Standardized Reporting of Data is not Much Use if the Data Themselves are Lousy
Keith Tipton
Department of Biochemistry, Trinity College, Dublin, Ireland
Efforts to ensure standardized reporting of data will be to little avail if the data themselves are obtained by flawed procedures or analysed incorrectly. The importance of ensuring that initial rate conditions apply to enzyme data, which has been stressed on many occasions, has become even more important now that high-throughput single time-point assays are increasingly used. Similarly, the use of progress curve analysis to determine kinetic parameters will only be valid under defined mechanisms of activity loss. Such considerations necessitate the use of adequate controls, which may, unfortunately, be neglected in the rush to obtain ‘publishable’ results. Such problems are often compounded by inadequate statistical analysis.
Unfortunately, even a cursory examination of the biochemical and related literature will show many examples of such inadequacy. Distinctions are not made between the types or error estimates presented. Parameters are frequently reported with the implication that they are error-free. Even when errors are reported for Km and Vmax (or kcat), it is common to find that no error estimates are associated with the ratio Vmax/Km (or kcat/Km), presumably because the authors do not know how these should be calculated and the journal editors do not care.
The, undeserved, popularity of the bar chart for presenting data has led authors to express their results as fractional or percentage changes, of the form100 (y/x) = 100 [(treated level) - (basal level)]. This means taking a ratio from a set of observations on the numerator (y) and another set on the denominator (x). These ratios may be paired observations, with some degree of correlation between the two; or x and y taken from two independent determinations with no question of correlation. The validity of using the ratio is based on the assumption of proportionality between these two values.
However, in many instances, one can expect the amount change following treatment may be additive, rather than proportional to basal levels. This may lead to artefacts in the magnitude of the effect induced by stimulation. Some specific examples of the confusion that can result from such problems will be discussed.
Thermodynamics of Enzyme-catalyzed Reactions
Robert A. Alberty
Department of Chemistry, Massachusetts Institute of Technology, Cambridge, MA, United States of America
My main interest is in the thermodynamics of reactions catalyzed by enzymes, but some of the things I am going to talk about apply to the kinetics of enzyme-catalyzed reactions. As a chemist I want to remind you that both of these fields are based on chemistry, which is concerned with reactions written in terms of species. The thermodynamics of reactions written in terms of species is described quantitatively by means of the Gibbs energy G, enthalpy H, and entropy S at specified temperature and pressure. When reactions are described in terms of sums of species, like ATP, their thermodynamics depend on the pH and are described quantitatively by means of the transformed Gibbs energy G', transformed enthalpy H', and transformed entropy S' at specified temperature, pressure, and pH. I have developed a database (BasicBiochemData3) that gives the standard Gibbs energies of formation DfG0 and standard enthalpies of formation DfH0 of species of biochemical interest at 298.15 K and zero ionic strength. DfG0 values are currently known for species of 199 reactants, but DfH0 are known for species of only 94 reactants. This database can be used to calculate standard transformed Gibbs energies of formation DfG'0 of reactants in the pH range 5 to 9 and ionic strengths from zero to about 0.35 M. These DfG'0 have been used to calculate standard transformed Gibbs energies of reaction DrG'0 and apparent equilibrium constants K' for 229 enzyme-catalyzed reactions. When DfH0 are known for all the species of a reactant, standard transformed Gibbs energies DfG'0 and standard transformed enthalpies DfH'0 of reactants can be calculated in the temperature range 273.15 K to 313.15 K, pH range 5 to 9 and ionic strengths from zero to about 0.35 M. This information has been used to calculate DrG'0, DrH'0, and apparent equilibrium constants K' for 90 reactions.
This has all been possible because biochemists and chemists have determined apparent equilibrium constants and enthalpies of enzyme-catalyzed reactions. The literature data has been summarized and evaluated by Goldberg and Tewari in six survey papers in J. Phys. Chem. Ref. Data.
It is important to emphasize the importance of ionic strength in the thermodynamics of enzyme-catalyzed reactions. According to the Debye-Huckel theory, the logarithm of the activity coefficient of an ion in water is proportional to its charge squared.
This means that the ionic strength effect for the species ATP4- is 16 times that for a chloride ion, a huge effect. It is also important to emphasize the difficulty in including the effect of magnesium ions. People determining apparent equilibrium constants of reactions in the presence of magnesium ion often give the total magnesium concentration, but it is pMg that affects the value of K'. The effect of pMg on the hydrolysis of ATP has been calculated, but this is about the only reaction for which there is sufficient information about the dissociation of magnesium complex ions. The equations needed to make the calculations discussed here are sufficiently complicated that it is not practical to write them out by hand or to calculate properties for specified conditions by hand. Mathematica is especially convenient for both deriving and evaluating functions of temperature, pH, and ionic strength that yield various thermodynamic properties.
