<%BANNER%>

Cytoskeletal Mechanics in Endothelial Cells

Permanent Link: http://ufdc.ufl.edu/UFE0042874/00001

Material Information

Title: Cytoskeletal Mechanics in Endothelial Cells
Physical Description: 1 online resource (147 p.)
Language: english
Creator: RUSSELL,ROBERT J
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ABLATION -- ACTIN -- ENDOTHELIAL -- FIBERS -- LASER -- MECHANICS -- MYOSIN -- SARCOMERES -- STRESS
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: An important function of normal endothelial cells is the ability to sense and transduce applied mechanical forces due to blood flow through the vasculature. Defects in endothelial mechano-sensing result in numerous pathologies including atherosclerosis. The actomyosin and microtubule cytoskeleton gives endothelial cells structural support and the ability to both sense and respond to external forces. This dissertation focuses on how the actomyosin and microtubule cytoskeleton generate and respond to forces in endothelial cells. Tensile force within endothelial cells is generated in actomyosin stress fibers, which are composed of contractile units called sarcomeres. Using femtosecond laser ablation, we severed living stress fibers and measured sarcomere contraction under zero tension. The time dependent trajectories of contracting sarcomeres suggested three distinct phases: an instantaneous initial response, a sustained change in length at constant velocity and finally a steady state. We proposed a novel model for tension generation in the sarcomere where myosin generated forces in adjacent sarcomeres are directly in balance. Through live cell imaging we observed that the number of sarcomeres and sarcomere lengths dynamically change in the cell. We show that sarcomere lengths continually fluctuate, new sarcomeres formed at focal adhesions are convected into the stress fiber and that sarcomeres disappear at specific points or ?sinks? along the stress fiber. These results show that stress fibers are highly dynamic structures despite their relatively static morphology. Taken together the observation of sarcomere fluctuations and the sarcomere contraction results suggest a model where tension fluctuations determine the distribution of sarcomere lengths. Microtubules are the stiffest of the cytoskeletal filaments but are consistently observed to exist in bent conformations in living cells and are assumed to be under compression. Individual buckled microtubules were severed using a femtosecond laser to directly determine the nature of the microtubule force balance. Newly formed minus-ended microtubules did not straighten to release bending energy but rather increased in curvature. Interestingly in dynein inhibited cells newly formed minus-ended microtubules were observed to straighten. These results suggest a model where dynein both stabilizes and enhances microtubule buckles. Overall this work considerably advances the understanding of cytoskeletal mechanics in endothelial cells.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by ROBERT J RUSSELL.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Lele, Tanmay.
Local: Co-adviser: Dickinson, Richard B.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042874:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042874/00001

Material Information

Title: Cytoskeletal Mechanics in Endothelial Cells
Physical Description: 1 online resource (147 p.)
Language: english
Creator: RUSSELL,ROBERT J
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ABLATION -- ACTIN -- ENDOTHELIAL -- FIBERS -- LASER -- MECHANICS -- MYOSIN -- SARCOMERES -- STRESS
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: An important function of normal endothelial cells is the ability to sense and transduce applied mechanical forces due to blood flow through the vasculature. Defects in endothelial mechano-sensing result in numerous pathologies including atherosclerosis. The actomyosin and microtubule cytoskeleton gives endothelial cells structural support and the ability to both sense and respond to external forces. This dissertation focuses on how the actomyosin and microtubule cytoskeleton generate and respond to forces in endothelial cells. Tensile force within endothelial cells is generated in actomyosin stress fibers, which are composed of contractile units called sarcomeres. Using femtosecond laser ablation, we severed living stress fibers and measured sarcomere contraction under zero tension. The time dependent trajectories of contracting sarcomeres suggested three distinct phases: an instantaneous initial response, a sustained change in length at constant velocity and finally a steady state. We proposed a novel model for tension generation in the sarcomere where myosin generated forces in adjacent sarcomeres are directly in balance. Through live cell imaging we observed that the number of sarcomeres and sarcomere lengths dynamically change in the cell. We show that sarcomere lengths continually fluctuate, new sarcomeres formed at focal adhesions are convected into the stress fiber and that sarcomeres disappear at specific points or ?sinks? along the stress fiber. These results show that stress fibers are highly dynamic structures despite their relatively static morphology. Taken together the observation of sarcomere fluctuations and the sarcomere contraction results suggest a model where tension fluctuations determine the distribution of sarcomere lengths. Microtubules are the stiffest of the cytoskeletal filaments but are consistently observed to exist in bent conformations in living cells and are assumed to be under compression. Individual buckled microtubules were severed using a femtosecond laser to directly determine the nature of the microtubule force balance. Newly formed minus-ended microtubules did not straighten to release bending energy but rather increased in curvature. Interestingly in dynein inhibited cells newly formed minus-ended microtubules were observed to straighten. These results suggest a model where dynein both stabilizes and enhances microtubule buckles. Overall this work considerably advances the understanding of cytoskeletal mechanics in endothelial cells.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by ROBERT J RUSSELL.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Lele, Tanmay.
Local: Co-adviser: Dickinson, Richard B.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042874:00001


This item has the following downloads:


Full Text

PAGE 1

1 CYTOSKELETAL MECHANICS IN ENDOTHELIAL CELLS By ROBERT JOHN RUSSELL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOS OPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Robert John Russell

PAGE 3

3 To my parents and my lovely wife

PAGE 4

4 ACKNOWLEDGMENTS I would lik e to acknowledge the people who supported me during m y time at the University of Florida and through my doctoral studie s. I first would like to acknowledge my advisor Professor Tanmay Lele and my co advisor Professor Richard Dickinson. Professor Lele was supportive and patient throughout my time here and his expertise was extremely helpful in both selecting problems and co mpleting this work. Professor Dickinson provided an insightful perspective for my research and provided me experience with applying models to biological systems. B oth advisors have provided me a great opportunity t o learn how to conduct research and improv e my knowledge of chemical engineering I would like to thank the other members of my supervisory committee. Professor Benjamin Kes e lowsky has been supportive and offered helpful advice. I would like Professor Yiider Tseng for allowing me to start on my re search in his lab as well being willing to discuss problems and offer advice as needed. I would also thank Professor Daniel Purich for holding the Cytoskeleton Journal Club through which I gained much insight into cytoskeletal processes. Professor Shen Lin g Xia provided me access and help with the multi photon confocal microscope at the VA hospital. I would also like my coworkers in the Lele Lab and other students that helped throughout my time at the University of Florida. First I would like to express my appreciation to Dr. Hengyi Xiao who taught me the necessary biology lab skills to get started doing research. Pei Hsun Wu, T J Chancellor, Jun Wu and Jiyeon Lee were always helpful in getting the lab running, figuring out new experimental techniques trouble shooting problems and discussing results. Jun and I had enjoyable time working on the microtubule project together. David Lovett and Alex Grubbs were supportive and helpful during my the end of my time in the lab. I would like to thank the undergraduates

PAGE 5

5 t hat worked with me over the years most specifically Sunil Magroo and Sandra Nakasone. I would like to thank all of my family members that supported me throughout the years and made this a reality. My parents have always been encouraging, loving and have mo tivated me to achieve my goals. They have always been willing to listen to me and give good advice on how to move forward. I would like to thank my brother for being a great friend and being supportive during my time in Florida. been helpful to us during our time in Gainesville which we both appreciate. Lastly I would I like to acknowledge the support and love of my wife Rose. Her patience, resourcefulness and success have always inspired me, and her encouragement and understanding ha ve always motivated me. Without Rose I am certain that this would not have been possible and certainly not as enjoyable. I look forward to our future as we take another step together in life.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF TERMS ................................ ................................ ................................ ........... 11 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 16 Function of the Endothelium ................................ ................................ ................... 16 Endot helial Cytoskeleton ................................ ................................ ........................ 17 Actin Cytoskeleton ................................ ................................ ............................ 17 Microtubule Cytoskeleton ................................ ................................ ................. 18 Evidence for Stress Fibers in the Intact Endothelium ................................ ............. 19 Microstucture of Endothelial Stress Fibers ................................ .............................. 21 Mechanics and Dyn amics of Stress Fibers ................................ ............................. 23 Sarcomere Mechanics Models ................................ ................................ ................ 25 Mechanics of Microtubules ................................ ................................ ..................... 26 Organization of this Document ................................ ................................ ................ 27 2 SARCOMERE MECHANICS IN CAPILLARY ENDOTHELIAL CELLS ................... 33 Materials and Meth ods ................................ ................................ ............................ 36 Cell Culture and Transfection ................................ ................................ ........... 36 Laser Ablation of Stress Fibers ................................ ................................ ........ 37 Image Correlation Based Tracking of Dense Bodies ................................ ........ 39 Parameter Estimation for Sarcomere Model ................................ ..................... 41 Results and Discussio n ................................ ................................ ........................... 42 Sarcomeres Contract in the Severed Stress Fiber ................................ ........... 42 Potential Energy is not Stored in Severed Stress Fibers ................................ .. 44 A Mechanical Model for the Sarcomere ................................ ........................... 45 Insight into the Mechanics of Stress Fibers ................................ ...................... 48 Summary of Findings ................................ ................................ .............................. 49 3 SARCOMERE LENGTH FLUCTUATIONS AND FLOW IN CAPILLARY ENDOTHELIAL CELLS ................................ ................................ ........................... 58 Materials and Method s ................................ ................................ ............................ 59 Cell Culture and T ransfections ................................ ................................ ......... 59 Micropatterning of Bovine Capillary Endothelial Cells ................................ ...... 60 Time Lapse Imaging of Labeled Sarcomeres ................................ .................. 61

PAGE 7

7 Image Analysis ................................ ................................ ................................ 61 Results ................................ ................................ ................................ .................... 62 Fluctuations in Sarcomere Length ................................ ................................ .... 62 Flow of Nascent Sarcomeres from Focal Adhesions ................................ ........ 63 Sarcomere Flow Velocity from Focal Adhesions is Tension Independent ........ 64 ................................ .............................. 64 Discussion ................................ ................................ ................................ .............. 65 Summary of a Comprehensive Sarcomere Model ................................ .................. 67 4 EFFECTS OF DYNEIN ON MICROTUBULE MECHANICS ................................ ... 78 Materials and Methods ................................ ................................ ............................ 79 Cell Culture, Plasmids and Transfection ................................ .......................... 79 Laser Ablation ................................ ................................ ................................ .. 80 Root Mean Squared Curvature Calculations ................................ .................... 82 Results ................................ ................................ ................................ .................... 84 Dynamics of Severed Micro tubules ................................ ................................ .. 84 Dynein Inhibition Alters the Dynamics of Severed Microtubules ....................... 86 Discussion ................................ ................................ ................................ .............. 87 Summary of Findings ................................ ................................ .............................. 89 5 CONCLUSIONS ................................ ................................ ................................ ..... 96 Summary of Findings ................................ ................................ .............................. 96 Stress Fiber Sarcomere Mechanics ................................ ................................ 97 Sarcomere Dynamics in Endothelial Cells ................................ ........................ 98 Effects of Dynein on the Microtubule Force Balance ................................ ...... 100 Future Work ................................ ................................ ................................ .......... 100 Biophysical Analysis of Mechanical Factors Leading to Atherosclerosis ........ 100 Force Transmission B etween Endothelial Cells in a Mono Layer ................... 102 Mechanics of Myofilament Repair in C2C12 Myoblasts ................................ 104 APPENDIX A MATLAB CODE FOR ANALYZING SARCOMERE DYNAMICS .......................... 107 Dense Body Tracking Software ................................ ................................ ............ 107 Correlation Processing Function ................................ ................................ .... 113 Paraboloid Fitting Function ................................ ................................ ............. 113 Averaging Position Software ................................ ................................ ................. 115 Sarcomere Length Calculation ................................ ................................ .............. 119 Model Fitting Code ................................ ................................ ................................ 122 Four Parameter Fitting ................................ ................................ .................... 123 Two Parameter Fitting ................................ ................................ .................... 125 B MATLAB CODE FOR DETERMINING MICROTUBULE CURVATURE ............... 127 Microtubule Tracing Code ................................ ................................ ..................... 127

PAGE 8

8 Linear Gaussian Fitting Function ................................ ................................ .... 133 Curvature and Bending Energy Calculatio n Code ................................ ................ 134 LIST OF REFERENCES ................................ ................................ ............................. 1 37 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 147

PAGE 9

9 LIST OF FIGURES Figure page 1 1 Schematic of forces exerted on the endothelium in the vasculature .................. 29 1 2 Elements of the cellular cytoskeleton ................................ ................................ 30 1 3 Microstructure of stress fibers consists of sarcomeric subunits ......................... 31 1 4 Microtubules exhibit short wavelength buckling at the c ell periphery ................. 32 2 1 Sarcomeres contract in a severed fiber ................................ ............................. 50 2 2 Cartoon schematic of microscope system used for ablation an d imaging ......... 51 2 3 Time dependent length change of a sarcomere ................................ ................ 52 2 4 Histograms of sarcomere contraction parameters ................................ ............. 53 2 5 Potential energy is not stored in a severed stress fiber ................................ ..... 54 2 6 Second example of potential energy not sto red in a severed stress fiber .......... 55 2 7 Fully contracted severed stress fibers do not relax when treated with ML7 ....... 56 2 8 Proposed mec hanical model for th e sarcomere ................................ ................ 57 3 1 Sarcomeres unde rgo dynamic length fluctuations ................................ ............. 71 3 2 Sarcomeres undergo length fluctuations in cells with confined shaped ............. 72 3 3 Nascent sarcomere s flow from all focal adhesions ................................ ............ 73 3 4 Nascent sarcomeres flow inwards from all fo cal adhesions in patterned cells .. 74 3 5 Sarcomere flow rate is i ndependent of mechanical stress ................................ 75 3 6 Sarcomeres are consumed at sinks and join end on to maintain tension .......... 76 3 7 Sarcomeres a re consumed to maintain tension ................................ ................. 77 4 1 Depolymerization rates dif fer between plus end ed and minus ended microtubules ................................ ................................ ................................ ....... 90 4 2 Minus ended microtubule fragments increase in curvature after severing ......... 91 4 3 There is no correlation observed between the change in RMS curvature and either the initial RMS curvature or position of the microtubule ............................ 92

PAGE 10

10 4 4 Bent microtubules are stable over si gnificant time scales ................................ 93 4 5 Severed microtubules in dynein inhibited cells were observed to straigthen ..... 94 4 6 Cartoon schemat ic of proposed model of dynein force generation in microtubules ................................ ................................ ................................ ....... 95 5 1 Severed myofilaments in C2C12 myobl asts display a repair mechanism ........ 106

PAGE 11

11 LIST OF TERMS ATP A denosine T ri phosphate N ucleotide BCE Bovine Capillary Endothelial Cells Autocorrelation function of sarcomere fluctuations CCD Charge Coupled Device Effective diffusion coefficient of sarcomere fluctu ations DBS Donor Bovine Serum DMEM EC Endothelial Cells ECM Extracellular Matrix EM Electron Microscopy EGFP Enhanced Green Fluorescent Protein EYFP Enhanced Yellow Fluorescent Protein FGF Fibroblast Growth Factor Force generated by myosin molecules Force exerted by stiff elastic element in series with myosin molecules G Actin Globular Actin HEPES 4 (2 hydroxyethyl) 1 piperazineethanesulfonic acid Ini tial length of stress fiber Time dependent length of stress fiber Minimum length of stress fiber after retraction MEM Minimum Essential Medium Eagle MTOC Microtubule Organizing Center NA Numerical Aperture

PAGE 12

12 Probability distribution of contraction distances PDMS Poly(dimethyl)siloxane PIV Particle Image Velocimetry Ptk2 Male Rat Kangaroo Kidney Epithelial Cells Relative retraction length L ine segment lengths used for curvature calculation SF Stress Fibers Instantaneous stress fiber tension Mean stress fiber tension Instantaneous tension in ith sarcomere External tension in stress fiber Tension at which contraction stalls U2OS Human osteosarcoma cell line UV Ultraviolet Linear retraction velocity Instantaneous velocity of ith s arcomere Drift velocity Instantaneous velocity of ith sarcomere VAS P Vasodilator Stimulated Phosphoprotein Linear contraction distance Initial sarcomere length Steady state sarcomere length upon reaching barrier Time dependent sarcomere length

PAGE 13

13 Z lines Borders that link and separate penultimate muscle sarcomeres Proportionality constant f or sarcomere tension velocity relationship Mean contraction distance Initial contraction distance after release of elastic element Local curvature of a microtubule Standard deviation of experimentally observed sarcomere length fluctuations Variance of sarcomere tension fluctuations Variance of external tension fluctuations Variance of velocit y fluctuations Relaxation time constant of sarcomere length fluctuations Relaxation time constant of sarcomere velocity fluctuations

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the Universi ty of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CYTOSKELETAL MECHANICS IN ENDOTHELIAL CELLS By Robert John Russell May 2011 Chair: Tanmay Lele Cochair: Richard Dickinson Major: Chemical E ngineering An important function of normal endothelial cells is the ability to sense and transduce applied mechanical forces due to blood flow through the vasculature. Defects in endothelial mechano sensing result in numerous pathologies including atheros clerosis. The actomyosin and microtubule cytoskeleton gives endothelial cells structural support and the ability to both sense and respond to external forces. T his dissertation focuses on how the actomyosin and microtubule cytoskeleton generate and respond to forces in endothelial cells. Tensile force within endothelial cells is generated in actomyosin stress fibers, which are composed of contractile units called sarcomeres. Using femtosecond laser ablation, we severed living stress fibers and measured sarc omere contraction under zero tension. The time dependent trajectories of contracting sarcomere s suggested three distinct phases: an instantaneous initial response a sustained change in length at constant velocity and finally a steady state. We proposed a novel model for tension generation in the sarcomere where myosin generated forces in adjacent sarcomeres are directly in balance

PAGE 15

15 Through live cell imaging we observed that the number of sarcomeres and sarcomere lengths d ynamically change in the cell. We s how that sarcomere lengths continually fluctuate n ew sarcomeres formed at focal adhesions are convected into the stress fiber and that sarcomeres fiber. These results show that stress fibers are hig hly dynamic structures despite their relatively static morphology Taken together the observation of sarcomere fluctuations and the sarcomere contraction results suggest a model where tension fluctuations determine the distribution of sarcomere lengths. Mi crotubules are the stiffest of the cytoskeletal filaments but are consistently observed to exist in bent conformations in living cells and are assumed to be under compression. Individual buckled microtubules were severed using a femtosecond laser to direct ly determine the nature of the microtubule force balance. Newly formed minus ended microtubules did not straighten to release bending energy but rather increased in curvature. Interestingly in dynein inhibited cells newly formed minus ended microtubules we re observed to straighten. These results suggest a model where dynein both stabilizes and enhances microtubule buckles. Overall this work considerably advances the understanding of cytoskeletal mechanics in endothelial cells.

PAGE 16

16 CHAPTER 1 INTRODUCTION Fun ction of the Endot helium The endothelium serves as a semi permeable membrane that lines the whole of the vasculature (1) The primary functions of the endothelium are to regulate transport between the blood stream and the surrounding ti ssues, and provide a smooth boundary condition to reduce turbulent blood flow (1 4) The endothelium is hetero gen e ous in function depending on the location with varying degrees of permeability (2 4) i e extremely low permeability at blood brain barrier (5) and higher permeability elsewhere (6 8) Interestingly all of this regulation that is crucial for normal physiological function is accomplished with a single la yer of specialized endothelial cells Alterations in the permeability of the vascular endothelium can lead to significant pathologies including acute lung injury, acute respiratory distress syndrome and sepsis (9) Endothelial dysfunction has also been implicated in hypercholesterolemia, atherosclerosis myocardial ischemia, and chronic heart failure (10) While biochemical signaling pathways have be en identified in static cell culture that lead to endothelial dysfunction (11) increasingly the importance of the endothe lial cells ability to respond to mechanical stress is being considered (12,13) The endothelium experiences varying degrees of mechanical stresses depending on the location and function (9) Generally endothelial cel ls are exposed to shear stress from blood flow and a cyclic wal l strain from the compliant vessels undergoing pulsatile blood flow (Figure 1 1 illustrates these external stresses) (13) Endothelial cells in the microvasculature must also resist frequent collisions from red and white blood cells (14) Evidence from both in situ observations and in vitro experiments suggest t hat

PAGE 17

17 endothelial cells dynamically reorient themselves to minimize exposure to stresses (15,16) The mechanisms by which endothelial cells adapt to these app lied stresses are not well understood (9) The endothelial cytoskeleton provides structural support and the ability to transduce external mechanical forces. As a result the cytoskeleton pla ys a crucial role in how endothelial cells respond and adapt to mechanical forces. Endothelial Cytoskeleton The endothelial cellular cytoskeleton consists of three distinct subsystems of biopolymer filaments which have varying mechanical properties an d functional roles. These subsystems include actin microfilaments, microtubules, and intermediate filaments, visual e xamples are provided in Figure 2 2. Two of these systems, actin filaments and microtubules, have associated molecular motors that both gene rate force in these systems and transport cargo. Actin C ytoskeleton Actin is a globular protein of approximately 42 kDa found in abundance in all eukaryotic cells (17) The ability of actin to directionally polymerize i nto helical filaments results in it being a key structural protein for almost all types of cells (18) Actin filaments and their associated proteins such as the motor protein myosin and the cross linking protei actinin make up the actin cytoskeleton (19) The actin cytoskeleton has many importan t physiological functions both at the cellular and tissue level (2) The actin cytoskeleton plays a substa ntial role in endothelial function including regulating permeability (2) There are generally considered t o be three distinct subsections of the actin cytoskeleton in the intact endothelium including the cortical actin rim, the membrane skeleton and actin stress fibers (2) In tissue culture these structures are

PAGE 18

18 similar except the cortical actin rim is replaced by region s of active actin polymerization known as lamellipodi a (20) T his document will focus on stress fibers which are force generating and transmitting structures that provide the endothelial cells with structural support and the ability to resist mechanical stresses. Stress fibers have been observed in almost all cell types in tissue culture and appear as continuous cables when stained with actin (Figure 2 2A) (21) Stress fibers either connect with other stress fibers or terminate at focal adhesions (Figure 2 2 A) (22,23) Focal adhesions serve to connect the cell through integrin linkages to the ECM and promote strong adhesion to the basement membrane (24) Furthermore focal adhesions play both a mechanosensory and signaling role for adhesion dependent cell types (24) While stress fibers and focal adhesions are ob served in almost all cell types in tissue culture, stress fibers are observed only in endothelial cells in vivo (25) Microtubule C ytoskeleton Microtubules are long tubular biopolymers (Figure 2 2B) that are approximately 25 nm in diameter In mammalian cells microtubules generally consist of 13 protofilaments arranged radially around the hollow core (26) Protofilaments are made up of tubulin monomer tubulin monomer (27) The tubulin heterodimers are arranged such that an tubulin monomers persists throughout the length of the protofilaments (27) This polarity results in the microtubule as a whole being polarized such that tubulin preferentially adds t o one end ( referred to as the plus end ) over the other end ( referred to as the minus end ) (26) The depolymerizat ion rate of the plus end is significa ntly faster then the minus end (28) The plus end of a

