The Pennsylvania State University

The Graduate School

Eberly College of Science

SELF-PROPELLED SYSTEMS FOR VERSATILE APPLICATIONS

A Dissertation in

Chemistry

by

Vinita Yadav

2015 Vinita Yadav

Submitted in Partial Fulfillment of the Requirements

for the Degree of

Doctor of Philosophy

May 2015

The dissertation of Vinita Yadav was reviewed and approved* by the following:

Ayusman Sen Distinguished Professor of Chemistry Dissertation Advisor Chair of Committee Thomas E. Mallouk Evan Pugh Professor of Chemistry, Physics, Biochemistry and Molecular Biology Associate Head of the Chemistry Department Associate Director, Penn State MRSEC Director, Center for Solar Nanomaterials

John Badding Professor of Chemistry Associate Head for Equity and Diversity Director of Graduate Recruiting

James H. Adair Professor of Materials Science and Engineering Biomedical Engineering and Pharmacology Barbara J. Garrison Head of the Chemistry Department Shapiro Professor of Chemistry

*Signatures are on file in the Graduate School

iii

ABSTRACT

A decade ago, the first examples of self-propelled motion at the nano and

microscale by synthetic objects were discovered. This was the first step towards the

design of autonomous nano and micro-machines and robots. Nature has been using

nanoscale motors and pumps to power its numerous creations and this has inspired the

scientific community to emulate such systems. The focus of this thesis is on colloidal

systems - both biological and inorganic, whose constituents move and respond to each

other and their surroundings through a specific mechanism: diffusiophoresis.

This thesis begins with an introduction on diffusiophoresis - the electrolyte and

non-electrolyte versions along with other competing or complementing propulsion

mechanism reported for colloidal systems.

The first system discussed in this thesis is an inorganic scheme that displays a

one of its kind ‘on/off’ switch that controls colloidal transport. Additional built-in levels of

regulation allow for both rectification and amplification of particle motion.

A biological system is discussed next that utilizes the phenomenon of electrolyte

diffusiophoresis to detect and repair cracks in bones. This represents one of the few

viable examples of utilizing nanomotors towards a medical treatment. Repair of

damaged tissues has also been expanded to curing dental ailments. Dental caries or

bacterial cavities can also be detected and cured using the same underlying mechanism.

This approach also offers the first explanation on why fluoride treatment works for

general dental well-being.

Restoration of biological cracks has also been expanded onto polymerized

surfaces. The mechanism involved varies from diffusiophoresis, in that it is density

iv

driven rather than being electric field driven. Complete repair of cracked surfaces is

observed in real time.

Diffusiophoretic motion is then applied to polymeric systems where a fluoride ion

triggered colloidal pump is designed. The pump is a versatile starting ground that is

expanded into designing a bacteria scavenging material as well as systems that show

first signs of memory.

This thesis concludes with perhaps the most exciting chapter that brings new

light to enzymatic cascades, the intricate systems that allow for the perpetuity of life on

earth. Earlier work done on enzyme substrate interactions is expanded to solve the

mechanistic mystery behind cascades that has eluded enzymologists and biologists for

long.

v

TABLE OF CONTENTS

List of Figures…………………………………………………………........………...…..viii

List of Tables……………………………………………………………………….…......xvi

List of Multimedia Files………………………………………………………….…...…...xvii

Acknowledgements…………………………………………………………...................xix

Chapter 1 Difusiophoresis- An Introduction ........................................................ 1

1.1 Reynolds Number and Brownian Motion ..................................................... 1 1.2 Role of Debye Length in Phoretic Transport .............................................. 4 1.3 Mechanisms of Motility ............................................................................... 6

1.3.1 Self-Electrophoresis .......................................................................... 7 1.3.2 Self-Diffusiophoresis ......................................................................... 11 1.3.3 Electrolyte Diffusiophoresis ............................................................... 11 1.3.4 Non-Electrolyte Self-Diffusiophoresis ................................................ 14 1.3.5 Self-Electrophoresis vs Electrolyte Self-Diffusiophoresis .................. 15 1.3.6 Enzyme motors ................................................................................. 16 1.3.7 Chemotaxis ....................................................................................... 17 1.3.8 Enzyme Pumps ................................................................................. 20

1.4 Other Mechanisms...................................................................................... 22 1.4.1 Bubble Propulsion ............................................................................. 22 1.4.2 Magnetically-driven Motors ............................................................... 24 1.4.3 Acoustically-powered Motors ............................................................ 27

1.5 Conclusion .................................................................................................. 29 1.6 References ................................................................................................. 30

Chapter 2 Triggered “On/Off” Micro-Pumps and Colloidal Photo-Diode ........... 36

2.1. Introduction ................................................................................................ 36 2.2 Design of Smart Micro-Pumps .................................................................... 36 2.3 Propulsion Mechanism ............................................................................... 37 2.4 Switchable Photoacid Pump ....................................................................... 39

2.4.1 Experimental Set-Up ......................................................................... 39 2.4.2 ‘On/Off’ Pump in Action ..................................................................... 39 2.4.3 Separation of Diffusiophoretic and Electroosmotic Motion ................ 42 2.4.4 Self- Assembled Patterns ................................................................. 44

2.5 pH Controlled Polymer Pump ..................................................................... 46 2.6. Photo-Colloidal Diode ................................................................................ 50

2.6.1 Experimental Set-Up ......................................................................... 51 2.6.2 Spatial and Temporal Regulation of Colloidal Transport ................... 52

2.7 Conclusion .................................................................................................. 54 2.8 Acknowledgement ...................................................................................... 54 2.9 References ................................................................................................. 55

vi

Chapter 3 Bone-Crack Detection, Targeting and Repair Using Ion Gradients .. 57

3.1. Introduction ................................................................................................ 57 3.2 Motivation ................................................................................................... 57 3.3 Generation of Local Electric Fields ............................................................. 58 3.4 Experimental Design ................................................................................... 61 3.5 Diffusiophoresis led Damage Detection ...................................................... 62 3.6 Disfussiophoresis Guided Targeted Protein delivery ................................... 66

3.6.1 Fluorescence Microscopy analysis .................................................... 66 3.6.2 Raman Spectroscopy Analysis ......................................................... 66

3.7 Targeted Drug Delivery ............................................................................... 69 3.7.1 Synthesis of Alendronate Nanoparticle ............................................. 69 3.7.2 Drug load Calculation........................................................................ 69 3.7.3 Particle Characterization ................................................................... 70 3.7.4 Drug Delivery and Cell Proliferation Assay ........................................ 72

3.8 Expansion of the detection and repair technique ........................................ 75 3.8.1 Present therapeutic techniques ......................................................... 75 3.8.2 Detection using FDA approved diagnostic dye .................................. 77 3.8.3 Mechanism of Fluoride treatment ...................................................... 79

3.9 Application on Synthetic Surfaces- Polymer Repair .................................... 81 3.9.1. Motivation ........................................................................................ 81

3.9.2 Density Driven Flows ........................................................................ 82 3.9.3 Synthesis of repair agents ................................................................ 83 3.9.4 Polymer Repair ................................................................................. 83 3.9.5 Enzymatic repair ............................................................................... 85

3.10. Conclusion ............................................................................................... 89 3.11. Acknowledgements .................................................................................. 89 3.12 References ............................................................................................... 90

Chapter 4 A Self-Powered Polymeric Material that Responds Autonomously and Continuously to Fleeting Stimuli ............................................................ 94

4.1 Introduction ................................................................................................. 94 4.2. Experimental Design .................................................................................. 94 4.3. Results and Discussion ............................................................................. 97

4.3.1 Colorimetric Analysis ........................................................................ 99 4.3.2. Stimuli Responsive Pumping Behaviour........................................... 101 4.3.3 Memory based Pumping in the Absence of Stimuli ........................... 103

4.4 Difusiophoretic Pumping- Scavenger Design .............................................. 107 4.5 Conclusion .................................................................................................. 110 4.6 Acknowledgements..................................................................................... 111 4.7 References ................................................................................................. 112

Chapter 5 Substrate-driven Chemotatic Assembly in Enzyme Cascades ......... 114

5.1 Introduction ................................................................................................. 114 5.2 Motivation ................................................................................................... 115 5.3 Experimental Design ................................................................................... 115

5.3.1 Microfluidic device fabrication ........................................................... 117

vii

5.3.2 Fluorescent tagging of HK and Ald.................................................... 118 5.3.3 Fluorescence Correlation Spectroscopy............................................ 120 5.3.4 Statistical Significance Analysis of FCS data .................................... 123 5.3.5 Confocal Microscope Imaging ........................................................... 123 5.3.6. Detailed Investigation into Hexokinase Chemotaxis Behavior .......... 124 5.3.7 Substrate Triggered Chemotaxis ...................................................... 124 5.3.8 Binding Affinity VS Turnover Rate ..................................................... 127 5.3.9 Enzyme activity assays ..................................................................... 127 5.3.9 Investigation into Aldolase Chemotaxis ............................................. 128 5.3.10 Why Chemotaxis? Enhanced Diffusion Model ................................. 128 5.3.11 Inadequacy of the Diffusion Model to Explain Chemotaxis .............. 129

5.4. Enzyme Cascade Investigation ................................................................. 133 5.4.1 Cascade In-situ ................................................................................. 136 5.4.1.1 Progress Curve Simulation ............................................................ 136 5.4.2 Competitive Substrates ..................................................................... 139

5.5. Chemotaxis and Metabolons ..................................................................... 139 5.5.1. Chemotaxis in Cytosolic Conditions ................................................. 142

5.6 Conclusion .................................................................................................. 143 5.7 Acknowledgements .................................................................................... 143 5.8 References ................................................................................................. 144

Chapter 6 Bringing discipline into enzyme motors ............................................. 145

6.1. Introduction ................................................................................................ 145 6.2 Motivation ................................................................................................... 145 6.3. Experimental Design .................................................................................. 146

6.3.1 Test Subject 1: Catalase ................................................................... 147 6.3.2 Test Subject: Urease ........................................................................ 149

6.4 Results and Discussion .............................................................................. 155 6.5 Conclusions ................................................................................................ 156 6.6 References ................................................................................................. 158

Chapter 7 Conclusions.......................................................................................... 159

viii

LIST OF FIGURES

Figure 1-1. The electric double layer of a charged particle in a polar solution. The counter ions from the solution come near the charged particle surface to neutralize the charge and this fluid layer remains diffused around the particle. The ζ-potential is the electric potential at the shear plane or outer edge of the Stern layer. A non-spherical charged surface behaves the same way.............................................................................................……5

Figure 1-2. Propulsion of bimetallic Au-Pt rods in hydrogen peroxide solution powered

by self-electrophoresis. Catalytic redox reaction on the two metallic ends generates the local electric field…………………………………...........…….8

Figure 1-3. An immobilized bimetallic surface can generate fluid flow in its vicinity by

the generating a local electric field in the same manner as a bimetallic motor. The schematic describes electrochemical conversion of hydrogen peroxide on the two metallic surfaces- gold and silver, the generated electric field and the directional motion imparted to positively charged carboxyl functionalized polystyrene (carboxy-PS) and negatively charged amidine functionalized polystyrene (amidine-PS) particles…………..…...10

Figure 1-4. Schematic depiction of diffusiophoretic motion. The difference in diffusivity

of the ions generated from the source causes a local electric field. The double layer around the particles as well as the wall responds to the thus formed electric field leading to electrophoretic and electroosmotic motion respectively. In the example in the Figure above, the anion diffuses faster than the cation generating an electric field from right to left. The electrophoretic motion of a negatively charged particle is from left to right. Correspondingly, the electroosmotic flow along the negatively charged wall is from right to left. The concentration gradient also leads to thickness gradient of double layers on the surfaces of the particle and wall, and in-turn a pressure difference that propels particles from left to right.......…..13

Figure 1-5. Collective behavior demonstrated by synthetic motors. Au-Pt bimetallic

nanomotors chemotax towards the source of hydrogen peroxide fuel (the gel in the upper left side), as depicted by an increase in the number of rods over time………………….……………………………………………..………18

Figure 1-6.Schematic depiction of fabrication and functioning of enzymatic

micropumps. (a) Au patterned on a PEG-coated glass surface is functionalized with a quaternary ammonium thiol, which electrostatically binds to the negatively charged groups on the enzyme. Triggered fluid pumping is initiated by introducing enzyme specific substrate. (b) Cascading fluid pumping is observed when enzyme catalase is actuated by production of its substrate in situ by enzyme glucose oxidase and its substrate glucose enabling microfluidic regulation and logic…...…………21

ix

Figure 1-7. Bubble propulsion mechanism. Oxygen microbubbles are generated through decomposition of hydrogen peroxide. As the bubbles detach from the motors, the associated recoil force pushes motors in the opposite direction………………………………………………….………………...……23

Figure 1-8. Magnetic manipulation of cage-like micromotors for transportation of cells.

(a) SEM image of a hexahedral microrobot after cell culture and (b) an enlarged SEM image. Confocal microscope images of the (c) hexahedral and (d) cylindrical microrobots after staining of the cells…………………..26

Figure 1-9. Acoustic powered self-propelled motors. (a) Propagation and assembly of

bimetallic rods under acoustic fields. (b) Navigation of an acoustically-powered motor towards a HeLa cell under magnetic field-guidance…….28

Figure 2-1. A schematic depiction of PAG pumping mechanism. The negative surface

charge of the glass creates a positive double layer, which in response to the generated ions causes an inward electroosmotic flow. The negatively charged tracers (S-PS particles) move opposite to the direction of the electric field, competing against the electroosmotic flow while the positively charged tracers (NH2-PS particles) move along the electric field direction aided by the electroosmotic flow……………………………………………..38

Figure 2-2. Optical microscope images of particle motion. (a) and (b) show the

distribution of the positively charged tracers (NH2-PS) around the photoacid (PAG-1) microcrystallites with UV off (control) and after 1 min of UV illumination respectively. (c) and (d) display the same for the negatively charged tracers (S-PS). Each of the tracer particles seen is 2 µm. (Also see Supporting Video 2-1 and 2-2)…………….……..…….….41

Figure 2-3. Velocity distribution histograms obtained for (a) NH2-PS particles and (b)

S-PS particles using the PAG pump……………………………………...….43 Figure 2-4. Patterns induced by PAG pumping. (a) Control image, NH2-PS particle

distribution around a single photoacid crystallite with UV off. (b) Self-assembled NH2-PS particle pattern with UV on. Each of the tracer particles seen is 2 µm……………………………………………………………….……44

Figure 2-5. A schematic depiction of the pattern. The pump pulls the NH2-PS particles

out from the large reservoirs (10 x 10 mm2) into the micro-channels (4 x 1 mm2), towards the PAG chambers (1 x 1 mm2) on either side of the channel...………………………………………………………………………..45

Figure 2-6. Schematic depiction of PFA-S pumping mechanism. The local electric field

points outwards away from the polymer film and the negatively charged tracers COOH-PS particles move inwards, towards film………..…………46

Figure 2-7. Optical microscope images of PFA-S film pumping away HOOC-PS tracers

(6 µm). (a) Image taken 0 s after exposure to 1 M HCl in deionized water at 25 °C, and (b) 1200 s after exposure. ……………………………………47

x

Figure 2-8. Velocity distribution histograms of HOOC-PS tracers as a function of the

acid concentration for the PFA-S pump……………………………………..48 Figure 2-9. Velocity distribution histograms of HOOC-PS tracers at 100 to 1100 µm

away from the PFA-S pump upon addition of 1 M HCl to the PFA-S film at 25 °C, demonstrating long range pumping……………...…………………..49

Figure 2-10. Schematic depiction of source (PAG)-drain (PFA-S) based colloidal photo-

diode indicating both the rectification and the direction of movement of S-PS particles…………………………………………………………….……….50

Figure 2-11.Spatial and temporal regulation of velocity (S-PS particles) attained using

the source-drain photo-diode. Distance is measured from edge of the PAG and time is measured from when the UV is turned on. For velocity vs time plot, distance = 150 µm; for velocity vs distance, time = 20 s…………….53

Figure 3-1. Schematic depiction of ion gradient-induced electric field and the resultant

particle migration. The length of the arrows next to the ions represent their relative mobilities. The generated electric field points outwards away from the crack. Accordingly, the negatively charged particles move towards and positively charged particles move away from the crack……………..…….60

Figure 3-2. Increasing quantum dot intensity within the crack on bone surface (a) and

PDMS surface (b) demonstrating an effective damage detection scheme. Scale bar is 60 µm. Right panel shows calculated intensities inside the damage (averaged over entire damaged area) for HOOC Q-Dots, amine Q-Dots and control, using Image J software, for bone surface (c) and PDMS surface (d)………………………………………………………………63

Figure 3-3. Analysis of the crack detection scheme using confocal microscopy.

Intensity study within the crack on bone surface (a) and PDMS surface (b) using amine functionalized quantum dots. Control images showing no intensity change on bone (c) & PDMS (d). Scale bar is 130 µm……...….65

Figure 3-4. (a) Raman spectra obtained on the bone and enzyme separately, overlaid

with one collected on the bone exposed to the enzyme. (b) Raman spectra at increasing distances from the crack depicting the preferential enzyme migration towards the crack…………………………………………………..68

Figure 3-5. Electron microscopy analysis of drug loaded particles: SEM images of

PLGA nanoparticles coated with Au/Pd sputter coating for visualization...............................................................................................71

Figure 3-6. Increasing fluorescence intensity within the crack indicates active migration

of Nile-red tagged drug loaded PLGA particles to the crack site demonstrating an effective drug delivery protocol. Scale bar is 100 µm..............................................................................................................72

xi

Figure 3-7. Proliferation of MG-63 cells treated with PLGA nanoparticles containing 10-

6, 10-8 and 10-10 M alendronate for 48 hours, expressed as percentage optical density relative to the negative control of 100%, using a colorimetric MTS cell proliferation assay. (Graph expressed as Mean ± SD; Significance (*P < 0.05) compared with negative control group (medium alone))…………………………………………….……………………………..74

Figure 3-8.Damage detection in cracked teeth. Negatively charged amine

functionalized quantum dots move in towards the crack leading to an increase in fluorescence intensity (a) while positively charged carboxyl functionalized quantum dots move away from the crack leading to decrease in fluorescence intensity (c). Images (b) & (d) are the bright field images of the tested crack……………………………….……………………76

Figure 3-9. (a) Increasing fluorescein intensity within the dental crack in a tooth slice

leads to detection. (b) Fluorescence intensity analysed inside the damage (averaged over entire damaged area) through Image J……………...……77

Figure 3-10. Damage detection on a whole tooth using fluorescein dye………..….…..78 Figure 3-11.EDS measurements at increasing distances from the crack, show a

decreasing fluoride signal……………………………………………….…….79 Figure 3-12. EDS maps generated at the crack site show a heavy deposition of sodium

and fluoride at the crack site. The presence of the crack can be noted by the scarcity of calcium, phosphate and oxygen at the same site, the primary components of hydroxyapatite Scale bar is 400 µm………….….80

Figure 3-13.ESEM images of polymer deposition at the damage site. The strategy

works well for both single (a, b) and multiple cracks (c, d). (a, c) The image of cut polymer with no salt after 1 hr. exposure to emulsions. (b, d) PDMS/CaCl2 in inverted setup after 1 hr. exposure to emulsions…….….84

Figure 3-14.Schematic of a surface healing system using a salt/PDMS film. The urease

enzymes (blue) and urea molecules (grey) move over the crack due to density driven flows. While this occurs, the urea is converted by the urease to carbonate ions (pH~10.3). The carbonate ions then react with the leaching calcium ions forming solid calcium carbonate…………………...85

Figure 3-15. (a) ESEM images showing the control (left) and sample (right) where the

crack was exposed to the urease-urea mixture without and with the underlying calcium chloride layer, respectively. Scale bar is 100µm. (b) XRD Analysis of the crack site confirming the presence of calcite (red bars-standard) and aragonite (blue bars-standard). The amorphous halo at lower two-theta values is due to PDMS……………………………………...87

Figure 3-16.(a)SEM image of the precipitated material within the crack showing

aragonite and calcite like morphology. Scale bar is 20 µm. (b) Carbonate vibration bands61 around 1460 (symmetric stretching) and 880 (out-of-

xii

plane bending) cm-1 confirms the presence of the precipitated calcium carbonate……………………………………………………….….……………88

Figure 4-1. Schematic depiction of polymer microsphere pump that induces the

movement of fluid that surrounds the pump in response to a specific stimulus, even after the stimulus has been removed. The blue arrows represent the direction of fluid movement, and the sizes of the arrows illustrate an approximation of the relative magnitude of fluid flow when the signal is present or absent. When the UV light is off, a self-propagating reaction enables the microsphere to continue generating a concentration gradient of products that drive the pumping response. The signal transduction reagents (fluoride ion) translate the first reaction with UV light to initiation of the self-propagating reaction. The byproduct of the reactions (3) is yellow/orange in color and, thus, turns the microsphere from colorless to yellow to orange over the course of the pumping response………………………………………………………………………...96

Figure 4-2. Structures and reactions of reagents 1 and 2 that are grafted onto a 300

µm-diameter TentaGel microsphere. (a) A microsphere that contains a 1:1 ratio of reagents 1 and 2. (b) Exposure of this microsphere to UV light causes the activity-based detection reagent (1) to release fluoride, compound 3, and protons (exist as pyridinium ions). (c) The released fluoride then reacts with 2 to initiate a self-propagating reaction that amplifies fluoride, 3, and protons (exist as pyridinium ions). The gradient of these small molecules causes fluid movement around the microsphere (i.e., pumping). The notation “n” refers to the number of cycles of the autoinductive reaction in (c)……………………………………….………....98

Figure 4-3. Colorimetric response of a TentaGel microsphere that contained 100% of

2. (a) The procedure for testing the autoinductive, self-propagating reaction that is mediated by 2. The product of the autoinductive reaction is 3 (Figure 4-2c), which turns the microspheres a yellow/orange color (b). (c) This color reflects the extent of the autoinductive reaction,15,19 and can be quantified by photographing the microspheres and using image processing software to measure the intensity of color. Exposure of the microspheres to substoichiometric quantities of fluoride (relative to the loading level of the microspheres) reveals sigmoidal kinetics characteristic of autoinductive reactions.15,19 Note that the scale of the x-axis changes after the break………………………………………….……………………..100

Figure 4-4. Average pumping speeds caused by TentaGel microspheres exposed to

365 nm light. (a) The pumping action can be switched on and off for a microsphere functionalized with 1 only (blue data), whereas no pumping was observed for microspheres functionalized with only 2 (orange data). In contrast, the pumping speed could be varied (but not turned off) for microspheres functionalized with both 1 and 2 by turning on and off the UV light (black data). The pumping speeds reflect the averages of measurements acquired over 30 s intervals that span the length of the data bars. (b) Continuous pumping also is possible using microspheres

xiii

that are functionalized with 1 and 2 once the microspheres are exposed to UV light for 20 min. For both (a) and (b), the average pumping speeds were obtained by tracking the distance that 30 tracer beads traveled over time………………………………………………………..……………………102

Figure 4-5. Optical microscope images of the microsphere functionalized with reagent

1 alone, triggered with UV light that induces motion to the surrounding (a) negatively charged polystyrene particles at 5X magnification. A zone of exclusion can clearly be seen around the microsphere where tracer particles have been pushed away. (b) Positively charged particles were pulled in towards the microsphere, eventually getting trapped inside at 20X magnification. The imaged polystyrene particles are each 2 µm in diameter………………………………………………….…………………….108

Figure 4-6. (a) Electron microscope images of the polystyrene particles trapped within

the microsphere. (b) Shows a zoomed in image of the chipped part confirming the particles to be trapped inside the permeable body and not just on the surface of the microsphere. The scale bar is 50 µm.………109

Figure 5-1.The glycolysis cycle.7 The first four enzymes, hexokinase (HK),

phosphogluco isomerase (Iso), phosphofructokinase (PFK) and aldolase (Ald) were examined for their ability to undergo chemotactic assembly..................................................................................................117

Figure 5-2. Photo-lithographically fabricated flow based microfluidic gradient generator

for studying enzyme chemotaxis. The length of the channels is either 20 or 40 mm, width 360 μm, and the height is 100 μm. Considering laminar flow, the width of each channel is 120 µm. Fluorescence intensities were analyzed along a vertical line as shown in the figure leaving off 20 µm next to the sidewalls……………………………………….……………….………118

Figure 5-3. Fluorescence correlation spectroscopy (FCS) results showing an enhanced

diffusion coefficient for Ald (a) and HK (b) in the presence of their respective substrates…………………………….…………………….…….123

Figure 5-4.Chemotactic response observed for hexokinase (HK). HK shows

chemotactic shift only in presence of a gradient of its substrate, D-glucose (D-Glu) and is unaffected by the presence of L-glucose (L-Glu). Also, hexokinase shows a greater chemotactic shift towards its substrate of choice D-glucose (D-Glu) compared to mannose (Mann) which it phosphorylates at a significantly lower rate. Experimental conditions: Starting enzyme concentration = 200 nM (100%) Flow rate = 200µl/h, distance = 30 mm, interaction time = 6.48 s; Percentage of enzyme migration into the left D-glucose channel is 4.59 ± 0.4 % and towards the right D-glucose channel is 4.54 ± 0.3 %. Percentage of enzyme migration into mannose channel is 2.85 ± 0.5 %. Inset on the top and bottom shows a clearer migration towards preferred channels. Note that the percent enzyme migration into adjoining buffer channels due Brownian diffusion alone is ~ 2%...........................................................................................127

xiv

Figure 5-5. Substrate-induced enzyme focusing. The normal diffusional spreading of HK (1 µM) towards the flanking channels that contain buffer is reduced if the composition in the middle channel is switched from HK in buffer to HK in 70 mM D-glucose. The net reduction in area is 13.4 ± 3.0%. Experimental conditions: Flow rate = 100µl/h, distance = 18 mm, interaction time = 7.78 s………………………………………….………….131

Figure 5-6. Cofactor-induced enzyme focusing. The enzyme (1 µM) switches from an

equilibrium distribution to a non-equilibrium one when cofactors ATP (50 mM) and MgCl2 (100 mM) are introduced in the middle channel. This is analogous to reported cellular responses in the cytosol where enzyme association is regulated by oxygenation and phosphorylation requirements. Experimental conditions: Flow rate = 30µl/h, distance = 19 mm, interaction time = 24.7 s…………………………………………………………………..132

Figure 5-7. Restricted chemotaxis in the absence of substrate gradient. The normal

diffusional spreading of HK (200 nM) towards the flanking substrate channels is reduced if the substrate is also introduced within the middle channel flowing the enzyme. Experimental conditions: Flow rate = 100µl/h, distance = 20 mm, interaction time = 8.64 s………………………….……133

Figure 5-8. (a) Experimental set-up to study the chemotactic response of Ald (green

channel) towards the channel that generates its substrate in situ. (b) Fluorescence intensity measured across the channels plotted against the width of the channels. The dotted lines represent the approximate center channel boundaries. When compared to Ald’s movement towards buffer, the enzyme shows enhanced migration into the channel that generates its substrate in situ. (c) Experimental set up that allows the entire enzymatic reaction cascade to occur in-situ. Substrate (D-glucose) for enzyme 1, HK (red channel), was provided in the middle channel to trigger the cascade. (d) Ald (green bars) shows time-delayed chemotactic response compared to HK (red bars) as expected based on the sequence of reactions. When mannose was introduced along with D-glucose, HK shows reduced chemotaxis (orange bars) corresponding to slower rate of mannose phosphorylation……………………………………….………………………135

Figure 5-9. (a) While Ald chemotaxes towards its substrate gradient (Figure 5-7b), HK

flowing along with its substrate in its own channel, shows no movement into the adjacent channel. (b) Control experiments performed for studying the chemotactic response of Ald towards its substrate precursors. Ald shows no movement towards the channel flowing the recipe for its substrate when any one of the ingredients is missing……………………136

Figure 5-10. The simulated substrate and product progress curves through the first four

enzymes in the glycolytic cascade, assuming steady state concentrations.........................................................................................139

xv

Figure 5-11. Linear relationships between fluorescence intensity (arbitrary units) and concentration for both HK and Ald. This enables directly correlating fluorescence intensity to the concentration of enzyme………..…………141

