Modeling Dynein: The Gear-Shifting Motor step Manoranjan Singh, Roop Mallik, Steve Gross, and Clare Yu University of California, Irvine A Cell Is Like a City • • • • • • • • • • Workers Power Plant Roads Trucks Factories Library Recycling center Police Post office Communications • • • • • • • • • • Proteins Mitochondria Actin fibers, microtubules Kinesin, dynein, myosin Ribosomes Genome Lysosome Chaperones Golgi apparatus Signaling networks Intracellular Traffic + Nucleus + + © Scientific American + + Nucleus + + + Kinesin Dynein Mitochondria Vesicle Microtubules (MT) are like freeways and actin filaments are like local surface streets. Filaments Actin filament • 10 nm diameter • 2.77 nm rise • 26 subunits/74 nm repeat - end + end + end Microtubule 25 nm diameter 13 protofilaments - end Motors Walk Along Filaments Yildiz et al. Science 2003. Motor proteins move cargo along filaments Molecular Biology of the Cell, 3rd Ed, 1994 Biochemistry, 4ht Ed, 1995 Herpes Virus Transport in Neurons Along Microtubules • Virus Movie: VirusMov.mov Kinesin Myosin-V Dynein Cargo Cargo KLC Pi KR1 Dynactin binding KR2 MR2 Ca2+ Pi Head (ATPase) Stalk 1 c 3 6 4 5 2 KAPP KR3 MR1 Lever (?) KHC Head (ATPase) MT binding The anatomy of a dynein molecule Microtubule Stalk 4 5 3 6 2 7 ATP 1 ADP+Pi N.Hirokawa, Science, 279, 519 (1998) Figure edited to show stalk Stem Burgess et. al. Nature 421, 715 (2003) Dynein Can Shift Gears Dynein Dynactin binding Roop Mallik1, Brian Carter1, Stephanie Lax2, Stephen King2, Steven Gross1 Head 1UC Irvine (ATPase) 2Univ. of Missouri-Kansas City 2 1 3 Stalk c 6 4 5 Lever (?) The basic experiment Optical trap Plastic bead (450 nm) Burgess et al, Nature 2003 FT FM Dynein Load FTrap kx Bead displacement proportional to backward force …. Calibrate and measure !!! Typical trace of motion F = - kx Optical trap stiffness = spring constant = k = 0.011 pN/nm Mallik et al., Nature 427, 649 (2004). Dynein can change the size of its steps as it walks along microtubules depending on • Load • ATP Concentration • Possible step sizes ≈ 8, 16, 24, 32 nm • High load → small steps • Low load → large steps Step Size as a Function of Load at High [ATP] High load 8nm steps Intermediate load 15 nm steps Low load ~25 nm steps Very Low ATP, No Load Position vs. time Step Size Distribution Mostly 24, 32 nm steps Optical trap Stalling Force bead FT FM • • • • Motor attached to bead Motor walks along microtubules Dynein Laser tweezers pull on bead As bead moves a distance x from center of trap, it feels a spring force = F = -kx • If force = stalling force, motor cannot pull bead. Stalling force strongly dependent on available ATP Linear force-ATP curve Histogram of stall forces, 1 mM ATP Goal of Theoretical Modeling To reproduce dependence of step size, stalling force and velocity on [ATP] and load F • Traditional approach : Coupled differential eqns. • Alternative approach: Monte Carlo Simulations (Advantage: easy to deal with complicated nonlinear dependencies) MT Stalk 4 5 3 6 2 7 ATP 1 ADP+Pi stem Desired Features of ATP Binding MT Stalk • ATP can bind to (or unbind from) sites 1-4 • Step size decreases as number of bound sites increases # ATP 1 2 3 4 bound Step 32 24 16 8 size nm nm nm nm 5 6 7 4 ATP 3 2 1 ADP+Pi stem • Binding probability increases with [ATP] • Problem: High [ATP], no load → small step size (nonsense) • Solution: Different ATP binding affinities on different sites 1 3 kon kon kon4 kon2 • Dictates sequential binding: site #1, then #3, then #4, then #2 • Binding probability on sites 2-4 increases with load F (need fuel to haul cargo) ATP Hydrolysis needed for step to occur MT Stalk 5 6 7 4 ATP • Sites 1 and 3 hydrolyze ATP, but energy for step probably comes from hydrolysis at site 1 • Problem: Why is there a mixture of 24 and 32 nm steps at low [ATP] and no load? • Answer: Probability to hydrolyze ATP at site 1 increases if ATP bound to other sites, esp. site 3 • Question: What produces stalling? • Answer: Probability to hydrolyze ATP decreases with load (Harder to walk with load) • Reverse hydrolysis can occur (ADP + Pi → ATP) (reversal rate increases with load) 3 2 1 ADP+Pi stem Monte Carlo Simulation n = number of sites bound = {0, 1, 2, 3, 4} d = step size = (5 – n) · 8 nm (conjecture) 1. Bind (or unbind) ATP: Probability