PAGE 19

19 microtubule randomly switches from growing to shrinking phases in a process known as dynamic instability (28) Microtubules play critical roles in normal cell functions including the formation of mitotic and meotic spindles, stabili zation of protrusions, and polarization of cells (29) The directionally persistent molecular motors dynein (minus end directed) and kinesin (plus end directed) are responsible for the transport of organelles and proteins (29) In mammalian cells it has been established that most microtubules emanate fr om a central microtubule organizing center, MTOC, or the centrosome (29,30) At the MTOC, minus ends of the microtubule are both nucleated and stabilized resulting in a star shaped radial array of microtubules (Figure 2 2B) (30) Evidence for Stress Fibers in the Intact Endothelium There have been numerous studies that have reported the presence of stress fibers in the intact endothelium and have investigated the conditions under which they form (25,31,32) Herman and coworkers published evidence and novel observations of stress fibers in vascular endothelial cells in vertebrate tissues in 1982 (25) Using fluorescence microscopy of fluo rescently tagged phallacidin, anti myosin antibodies and anti actinin the authors searched for actin stress fibers in a variety of vertebrate tissues. Ultimately stress fiber like structures were found only in the vascular endothelial cells of cows, chic kens, dogs, rats and a cat. Specifically in regions of the vasculature subjected to high velocity and turbulent blood flow stress fibers were confirmed to show actinin and were oriented parallel with the flow direction The authors make some astute comments about the possible role stress fibers play physiologically including the actors that may govern their formation The authors suggest that stress fibers likely transmit forces to the basement membrane and that hemody namic forces may be responsible for the formation of stress fibers (25)

PAGE 20

20 Physiological factors influencing expression of stress fibers in vascular endothelial cells were investigated by Fujiwa ra and coworkers in 1986 (31) Their i nvestigation confirmed the existence of axially orient ed stress fibers in vascular endothelial cells specifically mouse and rat thoracic aorta. T hey reported an increase in stress fiber expression in male rats versus female rats and in same sex normotensive versus spontaneous hypertensive rats. These findings are consistent with the hypothesis that increased hemodynamic forces (male rats generally have higher blood pressure) result in the expression of stress fibers in vascular endothelial cells (31) Drenckhahn and Wagner investigated in situ stress fibers in splenic sinus (a channel of endothelial cells with increased permeability to allow the se lective passage of red and white blood cells) endothelial cells (32) This study was one of the first to demonstrate the in situ existence of stress fibers in endothelial cells outside of the large vessel vasculature. The authors established with electron m icroscopy that the molecular structure of stress fibers from isolated in situ endothelial cells matches that of cultured EC in that they displayed regions of alternating polarity actin filaments separated with myosin fragments. Using immunohistochemistry t he authors observed that fluorescently labeled antibodies to actin, m actinin localize strongly to the stress fibers while extracellular matrix proteins (laminin, fibronectin and collagen) were localized to the periphery of the cells. Furthermore the authors demonstrated that stress fibers are contra ctile in the presence of ATP and Mg 2+ in permeabilized cells (32) Further evidence that hemodynamic forces modulate the expression of EC stress fibers was presented by Gotlieb and coworkers through examination of in situ rabbit aorta bifurcations (33) Prior to bifurcations actin localized in EC to two distinct

PAGE 21

21 structures, a peripheral actin and ring and variable numbers of stress fibers. Along the divider wall EC s exhibited more of the stress fiber phenotype and less peripheral actin and with increasing distance from the bifurcation reverted to an even distrib ution in periphe ral and stress fiber actin. Kim et al. showed that increased hemodynamic forces caused an increased expression of stress fibers in an excised rabbit mid abdominal aorta (34) Briefly the authors altered the flow fields by surgically constricting flow in the mid abdominal aorta. In non constricted vessels as well as prior to the constriction actin was distributed into a peripheral ring and in a variable numbers of stress fibers Immediately after the constriction some EC cells lost directionality and all cells displayed actin stress fibers. Further (3 13 m m) from the constriction point cells recovered their directionality but still expressed stress fibers (34) While most studies init ially reported the expression of stress fibers in large arteries and heart chambers Nehls and Drenckhahm r eported the first evidence of stress fibers in microvascular endothelial cells (14) Their results suggest that stress fibers exist at all level s of the microvasculature including capillaries and venules (feeders to main vein) and hypothesize that microvascular endothelial stress fibers may protect the cells from shear stress and collisions of red and white blood cells (14) Microstucture o f Endothelial Stress Fibers Stress fibers are contractile cables that display a continuous staining for actin until their ultimate termination into focal adhesions see example in Figure 2 2A (35) There has been significant work done in the past to understand the microstructure of stress fibers from a variety of cells types (36 39) While some conflicting results have been reported between motile and non motile cell types the general findings for non motile cells hold true for the endothelial cells discussed in this document (40) The cross

PAGE 22

22 linking protein actinin is responsible for bundling the 10 30 parallel actin filaments (40) that make up the fiber together. Interestingly it was found through immunohistochemistry that anti actinin antibodies localize to stress fibers in a punctate pattern as illustrated in Figure 2 3 (37) These punctate spots were originally referred to in the literature as dense bodies (Fi gure 2 3 contains schematic) as they appeared as electron dense regions similar to muscle Z lines (35) Further studies have identified multiple other proteins that localize to stress fiber dense bodies including filamin (41) fascin (42) and VASP (43) Staining for myosin or tropomyosin reveals a punctate pattern that is offset from the dense bodies (36,38,44) These findings led to the hypothesis that stress fibers were composed of contractile subunits analogous to muscle sarcomeres. The idealized sarcomeric model of the stress fiber is such that each dense body serves as the boundary be tween adjacent sarcomeres (Figure 2 3 contains a schematic described below ) (36) A single sarcomere consists of opposite polarity actin filaments emanating inwards from t he respective dense bodies where the barbed ends are located (45) Bipolar myosin filaments cross link the opposing bundles of actin filaments (44) Contractile force is generated as the bipolar myosin filaments translocate in opposite directions towards the dense bodies. There exists significant experimental evidence that the idealized sarcomere structure is in fact realized in stress fibers of non motile cells but maybe more complicated for some stress fibers in motile cells (40,45) Kries and Birchmeier initially reported that stress fiber sarcomeres are contractile in cultured fibroblasts suggesting that a sliding filament mechanism must be pre sent (46) A key requirement of the sliding filament mechanism is the alternating polarity of

PAGE 23

23 actin filaments. Cramer et al. conducted an electron microscopy study of actin filament orientation in both motile fibroblasts and non m otile PtK2 cells (40) While an earlier study had reported alterna ting actin polarity from EM studies of PtK2 cell stress fibers (47) this study demonstrated that actin polarity alternated as a function of length with an average perio d of 0.6 m (40) Stress fibers in m otile cells were reported to contain a more complicated arrangement of actin filaments including both graded and unifo rm polarities While this was certainly an interesting finding stress fibers in motile cells are beyond the scope of th is document (40) Actin filaments are dynamic in that there is directional turnover (assembly at plus ends and disassembly at minus ends) of actin subunits with the unpolymerized pool of G actin. D ense bodies are locations of actin polymerization as demonstrated by imaging the stress fiber immediately after microinjection of labeled rhodamine actin (45) This findi ng is consistent with t he sarcomeric model which requires that the plus ends to be located in the dense bodies, in order for the directional myosin motor to function. Mechanics and Dynamics of Stress Fibers There has been considerable int erest in understanding the mechanical properties of stress fibers due to their importance in normal endothelial function. U ntil recently studying mechanics of stress fibers in living cells was difficult with earlier work done in permeabilized cells or wit h isolated stress fibers (21,48) The first evidence of the possibility of contractile stress fibers was in permeabilized systems containing stress fibers and cortical actin where a change in cellular shape was obse rved upon addition of Mg 2+ and ATP (46) As this study could not rule out the possibility of other cortical actin being responsible for the observed contraction Katoh et al. showed that isolated single stress fibers were indeed contractile when treate d with Mg ATP solution (21) When the

PAGE 24

24 contracted fibers were washed of the Mg ATP solution no relaxa tion or lengthening of the fiber was observed over the time course of 20 minutes (21) The authors a lso reported that treatment of isolated stress fibers with the myosin light chain kinase inhibitor KT5926 were not contractile in the Mg ATP solution suggesting that the stress fiber contraction is a myosin driven phenomen on (21) With the advent of transfectable fluorescent protein fusions t he ability to study stress fibers in living cells became a reality. In a recent paper Kumar et al. demonstrated for the first time that stress fibers are contractile in living cells (49) Stress fibers l abeled with EYFP actin in bovine capillary endothelial cells were severed using a femtosecond laser nanoscissor. The authors demonstrated that the observed retraction of cut fiber ends resulted from fiber shortening by cutting before branch points and phot obleaching fiduciary markers in the fiber. The authors found that the gap between retracting ends of the fiber followed exponential kinetics as would be expected by the severing of a stretched viscoelastic cable. Thus the characteristic time constant of th e T he recoil of severed stress fibers was partially inhibited when cells were treated with the rho associated protein kinase inhibitor Y27632 and was completely abolish ed with treatment with myosin light chain kinase inhibitor ML7. In summary this work was significant in that it clearly demonstrated the stress fibers are contractile in living cells and that stress fiber force generation is myos in dependent (49) Yin and coworkers used atomic force microscopy to probe the mechanics of stress fibers in living cells and found that myosin plays a integral role in de termining the apparent stiffness of the stress fiber (50) Stress fibers with decreased contractility, cells treated the myosin A TPase

PAGE 25

25 inhibitor blebbistatin, had decreased stiffness while fibers with increased contractility, cells treated with the serine/threonine phosphatase inhibitor calyculin A had an increase in stiffness (50) While these studies provided important insight into the mechanics of stress fibers in living cells they did not provide detail into the mechanics of individual stress fiber sarco meres. Peterson and coworkers showed that cells treated with calyculin A, a serine/threonine phosphatase inhibitor displayed both c ontraction and expansion of sarcom eres with in the same stress fibers (51) a result that prompted explanations based on theoretical mechanical models of stress fiber sarcomeres (52,53) In the work by Peterson and coworkers, actinin and myosin II regulatory myosin light chain transfected fibroblasts and fibroma cells were treated with calyculin A which resulted in overall stress fiber shortening (51) T he author reported that while sarcomeres at periphery shorted upon calyculin A treatment, central sarcomeres lengthened (51) Sarcomere Mechanics Models Two theoretical modeling studies were published to explain the inhomogeneous behavior of stress fibers reported by Peterson and coworkers (51) Besser and Schwarz proposed a mechanical model of the sarcomere consisting of an active force generating myosin element in parallel with both an elastic and viscous element (52) This study included a biochemical feedback loop for myosin activation from force induced signaling mod el that also included an elastic element in series with active myosin force generation but excluded the viscous element (53) A complication was added to this model in that a repulsive force of overlapping filaments was also included. While these

PAGE 26

26 studies were novel attempts to model the stress fiber as sarcomeric units in series, there is limited experimental evidence to support these mechanical models. Mechanics of Microtubule s Because of their hollow geometry microtubules are approximat ely 100 times stiffer then either actin filaments or intermediate filaments (54,55) Microtubules have been established to generate pushing force s as a result of their directional polymerization (56) Despite a thermal persistent length in the millimeter range, m icrotubules have been widely observed and documented to exist in bent shapes in cultured cells (29,55) see Figure 2 4 A Thus these buckled microtubules have been widely assumed to be under compressive loading. Polymerizing microtubules have been shown to buckle upon i mpinging on an immovable boundary in in vitro studies (56) and in vivo experiments (55) see Figure 2 4B I t has been hypothesized that microtubules play can balance the te nsional forces generated by acto myosin systems (57 59) An issue with this hypothesis is that microtubules reach significant lengths in cells which cannot support the large forces generated. An opposing viewpoint is that microtubules do not a play a structural role in interphase animal cells but only ul ar tra f ficking. Recently there has been increasing interest in determini ng the cause for the bent shapes of microtubules (29,55,60) Microtubules have been hypothesized to bend due to polymerization of a microtubule against an imm ovable boundary (55,56) actomyosin contractility (61,62) and microtubule based motors interacting with cortical actin (63 67) Recently Wei tz and coworkers proposed that bu ckled microtubules could support larger compressive loads if they were laterally supported by an elastic medium (55) Their findings suggest that

PAGE 27

27 single microtubules buckle with shorter wavelengths then expected for classic Euler buckling under a compress ive load. Their conclusion was that single microtubules supported by an elastic matrix would be able to support significantly larger compressive loads and exhibit the observed short wavelength buckles (55) Another study investigated the mechanism involved in microtubule buckling through direct observ ation of buckling microtubules in living cells (29) By performing fluorescence speckle microscopy Odde and coworkers found th at buckling microtubules move in the direction of the cell periphery, instead of towards the cell center which would be suggestive of polymerization drive n buckling. Furthermore the authors found that the F actin surrounding buckling microtubules was stationary which suggests that the actomyosin activity or actin retrograde flow do not play a substantial role in buckling microtubules. The authors hypothesiz e d that the anterograde driven buckling is indicative of microtubule based molecular motors (29) While this paper provides evidence that microtubule based motors are responsible for buckling it is still unclear which motor is involved and the mechanism by which the motor acts to buckle microtubules. Organization of t his Document The endothelial cytoskeleton plays a crucial role in normal cell function but fundamental questions remain about how ensembles of molecular motors generate forces in both actomyosin and microtubule cyto skeleton. This dissertation focuses on how the actomyosin and microtubule cytoskeleton generate and respond to forces in endothelial cells. In Chapter 2, a novel technique of femtosecond laser ablation is applied to study the mechanics of stress fiber sarc omeres in Bovine Capillary Endothelial Cells. A novel mechanical model for the stress fiber sarcomere is proposed

PAGE 28

28 and the implications of this model on the overall behavior of stress fibers are discussed. In Chapter 3 novel observations of the dynamics of stress fiber sarcomeres are reported and these observations are discussed in the terms of the model proposed in Chapter 2. In Chapter 4 the dynamics of severed microtubules and the role of dynein in stabilizing and enhancing microtubule buckles will be dis cussed. I n Chapter 5 the results from this document are summarized and future work is proposed to continue this study.

PAGE 29

29 Figure 1 1. Schematic of forces exerted on the endothelium in the vasculature. Pressure driven blood flow results in shear stress and a normal force against the endothelium The complia nt vasculature undergoes cyclical circumferential wall strain as result of the pulsatile contractions of the heart.

PAGE 30

30 Figure 1 2 Elements of the cellular cytoskeleton (A) Micrograph of BCE cell st ained with phalloidin for actin stress fibers (red) and immunostained for vinculin (green) in focal adhesions. (B) Micrograph of BCE cell infected with adenoviral EGFP tubulin displaying microtubules emanating from the centrosome. (C) Micrograph of BCE cell transfected with EGFP vimentin to visualize the intermediate filament network. Scale Bars

PAGE 31

31 Figure 1 3 Microstructure of s tress fibers consists of sarcomeric subunits A micrograph of BCE cell transfected with EGFP actinin which results in a punctate staining of dense bodies. The cartoon schematic illustrates the microstructure of a single stress fiber sarcomere which marke d by two penultimate dense bodies.

PAGE 32

32 Figure 1 4. Microtubules exhibit short wavelength buckl ing at the cell periphery. (A) Microtubules buckle with short wavelengths (~ 2 3 ) at the periphery. (B) Time series of a single microtubule a s it contacts the periphery and undergoes buckling. Scale bars are 2 m.

PAGE 33

33 CHAPTER 2 SARCOMERE MECHANICS IN CAPILLARY ENDOTHELIAL CELLS Tension generation inside cells and its transmission to the extracellular matrix at cell matrix adhesions enables cells to adhere, spread, migrate and maintain tissue form Tension is generated in vascular endothelial cells by actomyosin stress fibers. While stress fibers are observed in many cell types cultured in vitro stress fibers in vivo are observed in endothelium of intact tissues such as the aorta (25) the spleen (32) and the eye (68) Tension in stress fibers promotes strong adhesion between endothelial cells and the basement membrane (32) This adhesion allows endothelial cells to resist blood flow induced mechanical stresses including cyclic strain, hydrostatic pressure and sh ear flow (25,32,68 71) Endothelial cells respond to applied mechanical stresses by changing their orientation both in tissue (69,70) and culture (71 74) which depends on remodeling of stress fibers (71) Thus stress fibers are necessary for resisting and sensing mechanical stresses in vascular endothelial cells Stress fiber tension is also greatly altered in tumor endothelial cells which have aberrant response to mechanical stresses (75) Stress fiber tension is implicated in abnormal endothelial response to altered hemodynamic forces that cause atherosclerotic plaques at vascular bifurcations (76) Owing to their key importance as a contractile structure that enables normal endothelial cell function, there is considerable rec ent interest in understanding the mechanical properties of stress fibers (48,49,52,53) However, it has remained a fundamental challenge to measure these properties in the context of a living, func tioning cell. Stress fibers are composed of repeating units termed sarcomeres after the analogous muscle structures (36 38,4 7,77,78) Using electron microscopy, Mitchison

PAGE 34

34 and co workers confirmed the sarcomeric microstructure of stress fibers in non motile PtK2 cells by showing that they contain actin filaments of alternating polarity along the entire length of the fiber (40) Tension is generated in the sarcomere by walking of bipolar myosi n heads in opposite directions, resulting in contraction of the sarcomere ( (32,46,79) and see Figure 2 1 A for a schematic). Neighboring sarcomeres are connected by dense bodies containing actinin, fila min and VASP (Figure 2 1 B). Physical coupling between stress fibers and focal adhesions allows transfer of the tension to the substrate (80) Recently proposed mechanical models for stress fiber sarcomeres (52,53) assume that velocity dependent myosin contraction operates in parallel with an elastic element, analogous to the Kelvin Voigt model (spring and dashpot in parallel) for viscoelasticity A feature of these models is that ex panded or compressed elastic elements partially balance the tension and account for the distribution of sarcomere sizes (52,53) These mechanical models have been used to explain the experimental observation of Peterson et. al (51) that sarcomeres contract near focal adhesions on activation of myosin li ght chain kinase (52,53) However, to date no direct experimental evidence has been offered to test current mechanical models of SF sarcomeres. While the majority of experimental data on stress fiber mechanics is from in vitro experiments of extracted stress fibers (21,48) a recent study used a femtosecond laser to sever individual stress fibers allowing the first in v ivo measurements of stress fiber mechanics (49) Cutting the fiber caused the severed ends to retract (see Figure 2 1 C for an example). The retr action of the fiber exhibited an exponential relaxation characteristic of that following tension loss in a Kelvin Voigt element (49) with a

PAGE 35

35 rel axation time constant of ~6 seconds. Myosin inhibition completely eliminated retraction in severed stress fibers (49) While this work pioneered a new approach to measure mechanics of stress fibers in living cells, it did not provide a mechanistic explanation for the observed mechanical response. For example, the molecular origin of the apparent viscosity of the fiber is unknown, and it is unclear whether the observed lack of retraction under myosin inhibition can be explained by existing sarcomere models. To explain this viscoelastic retraction of the severed ends, a recent model represented the stress fiber as a tensegrity structure containing te nsed and compressed elements (81) However, this model did not involve a sarcomeric description of the stress fiber. W e used femtosecond ablation to sever individual fibers and dire ctly measured sarcomere contraction by tracking the position of actinin labeled dense bodies in the severed fiber. The lengths of individual sarcomeres decreased in two phases: an instantaneous initial decrease, followed by a linear contraction at consta nt speed. The latter phase, interpreted as active myosin mediated contraction, ceased abruptly after a minimum sarcomere length was achieved. The linear response suggests that there is no increase in resisting force (for example, due to a spring element) d uring contraction. In addition, subsequent inhibition of myosin following severing and contraction yielded no elastic recovery, suggesting that elastic potential energy is not stored in the contracted fiber. Together, these observations argue against model s that have elastic elements in parallel with myosin contractile elements. Based on these findings, we propose a new and simpler mechanical model for the sarcomere in which stress fiber tension is determined only by myosin contraction in series with a stif f elastic element. The

PAGE 36

36 fundamental difference between this model and previous models is that all tension in the stationary stress fiber is borne entirely by the myosin contractile elements in each sarcomere, with no apparent elastic element existing in par allel with the myosin contractile element. Materials and Methods Cell C ulture and T ransfection Bovine capillary endothelial (BCE) cells were used at passages 11 14 and were maintained at 37C in humidified 10% CO2 The BCE cells were cultured on tissue c ulture dishes in complete medium consisting of low bovine serum (Gibco), 1% 1M HEPES (Mediatech) and glutamine (0.292 mg/ml)/ penicillin (100 U/ml)/ stre ptomycin (100 g/ml) (Sigma) The growth media was supplemented with basic FGF ( final concentration 2 ng/ml; Sigma) per 1 ml of media. This culture method was described previously in (82) For experiments cells were transiently transfected with EGFP actinin plasmid (kindly provided by Prof. Carol Otey) or EGFP actin using the Effectene (Qiagen, Valencia, CA) reagent. Previous studies have shown that the fusion construct is functional and localizes with the endogenous protein (83) Transfections were done approximately 2 full days before any planned microscopy experiments. C ells were transfected for 6 hours in a 12 well cell culture dish using the Effectene transfection kit For transfections 1 sufficient for both the EGFP actinin and EFGP actin plasm ids. The manufacturer supplied protocol was optimized and used for all transfections. After 24 h ours the cells

PAGE 37

37 were trypsinized (Gibco) and plated sparsely on glass bottom dishes (MatTek, Ashland, TX). For imaging experiments the media was changed to a CO 2 independent me dia as described previously (84) Briefly the CO 2 independen t imaging medium (pH 7.3) was prepare d by dissolving Hanks balanced salts, which are phenol red and bicarbonate free, in 1000 ml of DD H 2 O. The f ollowing supplements were added; HEPES (20.0 mM ), 1% bovine serum albumin, and MEM essential and nonessential am ino acids (Sigma). Imaging media was prepared by adding 10% DBS and glutamine (0.292 mg/ml)/ penicillin (100 U/ml)/ streptomycin (100 g/ml) (Sigma) Laser A blation of S tress F ibers For laser ablation experiments, an inverted Zeiss Axiovert 200M laser scan ning confocal microscope (LSM 510 NLO, Thornwood, NY) was used with 63X, 1.4 NA Plan Approchromatic oil immersion lens ( Zeiss) In order to image E GFP actinin, the 488 nm laser lin e with the power attenuated to 5 10% and the appropriate E GFP bandpass filter was used Single stress fiber ablation was done with a Ti:Sapphire laser at 50 100 % transmission (Chameleon XR, Coherent, Santa Clara, CA ) as p reviously described (49) The Ti:Sapphire laser was focused through the objective and scanned a thin, ~0.14 m, rectangle orthogonally crossing the width of the stress fiber for 1 7 iterations. A wavelength of 790 nm was used with a laser head power of 1.5 W, pulse duration of 140 fs and repetition rate of 90 MHz After ablation confocal scans as described previously were collected using Zeiss LSM 510 4.2 software at 100 1000 ms/frame to capture the kinetics of sarcomere contraction. A cartoon schematic of the microscope used for the ablation and imaging is included in Figure 2 2 Detailed methods used for ablation of single stress fibers follow in the next paragraph.