Figure 5-12. D-glucose gradient-driven sequential movement of HK and Ald for the

entire enzymatic reaction cascade was observed even in Ficoll PM 70 (20% w/v) induced crowded environment mimicking cytosolic crowding conditions in cell. Ald (red bars) shows a time delayed chemotactic migration towards substrate channel compared to HK (blue bars) corresponding to the cascade reaction sequence……………...…………143

Figure 6-1. Photo- lithographically fabricated flow based microfluidic gradient generator

for studying enzyme chemotaxis. The length of the horizontal channels is 20 mm, width 360 μm and height is 100 μm……………………………….147

Figure 6-2. Shift in fluorescence intensity observed for catalase. The enzyme diffuses

away from the inhibitor (NaCN) and towards the substrate (H2O2) (Note the blue graph’s shift towards left when compared to the control (red))….…149

Figure 6-3. No shift in fluorescence intensity observed for Urease. Pyrochatechol is

unable to completely inhibit urease within the 4.32 s in the microfluidic channel, due to the slow inhibition rate. As a result, no shift is observed……………………………………………………………………….151

Figure 6-4. Diffusion based microfluidic gradient generation device designed using

Adobe illustrator, printed on acrylic surface using a CO2 laser printer and then cast on PDMS using soft lithography………………….…….………..152

Figure 6-5. Normalized fluorescence intensity measured across the substrate and

substrate + inhibitor channels in the microfluidic device. A) The fluorescence intensity within substrate (urea) and the S+I (urea + catechol) channel. The enzyme diffuses much faster and further into the substrate channel compared to the S+I channel. In case of S+I channel most of the enzyme concentration (fluorescence maxima) stays close to the starting position. B) Control experiment performed contained the substrate urea in both reservoirs and the fluorescence intensity indicates similar enzyme diffusion in both channels……………………………………………………154

Figure 6-6. Normalized fluorescence intensity measured across the substrate and

substrate + inhibitor channels in the microfluidic device over 5 hours. The fluorescence intensity within the (a) substrate (urea) and (b) the S+I (urea + catechol) channel. The enzyme diffuses much faster and further into the substrate channel compared to the S+I channel. In case of S+I channel most of the enzyme concentration (fluorescence maxima) stays close to the starting position…………………………………………….…………….155

Figure 6-7. Normalized fluorescence intensity measured in the buffer channel over

time. Only Brownian diffusion is observed…………………….…….……..157

xvi

LIST OF TABLES

Table 4-1. Average pumping speeds caused by TentaGel microspheres exposed to a on and off cycle of 365 nm light. Speeds correspond to the data represented in Figure 4-4a…………………..……………………………...…………...………105

Table 4-2. Average tracer particle speeeds caused by TentaGel microspheres exposed 20 min of continous UV exposure. Speeds correspond to the data represented in Figure 4-4b………………………………..……………………106

Table 5-1. Distance from the start of the channel converted into time spent inside the channel for specified channel geometry.................................................119

Table 5-2. Concentration of enzyme (HK or Ald) migrated into the central channel (containing either buffer only or 10 mM D-glucose + buffer) at specified time periods (see Figure 5-7c). The starting concentration of both enzymes was 200 nM……………………………………………………..……………………..142

xvii

LIST OF MULTIMEDIA FILES

Supporting Video 2-1. PAG pumping using amine functionalized tracer particles with UV on and off at 50X magnification……………………………….…..41

Supporting Video 2-2. PAG pumping using COOH functionalized tracer particles with UV

on and off at 50X magnification……………………………………41

Supporting Video 2-3. PAG pumping resulting in self-assembled patterns using amine functionalized tracer particles with UV on and off at 50X magnification………………………………….……………………. 44

Supporting Video 2-4. PAG pumping causing colloidal transport through micro-channels using amine functionalized tracer particles. Video captured at 5X magnification with the UV turned on-off-on. Video speeded 5 times using Virtualdub software……………………………….…..45

Supporting Video 2-5. PFA-S pumping using COOH functionalized tracer particles at pH 1 at 5X magnification. Video speeded 25 times using Virtualdub software…………………………………………………….………..46

Supporting Video 2-6. PFA-S control using COOH functionalized tracer particles at pH 7 at 5X magnification. Video speeded 25 times using Virtualdub software……………………………………………...……………….46

Supporting Video 2-7. PAG and PFA-S photo-diode’s colloidal transport using sulfate functionalized tracer particles with UV on at 5X magnification; Video captured in between the PAG and PFA-S films, showing both films along with transport direction. Video speeded 5 times using Virtualdub software……………………..……………………52

Supporting Video 4-1. 2 µm-diameter amine-functionalized polystyrene tracer particles showing directional fluid pumping when exposed to UV; moving in towards the bead functionalized with 100% reagent 1. The motion ceases as UV is turned off; 50X magnification………103

Supporting Video 4-2. 2 µm-diameter amine-functionalized polystyrene tracer particles showing only Brownian motion when exposed to UV in the presence of the bead functionalized with 100% reagent 2; 5X magnification……………………………………………………….103

Supporting Video 4-3. Microspheres functionalized with 50% each of reagent 1 and 2 initiate directional fluid pumping upon UV exposure. The pumping continues even when the UV is turned off; 50X magnification…………………………………………………….…104

Supporting Video 4-4. Continuous pumping using microspheres functionalized with 50% each of reagent 1 and 2; 50X magnification. Video captured from all sides of the bead………………………………….……………104

xviii

Supporting Video 4-5. Diffuiophoretic pumping using microsphere functionalized with 100% reagent 1 in aqueous solution. Amine functionalized polystyrene tracer particles with UV on at 50X magnification are seen to move towards and even inside the bead…………..…107

Supporting Video 4-6. Diffuiophoretic pumping using microsphere functionalized with 100% reagent 1 in aqueous solution. Sulfate functionalized polystyrene tracer particles with UV on at 50X magnification are seen to move away from the bead creating exclusion zones……………………………………………………….……….107

xix

ACKNOWLEDGEMENTS

First and foremost, I would like to thank my advisor Professor Ayusman Sen for

his constant guidance and support throughout my Ph.D. His willingness to devote his

time to hear out all my crazy ideas, his valuable suggestions every time I was stuck in a

project have certainly helped me bring my projects to completion. His lucid approach to

routine glitches and witty responses to any problem, big or small, have helped to always

keep the mood light and support a productive work environment. Besides technical

guidance, I have also learnt some efficient life lessons on dealing with both difficult

situations and difficult people in a light and positive manner. I would also like to thank

both him and Mrs. Sen for hosting the wonderful Christmas parties and delightful dinner

conversations.

I would also like to thank Professor Tom Mallouk for his approachable nature, his

inclination to attend to, and solve problems and sharing his curious ideas on enzyme

chemotaxis. Many thanks to Professor Adair for laying the early foundations of colloidal

chemistry during my graduate career and Professor Badding for providing his

constructive feedback on topics outside my area of expertise. A big thank you to my

funding source MRSEC for providing me the resources to have some fun both inside and

outside the lab!

I would like to thank my parents Lata and Suraj Yadav and my friend, husband

and confidante Rahul Thakar for their constant support and encouragement. I truly could

not have made it without them. Their belief in me and their constant motivation have

made me a better professional and a better person. The innumerable sessions of

scientific discussions with Rahul and scientific explanations to my parents have

undoubtedly made me a better speaker and presenter. I would also acknowledge my

xx

adorable little nephews Rudra and Yuvraj, Skype sessions with them have been great

stress busters!

I thank each of my lab mates, coworkers and collaborators for everything I learnt

from them and for keeping everyday interesting, specifically Ryan Pavlick for showing

me how things are done and Samudra Sengupta for being the comic relief, Wentao

Duan for being my problem solver, Hua Zhang for his interpretation of Bollywood, Matt

Baker for showing me there is always another side, Scott Biltek for the frequent

experimental abuses and Weiran Yang for the early Chinese lessons. I would also thank

Xi Zhao for letting me explore my mentoring and sometimes managerial skills,

continuous Chinese, Japanese and Korean lessons but more importantly for the

relentless laugh riots. A big thank you also goes out to the Penn state gym; an

irreplaceable part of my everyday graduate life survival. Outside of my lab, I would also

like to thank all my friends & family- M.L. Thakur, Vipin, Trupti, Mansi, Sriram, Kar,

Trivedy, Manasi, Naomi, Ravish and Gaurika, for keeping the summers active and the

winters warm but mostly for making the last four years memorable.

xxi

Dedicated to Mum; to me she’s perfection..... …..And to Rahul; my anchor, my rock

1

Chapter 1

Difusiophoresis- An Introduction

“Science is not only a disciple of reason but, also, one of romance and passion" - Stephen Hawking

Nano and microscale propulsion is ubiquitous in nature.1-4 Unicellular organisms

like bacteria are not only motile but can also sense food and toxin gradients, then

interact and communicate amongst themselves and respond accordingly.2 The ability to

sense one’s environment, advance towards food and away from toxins, and to

communicate is as vital to a bacterium’s survival as to a blue whale. However, as one

goes down the length scale, the applicable laws of physics change.5 As the radius of an

object scales down, the decrease in volume is greater than the decrease in surface area.

This implies volume dependent forces like inertia, which dominate higher up the scale,

lose relevance as we scale down. Instead, it is the surface forces that need to be

channeled in order to induce motion.6 This thesis focuses on motion at the nano and

micro-scale, the physics that governs this motion and the myriad feasible applications.

1.1 Reynolds Number and Brownian Motion

Reynold’s number (Re) refers to a dimensionless quantity often invoked when

performing scaling of fluid dynamics problems. It is defined as the ratio of the inertial and

viscous forces and helps to characterize fluid patterns under different fluid conditions:

𝑅𝑅𝑅𝑅 = 𝜌𝜌𝜌𝜌𝜌𝜌 ⁄ 𝜂𝜂 (1.1)

2

where ρ is the density of the fluid, V is the mean velocity relative to the fluid, 𝜌𝜌 is the

characteristic linear dimension or the travelled length of the fluid and η is the dynamic

viscosity of the fluid. Laminar flows occur at low Reynolds numbers, where viscous

forces are dominant, and are characterized by smooth, constant fluid motion while

turbulent flows occur at high Reynolds numbers and are dominated by inertial forces.

This principle is often used in designing microfluidic devices. It also helps to determine

which of the two forces - inertial or viscous would dominate. Bacteria and other

unicellular organisms are the finest examples of low Reynolds number swimmers (Re =

10-4). For comparison, an average sized human being has a Reynolds’s number of 104.

Inducing motion at low Reynolds number also requires introducing asymmetry in the

object to evade reciprocal motion, in accordance with the scallop theorem.7

Low Reynolds number represents the first challenge to nano and microscale

motion. However, it is not the only one. We know from classical statistical mechanics

that every molecule moves randomly in all three dimensions with an average kinetic

energy of KT/2; K being the Boltzmann constant and T the absolute temperature.

Therefore, micro-scale objects are subject to the rapid thermal “bumping” by solvent

molecules, and are driven into motion when collisions are uneven. This effect creates

what is known as Brownian diffusion where the objects diffuse and wander around in a

solution resulting in translational Brownian diffusion. Thermal “bumping” also causes an

object to rotate and randomly change orientation, known as rotational Brownian

diffusion. In contrast, active diffusion involves directed motion and requires an input of

energy.

Translational particle diffusion caused due to Brownian motion can be calculated

using Equation 1.2,

3

𝐷𝐷𝑡𝑡 = 𝑘𝑘𝑘𝑘 ⁄ 6𝛱𝛱𝜂𝜂𝛱𝛱 (1.2)

while the rotational particle diffusion is given by Equation 1.3

𝐷𝐷𝑟𝑟 = 𝑘𝑘𝑘𝑘 ⁄ (8𝛱𝛱𝜂𝜂𝛱𝛱3 ) (1.3)

where 𝐷𝐷𝑡𝑡 is the translational diffusion coefficient and 𝐷𝐷𝑟𝑟 is the rotational diffusion

coefficient of the particle, k is the Boltzmann constant, T is the absolute temperature, η is

the viscosity of fluid through which the particle moves and 𝛱𝛱 is the radius of such a

particle.

To examine the nature of particle motion, mean-squared-displacement (MSD)

over different time intervals (τ) is calculated by analyzing the trajectories of particles. For

several idealized types of motions, the MSD has been shown to increase as a function of

τ raised to some power, α. 8

MSD = Kτ α (1.4)

K is a constant whose value depends on the diffusion coefficient of the particle. For

particles undergoing a purely diffusive, two-dimensional random Brownian walk, K

equals four times the diffusion coefficient of the particles, and MSD increases linearly

with τ (i.e., α = 1).9 Since “normal” Brownian diffusion is by far the most commonly

observed motion, systems in which α does not equal 1 are often deemed as having

“anomalous” diffusive behavior. Values greater than 1 correspond to “superdiffusive”

systems, and values less than 1 correspond to “subdiffusive” systems.10, 11, 12 For

example, for the Brownian motion of an inert colloid suspended in a solvent, during a

given time interval τ, the MSD of the particle does indeed go as τ except when that time

interval is very small, e.g., time interval between collisions. At these very small

4

timescales, the particle may appear to be undergoing what is defined as ballistic motion

as it traverses its mean-free-path between solvent collisions.9 For particles that migrate

along a linear trajectory with a constant ballistic velocity, α is simply 2. On the other

hand, labelled messenger RNA molecules in a living E.coli cell undergo “subdiffusion”

with α around 0.7.12

Inducing directed motion at the nano and micron scale requires overcoming

these randomizing events and the following sections discuss mechanisms that have

been employed to accomplish the same.

1.2 Role of Debye Length in Phoretic Transport

One of the first mechanisms identified for autonomous motion was

electrophoresis. In this context, Anderson recognized a critical concept, the slip velocity,

at the solid-liquid interface and the role it plays in fluid dynamics, thus, laying down the

foundation for such phoretic transport mechanism.13 At low Reynold’s number, where

surface forces dominate, it is often processes occurring within this thin interfacial layer

that control the fluid dynamics. In a solution, the charge on a particle's surface is

balanced by a diffuse cloud of counter ions (Figure 1-114). The thickness of the double

layer is defined as the Debye screening length (K-1) and is dependent on the

concentration of ions in the surrounding fluid. The charge density within the cloud at a

distance y, ρe(y), decays exponentially in y at distances of the order of the Debye

screening length from the surface. Taken together, the surface charge and the diffuse

cloud, called the "double layer," are a neutral body. The Debye length plays an important

role in controlling the behavior of colloidal particles and is given by Equation 1.5,

5

𝐾𝐾2 = (2𝑍𝑍2 𝑅𝑅2 𝑐𝑐∞) ⁄ 𝜀𝜀𝑘𝑘𝑘𝑘 (1.5)

where 𝑍𝑍 is the absolute value of the valency of the ion, e is the charge on an electron,

and c is the concentration of the ions 𝜀𝜀 is defined as the dielectric constant of the

material, k is the Boltzmann constant, T is the absolute temperature.

14Figure 1-1. The electric double layer of a charged particle in a polar solution. The

counter ions from the solution come near the charged particle surface to neutralize the

charge and this fluid layer remains diffused around the particle. The ζ-potential is the

electric potential at the shear plane or outer edge of the Stern layer. A non-spherical

charged surface behaves the same way.

6

While low ionic strengths lead to high Debye lengths resulting in colloidal

stability, a high ionic strength solution implies a small Debye length, which leads to short

range van der Waals forces dominating and leads to particle aggregation.

Phoretic transport is defined as the movement of colloidal particles by a field that

interacts with the surface of each particle;13, 14, 15, 16 for instance, electrophoresis involves

an electric field gradient, thermophoresis involves a thermal gradient17-19 and

diffusiophoresis involves a gradient of ionic or non-ionic chemical species.20-22 Other

mechanisms like propulsion based on Marangoni effect,23-26 bubble propulsion,27-30 as

well as propulsion under magnetic31-38 or acoustic fields39-40 have also been identified.

1.3 Mechanisms of Motility

The generation of the propulsive force, asymmetry and, hence, motion can arise

from a variety of mechanisms, including ones that are based on chemical concentration

gradients such as self-electrophoresis and self-diffusiophoresis, and ones that are based

on the gradients of external fields. Motors and pumps are the two major synthetic

machines of interest, and both generate mechanical forces and cause directional

transport by converting energy from chemical fuels,21, 27, 28, 30, 41-49 or external fields

including magnetic,32, 33, 35, 50 electric,51, 52 light,53-56 acoustic,39, 57, 58 and thermal18, 19, 59.

Immobilized “motors” can transfer their force to the surrounding fluid; in effect,

functioning as micropumps. Unlike motors that propel themselves, pumps do not move

themselves, but induce the movement to nearby fluids and inert tracer particles. The

motors require a gradient (e.g. chemical concentration, temperature, surface tension, or

pressure) along the surface to induce motion. They are mostly designed as rods or

spheres with asymmetry in composition (e.g., Janus particles with active material on one

7

side and inert material on the other),27 activity (different chemical reaction rates at the

two ends)60 or shape (concave on one end and convex on the other).38 Early micropump

designs were based on the generation of local electric fields.44, 53, 61-64 Recent designs

include polymeric or enzymatic micropumps that pump fluids by generating chemical

concentration gradients.65-68

1.3.1 Self-Electrophoresis

Electrophoresis is a phenomenon that describes transport of charged species in

a liquid medium (mostly aqueous solution) under an electric field. In an electric field (E),

charged particles migrate with velocity (U) governed by the Smoluchowski equation for

particles with thin double layers.13, 69

U = ζpε

µE (1.6)

Here ζp is the zeta potential of the particle surface, which is related to the surface

charge, ε and µ are the permittivity and dynamic viscosity of the medium, respectively.

Unlike conventional electrophoresis that requires an external electric field, redox

reactions occurring at different parts of a particle surface can result in an ion

concentration gradient and hence local electric field that leads to the motion of the object

itself. This process is called self-electrophoresis, and has been exploited in various

synthetic micro- and nanomachine systems over the past decade.

The first such system was discovered60, 70 using gold (Au)-platinum (Pt) nanorods

(2-3 μm long and ~300 nm in diameter) that move autonomously in dilute hydrogen

peroxide (H2O2) (a few wt.%) with the Pt end leading at a speed of ~ 10 μm/s.

8

In self-electrophoresis, the charged microparticle moves in a self-generated

electric field as a result of an asymmetric distribution of ions. For example, in the case of

the Au-Pt bimetallic nanomotors, the oxidation of H2O2 occurs at the anode (Pt) end and

reduction of H2O2 at the cathode (Au) end lead to a proton concentration gradient

oriented from the Pt end to the Au end. Since the protons are positively charged, the

asymmetric distribution results in an electric field with the same direction (Figure 1-2).60

The negatively charged nanorod therefore moves with the Pt end forward, an effect

similar to traditional electrophoresis.

60Figure 1-2. Propulsion of bimetallic Au-Pt rods in hydrogen peroxide solution powered

by self-electrophoresis.60 Catalytic redox reaction on the two metallic ends generates the

local electric field.

9

The discovery of bimetallic motors has inspired the design of other synthetic

machines, including motors that are based on different shapes,71, 72 fuels41, 43 and power

sources.73

Micropumps that are based on self-electrophoresis have also been designed.

Since a motor moves through fluid, by inverse, immobilizing it will induce fluid flow in its

vicinity. The first examples of micropumps62-64, 74, 75 were developed using the same

principle as that of bimetallic Au-Pt motors mentioned above. With addition of fuel,

electrochemical reactions take place at the surface of the two metals, with the cathode

reducing fuel and consuming protons, and the anode oxidizing fuel and producing

protons (Figure 1-3).63 The redox reaction creates a proton gradient in solution over the

metals, and thus an electric field. The electric field acts both phoretically on charged

tracer particles, and osmotically on the electric double layer of charged metal surface

leading to fluid motion. For tracer particles that are suspended in the solution, only

electrophoretic effect matters, and for ones near to the metal surface, the combination or

competition of the two effects decides their moving direction. Changing the fuel can lead

to change in pumping direction.

10

63Figure 1-3. An immobilized bimetallic surface can generate fluid flow in its vicinity by

the generating a local electric field in the same manner as a bimetallic motor.63 The

schematic describes electrochemical conversion of hydrogen peroxide on the two

metallic surfaces- gold and silver, the generated electric field and the directional motion

imparted to positively charged carboxyl functionalized polystyrene (carboxy-PS) and

negatively charged amidine functionalized polystyrene (amidine-PS) particles.

11

1.3.2 Self-Diffusiophoresis

Similar to self-electrophoresis, self-diffusiophoresis is a mechanism that also

originates from chemical concentration gradients that are produced by surface chemical

reactions. Self-diffusiophoresis can be classified into two categories: electrolyte and non-

electrolyte self-diffusiophoresis, depending on whether the chemical species contributing

to the gradient are charged or uncharged, respectively.

1.3.3 Electrolyte Diffusiophoresis

Electrolyte self-diffusiophoresis is more commonly exploited in the synthetic

motor and pump systems. It operates when a gradient of electrolytes is formed across a

charged surface. For diffusiophoresis near a wall, there are two effects contributing to

the movement of a particle: an electrophoretic effect and a chemophoretic effect, and the

speed of the diffusiophoretic movement can be approximated by Equation (1.7),13

𝑼𝑼 = 𝜵𝜵𝜵𝜵𝜵𝜵𝟎𝟎

��𝑫𝑫+−𝑫𝑫−

𝑫𝑫++𝑫𝑫−� (𝒌𝒌𝑩𝑩𝑻𝑻𝒆𝒆

) 𝜺𝜺(𝜻𝜻𝒑𝒑−𝜻𝜻𝒘𝒘)𝜼𝜼

������������������𝑬𝑬𝑬𝑬𝒆𝒆𝜵𝜵𝑬𝑬𝑬𝑬𝑬𝑬𝒑𝒑𝑬𝑬𝑬𝑬𝑬𝑬𝒆𝒆𝑬𝑬𝑬𝑬𝜵𝜵 𝑻𝑻𝒆𝒆𝑬𝑬𝑻𝑻

+ 𝜵𝜵𝜵𝜵𝜵𝜵𝟎𝟎

�(𝟐𝟐𝜺𝜺𝒌𝒌𝑩𝑩𝟐𝟐 𝑻𝑻𝟐𝟐

𝜼𝜼𝒆𝒆𝟐𝟐 )�𝑬𝑬𝒍𝒍(𝟏𝟏 − 𝜸𝜸𝒘𝒘𝟐𝟐 ) − 𝑬𝑬𝒍𝒍�𝟏𝟏 − 𝜸𝜸𝒑𝒑

𝟐𝟐��� ���������������������������𝑪𝑪𝑬𝑬𝒆𝒆𝑻𝑻𝑬𝑬𝒑𝒑𝑬𝑬𝑬𝑬𝑬𝑬𝒆𝒆𝑬𝑬𝑬𝑬𝜵𝜵 𝑻𝑻𝒆𝒆𝑬𝑬𝑻𝑻

(1.7)

where U is the particle velocity, D+ and D- are the diffusion coefficients of the cation and

anion respectively, Z is the absolute value of the valences of the ions, e is the charge of

an electron, kB is the Boltzmann constant, T is the absolute temperature, ∈ is the

dielectric permittivity of the solution, η is the viscosity of the solution, ζ𝑃𝑃 is the zeta

potential of the particle, ζ𝑤𝑤 is the zeta potential of the wall, γ = tanh(Zeζ𝑃𝑃/4kT), 𝜵𝜵𝜵𝜵 is the

concentration gradient and c0 is the bulk concentration of ions at the particle location, as

12

if the particle was not there. The electroosmotic component, caused due to the wall

double layer, is given by a similar equation, with the particle zeta potential replaced by

the wall zeta potential.

The two parts of the equation signify the two components of diffusiophoresis, as

shown in Figure 1-4.76 The first half signifies electrophoresis. The electric field in this

case originates from the difference in diffusion between the cation and anion which

contributes to the ion gradient in a given direction. This difference leads to a net electric

field, which acts both electrophoretically on the nearby particles and electroosmotically

on the ions adsorbed in the double layer of the wall. The electroosmotic component

leads to fluid movement near the walls. Depending on the charge of the particle, the

electrophoretic and electroosmotic components can augment or allay each other. In

case of competition between the two, the zeta potential of particle or wall dominates and

leads to reduced velocities. However, when both electroosmotic and electrophoretic

motion are in the same direction, an enhancement in particle speed is observed.

Interplay between the osmotic and phoretic components can also lead to schooling and

exclusion patterns.

The second component is the chemophoretic effect. The concentration gradient

of the electrolytes causes a gradient in the thickness of the electric double layer, and

thus a “pressure” difference along the wall is created. As a result, the solution will flow

from the area of higher electrolyte concentration to that of lower concentration, known as

the chemophoretic effect.

13

76Figure 1-4. Schematic depiction of diffusiophoretic motion. The difference in diffusivity

of the ions generated from the source causes a local electric field. The double layer

around the particles as well as the wall responds to the thus formed electric field leading

to electrophoretic and electroosmotic motion respectively. In the example in the figure

above, the anion diffuses faster than the cation generating an electric field from right to

left. The electrophoretic motion of a negatively charged particle is from left to right.

Correspondingly, the electroosmotic flow along the negatively charged wall is from right

to left. The concentration gradient also leads to thickness gradient of double layers on

the surfaces of the particle and wall, and in-turn a pressure difference that propels

particles from left to right.

14

Also, as the thickness of electric double layers is influenced by ionic strength, the

concentration gradient of the electrolytes causes a gradient in the thickness of the

electric double layer. Higher “pressure” at thinner double layers drives fluid flow from the

area of higher electrolyte concentration to that of lower concentration, known as the

chemophoretic effect. In most cases, chemophoretic effect is negligible and

diffusiophoretic transport is governed by the electrophoretic effect, unless the diffusivities

of the cations and the anions are very similar.

The combination of electrophoretic and chemophoretic effects leads to an overall

diffusiophoretic flow, which powers the movement of particles. Electrolyte

diffusiophoresis, however, is not effective in high ionic strength media because of the

collapse of the double layer on the particle surface, as discussed in the previous section.

1.3.4 Non-Electrolyte Self-Diffusiophoresis

Non-electrolyte diffusiophoresis is caused by a gradient of uncharged solutes

and has no dependence on surface charge. This mechanism is able to function in high

ionic strength media, unlike, electrolyte diffusiophoretic transport, which is suppressed

by high electrolyte concentration and, as a result, synthetic machines powered by the

later mechanism cannot operate in highly concentrated ionic media. On the other hand,

in a low ionic strength medium, electrolyte diffusiophoresis is a more powerful

mechanism resulting in higher speeds. This is shown qualitatively by considering that the

chemophoretic component of electrolyte diffusiophoresis has similar origins as non-

electrolyte diffusiophoresis. Both of these mechanisms occur by the chemical species

responsible for the gradient being attracted to the surface either by electrostatic (ionic) or

through van der Waals (non-ionic) interactions. If these two effects are comparable, the

15

electrolyte diffusiophoresis is stronger because it has an additional electric field term

(Equation 1.7).

Although the propulsive forces generated here are generally weaker than from

the electrolyte analog, non-electrolyte diffusiophoresis based on neutral solute gradients

remains effective in powering motion at high ionic strength. Such systems can prevail in

high salt or low polarity solvents. This propulsion mechanism is observed in much fewer

systems and one example is discussed in chapter 4 and is based on a depolymerization

system in organic solvents.

1.3.5 Self-Electrophoresis vs Electrolyte Self-Diffusiophoresis

Self-electrophoresis and electrolyte self-diffusiophoresis are two most commonly

exploited mechanisms for the design of synthetic micro- and nanomachines. Both

mechanisms are based on surface chemical reactions, and the generation of chemical

gradients and local electric fields. The differences between the two mechanisms and the

associated systems can be summarized in three major points. First, electric fields

generated by self-electrophoretic motors are more localized and do not spread as far as

those from electrolyte self-diffusiophoretic systems. As a result, interactions between the

self-electrophoretic motors are short range, and only lead to assembly of doublets or

triplets, while electrolyte self-diffusiophoretic interactions can lead to formation of

collective patterns like “schools”. Secondly, interactions between self-electrophoretic

motors are anisotropic and highly influenced by the relative position or orientation

between the motors. This is significantly different from the case of electrolyte self-

diffusiophoretic motors, which emits and receives chemical signals in an isotropic

16

manner. Lastly, formation of electric fields requires self-electrophoretic systems to be

conductive77, which is not necessary for the electrolyte self-diffusiophoretic counterparts.

