The post-genomic era is significantly characterized by a high integration and interdisciplinary of research resources from such diverse fields as computational biology, bioinformatics, functional genomics, structural biology, and proteomics. In this perspective, established biological systems can be comprehensively investigated in terms of interactions of individual or groups of proteins and enzymes as well as the behaviour of collective networks of such interactions. On the other hand, these systems can be re-examined in the light of new results that suggest novel associations between otherwise unrelated pathways and individual proteins.

Modern experimental technologies are providing seemingly endless opportunities to generate massive amounts of sequence, expression and functional data. Continuous advances and improvements have enabled proteome analyses to proceed with increased depth and efficiency. To capitalize on this enormous pool of information and in order to understand fundamental biological phenomena it is essential to collect, organize, categorize, analyse, and share data and results.

However, whilst the large international genome sequencing projects elicited considerable public attention with the creation of huge sequence databases, it has become increasingly apparent that functional data for the gene products, in particular for enzymes, has either limited accessibility or is unavailable. Additionally, although enzyme structural information has been rapidly accumulated in databases, little effort has been invested toward systematic characterization of enzyme functions.

The problem is twofold; deriving data from experimental work is expensive and very time consuming and it is inherently very difficult to collect, interpret and standardize published data since they are widely distributed among journals covering a number of fields, and the data itself is often dependent on the experimental conditions.

For these reasons a systematic and standardized collection of functional enzyme data is essential for the interpretation of the genome information.

The first ESCEC meeting in 2003 resulted in a general agreement that standardization of experiments and methods for enzyme characterization is definitely necessary and in the formation of the STRENDA Commission. STRENDA stands for Standards for Reporting Enzyme Data and the commission accompanies the upcoming series of ESCEC symposia.

This 2nd ESCEC symposium provided a platform to discuss a number of checklists worked out and presented by the STRENDA Commission. In general, these lists are intended to support the improvement of reporting enzyme data and can be found here.

We would like to thank particularly the authors who provided us with written versions of the papers that they presented. Special thanks go to all those involved with the preparation and organization of the workshop, to the chairmen who piloted us successfully through the sessions and to the speakers and participants for their contribution in making this workshop a success.

Frankfurt/Main, July 2007

Carsten Kettner
Martin G. Hicks

Since the pH is treated as an independent variable in biochemical thermodynamics, the Gibbs energy G does not provide the criterion for equilibrium, but the transformed Gibbs energy G’ does. The standard transformed Gibbs energy of formation Δ_fG’⁰ of a reactant (sum of species) can be calculated at the desired temperature, pH, and ionic strength if the standard Gibbs energies of formation Δ_fG⁰and standard entropies of formation Δ_fH’⁰of the species that make up the reactant are known. BasicBiochemData3 in MathSource provides species properties for 199 biochemical reactants and Mathematica programs for calculating apparent equilibrium constants K’ and other transformed thermodynamic properties of enzyme-catalysed reactions are given. This database can be extended, and the number of reactions for which apparent equilibrium constants can be calculated increases exponentially with the number of reactants in the database.

The aim of this study was to standardize the cultivation conditions of Saccharomyces cerevisiae and the in vitro enzyme activity assays between different laboratories. Furthermore, the conditions under which the enzyme activity measurements were carried out were adapted such that they would be as close as possible to in vivo conditions, thus yielding results which are relevant for Systems Biology. This approach is different from the classical enzymologists’ approach which is to optimize for the highest catalytic activity.

Saccharomyces cerevisiae strain CEN.PK113–7D was cultured in aerobic, glucose- limited chemostats under standardized conditions. It was shown that, in accordance with earlier interlab comparisons, the main culture characteristics, including biomass, dry weight, glucose flux, and mRNA levels of glycolytic enzymes were comparable between five different laboratories.

As could be expected, the V_max values of the glycolytic enzymes were lower when measured under in vivo-like conditions than in optimized assays, but still sufficient to account for the glycolytic capacity of the cells. The addition of a crowding agent (polyethylene glycol) hardly affected the measured enzyme activities.

The conventional assay method for the majority of enzymes envisages a reaction between substrates in aqueous solution. A measurable concentration of product is accumulated over time. This paradigm has served well for the characterization of many enzymes. Variations of the method, often using chromogenic or fluorogenic substrates, have been developed and are widely used for purposes such as clinical diagnosis and screening. There are some metabolically important enzymes for which the only published assay methods use artificial substrates. Some of these are oxidoreductases that use artificial mediators, and are listed in the EC list under EC 1.x.99. For computational reconstruction of the metabolism of a cell, however, it is necessary to use kinetic data from assays that reflect the physiological function in the cell, and the physiological substrates. For some oxidoreductases it is known, or considered likely that the acceptors are water-insoluble membrane-bound quinones such as ubiquinone or menaquinone, which present particular problems for measurement of kinetic parameters. Succinate dehydrogenase/fumarate reductase is considered as an example. The oxidoreductases from membranes must be rendered soluble by detergents, which alter their kinetic behaviour. Uncertainty about the way of measuring activity of such enzymes has led to confusion in textbooks and metabolic maps, such as the persistent myth that free FAD is the acceptor for succinate dehydrogenase and related enzymes. New strategies are discussed to measure electrontransfer flux, under conditions that reflect the physiological activity of membrane-associated oxidoreductases. An example is direct electrochemistry of enzymes adsorbed onto carbon surface. In favourable cases this method is able to observe electron flux both within and through individual enzyme molecules. The kinetic parameters and substrate specificity of membrane-bound oxidoreductases may be obtained in this way.

Recommendations on the symbolism and terminology of enzyme kinetics were approved by the International Union of Biochemistry in 1981. They were primarily necessitated by the need for a systematic treatment of reactions of more than one substrate, but some important omissions have subsequently become evident, and a decision is needed as to whether these warrant the preparation of new recommendations, and if so whether these should constitute a complete revision of the entire document, or just the preparation of some new sections.