PAGE 38

38 Once cells were placed on the microscope and the co rrect focal plane was found the laser parameters for cutting were optimized. The separate lasers that were used for ablation and imaging must be aligned correctly to focus on the exact same position in the z direction for optimal cutting. Cha nges in the alignment resulted in slightly different laser parameters being needed for optimal ablation from day to day. Before experiments were conducted a cell expressing clear stress fibers was f ound and used to test laser parameters for cutting. T he FRAP module included with the Ze i ss microscope software was used to conduct ablation by scanning with the Ti:Sapphire laser at 790 nm. To optimize laser parameters the laser was initially set at 100% transmission power and to scan for one iteration. These conditions were tested by ablating a single fiber. In the case w h ere a fiber was cut and no collateral damage was observed in the DIC image then experiments were continued using these settings. When c oll ateral damage was observed the laser power was reduced incrementally by 10 % until a clean cut occur red. If the fiber would not cut at 100% transmission multiple scanning iterations were used. No more than approximately 7 scans were used because the cu t ting scans took too long which resulted in unacceptable time resolution. In the case where it was not possible to sever the stress fiber with multiple iterations the 488 nm Argon and the Ti:Sapphire lasers were realigned. To realign the lasers, a cell wa s focus ed using the 488 nm laser and then the imaging conditions were switched to image the cell using the Ti:Sapphire laser at 790 nm The vertical position of the TI:Sapphire laser was slowly adjusted until the cell came into focus. For myosin inhibitio minutes after stress fibers were severed and allowed to reach steady state. Confocal

PAGE 39

39 images were taken to determine if stress fibers were able to lengthen or relax after myosin poisoning. For all bl ebbistatin experiments exposure to the 488 nm laser was minimized to prevent photoinactivation or phototoxicity. Image C orrelation B ased T racking of D ense B odies To measure the change in length of the sarcomeres during stress fiber retraction an image correlation based tracking method was developed in MATLAB 7.2 (The MathWorks, Natick, MA) This method was used to determine the positions o f GFP actinin labeled dense bodies to sub pixel accuracy. Image sequences from the Zeiss software were exported to ImageJ (NIH) for image processing, images were rotated to make the stress fibers horizontal and then smoothed to remove noise. MATLAB was u sed for contrast stretching (MATLAB function strechlim) on the image sequences before analysis. The commented MATLAB code for tracking single dense bodies is provided in Appendices A.1 and A.2. The image correlation based method works as described previou sly in references (85,86) Briefly, a kernel was chosen containing the GFP actinin labeled dense body from the image at time t which was used for a correlation calculation (MATLAB function the correlation matrix. A paraboloid was fit to the correlation function around the peak in order to achieve sub pixel accuracy for the offset. In order to assign positions to the feature through time, an intensity weighted centroid calculation was done on the first frame to provide initial conditions to be updated by offsets from the correlation calculations. To minimize any error due to changes in the dense body shape, averaged positions were found by correlating each

PAGE 40

40 image with every other image and aver aging the result. Tracking a simulated particle indicated that the error was minimal, less than 1%, for the spatial offsets. As mentioned above, programs were used to track individual dense bodies in severed stress fibers. These programs were semi automat ed and require d select ion of the dense body in the initial frame but then track ed the feature through specified number of frames. Images were prepared for analysis in ImageJ as mentioned above by first cropping out a single stress fiber and then rotating i t until it wa s aligned horizontally. The Dense Body Tracking program provided in Appendix A.1 was run first on the data to get initial position estimates for the dense body in all frames. To run this program a single dense body was selected by drawing a r ectangle around it with the mouse. The program then use d this template to track the movement of the dense body through the time series. The initial tracking program prepare d a series of images w h ere the surrounding area of the tracked dense body is blacked out. The Averaging Position Software provided in Appendix A.2 was then run on the modified image series to get accurate position estimates. To run this code the dense body must again be selected by drawing a rectangle with the mouse. The output of this co de was the final averaged coordinates for the position of the dense body at each time points. The output was saved for later reference and directions on how to save the data generated were commented in to the code in Appendix A.2. After one dense body was r un through both programs neighboring dense bodies were then analyzed and the results were saved for each After multiple dense bodies were analyzed (between 2 5) the sarcomere trajectories with respect to time were found using the code provided in Append ix A.3.

PAGE 41

41 This code provide d both a composite image of all dense bodies labeled and the time dependent lengths of all the sarcomeres. The composite image generated by this program allow ed the user to check the fidelity of the tracking of each dense body. If at this step the tracking of one dense body was found to jump to a neighbor for one or multiple time frames then that dense body was reanalyze d Parameter E stimation for S arcomere M odel The sarcomere lengths measured during contracti on were fit with the four parameter model by minimizing the sum of the square of the residuals (MATLAB function fminsearch) to estimate the parameter set, The code used for fitting the four parameter model is provided in Appendix A.4.1. In the ca se where no sustained linear decrease in length was observed and a line could not be fit, a two parameter model was fit to the data with parameter set, The code used for fitting the four parameter model is provided in Appendix A.4. 2. In order to determine the uncertainty and correlation in the parameter estimates, the variance covariance matrix was found for each data set. To calculate the v ariance covariance matrix, the Jacobian m atrix, was constructed where This was used to calculate the variance covariance of the parameters according to this formula: (2 1 ) where are the resid uals. The parameter variances, lie along the diagonal of and were used to calculate a weighted mean and standard deviation:

PAGE 42

42 (2 2 ) (2 3) where the weighting factor was calcul ated using the parameter variances cal culated from Eq. 2 1 and the form Results and Discussion Sarcomeres C ontract in the S evered S tress F iber To measure sarcomere contraction in living BCEs, we expressed GFP actinin, which lab els stress fibers at junctions of neighboring sarcomeres (Figure 2 1B). Next, we severed individual stress fibers in living cells using a recently developed femtoseco nd laser ablation technique (49) A key feature of femtosecond laser ablation is that it minimizes collateral damage outside the ablated spot (34). On severing a stress fiber, individual sarcomere units in the fib er contracted (kymogr aph in Figure 2 1C). The severed end appeared to retra ct exponentially with time (Figure 2 1D), consistent with a pr evious study by Kumar et al. (49) Thus, t he net retraction of the severed edge corresponded to the contractions of individual sarcomeres as measured actinin labeled dense bodies. Although the severed edge of the cut fiber appeared to retract exponentially, the contract ion of individual sarcomeres was quite variable but shared qualitatively similar

PAGE 43

43 non exponential behavior. As shown in Fig ure 2 3 A, the contraction of individual sarcomeres from their initial length, to their final contracted lengt h, occurred in two distinct phases. First, a nearly instantaneous contraction occurred of distance, ~0 0.4 m at nearly constant speed. After the linear phase, the contraction ceased abruptly and the sarcomere remained constant at the final length for the duration of the experiment. Noting this consistent behavior, all sarcomere contraction trajectories were fit by least squares regression to the following piecewise linear model, (2 4) The fit all owed us to estimate the sa rcomere contraction parameters and The characteristic time for the transition from linear contraction to steady state was calculate d as For all analyzed sarcomeres from different stress fibers, normalized contraction was plotted against excluding initial sarcomere lengths, (Figure 2 3 B). As seen in Figure 2 3 B, t here is a clear linear trend followed by a stationary non contracting phase in the pooled, normalized contraction data. The initial and final lengths ( and respectively) were approximately Gaussian distribu 2 4 are the distributions of contraction distances. The distribution of the initial contraction distance was approximat ely

PAGE 44

44 symmetric with a mean and standard devia 2 4 A). The linear contraction distance was approximately exponentially distributed, with mean ure 2 4 B). The net contraction had a mean and standard devi 2 4 C) For sarcomeres with a signifi cant linear contraction regime, the fit to Eq. 1 yielded the velocity of the slower contraction phase. The mean and standard deviation of the contraction Poten tial E nergy is not S tored in S evered S tress F ibers To confirm that stress fiber retraction is myosin dependent, we severed stress fibers in cells pr e treated with blebbistatin for 30 minutes. As shown in Figure 2 5 A, stress fibers in cells treated with blebbistatin retracted insignificantly compare d to non treated cells (see Figure 2 5 B). This confirms the results of as reported by Kumar et. al. (49) that retraction in the sever ed fiber is myosin dependent. To determine the extent to which the sarcomere contraction could be reversed, we quan tified recovery of sarcomere length upon inhibition of myosin. Stress fibers were severed and sarcomeres were allowed to contract to a new steady state. Next, myosin activity was poisoned by treatment with blebbistatin (Figure 2 5B and 2 6 ) or ML7 (Figure 2 7 ) The severed fiber length was determined after myosin inhibition and compared with the severed fiber length bef ore inhibition. As seen in Figure 2 5 B, there was no visible lengthening of the contracted stress fiber several minutes time after blebbista tin treatment (see Fig ure 2 6 for another example ). This result was confirmed when cells were treated with the myosin light chain kinas e inhibitor ML7 (see Figure 2 7 ). These results show that there is no elastic energy stored in contracted sarcomeres

PAGE 45

45 whic h can be recovered upon inhibition of myosin contraction. This finding argues against the presence of an elastic element in parallel with myosin force generation. This finding is also in agreement with the results of Katoh et al. (21) who showed that isolated stress fibers cannot relax after contracting in a myosin dependent fashion. A M echanical M odel for the S arcomere The behavior of contracting sarcomeres following loss of tension, i.e. a near instantaneous initial retraction followed by slower linear contraction until a minimum length is reached can be explained with the simpl e mechanical model show n in Figure 2 8 In this model, an elastic element is in series with a myosin contractile element. Following severing, the sarcomere initially relaxes elastically, followed by slower myosin mediated c ontraction at a constant speed, which is likely limited by the maximum (unhindered) myosin motor velocity. The abrupt cessation of contraction suggests that either a strong resistance to further contraction is suddenly encountered, which could reflect a rigid steric barrier, or that myo sin motors have reached a limiting minimal distance from the dense bodies where the motors can no longer walk on actin filaments. Because myosin motors are sensitive to load, the uniform speed of the contraction phase suggests that elastic forces do not in crease (and are not relieved) during contraction. This observation, combined with the observation that elastic energy is not stored in contracted stress fibers, suggests that there is no elastic element in parallel with the contractile element, contrary to other recent models for stress fiber sarcomere mechanics (14,15).

PAGE 46

46 Based on the model in Figure 2 8 corresponds to the zero load working velocity of myosin motors in living stress fibers. The mean velocity of myosin walking is fou nd to be 0.0099 m/sec (in vitro measurements of myosin velocity are in the range of 0.14 m/sec, Umemoto et al. (87) ). While this speed may be limited by the ma ximum speed of the ensemble of myosin motors, we cannot rule out other sources of internal friction within the sarcomere operating in parallel with myosin motors. However, the near instantaneous initial retraction argues against a significant external visc ous drag limiting the rate of contraction after severing. Also, any internal or external viscous drag does not alter the main conclusions of the model: that spring elements do not contribute to the force balance in the stress fiber, and that there are inte rnal barriers present which limit contraction under zero tension. Based on the exponential distribu tion of contraction distances, reported in Figure 2 4 B, the most probable state of a sarcomere following the initial instantaneous contraction is at its minimum length. This surprising conclusion again argues against models that invoke tensile or compressive elastic elements within the sarcomere to position the sarcomere length at some optimal value. A key question remains: how does t he observed exponential distribution of contraction distances arise in a steady state stress fiber? One possibility is that the tension in the stress fiber fluctuates in way that results in this distribution in lengths, as demonstrated by the following mod el. If tension is balanced by myosin contractile forces only, and the tension velocity relation is linear, then the contraction velocity has the form (2 5 )

PAGE 47

47 Where is the maximum myosin contraction spe ed at zero tension, as obtained from the sarcomere contraction measurements, and is the fiber tension required to stall contraction. Assume that the mea n tension in the stress fiber, is less than the stall t ension for myosin contraction in the sarcomere, such that the sarcomeres ten d to contract, thus decreasing with an effective drift velocity (2 6 ) However, fluc tuations in the instantaneous could lead to transient increases in analogous to diffusion. The ef fective diffusion coefficient for arising from fluctuations in is given by (2 7 ) Where is the autocorrelation function for the fluctuations. Under th ese assumptions, the distance is governed by a diffusion process with constant drift toward an impenet rable barrier for which the s tationary probability density, has the exponential form (2 8 ) Therefore, this model predicts an exponential distribution in as observed experime ntally, and the mean value of is equal to In addition to explaining the distribution of contraction distances, the above model provides an alternative mechanical explanation for the ob servation that stimulation of myosin contraction of sarcomeres at a cell b oundary by calyculin A (51) causes

PAGE 48

48 shortening of the stimulated sarcomeres and a corresponding lengthening of di stal sarcomeres. The increase in myosin activity causes an overall increase the stress fi ber tension, and an overall increase in However, a greater stimulation of peripheral sarcomeres leads to a correspond ing increase in their value of In this interpretation, the calyculin A treatment decreases the ratio and hence the mean of in peripheral sarcomeres, while increasing this ratio in di stal sarcomeres. This can cause the observed distribution of sarcomere lengths in cells treated with calyculin A. Insight i nto the M echanics of S tress F ibers A key prediction of the proposed model is that, after the initial elastic retraction, sarcomere te nsion is entirely balanced by myosin contraction, without invoking elastic elements in parallel or a dependence of tension on sarcomere length. Removal of tension causes linear contraction against zero external resistance until the minimum distance is reac hed. Because the dynamics of individual sarcomeres in the linear retraction model differs from the stress fiber end retraction as a whole, a question arises as to whether this linear retraction model is consistent with the exponential retra ction curves rep orted here (Figure 2 1D) and by Kumar et al. (49) In fact, an approximately exponential retraction of the severed fiber end does arise from the linear contraction of multiple sarcomeres in series, if contraction distances come from an exponential distribution, as we found in our measurements (Figure 2 4 B). The measurements suggest that has a prob ability density that can be approximated as an exponential density, such that (2 9 )

PAGE 49

49 is the probability of a contraction distance between and and is t he mean contraction distance ( in the above model). From Equation 2 4 the time dependent mean sarcomere length is thus exponential, given by (2 10 ) For a stress fiber consisting of N sarcomeres, the mean init ial length is Following cutting and the initial elastic retraction, the mean time dependent length is such that relative retraction length is (2 11 ) The solid line in Figure 2 1 D is a least squares regression fit of Equation 2 11 to these data, illustrating the good agreement of the exponential model, with initial retraction, to the stress fiber retraction dat a. If the initial retraction h appens to be smal l relative to then this equation becomes approximately consistent with the exponential distributi on reported by Kumar et al. (49) Therefore, there is no inconsistency between the linear contraction model and the apparent exponential retraction curves of the severed ends of the stress fiber. Summary of Findings In summary, we have performed the f irst measurements of sarcomere contraction in stress fibers formed by living endothelial cells. Our results suggest that tension in the fiber is established entirely by myosin activity, and is not influenced by spring elements as is currently believed. Tak en together, our experiments and analysis shed new light into the behavior of living sarcomeres and suggest a new model for stress fiber

PAGE 50

50 sarcomere mechanics. Future avenues of investigation will need to identify molecular players that contribute to the sti ffness of sarcomeres and determine the barrier position. Figure 2 1. Sarcomeres contract in a severed fiber. (A) Schematic showing the structure actinin marking the ends and polymeri zation competent ends of actin filaments pointing inwards joined by bipolar myosin filaments. (B) An epi fluorescence micrograph illustrating the punctate staining of EGFP actinin in dense bodies of bovine capillary endothelial cell stress fibers (Scale A kymograph showing the results for one half of a severed fiber cut at the thin arrow. Notice that sarcomeres contract and that the contraction is not uniform is 8 40 ms). (D).The distance the severed edge moves follows an exponential form; the solid line is a least squares regression fi t of the retraction model (Eq. 2 11 ) to the data.

PAGE 51

51 Figure 2 2. Cartoon schematic of microscope system used for ablation and ima ging. A mode locked ultrafast Ti Sapphire laser at 780 nm is used to selectively sever GFP actinin labeled stress fibers in Bovine Capillary Endothelial Cells. After the fiber is severed the confocal images are continuously recorded at 488 nm until retraction is complete

PAGE 52

52 Figure 2 3 Time dependent length change of a sarcomere. (A) Representative example of sarcomere length change in a severed fiber. The contraction occurs in two distinct phases: first a quick initial drop (marked on plot) followed by sustain ed contraction at nearly constant speed. After some time the sarcomere reaches a steady state length (arrow) and remains there for the remainder of the experiment. (B) Pooled data from contraction of 28 sarcomeres in 18 cells was normalized and plotted tog ether excluding initial sarcomere lengths (see text for details). A clear linear trend is visible in the normalized data.

PAGE 53

53 Figure 2 4 Histograms of sarcomere contraction parameters. (A) The distribution of the initial length of sarcomeres appears symm etric with a mean and standard the distribution of the linear contraction appears to be approximately exponential (grey curve is drawn for visualization) with a mean and standar d distribution of the net contraction appears to be a shifted exponential with a cells).

PAGE 54

54 Figure 2 5 Potential ener gy is not stored in a severed stress fiber. (A) A stress fiber 20 sec) indicating that myosin activity is responsible for the stress fiber t stress fiber that is allowed to reach steady state (marked by arrow) before myosin poisoning does not recover its length after blebbistatin treatment for several minutes. This indicates that potential energy was not stored in the contracted fiber. (Scale

PAGE 55

55 Figure 2 6 Second example of potential energy not stored in a severed stress fiber. Stress fibers labeled with GFP actin in bovine capillary endothelial were severed using a femtosecond laser and allowed to reach steady state (Scale bar 1 was poisoned with blebbistatin and a confocal micrograph was taken after 35 minutes. Stress fibers did not significantly lengthen or relax during the time course indicating that po tential energy was not stored in the contracted fiber.

PAGE 56

56 Figure 2 7 Fully contracted severed stress fibers do not relax when treated with ML7. (A) Stress fibers labeled with GFP actin in bovine capillary endothelial were severed using a femtosecond lase r and allowed to reach steady state (Sca le activity was poisoned with ML 7 and confocal micrograph s were taken every 10 minutes. Stress fibers did not lengthen or relax during the time course indicating that potential en ergy was not stored in the contracted fiber.

PAGE 57

57 Figure 2 8 Proposed mechanical model for the sarcomere. The tension, T, in the stress fiber is only determined by myosin forces, F M in series with a stiff elastic element, F S An impenetrable barrier pr events further sarcomere contraction at some minimal sarcomere length.

PAGE 58

58 CHAPTER 3 SARCOMERE LENGTH FLUCTUATIONS AND FLOW IN CAPILLARY ENDOTHELIAL CELLS Generation and maintenance of intracellular tension is necessary for endothelial cells to adhere to th e basement membrane and resist external mechanical stresses such as those due to blood flow. Tension in endothelial cells is generated in stress fibers which are observed not only in vitro but also in vivo (25,32) in endothelial ce lls. Endothelial stress fibers are made from tens of repeating contractile units called sarcomeres assembled end to end (37,44) Stress fiber sarcomeres consist of periodic actomyosin contractile elements separated by crosslinked regions of actin filaments actinin. Translocation of bipolar myosin filaments along the actin filament s generates force in the sarcomere. This tension is balanced by the extracellular matrix at focal adhesions (23) Stress fibers are extremely dynamic structures that fo rm as a result of continual generation of F actin at focal adhesions and the cell membrane (88,89) and subsequent crosslinking and bundling of microfilaments (90) Dynamic assembly and disassembly of (72) Recently it has been reported that stress fibers undergo force induced thinning events that result in stress fiber breakage (91) Surprisingly, the dynamic behaviors of sarcomeres, which are building blocks of stress fibers, have not received such attention. It is known that sarcomeres respond dynamically on treatmen t with calyculin A with shortening of peripheral sarcomeres and lengthening of central sarcomeres (51) Nascent sarcomeres have also been shown to flow in at focal adhesions at the cell periphery (88,89) Using femtosecond laser ablation to sever individual stress fibers, we

PAGE 59

59 recently explained the measured sarcomere contraction length distributions by hypothesizing that sarcomeres in a stress fiber may continually fluctuate with time (92) However, we are not aware of studi es which have actually demonstrated fluctuations in sarcomere lengths in living cells. W e found that sarcomere lengths indeed continually fluctuate in living endothelial cells. Nascent sarcomeres were observed to flow into pre existing fibers at focal adh esions throughout the cell with flow velocities that do not correlate with focal stress fiber junctions. Together, these results shed new light into the dynamics of stress fibe r sarcomeres in endothelial cells. Materials and Methods Cell C ulture and T ransfections Bovine capillary endothelial (BCE) cells passages 10 13 were maintained at 37C in humidified 10% CO 2 The BCE cells were cultured on tissue culture dishes in compl ete medium consisting of low (DMEM; Mediatech, Manassas, VA) supplemented with 10% donor bovine serum (Gibco Grand Island, N Y), 1% 1M HEPES (Mediatech), glu tamine (0.292 mg/ml)/ penicillin (100 U/ml)/ streptomyc in (100 g/ml) (Sigma, St. Louis, MO) and basic FGF (2 ng/ml; Sigma) This culture method was described previously in (92) and in great detail in Chapter 2. For experiments, cells were transiently transfected with an EGFP actinin 1 plasmid (kindly provided by Prof. Carol Otey) using the Effectene (Qiagen, Valencia, CA) reagent. Previous studies have rigorously shown that the fusion construct is functional and localizes with the endogenous protein (83) Cells were transfected for 6

PAGE 60

60 hours in 12 well cell culture dish. After 24 hours the cells were trypsinized (Gi bco) and plated sparsely on glass bottom dishes (MatTek, Ashland, TX). Cells were imaged approximately 24 hrs after seeding on glass bottom dishes. The transfection protocol used for these studies is described in Chapter 2. Because the epi fluorescence mic roscope used for these imaging experiments was outfitted with an environmental chamber capable of keeping a 10% CO 2 environment normal culture media was used except for one change. C ulture media was replaced with imaging media consisting of low glucose phe 1% 1M HEPES (Mediatech) and glutamine (0.292 mg/ml)/ penicillin (100 U/ml)/ streptomycin (100 g/ml) (Sigma). Micropatterning of B ovine C apillary E nd othelial C ells To determine if sarcomere dynamics were dependent on cellular shape changes, BCE cells were seeded onto micropatterned islands of fibronectin. Microcontact printing was done according to previously published methods (93) M olds for the stamps w ere produced using standard UV lithography techniques. PDMS (Sylgard 184 kit, Dow Corning, MI) stamps were created by casting photo resist mold using a 10:1 ratio (w/w) The se PDMS stamps can be used for many experiments and should be kept in a clean environment. To micropattern BCE cells the following steps w ere used. Human fibronectin coatTM, Franklin Lakes, NJ) was adsorbed on the PDMS stamps by c oating stamps with FN solution and allow ed to incubate overnight at 4C

