that proteins and homopolymers do indeed differ in their hydrodynamic properties in shear ow. Furthermore, it was also noted that the array of metastable conformations observed during protein unfolding in uniform ow was absent in mechanical pulling measurements in a force clamp apparatus such as the atomic force microscope.

742

Bekard et al.

A subsequent study examined the inuence of hydrodynamic interactions, which facilitate the folding of protein molecules into compact native structures, on the mechanical stretching of proteins using ubiquitin as a model system.94 Mechanical stretching at constant speed, at constant force, and through uid ow was examined. It was found that hydrodynamic interactions (cohesive forces) between amino acid residues facilitated unfolding (stretching) when a constant speed or force was applied to grafted ubiquitin. This observation was explained on the basis that an amino acid residue, which is pulled away from the bulk protein, creates a ow which drags other residues with it. In contrast, these cohesive forces were found to counter uid forces by screening monomers concealed in the protein matrix from the drag in the ow eld, thereby hindering ubiquitin unfolding in both uniform and elongational ow elds.94 It has been reported that hydrodynamic interactions increase with increasing protein size and hence inuence the critical shear rate required to unfold a protein of radius R.88 However, the authors note that this prediction pertains to solutions with high solvent viscosity and low interfacial tension. Recently, Sulkowska and Cieplak applied the coarsegrained MD model to investigate the resistance of 7510 proteins to mechanical stretching at a constant speed.95 Pulling was done by the termini to determine the maximum force of resistance, Fmax, of the protein chains. They restricted their study to non-fragmented protein chains deposited in the Protein Data Bank by August, 2005 and between 40 and 150 amino acid residues long. For specic sequential lengths of proteins, the average Fmax increased with increasing chain length. That is, longer protein chains showed better resistance relative to shorter chains. In relation to conformational classes, a-proteins were found to be weak and generated small Fmax values relative to b- and ab-proteins. Specifically, immunoglobulins and transport protein homologies were shown to yield larger Fmax values. However, the authors acknowledged that proteins with multi-a-domains may show a signicantly high resistance to mechanical stretching. In total, 134 proteins were considered to be strong with a threshold force of 170 pN. The authors proposed that the mechanical strength of the strongest proteins arises from a clamp consisting of long parallel b-strands which are stabilized by a network of intra- and intermolecular interactions.

Implications in Physiology and Bioprocessing

Proteins are known to show a strong conformation-function relationship. Destabilization of the native protein structure can potentially lead to the evolution of dysfunctional, aggre-

gation-prone species with undesirable effects.4 Reports in the literature demonstrate that shear, which is a common stress in both physiology20,58,96 and bioprocessing,13,18 can induce protein deformation.97 Therefore, it is important to analyze the relevance of shear stress to protein molecules under physiological ow and in bioprocessing. Pulsatile blood ow generates non-uniform shear stress throughout the arterial system during circulation.96 Arterial shear rates peak at 1640 s21, 96 reaching 104 s21, 58 in partly clogged coronary arterioles. The literature contains a signicant number of reports98107 (and references therein) linking the haemodynamic stress generated on arterial walls during microcirculation, specically in regions of relatively low shear stress, to the pathogenesis of atherosclerosis through the modulation of endothelial function.108 A characteristic hallmark of this disease state is the deposition of amyloidlike protein aggregates as plaque on arterial walls. Arterial amyloid-plaques may trigger vascular inammation, weaken the elasticity of blood vessels, and alter biochemical reactions such as lipid metabolism.109 Interestingly, it has been shown that molecular connement enhances the deformation of entangled polymers under squeeze ow,110 a condition which models pulsatile blood ow. Similarly, there are a considerably number of reports27,34,35,42,59,60,111113 citing haemodynamic shear stress as the primary facilitator of the proteolytic cleavage of large vWF multimers freshly released from endothelial cells as well as vWF-mediated platelet adhesion during haemostasis. The accumulation of large vWF multimers in the plasma may lead to disease states including thrombotic thrombocytopenic purpura, microvascular occlusion and atherosclerosis, due to the development of microvascular thrombi resulting from the complexation of the hyperactive vWF multimers and platelets at regions of low physiological shear.63,114 On the other hand, the smaller size of circulating vWF multimers, resulting from proteolysis, enhance their haemostatic potential in high shear regimes as occurs during arteriolar bleeding.62 The inuence of shear stress on the activity of platelets and vWF has been discussed in other review articles.57,58,63,115 Others have linked the deterioration in cerebral microcirculation (cerebral hypoperfusion) to the onset of protein conformational disorders such as Alzheimers disease.116118 It is suggested that reduced cerebral blood ow may replace old age as a risk factor for Alzheimers disease.118 It also proposed that haemodynamic shear stress regulates insulin-like growth factor-I activity indirectly, and possibly its effect on vascular pathologies.119 It is noteworthy that although protein conformational diseases and vascular disorders have received considerable attention in research, the role of physiological shear stress in
Biopolymers