I have just completed a book “Biochemical Thermodynamics” written in Mathematica, that contains computer programs and calculations for utilizing the species data in BasicBiochemData3.
Since biochemists need thermodynamic properties at specified temperatures, pHs, and ionic strengths, tables and plots cannot satisfy these needs. This new book contains a CD that makes available 774 mathematical functions for these properties of reactants. These functions can be added and subtracted to obtain changes in thermodynamic properties in biochemical reactions and apparent equilibrium constants. It also contains the functions of temperature, pH, and ionic strength, that yield the standard transformed Gibbs energies of reaction for 90 enzyme-catalyzed reactions. By taking derivatives of these functions, the standard transformed enthalpies, standard transformed entropies, and changes in the binding of hydrogen ions can be calculated, plotted, or be made into tables.
In closing I want to make some specific recommendations about how to write enzyme-catalyzed reactions:
(1) Hydrogen ions or other ions should not be shown.
(2) It is better to use NADox and NADred, rather than NAD+ and NADH because hydrogen ions are not to be balanced.
(3) It is better to use CO2tot, rather than CO2(g), because there is no gas phase in a cell. (When this change is made, H2O has to be added to the other side of the biochemical equation.)
(4) It is better to use ammonia, rather than NH4+.
The Current Status of the IUBMB Recommendations (1981) on Symbolism and Terminology in Enzyme Kinetics
Athel Cornish-Bowden
CNRS – Bioenergétique et Ingénierie des Proteines, Marseille, France
As long as biochemists were mainly concerned with single-substrate reactions there was little necessity for a standardized system of symbols and terminology. If one person used the symbol k-1 for a rate constant that another called k2 then the small amount of confusion generated was easily resolved. With the development in the 1950s of serious interest in reactions of two or more substrates, the literature became much more difficult to read, however, because numerous symbols were needed and translation from one system to another was neither obvious nor trivial: the KAB of Bloomfield and co-workers was the same as Alberty's KAB, but their KA was Alberty's KAB/KA; Alberty's KA was the same as Cleland's Ka, but his KAB was Cleland's KiaKb; Dixon and Webb, in a system that became the basis of the Recommendations made by IUB in 1961, usedfor what Alberty called KA, etc. meanwhile Dalziel used a system that in many ways appeared quite different from any of these: the limiting rate, variously expressed as Vf, VAB, V1 or V in the systems mentioned, became f0/f12 in Dalziel's system.
In 1972 the IUB retreated, and although the 1961 Recommendations are mentioned (without a reference!) in the edition of Enzyme Nomenclature of that year, the section on Symbols of Enzyme Kinetics is brief, and includes nothing about symbols for reactions of more than one substrate. The next edition of Enzyme Nomenclature, which appeared in 1978, retreated still further and omitted the section entirely (without giving any indication of the reasons). The problems had not disappeared, however, and in 1978-1979 the views of numerous biochemists interested in kinetics were solicited. The Recommendations on Symbolism and Terminology in Enzyme Kinetics that were approved by IUB in 1981 were an attempt at a synthesis of these, preserving existing practices of biochemists as far as possible, while at the same time trying to bring biochemical practice into closer accord with the Report on Symbolism and Terminology in Chemical Kinetics that IUPAC had produced in 1981. Although IUB claimed in 1972 that their recommendations of 1961 had been “widely followed”, this assessment was more wishful thinking than fact.
Likewise, although the 1981 recommendations have had some influence on biochemical practice they have by no means been overwhelmingly adopted. After almost a quarter of a century there is no very strong reason to alter the existing recommendations, but they are certainly in need of extension. Increased interest in metabolic modelling has made it vital to consider reversible reactions, for example, but these are barely mentioned.
A Universal Rate Equation for Systems Biology
Johann M. Rohwer, Jan-Hendrik S. Hofmeyr and Arno J. Hanekorn
Department of Biochemistry, University of Stellenbosch, Stellenbosch, South Africa
Classical enzyme kinetics, as developed in the 20th century, had as primary objective the elucidation of mechanism of enzyme catalysis. In systems biology, however, the precise mechanism of an enzyme is less important; what is required is a description of the kinetics of enzymes that take into account the systemic context in which each enzyme finds itself. At the previous ESCEC symposium we presented the uni-substrate reversible Hill mechanism as an example of a rate equation that take into account (i) the kinetic, regulatory and thermodynamic properties of enzyme-catalysed reactions, (ii) the reversibility of all reactions, and (iii) modification of enzyme activity by allosteric effectors. Of particular significance was the result that the allosteric effect is abolished at saturating substrate concentrations, in the Hill mechanism, in contrast to the classical Monod-Wyman-Changeux (MWC) mechanism, where the allosteric effect does not saturate.