PAGE 61

61 Stamp s were then thoroughly washed by pipetting 1 ml of DD H 2 O onto the stamp 5 times. After washing the water was aspirated and the stamps were blown dry with N 2 Using tweezers the stamp was carefully placed feature side down onto an Ibidi dish A small weight was placed onto the stamp to enhance transfer of the FN from the stamp to the dish. After 5 minutes the stamp and the weight were removed with care taken not to slide the stam p. The uncontacted area was then blocked with PLL g Poly ethylene glycol (SuSoS AG) for 15 minutes preventing protein adsorption and cell attachment. After treatment the surface was washed and trypsinized cells are plated Time L apse I maging of L abeled S arcomeres Imaging was done on an inverted Nikon TE 2000 equipped with a t emperature and CO 2 controlled environmental chamber. The environmental chamber allowed experiments to be conducted at the culture conditions described above (37C and humidified 10% CO2). Time lapse live cell widefield epi fluorescent imaging was conducted with a 63X, 1.49 NA or a 40X 1.4 NA oil immersion objective. Images were captured at 3 minute intervals with a cooled CCD camera controlled by Nikon Elements software. To minimize photobleaching and toxic effects, the exposure time was kept below 500 ms and less th a n 50% of light intensity was used. In order to observe significant dynamics each cell was imaged between 2 8 hours. Image A nalysis To measure sarcomere fluctua tions in stress fibers during time lapse imaging, custom tracking software describe previously (92) was used to determine the positions of GFP actinin labeled dense bodies to sub pixel accuracy. Image sequences from the Nikon Elements software package were exported to Im ageJ (NIH) for image

PAGE 62

62 processing. Tracking a simulated particle indicated that the error was minimal, less than 1 percent. The procedure for using the software is identical to what is described in Chapter 2. Kymographs were prepared in ImageJ by first iden tifying focal adhesions in the time lapse image series that displayed clear production of dense bodies. Focal adhesions of interest were rotated to align vertically and then cropped from the image. function of ImageJ was then used to create ky mographs. Flow velocities were calculated by identifying the centroid of a single dense body at multiple time points. Then a line was fit using linear regression whose slope was used as the flow velocity after correcting the units. Particle Image Veloci metry was performed in MATLAB using MatPIV (94) for all images (20) of square cells collected over 1 hr. The individual velocity fields were then averaged to obtain a time averaged velocity field for the cell. Vectors were plotted at 12 of the time series to illustrate the time averaged flow field. Results Fluctuations in S ar comere L ength To investigate sarcomere dynamics in endothelial cells, we imaged the dynamics of EGFP actinin transfected BCE cells. Sarcomeres in the same fiber were observed to undergo significant changes in length on the time scale of several minutes (Figure 3 1 A and B ) in cells cultured in normal growth medium. Individual sarcomeres in a given fiber exhibited different behaviors, with some sarcomeres shorteni ng or lengthening while others fluctuated around th e mean (Figure 3 1B). The length changes were random because the deviation from the mean length remains close to zero throughout

PAGE 63

63 the time course of the measurement (Figure 3 1C). The autocorrelation of the pooled population of measured sarcomeres revealed that correlations in sarcomere length changes die out over a time scale of approximately 20 minutes. The mean sarcomere length from the autocorrelation was found to be 1.3 m and a standard deviation of t h e fluctuations was 0.2 m (Figure 3 1D). We next asked if sarcomere length fluctuations resulted in perturbations to the stress fiber network from cellular shape changes. To control cell shape, we confined minima lly affect cellular viability (95) BCE cells transfected with EGFP actinin displayed an ordered network of stress fibers with mos t fibers aligned along the diagonal (Figure 3 2A). Sarcomere lengths fluctuated continuously even i n confined patterned cells (Figure 3 2A), suggesting that fluctuations were not dependent on overall cell shape changes. Flow of N ascent S arcomeres f rom F oc al A dhesions Previous studies have shown that nascent sarcomeres flow out from focal adhesions at the cell periphery (88,89) In BCEs, we found that nascent sarcomeres flowed out of adhesions at the periphery as well as at adhesions in the interior of the cell. (Figure 3 3A, 3 4C ). In many cases, the sarcomeres did not have coherent dense bodies, but became more organized as they flow ed into the cell (Figure 3 3A). Focal adhesions remain stationary as the dense bodies moved inward which established that the sarcomeres were in deed flowing in (and not just sliding bec ause of focal adhesion movement ) In the majority of cases, the velocity of nascent dense bodies was constant with increasing distan ce from the focal adhesion (Figure 3 3E). When the trajectories for

PAGE 64

6 4 successive emer ging sarcomeres from a single focal adhesion were tracked, the velocity was found to be remarkably uniform (Figure 3 3E). The measured mean sarcomere velocity in endothelial cells was 0.075 m/min +/ 0.0066 m/min which was roughly three times less than t hat measured in U2OS cells (89) and podocytes and fibroblasts (88) Focal adhesions in cells confined to square micropatterned islands of fibronectin were found to exist at both the periphery as reported before (96) and i n the interior of the cell (Figure 3 4A ). Particle image velocimetry (PIV) was used to determine the flow pattern of dense bodies in the micropatterned BCE cells. The results indicated that there was continual flow of nascent dense bodies from all focal adhesions towards the center of the cell (Fig ure 3 4B ). Flow of dense bodies from an interior focal adhesion is illustrated in the kymograph of Figure 3 4C Sarcomere F low V elocity f rom F ocal A dhesion s is T ension I ndependent Dense bodies were observed to flow from focal adhesions of all sizes throughout the cell. It has been previously established that focal adhesion size is correlated with the net tensile force on the adhesion (22,97,98) To determine whether the rate of sarcomere product ion was dependent on focal adhesion size and thus tension, the flow velocity was plotted against size of the associated focal adhesion (Figure 3 4). No correlation was found between dense body flow rate and focal adhesion size suggesting that flow does not depend on tension. Sarcomeres are C S Since new sarcomeres are continually incorporated at focal adhesions, it stands to reason that sarcomeres must be consumed somehow along the fiber length. We

PAGE 65

65 observed sarcomeres being consumed at disc fibers, most often (but not always) at junctions betw een adjacent stress fibers (Figure 3 5A and 3 6 ). In some cases sarcomeres were seen to join end on as illustrated in Figure 3 5B. Taken together these results i ndicate that multiple mechanisms for disassembly of sarcomeres exist in capillary endothelial cells. Discussion While stress fibers have been shown to be dynamic and responsive to mechanical stimuli, little is known about the dynamics of the sarcomeric subunits of stress fibers. In this paper, we present observations of a rich array of dynamic behaviors exhibited by sarcomeres in living cells. Stress fiber sarcomeres were observed to fluctuate in length, cells confined to micropatterned islands of fibronectin. The rate of new sarcomere formation and incorporation from focal adhesions was found not to be correlated with focal adhesion size. Interestingly, within the sa me stress fiber, some sarcomeres elongated while other nearby sarcomeres simultaneously shrunk. Since tension is commonly assumed to be uniform along the length of the stress fiber, such fluctuations are difficult to explain with models which assume that s arcomere length is elastically coupled to tension. On the other hand, if tension and length were weakly coupled and instead only contraction/expansion speeds were governed by tension, then small changes in local myosin activity could lead to large changes in sarcomere length, while maintaining a uniform overall tension. We recently proposed such a model to explain the constant contraction speed and the exponential distribution of contraction distances following removal of tension by la ser severing of stress fibers (92)

PAGE 66

66 Previous studies have reported that sarcomeres are assembled and flow out of focal adhesions in podocytes, fibroblasts (88) and U2OS cells (89) Because stress fibers are anchored to the substrate at focal adhesions and are under isometr ic tension, any new sarcomeric units added to the fiber must form under tension. We found that sarcomere flow velocities are independent of focal adhesion size; it has been shown that focal adhesion size correlates with tension ( (22,97,98) ). It has been previously demonstrated that inhibition of myosin stops the flow of sarcomeres into the fiber (88) Taken together these findings suggest that while myosin activity is necessary for flow, the flow is not influenced by the magnitude of tension. This independence of actin assembly rate on tension is consistent with an end tracking motor model for insertional polymerization of focal adhesion attached filaments plus ends (99,100) but not models where actin assembly of attached filaments is enhanced by tension (101,102) Moreover, the slow, uniform speed of actin assembly that is independent of focal adhesion size argues against an ass embly mechanism limited by the rate of monomer diffusion, and suggests that the assembly rate is governed by a molecular timer with a rate constant of around 0.5 s 1 (based on the measured speed divided by the 2.7 nm added length per actin monomer). The PIV analysis in patterned cells suggests that there is continual flow of dense bodies from all focal adhesions directed towards the diagonal of the square cell. Because the steady state length of the stress fiber is approximately constant, in the absence of any sinks the length of each sarcomere would decrease continuously along maintain a steady state stress fiber length. Our finding of sarcomere fusion along the

PAGE 67

67 stress fiber length, therefore suggests a mechanism by which the cell controls sarcomere length at optimal values. Summary of a Comprehensive Sarcomere Model We have demonstrated that endothelial stress fiber sarcomeres undergo a variety of dynamic behaviors inc luding fluctuations, generation and consumption. Stress fibers form a dynamic and continually evolving connected network that has mechanisms to both increase and decrease tension. Further work is needed to determine the molecular mechanisms of sarcomere as sembly at focal adhesions and to parse out the mechanisms underlying tensional fluctuations. O ur results of sarcomere contraction and our observations of sarcomere fluctuations in intact stress fibers are enough to lead to a self consistent model of the sa rcomere. As reported in Chapter 2 sarcomeres in severed fibers contracted linearly at constant velocity, following a fast initial contraction The sustained linear contraction ended at minimum sarcomere leng th, suggesting a barrier to prevent further contraction. While the initial contraction distances were distributed randomly, distribution of the linear contraction distance was exp onential with a mean (Figure 2 4B). In this chapter we report observations of fluctuating sarcomeres in intact stress fibers with standard deviation This estimation is consistent with the mean linear contracted distance reported in Chapter 2 because for an exponential distribution the variance is equal to the mean squared ( ) The autocorrelation function of the sarcomere fluctuations gave a relaxation time, sec.

PAGE 68

68 Taken together these observations suggest a model where tension fluctuations determine the distribution of sarcomere lengths. What follows is a description of a sarcomere model which meets this criterion. First we assume that the instantaneous contraction speed in the ith sarcomere is proportional to the net tension in the sarcomere and thus follows this form: ( 3 1 ) where is the ith sarcomere length, is the tension generated by myosin within t he sarcomere, and is the tension on stress fiber (assumed uniform along the length of the stress fiber). Assuming that there exists instantaneous fluctuations in myosin activity in the sarcomere then can be as sumed to be a fluctuating random variable with a mean of and variance The external tension, on a stress fiber at constant length consisting of sarcomeres can b e described as hence and Thus fluctuations in are expected to be small (by factor 1/n) relative to fluctuations in Tension fluctuation s lead to diffusive behavior of the sarcomere length, with probability density described by the Fokker Planck equation, (3 2) where is the drift velocity, and D is a diffusion coefficient. The diffusion coefficient can related to fluctuations in the myosin contraction velocity by integrating the area under the velocity velocity autocorrelation function and thus described as:

PAGE 69

69 (3 3) where is the variance in velocity fluctuations (equal to ), and is the relaxation time in speed fluctuations. The relaxation time of the speed fluctuations, can be described acc ording to the following form. (3 4) The steady state solution to Equation 3 2, imposing the boundary condition for is: for ( 3 5) Comparing Equation 3 5 to the observed distribution of the linear contraction distance, implies that From Equation 3 2 we get that the relaxation time constant for length fluctuations is Hence we can solve for and in terms of experimentally estimated parameters and where and We have estab lished earlier that for an exponential distribution and thus and can be estimated using only from parameters obtained from the autocorrelation of sarcomere length fluctuations. Using th e estimates of and from the autocorrelation reported in Figure 3 1 we find that m 2 /sec and m/sec. In summary we have shown that our sarcomere contraction experiments and our observations of sarcomere fluctuations are consistent with each other and lead to a model w h ere sarcomere lengths are dependent on tens ion fluctu a t ions We have demonstrated that the autocorrelation function of fluctuating sarcomeres is itself

PAGE 70

70 sufficient to estimate the parameters that govern the distribution of sarcomere lengths. In the future further breakdown of into variance and relaxation time in speed (Equation 3 3 ) can be obtained from the autocorrelation in the fluctuation speed on the shorter time scale.

PAGE 71

71 Figure 3 1 Sarcomeres undergo dynamic length fluctuations. (A) A kymograph showi ng sarcomeres undergoing length fluctuations during time lapse epi fluorescence imaging (Scale bar is 1.4 m and time between frames is 3 min). Note that sarcomeres do not behave uniformly during the time course. (B) Four sarcomere lengths from the same st ress fiber are plotted with respect to time. The trajectories show the heterogeneous nature of sarcomere fluctuations. (C) The deviation from the mean for 13 sarcomeres is plotted with respect to time. The fluctuations can be assumed to be random because t he deviation fluctuates around zero. (D) The pooled autocorrelation was calculated for 13 sarcomeres indicating that correlations in the fluctuations die out after approximately 30 minutes.

PAGE 72

72 Figure 3 2. Sarcomeres undergo length fluctuations in cells with confined shaped. (A) BCE cells transfected with EGFP actinin were seeded onto square micropatterned islands of fibronectin. Shown here is an epi fluorescence micrograph illustrating the stress fiber network of micropatterned BCE cells (Scale Bar 10 m) (B). This kymograph sh ows the dynamics of a stress fiber (labeled with box in A) in a micropatterned BCE cell. The fiber undergoes qualitatively similar fluctuations to stress fibers from unpatterned cells suggesting that transient shape changes are not the cause of fluctuation s (Scale Bar 4 m and 18 min).

PAGE 73

73 Fig ure 3 3 Nascent sarcomeres flow from all focal adhesions. (A) A kymograph illustrating the continual flow of nascent sarcomeres from focal adhesions in BCE cells. The focal adhesion remains stationary and approxima tely the same size ruling out focal adhesion sliding. (Scale bar 5 m and 6 min) (B ) The trajectories of four dense bodies flowing from a single focal adhesion are plotted with respect to time. Sarcomeres were observed to flow from focal adhesions at const ant velocity as evidenced by the constant slope of each dense body. Successive sarcomeres displayed almost identical velocities as evidenced by the parallel trajectories

PAGE 74

74 Figure 3 4 Nascent sarcomere s flow inwards from all focal adhesions in patterned cells (A ) A fluorescence micrograph of EGFP actinin transfected BCE cells confined to square micropatterned islands of fibronectin (Scale ba r 10 m). (B) Particle image velocimetry was used to determine bulk flow patterns throughout the entire micropatt erned BCE cell. The PIV results indicate that nascent dense bodies flow from focal adhesions throughout the entire cell and not just at the periphery. (C ) A kymograph of flow of dense bodies from an interior focal adhesion (labeled with box in B) illustrat es that flow occurs from interior focal adhesions (Scale bar 5 m and 6 min).

PAGE 75

75 Fig ure 3 5 Sarcomere flow rate is independent of mechanical stress. (A) Sarcomere flow velocities from single focal adhesions are plotted with respect to the focal adhesion size (n= 51 FA from 7 cells). There is little correlation between sarcomere flow rate and focal adhesion size (correlation coefficient of 0.23).

PAGE 76

76 Fig ure 3 6 Sarcomeres are consumed at sinks and join end on to maintain tension (A) Time lapse images of E GFP actinin transfected BCE cells reveal that until two disappear during the course of 18 minutes (Sca le Bar 5 m). (B) Dense bodies were observed to join end on in stress fibers during time lapse imaging. A sarcomere shortens continuously until the two dense bodies (solid black circles) are indistinguishable while the neighboring sarcomeres (between solid and dotted black circles) maintain their integrity. (Scale Bar 2.5 m).

PAGE 77

77 Figure 3 7 Sarcomeres are consumed to maintain tension. (A) Time lapse images of EGFP actinin transfected BCE cells reveal that sarcomeres are consumed during the course of 36 minutes (Scale Bar 2.5 m). (B) A second example of of 18 minutes.

PAGE 78

78 CHAPTER 4 EFFECTS OF DYNEIN ON MICROTUBULE MECHANICS Microtubules are long stiff biopolymers o bserved in eukaryotic cells emanating outward from a central organizing center called the centrosome (29) Microtubules are known to play a key role in normal cell processes including transport of proteins and organelles, forming the mitotic spindle and positioning the centrosome for polarized cellular migration (29) These functions require force generation and transmission by microtubule s Despite their relatively high flexural rigidity (microtubules are the stiffest cytoskeletal filament by two orders of magnitude (54) ) microtubu les are frequently observed in buckled conformations at the cell periphery (29,55) As a result microtubules are generally assumed to be under compressive loading from force generated through polymerization at the plus end tip (55) Based on observations of buckled microtubules, it has been hypothesized that microtubules act as cellular struts resisting tensile forces generated by the acto myosin cytoskeleto n (57 59) Compressive forces have also been hypothesized to be responsible for centrosome centering (103 105) An issue with this hypothesis is that microtubules in vivo are in long conformations such that the critical buckling force is approximately 1 pN wh ich even a single motor protein can exceed (56,106) Recently Weitz and coworkers suggested that microtubules supported laterally by an elastic medium could support larger forces (55) Increasingly the minus end directed motor dynein has been implicated in force generation in microtubules (63,65,107) Immobilized dynein on the actin cortex has been hypothesized to genera te tensile forces on the centrosome by e (107 109) A recent study by Odde and coworkers

PAGE 79

79 implicated dynein in the anterograde motion of microtubules during microtubule buckling (29) In this chapter we investigated the role of dynein in de termining buckled shapes of microtubules in living cells. I ndividual buckled microtubules were severed with a femtosecond laser to probe the nature of the force balance in a buckled microtubule. The purpose of severing the microtubule was to perturb the ov erall force balance and to use the time dependent dynamics of the newly freed ends to determine the key forces involved in microtubule buckling. U pon severing, the newly created minus ended microtubule s did not release their stored bending energy by straig htening as expected of a filament under compression. Rather the slower depolymerizing minus ended microtubules actually increased in curvature and bending energy. In those microtubules where a relatively small increase in curvature was observed, the minus end depolymerized while maintaining shape. I ndividual microtubules were next severed in dynein inhibited cells. In dynein inhibited cells newly freed minus ended microtubules were observed to release stored bending energy and straighten. These results sugg est a new model in which where dynein generates tangential tensile forces on microtubules and frictionally resists lateral motion of microtubules. Materials and Methods Cell Culture, Plasmids and T ransfection Bovine capillary endothelial (BCE) cells were used at passages 11 14 and were maintained at 37C in humidified 10% CO2. The BCE cells were cultured on tissue culture dishes in complete medium consisting of low th 10% donor bovine serum (Gibco), 1% 1M HEPES (Mediatech) and glutamine (0.292 mg/ml)/

PAGE 80

80 penicillin (100 U/ml)/ streptomycin (100 g/ml) (Sigma) The growth media was basic FGF per 1 ml of media ( final concentration 2 ng/ml; Sigma) This culture method was described previously in (82) For control experiments cells were i nfected with a denoviral EGFP tubulin was provided by Prof. Donald Ingber. For dynein inhibition studies cells were transfected with DsRed CC1 plasmid that was p rovided by Prof. Trina Schroer and infected with a denoviral EGFP tubulin To confirm that dy nein inhibition by transfected CC1 was effective cells were fixed and immunostained with mouse monoclonal Golgi marker (Abcam) and Hoechst 33342. Transient transfection of plasmids into BCEs was done with Effectene Transfection Reagent (QIAGEN). For micro tubule stabilization experiments Paclitaxel (Taxol, SIGMA ALDRICH) was added to the media to a final concentration of 1M one hour before the experiment. I maging was conducted with a CO 2 independent me dia as described previously (84) Briefly the CO 2 independent imaging medium (pH 7.3) was prepared by dissolving Hanks balanced s alts, which are phenol red and bicarbonate free, in 1000 ml of DD H 2 O. The following supplements were added HEPES (20.0 mM ), 1% bovine serum albumin, and 1% of both MEM essential and nonessential amino acids (Sigma). Imaging media was further supplemented w ith 10% DBS and glutamine (0.292 mg/ml)/ penicillin (100 U/ml)/ streptomycin (100 g/ml) (Sigma) Laser A blation For laser ablation experiments, an inverted Zeiss Axiovert 200M laser scanning confocal microscope (LSM 510 NLO, Thornwood, NY) was used with 63 X, 1.4 NA Plan Approchromatic oil immersion lens ( Zeiss). In order to image EGFP tubulin the 488 nm laser lin e with the power attenuated to 5 10% and the appropriate GFP bandpass

PAGE 81

81 filter was used. For dynein inhibition experiments cells were confirmed to be expressing DsRed CC1 using epifluorescence microscopy through the eyepiece and then a single confocal image was recorded using the 543 nm laser. To increase the quality of microtubule images, two frames were taken per time point and averaged. Single microtubule ablation was done with a Ti:Sapphire laser at 50 100 % transmission (Chameleon XR, Coherent, Santa Clara, CA ) as previously described (49,92) The Ti:Sapphire laser was focused through th e objective and scanned a thin, ~0.14 m, rectangle orthogonally crossi ng the width of the microtubule for 1 2 iterations. A wavelength of 790 nm was used with a laser head power of 1.5 W, pulse duration of 140 fs and repetition rate of 90 MHz. After abla tion confocal scans as described previously were collected using Zeiss LSM 510 4.2 software at 1 3 s econds / frame to capture the dynamics of micotubules after severing A cartoon schematic of the microscope used for the ablation and imaging is included in F igure 2 2. Detailed methods used for ablation of single microtubules follow in the next paragraph and are similar to those discussed in Chapter 2. Once cells were placed on the microscope and the co rrect focal plane was found the laser parameters for cutti ng were optimized. The separate lasers that were used for ablation and imaging must be aligned correctly to focus on the exact same position in the z direction for optimal cutting. Cha nges in the alignment resulted in slightly different laser parameters be ing needed for optimal ablation from day to day. Before experiments were conducted a cell expressing clear microtubules was found and used to test laser parameters for cutting. Microtubules were generally most clearly visible in either thin protrusions aro und the periphery of the cell of under the nucleus. T he FRAP module