In addition to phoretic transport mechanisms, recently discovered biologically

relevant enzymatic motors78 and their collective chemotaxis behavior has brought a new

elixir of life to the field of nanomachines.

1.3.6 Enzyme motors

It has been demonstrated that like other chemically-driven motors, enzymes are

also able to power their own motion by turnover of their respective substrates.78 This is

manifested in the form of substrate-dependent enhancement in diffusivity, as measured

at the single molecule level using fluorescence correlation spectroscopy (FCS). The

observed diffusion enhancement disappears upon the addition of an inhibitor. The

precise mechanism for the turnover-induced enhanced diffusivity remains to be

established. However, a number of mechanistic possibilities have been suggested in the

literature. In one proposal, enzymes propel themselves in solution during substrate

turnover by going through a sequence of non-reciprocal conformational changes during

the substrate binding and product release steps.79 Alternatively, Kapral et al. have

suggested that molecular-scale catalysts can propel themselves through the production

of products that can interact with the catalyst via Lennard-Jones interaction potentials.80

Spatially asymmetric catalysis can lead to inhomogeneous distribution of products. This

non-homogeneous product distribution creates a concentration gradient that can cause

propulsion, depending on features of the products and the solvent (self-diffusiophoresis).

Finally, heat generation through reaction exothermicity may also lead to enhanced

diffusion. However, in several instances the bulk rise in solution temperature due to

17

enzymatic catalysis has been estimated and found to be in the micro-Kelvin range; too

small to account for the observed enhanced diffusion.78, 81 Moreover, to be discussed in

chapter 5 are recent results showing catalysis induced enhancement of diffusion

coefficient for the endothermic turn-over of fructose-bis-phosphate by Aldolase (∆G = +

5.73 Kcal/mol) that argue against the exothermicity hypothesis.

In case of non-motor proteins like urease and catalase, it was determined, using

Langevin/Brownian dynamics simulations, that forces of 12 pN and 9 pN respectively per

turnover were sufficient to cause the enhancement in diffusion. These forces are

comparable to that produced by myosin, kinesin, and dynein motors (about 10 pN)82 and

other molecular scale systems83, 84, and within the range to activate integrins,85 biological

adhesion molecules responsible for mechanosensation by cells, making force production

by enzyme catalysis a potentially novel mechanobiology-relevant event.

1.3.7 Chemotaxis

In the presence of a gradient of substrate concentration, the enzyme molecules

migrate towards higher substrate concentration regions, a form of molecular

chemotaxis;78 another important propulsion mechanism covered in this thesis.

Chemotaxis has long been observed in biological systems,2 and recently in artificial

systems and enzymes in vitro as well.12, 86, 45, 78 The mechanism, however, is not as well

understood in the latter case, unlike the previously discussed phoretic propulsion

mechanisms. In inorganic/synthetic systems, chemotaxis is defined as the preferential

migration in the direction of an externally applied chemical gradient. Hong et al.

proposed that catalytic motors preferentially diffuse up concentration gradients of fuel to

regions with higher diffusivities87, and similar theory has recently been proposed by

18

Saha et al.88 When Pt/Au nanorods are placed in a gradient of hydrogen peroxide, they

gradually diffuse to the source of the chemical, using a combination of active and

stochastic diffusion as demonstrated in Figure 1-5.86 A similar behavior was also

discovered in the polymerization motor system 49, as well as bubble-propelled catalytic

micro-engines.87

86Figure 1-5. Collective behavior demonstrated by synthetic motors. Au-Pt bimetallic

nanomotors chemotax towards the source of hydrogen peroxide fuel (the gel in the

upper left side), as depicted by an increase in the number of rods over time.

19

It has been suggested that the chemotactic behavior of the enzyme molecules

arises from the enhanced diffusion mechanism, since the substrate concentration

changes continuously as the enzyme diffuses along the gradient. Thus, at every point in

space, the diffusion rate increases on moving up the gradient and decreases on moving

down the gradient. A higher diffusion coefficient leads to a greater spreading of the

enzyme molecules on the side of the higher substrate concentration. Thus, the “center of

gravity” of the enzyme ensemble moves towards higher substrate concentration. As with

any non-equilibrium system, a continuous energy input is required for the directional

movement, in this case, to maintain the substrate gradient. The proposed mechanism is

stochastic in nature and is different from biological chemotaxis, which requires temporal

memory of the concentration gradient. The observed chemotactic behavior of single

enzymes suggests that an enzyme that acts on the products of a second, nearby

enzymatic reaction might exhibit collective movement up the substrate gradient towards

this second enzyme; an example of collective behavior at the molecular level.

Chapter 5 discusses such an enzymatic cascade- glycolysis as well as new

insights into the proposed enhanced diffusion controlled mechanism. Several new

control experiments suggest other possible mechanisms or factors such as binding

affinity and turnover rate that play a crucial role in the observed chemotaxis. Other

previously mentioned factors like enzyme conformation, orientation, locally produced

temperature gradients are currently under investigation. These results will shine new

light on the riveting process.

20

1.3.8 Enzyme Pumps

Similar to synthetic pumps, surface-anchored enzymes also transfer their

chemically-generated force to the surrounding fluid; in effect, generating micropumps in

the presence of enzyme-specific substrates.66 Thus, enzymes transduce chemical

energy from substrate turnover into fluid motion. This discovery enables the design of

non-mechanical, self-powered enzyme-based devices that act both as sensor and pump,

precisely controlling flow rate and turning on and off in response to specific analytes.

Most of the enzyme pumps studied so far (glucose oxidase, catalase, lipase, DNA

polymerase) catalyze exothermic reactions and therefore pump fluid and tracer particles

inward along the bottom surface of a microchannel through thermal gradients, as

illustrated in Figure 1-6.

21

66Figure 1-6. Schematic depiction of fabrication and functioning of enzymatic

micropumps. (a) Au patterned on a PEG-coated glass surface is functionalized with a

quaternary ammonium thiol, which electrostatically binds to the negatively charged

groups on the enzyme. Triggered fluid pumping is initiated by introducing enzyme

specific substrate. (b) Cascading fluid pumping is observed when enzyme catalase is

actuated by production of its substrate in situ by enzyme glucose oxidase and its

substrate glucose enabling microfluidic regulation and logic.

22

However, urease (which hydrolyzes urea to bicarbonate and ammonium ions)

increases the solution density and thus pumps fluid outward. These experiments

establish two important findings: 1) essentially all surface-anchored enzymes act as

pumps when turning over their substrates, 2) these pumps are selective for the substrate

or promoter of a particular enzyme.

As with the diffusivity of freely swimming enzymes, the pumping velocity of the

enzyme pumps increases with increasing substrate concentration and reaction rate.

Similar pumping can occur in gel particles in which the enzymes are immobilized. For

example, bound glucose oxidase pumps insulin out of gel particles when glucose is

added to solution.78

1.4 Other Mechanisms

The phoretic mechanisms - electrophoresis and diffusiophoreis, as well as

chemotaxis comprise the focus of this thesis and will be discussed in great details, with

example applications, in the chapters that follow. However, there are other propulsion

mechanisms that have been identified and applied in synthetic systems. The following

sections briefly discuss a few such mechanisms.

1.4.1 Bubble Propulsion

Bubble propulsion is another mechanism, like non-electrolyte diffusiophoresis, that can

power motion at high ionic strengths. In this case, oxygen or hydrogen microbubbles are

generated through decomposition of hydrogen peroxide or reduction of water (Figure 1-

7).89 When bubbles detach from the motors, the associated recoil force pushes motors in

23

the opposite direction. Through surface modification and functionalization, bubble-

propelled motors can sense, capture, and transport biological analytes ranging from

molecules to cells.90 Identification, separation, and isolation of target analytes, such as

specific proteins, nucleic acids, or other biomarkers, are extremely important in

biomedical research. Receptor-modified tubular micro-engines have been demonstrated

to selectively isolate a wide range of target bioanalytes, including bacteria,91 DNA

molecules46 and cancer cells.92

89Figure 1-7. Bubble propulsion mechanism. Oxygen microbubbles are generated

through decomposition of hydrogen peroxide. As the bubbles detach from the motors,

the associated recoil force pushes motors in the opposite direction.

24

For example, with the outer surface functionalized with Concanavalin A (Con A)

lectin receptor, catalytic micro-engines can recognize carbohydrate constituents of

bacterial surface, and selectively bind to them.91 As a proof of concept, E.coli was

demonstrated to be isolated from untreated seawater and drinking water samples.

Tubular catalytic microengines can also function as concentrating systems,93 and

achieve directional transport and delivery of cells with the help of external magnetic

fields.94

By coating the surface with polymeric layers, it is also possible to achieve

controlled-drug release via bubble-propelled motors. Mg/Pt Janus motors, when coated

with thermo-responsive poly(N-isopropylacrylamide) (PNIPAM) hydrogel layer, have

been reported to release drug molecules in response to a temperature change.95

Despite these potential applications, in-vivo application of bubble-propelled

motors are hindered by the fact that their motility is attenuated by electrolytes and blood

plasma.96, 97

1.4.2 Magnetically-driven Motors

A problem for bubble-propelled motor is their general lack of directionality, due to

Brownian randomization at longer time scales. One way to overcome this problem is to

introduce magnetic components into motors. Such motors, although still powered by

chemical fuels, are subject to guidance by external magnetic fields.

Another method is to simply replace the power source with external magnetic

fields. These motors, when actuated, can be employed both in vitro and in vivo .98 Using

this technique, Nelson group has reported several examples of cell transportation and

drug delivery by artificial flagella.99-101 For cell transportation, cage-like micromotors

25

(Figure 1-8)100 were fabricated and cells were allowed to grow inside them. These

motors were subsequently activated and propelled using an external rotating magnetic

field. For drug delivery, motor surfaces were modified with drug-loaded chitosan or

liposomes 99, 101; motors then migrated towards targets and released drugs.

Wireless manipulation of micromotors inside eye cavity through OctoMag

electromagnetic control system has also been reported. 102, 103 The OctoMag can control

motors of a human eye. Micromotors are injected into eyes through a 23G-needle

syringe and, once inside, are powered and manipulated by the magnetic fields.

26

100Figure 1-8. Magnetic manipulation of cage-like micromotors for transportation of cells.

(a) SEM image of a hexahedral microrobot after cell culture and (b) an enlarged SEM

image. Confocal microscope images of the (c) hexahedral and (d) cylindrical microrobots

after staining of the cells.

27

1.4.3 Acoustically-powered Motors

Low-power acoustic waves are safe and used extensively for in vivo imaging,

and are thus useful for powering motors. In an acoustic field, suspended microparticles

experience acoustic radiation forces, which are strongest when standing waves are

formed under acoustic excitation.

Recently Wang et al. reported a MHz-frequency ultrasound-powered

autonomous micromotor system.6 In the system, bimetallic microrods are suspended in

water, and levitated to a plane at the midpoint of the cell by a vertical standing wave, as

demonstrated in Figure 1-9a.38 In the plane, the rods exhibit axial propulsion at speeds

up to 200 µm/s (~100 body lengths/s), and also form patterns in the nodal plane. Motion

of the motors are significantly affected by their composition, as only metallic rods

showing fast axial motion, and polymeric rods do not.

Self-acoustophoresis has been proposed as the mechanism of motility. The

acoustic motors, under guidance of magnetic fields, can be steered to capture and

transport various bioanalytes like cells, as demonstrated in Figure 1-9b.38 The motion of

acoustic motors inside living HeLa cells, has also been reported, the first example of

artificial motors inside living cells.57 The motors attach strongly to the external surface of

the cells, and are readily internalized through incubation for periods longer than 24 h.

Actuated at 4 MHz, these motors exhibit axial propulsion and spinning while the cells

remaining viable. Such systems can provide a new tool for probing the response of living

cells to internal mechanical excitation and related biomedical applications.

28

Figure 1-9. Acoustic powered self-propelled motors.38 (a) Propagation and assembly of

bimetallic rods under acoustic fields. (b) Navigation of an acoustically-powered motor

towards a HeLa cell under magnetic field-guidance.

29

1.5 Conclusion

This chapter gives a general introduction to the theme of nanoscale propulsion.

However, the focus of this thesis continues to be diffusiophoretic mechanisms. The

following chapters discuss diffusiphoretic mechanism utilized to design application

oriented triggered, self-propelled micro-pumps, followed by new insights into enzymatic

chemotaxis.

30

1.6 References

1. Feynman, R. P. Engineering and Science 1960, 23, 22-36. 2. Berg, H. C.; Brown, D. A. Nature 1972, 239, 500-504. 3. Block, S. M. Cell 1998, 93, 5-8. 4. Bonabeau, E.; Dorigo, M.; Theraulaz, G., Swarm intelligence: from natural to

artificial systems. Oxford University Press, Inc.: 1999; p 307. 5. Purcell, E. M. Am. J. Phys. 1977, 45, 3-11. 6. Wang, W.; Duan, W.; Ahmed, S.; Mallouk, T. E.; Sen, A. Nano Today 2013, 8,

531-554. 7. Lauga, E. Phys. Rev. Lett. 2011, 106, 178101. 8. Metzler, R.; Klafter, J. Phys. Rep. 2000, 339, 1-77. 9. Pusey, P. N. Science 2011, 332, 802-803. 10. Duan, W.; Ibele, M.; Liu, R.; Sen, A. Eur. Phys. J. E 2012, 35, 1-8. 11. Dunderdale, G.; Ebbens, S.; Fairclough, P.; Howse, J. Langmuir 2012, 28, 10997-

11006. 12. Metzler, R.; Jeon, J.-H.; Cherstvy, A. G.; Barkai, E. Phys. Chem. Chem. Phys.

2014, 16, 24128-24164. 13. Anderson, J. L. Annu. Rev. Fluid Mech. 1989, 21, 61-99. 14. Adair, J. H.; Suvaci, E.; Sindel, J. Encyclopedia of Materials: Science and

Technology, Elsevier Publishing, The Netherlands, 2001, 8996-9006. 15. Ruckenstein, E. J. Colloid Interface Sci. 1981, 83, 77-81. 16. Golestanian, R.; Liverpool, T.; Ajdari, A. New J. Phys. 2007, 9, 126. 17. Rasuli, S. N.; Ramin, G. J. Phys.: Condens. Matter 2005, 17, S1171. 18. Jiang, H.-R.; Yoshinaga, N.; Sano, M. Phys. Rev. Lett. 2010, 105, 268302. 19. Baraban, L.; Streubel, R.; Makarov, D.; Han, L.; Karnaushenko, D.; Schmidt, O.

G.; Cuniberti, G. ACS Nano 2013, 7, 1360-1367. 20. Córdova-Figueroa, U. M.; Brady, J. F. Phys. Rev. Lett. 2008, 100, 158303.

31

21. Golestanian, R.; Liverpool, T. B.; Ajdari, A. Phys. Rev. Lett. 2005, 94, 220801. 22. Ebbens, S.; Tu, M.-H.; Howse, J. R.; Golestanian, R. Phys. Rev. E 2012, 85,

020401. 23. Okawa, D.; Pastine, S. J.; Zettl, A.; Fréchet, J. M. J. J. Am. Chem. Soc. 2009,

131, 5396-5398. 24. Lauga, E.; Davis, A. M. J. J. Fluid Mech. 2012, 705, 120-133. 25. Sharma, R.; Chang, S. T.; Velev, O. D. Langmuir 2012, 28, 10128-10135. 26. Zhang, H.; Duan, W.; Liu, L.; Sen, A. J. Am. Chem. Soc. 2013, 135, 15734-15737. 27. Gibbs, J. G.; Zhao, Y.-P. Appl. Phys. Lett. 2009, 94, 163104. 28. Solovev, A. A.; Mei, Y.; Bermúdez Ureña, E.; Huang, G.; Schmidt, O. G. Small

2009, 5, 1688-1692. 29. Sanchez, S.; Solovev, A. A.; Mei, Y.; Schmidt, O. G. J. Am. Chem. Soc. 2010,

132, 13144-13145. 30. Gao, W.; Sattayasamitsathit, S.; Orozco, J.; Wang, J. J. Am. Chem. Soc. 2011,

133, 11862-11864. 31. Tierno, P.; Golestanian, R.; Pagonabarraga, I.; Sagués, F. J. Phys. Chem. B

2008, 112, 16525-16528. 32. Ghosh, A.; Fischer, P. Nano Lett. 2009, 9, 2243-2245. 33. Gao, W.; Sattayasamitsathit, S.; Manesh, K. M.; Weihs, D.; Wang, J. J. Am.

Chem. Soc. 2010, 132, 14403-14405. 34. Fischer, P.; Ghosh, A. Nanoscale 2011, 3, 557-563. 35. Tottori, S.; Zhang, L.; Qiu, F.; Krawczyk, K. K.; Franco-Obregón, A.; Nelson, B. J.

Adv. Mater. 2012, 24, 811-816. 36. Masoud, H.; Alexeev, A. Soft Matter 2010, 6, 794-799. 37. Keaveny, E. E.; Walker, S. W.; Shelley, M. J. Nano Lett. 2013, 13, 531-537. 38. Wang, W.; Castro, L. A.; Hoyos, M.; Mallouk, T. E. ACS Nano 2012, 6, 6122-

6132. 39. Ahmed, S.; Wang, W.; Mair, L. O.; Fraleigh, R. D.; Li, S.; Castro, L. A.; Hoyos, M.;

Huang, T. J.; Mallouk, T. E. Langmuir 2013, 29, 16113-16118.

32

40. Garcia-Gradilla, V.; Orozco, J.; Sattayasamitsathit, S.; Soto, F.; Kuralay, F.; Pourazary, A.; Katzenberg, A.; Gao, W.; Shen, Y.; Wang, J. ACS Nano 2013, 7, 9232-9240.

41. Liu, R.; Sen, A. J. Am. Chem. Soc. 2011, 133, 20064-20067. 42. Pantarotto, D.; Browne, W. R.; Feringa, B. L. Chem. Commun. 2008, 1533-1535. 43. Mano, N.; Heller, A. J. Am. Chem. Soc. 2005, 127, 11574-11575. 44. Kline, T. R.; Paxton, W. F.; Mallouk, T. E.; Sen, A. Angew. Chem. 2005, 117, 754-

756. 45. Paxton, W. F.; Baker, P. T.; Kline, T. R.; Wang, Y.; Mallouk, T. E.; Sen, A. J. Am.

Chem. Soc. 2006, 128, 14881-14888. 46. Wu, J.; Balasubramanian, S.; Kagan, D.; Manesh, K. M.; Campuzano, S.; Wang,

J. Nat. Commun. 2010, 1, 36. 47. Howse, J. R.; Jones, R. A. L.; Ryan, A. J.; Gough, T.; Vafabakhsh, R.;

Golestanian, R. Phys. Rev. Lett. 2007, 99, 048102. 48. Stock, C.; Heureux, N.; Browne, W. R.; Feringa, B. L. Chem. Eur. J. 2008, 14,

3146-3153. 49. Pavlick, R. A.; Sengupta, S.; McFadden, T.; Zhang, H.; Sen, A. Angew. Chem.

2011, 123, 9546-9549. 50. Dreyfus, R.; Baudry, J.; Roper, M. L.; Fermigier, M.; Stone, H. A.; Bibette, J.

Nature 2005, 437, 862-865. 51. Calvo-Marzal, P.; Sattayasamitsathit, S.; Balasubramanian, S.; Windmiller, J. R.;

Dao, C.; Wang, J. Chem. Commun. 2010, 46, 1623-1624. 52. Chang, S. T.; Paunov, V. N.; Petsev, D. N.; Velev, O. D. Nat. Mater 2007, 6, 235-

240. 53. Hong, Y.; Diaz, M.; Córdova-Figueroa, U. M.; Sen, A. Adv. Funct. Mater. 2010, 20,

1568-1576. 54. Liu, M.; Zentgraf, T.; Liu, Y.; Bartal, G.; Zhang, X. Nat. Nanotechnol. 2010, 5, 570-

573. 55. Ibele, M.; Mallouk, T. E.; Sen, A. Angew. Chem., Int. Ed. 2009, 48, 3308-3312. 56. Abid, J.-P.; Frigoli, M.; Pansu, R.; Szeftel, J.; Zyss, J.; Larpent, C.; Brasselet, S.

Langmuir 2011, 27, 7967-7971.

33

57. Wang, W.; Li, S.; Mair, L.; Ahmed, S.; Huang, T. J.; Mallouk, T. E. Angew. Chem., Int. Ed. 2014, 53, 3201-3204.

58. Kagan, D.; Benchimol, M. J.; Claussen, J. C.; Chuluun-Erdene, E.; Esener, S.;

Wang, J. Angew. Chem. 2012, 124, 7637-7640. 59. Qian, B.; Montiel, D.; Bregulla, A.; Cichos, F.; Yang, H. Chem. Sci. 2013, 4, 1420-

1429. 60. Paxton, W. F.; Kistler, K. C.; Olmeda, C. C.; Sen, A.; St. Angelo, S. K.; Cao, Y.;

Mallouk, T. E.; Lammert, P. E.; Crespi, V. H. J. Am. Chem. Soc. 2004, 126, 13424-13431.

61. McDermott, J. J.; Kar, A.; Daher, M.; Klara, S.; Wang, G.; Sen, A.; Velegol, D.

Langmuir 2012, 28, 15491-15497. 62. Jun, I.-K.; Hess, H. Adv. Mater. 2010, 22, 4823-4825. 63. Kline, T. R.; Paxton, W. F.; Wang, Y.; Velegol, D.; Mallouk, T. E.; Sen, A. J. Am.

Chem. Soc. 2005, 127, 17150-17151. 64. Ibele, M. E.; Wang, Y.; Kline, T. R.; Mallouk, T. E.; Sen, A. J. Am. Chem. Soc.

2007, 129, 7762-7763. 65. Sengupta, S.; Spiering, M. M.; Dey, K. K.; Duan, W.; Patra, D.; Butler, P. J.;

Astumian, R. D.; Benkovic, S. J.; Sen, A. ACS Nano 2014, 8, 2410-2418. 66. Sengupta, S.; Patra, D.; Ortiz-Rivera, I.; Agrawal, A.; Shklyaev, S.; Dey, K. K.;

Córdova-Figueroa, U.; Mallouk, T. E.; Sen, A. Nature Chem. 2014, 6, 415-422. 67. Zhang, H.; Yeung, K.; Robbins, J. S.; Pavlick, R. A.; Wu, M.; Liu, R.; Sen, A.;

Phillips, S. T. Angew. Chem., Int. Ed. 2012, 51, 2400-2404. 68. Yadav, V.; Zhang, H.; Pavlick, R.; Sen, A. J. Am. Chem. Soc. 2012, 134, 15688-

15691. 69. Solomentsev, Y.; Anderson, J. L. J. Fluid Mech. 1994, 279, 197-215. 70. Fournier-Bidoz, S.; Arsenault, A. C.; Manners, I.; Ozin, G. A. Chem. Commun.

2005, 441-443. 71. Gibbs, J. G.; Fragnito, N. A.; Zhao, Y. Appl. Phys. Lett. 2010, 97, 253107. 72. Wheat, P. M.; Marine, N. A.; Moran, J. L.; Posner, J. D. Langmuir 2010, 26,

13052-13055. 73. Loget, G.; Kuhn, A. J. Am. Chem. Soc. 2010, 132, 15918-15919.

34

74. Kline, T. R.; Iwata, J.; Lammert, P. E.; Mallouk, T. E.; Sen, A.; Velegol, D. J. Phys. Chem. B 2006, 110, 24513-24521.

75. Kline, T. R.; Sen, A. Langmuir 2006, 22, 7124-7127. 76. Yadav, V.; Duan, W.; Sen, A., Diffusiophoretic Nano and Microscale Propulsion

and Communication. In Engineering of Chemical Complexity II, pp 73-91. 77. Laocharoensuk, R.; Burdick, J.; Wang, J. ACS Nano 2008, 2, 1069-1075. 78. Sengupta, S.; Dey, K. K.; Muddana, H. S.; Tabouillot, T.; Ibele, M. E.; Butler, P. J.;

Sen, A. J. Am. Chem. Soc. 2013, 135, 1406-1414. 79. Sakaue, T.; Kapral, R.; Mikhailov, A. S. Eur. Phys. J. B 2010, 75, 381-387. 80. Peter, H. C.; Raymond, K. Europhys. Lett. 2014, 106, 30004. 81. Shah, A. S.; Ben-Shahar, Y.; Moninger, T. O.; Kline, J. N.; Welsh, M. J. Science

2009, 325, 1131-1134. 82. Spudich, J. A.; Rice, S. E.; Rock, R. S.; Purcell, T. J.; Warrick, H. M. Cold Spring

Harb Protoc. 2011, 2011, 066662. 83. Mahadevan, L.; Matsudaira, P. Science 2000, 288, 95-99. 84. Mehta, A. D.; Rief, M.; Spudich, J. A.; Smith, D. A.; Simmons, R. M. Science

1999, 283, 1689-1695. 85. Wang, N.; Butler, J.; Ingber, D. Science 1993, 260, 1124-1127. 86. Hong, Y.; Blackman, N. M. K.; Kopp, N. D.; Sen, A.; Velegol, D. Phys. Rev. Lett.

2007, 99, 178103. 87. Baraban, L.; Harazim, S. M.; Sanchez, S.; Schmidt, O. G. Angew. Chem. 2013,

125, 5662-5666. 88. Saha, S.; Golestanian, R.; Ramaswamy, S. Phys. Rev. E 2014, 89, 062316. 89. Abdelmohsen, L. K.; Peng, F.; Tu, Y.; Wilson, D. A. J. Mater. Chem. B 2014, 2,

2395-2408. 90. Solovev, A. A.; Sanchez, S.; Pumera, M.; Mei, Y. F.; Schmidt, O. G. Adv. Funct.

Mater. 2010, 20, 2430-2435. 91. Campuzano, S.; Orozco, J.; Kagan, D.; Guix, M.; Gao, W.; Sattayasamitsathit, S.;

Claussen, J. C.; Merkoçi, A.; Wang, J. Nano Lett. 2012, 12, 396-401.

35

92. Balasubramanian, S.; Kagan, D.; Jack Hu, C.-M.; Campuzano, S.; Lobo-Castañon, M. J.; Lim, N.; Kang, D. Y.; Zimmerman, M.; Zhang, L.; Wang, J. Angew. Chem., Int. Ed. 2011, 50, 4161-4164.

93. Restrepo-Perez, L.; Soler, L.; Martinez-Cisneros, C.; Sanchez, S.; Schmidt, O. G.

Lab Chip 2014, 14, 2914-2917. 94. Sanchez, S.; Solovev, A. A.; Schulze, S.; Schmidt, O. G. Chem. Commun. 2011,

47, 698-700. 95. Mou, F.; Chen, C.; Zhong, Q.; Yin, Y.; Ma, H.; Guan, J. ACS Appl. Mater.

Interfaces 2014, 6, 9897-9903. 96. Wang, H.; Zhao, G.; Pumera, M. Phys. Chem. Chem. Phys. 2013, 15, 17277-

17280. 97. Zhao, G.; Wang, H.; Khezri, B.; Webster, R. D.; Pumera, M. Lab Chip 2013, 13,

2937-2941. 98. Peyer, K. E.; Zhang, L.; Nelson, B. J. Nanoscale 2013, 5, 1259-1272. 99. Fusco, S.; Chatzipirpiridis, G.; Sivaraman, K. M.; Ergeneman, O.; Nelson, B. J.;

Pané, S. Adv. Healthc Mater. 2013, 2, 1037-1044. 100. Kim, S.; Qiu, F.; Kim, S.; Ghanbari, A.; Moon, C.; Zhang, L.; Nelson, B. J.; Choi,

H. Adv. Mater. 2013, 25, 5863-5868. 101. Mhanna, R.; Qiu, F.; Zhang, L.; Ding, Y.; Sugihara, K.; Zenobi-Wong, M.; Nelson,

B. J. Small 2014, 10, 1953-1957. 102. Marino, H.; Bergeles, C.; Nelson, B. J. IEEE Trans. Autom. Sci. Eng. 2014, 11,

310-316. 103. Kummer, M. P.; Abbott, J. J.; Kratochvil, B. E.; Borer, R.; Sengul, A.; Nelson, B. J.

