Introduction The Biochemical Simulation Environment (BISEN) package provides a framework for constructing mathematical models to simulate and analyze the kinetics of biochemical systems, integrating data on biochemical thermodynamics and kinetics into a set of model equations and computer codes (see figure 1). This package can generate systems of differential equations for user-specified multicompartment systems of enzymes and transporters accounting for proton and metal cation binding, electrophysiology, distribution of biochemical reactants (e.g. [ATP]) into multiple rapidly converting chemical species (e.g. [H-ATP], [K-ATP], [Mg-ATP]), buffering of protons and metal cations, and the effect of temperature, ionic strength and reaction thermodynamics [2, 3]. For large systems, manual construction of such models is error prone. Fig 1: A graphical illustration of the different sources of information. Included in the BISEN package are databases of physicochemical constants and kinetic models for enzymes and transporters. Included is a script that parses models specified in biochemical scripting language (BSL) into differential equations coded in a MATLAB M-file. The BSL syntax invokes structured lists of biochemical reactions (and associated enzymes) in different compartments, and lists of transporters between compartments. Small scale example To illustrate the effect of accounting for the aforementioned phenomena consider the small scale example depicted in figure 2. Fig 2: A small scale example demonstrating transport. Left: Cartoon of the model. Right: BSL input file for the BISEN model builder. Simulated time courses are shown in figure 3. In the example, we initially observe an exchange of ATP and ADP where ATP increases in the matrix. This exchange of matrix ADP for cytoplasmic ATP is driven by a concentration gradient and is associated with the reverse operation of the ANT transporter. Subsequently the ATP concentration in the matrix increases leading to reverse operation of the F1F0- ATPase transporter, leading to a net transfer of positive charge from the matrix side to the cytoplasmic side of the membrane. This results in a membrane potential, defined as the potential difference between cytoplasm and matrix. As ATP is consumed, the membrane potential diminishes again. Fig 3: Simulated time courses for the example shown in figure 2. BISEN: Biochemical Simulation Environment J. Vanlier 1 , F. Wu 2 , F. Qi 2 , K.C. Vinnakota 2 , Y. Han 2 , R.K. Dash 2 , F. Yang 3 , N.A.W. van Riel 1 , J.A.L. Jeneson 1 , P. A. J. Hilbers 1 and D.A. Beard 2* 1 BioModeling and BioInformatics, BioMedical Engineering, Eindhoven University of Technology, The Netherlands. 2 Department of Physiology, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226. email:[email protected] 3 Lilly-Singapore Centre for Drug Discovery Pte Ltd, Singapore / BioModeling and BioInformatics Large scale model To put BISEN to the test, a model of the tricarboxylic acid cycle [3] and oxidative phosphorylation (see figure 4) was successfully reproduced as a BISEN model. Fig. 4: Left: Schematic of the TCA cycle model (Wu, et al 2007). Right: Simulations (lines) and data (symbols) for state 2 and state 3 respiration. For the large scale example the input file consisted of roughly 50 declarations while the output model corresponded to 1200 lines of MATLAB code. Models of this size (and larger) cannot realistically be built without a tool like this. The modularity of the approach enables the user to compare different models for different enzymes and transporters while maintaining a database of model components. Models are collected at http://www.biocoda.org (see figure 5). Fig. 5: Biological components databank website Future work Features to export BISEN models in the Systems Biology Markup Language (SBML) and CellML formats are planned for future updates of the package. We are currently working with the CellML developers so that the next CellML specification is engineered to interface with the thermodynamic and ion dissociation databases. Availability BISEN can be obtained at: http://bbc.mcw.edu/BISEN References [1] Beard, D.A. and Qian, H. (2008) Biochemical reaction networks. In, Chemical Biophysics: Quantitative Analysis of Cellular Processes. Cambridge University Press, Cambridge, UK, 128-161. [2] Vinnakota, K.C., Wu, F., Kushmerick, M.J. and Beard, D.A. (2009) Multiple ion binding equilibria, reaction kinetics, and thermodynamics in dynamic models of biochemical pathways, Methods in Enzymology, 454, 29-68. [3] Wu, F., Yang, F., Vinnakota, K.C. and Beard, D.A. (2007) Computer modeling of mitochondrial tricarboxylic acid cycle, oxidative phosphorylation, metabolite transport, and electrophysiology, J Biol Chem, 282, 24525-24537. Funding: This work was funded by NIH grant HL072011. Kinetic models Thermodynamic constraints Large scale model Isolated enzyme studies Thermodynamic studies ‘Whole system’ studies Ion-ligand stability constants compartment cytoplasm 0.8425 0.4970 ATPASE E.ATPASE.0 CK E.CK.0 compartment matrix 0.6514 0.2106 transport cytoplasm matrix ANT T.ANT.1 F1F0ATPASE T.F1F0ATPASE.0 EOF Ion binding Water fraction Volume Model Identifier Compartment Model Identifier