PAGE 82

82 included with the Ze i ss microscope software was used to conduct ablation by scanning with the Ti:Sapphire laser at 790 nm. Because microtubules were generally found in thin protrusions su rrounding the cell periphery great care was taken to prevent collateral damage to the membrane. To optimize laser parameters the laser was initially set at 100% transmission power and to scan for one iteration. These conditions were tested by ablating a si ngle microtubule. In the case w h ere the microtubule was severed and no collateral damage was observed in the accompanying DIC image then experiments were continued using these settings. When coll ateral damage was observed the laser power was reduced increm entally by 10 % until a clean cut occur red. If the microtubule would not cut at 100% transmission multiple scanning iterations were use d. No more then approximately 2 scans were used because the extended cut ting scans resulted in unacceptable time resoluti on. In the case where it was not possible to sever the microtubule with multiple iterations then the 488 nm Argon and the Ti:Sapphire lasers were realigned. To realign the lasers, a cell was focus ed using the 488 nm laser and then the imaging conditions we re switched to image the cell using the Ti:Sapphire laser at 790 nm The vertical position of the TI:Sapphire laser was slowly adjusted until the cell came into focus. Root M ean S quared C urvature C alculations To measure the root mean squared curvatures o f buckled microtubules before and after severing a custom program was developed in MATLAB 7.2 (The MathWorks, Natick, MA) Before curvature measurements in MATLAB were pe r formed i mage sequences were processed by exporting image sequences to ImageJ (NIH). Images were then rotated to align in the horizontal direction and then processed using built in ImageJ functions to reduce noise, enhance contrast and smooth the image The

PAGE 83

83 commented MATLAB code for determining root mean squared microtubule curvatures is p rovided in Appendices B.1 and B.2. The curvature of microtubules was estimated from traces of microtubules using the method described by Odde and coworkers as the three point method (29,110) Custom Matlab code was developed to determine coordinates for the microtubule traces with sub pixel accuracy adapting methods described previously (110,111) A description of the process for determining microtubule coordinates and curvature is described below. An initial trace of the microtubule to be analyzed was first provided to the program listed in A ppendix B.1 by using the mouse to select points along the contour of the microtubule. The initial points were then used to segment the microtubule and local maxima were identified either row wise or column wise for each segment depending on the orientation of microtubule in the segment. The coordinates of the local maxima were refined to subpixel resolution by fitting a one dimensional Gaussian approximately orthogonally across the microtubule. The sub pixel coordinates determined by Gaussian fitting were t hen provided to the program listed in Appendix B.2 eliminate short wavelength measurement error and preserve long range microtubule buckles. The coordinates were then coarse grained s uch that coordinates a defined number, apart were connected with a line segment. The series of line segments were then used to estimate the local curvature according the three point method described by Odde and coworkers (29,110) The three point method estimates local curvatures according to the following equation: (4 1)

PAGE 84

84 where is the local curvature estimate and is the tangent angle as function of contour length The estimation of is done by calculating the a ngle between two coarse grained line segments and dividing by the average length of the line segments and (29,110) After local curvature estimates were calculated the root mean squared curvature was determined for the microtubule contour in question. Results Dynamics of Severed M icrotubul es Recently Weitz and coworkers suggested that microtubules buckle as a result of compressive loading through a com b ination of experiments and modeling (55) The authors noted that microtubules buckle in much smaller wavelengths then would be expected from typical Euler buckling. Through both observati on of polymerizing microtubules impinging on the cell periphery and the use of exogenous force the authors confirmed that the short wavelength buckles were the result of compressive loading. Through modeling and experiments the authors show that individual microtubules supported laterally by an elastic medium would exhibit the observed short wavelength buckles and could support larger compressive forces th a n previously thought, on the order of 100 pN (55) In this model the stored energy in both the bent microtubule and the surrounding elastic medium is balanced by energy generated by microtubule polymerization. These results and conclusions are in line with the hypothesis that microtubules bear compressive loads and serve as struts to balance acto myosin generated tensional forces in the cell (57 59) To determine if compressive stresses are significant components of the force balance on bent microtubules, we used femtosecond laser ablation to sever individual

PAGE 85

85 microtubules. Upon severing a single microtubule b oth a new plus ended and minus ended microtubule fragment were formed. T he newly formed fragments were distinguished by differences in depolymerization rate ; p revious studies (112) and our own results (Figure 4 1) have reported markedly larger depolymerization rates at the plus end versus the minus end. An example of the time dependent evolution of a severed buckled microtubule is illustrated in Figure 4 2A. As result of the three fold faster depolymerization kinetics the newly formed plus ended microtubule (top portion) was observed to rapidly depolymerize along the original contour of the microtubule. Interestingly the newly formed minus ended microtubule (bottom portion) was observed not to straighten as one would expect of a buckled filament under compression but rather increased in curvature. Time dependent shape traces fr om the images in Figure 4 2B illustrate the differing dynamics of the two newly freed ends. The root mean squared (RMS) curvature was found to increase with time in the newly formed minus ended filament as illustrated in Figure 4 2C. These differing beha viors of the newly formed filaments were found to be reproducible across many experiments ; another example of a buckled microtubule at the cell periphery is included as Figure 4 2D. Overall newly freed plus ended microtubules were observed to rapidly depo lymerize along the original contour without relaxing. Newly freed minus ended microtubules were always observed to either preserve the original contour or increase in RMS curvature. T he increase in RMS curvature varied from microtubule to microtubule but w as not found to be correlated with the initial RMS curvature of the microtubule, Figure 4 3A. The change in RMS curvature

PAGE 86

86 was also found not be correlated with the position in the cell as illustrated in Figure 4 3B. In experiments where small increases in RMS curvature were observed ( see Figure 4 4A ) the slowly depolymerizing minus ended filament was observed to depolymerize around the original contour. To determine if the relaxation kinetics of buckled microtubule were slower th a n the possible observation time before depolymerization ( approximately 10 15 seconds ), the experiments where performed with microtubules stabilized with Taxol treatment. A severed buckled microtubule in a Taxol treated cell was observed to hold the original contour for tens of seco nds, Figure 4 4B. Dynein Inhibition Alters the Dynamics of Severed M icrotubules Recently Odde and coworkers reported results suggesting that the molecular motor protein dynein may play a fundamental role in buckling microtubules (29) Through direct observation of buckling mic rotubules using fluorescent speckle microscopy they reported that during buckling microtubules move anterogradely, towards the cell periphery and the microtubule plus end (29) To directly test the role of dynein in buckling and stabilizing bent microtubules individual microtubul es were severed in cells transfected with the DsRed CC1 plasmid. The DsRed CC1 fusion protein competitively binds to dynein rendering it inactive. DsRed CC1 inhibition of dynein activity was confirmed by immunostaining for the Golgi complex, see Figure 4 5 A. In DsRed CC1 transfected cells the Golgi complex was observed to diffusively surround the nucleus instead of being closely packed in control cells. This has been reported in the literature as a positive control for dynein inhibition (113)

PAGE 87

87 In DsRed CC1 transfected cells newly freed minus ends were not observed to hold the contour or increase in RMS curvature but rather were observed to str aighten, Figure 4 5B. T he straightening occurred in a relatively short time scale, 4 seconds for the case of the experiment in Figure 4 5B. This striking ly different result compared to normal cells occurred reproducibly in CC1 transfected cells over multip le experiments. On average newly freed minus ended microtubules in control cells were observed to increase in RMS curvature while in CC1 treated cells an average decrease in RMS curvature was observed, Figure 4 5C. D iscussion The results of this ch apter provide new insights into the forces generated by dynein on microtubules Generally buckled microtubules have been assumed to be under compressive loading due to forces generated by polymerization (55) Weitz and coworkers have proposed that buckled microtubules are supported laterally by deform ing the surrounding elastic medium and as a result can support larger compressive forces (55) If microtubules are indeed under compression then one would expect that severed buckled microtubules to relax to release the stored bending energy. Furthermore if the surrounding elastic matrix is deformed th is should increase the energy available for newly freed microtubule s to straighten. Surprisingly when individual buckled microtubules were severed using a femtosecond laser both newly freed ends did not straighten. Rather both freed ends were observed to h old the original contour or actually increase in RMS curvature in the case of the minus ended microtubule. The increase in curvature of the minus ended microtubule suggests that dynein tangentially pulls on the microtubule as it walks toward the minus end. In the newly formed minus ended mictotubule, a n ensemble of dynein molecules increases the bending energy of

PAGE 88

88 by translating the freed end towards the plus end. Recently Odde and coworkers reported that microtubules move anterogradely during buckling and h ypothesized that minus end directed motor immobilized to the cortical actin network could be responsible (29) Here we have provided direct experimental evidence that suggests that dynein provides the major contribution to the stress in microtubules. Furthermore severed microtubu les in dynein inhibited cells were observed to straighten on time scales of approximately 4 seconds. This suggests that dynein provides the dominant frictional resistance to microtubule straightening Th ese conclusion s ha ve interesting implications on the directionality of force that individual microtubules apply to the centrosome. Generally it has been assumed that buckled microtubules are under compression and exert pushing forces on the centrosome. Our experimental results lend themselves to a model wher e dynein generated stresses in the microtubule lead to tensile forces on the centrosome. In this model, a schematic is included as Figure 4 6, each microtubule is connected at the minus end to the immovable centrosome and plus end is free to polymerize at the cell periphery. The centrosome is considered stationary because of the radial symmetry of the microtubules emanating outwards. An ensemble of dynein molecules immobilized on the cortical actin network bind to the microtubule and pull it taut against th e centrosome. At some boundary point near the periphery the force in the microtubule switches from tensile to compressive as slack generated from polymerization is compressed against the cell periphery. The dynein motors exerting an anterograde force resul ts in concentrating buckles near the cell periphery.

PAGE 89

89 Summary of Findings In summary our results have provided novel insight into the mechanics of microtubules and the role that dynein plays in these mechanics. Through relatively few but surprisingly infor mative experiments we can provide a qualitative picture of how dynein contributes to the force balance of a microtubule. Further quantitative analysis of the mechanics of dynein in microtubules and how these impact the centrosome are discussed in (114,115)

PAGE 90

90 Figure 4 1. Depolymerization rates differ between plus ended and minus ended microtubules. The depolymerization rates of newly formed plus ended and minus ended was measured in severed microtubules. The plus ended microtubule fragments were f ound to have a newly three fold higher depolymerization rate then minus ended fragments.

PAGE 91

91 Figure 4 2. Minus ended microtubule fragments increase in curvature after severing. ( A ) Upon severing (at the black arrow) microtubule (highlighted by crosses) wer e not observed to straighten. The quickly depolymerizing plus ended fragment was observed to depolymerize along the original contour. Surprisingly the minus ended fragment was observed to increase in curvature. Scale bar is 2 m. (B) Shape traces of the tw o fragments created by severing illustrate the differing dynamics. (C) The RMS curvature was found to increase in the minus ended fragment with time and in this example increased almost 7 fold. The plus ended microtubule fragments were not observed to chan ge in curvature. (D) Increases in the RMS curvature were found to be reproducible across many experiments. A representative example of experiments where a single microtubule, near the cell periphery (white arrow), was severed. Images show increased bending of minus ended microtubules after severing (cut at black arrow). Note that the plus end depolymerizes but does not show a change in curvature. Scale bar is 1 m.

PAGE 92

92 Figure 4 3. There is no correlation observed between the change in RMS curvature and eithe r the initial RMS curvature or position of the microtubule. (A) The maximum change in RMS curvature for 1 8 experiments is plotted with respect to the initial curvature. In all experiments that were analyzed the change was positive and th ere exists no corre lation. (B) The change in RMS curvature was plotted with respect to the shortest distance from the cut location to the cell periphery. No correlation was observed suggesting that the increase in curvature is location dependent.

PAGE 93

93 Figure 4 4. Bent microtub ules are stable over significant time scales. ( A ) In experiments were minimal increases in curvature were observed, the minus ended fragment depolymerized along the original contour. This data indicates that microtubules in living cells are pinned to the c ytoplasm. Scale bar is 2 m. (B) In taxol treated cells bent microtubules in severed minus ended fragments were stable for tens of seconds. Scale Bar is 1 m.

PAGE 94

94 Figure 4 5 Severed microtubules in dynein inhibited cells were observed to straigthen. (A) BCE cells were transfected wi th DsRed CC1 to inhibit dynein. To confirm that dynein was inhibited the cells were immunostained for the Golgi complex. Control cells (left) show a compact Golgi complex (green) near the nucleus (blue), while DsRed CC1 transfe cted cells (right) show a dispersed Golgi complex. Scale bars are 10 m. (B) Severed minus ended microtubules were observed to straighten in dynein inhibited cells on a time scale of seconds. This suggests the dynein plays a fundamental role in providing f riction for the microtubule. Scale bar is 2 m (C) On average newly formed minus ended microtubules were observed to increase in RMS curvature while in dynein inhibited cells a decrease in RMS curvature was observed.

PAGE 95

95 Figure 4 6. Cartoon schematic of p roposed model of dynein force generation in microtubules. Microtubules that are buckled near the periphery are under compressive stress but pinned to the cytoplasm. The compressive stress results from the polymerizing plus end tip impinging on the cell per iphery. Along the length of the microtubule an ensemble of dynein molecules immobilized on the cortical actin network exert tensile forces that eventually overcome the compressive stresses. As result the individual microtubule exerts a net tensile force on the centrosome.

PAGE 96

96 CHAPTER 5 CONCLUSIONS Summary of Findings The endothelium lines the vasculature and consists of a single monolayer of specialized endothelial cells. Endothelial cells are subjected to a continuous external mechanical stress due to b lood flow. E ndothelial cell function is critical for normal function of the vasculature and numerous pathologies have been shown to result from altered EC function including hypercholesterolemia, atherosclerosis myocardial ischemia, and chronic heart fail ure (10) P ast work has focused on pathological biochemical pathways (11) I ncreasin gly altered mechanical forces on endothelial cells is being studied as an initial inciting event in atherosclerosis (76) For example it has been shown that disturbed blood flow at vasculature bifurcations leads to the expression of pro atherosclerotic genes in endothelial cells (76) Thus understanding the mechanisms by which endothelial cell s sense mechanical forces and transduce them into intracellular response is cr itical for developing effective therapies. Furthermore endothelial cell mechanotransduction has broad implications for regenerative medicine which seeks to create engineered organs in vitro Creating an adequate microvasculature in engineered tissues is a key challenge in regenerative medicine, and depends on tuning endothelial function through biochemical and mechanical factors (116) Future engineered tissues a nd tissue substitutes will have to recapitulate the in vivo mechanical environment for proper endothelial cell function (117) The cytoskele ton is key to endothelial mechanotransduction. Therefore, this thesis focused on understanding the mechanical properties of the cytoskeleton in endothelial cells.

PAGE 97

97 Speci fically, o ur findings considerably advance fundamental understanding of the endothelial c ytoskeleton as summarized below. Stress Fiber Sarcomere M echanics Stress fibers are tension generating actomyosin structures in endothelial cells that allow endotheli al cells to resist shear stresses from blood flow. Stress fibers are composed of linearly arranged contractile subunits referred to as sarcomeres. While theoretical mechanical models have been proposed for the stress fiber sarcomere there exists no direct experimental evidence to support these sarcomere models Understanding the mechanics of the sarcomere is critical for a comprehensive mechanical model of the stress fiber. A novel technique of femtosecond laser ablation was applied to study the mechanics of stress fiber sarcomeres (92) Upon severing stress fibers were observed to retract and the ends followed exponential kinetics as reported in the literature. I ndividual sa rcomeres were observed to shorten in three distinct phases. These phases included a rapid contracti on followed by a sustained linear contraction at constant speed and finally a steady state. Fully contracted stress fibers were shown to not release potential energy after treatment with myosin inhibitors. This result suggests that mechanical energy is sto red in a fully contracted stress fiber. Based on these observations a mechanical model of the stress fiber sarcomere was proposed consisting of an elastic element in series with active myosin contraction that stops upon reaching a barrier. This sarcomere m odel lends itself to a stress fiber model where tension in the fiber is directly determined by myosin activity and that the force in a sarcomere is balanced by the neighboring sarcomeres An interesting result from these experiments was the existence of an exponential distribution of linear contraction distances the distance which the sarcomere contracted

PAGE 98

98 a t constant velocity. This finding suggests that the most probable state for a sarcomere is fully contracted at the barrier. Furthermore the reported vis coelastic retraction of the stress fiber as a whole can be adequately modeled with the exponential distribution of contraction distances. Sarcomere fluctuations were hypothesized as one possible explanation of this exponential distribution which were obser ved through live cell imaging and are discussed in the following section. Sarcomere Dynamics in Endothelial C ells The number of sarcomeres and sarcomere lengths dynamically change in the cell but the mechanisms by which these processes occur are not un derstood. Novel observations of sarcomere behavior in BCE cells were observed and interpreted through our proposed mechanical model. These observations suggest that sarcomeres undergo a variety of dynamic behaviors including fluctuations, generation at f oc al adhesions and consumption (118) These behaviors are not dependent on cell shape changes as they were observed in cells confined to micropatterned islands of fibronectio n. T he rate of generation at focal adhesions was shown to be independent of focal a dhesion area suggesting the rate of generation of sarcomeres is tension independent. T he fluctuating length of individual sarcomeres under constant tension is consistent wi th our mechanical model whereby sarcomere contraction/expansion speed, rather than sarcomere length, is modulated by tension. Furthermore our results from the sarcomere contraction experiments were found to be consistent with our observations of sarcomere length fluctuations. These findings lend support to a model where the distribution of sarcomere lengths is determined by tension fluctuations. We demonstrated mathematically that the mechanical model

PAGE 99

99 parameters of effective sarcomere diffusion coefficient and drift velocity can be obtained from only the autocorrelation function of sarcomere length fluctuations. Our experimental results lead to a self consistent mechanical model of the stress fiber that recapitulates observations in the literature, such as seemingly viscoelastic retraction of a severed fiber and inhomogeneous sarcomere contractions. Recently and coworkers (119) published a model for a stress fiber con traction mechanism to capture the dynamics of stress fiber retraction reported by Kumar et al (49) A key finding of this model was that substa ntial drag forces act on the severed fiber that could result from specific or non specific interactions. Our model is consistent with this study in that we assume that the velocity of contraction of a sarcomere is proportional to the net tension on that sa rcomere. Thus the proportionality constant could include contributions from both internal and external friction A recent study by Stelzer and coworkers suggests that stress fibers in epithelial cells and fibroblasts are connected to the ventral cellular surface (120) This result was based on the observation that in severed fibers retraction is limited to close proximity to the cut and that new focal adhesions were formed at the severed stress fiber end (120) This finding is not in line with our observations of stress fiber retraction in BCE cells where fibers were regularly observed to retract across the entirety of the cell, see Figure 2 6 for an example and to our knowledge focal adhesions were not formed at severed ends. These different findings suggest that that there may be cell type variations in stress fiber mechanics and that possible future studies in other types of endothelial cells could be useful.

PAGE 100

100 Effects of Dynein on the Microtubule Force B alance Microtubules ar e stiff biopolymers that emanate outwards from a central organizing center known as the centrosome. Microtubules are consistently observed to exist in bent conformations surrounding the cell periphery and are assumed to be under compression. Recently the m inus ended directed motor protein has been implicated in the buckling of microtubules. To determine the role of dynein in the force balance of microtubules, individual buckled microtubules were severed in living cells using femtosecond laser ablation. Surp risingly buckled microtubules were not observed to straighten as would be expected for a compressed filament. Rather newly formed minus ended microtubule fragments were observed to increase in bending energy. I n dynein inhibited cells newly formed minus e nded microtubule fragments were observed to straighten releasing stored bending energy. These results suggest that dynein contributes significantly to the microtubule force balance. Furthermore this points to a model where microtubules are both under tensi on and compression. The net result of this model is that microtubules exert a tensile force on centrosome but are compressed at the cell periphery. Future Work The purpose of this section is to outline possible future studies to expand on our models and propose new future work based on the methods and ideas contained in this dissertation. Bi ophysical Analysis of Mechanical Factors Leading to A therosclerosis Recently the role of altered shear stress has been implicated in the expression of pro atheroscler otic genes in endothelial cells (76) Genera lly atherosclerotic plaques first form n ear vasculature bifurcations with disturbed blood flow patterns (121) A rterial

PAGE 101

101 stiffening has been demonstrated to lead to vascular pathologies in the aging population (122) Both of these results suggest that understanding how endothelial cells react to a changing mechanical environment is crucial to understanding t he pathology of atherosclerosis. Stress fibers are believed to play an integral role in the mechanotransduction in endothelial cells (121) A natural progression for the results in this work is to use our stress fiber model to determine how mechanical forces are transduced and affect the mechanics of the stress fiber. This is the first step in delineating how shear stress and arterial stiffening may incite atherosclerotic phenotypes Furthermore gene expression assays f or pro atherosclerotic genes will allow the d etermination of a biophysical phenotype for atherosclerosis. The experimental approach for this work would be to first characterize stress fiber mechanics under differing mecha nical conditions and then determine expression levels of pro atherosclerotic gen es under matching mechanical conditions. A brief outline of the experimental procedure follows. Our initial work on stress fiber mechanics was conducted on glass substrates with static flow conditions. To gain an understanding of how stress fibers transduc e mechanical forces from the external environment we would culture cells on compliant poly acrylamide gels or apply shear stresses to cells cultured on glass. Both of these methods have been well established in the literature (123,124) Sarcomere fluctuations and production from focal adhesions would be observed using time lapse microscopy to obtain model parameters discussed in this document. Samples of ECs exposed to matching mechanical stresses would be processed for RT P CR quantification of pro atherosclerotic gene expression. This would allow us to determine what mechanical

PAGE 102

102 stresses result in maximum expression of pro atherosclerotic genes and understand the state of the stress fiber network. A discussion of how model pa rameters may be affected by changes in cellular tension follows. The impact of overall cell tension determined by external mechanical forces on the model parameters is not easily interpretable without experimental results. The external tension, of the sarcomere is expected to change but it is not clear how the range of tens ion, that myosin in sarcomere can generate would affected If it is assumed that the range of is fixed based on the stall force of myosin and constant for all conditions then for increasing the drift velocity would be expected to decrease. Under the same assumption an increase in should lead to a decrease in the variance of tension fluctuations and thus velocity fluctuations, and as a result decrease Force Transmission B etween Endothelial Cells in a Mono L ayer Vascular endothelial cells regulate t he permeability of the endothelium by binding tightly to neighboring cells through specialized multi protein cell cell adhesion structures, tight junctions and adherens junctions (125) The endothelium must dynamically adjust in permeability to allow the passage of leukocytes from the blood stream to tissues (126) Actin stress fibers have been identified as key players in leukocyte transmigration and regulation of small solute permeability (126) A recent paper by Ridley an d coworkers suggests that stress fibers in adjacent cells are connected through specialized adhesions called adherens junctions (126) While this result suggests tha t there is mechanical continuity between endothelial cells, little is known about how forces are transmitted between cells and what the key molecular players are in this process. A possible direction for continuation of this work would be to