IEEE Trans. Robot. 2010, 26, 1006-1017.

36

Chapter 2

Triggered “On/Off” Micro-Pumps and Colloidal Photo-Diode

2.1. Introduction

An important challenge in designing nano/micromotors and pumps involves

achieving targeted transport to a precise destination.1-8 Such fine-tuned motion is

essential for complex functions in microfluidic chips, cargo delivery systems, and self-

assembly applications. Further, it is also highly desirable that these pumps be capable of

being turned on by a specific external signal. This chapter discusses a set of

autonomous micropumps based on simple acid-base and photochemistry induced ion

gradients that result in spatio-temporal control of fluid flow. Earlier designs of

micropumps that cause fluid flow in response to specific fuels or chemical signals 9-15

typically lack control since they cannot be readily turned off and on again.16 Having an

on/off switch is important for it allows the pump to respond to changes in the

environment, which is useful for the design of sensors and logic gates.

2.2 Design of Smart Micro-Pumps

This chapter describes a photo-acid generator (PAG) based system that was

used to demonstrate a UV-initiated pump with an “on/off” switch that can further be used

for patterning. In addition the acid-catalyzed hydrolysis of a polymeric imine17-20 was

utilized to create a pump with pH regulated velocities that was combined with the PAG

37

pump to create a source-drain based “photo-diode” that gives spatio-temporal control

over colloidal mobility.

2.3 Propulsion Mechanism

As shown in Figure 2-1, upon illumination at wavelength 365 nm,, the solid

photoacid generator, N-hydroxyphthalimide triflate (PAG-1), forms N-

hydroxyphthalimide, and two ions: proton and triflate anion. The large difference in

diffusion coefficients between the small cation and the large anion, establishes a

diffusion-induced electric field pointing towards the PAG that is responsible for the

observed diffusiophoretic motion.

As discussed in the previous chapter, in an unbounded solution of a

symmetrically charged binary electrolyte with a uniform concentration gradient ∇𝑐𝑐, the

diffusiophoretic velocity of a charged particle, U, is given by Equation 2.121

U = 𝜀𝜀kT2𝑒𝑒𝑒𝑒

��𝐷𝐷+−𝐷𝐷−

𝐷𝐷++𝐷𝐷−� ζ𝑃𝑃 − 2kT𝑍𝑍𝑒𝑒

𝜌𝜌𝑙𝑙(1 − 𝛾𝛾2) ∇𝑐𝑐c0

� (2.1)

where D+ and D- are the diffusion coefficients of the cation and anion

respectively, Z is the absolute value of the valences of the ions, e is the charge of

an electron, k is the Boltzmann constant, T is the absolute temperature, ε is the

dielectric permittivity of the solution, η is the viscosity of the solution, ζ𝑃𝑃 is the

zeta potential of the particle, γ = tanh(Zeζ𝑃𝑃/4kT), and c0 is the bulk concentration

of ions at the particle location, as if the particle was not there. The electroosmotic

component is given by a similar equation, with the particle zeta potential replaced

by the wall zeta potential.

38

The net velocities in this system result from the competition between the

diffusiophoresis and the electroosmosis (Figure 2-1). With the proton diffusing faster (D

= 9.31 x 10-5 cm2 s-1) than the larger triflate anion (estimated D = 1.38 x 10-5cm2 s-1

assuming a sphere), a local electric field is set up pointing inwards. Owing to the electric

double layer on the negatively-charged sodium borosilicate glass slide, an

electroosmotic flow is generated, which is also inwards in the direction of the local

electric field.

Figure 2-1. A schematic depiction of PAG pumping mechanism. The negative surface

charge of the glass creates a positive double layer, which in response to the generated

ions causes an inward electroosmotic flow. The negatively charged tracers (S-PS

particles) move opposite to the direction of the electric field, competing against the

electroosmotic flow while the positively charged tracers (NH2-PS particles) move along

the electric field direction aided by the electroosmotic flow.

UV

N

O

O

O S

O

CF3

O

hvN

O

O

OH

CF3SO3-

+

+ H+

39

2.4 Switchable Photoacid Pump

2.4.1 Experimental Set-Up

As displayed in Figure 2-1, to construct the PAG pump, crystallites of PAG-1

were placed on a glass slide and covered with hybridization chambers. An HBO (H for

Hg or mercury, B is the symbol for luminance, O for unforced cooling) 100 lamp attached

to the microscope was used for UV exposure. The light wavelength was predominantly

365 nm with a maximum power of 2.5 Wcm-2 at the center. In a typical experiment a 9

mm diameter, 0.12 mm thick imaging hybridization chamber was placed on a

microscope slide and the appropriate solution, typically a water suspension of sulfate,

carboxylate, or amine functionalized polystyrene tracers (2 µm diameter) was added to

the chamber. The deionized (DI) water that was used for all experiments had a specific

resistance of 18 MΩ cm. The videos were captured using a CCD camera attached to an

optical microscope (Zeiss Axiovert 200 reflectance/ transmission) at 50x magnification.

To calculate particle velocities, 30 randomly selected particles were tracked using

tracker software for 6 s.

2.4.2 ‘On/Off’ Pump in Action

Positively charged tracers (amino functionalized polystyrene particles, NH2-PS)

were observed to move towards the photoacid, aided by both the diffusiophoretic and

electroosmotic flows, attaining an average velocity of 7.2 ± 2.4 µm/s. The negatively

charged tracers (sulfate functionalized polystyrene particles; S-PS and carboxylate

functionalized polystyrene particles; HOOC-PS) on the other hand, move in the direction

40

of the diffusiophoretic flows, out-winning the electroosmotic flow, owing to their higher

zeta potential in comparison to that of the glass.22, 25 Also, in this system, the pH

changes from 6.8 to 2 and the ionic strength changes from 10-6 to 10-3 M. In this range,

the zeta potential of glass and involved tracers are: Glass, -30 to -60 mV22; NH2-PS,

approx. +55 mV23; S-PS, -100 to -180 mV24; HOOC-PS, -120 to -160 mV25, the exact

value depending on the ionic strength and pH. The S-PS particles attained an average

velocity of 4.8 ± 1.3 µm/s and that for HOOC-PS particles was calculated to be 4.1 ± 0.9

µm/s. Note that all the reported velocities were measured close to the wall (glass

surface). As expected, due to fluid continuity, the direction of motion is reversed when

observing particles several hundred microns above the wall. The motion ceases when

UV light is turned off but can be reinitiated upon re-illumination. Only Brownian diffusion

is observed in the absence of UV light. Another control was performed by first

exhausting the photoacid by prolonged UV illumination, until it no longer produces

protons (no measurable change in pH). Here again, no powered motion was observed

even under illumination. Hence, the role of a thermal gradient in causing flow can be

ruled out. Figure 2-2 displays the tracer particle distributions without UV and after 1 min

of UV irradiation.

41

Figure 2-2. Optical microscope images of particle motion. (a) and (b) show the

distribution of the positively charged tracers (NH2-PS) around the photoacid (PAG-1)

microcrystallites with UV off (control) and after 1 min of UV illumination respectively. (c)

and (d) display the same for the negatively charged tracers (S-PS). Each of the tracer

particles seen is 2 µm. (Also see Supporting Video 2-1 and 2-2)

42

2.4.3 Separation of Diffusiophoretic and Electroosmotic Motion

In the interest of separating the diffusiophoretic and electroosmotic components,

the same experiments were performed on a polystyrene surface, which has minimal

surface charge. The particles (2 µm) were tracked over a 6 s period using tracker

software. The velocities of the positively charged tracers were impeded in the absence

of the aiding electroosmotic force (average velocity = 4.0 ± 0.4 µm/s), while those of the

negatively charged tracers were enhanced due to the absence of the opposing

electroosmotic force (S-PS particle, average velocity = 7.9 ± 1.1 µm/s and HOOC-PS

particle, average velocity = 6.6 ± 0.3 µm/s), both approximately by a factor of two. Thus,

the estimated contributions of diffusiophoretic and electroosmotic components of velocity

are approximately equal (Figure 2-3).

43

Figure 2-3. Velocity distribution histograms obtained for (a) NH2-PS particles and (b) S-

PS particles using the PAG pump.

44

2.4.4 Self- Assembled Patterns

As shown in Figure 2-4, diffusiophoretic pumping by PAG upon illumination

resulted in patterned self-assemblies of tracer particles (2 µm) (Supporting Video 2-3).

This outcome was utilized to design the photoacid pump induced guided motion through

microchannels to achieve patterning at the micron scale using photo-induced ion

gradients. The lithographic mask was printed on a polyacrylate surface using epilog mini

laser printer and a PDMS layer was cast on it (Figure 2-5). The PDMS layer was

subsequently peeled off after curing and was used as our pattern. The reservoirs on

each end of the pattern were 10 x 10 mm. PAG-2 crystallites were packed in two 1 x 1

mm compartments on either side of a 4 x 1 mm channel connecting the two reservoirs.

Figure 2-4. Patterns induced by PAG pumping. (a) Control image, NH2-PS particle

distribution around a single photoacid crystallite with UV off. (b) Self-assembled NH2-PS

particle pattern with UV on. Each of the tracer particles seen is 2 µm.

Control Self-assembly (a) (b)

45

N-Hydroxy-5-norbornene-2,3-dicarboximide perfluoro-1-butanesulfonate, PAG-2

was used for this experiment due to its granular morphology and its ease of packing in

the 1 mm x 1 mm compartment within the pattern. Upon UV illumination, tracers (NH2-

PS) were pulled out of the large reservoirs and were made to follow the designed micro-

channel in the pattern (Supporting Video 2-4). When the UV was switched off, the

particle motion relaxed and picked up again as the UV was turned back on. Thus, the

motion can be initiated and stopped repeatedly.

Figure 2-5. A schematic depiction of the pattern. The pump pulls the NH2-PS particles

out from the large reservoirs (10 x 10 mm2) into the micro-channels (4 x 1 mm2), towards

the PAG chambers (1 x 1 mm2) on either side of the channel.

46

2.5 pH Controlled Polymer Pump

Figure 2-6. Schematic depiction of PFA-S pumping mechanism. The local electric field

points outwards away from the polymer film and the negatively charged tracers COOH-

PS particles move inwards, towards film.

The acid-catalyzed hydrolysis of a polymeric imine, poly(4-formyphenyl acrylate)

aniline Schiff base (PFA-S) film cast on a glass slide was next used to design another

diffusiophoretic pump, only this time with the direction of the local electric field reversed

due to the higher diffusivity of the anion (DCl- = 2.032 x 10-5 cm2 s-1) relative to the much

larger cation (Figure 2-6). (Supporting Videos 2-5 and 2-6). Accordingly, the

movement of the tracers was observed to be reversed with the negative HOOC-PS

47

tracers moving inwards competing against the electroosmotic component and the

positively charged NH2-PS particles moving outwards, aided by the electroosmotic

component. The experimental design of the PFA-S polymer pump involved dissolving

the polymer in dichloromethane (20 mg/mL) and drop casting the solution onto a glass

slide. The polymer film was vacuum dried overnight and then carefully cut into

approximately 0.3 x 0.3 mm and excess polymer was scraped off. The PFA-S film was

covered with hybridization chamber and tracers (6 µm diameter) in HCl (1M, 0.1M and

0.01M) or phosphate buffer solution (pH = 7, control) were added in the chamber. The

ionic strength of all solutions was adjusted to 1M by adding extra NaCl if necessary. The

optical imaging was performed at 5X magnification.

Figure 2-7. Optical microscope images of PFA-S film pumping away HOOC-PS tracers

(6 µm). (a) Image taken 0 s after exposure to 1 M HCl in deionized water at 25 °C, and

(b) 1200 s after exposure.

(a) (b)

48

Figure 2-7 shows optical images of the pumping by PFA-S film. In accordance

with equation (1), the velocities attained by the tracer particles were noted to be

dependent on the concentration gradient of the formed electrolytes. Accordingly, varying

concentrations of HCl were introduced into the PFA-S system and the HOOC-PS particle

velocities measured for each concentration. In each experiment, 20 randomly selected

particles were tracked using Tracker software for 6 consecutive time-steps of 5 s (120

data points for each distribution plot). As expected, the particle velocities showed an

increase with decreasing pH reaching an average velocity of 3.2 ± 0.8 µm/s in 1 M HCl.

Figure 2-8. Velocity distribution histograms of HOOC-PS tracers as a function of the

acid concentration for the PFA-S pump.

(µm/s)

49

The particle velocities were measured at distances from 100 µm to 200 µm away

from the pump, 0 to 50 s after exposure to acid. The velocity distribution histograms at

each proton concentration, over the specified range of distance and time, are shown in

Figure 2-8. The ionic strength at each acid concentration (1M, 0.1M, 0.01M HCl) was

kept the same at 1M, to avoid any changes to the electric double layer.

Figure 2-9. Velocity distribution histograms of HOOC-PS tracers at 100 to 1100 µm

away from the PFA-S pump upon addition of 1 M HCl to the PFA-S film at 25°C,

demonstrating long range pumping.

50

Further, as expected of a diffusiophoretic pump, the pumping velocities also

showed a dependence on the distance from the pump source. Average particle velocity

was noted to change as a function of distance for the PFA-S pump over the time period

of 0 s to 5 s after exposure to 1M HCl at 25 °C (Figure 2-9). The PFA-S pump is capable

of providing minimum average pumping velocities of 2.0 ± 0.6 µm/s at distances over 1

mm from the pump.

2.6. Photo-Colloidal Diode

Figure 2-10 Schematic depiction of source (PAG)-drain (PFA-S) based colloidal photo-

diode indicating both the rectification and the direction of movement of S-PS particles.

UV

51

In the pursuit to design multiple levels of regulation over colloidal transport, the

two individual pumps were combined to create a source-drain based colloidal photo-

diode which uses UV as the input to regulate the direction and speed of particle

transport. The PAG and the PFA-S were cast into separate films, about 300 µm away

from each other to create an emitter (PAG, proton generator)-collector (PFA-S, proton

consumer) system (Figure 2-10). The experiments were performed on glass surface

using S-PS tracers (2 µm) owing to their high zeta potential and desired direction of

motion: away from PAG and towards PFA-S. This push-pull mechanism was observed to

results in rectification of particle motion.

2.6.1 Experimental Set-Up

PFA-S polymers and PAG were dissolved in dichloromethane (20 mg/mL for

PFA-S, 100 mg/mL for PAG-2). One drop of PFA-S solution was put onto a glass slide

and one drop of PAG solution was put next to PFA-S droplet. The glass slide was

vacuum dried overnight. Then the PFA-S film (drain) and PAG (source) were carefully

cut into approx. 0.3 × 0.3 mm with the distance of approximately 300 µm apart. The

PFA-S/PAG was covered with hybridization chamber and a water suspension of sulfate

tracers (2 µm diameter) was added in the chamber. The videos were captured at 20X

magnification and 30 randomly selected particles were tracked for two consecutive 0.5 s

time-steps at 10 s intervals for 40 s total using tracker software (60 data points every 10

s).

52

2.6.2 Spatial and Temporal Regulation of Colloidal Transport

Before each experiment, a control video was recorded with UV lamp off. The

tracers show no directional movement, only Brownian motion was observed. Upon UV

illumination, the PAG initiated the diffusiophoretic motion pushing the negatively charged

tracers away in all directions (Supporting Video 2-7). When the acid formed by the

photolysis of PAG reached the PFA-S film, it started imine hydrolysis; the PFA-S film

then actively pulls the tracers precisely towards itself, and further enhances the tracer

velocities. As a result, the particle velocities in between the source and the drain were

observed to be much faster than that on opposite sides of either of them. Thus, upon

illumination, the particles began moving away from the PAG slowly and their velocities

steadily increased as they approached the drain; PFA-S. Moreover, the velocities in

between the source and drain were time-dependent owing to the delay in the PFA-S’s

hydrolysis which depends on the diffusion of the acid from the illuminated PAG. Figure

2-11 depicts the spatial and temporal control achieved using the source-drain set-up,

with the particles reaching a maximum average velocity of 3.9 µm/s. Effectively, the

designed colloidal photo-diode is capable of amplifying the velocity, as well as directing

the particle flow in one direction, without introducing a third “base/gate” terminal as

required by a typical transistor. This opens the door to creating more complex colloidal

logic systems.

53

Figure 2-11. Spatial and temporal regulation of velocity (S-PS particles) attained using

the source-drain photo-diode. Distance is measured from edge of the PAG and time is

measured from when the UV is turned on. For velocity vs time plot, distance = 150 µm;

for velocity vs distance, time = 20 s.

54

2.7 Conclusion

In conclusion, this chapter describes a set of smart, self-powered microscale

pumps capable of converting chemical/photochemical energy directly into mechanical

motion. The photo-triggered pump can be turned on and off repeatedly. Moreover, the

design of a colloidal photo-diode26 opens up further avenues to amplify and attain spatio-

temporal control over colloidal transport.

2.8 Acknowledgement

The author acknowledges the role of Hua Zhang in the synthesis of the Schiff’s

base polymer and his help in carrying out the pump experiments with it.

55

2.9 References

Parts of this chapter have been adapted from “Yadav, V.; Zhang, H.; Pavlick, R.; Sen, A. J. Am. Chem. Soc. 2012, 134, 15688-15691”. 1. Paxton, W. F.; Sundararajan, S.; Mallouk, T. E.; Sen, A. Angew. Chem. Int. Ed.

2006, 45, 5420-9. 2. Hong, Y.; Velegol, D.; Chaturvedi, N.; Sen, A. Phys. Chem. Chem. Phys. 2010,

12, 1423. 3. Sanchez, S.; Pumera, M. Chem. Asian. J. 2009, 4, 1402. 4. Mei, Y. F.; Solovev, A. A.; Sanchez, S.; Schmidt, O. G. Chem. Soc. Rev. 2011,

40, 2109. 5. Wang, J. ACS Nano. 2009, 3, 4. 6. Wang, J.; Manesh, K. M Small, 2010, 6, 338. 7. Mirkovic, T.; Zacharia, N. S.; Scholes, G. D.; Ozin, G. A. Small 2010, 6, 159. 8. Ebbens, S. J.; Howse, J. R. Soft Matter, 2012, 8, 3077. 9. Jun, I.K.; Hess, H. Adv. Mat. 2010, 22, 4823. 10. Kline, T. R.; Paxton, W. F.; Wang, Y.; Velegol, D.; Mallouk, T. E.; Sen, A. J. Am.

Chem. Soc. 2005, 127, 17150. 11. Ibele, M. E.; Wang, Y.; Kline, T. R.; Mallouk, T. E.; Sen, A. J. Am. Chem. Soc.

2007, 129, 7762. 12. Pavlick, R. A.; Sengupta, S.; McFadden, T.; Zhang, H.; Sen, A. Angew. Chem. Int.

Ed. 2011, 50, 9374. 13. Zhang, H.; Yeung, K.; Robbins, J. S.; Pavlick, R. A.; Wu, M.; Liu, R.; Sen, A.;

Phillips, S. T. Angew. Chem. Int. Ed. 2012, 51, 2400. 14. Smith, E. J.; Xi, W.; Makarov, D.; Monch, I.; Harazim, S.; Quinones, V. A. B.;

Schmidt, C. K.; Mei, Y.; Sanchez, S. ; Schmidt, O. G. Lab Chip, 2012, 12, 1917. 15. Solovev, A. A.; Sanchez, S.; Mei, Y.; Schmidt, O. G. Phys. Chem. Chem. Phys.

2011, 13, 10131.

56

16. Solovev, A. A.; Smith, E. J.; Bof Bufon, C. C.; Sanchez, S. Schmidt, O. G. Angew. Chem. Int. Ed. 2011, 50, 10875.

17. Tauk, L.; Schröder, A. P.; Decher, G.; Giuseppone, N. Nature Chem. 2009, 1,

649. 18. Deng, G.; Tang, C.; Li, F.; Jiang, H.; Chen, Y. Macromolecules 2010, 43, 1191. 19. Gu, J.; Cheng, W. P.; Liu, J.; Lo, S. Y.; Smith, D.; Qu, X.; Yang, Z.

Biomacromolecules 2007, 9, 255. 20. Wang, C.; Wang, G.; Wang, Z.; Zhang, X. Chem. Europ. J. 2011, 17, 3322. 21. Anderson, J. L. Annu. Rev. Fluid Mech. 1989, 21, 61. 22. Gu, Y.; Li, D. J. Colloid Interface Sci. 2000, 226, 328. 23. Delair, T.; Meunier, F.; Elaissari, A.; Charles, M. H.; Pichot, C. Colloids Surf., A

1999, 153, 341. 24. Bastos, D.; De Las Nieves, F. J. Prog. Colloid Polym. Sci. 1993, 93, 37. 25. Gracia-Salinas, M. J.; Romero-Cano, M. S.; De Las Nieves, F. J. Prog. Colloid

Polym. Sci. 2000, 115, 112. 26. Yadav, V,; Zhang, H.; Pavlick, R. A.; Sen, A. J. Am. Chem. Soc. 2012, 134,

15688-15691.

57

Chapter 3

Bone-Crack Detection, Targeting and Repair Using Ion Gradients

3.1. Introduction

Self-powered nanomotors and pumps are increasingly being explored for

biological applications given the advances in basic motor design and functionality over

the last decade.1-13 Such autonomous devices, requiring no external power supply, offer

a broad range of potential bio-medical applications ranging from targeted drug delivery

to minimally invasive surgeries. Ion gradients can cause diffusiophoretic transport of fluid

and particles and provide a technique for directing movement towards specific targets.

This chapter describes a biological-synthetic hybrid micropump-based strategy for

detection of bone lesions or dental cracks by utilizing the damaged matrix itself as both

the trigger and the fuel. This strategy is also applicable to synthetic surfaces with equal

efficiency, a case in point being the polymer repair system described in the last section

of this chapter. A crack in a mineral-rich material generates ion-gradient driven electric-

fields which can be utilized for active targeting and treatment.

3.2 Motivation

The hybrid approach described in this chapter complements but is orthogonal to

current methods that promote healing by delivery of a therapeutic agent to the bone via

passive diffusion.14-18 The current clinical treatments include systemic anti-resorptive

(bisphosphonates19) or anabolic therapies (parathyroid hormone therapy20), which are

useful for general increase in mineralization and bone strength in patients. However,

58

since bone diseases like osteoporosis vary in degree of degeneration at different

skeletal sites, fractures of vulnerable areas like the hip, spine and wrist are common

even with preventative therapies.21 Consequently, a variety of new targeting strategies to

increase drug delivery to the bone, are currently being investigated and include, for

example environmentally sensitive cleavable linker systems, and fusion proteins or

nanoparticles with bone targeting moieties.22 While these treatments enhance specificity

to bone, mechanisms for active delivery of agents to target sites most at risk for fracture

or of active degeneration remain elusive and are highly desired. Described in this

chapter, is the active detection of ex vivo human bone cracks using charged quantum

dots and strategy for repair, based on the phenomenon of diffusiophoretic motion. The

role of electric fields and ensuing electrophoresis as a mechanism for the directional

movement of motile cells has also previously been illustrated.23

3.3 Generation of Local Electric Fields

The strategy is based on the generation of ion gradients from freshly exposed

mineral surfaces, which results in a local electric field that can be exploited for targeting

and treatment. Bone is a composite material that supports load. It is composed of

collagen and a mineral matrix most closely resembling hydroxyapatite.24 The mineral is

also used in cements for bone repair as well as an implant coating for improved

biocompatibility and integration of medical devices. Hydroxyapatite, at the physiological

pH, undergoes hydrolysis as follows:

Ca10(PO4)6(OH)2 + 12H2O 10Ca2+ + 6H2PO4- + 14OH-

59

A crack in the bone releases ions into the surrounding solution. The large

difference in diffusion coefficients between the cation (Ca2+) and the faster anion (OH-)

[D(Ca2+) = 0.789x10-5 cm2s-1, D(OH-) = 5.273x10-5 cm2s-1, D(H2PO4-) = 0.959x10-5 cm2s-1]

induces a local electric field oriented outwards, away from the crack in the bone surface

(i.e., the ion source). Charged moieties introduced in the system respond to this electric

field undergoing diffusiophoretic motion. As has been described in the previous two

chapters, in an unbounded solution of a symmetrically charged binary electrolyte with a

uniform concentration gradient ∇𝑐𝑐, the diffusiophoretic velocity of a charged particle, U, is

given by Equation 3.1,25

(3.1)

where D+ and D- are the diffusion coefficients of the cation and anion

respectively, Z is the absolute value of the valences of the ions, e is the charge of an

electron, k is the Boltzmann constant, T is the absolute temperature, ε is the dielectric

permittivity of the solution, η is the viscosity of the solution, ζ𝑃𝑃 is the zeta potential of the

particle, γ = tanh(Zeζ𝑃𝑃/4kT), and c0 is the bulk concentration of ions at the particle

location, as if the particle was not there. Typical electric fields generated in

diffusiophoretic systems are 1-10 V/cm, sufficient to cause directed motion of charged

particles. Electric fields of similar magnitude are also known to cause galvanotaxis of

motile cells (reorientation and migration along the direction of the electric field).23

U =𝜀𝜀kT𝑍𝑍𝑅𝑅𝜂𝜂 ��

𝐷𝐷+ − 𝐷𝐷−

𝐷𝐷+ + 𝐷𝐷−� ζ𝑃𝑃 −2kT𝑍𝑍𝑅𝑅 𝜌𝜌𝑙𝑙(1 − 𝛾𝛾2)

∇𝑐𝑐

c0�

60

Figure 3-1. Schematic depiction of ion gradient-induced electric field and the resultant

particle migration. The length of the arrows next to the ions represent their relative

mobilities. The generated electric field points outwards away from the crack.

Accordingly, the negatively charged particles move towards and positively charged

particles move away from the crack.

61

In the described system, anionic or cationic moieties are expected to respond to

the diffusion-induced electric field generated by slow dissolution of hydroxyapatite by

moving towards or away from the source (the crack), respectively (Figure 3-1). The

hypothesis was first tested out by evaluating the mobility of negatively and positively

charged quantum dots in the presence of a cracked substrate.

3.4 Experimental Design

Bone from human tibia and femur, obtained from the Boston medical school were

cut using an IsoMet 3000 (Beuhler, IL) with a diamond metal bonded, wafering blade.

Samples were cut at approximately 500 micrometer thickness at low speed using saline

lubrication bath. Bone samples were stored at 4 °C in saline and washed with DI prior to

analysis. The experimental set-up involved introducing a fluorescent quantum dot

solution into a 9 mm diameter, 0.6 mm thick imaging spacer, covering the cracked

bone/PDMS samples. The quantum dots were 20 nm crystals of a semiconductor

material (CdSe), which are shelled with an additional ZnS semiconductor layer. This

core-shell material is coated with a polymer layer that has –COO- or –NH3+ surface

groups. 10 μL of a 8 μM stock of quantum dot solution was diluted to 500 μL with DI

water. The setup was sealed, inverted and placed on the confocal microscope stage

(Leica TCS SP5 confocal instrument). Ar laser was used for the imaging of the quantum

dots. The fluorescence intensity was monitored at the crack with a 10X objective every

10 min for 2 h following the time taken to set up the experiment (~ 5 min). T = 0 is

defined as the point of initial observation.

62

3.5 Diffusiophoresis led Damage Detection

When the carboxylate-functionalized negatively-charged quantum dots were

added to a freshly cracked bone within the confines of a hybridization chamber and

monitored on a confocal microscope the quantum dot intensity was observed to increase

within the crack, and along its edges due to the expected diffusiophoretic movement

(Figure 3-2a). In contrast, the amine-functionalized, positively-charged quantum dots

move outwards, away from the crack (Figure 3-3a.). Image J software was used for

intensity calculation. The fluorescence intensity was averaged over the entire crack area

and the values obtained from different experiments were normalized (maximum = 1) for

comparison and representation on a common scale.