Here, we present three extensions of the above work. First, the reversible Hill equation is generalised to an arbitrary number of substrate-product pairs as well as to uni-bi and bi-uni reactions. The equation is further generalised to account for an arbitrary number of independently binding modifiers for the uni-uni, bi-bi and three substrate-three-product cases. This provides a universal rate equation for cooperative and non-cooperative reactions with or without allosteric modifier.
Second, the equation is evaluated by comparison to other models in current use. The bisubstrate reversible Hill equation was able to account for substrate-modifier saturation as in the uni-substrate case, in contrast to the MWC equation which did not. Moreover, setting the Hill coefficient to one yielded a universal equation that could successfully mimic the behaviour of various non-cooperative mechanistic models (ordered and random binding, ping-pong catalysis).
Finally, we present experimental data from Bacillus stearothermophilus pyruvate kinase with inorganic phosphate as allosteric inhibitor. These data show that the allosteric effect does indeed saturate, providing support for the reversible Hill equation and invalidating the MWC mechanism.
In conclusion, the proposed reversible Hill equations are all independent of underlying enzyme mechanism, they contain parameters that have clear operational meaning and can all be transformed to non-cooperative equations by setting the Hill coefficient equal to one. These equations are of great use in computational models, and we hope that they will lay the groundwork for a “new” enzyme kinetics for systems biology.
Assaying Enzymes from Hyperthermophiles
Wilfred R. Hagen
Department of Biotechnology, Delft University of Technology, Delft, The Netherlands
Extremophiles are forms of life that can exist under – from an antropocentric viewpoint – extreme environmental conditions. Some of such conditions, notably high temperature and also high pressure, cannot be prevented from propagation into the interior of the living cell, and thus to be a key determinant of extremophilic enzymology. Hyperthermophiles are microorganisms (archaea and bacteria) that grow optimally (i.e. with the shortest doubling time) above 80 C. Several species have optimal growth rate at T larger than 100 C. The maximum (sub-optimal) temperature of growth is presently documented at 121 C.
Biochemical studies suggest that the dogma of unity in biochemistry is unshaken: enzyme catalysis “in boiling water” is not fundamentally different from that of mesophiles living at moderate temperatures. Hyperthermophilic and mesophilic enzymes frequently exhibit homology in primary, secondary, tertiary, and quaternary structure, and hyperthermophiles apparently use the same substrates, coenzymes, and cofactors as mesophiles. Hyperthermophilic enzymes have increased thermal stability, which is clearly not explainable in terms of a single cause, but which rather seems to be the combined result of a plethora of small changes in structural elements (salt bridges, hydrogen bonds, shortened loops, modified amino acid usage, b-strand content, etc.), making rational design of thermo-stable enzymes difficult.
Hyperthermophiles do not grow at mesophilic temperatures, and enzymes purified from them usually have very low activity at room temperature. Apparently, hyperthermophilic enzymes are ‘stiff’ at low temperatures, and they acquire their optimal flexibility for catalysis only near the temperature of optimal cell growth. At this high temperature their kcat/KM may well be comparable to that of mesophilic homologues at mesophilic temperatures.
Assaying purified enzymes from hyperthermophiles at physiological temperature, e.g., 100 C, poses several problems, e.g.:
i. the purified enzyme may not retain its thermal stability in a standard buffer solution; substrates (example: glyceraldehyde 3-phosphate) and coenzymes (example: NADPH) are extremely unstable in dilute solutions at 100 C;
ii. the rate of the uncatalyzed reaction (or of uncatalyzed side reactions) may be high at elevated temperatures;
iii. it is not always a sinecure to modify and run assay instrumentation, such as a UV spectrometer, at 100 C. In practice, hyperthermophilic enzymes are frequently assayed, or used for synthetic purposes, at a ‘compromise’ temperature of circa 60 C, and this may compromise assignment of physiological function.
Assay of Enzymes with Difficult or Unknown Substrates: The “99” Oxidoreductases as an Example
Richard Cammack
Department of Chemistry, King's College, London, United Kingdom
In order to carry out metabolic reconstruction from enzyme kinetic data, it is necessary to measure catalytic activities that reflect the physiological function. The majority of the assays for use in diagnostics are for enzymes in Class 1, oxidoreductases, and Class 3, hydrolases. They use chromogenic or fluorogenic substrates designed for convenience of measurement, which exploit the breadth of specificity of the enzyme. In some cases however, particularly for enzymes from the older literature, the physiological substrate may be unidentified or misidentified.
In the EC list of enzymes there are some that are conventionally assayed with artificial substrate, for example oxidoreductases that are assayed with dyes or other small molecules as acceptors. Such activities may reflect the activity of parts of enzymes, or enzymes that are degraded. This has led to misunderstandings in textbooks and metabolic maps, such as the belief that free FAD is the acceptor for succinate dehydrogenase.