PAGE 103

103 investigate fo rce transmission mechanisms between endothelial cells in a mono layer to gain a greater understanding of mechanisms for regulating paracellular permeability. Recently poly acrylamide hydrogels have been used to culture cells reliably on substrates of diff ering stiffness (123) and the inclusion of fluorescent micro beads has allowed the researchers to calculate traction stresses that cells exert on the substrate (127) Using this technique along with femtosecond laser ablation a future study could determine how forces are transferred t hrough cell cell adhesions and the key molecular players. A brief description of a basic experiment for this study follows. A mono layer of ECs would be grown on a complaint poly acrylamide substrate with embedded blue green microspheres. The EC monolayer would then be transfected with EGFP actinin and mCherry VE cadherin (a critical AJ protein). Single stress fibers would be ablated in one cell and the surrounding cells would be on image on three channels (red, green, and blue). This experiment would yie ld useful data about how forces are transmitted either between cells or to the substrate. Upon releasing tension in the severed stress fiber there are three possible outcomes. One of these is that all dissipated tension is transferred to the neighboring ce lls in which case the AJs and sarcomeres would translate. Another possible outcome is that all dissipated tension would be transferred to the substrate in which case the AJs and sarcomeres remain stationary and the microspheres translate. The third outcome is a hybrid of both of these in which some dissipated tension is transferred to both neighboring cells and the substrate. One would expect that for effective regulation of vascular permeability the majority of tension would be transmitted between cells. T his would allow for the

PAGE 104

104 endothelium to directly remodel itself to regulate permeability without wasting energy transmitting too much force to the substrate. Further experiments could be done at differing substrate compliances to determine the affect of s ubstrate compliance on force transfer between cells. The use of small interfering RNAs, si RNA, has allowed researchers to knockdown the expression of specific proteins of interest. In this proposed study si RNAs directed towards AJ proteins could be used to determine the key molecular factors in efficient force transmission between cells. Overall this proposed project would contribute to the understanding of how the endothelium regulates permeability both mechanistically and by identifying proteins respon sible for this functionality. Mechanics of M yofi lament R epair in C2C12 M yoblasts Cellular damage occurs in mammalian cells during normal physiological functions as well as the result of local trauma (128) There exists cellular mechanisms for repair but these mechanisms be hindered in pathologies such as muscular dystro phy (129) A recent paper by Beckerle and co workers suggests a zyxin mediated repair mechanism for stress fibers in non muscle cells (91) Unfortunately there are no good model systems where one can create subcellular damage and follow the repair process. Recently we (92,115) and others (49) have used to femtosecond laser ablation to selectively sever individual cyt oskeletal structures in BCE cells. W hen myofilament bundles were severed with a femtosecond laser in C2C12 myoblasts transfected with EYFP actinin the myofilaments were observed to repair themselves over the course of 1 2 minutes, Figure 5 1. After ablation of a small section of the actomyosin structure the newly severed ends were observed to pull apart as expected from other studies, Figur e 5 1A. T he wound was observed to be repaired

PAGE 105

105 upon a local recruitment of EYFP actinin, Figure 5 1B and C. After 150 seconds the myofilaments appear to have been completely repaired and resemble the same structure as before ablation, Figure 5 1D. To dete rmine if this structure was still functioning and contractile the myofilaments were severed again in the same spot and minimal retraction was observed. These results suggest that a mechanism for myofilament repair exists in C2C12 myoblasts and that further studies are needed to determine the mechanics and key molecular factors for this mechanism. This proposed project would be to first study the mechanics and kinetics of the repair mechanism in C2C12 myoblasts by ablating small sections and observing the dy namics of repair in cells transfected with EYFP actinin. A second phase of this project would be to transfect cells with different sarcomeric fluorescent fusion proteins such as zyxin or VASP to determine if they play a role and their kinetics in the rep air mechanism. Finally si RNA knockdowns of key sarcomeric proteins will be used to determine key molecular players for this process. Overall this proposed project will enhance the collective understanding if the mechanics and molecular players for subcell ular repair.

PAGE 106

106 Figure 5 1. Severed myofilaments in C2C12 myoblasts display a repair mechanism. (A) Myofilaments in C2C12 myoblasts transfected with EYFP actinin were severed using a femtosecond laser (white line) and upon severing the ends of fiber retracted away from each other. (B) A kymograph of a section of severed myofilament reveals that shortly after severing (at arrow) there is a transient actinin concentration resulting in repair of the myofilaments. (C) actinin intensity with time in the severed myofilament. Intensity of actinin peaks at approximately 60 seconds after the cut (D) Comparing images before cutting and at the end of the time series suggests that the sarcomeric structure has been repaired in the cut region. (E) Ablation (white line) in the repaired zone reveals that the myofilaments are again generating tension and functioning normally. Scale bars are 5

PAGE 107

107 APPENDIX A MATLAB CODE FOR ANAL YZING SARCOMERE DYNA MICS Included in this appendix are the MATLAB programs used for most of the data analysis in this document. These programs were written as semi automated methods to reproducibly analyze dynam ics of cytoskeletal structures. Various image processing methods were included and are explained in the method sections of appropriate chapters. Dense Body Tracking Software This code was written and used for tracking dense bodies using a correlation based method but could be used for tracking any distinguishable feature. To improve accuracy a secondary centroid based correction method is also included. The code is semi automated in that the user must select the feature in the initial frame and the code wil l compute positions for a user defined number of frames. The code calculates two separate coordinates for each frame which are calculated by distinct methods that are discussed in Chapter 2. For some applications a second program, discussed in A.2, can be run using the data generated from this program to ensure greater accuracy and refine the coordinate calculations. %Start of densebodytracking.m %This code tracks features in fluorescent images based on a correlation. clear all close all %%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %set program parameters n=1 ; %starting image for analysis m=20; %ending image for analysis threshold=.6; % Set thresh holding parameter. This will depend on the quality of the images being analyzed a nd is best determined by trial and error for the experiments being analyzed

PAGE 108

108 num=3; % This is the size around the correlation peak to which we fit a accurate results %Import image f iles and process them matfiles =dir(fullfile('H:','Sept232010', 'cell8','ForPIV', '*.tiff')); size_matfiles=size(matfiles); for i =1:size_matfiles(1) %temp_image = imread(matfiles(size_matfiles(1) i+1).name); % Sometimes need to reverse the sequence of images so that we start from bad photobleached images and move on to better ones. temp_image = imread(matfiles(i).name); %temp_image=rgb2gray(temp_image); temp_image = imadjust(temp_image, stretchlim(temp_image), [0 1]); %This line contrast stretches the images to improve the ability to track size_image=size(temp_image); [p,q]=size(temp_image); %Use this code to add extra rows or columns to the image. In cases where the feature approaches the image boarder this will help prevent the tracking program from losing the feature D=zeros(p+20,q+20); D(11:p+10,11:q+10)=temp_image; imag{i}=D; imag{i}=temp_image; end %The user must select the feature in the initial frame numbered n. [template, rectcrop] =imcrop(imag{n},[min(min(imag{n})),max(max(imag{n}))]); r ectcrop(1)=round(rectcrop(1)); rectcrop(2)=round(rectcrop(2)); width=round(rectcrop(3)); height=round(rectcrop(4)); [template, rectcrop]=imcrop(imag{n},[rectcrop(1),rectcrop(2),width,height]); %To ensure accurate calculations the positions of the selecti on are rounded to integer values and stored save_template{1}=template; save_rectcrop{1}=rectcrop; %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %First must calculate the centroid of the feature in the initial frame. Because correlation only gives displacements this will be the starting point for all coordinates that are calculated. imag_centr{1}=template; imag_max=max(max(imag_centr{1})); %Calculate maximum intensity value in selection imag_max=double(imag_max);

PAGE 109

109 size_imag= size(template); %Threshold selection to try eliminate noise or influence of neighboring features for i = 1:size_imag(1) for j = 1:size_imag(2) if imag_centr{1}(i,j)rectcrop(2)+height+7) imag1{n}(i,:)=0; end end for i= 1:sizeofimage(2) if (i < rectcrop(1) 10) imag1{n}(:,i)=0; end

PAGE 110

110 if (i >rectcrop(1)+width+10) imag1{n}(:,i)=0; end end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%% %Run the correlation code on each frame to update positions of the feature and save coordinates for g=n+1:m %To prevent the code from jumping to similar neighboring features the code blacks out the area far away from the feature. The use r may need to adjust these parameters to ensure that code tracks the correct particle. for i= 1:sizeofimage(1) if (i < (rectcrop(2)) 10) imag1{g}(i,:)=0; end if (i >(rectcrop(2)+height+10)) imag1{g}(i,:)=0; e nd end for i= 1:sizeofimage(2) if (i < rectcrop(1) 7) imag1{g}(:,i)=0; end if (i >rectcrop(1)+width+5) imag1{g}(:,i)=0; end end image=imag1{g}; [c,corner_y(g),corner_x(g),xrefined,yrefined,cvalu e]=process(template,image,n um); %This gives the corner position in the ith image of the old template (i.e. from the i 1 image). The main code passes three things to the custom 1, image i and the parame ter num. x_offset(g)=round(rectcrop(1)) corner_x(g); y_offset(g)=round(rectcrop(2)) corner_y(g); %These are the measured offsets between images and is used to update the positions cvalues(g)=cvalue; %The correlation values are saved for troubleshooting if needed.

PAGE 111

111 %Update positions are calculated with the offsets and stored xcord(g)=xcord(g 1) x_offset(g); ycord(g)=ycord(g 1) y_offset(g); %Updated positions are used to determine the template position in image i templatex=round(xcord(g) width /2); templatey=round(ycord(g) height/2); %Cropping new template for the next correlation [template,rectcrop]=imcrop(imag{g},[templatex,templatey,width,height]); %Refine the new template position by calculating the centroid imag_centr{g}=template; imag _max=max(max(imag_centr{g})); imag_max=double(imag_max); size_imag=size(template); for i = 1:size_imag(1) for j = 1:size_imag(2) if imag_centr{g}(i,j)
PAGE 112

112 cord(:,1)=xcord; cord(:,2)=ycord; %Create new series of images where area surrounding feature is blacked out for Averaging Program for g=n:m for i= 1:sizeofimage(1) if (i < (yCent(g) ((height)/2))) imag1{g}(i,:)=0; end if (i >(yCent(g)+((height)/2))) ima g1{g}(i,:)=0; end end for i= 1:sizeofimage(2) if (i < (xCent(g) ((width)/2))) imag1{g}(:,i)=0; end if (i >(xCent(g)+((width)/2))) imag1{g}(:,i)=0; end end end %For each frame plot coordinates on image as well blacked out image for l=n:m figure(l) subplot(2,1,1) imshow(imag{l},[min(min(imag{n})),max(max(imag{n}))]) hold on plot(xcord(l), ycord(l), 'bx') hold on plot(xCent(l), yCent(l), 'rx') subplot(2,1,2) imshow(imag1{l},[min(min(imag{n})),m ax(max(imag{n}))]) end save tracking %End of densebodytracking.m

PAGE 113

113 Correlation P rocessing F unction This function uses the built in MATLAB function normxcorr2 to do the correlation calculation and then find the maximum. The integer values of the maximum are passed to the custom function refineparaboloidnew to refine to sub pixel resolution. %Start of process.m %This function takes in a template and an image and returns the position of the template in the new image. function [c,corner_y,corner_x,xrefine d,yrefined,cvalue]=process(template,image,num) c = normxcorr2(template, image); % find correlation of dense body with image [max_c, imax]= max (c(:)); %find the maximum correlation [ypk, xpk]= ind2sub(size(c), imax(1)); %Get coordinates for the maximum [xrefined,yrefined,cvalue]=refineparaboloidnew(ypk,xpk,num,c); %This is the predicted sub pixel position of the cropped rectangle in the image at the next time point. corner_y=yrefined size(template,1)+1; corner_x=xrefin ed size(template,2)+1; end %End of process.m Paraboloid F itting F unction This function refines the peak found in the correlation spectrum by fitting it to a paraboloid function using the method of least squares. It returns the refined x and y peak po sitions (sub pixel) as well as the correlation value at the peak. %Start of refineparaboloidnew .m function [xrefined,yrefined,cvalue,coeff]=refineparaboloidnew(ypk,xpk,num, c) %Create matrices for the calculation based on the least squares method. A=zer os(2,2); B=zeros(2,1); D=zeros(6,6); E=zeros(6,1); coeff=zeros(6,1); Z=double(c);

PAGE 114

114 %Populate matrices for i = xpk num:xpk+num for j = ypk num:ypk+num D(1,1)=D(1,1)+1; D(1,2)=D(1,2)+i; D(2,1)=D(1,2); D(2,2)= D(2,2)+i*i ; D(5,1)=D(2,2); D(1,5)=D(2,2); D(1,3)=D(1,3)+j; D(3,1)=D(1,3); D(3,3)= D(3,3)+j*j; D(1,6)=D(3,3); D(6,1)=D(3,3); D(2,3)=D(2,3)+i*j; D(3,2)=D(2,3); D(4,1)=D(2,3); D(1,4)=D(2,3); D(4,2)=D(4,2)+i*i*j; D(2,4)=D(4,2); D(5,3)=D(4,2); D(3,5)=D(4,2); D(5,2)=D(5,2)+i*i*i; D(2,5)=D(5,2); D(3,4)=D(3,4)+i*j*j; D(4,3)=D(3,4); D(6,2)=D(3,4); D(2,6)=D(3 ,4); D(4,4)=D(4,4)+i*i*j*j; D(3,6)=D(3,6)+j*j*j; D(6,3)=D(3,6); D(5,4)=D(5,4)+i*i*i*j; D(4,5)=D(5,4); D(6,4)=D(6,4)+i*j*j*j; D(4,6)=D(6,4); D(5,6)=D(4,4); D(6,5)=D(4,4); D(5,5)=D(5,5)+i*i*i*i; D(6,6)=D(6,6)+j*j*j*j; E(1)=E(1)+Z(j,i); E(2)=E(2)+Z(j,i)*i; E(3)=E(3)+Z(j,i)*j; E(4)=E(4)+Z(j,i)*i*j; E(5)=E(5)+Z(j,i)*i*i; E(6)=E(6)+Z(j,i)*j*j;

PAGE 115

115 end end %Ca lculate the coefficients using least squares method coeff=inv(D)*E; %Find the maximum A=[ 2*coeff(5) coeff(4) coeff(4) 2*coeff(6)]; B=[ coeff(2) coeff(3)]; maxpositions=inv(A)*B; xrefined=maxpositions(1); yrefined=maxpositions(2); %Check t o make sure the maximum is reasonable and if not return integer values if (abs(xrefined xpk)>3) xrefined=xpk; end if (abs(yrefined ypk)>3) yrefined=ypk; end %Calculate cvalue at maximum from coefficients for troubleshooting if needed cvalue=coe ff(1)+(coeff(2)*xrefined)+(coeff(3)*yrefined)+(coeff(4)*xrefined*yr efined)+(coeff(5)*xrefined*xrefined)+(coeff(6)*yrefined*yrefined); %End of refineparaboloidnew .m Averaging Position Software This code was written and used for refining the positions of d ense bodies by taking into account possible shape changes during the time series. To improve accuracy the correlation calculation was done using a template from each frame compared to every other frame. These positions are then all averaged to give an aver aged position. The code is semi automated in that the user must select the feature in the initial frame and the code will compute positions for a user defined number of frames. The code calculates two separate coordinates for each frame as mentioned in A.1 %Start of averagingtracking.m clear all close all

PAGE 116

116 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %set program parameters n=1; %starting image for analysis threshold=.6; % set thresh holding parameter num=4; % this is the size ar ound the correlation peak to which we fit a paraboloid %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %Import file generated by Tracking Program load tracking n k=m; imag=imag1; %Select template in first image like before [temp late, rectcrop] =imcrop(imag1{1},[min(min(imag{1})),max(max(imag{1}))]); rectcrop(1)=round(rectcrop(1)); rectcrop(2)=round(rectcrop(2)); width=round(rectcrop(3)); height=round(rectcrop(4)); %Round template positions for accurate coordinate calculations [ template, rectcrop]=imcrop(imag1{1},[rectcrop(1),rectcrop(2),width,height]); %Store template and template position save_template{1,1}=template; save_rectcrop{1,1}=rectcrop; %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %Find the centroid of the particle in the first image imag_centr{1}=template; imag_max=max(max(imag_centr{1})); imag_max=double(imag_max); size_imag=size(template); for i = 1:size_imag(1) for j = 1:size_imag(2) if imag_centr{1}(i,j)
PAGE 117

117 ycord(1,1)=s(I).Centroid(2)+rectcrop(2) 1; xCent(1,1)=xcord(1,1); yCent(1,1)=ycord(1,1); %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %Do Correlation for first particle to update positions of particles. This will use the blacked out images created in the Tracking Program. Since there are no neighboring particles the tracking should be very accurate. This also creates a template for each frame which is later used. for g=n:m image=imag1{g}; [c,corner_y(g),corner_x(g),xrefined,yrefined,cvalue]=process(template,image,n um); %This gives the corner position in the ith image of the old template (i.e. from the i 1st image) %This is the measured offset immediately between imag es x_offset(1,g)=round(rectcrop(1)) corner_x(g); y_offset(1,g)=round(rectcrop(2)) corner_y(g); cvalues(1,g)=cvalue; %Saves the correlation based positions for each frame based on correlation with the first frame. xcord(1,g)=xcord(1,n) x_offset( 1,g); ycord(1,g)=ycord(1,n) y_offset(1,g); %Calculate new template position for each frame and crop update template. templatex=round(xcord(1,g) width/2); templatey=round(ycord(1,g) height/2); [template1,rectcrop1]=imcrop(imag{g},[templatex,templatey, width,height]); %Find centroids of particle in each frame to use as starting position for later correlation calculation imag_centr{g}=template1; imag_max=max(max(imag_centr{g})); imag_max=double(imag_max); size_imag=size(template1); for i = 1:size_imag (1) for j = 1:size_imag(2) if imag_centr{g}(i,j)
PAGE 118

118 [maxval, I]=max(cell2mat(asd)); %Store centroid calculate position for each frame xCent(1,g)=s(1).Centroid(I)+rectcrop1(1) 1; yCent(1,g)=s(1).Centroid(I)+rectcrop1(2) 1; %Reset template positions with centroids and recrop templates templat ex=round(xCent(1,g) width/2); templatey=round(yCent(1,g) height/2); [template1,rectcrop1]=imcrop(imag{g},[templatex,templatey,width,height]); %Store templates and template coordinates for use later save_template{1,g}=template1; save_rectcrop{1,g}=re ctcrop1; end %Goes through and uses each template for correlation calculation against every frame. For example use template from frame 2 against frames n through m then move onto template from frame 3. for h=n+1:k %Load template and template coordina tes for respective frame. template=save_template{1,h}; rectcrop=save_rectcrop{1,h}; template1 =template; rectcrop1 =rectcrop; %Store coordinates calculated from first correlation calculation as reference point. xcord(h,h)=xcord(1,h); ycord(h,h)=ycord( 1,h); for g=n:m %Load blacked out image for frame g image=imag1{g}; %Do correlation calculation as mentioned before [c,corner_y(g),corner_x(g),xrefined,yrefined,cvalue]=process(template,image,n um); %Calculate offsets as mentioned before. Thi s represents the displacement of the particle from frame g to frame g x_offset(h,g)=round(rectcrop(1)) corner_x(g); y_offset(h,g)=round(rectcrop(2)) corner_y(g); cvalues(h,g)=cvalue; %Calculate coordinates of particle xcord(h,g)=xcord(h,h) x_o ffset(h,g); ycord(h,g)=ycord(h,h) y_offset(h,g);

PAGE 119

119 end end %Calculate average positions and save as variables (xfinal,yfinal) xsum=cumsum(xcord,1); xfinal=xsum(k,:)/(m n+1); ysum=cumsum(ycord,1); yfinal=ysum(k,:)/(m n+1); %Plot cloud of coordinat es as well as averaged coordinates for each frame for l=n:m figure(l) subplot(2,1,1) imshow(imag{l},[min(min(imag{n})),max(max(imag{n}))]) hold on plot(xcord(:,l), ycord(:,l), 'bx') hold on plot(xfinal(l), yfinal(l), 'rx') su bplot(2,1,2) imshow(imag1{l},[min(min(imag{n})),max(max(imag{n}))]) hold on plot(xfinal(l), yfinal(l), 'rx') end %Save the results with as averagingdensebody# where # is the number of the save averagingdensebody# %End of averagingtracking.m Sarcomere Length Calculation This code takes the positions of individual dense bodies calculated from the Averaging Position program and calculates sarcomere lengths with respect to time. The code is writ ten to handle up to five dense bodies/ four sarcomeres at time but can be expanded to handle more. In the case where all sarcomeres are not adjacent ignore spacings calculated between non adjacent dense bodies. For example if dense bodies 1,2, and 3 corres pond to penultimate sarcomeres but dense bodies 4 and 5 are

PAGE 120

120 somewhere else in the stress fiber ignore the spacing calculated between dense body 3 and 4. %Start of sarcomerecalculator.m clear all close all %Load files of all dense bodies tracked load aver agingdensebody1 load averagingdensebody2 load averagingdensebody3 load averagingdensebody4 load averagingdensebody5 %Input time increment t=543; %Input pixel size in microns pixelsize=.14; %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%% %Import image files and process them to create image with all dense bodies labeled matfiles =dir(fullfile('C:','Documents and Settings', 'BobRussell','Desktop', 'CellCutFiles','Dec5Experiments', 'Cell13top', '*.tif')); size_matfiles=size(matfiles); fo r i =1:size_matfiles(1) %temp_image = imread(matfiles(size_matfiles(1) i+1).name); % we reverse the sequence of images so that we start from 'bad particles' and move to betterones temp_image = imread(matfiles(i).name); temp_image=rgb2gray(temp_image); temp _image = imadjust(temp_image, stretchlim(temp_image), [0 1]); size_image=size(temp_image); [p,q]=size(temp_image); %Use this code to add extra rows or columns to D=zeros(p,q+10); D(1:p,6:q+5)=temp_image; imag{i}=temp_image(:,10:size_image(2)); imag{i}=D; e nd %Calculate Spacings based on each coordinate set calculated in Averaging program spacing45=((((x4 x5).^2)+ ((y4 y5).^2)).^.5)*pixelsize; spacing34=((((x3 x4).^2)+ ((y3 y4).^2)).^.5)*pixelsize; spacing23=((((x2 x3).^2)+ ((y2 y3).^2)).^.5)*pixelsize; spa cing12=((((x1 x2).^2)+ ((y1 y2).^2)).^.5)*pixelsize; %May want total spacing %spacingtotal=((((x1 x3).^2)+ ((y1 y3).^2)).^.5)*.14; %Now calculate mean spacing which is what is used for model fitting and further analysis