The experiments were carried out in an inverted set-up, eliminating the role of

gravity in the particle migration. Control experiments were performed by immersing

cracked bone slices in DI water for 3-4 weeks, till no further measurable change in

conductivity was recorded after showing an initial increase of ~15 µS/cm every 10

minutes. When exposed to quantum dot solution no change in intensity was observed in

this case (Figure 3-3c). Marangoni and other non-ionic gradient-driven flows can also

cause active movement of particles.26 However, our observation of opposite directional

migration of positively and negatively charged particles suggest diffusiophoresis to be

the dominant propulsion mechanism.

63

Figure 3-2. Increasing quantum dot intensity within the crack on bone surface (a) and

PDMS surface (b) demonstrating an effective damage detection scheme. Scale bar is 60

µm. Right panel shows calculated intensities inside the damage (averaged over entire

damaged area) for HOOC Q-Dots, amine Q-Dots and control, using Image J software,

for bone surface (c) and PDMS surface (d).

64

The mechanism described above is not surface specific, and its versatility can be

gauged from its effectiveness on synthetic mineral surfaces as well. To generalize the

crack detection mechanism, an artificial system was engineered by embedding

hydroxyapatite in between two 1 mm thick polydimethylsiloxane (PDMS) layers. The

procedure involved mixing PDMS elastomer base with the crosslinker, Sylgard® 184

Silicone Elastomer, in a 10:1 ratio. Then 3 g of this mixture was poured on to a 60 x 15

mm polystyrene Petri-dish and placed in a vacuum desiccator for 30 min. to remove

bubbles from the mixture. The dish was then placed in an oven at 60⁰C for 1 hr. to cure

the PDMS. A thin layer or line of solid hydroxyapatite was added on the PDMS base.

Then an additional 3 g of a mixture of 10:1 PDMS and crosslinker was poured on top of

the hardened PDMS. The coating was then again placed in the vacuum desiccators for

30 min. followed by curing at 60⁰C for 1 hr. The embedded PDMS was then cut out of

the Petri-dish and sealed onto a glass slide using a high frequency generator. A crack

was formed in this “artificial bone” using a scalpel and a similar quantum dot migration

study was performed. As expected, the negatively charged quantum dots migrated

towards the crack, as indicated by the increase in fluorescence intensity (Figure 3-2b),

while the positively charged ones migrated away from the crack (Figure 3-3b). Note that

the rate of ion release in solution is governed by the level of hydroxyapatite incorporation

and the hydrophobicity of the PDMS. Control experiments with a PDMS layer containing

no mineral showed no increase in the quantum dot intensity within the crack over similar

time periods (2 h) (Figure 3-3d). These data establish an effective and versatile crack

detection system utilizing ion gradient-induced electric fields. In principle, any underlying

layer of mineral can be effectively utilized to detect cracks on a surface, as long as the

cation and the anion have significantly different diffusivities.

65

Figure 3-3. Analysis of the crack detection scheme using confocal microscopy. Intensity

study within the crack on bone surface (a) and PDMS surface (b) using amine

functionalized quantum dots. Control images showing no intensity change on bone (c) &

PDMS (d). Scale bar is 130 µm.

20 min 0 min 40 min

0 min 20 min 40 min

0 min 20 min 40 min

0 min 40 min 20 min

66

3.6 Disfussiophoresis Guided Targeted Protein delivery

3.6.1 Fluorescence Microscopy analysis

To further demonstrate the generality and applicability of this approach, the

migration of an anionic protein was evaluated to the bone crack site. Urease enzyme

was chosen since it has an isoelectric point less than the physiological pH. Urease was

tagged with Dylight melamide 550 in 1 mM PBS introduced over the cracked bone

surface, and followed under a confocal microscope using a He-Ne laser for imaging.

Urease (type C-3) was tagged with a thiol-reactive dye, Dylight 550 (ex/em: 557/572).

The reaction of the fluorescence probe (40 μM) with urease (2 μM) was carried out in

150 mM phosphate buffer (pH 7) at room temperature for 4-5 h under gentle stirring. The

enzyme-dye complexes were further purified using membrane dialysis (10 kDa pores) to

reduce free-dye concentration. The number of dye molecules per catalase enzyme

molecule was ~2 as quantified with UV-Vis spectroscopy. Urease was observed to

consistently move towards the crack, increasing the dye intensity within the crack and at

its edges.

3.6.2 Raman Spectroscopy Analysis

In order to further support this finding, Raman data was acquired on the enzyme-

containing bone samples. Control spectra were collected for both the enzyme and the

bone, individually, and overlaid with the bone sample with the deposited enzyme (Figure

3-4). Raman spectra were acquired using a confocal Raman microscope equipped with

a 40 X (NA = 0.6) objective, utilizing a 785 nm diode laser for excitation. A 10 mM

67

solution of urease was introduced on to a cracked bone slice within the confines of an

imaging spacer. After an hour of exposure, the bone slice was taken out of solution and

allowed to air dry. The integration time for each spectrum was 30 sec. The spectra were

recorded by an And Or DV401-BV CCD camera attached to an Acton Research

Corporation SpectraPro-2300i spectrometer using the 600 g/mm diffraction grating. The

characteristic stokes lines for the human bone were identified at 965, 1075 and 1269,

1456, 1669 cm-1, indicating the presence of phosphate, carbonate and amide bonds

respectively (other notable peaks at 862, 596 and 436 cm-1).27 Urease enzyme showed a

broad peak centered around 379 cm-1. The presence of these characteristic peaks from

both the bone and the enzyme was noted in the analyzed sample, indicating the

presence of enzyme at the crack site (Figure 3-4a).

Conclusive evidence of the enzyme migration towards the crack site was noted

upon collection of Raman spectra at increasing distances away from it. Spectra recorded

at the precise crack site displays an intense enzyme peak along with a noticeable

characteristic bone peak (phosphate). As we moved away from the crack (in 20 µm

steps) the ratio of characteristic urease peak to that of bone consistently decreases

indicating the diffusiophoretic motion of the anionic enzyme towards the ion source

(crack) (Figure 3-4b).

68

Figure 3-4. (a) Raman spectra obtained on the bone and enzyme separately, overlaid

with one collected on the bone exposed to the enzyme. (b) Raman spectra at increasing

distances from the crack depicting the preferential enzyme migration towards the crack.

69

3.7 Targeted Drug Delivery

The motion of this self-propelled system was next explored as a targeting

mechanism, such as a drug delivery vehicle, transporting biomaterials to the bone-crack

site. Accordingly, negatively charged, fluorescently labeled poly(lactic-co-glycolic acid)

(PLGA) nanoparticle containing sodium alendronate were prepared and characterized.

3.7.1 Synthesis of Alendronate Nanoparticle

The drug loaded nanoparticles were synthesized by adding 50 mg of

alendronate sodium dissolved in 1mL of deionized water to a mixture of 200mg PLGA

(MW, 44k) and 1 mg Nile red dissolved in 5 mL dichloromethane followed by sonication

of the combined mixture for 2 min. 20 mL of 0.05 g/mL SDS solution was added and

again sonicated for 2 min. Following sonication, 100 mL of deionized water was added

and the solution was allowed to stir overnight, exposed to air to allow evaporation of the

organic solvent. The solution was centrifuged and the resulting pellets re-suspended in 5

mL deionized water and centrifuged again. The pellets were next re-suspended in 1mM

phosphate buffer for analysis and use.

3.7.2 Drug load Calculation

Alendronate concentration was determined by a fluorimetric assay of its complex

with fluorescamine. The PLGA nanoparticles were degraded in 1 M sodium hydroxide for

1 h and then the solution neutralized with 1 M hydrochloric acid. Alendronate was

reacted with fluorescamine in a pH 10 borate buffer. Fluorescence was compared to that

70

of known concentrations of alendronate to determine loading.28, 29 Drug loading of the

particles was found to be 70.3 ± 5.3 %.

3.7.3 Particle Characterization

Drug loaded nanoparticles were imaged by allowing a 10 µL droplet of 1 mg/mL

nanoparticle solution to dry on a silica wafer connected to an aluminum sample stub.

Samples were imaged on a Zeiss SUPRA 40VP (Carl Zeiss Microscopy LLC,

Oberkochen, Germany) field emission scanning electron microscope (FE SEM) using an

accelerating voltage of 2 kV (Figure 3-5).

The hydrodynamic size of the nanoparticles was measured by dynamic light

scattering (DLS) at 25 °C using the Brookhaven Instruments 90 Plus Particle Sizer.

Samples were prepared in 1mM PBS. Hydrodynamic size was based on 5 measurement

Z-average/-effective of intensity-based distribution. The effective diameter was

measured to be ~220 nm

The zeta potential of the nanoparticles was measured using the Brookhaven

Instruments ZetaPALS Zeta Potential Analyzer. The same samples used for DLS size

characterization were used to determine zeta potential. Each sample was subjected to 5

runs each consisting of 20 cycles with auto correction for voltage. The zeta potential of

the particles was measured to be -24.5 ± 1.1 mV.

71

Figure 3-5. Electron microscopy analysis of drug loaded particles: SEM images of

PLGA nanoparticles coated with Au/Pd sputter coating for visualization.

72

3.7.4 Drug Delivery and Cell Proliferation Assay

PLGA is a well-known biocompatible polymer used in medical devices,28 and

sodium alendronate is a bisphosphonate drug used for the clinical treatment of

osteoporosis. The experiments were all performed at the physiological pH in 1 mM PBS

and followed using confocal microscopy. Once again increased fluorescence intensity in

the crack indicated the active migration of the negatively charged drug loaded

nanoparticles towards the crack (Figure 3-6).

Figure 3-6. Increasing fluorescence intensity within the crack indicates active migration

of Nile-red tagged drug loaded PLGA particles to the crack site demonstrating an

effective drug delivery protocol. Scale bar is 100 µm.

73

To confirm that this drug delivery vehicle was indeed capable of delivering an

active agent, an in vitro cell proliferation assay30 was performed with human MG-63 cells

- an immortalized osteoblast cell line. The cells were maintained in Dulbecco’s Modified

Eagle Media supplemented with 10% bovine calf serum and 1% penicillin/streptomycin

in a humidified atmosphere at 37 °C and 5% CO2.30

For the MTS assay, MG-63 cells were plated at a density of 1 x 104 cells/well in

96 well plates. After overnight incubation at 37 °C, the media was replaced with

media/PLGA nanoparticle suspension containing 10-6, 10-8, or 10-10 M sodium

alendronate and the cells were allowed to incubate for 48 hours. Cell viability was tested

using a colorimetric MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-

sulfophenyl)-2H-tetrazolium) cell proliferation assay and absorbance read at 490 nm.

Data is expressed in Figure 3-7 as the percentage of optical density relative to medium

alone (control) which is taken as 100%.30

The colorimetric assay utilized, measures increase in cell proliferation, induced

by the drug, signified by increase in optical density. Indeed, an increase in cell density

was observed in cells treated with alendronate as compared to the control, non-treated

group, demonstrating increased cell proliferation and successful release of drug from the

PLGA nanoparticles. The increased cell growth (~10%) was consistent with other

reports30 and the clinical use of alendronate for bone regeneration and repair.

74

Figure 3-7. Proliferation of MG-63 cells treated with PLGA nanoparticles containing 10-6,

10-8 and 10-10 M alendronate for 48 hours, expressed as percentage optical density

relative to the negative control of 100%, using a colorimetric MTS cell proliferation

assay. (Graph expressed as Mean ± SD; Significance (*P < 0.05) compared with

negative control group (medium alone)).

75

3.8 Expansion of the detection and repair technique

The use of ion gradients and in-turn diffusiophoretic motion represents a new

approach to targeting a biological structure that augments current methods that are

focused on primarily biomacromolecular interactions involving small molecules, proteins

and nucleic acids. The above described active, self-propelled particle-based bone crack

detection, drug delivery, and repair strategy requires no external trigger or fuel supply,

and is based on ion gradients. The one challenge with this technique however remained

the high ionic strength that the drug loaded particles would have to withstand within the

body. This presented a major challenge to the in-vivo applicability of the scheme.

However, with the merits of the technique and the applicability of a self-guided technique

in the medical field in mind, the technique was expanded to investigate dental caries and

tooth decays.

3.8.1 Present therapeutic techniques

The presently available treatments to tooth decay or cracks remain inadequate,

similar to the bone lesion issue. Fluoride infused toothpastes or gels offer a remedy,

coating a layer of fluoroapatite on the tooth which is more resistant to acidic

environments, typical in case of bacterial decay. However, the mechanisms by which

fluoride acts on the teeth remain unknown.31 Moreover, less than 50% of the fluoride

actually reaches the target site and the rest is removed from the body. An excess

dosage of fluoride also increases the risk of fluorosis to the patient that is manifested as

joint pain in both upper and lower limbs, numbing and tingling of the extremities, back

pains, and knock-knees.32 Interestingly, tooth is also composed of hydroxyapatite, the

76

same mineral forming bone and therefore the diffusiophoretic mechanism, was applied

to the present problem. The fluorescent imaging was done as before, only this time in

the biologically relevant ionic strengths. Saliva typically has one-third to one half the

ionic strength of serum.33 Ionic strength dependent detection study revealed the

mechanism to be applicable upto ~60 µM salt concentrations. Preliminary testing with

quantum dots (Figure 3-8) gave expected results, with anionic particles leading to

detection.

Figure 3-8: Damage detection in cracked teeth. Negatively charged amine

functionalized quantum dots move in towards the crack leading to an increase in

fluorescence intensity (a) while positively charged carboxyl functionalized quantum dots

move away from the crack leading to decrease in fluorescence intensity (c). Images (b)

& (d) are the bright field images of the tested crack.

77

3.8.2 Detection using FDA approved diagnostic dye

Figure 3-9. (a) Increasing fluorescein intensity within the dental crack in a tooth slice

leads to detection. (b) Fluorescence intensity analysed inside the damage (averaged

over entire damaged area) through Image J.

(a)

0.5

0.6

0.7

0.8

0.9

1

1.1

0 20 40 60 80 100 120 140

Nor

mal

ized

Flu

ores

cenc

e In

tens

ity

Time (s)

(b)

78

Pursuing the biologically relevant theme, a biocompatible diagnostic dye

fluorescein, approved by the FDA for biological imaging in humans34, was used for

detection. The anionic dye, when tested on the confocal fluorescence microscope,

showed accumulation within the crack, enabling detection (Figure 3-9). Once detection

using fluorescein was confirmed on tooth slices, a crack was designed on a whole

human tooth and tested under similar conditions. The detection technique was found to

work well enabling the detection of a minor crack on the tooth (Figure 3-10). In-vivo

testing of the present technique is under way at the Boston Medical School.

Figure 3-10. Damage detection on a whole tooth using fluorescein dye.

79

3.8.3 Mechanism of Fluoride treatment

In addition to designing a detection scheme, this chapter also describes a

diffusiophoresis based mechanism behind the effectiveness of fluoride treatment on

teeth. A tooth sample was cracked and exposed to sodium fluoride solution. An EDS

analysis on the sample reveals an increasing ratio of fluoride ions as one approaches

the damage site (Figure 3-11). Moreover, EDS maps confirmed an excess of fluoride at

the damage site (Figure 3-12), leading to the postulate that diffusiophoresis is possibly

the predominant mechanism behind the fluoride medication efficacy.

Figure 3-11. EDS measurements at increasing distances from the crack, show a

decreasing fluoride signal.

Increasing distance from crack

80

Figure 3-12. EDS maps generated at the crack site show a heavy deposition of sodium

and fluoride at the crack site. The presence of the crack can be noted by the scarcity of

calcium, phosphate and oxygen at the same site, the primary components of

hydroxyapatite. Scale bar is 400 µm.

Fluoride Sodium

Calcium Phosphate Oxygen

81

3.9 Application on Synthetic Surfaces- Polymer Repair

Another facet of versatility of the described ion gradient generated

diffusiophoretic mechanism is the applicability to a variety of surfaces both biological and

synthetic. This section describes a follow-up of this study on synthetic polymeric

surfaces.

3.9.1. Motivation

Developing methods to detect and repair damage in polymers is an active area of

research.35-47 Many of the previously described methods suffer from the lack of long-term

stability of the reagents, which are typically pre-incorporated into the polymer. Also, they

tend to be specific to certain types of polymeric materials. More recent systems have

utilized supramolecular interactions, photochemistry, and thermal heating to seal

cracks.41-43 However; these systems have long healing times. The ‘ion gradient triggered

motion’ approach was expanded into this domain and a general method for detection

and repair of cracks in polymers was designed. The process consists of embedding the

polymer with a salt that leaches out upon cracking or degradation to the polymer,

thereby powering flows and activating the detection or repair reagents in the fluid.

Detection is possible with fluorescent quantum dots, which aggregate at the crack site.

Repair is shown to occur through two different strategies. The first repair strategy

involves high ionic strength triggered destabilization of oil-in-water emulsions,

transporting polymerization agents, resulting in polymer deposition at damage site. The

second, more biocompatible strategy, involves using an enzyme, urease, and its

catalytic hydrolysis of urea to deposit solid calcium carbonate in the crack. The solution

82

of the detection or healing agent may be added “as needed” thereby overcoming the

problem of reagent instability

3.9.2 Density Driven Flows

After detailed mechanistic studies it was concluded that a salt driven density flow

works more effectively in the given system compared to a directional diffusiophoretic

particle motion. A density driven flow can actively transport active agents to the target

site. The flow occurs when the salt concentration increases sufficiently to cause a local

density change. Such density driven flows are very common in nature at mineral salt

deposits and estuaries. Local fluid density varies with salt concentration, causing flows

dictated by the direction of gravity. 48-52

The samples were prepared in the same manner as described in section 3.5,

only replacing the sparingly soluble hydroxyapatite with the highly soluble calcium

chloride (solubility, 74.5 g in 100 ml at 20°C53). When the surface was cracked, the

calcium chloride dissolved into the supernatant aqueous solution, creating a change in

the local density powering density driven flows. This was confirmed by inverting the

experimental set-up that led to a reversal of flow directions. Moreover, the motion was

observed to be independent of the charge of the motile specie. Both positively and

negative charged particles were led to the damage site with equal speeds. The following

section describes the synthesis of repair agents that get transported to the damage site,

through the triggered flows.

83

3.9.3 Synthesis of repair agents

Two 1 mL centrifuge tubes were filled with 450 μL nanopure H2O and 20 μL oleic

acid. One tube was then filled with 40 μL 5 wt% Grubbs Catalyst, 2nd Generation in

1,1,2-trichloroethane. To the other tube, 40 μL liquid dicyclopentadiene, heated to 43°C,

was added. The tubes were then emulsified for 180 sec each forming the emulsions.

Equal volumes of each emulsion mixture were then added to a container. The separation

of the monomer from the catalyst is itself important; for certain polymerizations occur too

fast to allow adequate mixing of reactants, as is seen in the well-known Grubbs’ ring

opening metathesis system.54-58 The emulsions were tested to be stable for upto 20

hours in low salt concentrations (<0.1M), but underwent immediate destabilization in

higher salt concentrations, thereby being ideal agents for the designed system.

3.9.4 Polymer Repair

As the damaged salt-embedded polymer released ions, density driven flows were

generated. As the flows brought the emulsions over the crack, the high ionic strength at

the damaged site caused the emulsions to break open and aggregate, allowing the

catalyst and reactant to mix. This mixing resulted in polymer formation and deposition at

the crack, repairing the damage. The spacers were then removed after 1 hr, the samples

were washed extensively with DI water, dried and studied with an environmental

scanning electron microscope (ESEM) (Figure 3-13). The samples showed substantial

filling of the crack with polymer deposition centered at the crack. Deposition in the crack

was more uniform in the inverted samples, since the flow at the surface was directed

towards the crack allowing for better delivery of the repair reagents.

84

Figure 3-13. ESEM images of polymer deposition at the damage site. The strategy

works well for both single (a, b) and multiple cracks (c, d). (a, c) The image of cut

polymer with no salt after 1 hr. exposure to emulsions. (b, d) PDMS/CaCl2 after 1 hr.

exposure to emulsions.

500µm 500µm

CRACK 1

CRACK 2

500µm 500µm

CRACK 1

CRACK 2

(a) (b)

(c) (d)

85

Micro AT-IR spectroscopy was used to characterize the polymer deposits. The IR

spectrum showed key peaks for poly-DCPD at 1665 cm-1 and 956 cm-1 for the acyclic

C=C and the acyclic =C-H stretching, respectively.59, 60 A control with the emulsion

solution over a cut in pure PDMS film showed no polymer deposition.

3.9.5 Enzymatic repair

Figure 3-14. Schematic of a surface healing system using a salt/PDMS film. The urease

enzymes (blue) and urea molecules (grey) move over the crack due to density driven

flows. While this occurs, the urea is converted by the urease to carbonate ions

(pH~10.3). The carbonate ions then react with the leaching calcium ions forming solid

calcium carbonate.

86

Having established this polymeric repair strategy, the possibility of a more bio-

friendly approach was explored. The concept involved the addition of a premixed

solution of the enzyme urease (2 µM) and its substrate urea (1 M) to a calcium ion-

leaching crack. At pH 10.3, carbonate ions are formed due to enzymatic hydrolysis of

urea. These carbonate ions would be expected to combine with the leached calcium ions

to deposit calcium carbonate at the crack (Figure 3-14).

A damaged PDMS/CaCl2 system was prepared as before. When the crack in this

system was exposed to a mixture of urease and urea (pH~10.3; a drop of 0.1 M

ammonium hydroxide added to raise the pH.) and left undisturbed for 1 hr, a white

precipitate was observed to deposit at the crack (Figure 3-15a). The precipitate showed

aragonite and calcite like morphology (Figure 3-16a). XRD analysis at the crack site

also confirmed the presence of calcium carbonate, in aragonite and calcite forms

(Figure 3-15b). XRD patterns were collected using PANalytical Empyrean theta-theta

goniometer with Cu-K-alpha radiation, and programmable divergence slit (2 mm, 1.0

degree anti-scatter, specimen length 10mm) and diffracted (2 mm, 0.02 mm nickel filter)

optics in reflection geometry. Data was collected at 45 kV and 40 mA from 5-70 degrees

2-theta using PIXcel 3D detector in scanning mode with a PSD length of 3.35 degrees 2-

theta, and 255 active channels for duration of ~0.5 hr. Resulting patterns were corrected

for both 2-theta and position by comparison to ICDD (calcite PDF #00-005-0586 and

aragonite PDF #00-041-1475) and analyzed with Jade+9 software.

Supporting IR spectra were also collected at the crack site, (Figure 3-16b) to

corroborate the XRD findings. Carbonate vibration bands61 at 1460 cm-1 (symmetric

stretching) and 880 cm-1 (out-of-plane bending) confirmed the presence of the

precipitated carbonate.

87

Figure 3-15. (a) ESEM images showing the control (left) and sample (right) where the

crack was exposed to the urease-urea mixture without and with the underlying calcium

chloride layer, respectively. Scale bar is 100 µm. (b) XRD Analysis of the crack site

confirming the presence of calcite (red bars-standard) and aragonite (blue bars-

standard). The amorphous halo at lower two-theta values is due to PDMS.

(a)

(b)

88

Figure 3-16. (a) SEM image of the precipitated material within the crack showing

aragonite and calcite like morphology. Scale bar is 20 µm. (b) Micro-IR: Carbonate

vibration bands61 at 1460 cm-1 (symmetric stretching) and 880 (out-of-plane bending)

cm-1 confirms the presence of the precipitated calcium carbonate.

(a)

(b)

89

3.10. Conclusion

To conclude, the versatility of the ion- gradient-powered system has been

demonstrated in its ability to perform both crack detection and repair, on a variety of

substrates. Moreover, the designed system enables delivery of nanoparticles, protein,

emulsions, quantum dots, diagnostic dyes with equal efficacy.62 The elimination of the

requirement of an external power source to power the motion makes the system

especially desirable. This chapter describes a system that has practical applicability in

medical diagnostics, polymer and coatings industry.62, 63 The involved methodology is

versatile and the detection and healing solutions can be assembled as needed. Finally,

the method should be useful for coatings on materials that are not easy to remove and

repair. A comprehensive assessment of various salt and polymer combinations and the

effect of salt impregnation on polymer properties are planned. Future systems will also

seek to improve the deposition process and allow for restoration of the polymer

properties.

3.11. Acknowledgements

The author would like to express her gratitude towards Jonathan Freedman for

his help in the synthesis of the drug loaded nanoparticles, providing the bone and teeth

samples, and carrying out the cell proliferation assay. The author would also like to

thank Ryan Pavlick for his valuable contribution on the polymeric repair system.

90

3.12 References

Parts of this chapter have been adapted from “Yadav, V., Freedman, J. D., Grinstaff, M.; Sen, A. Angew. Chem., Int. Ed., 2013, 52, 10997-11001” and “Yadav, V.; Pavlick, R. A.; Meckler, S.; Sen. A. Triggered Detection and Polymer Deposition: Towards the Repair of Microcracks, Chem. Mater., 2014 26, 4647-4652”.

1. Ozin, G. A.; Manners, I.; Fournier-Bidoz, S.; Arsenault, A. Adv. Mater. 2005, 17, 3011-3018.

2. Paxton, W.F.; Mallouk, T. E.; Sen, A. Chem. Eur. J. 2005, 11, 6462-6470. 3. Paxton, W. F.; Sundararajan, S.; Mallouk, T. E.; Sen, A. Angew. Chem., Int. Ed.

2006, 45, 5420-5429. 4. Wang, J. ACS Nano 2009, 3, 4-9. 5. Sanchez, S.; Pumera, M. Chem. Asian J. 2009, 4, 1402-1410. 6. Hong, Y.; Velegol, D.; Chaturvedi, N.; Sen, A. Phys. Chem. Chem. Phys. 2010,

12, 1423-1425. 7. Wang, J.; Manesh, K. M. Small, 2010, 6, 338-345. 8. Mirkovic, T.; Zacharia, N. S.; Scholes, G. D.; Ozin, G. A. Small 2010, 6, 159-167. 9. Mei, Y.; Solovev, A. A.; Sanchez, S.; Schmidt, O. G. Chem. Soc. Rev. 2011, 40,

2109-2119. 10. Yadav, V.; Zhang, H.; Pavlick, R. A.; Sen, A. J. Am. Chem. Soc. 2012, 134,

15688-15691. 11. Sengupta, S.; Ibele, M. E.; Sen, A. Angew. Chem., Int. Ed. 2012, 51, 8434-8445. 12. Patra, D.; Sengupta, S.; Duan, W.; Zhang, H.; Pavlick, R. A.; Sen, A. Nanoscale

2013, 5, 1273-1283. 13. Sengupta, S.; Dey, K. K.; Muddana, H. S.; Tabouillot, T.; Ibele, M.; Butler, P. J.;

Sen, A. J. Am. Chem. Soc. 2013, 135, 1406-1414. 14. Arns, S.; Gibe, R.; Moreau, A.; Monzur Morshed, M.; Young, R. N. Bioorg. Med.

Chem. 2012, 20, 2131-2140. 15. Fleurence, R. L.; Iglesias, C. P.; Johnson, J. M. PharmacoEconomics 2007, 25,

913-933.

91

16. Olszynski, W. P.; Davison, K.S. Expert Opin Pharmacother 2008, 9, 491-498. 17. Wang, D.; Miller, S. C.; Kopeckova, P.; Kopecek, J. Adv. Drug Deliv. Rev. 2005,

57, 1049-1076. 18. Zhang, G. et al Nat. Med. 2012, 18, 307-314. 19. Russell, R. G; Xia, Z; Dunford J. E.; Oppermann, U.; Kwaasi, A.; Hulley, P. A.; et

al. Ann N Y Acad Sci. 2007, 1117, 209–257. 20. Uihlein, A. V.; Leder, B. Z. Endocrinol. Metab. Clin. North. Am. 2012, 3, 507-525. 21. Colón-Emeric, C. Z. J. Am. Med. Assoc. 2006, 296, 2968-2969. 22. Luhmann, T.; Germershaus, O.; Groll, J.; Meinel, L. J. Control Release 2012,

161, 198-213. 23. Allen, G. M.; Mogilner, A.; Theriot, J. A. Curr. Biol. 2013, 23, 1-9. 24. Boskey, A.; Pleshko Camacho, N. Biomaterials 2007, 28, 2465-2478. 25. Anderson, J. L. Annu. Rev. Fluid Mech. 1989, 21, 61-99. 26. Zhao, G.; Stuart, E. J. E.; Pumera, M.; Phys. Chem. Chem. Phys., 2011, 13,

12755-12757. 27. Smith, R.; Rehman, I. J. Mater. Sci.: Mater. Med., 1995, 5, 775-778. 28. Cohen-Sela, E.; Chorny, M.; Koroukhov, N.; Danenberg, H. D.; Golomb, G. J.