In some cases the physiological substrate may be difficult to use in a conventional assay. The substrate and product may be highly insoluble, so the rate of change of concentration is difficult to measure. Some of the enzymes presently in class 1.x.y.99 have acceptors such as membrane-bound quinones or electron-transfer proteins. In these cases, strategies are needed in order to design assays that reflect the physiologically relevant activity. Novel methods such as direct electrochemistry may be used for assays.
Investigations of Proteases - Suggestions
Hartmut Schlüter
Department of Physiology, Charité – Humboldt University Berlin, Berlin, Germany
Protease reactions are characterized by the nature of the reaction, which they catalyze, including the substrate specificity, the mechanism of the catalysis and the resulting reaction products as well as by their kinetics. Mass spectrometry (MS) is a useful tool for the qualitative description of enzymatic reactions since in most cases the enzymatic conversion results in a change of the molecular mass of the reactant. Furthermore, the chemical identity of the reaction products can be validated by tandem mass spectrometry instruments (MS/MS). With respect to the aims of ESCEC, standardizing experimental conditions and the result reports, it is recommended for the qualitative description of enzymatic reactions to use the natural substrates, if available, to monitor enzymatic reactions by mass spectrometry and to publish the mass spectrometric data. Furthermore, a comprehensive analysis of the identities of the proteins, which are present in the protein preparation, used for the experimental characterization of a protease, should be mandatory.
Leveraging the Structure-Function Relationships of Mechanistically Diverse Enzyme Superfamilies
Scott Pegg
Department of Biopharmaceutical Sciences, University of California, San Francisco, United States of America
The prediction of protein function from structure or sequence data remains a problem best addressed by leveraging information available from previously determined structure-function relationships. In the case of enzymes, the study of mechanistically diverse superfamilies can provide a rich source of structure-function information useful in functional determination and enzyme engineering. To access these relationships using a computational resource, several issues must be addressed regarding the representation of enzyme function, the organization of structure-function relationships in the superfamily context, the handling of misannotations, and reliability of classifications and evidence. The Structure-Function Linkage Database (http://sfld.rbvi.ucsf.edu) was developed as a solution to many of these issues.
Kinetic Characterization of Alcohol Dehydrogenases and Matrix Metalloproteinases: A Reflection on Standardization of Assay Conditions
Jan-Olof Winberg
Department of Biochemistry, University of Tromsø, Tromsø, Norway
The present paper will focus on the kinetic characterization of enzymes from two different types of families, Short Chain Dehydrogenases / Reductases (SDR) and Matrix matalloproteinases (MMPs). The former family includes over 3000 enzymes, and I have mainly worked with different allelic variants of alcohol dehydrogenase (ADH) from the fruitfly Drosophila melanogaster and the AdDH in Drosophila lebanonensis. Until now, approximately 25 MMPs are known in humans. I will here focus on both similarities and differences in problems regarding standardization of assay conditions and parameters that I have experienced during my work with these two different enzyme systems and to what extent I think it is possible to achieve standard conditions for a single enzyme or a group of enzymes.
Does Enzymology Need its own Ontology?
Kirill Degtyarenko
European Bioinformatics Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge, United Kingdom
The Enzyme Nomenclature is the only accepted systematic nomenclature for biochemical reactions. However, it suffers from inherent confusion between enzyme (object) and enzymatic reaction (process) concepts that propagates into the scientific literature and biological databases. I am going to describe three ontologies that can be (re)used to describe entities, processes and methods in enzymology. Ontology of Physico-Chemical Processes (REX) incorporates the enzyme-catalyzed reactions as a subset of catalytic reactions. Complex processes such as ATP- or coenzyme-dependent reactions can be represented as linear combination of partial reactions. Chemical Ontology provides a basis for classification of molecular entities and therefore can aid the classification of novel biochemical reactions. FIX ontology consists of two parts. The systems such as molecule or cell possess Physico-Chemical Properties, while the Physico-Chemical methods are used to elucidate these properties.
The Systems Biology Ontology
Nicolas LeNovere
European Bioinformatics Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge, United Kingdom
The rising popularity of Systems Biology, and its recognition as a new field of life science, called for the development by the community of standards and guidelines able to maximise the diffusion and reuse of its scientific production. The design of standard formal languages to encode models, such as SBML or CellML was a first step. However, past the syntax, it is timely to tackle the semantics. The Systems Biology Ontology, developed by the international initiative www. biomodels.net, aims to strictly index and define terms used in quantitative biochemistry. The ontology is made up of four orthogonal vocabularies. A controlled vocabulary defines the roles of reaction participants, such as “substrate”, “catalyst”, “competitive inhibitor” etc. A taxonomy orders the quantitative parameters used in biochemistry such as “Hill coefficient”, “Michaelis constant”, but also “first order forward rate constant”. A precise classification of rate laws defines them, such as “first order reversible mass action kinetics”, “Hill function” etc. Each term contains a precise mathematical expression stored as a MathML lambda function. The variables of the function refer to the controlled vocabularies described above. This classification will permit to differentiate between rate-laws expressed with identical mathematical expressions but based on different assumptions, such as “Henri-Michaelis-Menten” and “Briggs-Haldane” kinetics. Finally, a list of simulation frameworks, such as “discrete” or “continuous” precises the validity of a rate-law. The Systems Biology Ontology can be used to increase the semantic content of quantitative models. SBO vocabularies can also be used to annotate results of biochemical experiments in order to facilitate their efficient reuse.