Effects of Shear Flow on Protein Structure and Function

743

the initiation and/or progression of these debilitating diseases remain to be fully understood. In the case of bioprocessing, shear stresses are present during the separation, purication, shipping, and handling of protein based products.13 For example, processing steps involving ultraltration reactors,120,121 stirred tanks,122,123 homogenizers,124 and pumps53,125 induce high shear stresses. Simulated lobe pump studies have shown that high shear stress in these processing steps is sufcient to cause minor changes to protein structure.125 These structural changes may trigger the exposure of hitherto buried hydrophobic regions to the aqueous environment, initiating intra/intermolecular hydrophobic interactions leading to aggregation. The simulation further suggests that shear stresses may have played a crucial role in the protein aggregation behavior observed in a previous lobe pump study.126 Indeed, protein aggregation and particle formation during processing can affect the efciency of protein recovery, efcacy, safety, and product shelf-life.11 Furthermore, therapeutic protein aggregates are associated with a range of clinical side effects including non-specic anti-complementary activity leading to anaphylactic shock.127,128 In addition, insoluble aggregates can cause immune responses, lodge in the lung capillary bed, and block blood vessels.129131 The design of experiments to establish the effect of uid shear stress alone on protein structure and function has proved difcult. Existing shear devices generate other stresses including surface interactions and enhanced proteinair contacts, which complicates data interpretation.121,132,133 However, it has been established that the shear-stability of a protein molecule depends on: (i) its primary structure and molecular weight (ii) the magnitude of shear strain and the duration of its application, and (iii) the viscosity of its surrounding medium.31,64,65 Therefore, these factors must be taken in consideration for a complete understanding of the shear-stability of protein systems. Advancing knowledge in this eld will be crucial for the design of processes, especially in the bioprocessing industry, to ensure protein stability, maintain functionality, and promote process yield. Future development of protein based products will increasingly require several areas of improvement, especially in-process-monitoring of the stability of protein solutions. This will require the development of experimental techniques to allow the measurement of shear stress on proteins of interest. This is imperative as several studies show that individual protein systems demonstrate unique ow properties in a given ow eld. Additionally, it will be essential to appreciate _ t ) thresholds beyond which particular protein the shear (c systems destabilize in solution. This will inform the development of new processing equipment to improve efciency
Biopolymers

during production. It is noteworthy that knowledge from such studies could contribute to our understanding of the purported shear-induced pathogenesis of vascular and protein conformational disorders in vivo.

CONCLUSION
Protein unfolding and aggregation is an issue of concern both in vitro and in vivo, specically in bioprocessing and in the pathogenesis of disease. Shear rates less than 103 s21 have been found to signicantly alter the three-dimensional structure of globular proteins, leading to aggregation and amyloid _ t ) is found to be an imbril formation. The strain history (c portant parameter in the shear-effect. Therefore, even in environments where ostensibly low shear is generated can lead to shear-induced protein structural change. Given the diversity of protein species, and the diversity of molecular sizes, conformations and nature of intramolecular interactions, the susceptibility of proteins to shear can vary. It is clear that high shear generating processes should be avoided in bioprocessing. Shear also exists in the circulatory system and could be a signicant contributor to the pathogenesis of certain diseases by facilitating the aggregation of aberrantly folded proteins.