Control Release, 2009, 133, 90-95. 29. Ullrich, O.; Reinheckel, T.; Sitte, N.; Hass, R.; Grune, T.; Davies, K. J. Proc. Natl.

Acad. Sci. USA 1999, 96, 6223-6228. 30. Xiong, Y.; Yang, H. J.; Feng, J.; Shi, Z. L.; Wu, L. D. J. Int. Med. Res., 2009, 37,

409-416. 31. Peckham, S.; Awofeso, N. Water Fluoridation: A Critical Review of the

Physiological Effects of Ingested Fluoride as a Public Health Intervention. The Scientific World Journal, Volume 2014, Article ID 293019.

32. Riggs, B. L.; Hodgson, S. F.; O’Fallon, M. W. et al N Engl. J. Med. 1990, 322,

802–809. 33. Boackle, R. J.; Suddick, R. P. Salivary proteins and oral health.; Menaker L., Ed.; Hagerstown: Harper & Row, 1980, p113-131. 34. Mérian, J.; Gravier, J.; Navarro, F.; Texier, I. Molecules 2012, 17, 5564-5591.

92

35. Jin, H.; Mangun,C. L.; Griffin, A. S.; Moore, J. S.; Sottos, N. R.; White, S. R. Adv. Mater. 2014, 26, 282-287.

36. Yue, H. B.; Fernández-Blázquez, J. P.; Beneito, D. F.; Vilatela, J. J. J. Mater.

Chem. A, 2014, 2, 3881-3887. 37. White, S. R.; Sottos, N. R.; Geubelle, P.H., Moore, J.S.; Kessler, M. R.; Sriram, S.

R.; Brown, E. N.; Viswanathan, S. Nature, 2001, 409, 794-797. 38. Kessler, M. R.; Sottos, N. R.; White, S. R. Composites: Part A, 2003, 34, 743-753. 39. Jones, A. S.; Rule, J. D.; Moore, J. S.; White, S. R.; Sottos, N. R. Chem. Mater.,

2006, 18, 1312-1317. 40. Rule, J. D.; Sottos, N. R.; White, S. R. Polymer, 2007, 48, 3520-29. 41. Wu, D. Y.; Meure, S.; Solomon, D. Prog. Polym. Sci., 2008, 33, 479-522. 42. Nosonovsky, M.; Amano, R.; Lucci, J. M.; Rohatgi, P. K. Phys. Chem. Chem.

Phys., 2009, 11, 9530-9536. 43. Wojtecki, R. J.; Meador, M. A.; Rowan, S. J. Nature Mater., 2011, 10, 14-27. 44. Pang, J. W. C.; Bond, I. P. Compos. Sci. Technol., 2005, 65, 1791-1799. 45. Pang, J. W. C.; Bond, I. P. Composites: Part A, 2005, 36, 183-188. 46. Kolmakov, G. V.; Revanur, R.; Tangirala, R.; Emrick, T.; Russell, T. P.; Crosby, A.

J.; Balazs, A. C. ACS Nano, 2010, 4, 1115-1123. 47. Kratz, K.; Narasimhan, A.; Tangirala, R.; Moon, S. C.; Revanur, R.; Kundu, S.;

Kim, H. S.; Crosby, A. J.; Russell, T. P.; Emrick, T.; Kolmakov, G.; Balazs, A. C. Nature Nanotech., 2012, 7, 87-90.

48. Middleton, G. V.; Can. J. Earth Sci., 1966, 3, 523-546. 49. Anderson, R. Y.; Kirkland, D. W. Geology, 1980, 8, 66-69. 50. Devantier, B. R.; Larock, B. E. Int. J. Numer. Meth. Fl., 1986, 6, 241-253. 51. Linden P. F.; Simpson. J. E. J. Fluid Mech., 1986, 172, 481-497. 52. Alavian, V. J. Hydraul. Eng. 1986, 112, 27-42. 53. Lide D. R.; CRC Handbook of Chemistry and Physics, 87th Ed., 8-54. 54. Ben-Asuly, A.; Tzur, E.; Diesendruck, C. E.; Sigalov, M.; Goldberg, I.; Lemcoff, N.

G. Organometallics, 2008, 27, 811-813.

93

55. Monsaert, S.; Vila, A. L.; Drozdzak, R.; Van Der Voort, P.; Verpoort, F. Chem. Soc. Rev. 2009, 38, 3360-3372.

56. Drozdzak, R.; Nishioka, N.; Recher, G.; Verpoort, F. Macromol. Symp. 2010, 293,

1-4. 57. Thomas, R. M.; Fedorov, A.; Keitz, B. K.; Grubbs, R. H. Organometallics, 2011,

30, 6713-6717. 58. Ernst, C.; Elsner, C.; Prager, A.; Scheibitz, B; Buchmeiser, M. R. J. Appl. Polym.

Sci., 2011, 121, 2551-2558. 59. Jeong, Y.; Duncan, B.; Park, M. H.; Kim, C.; Rotello, V. M. Chem. Commun.,

2011, 47, 12077-12079. 60. Dragutan, V.; Dragutan, I.; Dimonie, M., A Selective Route for Synthesis of Linear

Polydicyclopentadiene. In Green Metathesis Chemistry, Dragutan, V.; Demonceau, A.; Dragutan, I.; Finkelshtein, E., Eds. Springer Netherlands: 2010; pp 369-381.

61. Mecozzi, M. et al Analyst 2001. 126, 144-146. 62. Yadav, V., Freedman, J. D., Grinstaff, M.; Sen, A. Angew. Chem., Int. Ed., 2013,

52, 10997-11001. 63. Yadav, V.; Pavlick, R. A.; Meckler, S.; Sen. A. Triggered Detection and Polymer

Deposition: Towards the Repair of Microcracks, Chem. Mater., 2014 26, 4647-4652

94

Chapter 4

A Self-Powered Polymeric Material that Responds Autonomously and Continuously to Fleeting Stimuli

4.1 Introduction

This chapter describes the design of a polymeric material that performs a

macroscopic function continuously once exposed to a specific stimulus, even when the

stimulus is no longer present. The material is self-powered, requires no reagents from

solution, operates autonomously, and converts chemical energy into a mechanical

response. Thus, the design offers a combination of attributes that are not available

currently in smart polymeric materials.1-3

4.2. Experimental Design

To demonstrate these capabilities, modified TentaGel microsphere were

prepared that are capable of initiating pumping of the fluid surrounding the microsphere

(i.e., the macroscopic response),4-11 even after the applied signal (UV light, a model

stimulus) had been removed (Figure 4-1). The continuous pumping response was made

possible via networks of reactions on the surface of, and within, the microsphere.12 The

chemical reactions enabled not only selective responses to UV light, but also a means to

propagate the response, even when the light was removed, which was a level of control

that is analogous; in the regulated behavior, to externally-controlled polymerization

reactions.13, 14 The consequence of this network of reactions was the continuous

95

production of small molecule products that generate a gradient as they diffuse away

from the microsphere (Figure 4-1). This gradient was then able to induce movement of

the surrounding fluid towards the microsphere (i.e., the pumping response).

Figure 4-1. Schematic depiction of polymer microsphere pump that induces the

movement of fluid that surrounds the pump in response to a specific stimulus, even after

the stimulus has been removed. The blue arrows represent the direction of fluid

movement, and the sizes of the arrows illustrate an approximation of the relative

magnitude of fluid flow when the signal is present or absent. When the UV light is off, a

self-propagating reaction enables the microsphere to continue generating a

concentration gradient of products that drive the pumping response. The signal

transduction reagents (fluoride ion) translate the first reaction with UV light to initiation of

the self-propagating reaction. The byproduct of the reactions (3) is yellow/orange in color

and, thus, turns the microsphere from colorless to yellow to orange over the course of

the pumping response.

96

On the molecular level, the ability of the material to continue performing its

function arose from a self-propagating autoinductive reaction15 that utilized reagents

incorporated directly onto the polymer. This self-propagating autoinductive reaction

amplified the gradient of products, which was similar to the behavior of pumps based on

signal-induced depolymerization reactions.5

For the purpose of this chapter, the activity based reagent, 2-(4-((4-

(Difluoromethyl)-3-methoxyphenylcarbamoyloxy)methyl)-3-nitrophenoxy)acetic acid, will

be referred to as reagent 1, the autoinductive reagent, 3-(3-(tert-Butyldimethylsilyloxy)-4-

((4-(difluoromethyl)-3-methoxyphenylcarbamoyloxy) methyl) phenyl)propanoic acid, as

reagent 2 and 4-amino-2-methoxybenzaldehyde as 3. Both reagents 1 and 2 are grafted

on the microsphere in varying proportions (Figure 4-2a). The activity-based detection

reagent16 detects a pre-defined stimulus by reaction of the stimulus with specific

functionality in the detection reagent. In principle, this activity-based detection reagent

could be exchanged with other functionality to create modular materials that respond to

a variety of fleeting stimuli.16

Reagent 1 is designed to respond to 254 nm to 365 nm light17,18 to release two

equivalents of fluoride, carbon dioxide, 4-aminobenzaldehyde derivative 3, and two

protons (that exist predominantly as as pyridinium ions upon reaction with the solvent

constituent pyridine) (Figure 4-2b). The released fluoride is a signal transduction

reagent that translates the detection event into initiation of a self-propagating reaction.

This self-propagating reaction involves reagent 2, which responds to one equivalent of

fluoride and releases two additional equivalents of fluoride as well as more of 3 and a

proton (exist as pyridinium ions).15,19 The released fluoride is then available to react with

additional equivalents of 2 to continue amplifying the quantity of fluoride, 3, and protons

(exist as pyridinium ions) in the microsphere until all of 2 has been consumed (Figure 4-

97

2c). This self-propagating autoinductive reaction enables the continuous response of the

material, even when the stimulus is removed, since the self-propagating reaction

generates and amplifies an ion gradient and a small molecule gradient (Figures 4-2)

that induce movement of the surrounding fluid.4-11

4.3. Results and Discussion

Both the activity-based detection reagent (1) and the self-propagating reagent (2)

were grafted to the polymeric Tentagel bead (composed of polystyrene and polyethylene

network) bead in a presumed 1:1 ratio. A 1:1 ratio of reagents 1 and 2 was chosen to

ensure that the beads contained sufficient quantities of both reagents to sustain the two

halves of the reaction network, particularly since ~40%−60% of the reactions occur on

the surface of the microsphere12 where diffusion could interfere with the signal

transduction and autoinductive processes. Two control microspheres were prepared as

well using analogous chemistry, one containing 100% of 1 and the other 100% of 2.20

98

Figure 4-2. Structures and reactions of reagents 1 and 2 that are grafted onto a 300 µm-

diameter TentaGel microsphere. (a) A microsphere that contains a 1:1 ratio of reagents

1 and 2. (b) Exposure of this microsphere to UV light causes the activity-based detection

reagent (1) to release fluoride, compound 3, and protons (exist as pyridinium ions). (c)

The released fluoride then reacts with 2 to initiate a self-propagating reaction that

amplifies fluoride, 3, and protons (exist as pyridinium ions). The gradient of these small

molecules causes fluid movement around the microsphere (i.e., pumping). The notation

“n” refers to the number of cycles of the autoinductive reaction in (c).

UV ON

UV OFF

99

4.3.1 Colorimetric Analysis

Before testing whether the microsphere containing both 1 and 2 were capable of

“remembering” its pre-defined stimulus, control experiments were first conducted, to

determine whether the microspheres that contain 2 were capable of supporting the

autoinductive, self-propagating reaction. Treatment of the colorless microspheres with 2

mM cesium fluoride in 10:4:1 isopropanol-water-pyridine for 3 h yielded dark

yellow/orange microspheres (the color of 3) (Figures 4-3a, b). The color change was

caused by 3 being released into solution, a fact that was confirmed by injecting an

aliquot of the reaction mixture into an HPLC connected to a mass spectrometer

(LCMS).20 The time-dependent intensity of the yellow/orange colorimetric response was

quantified by photographing the beads over time (Figure 4-3b) and using image

processing software to measure the intensity of the color in the digital images.15, 19 The

change in color over time reflected the extent of completion of the autoinductive reaction,

and provides a visual indication that the material is performing its function (i.e., pumping,

see below). The resulting sigmoidal response curves based on the intensity of the color

were consistent with an autoinductive reaction (Figure 4-3c).15, 19

The autoinductive behavior of 2, was further verified, by exposing microspheres

containing only 2 to substoichiometric quantities of added fluoride relative to the loading

level of the TentaGel microsphere. Regardless of the quantity of fluoride used to initiate

the autoinductive reaction, all microspheres provided equal levels of color over time,

which is an expected result for a self-propagating reaction. As expected for an

autoinductive reaction, the time to reach completion when the microspheres were

exposed to lower quantities of fluoride was longer than when the microspheres were

exposed to higher quantities of fluoride (Figure 4-3c).

100

Control experiment was also performed on a microsphere containing only 1. In

this experiment, the ability of 1 to respond to UV light when the microsphere was

exposed to 365 nm light for 40 min was tested. As expected, the microspheres turned a

bright yellow/orange color, which was indicative of formation of 4-amino-2-

methoxybenzaldehyde (3), a fact that was verified by LCMS analysis.20

Figure 4-3. Colorimetric response of a TentaGel microsphere that contained 100% of

reagent 2. (a) The procedure for testing the autoinductive, self-propagating reaction that

is mediated by 2. The product of the autoinductive reaction is 3 (Figure 4-2c), which

turns the microspheres a yellow/orange color (b). (c) This color reflects the extent of the

autoinductive reaction,15,19 and can be quantified by photographing the microspheres

and using image processing software to measure the intensity of color. Exposure of the

microspheres to substoichiometric quantities of fluoride (relative to the loading level of

the microspheres) reveals sigmoidal kinetics characteristic of autoinductive

reactions.15,19 Note that the scale of the x-axis changes after the break.

101

Overall, the two sets of control experiments performed on microspheres

containing only 1 or 2 demonstrate that reagents 1 and 2 were capable of performing

their individual functions on a molecular level when grafted to the microspheres.

Whether they were capable of imparting a macroscopic response to the microspheres

was next tested by measuring whether they could induce pumping of the surrounding

fluid. The experiment involved placing the microspheres on a glass slide that was

immersed in 10:4:1 isopropanol-water-pyridine in a closed system and then exposing the

microspheres to UV light.

4.3.2. Stimuli Responsive Pumping Behaviour

The following experimental set-up was designed to the study the pumping

response of the tentagel beads. Microspheres (either functionalized with a 1:1 ratio of

reagents 1 and 2, just reagent 1, or just reagent 2) were exposed to a solution of i-

PrOH/H2O/pyridine (10:4:1 respectively) for approximately 1 h, which allowed them to

swell to a stable size. One microsphere was removed from the solution, placed on a

glass microscope slide and covered with a hybridization chamber. The hybridization

chamber was filled with a solution of 2 µm diameter, amine-functionalized polystyrene

tracer particles suspended in i-PrOH/H2O/pyridine (10:4:1 respectively). To ensure that

the maximum sustainable directional movement of the fluid was measured, the

measurements were made after ~19 min of exposure to UV light. Also, the microspheres

did not move during the pumping experiments. Their density was higher than the

surrounding solvent, therefore they remained on the surface of the glass slide in the test

chamber. The speeds of the tracer particles were measured in the x-direction (the

sphere was located to either the left- or right-hand side of the observation window in all

102

experiments) using the ‘tracker’ software provided by Open Source Physics. The videos

were recorded at 60 fps and the particles were tracked for the last 30 s of each 2 min

observation window.

Figure 4-4. Average pumping speeds caused by TentaGel microspheres exposed to

365 nm light. (a) The pumping action can be switched on and off for a microsphere

functionalized with 1 only (blue data), whereas no pumping was observed for

microspheres functionalized with only 2 (orange data). In contrast, the pumping speed

could be varied (but not turned off) for microspheres functionalized with both 1 and 2 by

turning on and off the UV light (black data). The pumping speeds reflect the averages of

measurements acquired over 30 s intervals that span the length of the data bars. (b)

Continuous pumping also is possible using microspheres that are functionalized with 1

and 2 once the microspheres are exposed to UV light for 20 min. For both (a) and (b),

the average pumping speeds were obtained by tracking the distance that 30 tracer

beads traveled over time.

103

When microspheres containing 100% of 1 were exposed to 365 nm light, the

surrounding fluid displayed directional movement towards the microsphere, as reflected

by the movement of 2 µm-diameter amine-functionalized polystyrene tracer particles

towards the microsphere with an average speed of 4.8 ± 0.3 µm/s (blue data, Figure 4-

4a; Supporting Video 4-1). This directional fluid movement (i.e., pumping) could be

turned on and off repeatedly by switching on and off the UV light exposed to the

microspheres. In contrast, when microspheres containing 100% of 2 were exposed to

365 nm light for 20 min, no directional fluid movement was observed and no switching

behavior was established (orange data, Figure 4-4a; Supporting Video 4-2).

4.3.3 Memory based Pumping in the Absence of Stimuli

Clearly only 1, as anticipated, was capable of responding to the stimulus (UV

light) and inducing a macroscopic response from the materials. However, 1 alone did not

enable the microsphere to “remember” the applied stimulus when the UV light was

removed. For that capability, 2 was introduced to the microsphere to complete the

designed network of reactions (Figures 4-2a). First a baseline pumping speed for the

microspheres containing a 1:1 ratio of 1 and 2 when exposed to 365 nm light for 20 min

was established. In this experiment, directional fluid movement was observed once

again, albeit at a reduced speed of 3.7 ± 0.5 µm/s compared with the microsphere that

contained 100% of 1 (4.8 ± 0.3 µm/s). This reduced pumping speed was to be expected

since the microspheres containing approximately equal quantities of 1 and 2 had less of

1 to react with the UV light than the microspheres containing 100% of 1.

Next, the UV light was cycled on and off, which yielded persistent fluid pumping

during the off cycles for the microspheres containing 1 and 2 (unlike the microsphere

104

containing only 100% of 1), with a pumping speed of 0.28 ± 0.07 µm/s, when the light

was off (black data, Figure 4-4a; Table 4-1; Supporting Video 4-3, 4-4). This 0.28 ±

0.07 µm/s pumping speed arose from the autoinductive reaction mediated by 2, which

was a slower chemical reaction (Figure 4-2c) than the direct photochemical reaction of 1

when the UV light was turned on.17, 18

Perhaps more revealing about the behavior of the microspheres containing 1 and

2 was their ability to respond continuously when they are exposed only once to the

stimulus, rather than periodically (Figure 4-4b; Table 4-2). Specifically, when

microspheres that contained a 1:1 ratio of 1 and 2 were exposed to 365 nm light for 20

min, and then the light was removed, the pumping speed dropped from 3.2 ± 0.3 µm/s to

0.33 ± 0.07 µm/s. This speed was maintained for ~8 min, at which point the pumping

speed began to decrease, likely as a result of consumption of 2 in the microspheres.

105

Table 4-1. Average pumping speeds caused by Microspheres exposed to on and off

cycle of 365 nm light. Speeds and colors correspond to the data represented in Figure 4-4a.

Microsphere Time (min) Speed (µm/s) Standard Deviation

Microsphere functionalized with

reagent 1

19.5-20 4.81 0.32

21.5-22 0.02 0.01

23.5-24 4.41 0.48

25.5-26 0.03 0.01

27.5-28 4.28 0.58

29.5-30 0.02 0.01

Microsphere functionalized with

reagent 2

19.5-20 0.02 0.01

21.5-22 0.02 0.01

23.5-24 0.04 0.02

25.5-26 0.02 0.01

27.5-28 0.02 0.01

29.5-30 0.03 0.02

Microsphere functionalized with a 1:1

ratio of reagent 1 & 2

19.5-20 3.71 0.48

21.5-22 0.28 0.08

23.5-24 3.20 0.66

25.5-26 0.32 0.07

27.5-28 3.45 0.29

29.5-30 0.25 0.07

106

Table 4-2: Average tracer particle speeeds caused by TentaGel microspheres exposed

20 min of continous UV exposure. Speeds correspond to the data represented in Figure 4-4b.

These combined data demonstrated that pumping speeds could be modulated

either by periodic exposure of the material to its pre-defined stimulus, or through fleeting

exposure. More importantly, these results showed that the microspheres continued

pumping even when the signal was removed, which was an unusual capability in the

context of stimuli-responsive materials, and one that was achieved by building into the

material the capacity for a self-propagating reaction.

The motility mechanism discussed thus far in this chapter is based on non-

electrolyte self-diffusiophoresis. The presence of the organic non-polar solvents (i-

PrOH/H2O/pyridine) prevent the formation of an electrostatinc double layer, thereby

generating motion primarily due to a gradient of small molecules.

Microsphere Time (min) Speed (µm/s) Standard Deviation

Microsphere functionalized with a

1:1 ratio of reagent 1 & 2

19.5-20 3.18 0.34

21.5-22 0.33 0.07

23.5-24 0.33 0.08

25.5-26 0.33 0.08

27.5-28 0.30 0.08

29.5-30 0.24 0.07

107

4.4 Difusiophoretic Pumping- Scavenger Design

The aqueous counterpart of the experiments, that would sustain substantial

Debye lengths and thereby allow for diffusiophoretic pumping, was also evaluated. Since

organic solvents were essential for the occurrence of the auto-inductive reaction

(reagent 2) (Figure 4-2c), only microspheres functionalized with reagent 1 were

analyzed. The activity based reagent responded to UV light in an aqueous environment,

generating two ions, two fluoride ions and two protons (absence of pyridine in solution

makes these protons freely available) (Figure 4-2b). The difference in the diffusivity of

the two ions, led to a local separation of charges thereby creating a local electric field

pointing inwards, towards the microsphere. Following a similar experimental set-up as

described in section 4.3.2, the microsphere functionalized with reagent 1 was triggered

using UV light and the induced movement to the surrounding charged tracer particles

was noted. Unlike the previous case with organic solvents, a directional response based

on the charge of the tracers was observed, as expected of a diffusiophoretic mechanism.

Negatively charged amine functionalized polystyrene particles were observed to move

away from the microspheres creating exclusion zones (Figure 4-5a; Supporting Video

4-5), while positively charged carboxyl functionalized polystyrene particles moved

towards and inside the microsphere (Figure 4-5b; Supporting Video 4-6) Moreover, the

positively charged particles moving towards the microsphere were observed to get

trapped within its body. The microsphere containing the trapped particles was frozen in a

sugar solution and cryotomed to expose the interior, and was imaged on an

Environmental scanning electron microscope (Figure 4-6). The images reveal a versatile

scavenger design that is capable of scavenging, in its present design, any positively

charged species.

108

Figure 4-5. Optical microscope images of the microsphere functionalized with reagent 1

alone, triggered with UV light that induces motion to the surrounding (a) negatively

charged polystyrene particles at 5X magnification. A zone of exclusion can clearly be

seen around the microsphere where tracer particles have been pushed away. (b)

Positively charged particles were pulled in towards the microsphere, eventually getting

trapped inside at 20X magnification. The imaged polystyrene particles are each 2 µm in

diameter.

109

The hypothesis was tested out using charged biological species like bacteria.

Preliminary test were carried out with live E. coli bacteria that moved away from the

microspheres. The design of the microsphere allows the switching of the direction of the

electric field to trap any charged moieties within the sphere, creating a versatile

scavenger.

Figure 4-6: (a) Electron microscope images of the polystyrene particles trapped within

the microsphere. (b) Shows a zoomed in image of the chipped part confirming the

particles to be trapped inside the permeable body and not just on the surface of the

microsphere. Scale bar is 50 µm.

110

4.5 Conclusion

In conclusion, this chapter describes a new approach for creating smart, stimuli-

responsive materials that are capable of remembering when they are exposed to a

stimulus, even when the stimulus is no longer present. The proof-of-concept application

to demonstrated in this chapter is a plastic microsphere-based fluidic pump, which is

capable of pumping the fluid surrounding the microsphere in response to UV light, as

well as continuous pumping even when the UV light is removed. The microspheres

provide the continuous pumping without using reagents supplied in solution and without

intervention by the user. Moreover, the pumping speed can be altered in magnitude if

the signal is present or absent, and the microsphere reveal that it is responding to the

signal by turning a yellow/orange color (Figure 4-3), the intensity of which loosely

correlates with the remaining lifetime of the pump (i.e., the quantity of 2 that has been

consumed). This level of autonomous function extends beyond existing smart

materials,1-3 where closely related examples include materials that have a memory for

their original shape,21-25 or have the ability to perpetuate an oscillatory response to an

applied signal using reagents from the surroundings.26-29

The particular application of the above described fluidic pump may prove useful

in a variety of contexts, ranging from collecting and concentrating select agents, to

biological sensing, to directing flow in microfluidic devices in response to specific signals.

Efforts are underway to expand the scope of the autoinductive chemistry that enables

the continuous responses,30-34 and to create other types of stimuli-responsive pumps

with improved response rates and duration of pumping. Parameters to explore in this

context include varying the shape,35 size, loading capacity, and porosity of the polymeric

material, as well as the ratio of reagents grafted to the polymer. More broadly, the

111

chemical concepts described in this chapter may enable the preparation of other types of

stimuli-responsive polymeric materials that perform continuous, macroscopic operations,

other than pumping, when exposed to a fleeting signal.

4.6 Acknowledgements

The author would like to thank Dr. Matthew Baker and Prof. Scott Phillips for

synthesizing and characterizing reagent 1 and 2, and carrying out the colorimetric

analysis.

112

4.7 References

Parts of this chapter have been adapted from “Baker, M. S.; Yadav, V.; Sen, A.; Phillips, S. T. Angew. Chem., Int. Ed. 2013, 52, 10295–10299”

1. Spruell, J. M.; Hawker, C. J. Chem. Sci. 2011, 2, 18–26. 2. Epstein, I. R.; Vanag, V. K.; Balazs, A. C.; Kuksenok, O.; Dayal, P.; Bhattacharya,

A. Acc. Chem. Res. 2012, 45, 2160–2168. 3. Stuart, M. A. C.; Huck, W. T. S.; Genzer, J.; Müller, M.; Ober, C.; Stamm, M.;

Sukhorukov, G. B.; Szleifer, I.; Tsukruk, V. V.; Urban, M.; Winnik, F.; Zauscher, S.; Luzinov, I.; Minko, S. Nature Mater. 2010, 9, 101–113.

4. Laser, D. J.; Santiago, J. G. J. Micromech. Microeng. 2004, 14, R35–R64. 5. Zhang, H.; Yeung, K.; Robbins, J. S.; Pavlick, R. A.; Wu, M.; Liu, R.; Sen, A.;

Phillips, S. T. Angew. Chem., Int. Ed. 2012, 51, 2400–2404. 6. Chang, S. T.; Beaumont, E.; Petsev, D. N.; Velev, O. D. Lab Chip 2008, 8, 117–

124. 7. Chang, S. T.; Paunov, V. N.; Petsev, D. N.; Velev, O. D. Nature Mater. 2007, 6,

235–240. 8. Kline, T. R.; Paxton, W. F.; Wang, Y.; Velegol, D.; Mallouk, T. E.; Sen, A. J. Am.

Chem. Soc. 2005, 127, 17150–17151. 9. Paxton, W. F.; Baker, P. T.; Kline, T. R.; Wang, Y.; Mallouk, T. E.; Sen, A. J. Am.