Molecular Simulations of Enzyme Catalysis
Martin J. Field
Laboratoire de Dynamique Moleculaire, Institut de Biologie Structurale, Grenoble, France
Molecular modeling and simulation techniques have proved powerful tools for helping to understand how proteins and other biomacromolecules function at an atomic level. The study of enzyme reactions is a particularly challenging application of these methods because of the variety of processes of differing length and time scales that can contribute to catalysis. Among these processes are bond- breaking and forming chemical steps, the diffusion of ligands into and out of the active site and possible conformational changes in the enzyme's structure.
This talk will treat the following topics:
1. Give an overview of the range of molecular simulation techniques that are available for studying enzyme reactions.
2. Describe what such studies can and cannot achieve.
3. Outline the limitations of current approaches and possible ways of overcoming them.
4. Consider the interface between the approaches described here and those of systems biology.
Standards for Reporting Experimental Procedures – Example of a Synergistic Approach
Susanna-Assunta Sansone
European Bioinformatics Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge, United Kingdom
Several groups work collaboratively toward the establishment of communication standards for reporting and exchanging the experimental procedures (metadata) and the data. I will present the newly formed standards initiative under the Metabolomics Society umbrella. In particular, I will focus on its interactions with the Microarray Gene Expression Data Society (MGED) and HUPO Proteomics Standard Initiative (PSI) aiming to develop an integrated ontology to support the consistent annotation of the experimental procedures, regardless of the particular field of study. Undoubtedly, specialized information is needed by certain applications, but a high level unified model for description of metadata would be able to encompass these applications.
Analysis of High-Throughput Data on the Basis of Metabolic Network Models
Hermann-Georg Holzhütter
Institute of Biochemistry, Charité – Humboldt University Berlin, Berlin, Germany
Mathematical modeling of metabolic networks in conjunction with the application of optimization principles may help to better rational high-throughput data coming from transcriptomics, proteomics and metabolomics. This presentation provides two examples for such an approach. The first example (1) deals with the calculation of optimal temporal enzyme distributions across metabolic networks. Considering different functional objectives of the metabolism and metabolic systems of different complexity the calculated enzyme patterns are compared with high-throughput data obtained from promotor activity analysis of amino acid synthesizing pathways in E. coli and transcriptional profiling in yeast cells under conditions of the so-called dioxic shift. The second example demonstrates the application of flux-balance methods to the computation of stationary fluxes (2). A method is presented that enhances the feasibility of such flux-balance calculation by including high-throughput data from metabolomics.
(1) Klipp, E. Heinrich, R. and Holzhütter, H.G. (2002). Prediction of temporal gene expression. Metabolic optimization by re-distribution of enzyme activities. Eur. J. Biochem. 269:5406-5413.
(2) Holzhütter, H.G. (2004). The principle of flux minimization and its application to estimate stationary fluxes in metabolic networks. Eur. J. Biochem. 271:2905-2922.
Functional Implications of Changes in Gene Expression
Jildau Bouwman
Department of Molecular Cell Physiology, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands
Functional behavior of cells largely occurs at the level of “fluxes”, i.e. rates of processes such as product formation, protein production/degradation and gene expression. Although the mechanisms of these processes are known in great detail, it has not yet been possible to predict how changes of transcription eventually affect metabolic fluxes. We hypothesize that this is due to simultaneous regulation of all processes involved. Thanks to genomics, it should be possible to evaluate the implications of processes at the level of transcription for functional fluxes. Glycolysis in yeast is a good model system to test this relation, as it is one of the few pathways for which the kinetic properties of the enzymes are known sufficiently to calculate the flux from the enzyme activities and yeast can be brought under the well-defined steady-state and transient conditions.
In the project steady-state yeast cultures are perturbed in different ways (for instance by shifting from aerobic to anaerobic conditions, by temperature-shift and by shifting from respiratory to fermentative growth). Using regulation analysis (Rossell et al., 2004) we will quantify to what extent these changes are caused by changes in transcription, translation and/or metabolism. Concentrations of RNA, protein and metabolites, and enzyme activities and fluxes are measured quantitatively. Cultivation as well as sample analysis is done in 6 different labs. Therefore, standardization is a key issue in this project.