PAGE 121

121 mean45=mean(spacing45,1); mean34=m ean(spacing34,1); mean23=mean(spacing23,1); mean12=mean(spacing12,1); %Calculate mean positions for each dense body. This corresponds to xfinal and yfinal from the averaging program but they will all be overwritten because they all have the same name x1mean=mean(x1,1); x2mean=mean(x2,1); x3mean=mean(x3,1); x4mean=mean(x4,1); x5mean=mean(x5,1); y1mean=mean(y1,1); y2mean=mean(y2,1); y3mean=mean(y3,1); y4mean=mean(y4,1); y5mean=mean(y5,1); %Calculate max and min spacing to generate plots meanmax1=ma x(mean12(1:m))+.1; meanmax2=max(mean23(1:m))+.1; meanmax3=max(mean34(1:m))+.1; meanmax4=max(mean45(1:m))+.1; meanmin1=min(mean12(1:m)) .1; meanmin2=min(mean23(1:m)) .1; meanmin3=min(mean34(1:m)) .1; meanmin4=min(mean45(1:m)) .1; tmax=(m*t)/1000; %Gener ate image at each time point with all tracked dense bodies labeled for l=1:m figure (l) subplot(1,3,2) imshow(imag{l},[min(min(imag{1})),max(max(imag{1}))]) hold on plot(x1mean(l), y1mean(l), 'bx') hold on plot(x2mean(l), y2mean(l), 'bx') hold on plot(x3mean(l), y3mean(l), 'bx') hold on plot(x4mean(l), y4mean(l), 'bx') hold on plot(x5mean(l), y5mean(l), 'bx') end %Plot spacings versus time for each sarcomere for l=2:m

PAGE 122

122 figure (m+1) title(' Sarcomere Spacing 1,2','FontSize',18) hold on plot(((l 2)*t)/1000,spacing12(l), 'bx') xlabel('Time (sec)','FontSize',16) ylabel('Spacing (microns)','FontSize',16) axis([0 tmax meanmin1 meanmax1]) figure(m+2) title('Sarcomere Spacing 2,3','FontSize',18 ) hold on plot(((l 2)*t)/1000,mean23(l), 'bx') xlabel('Time (sec)','FontSize',16) ylabel('Spacing (microns)','FontSize',16) axis([0 tmax meanmin2 meanmax2]) figure(m+3) title('Sarcomere Spacing 3,4','FontSize',18) hold on plot(((l 2)*t)/1000,mean34(l), 'bx') xlabel('Time (sec)','FontSize',16) ylabel('Spacing (microns)','FontSize',16) axis([0 tmax meanmin3 meanmax3]) figure(m+4) title('Sarcomere Spacing 4,5','FontSize',18) hold on plot(((l 2)*t)/1000,mean45(l), 'bx') xlabel('Time (sec)','FontSize',1 6) ylabel('Spacing (microns)','FontSize',16) axis([0 tmax meanmin4 meanmax4]) end %End of sarcomerecalculator.m Model Fitting Code Sarcomere trajectories after severing that displayed a sustained linear contraction were fi t to a piecewise linear function corresponding to the mechanical model discussed in Chapter 2. In cases where an initial elastic contraction was observed but no linear contraction trajectories were fit to a single phase model. The MATLAB function

PAGE 123

123 fminsearc h was used to minimize the sum of the square of the residuals to fit parameters. Four P arameter F itting %Start of fourparameterfitting.m clear all close all %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %Enter time data and spac ing data as column vectors %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% spacing=[ ]; t=[ ]; %Provide initial guesses for fminsearch. These are some characteristic values that should lead to a reasonably solution linear0=[.1, 9, .02, .1]; linear1=[.8, .1, .1]; %Perform optimization using fminsearch to get model parameters [linearparam, linearval] =fminsearch(@(linearparam)twolinefit(linearparam,t,spacing),linear0); %Get variables from row vector linearparam A=linearparam( 1); B=linearparam(2); m=linearparam(3); C=linearparam(4); sizet=length(t); %Calculate F(x) in order to get the residuals for i=1:sizet if t(i)==0 F1(i)=C+B+A m*t(i); end if t(i)>0 F1(i)=B+A m*t(i); end F2(i)=B; end f=[F1',F2']; g=max(f,[],2);

PAGE 124

124 %These are the time dependent values of the fit function linearspacing=g(:,1); %Calculate the residuals linearresid=spacing linearspacing; %Normalize the spacing as described in Chapter 2 and calculate the dimension less time normspacing=(spacing B)/A; tc=A/m; normtime=t/tc; %Plot data along with line fit figure (1) plot(t, spacing, 'o') hold on plot(t, linearspacing,'b') %Plot residuals figure(2) plot(t, linearresid, 'bx') %End of fourpar ameterfitting.m This is the function that describes the two phase linear model to which the data is fit. %Start of twolinefit.m function F =twolinefit(linearparam,t,spacing) A = linearparam(1); B=linearparam(2); m= linearparam(3); C=linearparam(4); s izet=length(t); %This loop creates the piecewise linear function. for i=1:sizet if t(i)==0 F1(i)=C+B+A m*t(i); end if t(i)>0 F1(i)=B+A m*t(i); end F2(i)=B; end f=[F1',F2']; g=max(f,[],2);

PAGE 125

125 F=g(:,1); %Determine t he sum of the square of the residuals, F, to minimize SR=(spacing F).^2; SSR=cumsum(SR); F=SSR(sizet); %Start of twolinefit.m Two P arameter F itting %Start of twoparameterfitting.m clear all close all %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%% %Enter time data and spacing data as column vectors %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% t=[]; spacing=[]; %Provide initial guesses for fminsearch. These are some characteristic values that should lead to a reasonably solution guess=[.1,.9] ; %Perform optimization using fminsearch to get model parameters [zeroparam, linearval]= fminsearch(@(zeroparam)zeroslopeline(zeroparam,t,spacing),guess); sizet=length(t); C=zeroparam(1); B=zeroparam(2); % Calculate F(x) for i=1:sizet if t(i)==0 F2(i)=C+B; end if t(i)>0 F2(i)=B; end end %Plot data along with line fit plot(t, spacing, 'o') hold on plot(t, F2,'k') %Start of twoparameterfitting.m

PAGE 126

126 This is the function t hat describes the single phase linear model to which the data is fit. %Start of zeroslopeline.m function F =zeroslopeline(zeroparam,t,spacing) C=zeroparam(1); B=zeroparam(2); sizet=length(t); %This loop creates the single phase function. for i=1:siz et if t(i)==0 F2(i)=C+B; end if t(i)>0 F2(i)=B; end end F=F2; %Determine the sum of the square of the residuals, F, to minimize SR=(spacing F').^2; SSR=cumsum(SR); F=SSR(sizet); %End of zeroslopeline.m

PAGE 127

127 APPE NDIX B MATLAB CODE FOR DETE RMINING MICROTUBULE CURVATURE These programs were written and used to determine the curvature of microtubules. The programs adapt previously published methods to estimate the coordinates of a microtubule to sub pixel accuracy and then determines the root mean squared curvature of buckled microtubules. Microtubule Tracing Code This code is used to reliably and reproducibly determine the coordinates of a microtubule to sub pixel accuracy. The code requires the user to trace the gen eral shape of the microtubule by clicking the mouse. The methods used to refine the coordinates are discussed in the Materials and Methods section of Chapter 4. This code works on only one frame at a time but could be made semi automated in the future. %St art of microtubuletracing.m clear all close all %Load series of images matfiles=dir(fullfile('H:','MTBendingNoChange','May8cell22nucleuscut','crop', '*.tif'); %Provide name for file name=''May8cell22nucleuscut '; size_matfiles=size(matfiles); %Loop to load all images for i =1:size_matfiles(1) temp_image = imread(matfiles(i).name); %These lines allow for manipulations to the make the image appear better but are probably not necessary for this application. %temp_image=rgb2gray(temp_image); %temp_image = imadjust(temp_image, stretchlim(temp_image), [0 1]); size_image=size(temp_image); [p,q]=size(temp_image); %Use this code to add extra rows or columns to %imag{i}=temp_image(:,10:size_image(2)); imag{i}=temp_image; end

PAGE 128

128 %Determines size of line scan t o fit Gaussian distribution for sub pixel position refining. Should not need to change this. num=2; %Subtract back ground and apply blur if needed. This uses a custom function to try to enhance the image to make the microtubule easier to analyze. %Se t the frame number of interest in either line %image=double(imag{4}); %[image2]=convolve(image); %If not needed use this image2=double(imag{4}); %The user must provide the program with an initial trace of the microtubule. In the future it could be p microtubule but the user input seems to work very well. Click with the mouse along the microtubule contour as many times as possible. End the trace by figure(1) imshow(image2,[min(min(image 2)),max(max(image2))]) compimage=zeros(size_image); [x,y]=ginput; %The initial selected points are displayed on the microtubule for later reference. imshow(image2) hold on plot(x,y,'r*') user defined points. The height and width are calculated by determining the distance between adjacent points and then adding and arbitrary number. The slices are then cropped out with cropping coordinates determined by the relative positions of adjacent y coordinates. All slices are saved. for i=1:length(x) 1 %calc height and width width(i)= round(abs(x(i) x(i+1)))+4; height(i)=round(abs(y(i) y(i+1)))+4; if y(i)
PAGE 129

129 savetemp{i}=double(template); saverect{i}=rectcrop1; end %All of the slices are compiled to make a composite image of only the microtubule excluding the surrounding area close all for i=1:length(width) dimen=saverect{i}; compimage(dimen(2):dimen(2)+dimen(4),dimen(1):dimen(1)+dimen(3))=savetemp{i}; end %Display the composite image for later reference figure(2) imshow(compimage,[min(min(image2)),max(max(image2))]) hold on %This is a rule based algorithm to check the alignment of microtubule. Does cumsum on the rows and then columns and then compares the STD for the cumsum s and determines if horizontal or vertical. Larger STD should result if mtube is aligned in opposite direction of cumsum. Seems to work very well. Save direction of each point count=1; for i=1:length(width) sizetemp=size(savetemp{i}); temp1=cumsum (savetemp{i},1); sdtemp1=std(temp1(sizetemp(1),:)); temp2=cumsum(savetemp{i},2); sdtemp2=std(temp2(:,sizetemp(2))); if sdtemp1>sdtemp2 direction(i)=1; else direction(i)=2; end %Now use directions to calculate the local maximum in a one pixel wide slice either horizontally or vertically across the microtubule. Coordinates are saved and counted for j=2:sizetemp(direction(i)) 1 if direction(i)==1 [C,I]=max(savetemp{i}(j,:)); xc(j)=I; yc(j)=j; xcord (count)=I+saverect{i}(1) 1; ycord(count)=j+saverect{i}(2) 1; dirsave(count)=direction(i); else [C,I]=max(savetemp{i}(:,j));

PAGE 130

130 xc(j)=j; yc(j)=I; xcord(count)=j+saverect{i}(1) 1; ycord(count)=I+saverect {i}(2) 1; dirsave(count)=direction(i); end count=count+1; end clear j end %Display points refined by maxima figure(3) imshow(image2,[min(min(image2)),max(max(image2))]) hold on plot(xcord,ycord,'xb') %The initial point refinement will result in many duplicate points as a result of the overlapping slices. This code is meant to eliminate the duplicates and spurious points. First all points are sorted with respect to the x direction which requires that the microtubule to appear as a f unction y=f(x) cords=[xcord', ycord',dirsave']; cords=sortrows(cords); xcord=cords(:,1); ycord=cords(:,2); dirsave=cords(:,3); count2=1; xcordnew=xcord(1); ycordnew=ycord(1); dirsavenew=dirsave(1); for i=2:length(xcord) %First check if the x co ordinate is unique. If it is unique but the ycord is not this is most likely a flat portion of the microtubule where the y position can be refined to subpixel accuracy. if xcord(i)~=xcord(i 1) if (ycord(i)==ycord(i 1))&&(dirsave(i)==2) count2=count2+1; xcordnew(count2)=xcord(i); ycordnew(count2)=ycord(i); dirsavenew(count2)=dirsave(i); else count2=count2+1; xcordnew(count2)=xcord(i); ycordnew(count2)=ycord(i); dirsavene w(count2)=dirsave(i); end else

PAGE 131

131 %If the xcord is not unique and ycord is then save these points. This is sharply sloped portion of the microtubule where the x positions can be refined to subpixel accuracy if (xcord(i)==xcord(i 1))&& (yc ord(i)~=ycord(i 1))&&(dirsave(i)==1) count2=count2+1; xcordnew(count2)=xcord(i); ycordnew(count2)=ycord(i); dirsavenew(count2)=dirsave(i); end end end %It is n ecessary to check that the direction associated with each point is optimal. This code checks each point by determining difference between the intensity at the maximum and the average intensity of the pixels two away in both the x and y direction. For a poi nt located in horizontal part of a microtubule there should be a larger difference in the vertical direction then the horizontal. If these values are too close to each other then the direction is undetermined and is noted. for i=1:length(xcordnew) tempinten=image2(ycordnew(i),xcordnew(i)); xslope=(image2(ycordnew(i),xcordnew(i)2)+image2(ycordnew(i),xcordnew(i)+2))/; xslope=tempinten xslope; yslope=(image2(ycordnew(i)2,xcordnew(i))+image2(ycordnew(i)+2,xcordnew(i)))/; yslope=tempi nten yslope; if (xslope/yslope<.05 || yslope/xslope<.05) dirsavenew(i)=3; else if yslope< xslope dirsavenew(i)=1; else dirsavenew(i)=2; end end end %Use the directions to fit a Gaussian a cross the microtubule on single pixel slice to get sub pixel resolution %Create a kernel for the Gaussian fitting and define slice size to send to fitting function. kernel1=zeros(2*num+1,1); imagepart=kernel1; kernel1=( num:1:num); %Start display to sh figure(4)

PAGE 132

132 hold on imshow(image2,[min(min(image2)),max(max(image2))]) hold on clear i %Perform Gaussian fitting across microtubule to refine coordinates to subpixel accuracy. If direction was undefined as menti oned above then fit in both directions and average offsets. for i=1:length(xcordnew) xguess=xcordnew(i); yguess=ycordnew(i); dir=dirsavenew(i); if dir==2 plot([xguess, xguess], [yguess num, yguess+num]) ima gepart=image2(yguess num:yguess+num,xguess); [offset]=gaussfit(imagepart,kernel1,num); yfinal(i)=yguess+offset; xfinal(i)=xguess; end if dir==1 plot([xguess num, xguess+num], [yguess, y guess]) imagepart=image2(yguess,xguess num:xguess+num); [offset]=gaussfit(imagepart,kernel1,num); xfinal(i)=xguess+offset; yfinal(i)=yguess; end if dir==3 plot([xguess num, xguess+num], [yguess, yguess],'g') plot([xguess, xguess], [yguess num, yguess+num],'g') imagepart1=image2(yguess num:yguess+num,xguess); [offset1]=gaussfit(imagepart1,kernel1,num); imagepart=image2(yguess,xgue ss num:xguess+num); [offset]=gaussfit(imagepart,kernel1,num); xfinal(i)=xguess+offset/2; yfinal(i)=yguess+offset1/2; end end

PAGE 133

133 %cords1=[xfinal', yfinal']; %cords1=sortrows(cords1); %xfinal=cords1(:,1); %yfinal=co rds1(:,2); xfinaln=xfinal; yfinaln=yfinal; %Plot final coordinates figure(4) hold on plot(xfinaln,yfinaln,'rx') %Save data to pass to next program to calculate Bending Energy or Root Mean Squared Curvature save cord image2 compimage name xfinaln yfinaln %End of microtubuletracing.m Linear G aussian F itting F unction This function refines the coordinates of the maximum of the correlation function to sub pixel accuracy by fitting a Gaussian to a line slice across the microtubule. %Start of gaussfit.m %This function refines the peak found in the correlation spectrum by fitting it to a Gaussian function. It returns the refined x and y peak positions(sub pixel) as well as the correlation value at the peakfunction [yrefined]=gaussfit(imagepart ,kernel) %Create matrices for the calculation function [offset]=gaussfit(imagepart,kernel1,num) A=zeros(2,2); B=zeros(2,1); D=zeros(3,3); E=zeros(3,1); coeff=zeros(6,1); Z=log(imagepart); %Populate matrix for the least squares fitting for i = 1:le ngth(kernel1) D(1,1)=D(1,1)+1; D(1,2)=D(1,2)+kernel1(i); D(2,1)=D(1,2); D(2,2)= D(2,2)+kernel1(i)*kernel1(i);

PAGE 134

134 D(3,1)=D(2,2); D(1,3)=D(2,2); D(3,2)=D(3,2)+kernel1(i)*kernel1(i)*kernel1(i); D(2,3)=D(3, 2); D(3,3)=D(3,3)+kernel1(i)*kernel1(i)*kernel1(i)*kernel1(i); E(1)=E(1)+Z(i); E(2)=E(2)+Z(i)*kernel1(i); E(3)=E(3)+Z(i)*kernel1(i)*kernel1(i); end coeff=inv(D)*E; %Determine sub pixel offset offset=coeff(2)/( 2*co eff(3)); %Check the offset to make sure it is reasonable if abs(offset)>num offset=0; end end %End of gaussfit.m Curvature and Bending Energy Calculation Code This code was written to calculate the curvature and the bending energy/ root mean squared curvature of microtubules from traces generated from the program B.1. This code uses a previously published method known as the three point method to calculate curvature. %Start of curvaturecalculator.m %This program is used to calculate the Be nding Energy or the RMS curvature from the microtubule trace calculated in the Microtubule Tracer close all clear all %Load coordinates generated from Microtubule Tracer load cord %Sort the coordinates by x one last time cords1=[xfinaln', yfinaln'] ; cords1=sortrows(cords1); xfinaln=cords1(:,1); yfinaln=cords1(:,2);

PAGE 135

135 %Smooth the coordinates to eliminate short range oscillations from imaging uncertainty but preserve long range microtubule buckles. x=smooth(xfinaln,5); y=smooth(yfinaln,5); %Specify the number of points for line segment nump=4; %Display image of microtubule figure (1) imshow(image2,[min(min(image2)),max(max(image2))]) hold on %Calculate segment lengths and slopes. Store starting and ending positions of each lines segment. Also count the line segments count=1 for i=1:nump:length(x) nump s(count)=sqrt((x(i) x(i+nump))^2+(y(i) y(i+nump))^2); m(count)=(y(i+nump) y(i))/(x(i+nump) x(i)); xpoint(count,:)=[x(i), x(i+nump)]; ypoint(count,:)=[y(i),y(i+nump)]; count=count+1; end %Plot the line segments on the microtubule plot(xpoint,ypoint,'r ') plot(xpoint,ypoint,'b*') %Calc curvature with three point method for i=1:length(s) 1 theta(i)=atan((m(i+1) m(i))/(1+(m(i)*m(i+1)))); curve2(i) =abs(theta(i)/((s(i)+s(i+1))/2)); end %Sum all segment lengths scum=cumsum(s); %Plot curvature a function of arclength figure(2) plot(scum(2:length(scum)),curve2, 'rx ') xlabel('arclength s, pixels'), ylabel('curvature') arcleng=scum(2:length(scum) ); %Calculate Bending Energy by numerically integrating curvature squared.

PAGE 136

136 BendingEnergy=trapz(arcleng,curve2.^2) Enperleng=BendingEnergy/arcleng(length(arcleng)) save(name) %End of curvaturecalculator.m

PAGE 137

137 LIST OF REFERENCES 1. Cines, D. B., E. S. Pollak, C. A. Buck, J. Loscalzo, G. A. Zimmerman, R. P. McEver, J. S. Pober, T. M. Wick, B. A. Konkle, B. S. Schwartz, E. S. Barnathan, K. R. McCrae, B. A. Hug, A. M. Schmidt, and D. M. Stern. 1998. Endothelial cells in physiology an d in the pathophysiology of vascular disorders. Blood 91:3527 3561. 2. Prasain, N. and T. Stevens. 2009. The actin cytoskeleton in endothelial cell phenotypes. Microvasc Res 77:53 63. 3. Aird, W. C. 2003. Endothelial cell heterogeneity. Crit Ca re Med 31:S221 230. 4. Aird, W. C. 2008. Endothelium in health and disease. Pharmacol Rep 60:139 143. 5. Gloor, S. M., M. Wachtel, M. F. Bolliger, H. Ishihara, R. Landmann, and K. Frei. 2001. Molecular and cellular permeability control at the b lood brain barrier. Brain Res Brain Res Rev 36:258 264. 6. Aird, W. C. 2007. Phenotypic heterogeneity of the endothelium: I. Structure, function, and mechanisms. Circ Res 100:158 173. 7. Gebb, S. and T. Stevens. 2004. On lung endothelial cell h eterogeneity. Microvasc Res 68:1 12. 8. Ofori Acquah, S. F., J. King, N. Voelkel, K. L. Schaphorst, and T. Stevens. 2008. Heterogeneity of barrier function in the lung reflects diversity in endothelial cell junctions. Microvasc Res 75:391 402. 9. Dudek, S. M. and J. G. Garcia. 2001. Cytoskeletal regulation of pulmonary vascular permeability. J Appl Physiol 91:1487 1500. 10. Drexler, H. and B. Hornig. 1999. Endothelial dysfunction in human disease. J Mol Cell Cardiol 31:51 60. 11. Wojciak Stothard, B. and A. J. Ridley. 2002. Rho GTPases and the regulation of endothelial permeability. Vascul Pharmacol 39:187 199. 12. Gimbrone, M. A., Jr., T. Nagel, and J. N. Topper. 1997. Biomechanical activation: an emerging paradigm in endothelial adh esion biology. J Clin Invest 99:1809 1813. 13. Chien, S., S. Li, and Y. J. Shyy. 1998. Effects of mechanical forces on signal transduction and gene expression in endothelial cells. Hypertension 31:162 169. 14. Nehls, V. and D. Drenckhahn. 1991. Dem onstration of actin filament stress fibers in microvascular endothelial cells in situ. Microvasc Res 42:103 112.

PAGE 138

138 15. Barbee, K. A., T. Mundel, R. Lal, and P. F. Davies. 1995. Subcellular distribution of shear stress at the surface of flow aligned and n onaligned endothelial monolayers. Am J Physiol 268:H1765 1772. 16. Galbraith, C. G., R. Skalak, and S. Chien. 1998. Shear stress induces spatial reorganization of the endothelial cell cytoskeleton. Cell Motil Cytoskeleton 40:317 330. 17. Steinmetz, M. O., D. Stoffler, A. Hoenger, A. Bremer, and U. Aebi. 1997. Actin: from cell biology to atomic detail. J Struct Biol 119:295 320. 18. Welch, M. D., A. Mallavarapu, J. Rosenblatt, and T. J. Mitchison. 1997. Actin dynamics in vivo. Curr Opin Cell Biol 9:54 61. 19. Matsudaira, P. 1994. Actin crosslinking proteins at the leading edge. Semin Cell Biol 5:165 174. 20. Small, J. V., K. Rottner, I. Kaverina, and K. I. Anderson. 1998. Assembling an actin cytoskeleton for cell attachment and movement. B iochim Biophys Acta 1404:271 281. 21. Katoh, K., Y. Kano, M. Masuda, H. Onishi, and K. Fujiwara. 1998. Isolation and contraction of the stress fiber. Mol Biol Cell 9:1919 1938. 22. Balaban, N. Q., U. S. Schwarz, D. Riveline, P. Goichberg, G. Tzur, I. Sabanay, D. Mahalu, S. Safran, A. Bershadsky, L. Addadi, and B. Geiger. 2001. Force and focal adhesion assembly: a close relationship studied using elastic micropatterned substrates. Nat Cell Biol 3:466 472. 23. Geiger, B. and A. Bershadsky. 2001. A ssembly and mechanosensory function of focal contacts. Curr Opin Cell Biol 13:584 592. 24. Brakebusch, C. and R. Fassler. 2003. The integrin actin connection, an eternal love affair. Embo J 22:2324 2333. 25. Wong, A. J., T. D. Pollard, and I. M. He rman. 1983. Actin filament stress fibers in vascular endothelial cells in vivo. Science 219:867 869. 26. Lee, T. Y. and A. I. Gotlieb. 2003. Microfilaments and microtubules maintain endothelial integrity. Microsc Res Tech 60:115 127. 27. Moritz, M. and D. A. Agard. 2001. Gamma tubulin complexes and microtubule nucleation. Curr Opin Struct Biol 11:174 181. 28. Wade, R. H. 2009. On and around microtubules: an overview. Mol Biotechnol 43:177 191.