Chem. Soc. 2006, 128, 14881–14888. 10. Hong, Y.; Diaz, M.; Córdova-Figueroa, U. M.; Sen, A. Adv. Funct. Mater. 2010, 20,

1568–1576. 11. Yadav, V.; Zhang, H.; Pavlick, R.; Sen, A. J. Am. Chem. Soc. 2012, 134, 15688–

15691. 12. McAlpine, S. R.; Schreiber, S. L. Chem. Eur. J. 1999, 5, 3528–3532. 13. Leibfarth, F. A.; Mattson, K. M.; Fors, B. P.; Collins, H. A.; Hawker, C. J. Angew.

Chem., Int. Ed. 2013, 52, 199–210. 14. Neilson, B. M.; Bielawski, C. W. Chem. Commun. 2013, 49, 5453–5455. 15. Baker, M. S.; Phillips, S. T. J. Am. Chem. Soc. 2011, 133, 5170–5173. 16. Cho, D. G.; Sessler, J. L. Chem. Soc. Rev. 2009, 38, 1647–1662.

113

17. Il’ichev, Y. V.; Schwörer, M. A. Wirz, J. J. Am. Chem. Soc. 2004, 126, 4581–4595. 18. Bochet, C. G. J. Chem. Soc., Perkin Trans. I 2002, 125–142. 19. Baker, M. S.; Phillips, S. T. Org. Biomol. Chem. 2012, 10, 3595–3599. 20. Baker, M. S.; Yadav, V.; Sen, A.; Phillips, S. T. Angew. Chem., Int. Ed. 2013,

52, 10295–10299. 21. Lendlein, A.; Jiang, H.; Jünger, O.; Langer, R. Nature 2005, 434, 879–882. 22. Sun, L.; Huang, W. M. Ding, Z.; Zhao, Y.; Wang, C. C.; Purnawali, H.; Tang, C.

Mater. Des. 2012, 33, 577–640. 23. Liu, Y.; Lv, H.; Lan, X.; Leng, J.; Du, S. Compos. Sci. Technol. 2009, 69, 2064–

2068. 24. Luo, X.; Mather, P.T. ACS Macro Lett. 2013, 2, 152–156. 25. Lendlein, A.; Kelch, S. Angew. Chem., Int. Ed. 2002, 41, 2034–2057. 26. Yoshida, R. Adv. Mater. 2010, 22, 3463–3483. 27. Kaminaga, A.; Vanag, V. K.; Epstein, I. R. Angew. Chem., Int. Ed. 2006, 45,

3087–3089. 28. Shinohara, S.; Seki, T.; Sakai, T.; Yoshida, R.; Takeoka, Y. Angew. Chem., Int.

Ed. 2008, 47, 9039–9043. 29. Kuhnert, L. Nature, 1986, 319, 393−394 30. Yeung, K.; Schmid, K. M.; Phillips, S. T. Chem. Commun. 2013, 49, 394–396. 31. Mohapatra, H.; Schmid, K. M.; Phillips, S. T. Chem. Commun. 2012, 48, 3018–

3020. 32. Perry-Feigenbaum, R.; Sella, E.; Shabat, D. Chem. Eur. J. 2011, 17, 12123–

12128. 33. Sella, E.; Weinstain, R.; Erez, R.; Burns, N. Z.; Baran, P. S.; Shabat, D. Chem.

Commun. 2010, 46, 6575–6577. 34. Scrimin, P.; Prins, L. J. Chem. Soc. Rev. 2011, 40, 4488–4505. 35. Jang, S. G.; Audus, D. J.; Klinger, D.; Krogstad, D. V.; Kim, B. J.; Cameron, A.;

Kim, S. W.; Delaney, K. T.; Hur, S. M.; Killops, K. L.; Fredrickson, G. H.; Kramer, E. J.; Hawker, C. J. J. Am. Chem. Soc. 2013, 135, 6649–6657.

114

Chapter 5

Substrate-driven Chemotatic Assembly in Enzyme Cascades

“Aerodynamically, the bumble bee shouldn't be able to fly, but the bumble bee doesn't know it so it goes on flying anyway.”

-Mary Kay Ash

5.1 Introduction

A motor is a machine that consumes energy in some form and coverts it into

mechanical work. Motion is an inextricable part of life and nature has employed several

different chemically-powered motors to sustain life. Some examples of molecular motors

include myosins, dyneins and kinesins, which are known as cytoplasmic motors. These

utilize ATP as their energy source and move on tracks (e.g., microtubules) to achieve

directionality. ATP hydrolysis causes a conformational change that is further amplified

and translated into mobility. In this respect, synthetic motors and pumps are similar to

their biological counterparts. Both expend energy; ATP hydrolysis in one case and

chemical, electrical, or magnetic in the other, and convert it into mechanical motion.

Resemblance can also be seen in the working mechanisms of the two: specifically,

proton gradients cause transport across membranes in living systems, and are also

responsible for the propulsion of bimetallic nanorods and fluid pumping in bimetallic

pumps. A recent study reveals that almost all enzymes, and not just the ones involved in

cytoplasmic motors, are capable of exhibiting motion. Further, they exhibit a rudimentary

form of chemotaxis in the presence of their substrate gradient. This chapter applies this

latest finding into solving the mysteries of enzymatic cascades that have baffled

biologists and enzymologists for a long time.

115

5.2 Motivation

Enzymatic catalysis is essential to cell survival. The interaction between

enzymes in living cells is an area of active research. In many instances, enzymes that

participate in reaction cascades have been shown to assemble in response to the

presence of the initial substrate to facilitate substrate channeling.1, 2 However, what

triggers the enzymes to form metabolons remains an open question. While the

mechanism that brings together enzymes in a cascade to promote substrate channeling

remain unknown, metabolon formation through non-covalent interactions has been

suggested but not demonstrated. The diffusivity of enzymes has been shown to increase

in the presence of their respective substrates in a concentration-dependent manner3, 4

and directional movement of individual enzymes in response to their respective

substrate gradients has been reported. Furthermore, enzymes are known to diffuse over

distances of microns within seconds in living cells.5 This chapter presents an

investigation into a well-known enzyme cascade- glycolysis and the possibility of a

chemotactic response to be guiding the metabolon formation.

5.3 Experimental Design

The sequential catalysis by the first four enzymes of the glycolytic cascade

(Figure 5-1)7, hexokinase (HK), phosphoglucose isomerase (Iso), phosphofructokinase

(PFK) and aldolase (Ald), was examined. A flow-based microfluidic channel was

designed to study the chemotactic enzyme migrations (Figure 5-2). The first and the last

enzyme of this four-step cascade, HK and Ald, were tagged with an amine reactive

(ex/em: 493/518) and a thiol reactive (ex/em: 638/658) Dylight dye, respectively, in order

116

to monitor their movement by confocal fluorescence microscopy. A low dye:enzyme ratio

at 0.4:1 in case of HK and 0.6:1 in case of Ald was maintained in order to not

compromise the enzyme activity. Fluorescence correlation spectroscopy (FCS)

measurements were performed to study the diffusion enhancements for HK and Ald as a

function of their respective substrate concentrations.

7Figure 5-1. The glycolysis cycle. \The first four enzymes, hexokinase (HK),

phosphogluco isomerase (Iso), phosphofructokinase (PFK) and aldolase (Ald) were

examined for their ability to undergo chemotactic assembly.

117

5.3.1 Microfluidic device fabrication

The microfluidic device was cast in polydimethylsiloxane (PDMS, Sylgard 184)

using standard soft lithography protocols.11 A 100-μm deep master pattern was created

on a silicon wafer (Silicon Quest) using SPR-220 resist (Microposit) and deep reactive

ion etching (Alcatel). The master was exposed to 1H,1H,2H,2H-perfluorooctyl-

trichlorosilane to minimize adhesion of PDMS during the peeling step. After the PDMS

was peeled off, the inlet and outlet regions were opened by drilling, and the device was

sealed to a glass coverslip. Fluid flow through the channel was controlled by a syringe

pump, connected by polyethylene tubing to the device.

Figure 5-2. Photo-lithographically fabricated flow based microfluidic gradient generator

for studying enzyme chemotaxis. The length of the channels is either 20 or 40 mm, width

360 μm, and the height is 100 μm. Considering laminar flow, the width of each channel

is 120 µm. Fluorescence intensities were analyzed along a vertical line as shown in the

figure leaving off 20 µm next to the sidewalls.

118

Table 5-1. Distance from the start of the channel converted into time spent inside the

channel for specified channel geometry.

5.3.2 Fluorescent tagging of HK and Ald

Hexokinase (from Saccharomyces cerevisiae) was tagged with an amine-

reactive dye, Dylight 488 (ex/em: 493/518). Hexokinase (44 μM; 15% protein) was

reacted with a threefold excess of the fluorescent probe and 1 M mannose in 50 mM

Hepes (pH 7.0) at 4°C for 2–4 h on a rotator. Aldolase (from rabbit muscle) was labeled

with a thiol-reactive dye Dylight 633 (ex/em: 638/658). Labeling of Aldolase (75 mM;

80% protein) was carried out with two fold excess of the fluorescent dye and 1 mM

EDTA on a rotator at 4°C for 2–3 h in 50 mM HEPES buffer (pH 7.4). The enzyme–dye

complexes were further purified using a Sephadex G-25 (GE Healthcare) size exclusion

column with 50 mM hepes buffer (pH 7.4) to reduce free-dye concentration. The number

Flow rate (µL/h) Distance (mm) Time (s)

50

10 8.6

20 17.3

30 25.9

40 34.6

30

10 14.4

20 28.8

30 43.2

40 57.6

119

of dye molecules per HK or ALD enzyme molecule was ∼0.4 or 0.6, respectively, as

quantified using UV–Vis spectroscopy.

Hexokinase activity before and after attachment of the fluorophore was

measured spectrophotometrically by coupling with glucose-6-phosphate dehydrogenase

and following the reduction of NADP+ at 340 nm. An assay mixture 1 mL in total volume

contained 1 mM glucose, 2 mM ATP, 10 mM MgCl2, 50 mM HEPES (pH 7.4), 0.5 mM

NADP+, 2 units glucose-6-phosphate dehydrogenase, and 5 nM hexokinase. All assays

were performed at 25 °C. The enzymatic activity was not significantly altered by the

attachment of the fluorophore.

Aldolase activity before and after attachment of the fluorophore was also

measured spectrophotometircally by coupling with α-glycerophosphate

dehydrogenase/triosephosphate isomerase and following the oxidation of NADH at 340

nm. An assay mixture 1 mL in total volume contained 2 mM fructose-1,6-disphosphate,

50 mM HEPES (pH 7.4), 0.1 mM NADH, 1.5 units α-glycerophosphate

dehydrogenase/triosephosphate isomerase (based on GDH units), and 50 nM aldolase.

All assays were performed at 25 °C. The enzymatic activity was not significantly altered

by the attachment of the fluorophore.

120

5.3.3 Fluorescence Correlation Spectroscopy

Spectroscopy measurements were performed on a custom-built microscope-

based optical setup, described previously.13 Briefly, a PicoTRAIN laser (High-Q Laser),

delivered 5.4 ps pulses of 532 nm light at 80 MHz frequency. This light was guided

through a fiber optic cable, expanded and directed through an IX-71 microscope

(Olympus), with an Olympus 60×/1.2-NA water-immersion objective. Emitted fluorescent

light from the sample was passed through a dichroic beam splitter (Z520RDC-SP-POL,

Chroma Technology) and focused onto a 50 μm, 0.22-NA optical fiber (Thorlabs), which

acted as a confocal pinhole. The signal from the photomultiplier tube was routed to a

preamplifier (HFAC-26) and then to a time-correlated single-photon counting (TCSPC)

board (SPC-630, Becker and Hickl). The sample was positioned with a high-resolution 3-

D piezoelectric stage (NanoView, Mad City Laboratories).

Fluorescent molecules moving into and out of the diffraction-limited observation

volume induce bursts in fluorescence collected in first-in, first-out mode by the TCSPC

board, which was incorporated in the instrument. Fluctuations in fluorescence intensity

from the diffusion of molecules were autocorrelated and fit by a multicomponent 3D

model to determine the diffusion coefficients of individual species. Fluctuations in

fluorescence intensity from the diffusion of molecules were autocorrelated and fit by a

single component 3D model to determine the diffusion coefficients of individual species.

Contributions to the autocorrelation curve from fluctuations in molecular fluorescence

intensity due to fast processes such as triplet state excitation occur were minimal.

Nevertheless, when the shape of the autocorrelation curve indicated the need to include

the triplet state in the fit, the alternative Equation 5.2 was used.

121

𝑮𝑮(𝝉𝝉) =𝟏𝟏𝑵𝑵

�𝟏𝟏 + �𝟏𝟏𝒘𝒘

�−𝟏𝟏

� �𝟏𝟏 + �𝟏𝟏𝒘𝒘

�𝟐𝟐

�𝝉𝝉

𝝉𝝉𝑫𝑫��

−𝟏𝟏𝟐𝟐

(5.1)

𝑮𝑮(𝝉𝝉) = �𝟏𝟏 +𝑻𝑻

𝟏𝟏 − 𝑻𝑻𝒆𝒆−𝝉𝝉 𝝉𝝉𝑻𝑻� �

𝟏𝟏𝑵𝑵

�𝟏𝟏 + �𝟏𝟏𝒘𝒘

�−𝟏𝟏

� �𝟏𝟏 + �𝟏𝟏𝒘𝒘

�𝟐𝟐

�𝝉𝝉

𝝉𝝉𝑫𝑫��

−𝟏𝟏𝟐𝟐

(5.2)

where N is number of molecules in the confocal volume, w is the structure factor (radius,

r, of confocal volume over its half height), τ is the correlation time, τD is the characteristic

diffusion time (where τD = r2/4D (D is diffusion coefficient), and T is the triplet fraction, τT.

FCS measurements were performed with 30 μW excitation power, and the

optical system (r and w of confocal volume) was calibrated before each experiment

using free Rhodamine 6G (R6G) dye (D = 2.8 × 10−6 cm2/s in water; (Life Technologies,

CA) in deionized water. Autocorrelation curves were fit to Equation 5.1 or 5.2 using

Levenberg−Marquardt nonlinear least- squares regression algorithm with Origin software

to determine N, T, and τD .

122

Figure 5-3. Fluorescence correlation spectroscopy (FCS) results showing an enhanced

diffusion coefficient for Ald (a) and HK (b) in the presence of their respective substrates.

(a)

(b)

123

As described in Figure 5-3, both HK and Ald exhibited substantial increase in

diffusivity: upto 140% for HK and 230% for aldolase. Note that the reaction catalyzed by

aldolase is endothermic and its substrate-induced enhanced diffusion is, therefore,

inconsistent with the mechanism recently proposed by Bustamante et al.6

5.3.4 Statistical Significance Analysis of FCS data

Diffusion coefficients of each enzyme for each substrate concentration were

entered into a table in Graphpad Prism software. Means and standard deviations were

calculated. After this, an analysis of variance (ANOVA) test was performed followed by

Tukey’s multiple comparisons of means. For HK (Figure 5-3a), all means except for 50

μM were statistically significantly greater than the values at 0 μM substrate. For Ald

(Figure 5-3b) the values at 100 μM and 1000 μM were significantly greater than the

value at 0 uM.

5.3.5 Confocal Microscope Imaging

Confocal images were acquired using a Leica TCS SP5 laser scanning confocal

inverted microscope (LSCM, Leica Microsystems) with a 10X objective (HCX PL APO

CS, 0.70 NA) incorporated in it. The plane of interest (along the z-axis) for confocal

imaging was chosen such that fluorescence intensity was captured from the plane that is

half of the height into the channel.

Videos were recorded and analyzed using Image J software. In each experiment,

the mean fluorescence intensity was calculated from three videos. Each video is a

collection of 667 images. A region of interest (ROI) was selected along the channel (as

124

indicated by the vertical line in Figure 5-2), and the stack-averaged fluorescence

intensity was plotted as a function of distance along the width of the channel.

5.3.6. Detailed Investigation into Hexokinase Chemotaxis Behavior

The various components of the four-step cascade were allowed to flow through

the microfluidic device using syringe pumps to control flow rates. All solutions were

prepared in buffer, i.e. HEPES, 50 mM, pH 7.4. The enzyme concentrations were kept

constant at 200 nM. 2 mM ATP and 10 mM Mg2+ were also added to experiments

involving the two kinases to enable phosphorylation. The initial substrate concentration

was 10 mM, unless specified otherwise. The fluorescently tagged enzymes, HK and Ald

were tracked for their movement transverse to the flow planes in the microfluidic device,

along a vertical line towards the end of the channel as indicated in Figure 5-2. By using

specific flow rates and channel geometry, the distance traversed to the point where

fluorescence was measured was converted into time spent inside the channel (Table 5-

1). The fluorescence measured from top to bottom along the vertical line is plotted from

left to right in all the fluorescence figures shown. The fluorescence intensity was

normalized (maximum = 1) for comparison and representation on a common scale.

5.3.7 Substrate Triggered Chemotaxis

The ability of the individual enzymes in the cascade to chemotax up the gradient

of their respective substrates was first established. Accordingly, the chemotaxis of HK in

competing gradients of its substrate D-glucose and the corresponding enantiomer, L-

glucose, which is not a substrate, was tested. The volumetric flow rate was fixed at 200

125

µL/h and fluorescence was noted 30 mm down the channel, allowing a total interaction

time of 6.48 s. Three sets of experiments were performed. HK was allowed to flow

through the middle channel with either buffer solution (control), D-glucose (10 mM) and

buffer, or L-glucose (10 mM) and buffer flowing through the two flanking channels. As

shown in Figure 5-4, the spreading of HK into the buffer only and L-glucose channels

was comparable. On the other hand, there was a significantly enhanced migration of HK

into the D-glucose channel, suggesting a chemotactic movement towards its substrate

beyond what is expected for Brownian diffusion as has previously been observed for

urease, catalase, RNA and DNA polymerases.3, 4, 6

126

Figure 5-4. Chemotactic response observed for hexokinase (HK). HK shows

chemotactic shift only in presence of a gradient of its substrate, D-glucose (D-Glu) and is

unaffected by the presence of L-glucose (L-Glu). Also, hexokinase shows a greater

chemotactic shift towards its substrate of choice D-glucose (D-Glu) compared to

mannose (Mann) which it phosphorylates at a significantly lower rate. Experimental

conditions: Starting enzyme concentration = 200 nM (100%) Flow rate = 200µl/h,

distance = 30 mm, interaction time = 6.48 s; Percentage of enzyme migration into the left

D-glucose channel is 4.59 ± 0.4 % and towards the right D-glucose channel is 4.54 ± 0.3

%. Percentage of enzyme migration into mannose channel is 2.85 ± 0.5 %. Inset on the

top and bottom shows a clearer migration towards preferred channels. Note that the

percent enzyme migration into adjoining buffer channels due Brownian diffusion alone is

~ 2%.

127

5.3.8 Binding Affinity VS Turnover Rate

To confirm the role of substrate turnover in the observed enhanced chemotactic

movement, HK was presented a choice between its usual substrate, D-glucose and

another competitive substrate, mannose. HK shows a higher binding affinity towards

mannose (Km = 40 µM) compared to D-glucose (Km = 120 µM); on the other hand,

pyruvate kinase/lactose dehydrogenase coupled assays for HK activity confirmed

mannose phosphorylation to be half as fast as D-glucose phosphorylation (see section

5.3.9 for details). As before, three sets of experiments were performed at the same flow

rate. HK was allowed to flow through the middle channel with either a buffer solution

(control), D-glucose (10 mM) and buffer, or mannose (10 mM) and buffer flowing through

the flanking channels. A significantly greater chemotactic shift was observed towards the

D-glucose channel compared to the mannose channel (Figure 5-4) suggesting that

catalysis, rather than simple substrate binding, is important for the observed chemotaxis.

5.3.9 Enzyme activity assays

The difference in hexokinase activity using glucose or mannose as the substrate

was measured spectrophotometrically by coupling with pyruvate kinase/lactic

dehydrogenase and following the oxidation of NADH at 340 nm. An assay mixture 1 mL

in total volume contained 1 mM glucose or mannose, 2 mM ATP, 10 mM MgCl2, 3.3 mM

phosphoenolpyruvate, 50 mM HEPES (pH 7.4), 0.2 mM NADH, 2 units pyruvate

kinase/lactic dehydrogenase (based on PK units), and 5 nM hexokinase. All assays were

performed at 25 °C. The enzymatic activity of hexokinase with mannose as the substrate

was approximately half the enzymatic rate with glucose as the substrate.

128

5.3.9 Investigation into Aldolase Chemotaxis

Experiments were also performed to probe the chemotactic movement of Ald

towards its own substrate: fructose 1,6-bisphosphate. Varying concentrations of the

substrate were tested: 1µM, 10µM, 100µM and 1mM. Once again, three sets of

experiments were performed. The enzyme, Ald, was allowed to flow through the middle

channel with buffer (control) and varying concentrations of its substrate in buffer flowing

in the adjoining channels. A chemotactic shift was observed when a solution of Ald was

exposed to 100 µM or higher concentration of fructose bisphosphate (Km = 60 µM) in an

adjoining channel. The response was obsrved to increase with increasing substrate

concentration.

5.3.10 Why Chemotaxis? Enhanced Diffusion Model

Chemotactic movement of the enzyme molecules towards higher substrate

concentrations can arise from substrate concentration-dependent enhanced diffusion, as

demonstrated previously.3, 4, 7 The substrate concentration changes continuously as the

enzyme diffuse along the substrate gradient. Thus, at every point in space, the diffusivity

increases on moving up the gradient and decreases on moving down the gradient. A

higher diffusivity leads to a greater spreading of the enzyme molecules on the side of the

higher substrate concentration. Thus, the “center of gravity” of the enzyme ensemble

moves towards the region of higher substrate concentration. Detailed modeling based on

substrate-induced enhanced diffusion predicted chemotatic shifts similar to those

observed, thereby supporting the above hypothesis.

129

5.3.11 Inadequacy of the Diffusion Model to Explain Chemotaxis

While the above experiments involved enhanced enzyme diffusion towards

regions of higher substrate concentration, it is also possible to prevent the normal

diffusional spreading of enzymes and force them to focus into a narrower region in

space in response to the presence of the substrate. As shown in Figure 5-5 and Figure

5-6, the spreading of HK from the middle to the flanking buffer channels was significantly

reduced when the buffer solution is substituted by a solution of D-glucose in the middle

channel. Moreover, while HK in buffer spreads into adjoining substrate channels, when

glucose is introduced within the HK channel, chemotaxis towards adjoining substrate

channels was again restricted and resembled HK behaviour towards flanking buffer

cahnnels (Figure 5-7). If a purely diffusion model was to hold true, this constriction

should not have been seen.

Although the experiment involving competition between mannose and D-glucose

for HK suggested a dominant role for catalysis in the observed chemotatic movement,

substrate binding itself can also cause enzymes to preferentially locate themselves in

regions of higher substrate concentrations.9 If the chemical potential of the enzyme-

substrate complex is lower than that of enzyme + substrate, then the enzyme would be

located in (thermodynamically favorable) regions of higher substrate concentration.

Indeed, active transport in response to chemical potential gradients has been reported

for nanoparticles, polymers, and dendrimers.

130

Figure 5-5. Substrate-induced enzyme focusing. The normal diffusional spreading of HK

(1 µM) towards the flanking channels that contain buffer is reduced if the composition in

the middle channel is switched from HK in buffer to HK in 70 mM D-glucose. The net

reduction in area is 13.4 ± 3.0%. Experimental conditions: Flow rate = 100µl/h, distance

= 18 mm, interaction time = 7.78 s.

131

Figure 5-6. Cofactor-induced enzyme focusing. The enzyme (1 µM) switches from an

equilibrium distribution to a non-equilibrium one when cofactors ATP (50 mM) and MgCl2

(100 mM) are introduced in the middle channel. This is analogous to reported cellular

responses in the cytosol where enzyme association is regulated by oxygenation and

phosphorylation requirements. Experimental conditions: Flow rate = 30 µl/h, distance =

19 mm, interaction time = 24.7 s.

132

Figure 5-7. Restricted chemotaxis in the absence of substrate gradient. The normal

diffusional spreading of HK (200 nM) towards the flanking substrate channels is reduced

if the substrate is also introduced within the middle channel flowing the enzyme.

Experimental conditions: Flow rate = 100 µl/h, distance = 20 mm, interaction time = 8.64

s.

133

5.4. Enzyme Cascade Investigation

Having demonstrated that individually both enzymes, HK and Ald, chemotax up

the gradient of their respective substrates, the behavior of the entire four enzyme

cascade was then examined. The first experiment was designed to examine the

response of Ald towards its substrate, fructose 1,6-bisphosphate (FBP), generated from

D-glucose by the successive actions of the first three enzymes. In a microfluidic device,

the Ald was allowed to flow through the middle channel. The first three enzymes, HK, Iso

and PFK, with Mg2+ and ATP (required by the kinases) were passed through one of the

flanking channels, along with 10 mM D-glucose, while buffer was passed through the

flanking channel on the opposite side. The setup allowed Ald an equal opportunity to

migrate towards either or both of the flanking channels (Figure 5-8a). The volumetric

flow rate per inlet was fixed at 50 µL/h, allowing a total interaction time of 17.3 seconds

in a 20 mm channel. 8.1 ± 0.9 % of Ald was observed to spread into the channel in

which its substrate was being formed in situ (Figure 5-8b). When the interaction time

was reduced to 8.6 s, the chemotactic migration correspondingly reduced to 4.5 ± 1.0 %.

As expected, HK did not show any movement into the adjoining channel (Figure 5-9a).

Additional control experiments were performed by removing either enzyme 2; Iso,

enzyme 3; PFK, or D-glucose from the channel containg HK, Mg2+ and ATP. In each

case, Ald showed no chemotactic shift (Figure 5-9b).

134

Figure 5-8. (a) Experimental set-up to study the chemotactic response of Ald (green

channel) towards the channel that generates its substrate in situ. (b) Fluorescence

intensity measured across the channels plotted against the width of the channels. The

dotted lines represent the approximate center channel boundaries. When compared to

Ald’s movement towards buffer, the enzyme shows enhanced migration into the channel

that generates its substrate in situ. (c) Experimental set up that allows the entire

enzymatic reaction cascade to occur in-situ. Substrate (D-glucose) for enzyme 1, HK

(red channel), was provided in the middle channel to trigger the cascade. (d) Ald (green

bars) shows time-delayed chemotactic response compared to HK (red bars) as expected

based on the sequence of reactions. When mannose was introduced along with D-

glucose, HK shows reduced chemotaxis (orange bars) corresponding to slower rate of

mannose phosphorylation.

135

Figure 5-9. (a) While Ald chemotaxes towards its substrate gradient (Figure 5-8b), HK

flowing along with its substrate in its own channel, shows no movement into the adjacent

channel. (b) Control experiments performed for studying the chemotactic response of

Ald towards its substrate precursors. Ald shows no movement towards the channel

flowing the recipe for its substrate when any one of the ingredients is missing.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340

Nor

mal

ized

Flu

ores

cenc

e In

tens

ity

Distance across channel (µm)

Buffer-Ald-Substrate precursors (HK) at 8.6 s

Buffer-Ald-Substrate precursors (HK) at 17.3 s

(a

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340

Nor

mal

ized

Flu

ores

cenc

e In

tens

ity

Distance across channel (µm)

Buffer-Ald-Substrate precursorw/o glucose

Buffer-Ald-Substrate precursorw/o Iso

Buffer-Ald-Substrate precursorw/o PFK

(b)

136

5.4.1 Cascade In-situ

After establishing the basic outline of the chemotactic assembly of the four-

enzyme cascade, the sequential chemotactic movement of HK and Ald when exposed to

D-glucose was examined. This is expected since D-glucose is the immediate substrate

for HK, while the substrate for Ald, fructose 1,6-phosphate, is only formed from D-

glucose through three successive enzymatic steps. The components of the cascade

were now separated into two batches consisting of the first two and the last two

enzymes, respectively. HK, ATP, Mg2+, Iso were flowed through one flanking channel,

while PFK, ATP, Mg2+, Ald were flowed through the other flanking channel. A solution of

D-glucose passed through the middle channel (Figure 5-8c). The flow rate was reduced

to 30 µL/h and the channel length was increased to 40 mm, allowing for a total

interaction time of 57.6 s within the channel. A reaction progress simulation (see section

5.4.2) was run to ensure the time was enough for the entire cascade to occur, given the

enzyme concentrations and the Kcats.