Investigations with Global Constraint-based Metabolic Models
Vincent Schachter
Genoscope, Centre national de séquençage, Evry, France
The availability of complete genomes, of the first metabolomics datasets, and the rise in expectations toward the explanatory and predictive powers of biological network models has given a new impulse to the study of metabolism, thought by many to be a well-understood field a few years ago. One important aim is metabolic reconstruction: comparative methods confirm that procaryote metabolism, in particular, exhibits huge variability, and there are significant gaps in our knowledge of the best known network of metabolic reactions, that of E. coli. Another aim is to better understand the global metabolic behaviour of a (bacterial) cell, seen as a biochemical transformation machine interacting with its environment. Classical models based on sets of differential equations have limited applicability here, both because of the rarity of experimentally determined kinetic parameters, and because of the size of the networks involved.
Constraint-based modeling of metabolism is a semi-formal framework dedicated to the modelling of metabolic processes at steady state, i.e. a global state of the metabolic network is defined as a distribution of fluxes within the network reactions. It emerged in the 90s as a radical simplification of kinetic models and was developed to allow tractable modelling of genome-scale metabolic networks. During the last 4 years, it has been applied successfully to a variety of reconstruction, structural analyses and predictive tasks on large metabolic networks in bacteria and yeast, yielding non-trivial biological insights.
The steady-state hypothesis positions the framework at a level of detail intermediate between description of static network structure and representation of network dynamics. The focus, rather than being on fully instantiated descriptions of the system’s behavior, is on sets of such descriptions, i.e. sets of flux distributions compatible with a set of constraints representing the current knowledge on the structure of the network, on thermodynamic and kinetic parameters, and on input/output relationships of the network with its environment. This solution set can be refined incrementally as new constraints are added, ensuring some robustness in structural analyses and metabolic behaviour predictions with respect to modifications of the model.
These models, albeit based on very strong simplifying assumptions, can be used to predict with a reasonable degree of accuracy a number of qualitative observables.
They also provide a basis for interpretation of additional experimental information (e.g. metabolic fluxes, metabolic concentrations) as these become available, as well as for the design of more detailed models. They may also be used for theoretical investigations into the plasticity and evolution of metabolic function.
We will start by introducing the steady-state metabolic flux modeling framework and its constraint-based version.
We will then focus on two applications of the framework.
The first applications fits within the context of the “Metabolic Thesaurus” project, an experimental effort aimed at understanding the metabolism of, Acinetobacter ADP1 (a versatile, highly competent, strictly aerobic, gram-negative soil bacterium with biodegradative capabilities ) using large-scale phenotypic and biochemical data.
We will describe the reconstruction of a global metabolic model of Acinetobacter ADP1, up to a point where good agreement was reached between model predictions and a significant set of experimental data on single-deletion mutant growth phenotypes.
The second project is a theoretical study on the variability of flux-coupling patterns across a large set of simulated metabolic environments. Flux-coupling relationships correspond to pairs of reactions for which the value of the flux in one reaction constrains the value of the flux in the other reaction, over the entire set of flux distributions. We will present the approach and preliminary results.
Strategy to Develop Large-Scale Kinetic Models on the Basis of Enzyme Structural and Functional Information: Progress and Problems
Oleg Demin
AN Belozersky Institute of Physico-Chemical Biology, Moscow State University, Moscow, Russia
In this presentation we describe a strategy for constructing and investigation of large-scale kinetic network models. In the framework of our approach we suggest a way to collect, mine experimental data and use them to build and verify kinetic models. Our approach is exemplified by kinetic model of purine biosynthesis and histidine biosynthesis pathway in Escherichia coli.
The development of kinetic models for metabolic systems is accomplished in several steps. The first step is to find out all cellular players, intermediates, enzymes, small molecules, co-factors, and all non-enzymatic processes in the cellular network. The second stage is to generate rate equations to describe the dependence of each reaction rate on concentrations of intermediates involved in the selected pathway and experimental conditions such as pH and temperature. Parameter estimation is the third stage of model development. To estimate the kinetic parameter values we us (i) experimentally measured dependencies of the initial reaction rates on intermediate concentrations, (ii) time series data from enzyme kinetics and (iii) pH and temperature dependencies of enzyme activity. The fourth stage of modelling ist to generate the corresponding differential equations for the pathway under investigation. Once the equations are generated, numerical integration can be accomplished to simulate a sequence of different biological scenarios and a variety of mathematical analyses are possible.
Systems Biology of Signal Transduction and Cancer
Ursula Klingmüller
Theodor-Boveri Group “Systems Biology of Signal Transduction”, German Cancer Research Center (DKFZ), Heidelberg, Germany
Dysregulation of cellular communication manifests in diseases such as the onset of cancer. Growth and differentiation of cells are regulated by the coordinated activation of signaling pathways. The components of many signaling cascades have been identified, yet it is largely unknown how information is processed and how decisions are taken. To elucidate this it is important to analyze timing, extent and duration of signal activation.