REFERENCES
1. Petsko, G. A.; Ringe, D. Protein Structure and Function; New Science Press: London, 2004. 2. Dobson, C. M. Trends in Biochemical Science 1999, 24, 329332. 3. Stefani, M.; Dobson, C. M. J Mol Med 2003, 81, 678699. 4. Dobson, C. M. Nature 2003, 426, 884890. 5. Chiti, F.; Dobson, C. M. Annu Rev Biochem 2006, 75, 333366. 6. Calamai, M.; Chiti, F.; Dobson, C. M. Biophys J 2005, 89, 42014210. 7. Wang, W. Int J Pharm 1999, 185, 129188. 8. Bryant, Z.; Pande, V. S.; Rokhsar, D. S. Biophys J 2000, 78, 584589. 9. Li, P.-C.; Makarov, D. E. J Chem Phys 2003, 119, 92609268. 10. Rohs, R.; Etchebest, C.; Lavery, R. Biophys J 1999, 76, 27602768. 11. Wang, W. Int J Pharm 2005, 289, 130. 12. Morel, M.-H.; Redl, A.; Guilbert, S. Biomacromolecules 2002, 3, 488497. 13. Thomas, C. R. In Chemical Engineering Problems in Biotechnology; Winkler, M. A., Ed.; Elsevier Applied Science: London, 1991, pp 2393. 14. Bekard, I.; Dunstan, D. E. J Phys Chem B Lett 2009, 113, 84538457. 15. Dunstan, D. E.; Hill, E. K.; Wei, Y. Macromolecules 2004, 37, 16631665. 16. Smith, D. E.; Babcock, H. P.; Chu, S. Science 1999, 283, 17241727.

744

Bekard et al. 48. Guyon, E.; Hulin, J. P.; Petit, L.; Mitescu, C. D. Physical Hydrodynamics; Oxford University Press: New York, 2001. 49. Taylor, G. I. Philosophical Transactions Royal Society London Series A 1923, 223, 289343. 50. van Wazer, J. R.; Lyon, J. W.; Kim, K. Y.; Colwell, R. E. Viscosity and Flow Measurement; Interscience Publishers: New York, 1963. 51. Charm, S. E.; Wong, B. L. Biotechnol Bioeng 1978, 20, 451453. 52. Bowen, W. R.; Gan, Q. Biotechnol Bioeng 1992, 40, 491497. 53. Virkar, P. D.; Narendranathan, T. J.; Hoare, M.; Dunnill, P. Biotechnol Bioeng 1981, 23, 425429. 54. Talboys, B. L.; Dunnill, P. Biotechnol Bioeng 1985, 27, 17301734. 55. Bekard, I. B.; Barnham, K. J.; White, L. R.; Dunstan, D. E. Soft Matter 2011, 7, 203210. 56. Dong, J.; Moake, J. L.; Nolasco, L.; Bernardo, A.; Arceneaux, W.; Shrimpton, C. N.; Schade, A. J.; McIntire, L. V.; Fujikawa, pez, J. A. Blood 2002, 100, 4033. K.; Lo 57. Sadler, J. E. Annu Rev Biochem 1998, 67, 395424. 58. Kroll, M. H.; Hellums, J. D.; McIntire, L. V.; Schafer, A. I.; Moake, J. L. Blood 1996, 88, 15251541. 59. Tsai, H. M.; Sussman, I. I.; Nagel, R. L. Blood 1994, 83, 2171. 60. Siedlecki, C. A.; Lestini, B. J.; Kottke-Marchant, K.; Eppell, S. J.; Wilson, D. L.; Marchant, R. E. Blood 1996, 88, 29392950. 61. Mody, N. A.; Lomakin, O.; Doggett, T. A.; Diacovo, T. G.; King, M. R. Biophys J 2005, 88, 14321443. 62. Zhang, X.; Halvorsen, K.; Zhang, C. Z.; Wong, W. P.; Springer, T. A. Science 2009, 324, 1330. 63. Sadler, J. E. Annu Rev Med 2005, 56, 173191. 64. Maa, Y.; Hsu, C. C. Biotechnol Bioeng 1997, 54, 503512. s, M. J Pharm a, A.; Farin a, J.; Llabre 65. Oliva, A.; Santoven Biomed Anal 2003, 33, 145155. 66. Ramirez-Alvarado, M.; Merkel, J. S.; Regan, L. Proc Natl Acad Sci 2000, 97, 89798984. 67. Ahmad, A.; Uversky, V. N.; Hong, D.; Fink, A. L. J Biol Chem 2005, 280, 42669. 68. Nielsen, L.; Khurana, R.; Coats, A.; Frokjaer, S.; Brange, J.; Vyas, S.; Uversky, V. N.; Fink, A. L. Biochemistry 2001, 40, 60366046. 69. Grainger, R. J.; Ko, S.; Koslov, E.; Prokop, A.; Tanner, R. D.; Loha, V. Appl Biochem Biotechnol 2000, 84, 761768. 70. Sluzky, V.; Tamada, J. A.; Klibanov, A. M.; Langer, R. Proc Natl Acad Sci USA 1991, 88, 93779381. 71. Sluzky, V.; Klibanov, A. M.; Langer, R. Biotechnol Bioeng 1992, 40, 895903. 72. Hamilton-Brown, P.; Bekard, I.; Ducker, W. A.; Dunstan, D. E. J Phys Chem B 2008, 112, 1624916252. 73. Sasahara, K.; Yagi, H.; Sakai, M.; Naiki, H.; Goto, Y. Biochemistry 2008, 47, 26502660. 74. Ohhashi, Y.; Kihara, M.; Naiki, H.; Goto, Y. J Biol Chem 2005, 280, 32843. 75. Sicorello, A.; Torrassa, S.; Soldi, G.; Gianni, S.; Travaglini-Allocatelli, C.; Taddei, N.; Relini, A.; Chiti, F. Biophys J 2009, 96, 22892298. 76. Stathopulos, P. B.; Scholz, G. A.; Hwang, Y. M.; Rumfeldt, J. A. O.; Lepock, J. R.; Meiering, E. M. Protein Sci 2004, 13, 30173027.