5.4.1.1 Progress Curve Simulation

The substrate depletion and product formation through the first four enzymes in

the glycolytic cascade were simulated using Global Kinetic Explorer software (version

4.0, KinTek Corporation).12 The steady-state reaction scheme assumed 1) substrate

binding rates at the diffusion limit for glucose binding to hexokinase since the initial

glucose concentration was sufficient to saturate the enzyme, and at kcat/Km for the

subsequent enzyme reactions because the substrates were the product of the previous

enzyme reaction and their concentrations did not reach the level of saturation; 2)

137

irreversible reaction rates fixed at kcat for each enzyme since the product of each

reaction would be pulled through the cascade by the presence of the downstream

enzymes preventing the reverse reaction or product inhibition; and 3) that product

release was not rate limiting for any individual reaction. The simulation input values were

10 mM for the starting glucose concentration; 74 nM for each starting enzyme

concentration; k1 = 1000 µM-1s-1 and kcat = 315 s-1 for hexokinase; kcat = 408 s-1 and Km =

700 µM for isomerase; kcat = 113 s-1 and Km = 30 µM for PFK; and kcat = 5 s-1 and Km =

60 µM for aldolase. The simulation assumes that all the enzymes and glucose are

combined in one reaction mixture; an enzyme concentration of 74 nM was chosen

because that is the amount of hexokinase determined to migrate into a channel

containing 10 mM D-glucose (Table 5-2). The progress curves from the simulation

indicated that the 100 µM threshold concentration of fructose 1,6-bisphosphate, the

substrate for aldolase, could be produced under the experimental conditions to promote

the migration of the enzyme up the substrate concentration gradient.

138

Figure 5-10. The simulated substrate and product progress curves through the first four

enzymes in the glycolytic cascade, assuming steady state concentrations.

139

As described above, the examination of the enzyme reaction rates confirmed that

the time available within the microchannel was sufficient for entire cascade of reactions

to occur (Figure 5-10). As discussed, hypothesis being tested was that HK would

respond first to its substrate gradient by moving into the D-glucose channel, thereby

producing the substrate for enzyme 2, Iso. The cascade would continue with PFK

participation, finally producing fructose 1,6-bisphosphate that in turn should prompt Ald

to chemotax towards the central channel. The fluorescence profiles for enzymes HK

and Ald were noted at different interaction times, 14.4 s, 28.8 s, 43.2 s and 57.6 s and

their chemotactic behavior is summarized in Figure 5-8d. As expected, a clear

sequential movement of HK, followed by Ald towards the central channel was observed.

5.4.2 Competitive Substrates

The above experiment was repeated, except that the solution of D-glucose (10

mM) in the middle channel was replaced by a solution consisting of D-glucose (10 mM)

and its competitive substrate, mannose (10 mM), which has a higher binding affinity but

reacts more slowly than glucose (see above). As shown in Figure 5-8d, HK now showed

a smaller chemotactic shift towards the central channel and no significant shift of Ald

was observed because of the slowing down of the reaction cascade that originates from

glucose (mannose phosphorylation does not initiate the glycolysis cascade).

5.5. Chemotaxis and Metabolons

For both HK and Ald, a linear relationship was observed between fluorescence

intensity and concentration (Figure 5-11). This allowed the estimation of the amount of

140

enzyme that had chemotaxed into a specific substrate channel. The results for the four

enzyme cascade experiment involving sequential movements of HK and Ald (Figures 5-

8c-d) are tabulated in Table 5-2. For HK, the results indicate that, in 58 sec, 37 % of the

starting 200 nM enzyme moves into the central channel containing D-glucose (10 mM)

compared to 7 % of the enzyme moving into the same channel when flowing only buffer

(see Table 5-2).

Figure 5-11. Linear relationships between fluorescence intensity (arbitrary units) and

concentration for both HK and Ald. This enables directly correlating fluorescence

intensity to the concentration of enzyme.

141

Table 5-2. Concentration of enzyme (HK or Ald) migrated into the central channel

(containing either buffer only or 10 mM D-glucose + buffer) at specified time periods (see

Figure 5-7c). The starting concentration of both enzymes was 200 nM.

Enzyme Time (s)

Enzyme conc. in buffer channel (% of 200 nM)

Enzyme conc. in glucose channel

(% of 200 nM)

HK

14.4 2.6 ± 0.7 12.1 ± 3.8

28.8 3.8 ± 1.6 19.6 ± 3.7

43.2 5.0 ± 0.4 28.5 ± 0.3

57.6 6.7 ± 1.3 37.0 ± 3.0

Ald

14.4 2.9 ± 0.4 3.6 ± 0.8

28.8 3.4 ± 1.0 5.1 ± 1.4

43.2 5.0 ± 2.0 7.4 ± 0.6

57.6 5.9 ± 1.0 8.9 ± 0.7

142

5.5.1. Chemotaxis in Cytosolic Conditions

Finally, to replicate the cytosolic crowding conditions that enzymes encounter in

cells due to the presence of other macromolecules, 20% w/v Ficoll PM 70 was added to

the experiments involving the entire cascade. Ficoll PM 70 is a highly branched

polysaccharide polymer that serves as a synthetic crowding agent by affecting the fluid

properties of the solution, such as increasing the viscosity and osmolality.10 As shown in

Figure 5-12, the presence of the crowding agent slowed down but did not stop the

chemotactic movement and assembly of the enzymes.

Figure 5-12. D-glucose gradient-driven sequential movement of HK and Ald for the

entire enzymatic reaction cascade was observed even in Ficoll PM 70 (20% w/v)

induced crowded environment mimicking cytosolic crowding conditions in cell. Ald (red

bars) shows a time delayed chemotactic migration towards substrate channel compared

to HK (blue bars) corresponding to the cascade reaction sequence. The error bars

represent the standard deviation.

0

2

4

6

8

10

12

1 2

% E

nzym

e C

once

ntra

tion

Time (s)

HK in glucose channel

Ald in glucose channel

43.2 57.6

143

5.6 Conclusion

The results discussed above, suggest that the observed assembly of enzymes

participating in a cascade in response to the presence of the initial substrate is a direct

result of enzymes undergoing chemotaxis in response to their specific substrates. The

extent of enzyme migration is proportional to the exposure time to the substrate gradient.

Significantly, the chemotactic migration of enzymes is fairly rapid even under conditions

that mimic cystolic crowding: > 0.5 microns/sec, a rate very similar to that reported for

enzyme diffusion in living cells.5 This mechanism obviates the need for direct interaction

between the enzymes to form complexes that promote substrate channeling.

Furthermore, the enzymes should revert back to their equilibrium distribution, once the

initial substrate is completely reacted and the substrate gradients for the individual

enzymes disappear. Evidence presented in this chapter suggest that enzymes in a

cascade assemble via chemotaxis. Each of the enzymes involved independently follows

the gradient of its own specific substrate, which in turn is produced as a product of the

preceding reaction. The chemotactic assembly of enzymes occurs even under cytosolic

crowding conditions. Sequential directional migration of enzymes participating in the

glycolysis cascade is observed in response to the initial substrate, D-glucose.

5.7 Acknowledgements

The author would like to thank Dr. Michelle Spiering and Prof. Stephen Benkovic

for their guidance with the enzyme tagging experiments, and sharing their insights on

enzyme catalysis. The author also expresses her gratitude towards Xi Zhao and Prof.

Peter Butler for their help with the fluorescence correlation spectroscopy measurements.

144

5.8 References

1. An, S.; Kumar, R.; Sheets, E. D.; Benkovic, S. J. Science 2008, 320, 103-106. 2. Percival Zhang, Y.-H. Biotechnology Advances, 2011, 29, 715-725.

3. Sengupta, S.; Dey, K. K.; Muddana, H. S.; Tabouillot, T.; Ibele, M. E.; Butler, P. J.;

Sen, A. J. Am. Chem. Soc.2013, 135, 1406-1414. 4. Sengupta, S.; Spiering, M. M.; Dey, K. K.; Duan, W.; Patra, D.; Butler, P. J.;

Astumian, R. D.; Benkovic, S. J.; Sen, A. ACS Nano 2014, 8, 2410-2418. 5. Baum, M.; Erdel, F.; Wachsmuth, M.; Rippe, K. Nature Comm. 2014, 5,

doi:10.1038/ncomms5494. 6. Riedel, C.; Gabizon, R.; Wilson, C. A. M.; Hamadani, K.; Tsekouras, K.;

Marqusee, S.; Presse´, S.; Bustamante, C. Nature, 2015, 517, 227-230. 7. Karp, G. Cell and Molecular Biology: Concepts and Experiments. Wiley, New

York, 2009, 5, 108. 8. Yu, H.; Jo, K.; Kounovsky, K. L.; De Pablo, J. J.; Schwartz, D. C. J. Am. Chem.

Soc., 2009, 16, 5722-5723. 9. F. Wu, L. N. Pelster, S. D. Minteer, Chem. Commun., 51, 1244-1247 (2015). 10. Vopel, T.; Makhatadze, G. I. PLOS One, 2012, 7, 1-6. 11. Xia, Y.; Whitesides, G. M. Annu. Rev. Mater. Sci.1998, 28, 153−184. 12. Johnson, K. A.; Simpson, Z. B.; Blom, T. Anal. Biochem. 2009, 387, 20-29. 13. Gullapalli, R. R.; Tabouillot, T.; Mathura, R.;Dangaria, J. H.; Butler, P. J. J Biomed

Opt. 2007, 12, 014012. 14. Muddana, H. S.; Sengupta, S.; Mallouk, T. E.; Sen, A.; Butler, P. J. J. Am. Chem.

Soc. 2010, 7, 2110-2111.

145

Chapter 6

Bringing discipline into enzyme motors

6.1. Introduction

The discovery of synthetic motors and pumps in 2004 opened up new avenues

for nanoscale transport that held other broader implications in the field of artificial

intelligence. Chapters 1, 2, 3 and 4 of this thesis discuss specific applications of such

systems in microfluidics, medical therapy, encoded-healing materials, and memory-

encrypted smart materials. More recently motor like behavior of non-motor proteins, in

the form of chemotaxis towards substrates, has been demonstrated. These naturally

occurring biological motors present new opportunities for designing an innovative class

of organic-synthetic-hybrid nanomachines. However, further investigation is needed to

fully understand the origins of the enzyme motor behavior, in order to design future

systems. This chapter focusses on developing an understanding of the genesis of

enzymatic chemotaxis by testing enzyme behavior in different environments.

6.2 Motivation

Diffusion of enzymes like urease, catalase, DNA polymerase etc. towards higher

substrate gradient has previously been reported.1 This phenomenon, defined as diffusion

based chemotaxis, is observed due to higher enzyme diffusivity in the substrate.

However, further investigation is needed to comprehend this behavior entirely and

compare with the better studied bacterial chemotaxis.2, 3 This chapter focusses on better

146

understanding the phenomenon by investigating enzyme behavior in complicated

environments and investigating the role of binding vs catalytic turn-over.

6.3. Experimental Design

A flow based microfluidic gradient generator was fabricated, similar to the

one described in chapter 5. The microfluidic device was a three inlet, one outlet device,

fabricated through photo-lithography that consisted of three parallel channels, 20 mm in

length, 100 µm height and 360 µm in width (Figure 6-1).

Figure 6-1. Photo- lithographically fabricated flow based microfluidic gradient generator

for studying enzyme chemotaxis. The length of the horizontal channels is 20 mm, width

360 μm and height is 100 μm.

Fluorescence measurement

S E

S + I

147

6.3.1 Test Subject 1: Catalase

The first enzyme investigated was catalase. Catalase is a robust enzyme with a

high Kcat of 105 s-1. Also, sodium cyanide is known to inhibit catalase in a competitive

manner. 4μM catalase enzyme was prepared and tagged with NHS-Rhodamine dye, 10

mM H2O2 was prepared as the substrate (S) and 0.1 M sodium cyanide solution in 10

mM peroxide solution as the substrate-inhibitor mix (S+I). All solutions were made in DI

water since catalase is best deactivated by cyanide in water. The S and S+I solutions

were passed through the two extreme inlets and fluorescently tagged enzyme through

the middle inlet as shown in Figure 6-1. For the control experiment, substrate was

passed through both extreme inlets with the tagged enzyme in the middle. The flow rate

was maintained at 200 μl/h. Lower flow rates ~50 μl/h increase the bubble formation,

due to oxygen release, as a result of the peroxide decomposition by catalase. A confocal

microscope was used to track the fluorescently tagged enzyme; along the vertical line

shown in Figure 6-1, and the data was analyzed using image J software, in the same

manner as described in chapter 5. As shown in Figure 6-2, a shift in fluorescence

intensity was observed towards the substrate gradient and away from the inhibitor

gradient. For the control experiment, the fluorescence intensity remained largely

centered across the channel. In other words, the enzyme diffusion was equal towards

both substrate channels as expected.

The lateral shift observed at half maximum fluorescence intensity, in Figure 6-2,

was comparable to previously reported substrate driven chemotaxis shift for catalase

~13 μm.1 The presence of the inhibitor restricted the enhanced migration of the enzyme

into the substrate channel.

148

Figure 6-2. Shift in fluorescence intensity observed for catalase. The enzyme diffuses

away from the inhibitor (NaCN) and towards the substrate (H2O2) (Note the blue graph’s

shift towards left when compared to the control (red)).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 50 100 150 200 250 300 350

Nor

mal

ized

fluo

resc

ence

inte

nsity

Distance across channel (µm)

H2O2-catalase-H2O2in NaCNH2O2-catalase-H2O2

149

6.3.2 Test Subject: Urease

A similar study was then conducted for urease enzyme. 1 M urea was prepared

as the substrate and 1 mM pyrocatechol in 1 M urea was prepared as the S+I solution.

Urease was tagged with dylight malemide 550 dye (Thermo Scientific). All solutions

were made in PBS buffer (saline, pH 7.2). However, no shift was observed in case of

urease (Figure 6-3). The reason for this is postulated to be the incomplete inhibition of

urease. Pyrocatechol inhibits urease in a time and concentration dependent manner.4 It

reacts with urease to first convert to an intermediary that finally inactivates urease by

attacking the cysteine units at the active site. Given the dimensions of the microfluidic

device, the inhibitor gets only 4.32 seconds to deactivate or inhibit the enzyme. While

this time may be sufficient for cyanide to finish its job, pyrocatechol is unable to do so

resulting in a fluorescence profile similar to the control, i.e. no observed shift.

150

Figure 6-3. No shift in fluorescence intensity observed for Urease. Pyrochatechol is

unable to completely inhibit urease within the 4.32 s in the microfluidic channel, due to

the slow inhibition rate. As a result, no shift is observed.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 50 100 150 200 250 300

Nor

mal

ized

Flu

ores

cenc

e In

tens

ity

Distance across channel (µm)

Urea-Urease-Urea inCatechol

Urea-Urease-Urea

151

Hence, in order to study urease’ chemotactic behavior, a new microfluidic

gradient generator was designed that could elongate the diffusion process. Agar gel has

previously been reported to be utilized to study DNA chemotaxis.5 Keeping this strategy

in mind, a diffusion based microfluidic gradient generator was then designed. After

testing several designs, the pattern shown in Figure 6-4 was found to work best. This

microfluidic device consists of two adjacent reservoirs, one for S and the other for S+I,

separated by a wall. These two were connected to a third chamber that stored the

fluorescently tagged enzyme.

Figure 6-4. Diffusion based microfluidic gradient generation device designed using

Adobe illustrator, printed on acrylic surface using a CO2 laser printer and then cast on

PDMS using soft lithography.

S + I

152

Both the S and the S+I chambers were 2 mm in diameter and the enzyme

chamber 1 mm, connect by two 2 mm wide channels. 1% agar gel (prepared in PBS

buffer, saline, pH 7.2; the same solvent as that used to prepare the rest of the solutions)

was first introduced in the device to allow the substrate and inhibitor solutions to diffuse

slowly through it. This helps to study the diffusion for a longer time period and enables

pyrocatechol to inhibit urease effectively. About 20 µl of S and S+I was injected into their

respective channels and left undisturbed for about 10 minutes, to allow the gradient to

build. Finally about 10 µl of dye tagged urease was introduced in its chamber and

allowed to slowly diffuse through the agar-buffer bed into the connected channels for 4-5

hours. The recorded fluorescence was then analyzed using image J software (Figure 6-

5, 6-6).

153

Figure 6-5. Normalized fluorescence intensity measured across the substrate and

substrate + inhibitor channels in the microfluidic device. (a) The fluorescence intensity

within substrate (urea) and the S+I (urea + catechol) channel. The enzyme diffuses

much faster and further into the substrate channel compared to the S+I channel. In case

of S+I channel most of the enzyme concentration (fluorescence maxima) stays close to

the starting position. (b) Control experiment performed contained the substrate urea in

both reservoirs and the fluorescence intensity indicates similar enzyme diffusion in both

channels.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 500 1000 1500 2000 2500

Nor

mal

ized

Flu

ores

cenc

e In

tens

ity

Distance (µm)

0 hrs

2 hrs Catechol

2 hrs Urea

(a)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 500 1000 1500 2000 2500

Nor

mal

ized

Flu

ores

cenc

e In

tens

ity

Distance (µm)

0 hrs

4 hrs-urea1

4 hrs-urea2

(b)

154

Figure 6-6. Normalized fluorescence intensity measured across the substrate and

substrate + inhibitor channels in the microfluidic device over 5 hours. The fluorescence

intensity within the (a) substrate (urea) and (b) the S+I (rea + catechol) channel. The

enzyme diffuses much faster and further into the substrate channel compared to the S+I

channel. In case of S+I channel most of the enzyme concentration (fluorescence

maxima) stays close to the starting position.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 500 1000 1500 2000 2500

Nor

mal

ized

Flu

ores

cenc

e In

tens

ity

Distance (µm)

(a)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 500 1000 1500 2000 2500

Nor

mal

ized

Flu

ores

cenc

e In

tens

ity

Distance (µm)

(b)

0 hrs 3 hrs 5 hrs

0 hrs 3 hrs 5 hrs

155

6.4 Results and Discussion

The enzyme diffused into the substrate channel, indicated by the shifting

fluorescence maxima; whereas in case of the inhibitor channel, the fluorescence

maximum stays largely at the starting position. The hydrodynamic radii (a) of the

diffusing species are: urease- 7 nm6, 7, urea- 0.18 nm8, pyrocatechol- 0.39 nm9. Taking

the viscosity of 1% w/w agar gel to be 4 X 10-3 Pa-s10, the diffusion coefficient of urease

in the gel at 298 K is calculated (using D= kT/6Πηa) to be 0.78 X 10-11 m2/s. Using this D

value, the average distance (L) urease diffuses in 5 hrs is calculated (using L2 = 2Dt) to

be 515 μm, urea: 3208 μm and pyrocatechol: 2180 μm. The enzyme diffusion in both

channels is noted to be significantly higher. The presence of the substrate in the

channels enables the enhanced diffusion observed. Some diffusion by the enzyme is

also seen in the S+I channel as expected, it however does not cross the channel into the

S+I chamber. The modest spreading of urease into the S+I channel could simply be due

to the slight pull experienced by it due to the faster diffusion of urea compared to

pyrocatechol from the S+I solution. The enzyme’s inhibition to travel further ahead in the

channel could possibly be attributed to its sensing the presence of the inhibitor.

Pyrocatechol, once bound on the enzyme, alters its structure, thus preventing it to

interact any more with the substrate i.e. urea. The results confirm that a combination of

substrate and inhibitor solution could be used to restrict or alter the enzyme motor. Also

urease’ ability to sense a substrate gradient and diffuse towards it over time is

comprehensively proved (Figure 6-5a, Figure 6-6a).

Another set of control experiments performed with urease being exposed to a

gradient of buffer in both reservoirs. The enzyme diffuses ~1400 µm into along the buffer

channels, as expected of Brownian diffusion (Figure 6-7).

156

Figure 6-7. Normalized fluorescence intensity measured in the buffer channel over time.

Only Brownian diffusion is observed.

6.5 Conclusions

The results discussed in this chapter open new avenues towards controlling the

motion of ‘the enzyme motor’. The presence of an inhibitor along with a substrate

gradient was observed to restrict the motility allowing for regulated motion. This

observation has been demonstrated on both catalase and urease in both flow based as

well as diffusion based gradient generating microfluidic devices respectively. This

indicates towards the generality of imposing an inhibitor gradient to control a wide variety

of ‘enzyme motors’. The experimental results, such as competitive binding events, also

widen the mechanistic understanding of the phenomenon of chemotaxis itself. Future

studies are planned with hydroxyurea, a structural analog of urea that functions as a

reversible and competitive inhibitor to urease; unlike pyrocatechol. While urease binds

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 500 1000 1500 2000 2500 3000

Nor

mal

ized

Flu

ores

cenc

e In

tens

ity

Distance (µm)

0 hrs8 hrs12 hrs

157

reversibly to hydroxyurea, it is not turned over. A comparison of enzyme’s chemotactic

response towards hydroxyurea versus urea could help gain important insights into the

role of binding versus catalysis.

Preliminary fluorescence correlation spectroscopy (FCS) results on hexokinase

with increasing substrate concentration without the presence of cofactors- ATP and

MgCl2 shows an increase in diffusion coefficient of the enzyme. These results suggest

that reversible binding without catalytic turn over could also lead to enhanced diffusion.

The future experiments planned with urease that investigate its chemotactic response

towards hydroxyurea over urea, in other words, reversible binding over catalysis, would

shed new light on the process.

158

6.6 References

1. Sengupta, S.; Dey, K. K.; Muddana, H. S.; Tabouillot, T.; Ibele, M. E.; Butler, P. J.; Sen, A. J. Am. Chem. Soc. 2013, 135, 1406−1414.

2. Adler, J. Annu. Rev. Biochem. 1975, 44, 341-356. 3. Tso, W. W.; Adler, J. J. Bacteriol. 1974, 118, 560. 4. Kot, M.; Zaborska, W. J. Enzyme Inhib. Med. Chem. 2003, 18, 413–417. 5. Yu, H.; Jo, K.; Kounovsky, K. L.; De Pablo, J. J.; Schwartz, D. C. J. Am. Chem.

Soc. 2009, 131, 5722–5723. 6. Follmer, C.; Pereira, F. V.; Da Silveira, N. P.; Carlini, C. R. Biophys. Chem. 2004,

111, 79–87. 7. Muddana, H. S.; Sengupta, S.; Mallouk, T. E.; Sen, A.; Butler, P. J. J Am Chem

Soc. 2010, 132, 2110–2111. 8. Schultz, S. G.; Solomon, A. K. J. General Physiology, 1961, 44, 1189-1199. 9. Rudyk, R. A.; Molina, M. A. A.; Gómez, M. I.; Blanco, S. E.; Ferretti, F. H. Internet

Electronic Journal of Molecular Design, 2004, 3, 11–28. 10. Folger, R.; Weiss, L.; Glaves, D.; Subjeck, J. R.; Harlos, J. P. J. Cell Sci., 1978,

31, 245-257.

159

Chapter 7

Conclusions

“Science never solves a problem without creating ten more” - Geroge Bernard Shaw

This thesis describes the new developments in the field of artificial

nanomachines powered predominantly by ion gradient led diffusiophoretic nanoscale

motion. Viable applications of this motion have been elucidated in each chapter. First,

regulation of colloidal transport for microfluidic based lab on chip style devices is

demonstrated. On/off switchability, pH-controlled motion and photo-triggered rectification

and amplification of colloidal transport are significant advances in this field.

Secondly, ion gradients generated from damaged mineralized substrates like

bone lesion or cracked polymers, have been shown to transform the damaged

substrates into power-houses generating remedies to cure themselves. The biological-

synthetic-hybrid micropump powers motion that allows repair agents to track and reach

the damage site.2 The ability to repair on-site, at ambient temperatures, without the use

of an external power source like electric or magnetic field has powerful implications in

medicine and coatings industry. The versatility of the approach allows easy quick-fix

applications for a variety of problems.

Thirdly, ion gradients have also been demonstrated to design smart materials

that show signs of memory; a level of autonomous function that extends beyond existing

smart materials. Diffusiophoretic motion generated from an engineered polymeric

material is used an example to demonstrate continuous pumping response, even in the

absence of the model stimuli; once triggered. The design of the material allows for

160

fabricating pre-programmed materials that could respond to a variety of stimuli, thus

satisfying an array of applications.

Finally, new insights are presented on the substrate gradient driven enzyme

motor-the next generation of biological-synthetic-hybrid-nanomachine. The ensemble

behavior of the enzyme motor- chemotaxis, i.e. preferential movement up a substrate

gradient, has been used to unravel the mystery of enzymatic cascades in cells. The

sequential directed movement of enzymes in the glycolysis cascade towards each other,

driven by the chemotactic response of the individual enzymes to their respective

substrate gradients is described. The latest data shines new light on the novel

phenomenon that is still not completely understood. While catalytic conversion being a

necessary and overriding factor generating chemotaxis is comprehensively proved, the

new data leaves the enhanced diffusion model insufficient to explain the observed

substrate driven enzyme focusing.

The endeavor to gain mechanistic insights into chemotaxis continues into chapter

6, where the role of (inhibitor) binding is again observed to be moderated by catalytic

conversion. Further investigation continues into studying other factors such as, the role

of temperature gradients by examining a temperature driven chemotaxis in the absence

of substrate gradients. Also, investigations into enzyme orientation driven chemotaxis is

being undertaken. At the same time, theoreticians are working in tandem to model the

unique phenomenon based on new experimental observations.

In conclusion, substantial progress has been made into unravelling the

phenomenon of chemotaxis and the baton has been passed to Xi Zhao for conquering

the mystery.

VITA

Vinita Yadav

Education Ph.D. Chemistry, The Pennsylvania State University GPA 3.9/4.0 M.S. Chemistry, University of Delhi First Class B.S., Chemistry, University of Delhi First Class with Honors

Awards and Honors • Young Investigator award, Baxter International, 2014• Miller Fellowship, The Pennsylvania State University, 2014• Invited Speaker, Gordon research conference, 2014• Very Important Paper, Angewante Chemie, 2013• Noteworthy Paper, Angewante Chemie, 2013• Selected for Leadership Workshop at Georgia Tech., NSF, Summer 2013• Travel Award, The Pennsylvania State University, Fall 2012, Spring 2014• Graduate Fellowship, The Pennsylvania State University, 2011-2012• Merit Award, University of Delhi, 2006• Gold Medalist, University of Delhi, 2004• Merit Award, University of Delhi, 2004

Publications

1. Yadav, V.; Zhang, H.; Pavlick, R.; Sen, A. J. Am. Chem. Soc., 2012, 134, 156882. Yadav, V.; Freedman, J.; Grinstaff. M.; Sen, A. Angew. Chem. Int. Ed., 2013,

52, 109973. Baker, M. S.; Yadav, V.; Sen, A.; Phillips, S. T. Angew. Chem. Int. Ed., 2013,

52, 10295

4. Yadav, V.; Duan, W.; Sen, A. invited book chapter, Engineering of chemicalcomplexity, 2014 Ed. (Edited by 2007 Chemistry Nobel Laureate Gerhard Ertl)

5. Yadav, V.; Pavlick, R. A.; Meckler, S.; Sen. A. Chemistry of Materials, 2014, 26,4647

6. Duan, W.; Wang, W.; Das, S.; Yadav, V.; Sen, A. Annual Review of AnalyticalChemistry, accepted, 2015

7. Yadav, V.; Duan, W.; Sen, A. Annual Review of Biophysics, accepted, 2015

8. Yadav, V.; Spiering, M., Zhao, X.; Scott, J.; Linderberg, K.; Gilson, M. K.; Butler, P.J.; Benkovic, S. J.; Sen, A. Science, submitted, 2015

9. Zhao, X.; Yadav, V.; Sen, A. Bringing Discipline into Enzyme Motors, manuscript inpreparation for Nature Chemistry, 2015