PAGE 139

139 29. Bicek, A. D., E. Tuzel, A. Demtcho uk, M. Uppalapati, W. O. Hancock, D. M. Kroll, and D. J. Odde. 2009. Anterograde microtubule transport drives microtubule bending in LLC PK1 epithelial cells. Mol Biol Cell 20:2943 2953. 30. Luders, J. and T. Stearns. 2007. Microtubule organizing centr es: a re evaluation. Nat Rev Mol Cell Biol 8:161 167. 31. White, G. E., M. A. Gimbrone, Jr., and K. Fujiwara. 1983. Factors influencing the expression of stress fibers in vascular endothelial cells in situ. J Cell Biol 97:416 424. 32. Drenckhahn, D and J. Wagner. 1986. Stress fibers in the splenic sinus endothelium in situ: molecular structure, relationship to the extracellular matrix, and contractility. J Cell Biol 102:1738 1747. 33. Kim, D. W., B. L. Langille, M. K. Wong, and A. I. Gotlieb. 1 989. Patterns of endothelial microfilament distribution in the rabbit aorta in situ. Circ Res 64:21 31. 34. Kim, D. W., A. I. Gotlieb, and B. L. Langille. 1989. In vivo modulation of endothelial F actin microfilaments by experimental alterations in she ar stress. Arteriosclerosis 9:439 445. 35. Pellegrin, S. and H. Mellor. 2007. Actin stress fibres. J Cell Sci 120:3491 3499. 36. Lazarides, E. 1975. Tropomyosin antibody: the specific localization of tropomyosin in nonmuscle cells. J Cell Biol 65:5 49 561. 37. Lazarides, E. and K. Burridge. 1975. Alpha actinin: immunofluorescent localization of a muscle structural protein in nonmuscle cells. Cell 6:289 298. 38. Weber, K. and U. Groeschel Stewart. 1974. Antibody to myosin: the specific visuali zation of myosin containing filaments in nonmuscle cells. Proc Natl Acad Sci U S A 71:4561 4564. 39. Sanger, J. W., B. Mittal, and J. M. Sanger. 1984. Interaction of fluorescently labeled contractile proteins with the cytoskeleton in cell models. J Cel l Biol 99:918 928. 40. Cramer, L. P., M. Siebert, and T. J. Mitchison. 1997. Identification of novel graded polarity actin filament bundles in locomoting heart fibroblasts: implications for the generation of motile force. J Cell Biol 136:1287 1305. 41 Wang, K., J. F. Ash, and S. J. Singer. 1975. Filamin, a new high molecular weight protein found in smooth muscle and non muscle cells. Proc Natl Acad Sci U S A 72:4483 4486.

PAGE 140

140 42. Adams, J. C. 1995. Formation of stable microspikes containing actin a nd the 55 kDa actin bundling protein, fascin, is a consequence of cell adhesion to thrombospondin 1: implications for the anti adhesive activities of thrombospondin 1. J Cell Sci 108 ( Pt 5):1977 1990. 43. Price, C. J. and N. P. Brindle. 2000. Vasodila tor stimulated phosphoprotein is involved in stress fiber and membrane ruffle formation in endothelial cells. Arterioscler Thromb Vasc Biol 20:2051 2056. 44. Langanger, G., M. Moeremans, G. Daneels, A. Sobieszek, M. De Brabander, and J. De Mey. 1986. T he molecular organization of myosin in stress fibers of cultured cells. J Cell Biol 102:200 209. 45. Turnacioglu, K. K., J. W. Sanger, and J. M. Sanger. 1998. Sites of monomeric actin incorporation in living PtK2 and REF 52 cells. Cell Motil Cytoskelet on 40:59 70. 46. Kreis, T. E. and W. Birchmeier. 1980. Stress fiber sarcomeres of fibroblasts are contractile. Cell 22:555 561. 47. Sanger, J. M. and J. W. Sanger. 1980. Banding and polarity of actin filaments in interphase and cleaving cells. J Ce ll Biol 86:568 575. 48. Deguchi, S., T. Ohashi, and M. Sato. 2006. Tensile properties of single stress fibers isolated from cultured vascular smooth muscle cells. J Biomech 39:2603 2610. 49. Kumar, S., I. Z. Maxwell, A. Heisterkamp, T. R. Polte, T. P. Lele, M. Salanga, E. Mazur, and D. E. Ingber. 2006. Viscoelastic retraction of single living stress fibers and its impact on cell shape, cytoskeletal organization, and extracellular matrix mechanics. Biophys J 90:3762 3773. 50. Lu, L., S. J. Oswald H. Ngu, and F. C. Yin. 2008. Mechanical properties of actin stress fibers in living cells. Biophys J 95:6060 6071. 51. Peterson, L. J., Z. Rajfur, A. S. Maddox, C. D. Freel, Y. Chen, M. Edlund, C. Otey, and K. Burridge. 2004. Simultaneous stretching and contraction of stress fibers in vivo. Molecular Biology of the Cell 15:3497 3508. 52. Besser, A. and U. S. Schwarz. 2007. Coupling biochemistry and mechanics in cell adhesion: a model for inhomogeneous stress fiber contraction. New Journal of Physi cs 9: 53. Stachowiak, M. R. and B. O'Shaughnessy. 2008. Kinetics of stress fibers. New Journal of Physics 10:

PAGE 141

141 54. Gittes, F., B. Mickey, J. Nettleton, and J. Howard. 1993. Flexural rigidity of microtubules and actin filaments measured from ther mal fluctuations in shape. J Cell Biol 120:923 934. 55. Brangwynne, C. P., F. C. MacKintosh, S. Kumar, N. A. Geisse, J. Talbot, L. Mahadevan, K. K. Parker, D. E. Ingber, and D. A. Weitz. 2006. Microtubules can bear enhanced compressive loads in living cells because of lateral reinforcement. J Cell Biol 173:733 741. 56. Dogterom, M. and B. Yurke. 1997. Measurement of the force velocity relation for growing microtubules. Science 278:856 860. 57. Ingber, D. E. 1993. Cellular tensegrity: defining ne w rules of biological design that govern the cytoskeleton. J Cell Sci 104 ( Pt 3):613 627. 58. Ingber, D. E. 1998. The architecture of life. Sci Am 278:48 57. 59. Stamenovic, D., S. M. Mijailovich, I. M. Tolic Norrelykke, J. Chen, and N. Wang. 2002 Cell prestress. II. Contribution of microtubules. Am J Physiol Cell Physiol 282:C617 624. 60. Brangwynne, C. P., F. C. MacKintosh, and D. A. Weitz. 2007. Force fluctuations and polymerization dynamics of intracellular microtubules. Proc Natl Acad Sci U S A 104:16128 16133. 61. Waterman Storer, C. M. and E. D. Salmon. 1997. Actomyosin based retrograde flow of microtubules in the lamella of migrating epithelial cells influences microtubule dynamic instability and turnover and is associated with micr otubule breakage and treadmilling. J Cell Biol 139:417 434. 62. Zhou, F. Q., C. M. Waterman Storer, and C. S. Cohan. 2002. Focal loss of actin bundles causes microtubule redistribution and growth cone turning. J Cell Biol 157:839 849. 63. Koonce, M P., J. Kohler, R. Neujahr, J. M. Schwartz, I. Tikhonenko, and G. Gerisch. 1999. Dynein motor regulation stabilizes interphase microtubule arrays and determines centrosome position. Embo J 18:6786 6792. 64. Dujardin, D. L., L. E. Barnhart, S. A. Stehm an, E. R. Gomes, G. G. Gundersen, and R. B. Vallee. 2003. A role for cytoplasmic dynein and LIS1 in directed cell movement. J Cell Biol 163:1205 1211. 65. Malikov, V., A. Kashina, and V. Rodionov. 2004. Cytoplasmic dynein nucleates microtubules to orga nize them into radial arrays in vivo. Mol Biol Cell 15:2742 2749.

PAGE 142

142 66. Baas, P. W., A. Karabay, and L. Qiang. 2005. Microtubules cut and run. Trends Cell Biol 15:518 524. 67. Ferenz, N. P. and P. Wadsworth. 2007. Prophase microtubule arrays undergo flux like behavior in mammalian cells. Mol Biol Cell 18:3993 4002. 68. Hergott, G. J., M. Sandig, and V. I. Kalnins. 1989. Cytoskeletal organization of migrating retinal pigment epithelial cells during wound healing in organ culture. Cell Motil Cytoske leton 13:83 93. 69. Remuzzi, A., C. F. Dewey, Jr., P. F. Davies, and M. A. Gimbrone, Jr. 1984. Orientation of endothelial cells in shear fields in vitro. Biorheology 21:617 630. 70. Deck, J. D. 1986. Endothelial cell orientation on aortic valve lea flets. Cardiovasc Res 20:760 767. 71. Dartsch, P. C. and E. Betz. 1989. Response of cultured endothelial cells to mechanical stimulation. Basic Res Cardiol 84:268 281. 72. Hayakawa, K., N. Sato, and T. Obinata. 2001. Dynamic reorientation of cultur ed cells and stress fibers under mechanical stress from periodic stretching. Exp Cell Res 268:104 114. 73. Wang, J. H., P. Goldschmidt Clermont, J. Wille, and F. C. Yin. 2001. Specificity of endothelial cell reorientation in response to cyclic mechanic al stretching. J Biomech 34:1563 1572. 74. Yano, Y., Y. Saito, S. Narumiya, and B. E. Sumpio. 1996. Involvement of rho p21 in cyclic strain induced tyrosine phosphorylation of focal adhesion kinase (pp125FAK), morphological changes and migration of end othelial cells. Biochem Biophys Res Commun 224:508 515. 75. Ghosh, K., C. K. Thodeti, A. C. Dudley, A. Mammoto, M. Klagsbrun, and D. E. Ingber. 2008. Tumor derived endothelial cells exhibit aberrant Rho mediated mechanosensing and abnormal angiogenesis in vitro. Proc Natl Acad Sci U S A 105:11305 11310. 76. Pradhan, S. and B. Sumpio. 2004. Molecular and biological effects of hemodynamics on vascular cells. Front Biosci 9:3276 3285. 77. Lazarides, E. 1975. Immunofluorescence studies on the struct ure of actin filaments in tissue culture cells. J Histochem Cytochem 23:507 528. 78. Gordon, W. E., 3rd. 1978. Immunofluorescent and ultrastructural studies of "sarcomeric" units in stress fibers of cultured non muscle cells. Exp Cell Res 117:253 260.

PAGE 143

143 79. Sanger, J. W., J. M. Sanger, and B. M. Jockusch. 1983. Differences in the stress fibers between fibroblasts and epithelial cells. J Cell Biol 96:961 969. 80. Shemesh, T., B. Geiger, A. D. Bershadsky, and M. M. Kozlov. 2005. Focal adhesions as m echanosensors: a physical mechanism. Proc Natl Acad Sci U S A 102:12383 12388. 81. Luo, Y. Z., X. Xu, T. Lele, S. Kumar, and D. E. Ingber. 2008. A multi modular tensegrity model of an actin stress fiber. Journal of Biomechanics 41:2379 2387. 82. Ch en, J., B. Fabry, E. L. Schiffrin, and N. Wang. 2001. Twisting integrin receptors increases endothelin 1 gene expression in endothelial cells. Am J Physiol Cell Physiol 280:C1475 1484. 83. Edlund, M., M. A. Lotano, and C. A. Otey. 2001. Dynamics of alp ha actinin in focal adhesions and stress fibers visualized with alpha actinin green fluorescent protein. Cell Motil Cytoskeleton 48:190 200. 84. Alenghat, F. J., S. M. Nauli, R. Kolb, J. Zhou, and D. E. Ingber. 2004. Global cytoskeletal control of mech anotransduction in kidney epithelial cells. Exp Cell Res 301:23 30. 85. Gelles, J., B. J. Schnapp, and M. P. Sheetz. 1988. Tracking kinesin driven movements with nanometre scale precision. Nature 331:450 453. 86. Cheezum, M. K., W. F. Walker, and W H. Guilford. 2001. Quantitative comparison of algorithms for tracking single fluorescent particles. Biophys J 81:2378 2388. 87. Umemoto, S. and J. R. Sellers. 1990. Characterization of in vitro motility assays using smooth muscle and cytoplasmic myos ins. J Biol Chem 265:14864 14869. 88. Endlich, N., C. A. Otey, W. Kriz, and K. Endlich. 2007. Movement of stress fibers away from focal adhesions identifies focal adhesions as sites of stress fiber assembly in stationary cells. Cell Motil Cytoskeleton 64:966 976. 89. Hotulainen, P. and P. Lappalainen. 2006. Stress fibers are generated by two distinct actin assembly mechanisms in motile cells. J Cell Biol 173:383 394. 90. Hirata, H., H. Tatsumi, and M. Sokabe. 2007. Dynamics of actin filaments du ring tension dependent formation of actin bundles. Biochim Biophys Acta 1770:1115 1127. 91. Smith, M. A., E. Blankman, M. L. Gardel, L. Luettjohann, C. M. Waterman, and M. C. Beckerle. 2010. A zyxin mediated mechanism for actin stress fiber maintenance and repair. Dev Cell 19:365 376.

PAGE 144

144 92. Russell, R. J., S. L. Xia, R. B. Dickinson, and T. P. Lele. 2009. Sarcomere mechanics in capillary endothelial cells. Biophys J 97:1578 1585. 93. Fink, J., M. Thery, A. Azioune, R. Dupont, F. Chatelain, M. Born ens, and M. Piel. 2007. Comparative study and improvement of current cell micro patterning techniques. Lab Chip 7:672 680. 94. Sveen, J. K. 2006. MATPiv. Available at http://www.math.uio.no/~jks/matpiv/. Accessed August 20, 2009 95. Chen, C. S., M Mrksich, S. Huang, G. M. Whitesides, and D. E. Ingber. 1997. Geometric control of cell life and death. Science 276:1425 1428. 96. Parker, K. K., A. L. Brock, C. Brangwynne, R. J. Mannix, N. Wang, E. Ostuni, N. A. Geisse, J. C. Adams, G. M. Whitesides and D. E. Ingber. 2002. Directional control of lamellipodia extension by constraining cell shape and orienting cell tractional forces. Faseb J 16:1195 1204. 97. Tan, J. L., J. Tien, D. M. Pirone, D. S. Gray, K. Bhadriraju, and C. S. Chen. 2003. Cells lying on a bed of microneedles: an approach to isolate mechanical force. Proc Natl Acad Sci U S A 100:1484 1489. 98. Chrzanowska Wodnicka, M. and K. Burridge. 1996. Rho stimulated contractility drives the formation of stress fibers and focal adhesions J Cell Biol 133:1403 1415. 99. Dickinson, R. B. and D. L. Purich. 2002. Clamped filament elongation model for actin based motors. Biophys J 82:605 617. 100. Dickinson, R. B., L. Caro, and D. L. Purich. 2004. Force generation by cytoskeletal filamen t end tracking proteins. Biophys J 87:2838 2854. 101. Mogilner, A. and G. Oster. 2003. Force generation by actin polymerization II: the elastic ratchet and tethered filaments. Biophys J 84:1591 1605. 102. Kozlov, M. M. and A. D. Bershadsky. 2004. Proce ssive capping by formin suggests a force driven mechanism of actin polymerization. J Cell Biol 167:1011 1017. 103. Inoue, S. and E. D. Salmon. 1995. Force generation by microtubule assembly/disassembly in mitosis and related movements. Mol Biol Cell 6:16 19 1640. 104. Tran, P. T., L. Marsh, V. Doye, S. Inoue, and F. Chang. 2001. A mechanism for nuclear positioning in fission yeast based on microtubule pushing. J Cell Biol 153:397 411.

PAGE 145

145 105. Howard, J. 2006. Elastic and damping forces generated by confin ed arrays of dynamic microtubules. Phys Biol 3:54 66. 106. Gittes, F., E. Meyhofer, S. Baek, and J. Howard. 1996. Directional loading of the kinesin motor molecule as it buckles a microtubule. Biophys J 70:418 429. 107. Dujardin, D. L. and R. B. Vallee 2002. Dynein at the cortex. Curr Opin Cell Biol 14:44 49. 108. Burakov, A., E. Nadezhdina, B. Slepchenko, and V. Rodionov. 2003. Centrosome positioning in interphase cells. J Cell Biol 162:963 969. 109. Zhu, J., A. Burakov, V. Rodionov, and A. Mogiln er. 2010. Finding the cell center by a balance of dynein and myosin pulling and microtubule pushing: a computational study. Mol Biol Cell 21:4418 4427. 110. Bicek, A. D., E. Tuzel, D. M. Kroll, and D. J. Odde. 2007. Analysis of microtubule curvature. Met hods Cell Biol 83:237 268. 111. Brangwynne, C. P., G. H. Koenderink, E. Barry, Z. Dogic, F. C. MacKintosh, and D. A. Weitz. 2007. Bending dynamics of fluctuating biopolymers probed by automated high resolution filament tracking. Biophys J 93:346 359. 11 2. Goodwin, S. S. and R. D. Vale. 2010. Patronin regulates the microtubule network by protecting microtubule minus ends. Cell 143:263 274. 113. Cole, N. B., N. Sciaky, A. Marotta, J. Song, and J. Lippincott Schwartz. 1996. Golgi dispersal during microtu bule disruption: regeneration of Golgi stacks at peripheral endoplasmic reticulum exit sites. Mol Biol Cell 7:631 650. 114. Misra, G. 2010. Multi scale modeling and simulation of semi flexible filaments [Dissertation ]. Gainesville: University of Florida 115. Wu, J., G. Misra, R. J. Russell, A. J. C. Ladd, T. P. Lele, and R. B. Dickinson. 2011. Effects of Dynein on Microtubule Mechanics and Centrosome Positioning. Proc Natl Acad Sci U S A In Review. 116. Shiu, Y. T., J. A. Weiss, J. B. Hoying, M. N. Iwamoto, I. S. Joung, and C. T. Quam. 2005. The role of mechanical stresses in angiogenesis. Crit Rev Biomed Eng 33:431 510. 117. Discher, D., C. Dong, J. J. Fredberg, F. Guilak, D. Ingber, P. Janmey, R. D. Kamm, G. W. Schmid Schonbein, and S. Weinbaum. 2009. Biomechanics: cell research and applications for the next decade. Ann Biomed Eng 37:847 859.

PAGE 146

146 118. Russell, R. J., A. Y. Grubbs, S. P. Mangroo, S. E. Nakasone, R. B. Dickinson, and T. P. Lele. 2011. Sarcomere length fluctuations and flow in capillar y endothelial cells. Cytoskeleton (Hoboken) 68:150 156. 119. Stachowiak, M. R. and B. O'Shaughnessy. 2009. Recoil after severing reveals stress fiber contraction mechanisms. Biophys J 97:462 471. 120. Colombelli, J., A. Besser, H. Kress, E. G. Reynaud, P. Girard, E. Caussinus, U. Haselmann, J. V. Small, U. S. Schwarz, and E. H. Stelzer. 2009. Mechanosensing in actin stress fibers revealed by a close correlation between force and protein localization. J Cell Sci 122:1665 1679. 121. Katoh, K., Y. Kano, and S. Ookawara. 2008. Role of stress fibers and focal adhesions as a mediator for mechano signal transduction in endothelial cells in situ. Vasc Health Risk Manag 4:1273 1282. 122. Shirwany, N. A. and M. H. Zou. 2010. Arterial stiffness: a brief review. Acta Pharmacol Sin 31:1267 1276. 123. Pelham, R. J., Jr. and Y. Wang. 1997. Cell locomotion and focal adhesions are regulated by substrate flexibility. Proc Natl Acad Sci U S A 94:13661 13665. 124. Chin, L. K., J. Q. Yu, Y. Fu, T. Yu, A. Q. Liu, and K Q. Luo. 2011. Production of reactive oxygen species in endothelial cells under different pulsatile shear stresses and glucose concentrations. Lab Chip 125. Bazzoni, G. and E. Dejana. 2004. Endothelial cell to cell junctions: molecular organization and role in vascular homeostasis. Physiol Rev 84:869 901. 126. Millan, J., R. J. Cain, N. Reglero Real, C. Bigarella, B. Marcos Ramiro, L. Fernandez Martin, I. Correas, and A. J. Ridley. 2010. Adherens junctions connect stress fibres between adjacent endoth elial cells. BMC Biol 8:11. 127. Dembo, M. and Y. L. Wang. 1999. Stresses at the cell to substrate interface during locomotion of fibroblasts. Biophys J 76:2307 2316. 128. Miyake, K. and P. L. McNeil. 2003. Mechanical injury and repair of cells. Crit C are Med 31:S496 501. 129. Clarke, M. S., R. Khakee, and P. L. McNeil. 1993. Loss of cytoplasmic basic fibroblast growth factor from physiologically wounded myofibers of normal and dystrophic muscle. J Cell Sci 106 ( Pt 1):121 133.

PAGE 147

147 BIOGRAPHICAL SKETCH Robert J. Russell was born in Philadelphia, PA to Robert F. Russell and Nancy Clarke. After his graduation from Downingtown High School in 2001 he enrolled at Carnegie Mellon University in Pittsburgh, PA. While at Carnegie Mellon he double majored in c hemical e ngineering and b iomedical e ngineering with a minor in c olloids, p olymers and s urfaces. During his time at Carnegie Mellon Robert interned with MAB Paints in Philadelphia, P A and Bayer Materials Science in Pittsburgh, PA. Upon graduation from Carn egie Mellon in May 2006 he entered the Department of Chemical Engineering at the University of Florida as a graduate student in A ugust 2006. He started work under Professors Tanmay Lele and Richard Dickinson in January 200 7. During his time with Professor s Lele and Dickinson he has studied force generation in actin stress fibers and the microtubule cytoskeleton of endothelial cells.