Detailed Three- Dimensional Modeling of Cellular Signaling M. - PowerPoint PPT Presentation

Detailed Three- Dimensional Modeling of Cellular Signaling M. Wittmann, A. Eder, J.S. Wiegert, C.P. Bengtson, A. Hellwig, M. Knodel, R. Geiger, L.H. Ge, D. Bucher, C.M. Schuster, H. Bading, G. Wittum, G. Queisser Gillian Queisser G-CSC University of Frankfurt

Detailed Modeling Describe biophysical processes in space and time • => Set up systems of partial differential equations. Resolve morphology, ideally from microscopy • image reconstruction => Discretization of computational domain. Simulate biophysical signal processing on detailed • morphologies in space and time => Numerics: Discretization schemes, fast solvers. Gillian Queisser G-CSC University of Frankfurt

1. Modeling Nuclear Calcium Dynamics 2. Modeling Synaptic Transmission at the Drosophila NMJ Gillian Queisser G-CSC University of Frankfurt

1. Modeling Nuclear Calcium Dynamics Wittmann et al. (2009) The Journal of Neuroscience 29(47):14687-14700 Gillian Queisser G-CSC University of Frankfurt

Infoldings are formed by a membrane bilayer Dissociated hippocampal culture Hippocampal brain slice Gillian Queisser G-CSC University of Frankfurt

Surface Reconstructions Cell nuclei from hippocampal neurons Gillian Queisser G-CSC University of Frankfurt

Volume Reconstruction • Create volume mesh by tetrahedra grid generation (TetGen, http://tetgen.berlios.de/). • Integrate TetGen with Simulation Environment UG. Gillian Queisser G-CSC University of Frankfurt

Simulations on Reconstructed Morphologies – Cellular Calcium Signaling Synapse to nucleus communication Gillian Queisser M. Wittmann, H. Bading (2006) G-CSC University of Frankfurt

Single CCT Simulation Gillian Queisser G-CSC University of Frankfurt

Reconstructed Nuclei – Measurements • Infolded nuclei have larger surface area compared to spherical nuclei. • Volume of nuclei is nearly constant independent of morphology. Nuclear pore counts by imunoreactivity Increased membrane area leads to increase in nuclear pore complexes (NPCs) Gillian Queisser G-CSC University of Frankfurt

Nuclei form nuclear signaling microdomains • Depending on the compartment size, nuclei show different calcium dynamics in their microdomains (A-D). • With increased compartment ratio microdomain calcium levels become significantly higher (E). Gillian Queisser G-CSC University of Frankfurt

Frequency dependent dynamics Changes of activity dynamics due to a change in frequency #!!" #!!" ,-.//"01-2.34-564" ,-.//"01-2.34-564" +!" +!" ,275380./"690/59," ,275380./"690/59," *!" *!" )!" )!" !"#$%&'()*+( !"#$%&'()*+( (!" (!" '!" '!" &!" &!" %!" %!" $!" $!" #!" #!" !" !" !" )" #&" $#" $*" %'" &$" !" %'" )!" #!'" #&!" #)'" $#!" #,-().+( #,-().+( #!!" 0.1 Hz 0.5 Hz -./00"12.3/45.675" +!" -386491/0"7:106:-" *!" )!" !"#$%&'()*+( (!" '!" &!" %!" $!" #!" !" !" %,'" )" #!,'" #&" #),'" $#" Gillian Queisser #,-().+( 1 Hz Queisser, G., Wiegert, S., Bading, H. G-CSC (in press) Nucleus University of Frankfurt

Small Compartments Better Resolve Oscillating Ca 2+ Signals ER Tracker Experiment Model Blue White DPX AM 0.05 ! F/F 1s 1s Large Small 0.01 ! F/F 0.2s 0.2s -70mV 5mV 12 0.2s 0.5 10 0.4 8 Power Power 0.3 -3 x10 -3 6 x10 0.2 Gillian Queisser 4 0.1 2 G-CSC 0.0 0 University of Frankfurt 2 4 6 8 10 2 4 6 8 10 Frequency (Hz) Frequency (Hz)

Integration vs. Detection Nuclear Read-Out Signal Integration Frequency Detection Queisser, G., Wiegert, S., Bading, H. (in press) Nucleus Gillian Queisser G-CSC University of Frankfurt

Some Results Signal-regulation of Nuclear Geometry • Neuronal nuclei are extremely plastic and their geometry is controlled by NMDA receptor activation. • Synaptic NMDA receptors promote the formation of infoldings. • Extrasynaptic NDMA receptors – they direct neurons towards degeneration and cell death – lead to loss of nuclear infoldings. • How calcium signals are translated into structural alterations is still unknown. • ERK-MAP kinase pathway is required for the formation of infoldings (not presented). Nucleoplasmic Reticulum - Yes or No? • 3D reconstructions and EM results show that the invaginations are lined by both the inner and outer nuclear envelope, therefore the invaginated space is filled with cytosol, NOT ER lumen. • Nucleoplasmic reticulum does not exist in hippocampal neurons. • Infoldings enhance nuclear calcium signaling: 1.larger nuclear surface and increased number of NPCs 2.diffusion distances from cytosolic to nuclear locations are smaller. 3.Compartmentalization occurs allowing microdomains to regulate calcium dynamics differently from one another. Gillian Queisser G-CSC University of Frankfurt