for n → n ± 1, Pbind(site i) = kion [ATP] Δt k2-4on= k2-4on(F=0)exp{Fa/kT} 2. Hydrolyze ATP at site 1: Phydrolyze = po exp{-αFd/kT} Δt where po → po/100 if n = 1 3. Reverse hydrolysis: Preverse = pr exp{(1- α )Fd/kT)} or take a step 4. Repeat Theoretical Position vs. Time high [ATP] low [ATP] Step Size Distribution with No Load Compare Theory and Experiment (low [ATP]) Monte Carlo results At various [ATP] Simulation Results of Step Size • Step size decreases as load increases • Agrees with experiment Low [ATP] = 100 μM High [ATP] = 1 mM Theoretical Predictions Velocity vs. [ATP] concentration at various loads Velocity vs. Load at various [ATP] concentrations Conclusions about Dynein • Dynein can change step size depending on load and [ATP] • Monte Carlo simulations in good agreement with experiment • No load: sites # 1 and # 3 bind ATP → large step size ~ 32 and 24 nm • Large load, 4 sites bind ATP → small step size ~ 8 nm Collaborators Dmitri Petrov, Steve Gross, Clare Yu, Manoranjan Singh (missing: Roop Mallik) THE END Kinesin: Comparison of Monte Carlo and Experiment Velocity vs. [ATP] Velocity vs. Load Open symbols: Experiment (Vissher et al., 1999); closed symbols: Monte Carlo; solid line: Michaelis-Menten formula Modeling Dynein: The Gear-Shifting Motor + Nucleus + + + Manoranjan Singh, Roop Mallik, Steve Gross, and Clare Yu University of California, Irvine Intracellular Traffic How is intracellular transport regulated? © Scientific American Motors Walk Along Filaments Motor proteins Myosin Kinesin Highway System of a Cell + + + + Nucleus + + Kinesin Mitochondria Dynein Vesicle MT ATP Fuels the Motor • • • • Stalk 4 5 3 6 2 7 1 ATP = Adenosine Triphosphate ATP ADP+Pi ATP has 3 phosphate ions Hydrolysis: 1 phosphate ions breaks off Stem ADP = Adenosine Diphosphate has 2 phosphate ions • Energy is released ATP → ADP + Pi Laser Tweezers • • • • • • • Focused laser beam Electric field E most intense at focal point E induces a dipole moment p in particle U=-p·E U minimized at focal point Particle trapped at focal point Particle moving away from focal point feels spring force: F = -kx = load Microtubules (MT) are like freeways and actin filaments are like local surface streets. How does the cell regulate the transport of vesicles? Position along microtubule (nm) Step size as function of load High load 8nm steps Intermediate load 15 nm steps Low load ~25 nm steps (High [ATP]) Time (sec) Stepsize changes as a function of load Intermediate load ~ 0.4 to 0.8 pN 15 nm steps Low load < 0.4 pN ~25 nm steps (High [ATP]) Motion at very low ATP No load Video tracking Step Size Distribution at No Load • • • • 40% 24 nm steps 40% 32 nm steps ~19% 16 nm steps ~1% 8 nm steps (?) • Low [ATP] Mixture of steps at no load Mean step-size is load-dependent Large Load step Small step Large Load step Small step Model for implementation of a gear Desired Features MT 4 • ATP can bind to sites 1-4 5 3 Stalk 6 • Step size decreases as number of bound sites 2 7 1 increases • Binding probability increases with [ATP] ADP+Pi ATP • Different ATP binding affinities on different sites dictates sequential binding (1, 3, 4, 2) • Binding probability on sites 2-4 increases with load F (need fuel to haul cargo) • ATP can unbind • Hydrolysis needed for step to occur • Sites 1 and 3 hydrolyze ATP, but energy for step probably comes from hydrolysis at site 1 • Probability to hydrolyze ATP at site 1 increases if ATP bound to other sites, esp. site 3 • Probability to hydrolyze ATP decreases with load (Harder to walk with load) • Reverse hydrolysis can occur (ADP + Pi → ATP) Monte Carlo Simulation • Low ATP • High ATP Theoretical Position vs. Time high [ATP] low [ATP] Step Size Distribution Low ATP High ATP Step Size Distribution Low ATP High ATP Step Size Distribution with No Load • Theory Experiment (low [ATP]) Step Size Distribution with No Load: Compare Theory and Experiment Step Size Distribution with No Load Compare Theory and Experiment (low [ATP]) Monte Carlo results At various [ATP] Step Size Distribution with No Load • Theory vs. Experiment • Low [ATP] • Monte Carlo Results • Varying [ATP] Simulation Results of Step Size • Step size decreases as load increases • Agrees with experiment Low [ATP] = 100 μM High [ATP] = 1 mM Predicted Velocity vs. Load