We are using a systems biology approach to examine the dynamic behavior of signaling pathways that are altered in cancers and predict targets for intervention. To generate novel biological information by a systems biology approach it is important that high quality quantitative data is used for mathematical modeling and that model predictions are experimentally verified. We are studying signaling pathways activated in hematopoietic and hepatocelllular system and have successfully used a systems biology approach to identified systems properties and predict steps suitable for effective perturbation.
SABIO-RK (System for the Analysis of Biochemical Pathways - Reaction Kinetics)
Isabel Rojas-Muijca, Martin Golebiewski, Renate Kania, Olga Krebs,
Andreas Weidemann and Ulrike Wittig
Scientific Databases and Visualization Group, EML Research gGmbH, Heidelberg, Germany
SABIO-RK is a web-accessible database designed to support researchers interested in information about biochemical reactions and their kinetics. It contains and merges information such as substrates, products and modifying compounds (i.e. inhibitors or activators) of reactions, information about the catalyzing enzymes, organism, tissue and cellular location where the reactions take place and the kinetic laws governing the reactions. The latter with their parameters and information about experimental conditions under which they were obtained (or measured). The source of each database entry is given in order to allow the user to refer to the original paper.
The user can search for reactions based on their reactants, characteristics or their involvement in reaction networks, as well as for the reaction's kinetics based on the experimental conditions under which they were determined. The system supports the creation of SBML (Systems Biology Mark-up Language) file containing the information selected by the user. This file can then be used as the basis for the definition of a simulation model in SBML.
The kinetic data contained in the database is manually extracted from literature and curated manually by biological experts. These curators are faced with problems like synonymic notations and descriptions, identification of inconsistencies in compounds and reactions within and between databases, multiplicity and lack of units, lacking information about assay proceedings and experimental conditions or the complexity in the description of methods and buffers. The lack of standards for publishing kinetic data impedes the efficient and automated extraction of these data, slowing down the population of the database.
Beyond Flat Files: Data Modelling, Editing, Archival and Interchange
Steffen Neumann
Department of Stress and Developmental Biology, Leipniz Institute for Plant Biochemistry, Halle/Saale, Germany
Though Flat Files and Spreadsheets were (and are) in widespread use, new projects in the life-science area should be (and many are) using structured data models for archive and interchange purposes.
This talk demonstrates some benefits of using the Universal Modelling Language (UML) and resulting Code, Editor, Database and Web Application generation. The examples used are current efforts in the Metabolomics area, where mzData and ArMet are proposed as (Meta-) Data Models.
New Developments at the BRENDA Enzyme Information System Host
Dietmar Schomburg
Cologne University Bioinformatics Center (CUBIC), University of Cologne, Cologne, Germany
BRENDA is the world's largest enzyme information system which is available via the online host in Cologne and as an in-house-system. At the Cologne host alone per month more than 100000 hits and more than 2000 visits per day are registered. The enzyme information system that originally was based solely on data manually extracted from the literature has recently been complemented by a number of software features that extends is usability significantly.
The new developments at the BRENDA host will be presented.
This includes
- an all-against-all sequence comparison of enzyme sequences with analysis of
- the enzyme sequence families with respect to enzyme function.
- the possibility to get information on membrane regions from the query system
- the development of an genome explorer for the BRENDA host
- the development of AMENDA, a text-mining extension of the enzyme database
For each of the new features the method and some examples will be discussed. In addition an analysis of the enzyme kinetic data stored in BRENDA with respect to the recommendations of the STRENDA committee will be presented.
Simulation and Parameter Estimation on the Basis of Experimental Enzymatic Data
Ursula Kummer
Department of Bioinformatics and Computational Biochemistry, EML Research gGmbH, Heidelberg, Germany
COPASI is a software tool which we have been developing in close collaboration with Pedro Mendes' group (VBI, USA). The software offers diverse simulation and analysis methods as well as parameter estimation. All of these methods have different demands in terms of the enzymatic data that form the basis for the computation. The methods and the respective demands are summarized in this talk.
Integrating Structural and Kinetic Enzymatic Information in Systems Biology
Matthias Stein, Razif R. Gabdoulline, Rebecca C. Wade
Molecular and Cellular Modelling Group, EML Research gGmbH, Heidelberg, Germany
The successful modelling of metabolic and signalling pathways in systems biology requires a detailed and consistent set of enzymatic kinetic parameters.
Sometimes, these parameters are not available from the literature or were only obtained under different experimental conditions.
Strategies and preliminary results will be presented for deriving quantitative structure-function relationships between protein structures and enzymatic rate constants.
In many cases, the enzyme's activity is related to the molecular interaction field between the substrate molecule and the enzyme protein. These interactions can be computed based on structural models for the protein.
One approach to estimate rate constants is then by computer simulations of e. g. the diffusion of substrate molecules to the active site of an enzyme [1].