Some Results Nuclear Geometry and Calcium Signaling • Compartments are often unequal in size, smaller ones better resolve high frequency calcium signals. • May be relevant for activation of calmodulin, the principal calcium sensor. • Calcium oscillations may be important to activate CREB-dependent transcription during LTP and memory formation. • This information relay may be optimized in small microdomains. • Spatial re-organization of chromosome territories may be caused due to changes in nuclear architecture and therefore affect gene transcription. Gillian Queisser G-CSC University of Frankfurt

2. Modeling Synaptic Transmission at the Drosophila NMJ Gillian Queisser G-CSC University of Frankfurt

Biological Framework The morphology of presynaptic specializations can vary • greatly ranging from classical single-release-site boutons in the central nervous system to boutons of various sizes harboring multiple vesicle release sites. Basis of this analysis were the well-characterized • glutamatergic synapses of larval neuromuscular junctions (NMJs) of Drosophila. Gillian Queisser G-CSC University of Frankfurt

Two types of Boutons Larval bodywall muscles of Drosophila are typically • innervated by two motor neurons of which one forms large ball-like Ib boutons and a second motor neuron forms smaller type Is boutons. Generate three-dimensional density profiles of presynaptic • vesicles to monitor the local distribution and dynamics of vesicles as a function of bouton morphology . Gillian Queisser G-CSC University of Frankfurt

Important Parameters Frequency • Vesicle output probability P o • Bouton size • Gillian Queisser G-CSC University of Frankfurt

Simulations Ib bouton, P o = 5% Is bouton, P o = 90% Simulation on 3D-reconstruction Gillian Queisser G-CSC University of Frankfurt

Synaptic Transmission at different input Frequencies In vivo In silico 50 � eEJP [mV] � 15 Hz � 40 � 30 Hz � 60 Hz � 30 � 80 Hz � 20 � 10 � 0 � 0 � 1 � 2 � 3 � 4 � 5 � Stimulus number [X1000] � Gillian Queisser G-CSC University of Frankfurt

Output Probability, Size and Frequency 10 � 2 � 10 � Av. ves/AP � Av. ves/AP � Av. ves/AP � 10% � 20% � Po = 5 % � Po = 5 % � 30% � 40% � Po = 40 % � Po = 40 % � 50% � 60% � Po = 90 % � Po = 90 % � 5 � 1 � 5 � 70% � 80% � 0 � 0 � 0 � 0 3 6 0 350 700 0 3 6 time [s] � time [s] � time [s] � 1 � 1 � Av. ves/AP � Av. ves/AP � It � Ib � Is � It � Ib � Is � 0,5 � 0,5 � 0 � 0 � 0 350 700 0 1000 2000 time [s] � time [s] � 1 � 1 � 10 Hz � 10 Hz � Av. ves/AP � Av. ves/AP � 20 Hz � 20 Hz � 32 Hz � 30 Hz � 40 Hz � 40 Hz � 50 Hz � 0,5 � 50 Hz � 0,5 � 64 Hz � 64 Hz � 80 Hz � 80 Hz � 100 Hz � 100 Hz � 0 � 0 � 0 600 1200 0 600 1200 time [s] � time [s] � Gillian Queisser G-CSC University of Frankfurt

Bouton Configurations 1,4 � Ib bouton, Po = 5 % � Averaged vesicle Optimal bouton release per AP � Is bouton, Po = 90 % � configurations 0,7 � predicted by model 0 � 0 � 100 � 200 � time [sec] � Native pattern – in vivo vs. in silico 40 � Ib bouton � Vesicle release � Is bouton � ! 30 � 20 � 13 10 � 0 � 0 � 0,4 � 0,8 � 1,2 � time [s] � Gillian Queisser G-CSC University of Frankfurt

Conclusions P o strongly affects the magnitude of total vesicle release at the beginning of • stimulation Bouton size primarily affects the endurance of active vesicle release suggesting • that larger boutons are better suited for long bouts of synaptic activity whereas smaller boutons are only reliable with short trains of stimulation. The model predicts that Ib boutons are utilized to transmit longer-lasting high- • frequency stimuli whereas Is boutons are better suited for few low-frequency events at high amplitudes. Taken together Is and Ib boutons complement each others functionality, • forming an efficient system across different time scales. Gillian Queisser G-CSC University of Frankfurt

Thanks To C.P. Bengtson D. Bucher A. Eder L.H. Ge R. Geiger A. Hellwig M. Knodel J.S. Wiegert H. Bading C.M. Schuster G. Wittum Gillian Queisser G-CSC University of Frankfurt