17. Elias, C. B.; Joshi, J. B. Adv Biochem Eng Biotechnol 1997, 59, 4772. 18. Yim, S. S.; Shamlou, P. A. Adv Biochem Eng Biotechnol 2000, 67, 83122. 19. Lin, J.; Meyer, J. D.; Carpenter, J. F.; Manning, M. C. Pharm Res 2000, 17, 391396. 20. Tangelder, G. J.; Slaaf, D. W.; Arts, T.; Reneman, R. S. Am J Physiol 1988, 254, H1059H1064. 21. Perkins, T. T.; Smith, D. E.; Chu, S. Science 1997, 276, 20162021. 22. Danker, G.; Vlahovska, P. M.; Misbah, C. Phys Rev Lett 2009, 102, 148102. 23. Munson, B. R.; Young, D. F.; Okiishi, T. H.; Huebsch, W. W. Fundamentals of Fluid Mechanics; Wiley: Hoboken, New Jersey, 2009. 24. Lee, A. T.; McHugh, A. J. Biopolymers 1999, 50, 589594. 25. Hill, E. K.; Krebs, B.; Goodall, D. G.; Howlett, G. J.; Dunstan, D. E. Biomacromolecules 2006, 7, 1013. 26. Dunstan, D. E.; Hamilton-Brown, P.; Asimakis, P.; Ducker, W.; Bertolini, J. Soft Matter 2009, 5, 50205028. 27. Singh, I.; Themistou, E.; Porcar, L.; Neelamegham, S. Biophys J 2009, 96, 23132320. 28. Dunstan, D. E.; Hamilton-Brown, P.; Asimakis, P.; Ducker, W. A.; Bertolini, J. Protein Eng Des Sel 2009, 22, 741746. 29. Imomoh, E.; Dusting, J.; Ashton, L.; Blanch, E.; Balabani, S. 14th International Symposium on Applications of Laser Techniques to Fluid Mechanics, Lisbon, Portugal, 2008. 30. Ashton, L.; Dusting, J.; Imomoh, E.; Balabani, S.; Blanch, E. W. Biophys J 2009, 96, 42314236. 31. van der Veen, M. E.; van Iersel, D. G.; van der Goot, A. J.; Boom, R. M. Biotechnol Progr 2004, 20, 11401145. 32. Charm, S. E.; Wong, B. L. Science 1970, 170, 466468. 33. Charm, S. E.; Lai, C. J. Biotechnol Bioeng 1971, 13, 185202. 34. Peterson, D. M.; Stathopulos, N. A.; Giorgio, T. D.; Hellums, J. D.; Moake, J. L. Blood 1987, 69, 625628. 35. Ikeda, Y.; Handa, M.; Kawano, K.; Kamata, T.; Murata, M.; Araki, Y.; Anbo, H.; Kawai, Y.; Watanabe, K.; Itagaki, I.; Sakai, K.; Ruggeri, Z. M. J Clin Investig 1991, 87, 12341240. 36. Thomas, C. R.; Dunnill, P. Biotechnol Bioeng 1979, 21, 2279 2302. 37. Thomas, C. R.; Nienow, A. W.; Dunnill, P. Biotechnol Bioeng 1979, 21, 22632279. 38. Maa, Y.; Hsu, C. C. Biotechnol Bioeng 1996, 51, 458465. 39. Tirrell, M.; Middleman, S. Biotechnol Bioeng 1975, 17, 299303. 40. Charm, S. E.; Wong, B. L. Biotechnol Bioeng 1970, 12, 11031109. 41. Jaspe, J.; Hagen, S. J. Biophys J 2006, 91, 34153424. 42. Schneider, S. W.; Nuschele, S.; Wixforth, A.; Gorzelanny, C.; Alexander-Katz, A.; Netz, R. R.; Schneider, M. F. Proc Natl Acad Sci 2007, 104, 78997903. 43. Chien, S.; Skalak, R. Biorheology 1981, 18, 307330. 44. Bell, D. J.; Dunnill, P. Biotechnol Bioeng 1982, 24, 12711285. 45. Cathey, C. A.; Fuller, G. G. J Non-Newtonian Fluid Mechanics 1990, 34, 6388. del, W. In Instrumentation and Sensors for the Food Indus46. Ro try; Kress-Rogers, E., Ed.; Butterworth-Heinemann: Oxford, 1993, pp 375415. 47. Hill, E. K.; Watson, R. L.; Dunstan, D. E. J Fluorescence 2005, 15, 255266.