An alternative to time-consuming computer simulations is to take an informatics approach to estimate rate constants. PIPSA (Protein Interaction Property Similarity Analysis) is a method that allows the automated analysis of a molecular interaction fields (for example the electrostatic potential) to classify the interaction properties of families of proteins [2].
We are developing quantitative PIPSA(qPIPSA) to correlate molecular interaction fields with reaction rate constants and derive estimates of missing enzymatic data.
For the application of this methodology, it is critical to have a reliable training set of known experimental rate constants. This aspect will be discussed in this contribution.
[1] R. R. Gabdoulline, U. Kummer, L. F. Olsen, R.C. Wade
Concerted simulations reveal how peroxidase compound III formation results in cellular oscillations. Biophys J. 85, 1421-1428 (2003).
[2] http://projects.villa-bosch.de/mcm/software/pipsa
Discovering Novel Enzymes and Pathways by Comparative Genomcis
Valérie de Crecy-Lagard
Department of Microbiology and Cell Sciences, University of Florida, Gainesville, FL, United States of America
Identifying the function of every gene in all sequenced organisms is one of the major challenges of the post-genomic era and is one of the obligate steps leading to systems biology approaches. This objective is far from reached. By different estimates, over 30-50% of the genes of any given organism are of unknown function, incorrectly annotated or given a broad nonspecific annotation.
Most genome functional annotations programs rely on homology based approach, using first simple Blast or FASTA scores then more elaborate, sensitive and precise algorithms stemming from the field of protein structure prediction. The inherent limitations of homology based approaches (only similar objects can be identified), has driven the development of non-homology based methods to link gene and function.
Integrative genome mining tools that can analyze gene clustering, phylogenetic distribution, or protein fusions on a multi-genome scale have been recently developed (such as the SEED database, theseed.uchicago.edu/FIG/index.cgi). These bioinformatics tools allow the experimental biologist to make predictions on unknown gene function that can be tested experimentally.
Applying these comparative genomic methods to the field of tRNA modification has allowed us to identify the function of eight enzyme families, unraveling novel enzyme activities, cases of orthologous displacements, novel pathways and potential drug targets. This work opens the problem of how to name enzymes discovered through comparative genomics methods and give them EC numbers as in general these enzymes have been very poorly described or were totally unknown. The number of enzymes discovered by these methods is steadily increasing and guidelines for the “gene discoverers” who are often not enzymologists need to be defined.
Project CyberCell: Calibrating Biophysical Data to the Complexity of the Intracellular Environment
Michael J. Ellison
Institute for Biomolecular Design, University of Alberta, Edmonton, Canada
The enormous capacity of the genomic, proteomic and metabolomic technologies to generate data, immediately shifted the challenge of discovery from the bench to the computer. At this formative stage, bioinformatics provided the foundation to organize, sort, store, relate and evaluate data significance, trends and patterns. More recently however, there has been an escalating effort to integrate the bench with the computer to predict dynamic cellular behaviour. Within this context, the computer is envisioned as a predictive instrument for both gaining insight and designing experiments centred on complex problems through model simulation. The experiment presents the means for validating and refining the evolving model as it has always done since the days of Newton. A major challenge for biochemical modeling is to accurately reflect the granular nature of the cell, since the effects of a crowded, irregular intracellular environment of rates and equilibria are poorly understood. By representing molecules and molecular complexes as a spatially distributed collection of discrete particles or agents, we can capture the discrete and probabilistic nature of molecular events that underpin much of cell physiology. In this approach, molecules are assigned properties of movement, size, location and activity appropriate to their chemical species. The type and magnitude of physical interactions and chemical reaction affinities are determined upon molecular collision from probability factors assigned to each step of the reaction path in the forward and reverse direction. By portraying chemical and physical interactions between discrete objects as simple stochastic events, it is possible to represent enzyme-based kinetics, thermodynamics and diffusion in complex spatial terms.
JWS Online; a Web-based Tool for Curation, Review, Storage and Analyses of Kinetic Models
Jacky Snoep
Department of Biochemistry, University of Stellenbosch, Stellenbosch, South Africa
With the increase in the precision and number of intracellular compounds that can be quantified, kinetic models have become important tools for the analysis of experimental data. Specifically in the field of systems biology many of the kinetic models are formulated at the level of the enzyme catalysed reaction step and the rate equations used in the models have a strong link to the enzyme kinetic mechanisms of the reactions.
JWS Online is a repository of published kinetic models and allows users to interrogate and analyze these models in a user-friendly environment using their web browser. In addition, JWS Online facilitates the reviewing of models described in manuscripts submitted to Microbiology, FEBS Journal, IEE proceedings Systems Biology and Metabolomics.
In my presentation I will highlight the functionality of JWS Online and indicate its links to other database initiatives and international research groups and journals.