Biopolymers

Effects of Shear Flow on Protein Structure and Function 77. Akkermans, C.; Venema, P.; Rogers, S. S.; van der Goot, A. J.; Boom, R. M.; van der Linden, E. Food Biophys 2006, 1, 144150. 78. Akkermans, C.; van der Goot, A. J.; Venema, P.; van der Linden, E.; Boom, R. M. Food Hydrocolloids 2008, 22, 1315 1325. 79. Alexander-Katz, A.; Netz, R. R. Europhys Lett (EPL) 2007, 80, 18001. 80. Bee, J. S.; Stevenson, J. L.; Mehta, B.; Svitel, J.; Pollastrini, J.; Platz, R.; Freund, E.; Carpenter, J. F.; Randolph, T. W. Biotechnol Bioeng 2009, 103, 936. 81. Adamson, A. W.; Gast, A. P.; NetLibrary, I. Physical Chemistry of Surfaces; Wiley: New York, 1997. 82. Best, R. B.; Brockwell, D. J.; Toca-Herrera, J. L.; Blake, A. W.; Smith, D. A.; Radford, S. E.; Clarke, J. Anal Chim Acta 2003, 479, 87105. 83. Agarwal, U. S.; Mashelkar, R. A. J Chem Phys 1994, 100, 60556061. 84. De Gennes, P. G. J Chem Phys 1974, 60, 50305042. 85. Greeneld, N. J.; Fasman, G. D. Biochemistry 1969, 8, 4108 4116. 86. JiJi, R. D.; Balakrishnan, G.; Hu, Y.; Spiro, T. G. Biochemistry 2006, 45, 3441. 87. Immaneni, A.; McHugh, A. J. Biopolymers 1998, 45, 239 246. 88. Alexander-Katz, A.; Netz, R. R. Macromolecules 2008, 41, 33633374. 89. Buguin, A.; Brochard-Wyart, F. Macromolecules 1996, 29, 49374943. 90. Lemak, A. S.; Lepock, J. R.; Chen, J. Z. Y. Proteins Struct Funct Genet 2003, 51, 224235. 91. Szymczak, P.; Marek, C. J Chem Phys 2006, 125, 164903. 92. Szymczak, P.; Cieplak, M. J Chem Phys 2007, 127, 155106. 93. Agarwal, U. S.; Rohit, B.; Mashelkar, R. A. J Chem Phys 1998, 108, 16101617. 94. Szymczak, P.; Cieplak, M. J Phys Condensed Matter 2007, 19, 285224285235. 95. Sulkowska, J. I.; Cieplak, M. Biophys J 2008, 94, 613. 96. Stroev, P. V.; Hoskins, P. R.; Easson, W. J. Atherosclerosis 2007, 191, 276280. 97. McNally, E.; Lockwod, C. In Protein Formulation and Delivery; McNally, E., Ed.; Marcel Dekker: New York, 2000, pp 111138. rge, G.; Haude, M.; Meyer, J. Br Heart J 98. Ge, J.; Erbel, R.; Go 1995, 73, 462465. 99. Ku, D. N.; Giddens, D. P.; Zarins, C. K.; Glagov, S. Arterioscler Thromb Vasc Biol 1985, 5, 293302. 100. Malek, A. M.; Alper, S. L.; Izumo, S. JAMA 1999, 282, 20352042. 101. Cunningham, K. S.; Gotlieb, A. I. Lab Invest 2004, 85, 9 23. 102. Krams, R.; Wentzel, J. J.; Oomen, J. A. F.; Vinke, R.; Schuurbiers, J. C. H.; De Feyter, P. J.; Serruys, P. W.; Slager, C. J. Arterioscler Thromb Vasc Biol 1997, 17, 2061. 103. Gibson, C. M.; Diaz, L.; Kandarpa, K.; Sacks, F. M.; Pasternak, R. C.; Sandor, T.; Feldman, C.; Stone, P. H. Arterioscler Thromb Vasc Biol 1993, 13, 310.

745

104. Shaaban, A. M.; Duerinckx, A. J. Am J Roentgenol 2000, 174, 1657. 105. Vita, J. A.; Treasure, C. B.; Ganz, P.; Cox, D. A.; David Fish, R.; Selwyn, A. P. J Am Coll Cardiol 1989, 14, 11931199. 106. Davies, P. F.; Remuzzi, A.; Gordon, E. J.; Dewey, C. F.; Gimbrone, M. A. Proc Natl Acad Sci USA 1986, 83, 2114. 107. Irace, C.; Cortese, C.; Fiaschi, E.; Carallo, C.; Farinaro, E.; Gnasso, A. Stroke 2004, 35, 464. 108. Jiang, Y.; Kohara, K.; Hiwada, K. Stroke 2000, 31, 23192324. 109. Howlett, G. J.; Moore, K. J. Curr Opin Lipidol 2006, 17, 541. 110. Rowland, H. D.; King, W. P.; Cross, G. L. W.; Pethica, J. B. ACS Nano 2008, 2, 419428. 111. Alevriadou, B. R.; Moake, J. L.; Turner, N. A.; Ruggeri, Z. M.; Folie, B. J.; Phillips, M. D.; Schreiber, A. B.; Hrinda, M. E.; McIntire, L. V. Blood 1993, 81, 1263. 112. Goto, S.; Salomon, D. R.; Ikeda, Y.; Ruggeri, Z. M. J Biol Chem 1995, 270, 23352. 113. Chow, T. W.; Hellums, J. D.; Moake, J. L.; Kroll, M. H. Blood 1992, 80, 113. 114. Tsai, H. M. Semin Thromb Hemost 2003, 29, 479488. 115. Reininger, A. J. Haemophilia 2008, 14, 1126. 116. de la Torre, J. C. Lancet Neurol 2004, 3, 184190. 117. de La Torre, J. C. Neurol Res 2004, 26, 517524. 118. Crawford, J. G. Med Hypotheses 1998, 50, 2536. 119. Elhadj, S.; Akers, R. M.; Forsten-Williams, K. Ann Biomed Eng 2003, 31, 163170. 120. Ahrer, K.; Buchacher, A.; Iberer, G.; Jungbauer, A. J Membr Sci 2006, 274, 108115. 121. Kim, K. J.; Chen, V.; Fane, A. G. Biotechnol Bioeng 1993, 42, 260265. 122. Cutter, L. A. AIChE J 1966, 12, 3545. 123. Kresta, S. M.; Wood, P. E. AIChE J 1991, 37, 448460. 124. Floury, J.; Bellettre, J.; Legrand, J.; Desrumaux, A. Chem Eng Sci 2004, 59, 843853. 125. Gomme, P. T.; Prakash, M.; Hunt, B.; Stokes, N.; Cleary, P.; Tatford, O. C.; Bertolini, J. Biotechnol Appl Biochem 2006, 43, 113120. 126. Gomme, P. T.; Hunt, B.; Tatford, O. C.; Johnston, A.; Bertolini, J. Biotechnol Appl Biochem 2006, 43, 103111. 127. Thornton, C. A.; Ballow, M. Arch Neurol 1993, 50, 135136. 128. Wang, Y.-C. J.; Hanson, M. A. J Parenteral Sci Technol 1988, 42, S3S26. 129. Katakam, M.; Banga, A. K. J Pharm Pharmacol 1995, 47, 103107. 130. Cleland, J. L.; Powell, M. F.; Shire, S. J. Crit Rev Ther Drug Carrier Syst 1993, 10, 307377. 131. Dorris, G. G.; Bivins, B. A.; Rapp, R. P.; Weiss, D. L.; DeLuca, P. P.; Ravin, M. B. Anesth Analg 1977, 56, 422427. 132. Meireles, M.; Aimar, P.; Sanchez, V. Biotechnol Bioeng 1991, 38, 528534. 133. Narendranathan, T. J.; Dunnill, P. Biotechnol Bioeng 1982, 24, 21032107.

Reviewing Editor: Stephen Blacklow

Biopolymers