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 being 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 an 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 analyse gene clustering, phylogenetic distribution, or protein fusions on a multi-genome scale have been developed recently. These bioinformatics tools allow the experimental biologist to make predictions on unknown gene function that can be tested experimentally and discover novel enzymes, regulators and transporters that expand our knowledge of metabolism in all species.

Molecular modelling 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 are the bond-breaking and forming chemical steps, the diffusion of ligands into and out of the active site and conformational changes in the enzyme's structure.

This contribution gives a synopsis of the range of molecular simulation techniques that are available for studying enzyme reactions with particular emphasis on methods designed for the investigation of the chemical catalytic steps. The capabilities and limitations of current approaches will be described and possible future developments discussed. Special attention is given to the interface between molecular simulation and systems biology modelling and to how the STRENDA guidelines would need to be adapted to allow the reporting of enzyme data determined from simulation.

The determination of kinetic and thermodynamic data from hyperthermophilic enzymes at physiological temperature (i. e. ‡ 80 C) raises a number of technical and fundamental problems. Based on studies of purified enzymes from the model organism Pyrococcus furiosus several of these problems are identified and explored here. It is proposed that kinetic and thermodynamic data on hyperthermophilic enzymes be reported at the organism's growth temperature or, alternatively, at a lower temperature compatible with practical assay conditions with additional data obtained at yet lower temperatures to allow for extrapolation.

The dynamic behaviour of metabolic networks is determined by the kinetic properties and the cellular levels of the enzymes and transporters involved. Changes in the concentrations of enzymes can be assessed by proteomics measurements or – more indirectly – by gene expression analyses. However, a straightforward interpretation of such data with respect to metabolic functions of the cell is difficult as a simple correlation between changes of enzyme levels and changes of fluxes in a metabolic network does not exist. Here we outline a theoretical concept to exploit information on changes of enzyme concentrations for predicting changes of stationary fluxes and this way to characterize changes in the functional status of cells or tissues. The basis of our concept is a novel variant of flux-balance analysis which we call MinMode-decomposition. The basic idea of this concept is to approximate flux distributions in metabolic networks as linear combinations of functionally motivated minimal flux modes (MinModes). They are defined as minimal flux modes supporting a unit flux through only one of the target reactions of the network. This theoretical concept will be applied to metabolic networks of bacteria (Methylobacterium extorquens) and human red blood. Based on simulated data we demonstrate that a good prediction of observed flux changes can be achieved if the decomposition of flux changes into MinModes is performed such that a maximal correlation with observed changes in enzyme activities is accomplished.

Modelling, simulation and computational analysis have become important tools in modern biochemistry. Moreover, their tight integration with experimental approaches has become an integral part of systems biology which has attracted scientific and political interest all over the world. However, published enzymatic data often does not take a modeller's viewpoint into account, even though in many cases this would only demand minor adjustments and would serve the community a great deal. Supporting users by automating some of the steps in modelling and simulation adds even more requirements. In the following we would like to emphasize a few points that we feel should be further supported or that have been neglected in the discussion about the standardization of enzymatic data, but would be valuable for modellers.

The need to exchange and integrate models drove the community to design common data format such as SBML. However, as important as was the definition of a common syntax, we also need to tackle the semantics of the models. The community recently proposed MIRIAM, the Minimal Information Requested in the Annotation of Models, a set of rules for curating quantitative models of biological systems. This standard lists the condition an encoded model has to meet to fully correspond to its reference description, and describe how to annotate each of its components. The Systems Biology Ontology (SBO) aims to strictly index, define and relate terms used in quantitative modelling, and by extension quantitative biochemistry. SBO is currently made up of five different vocabularies: quantitative parameters, participant roles, modelling frameworks, mathematical expressions – that refers to the three previous branches – and events. SBO can be used not only to annotate quantitative models, but also biochemical experiements. It is expected that the adoption of those two semantic layers will favour the reusability of quantitative biochemical descriptions, whether parameters or models.

Software engeneering today provides tools which minimize the need for manual coding of the typical components of an application, such as database, frontend and web application. Visual modelling brings together users and developers, and allows quick and direct communication about the topic. In the metabolomics community data models and XML formats for data interchange such as mzData are currently emerging. Using these standards as a show case, we present an infrastructure to support the use of these data standards and the process of getting there.

Computational representation of enzyme function should include the structural elements of enzymes which deliver catalytic ability. This is especially important in mechanistically diverse enzyme superfamilies, whose members catalyze different overall reactions. In such superfamilies, evolutionarily conserved elements of structure can be correlated with only conserved aspects of function. The representation of enzyme function in the Structure-Function Linkage Database, in particular the specific structure-function relationships, at multiple levels of evolutionary conservation, aids in the annotation of enzyme function and in designing enzyme engineering experiments.

Classical enzyme kinetics, as developed in the 20th century, had as a primary objective the elucidation of the 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 takes into account the systemic context in which each enzyme is found. In this paper we present the generalized reversible Hill equation as a universal rate equation for systems biology, in that it takes into account (i) the kinetic and regulatory properties of enzyme-catalysed reactions, (ii) the reversibility and thermodynamic consistency of all reactions, and (iii) the modification of enzyme activity by allosteric effectors. Setting the Hill coefficient to one yields a universal equation that can successfully mimic the behaviour of various detailed non-cooperative mechanistic models. Subsequently, it is shown that the bisubstrate Hill equation can account for substrate-modifier saturation, in agreement with experimental data from Bacillus stearothermophilus pyruvate kinase. In contrast, the classical Monod–Wyman–Changeux (MWC) equation cannot account for this effect. The proposed reversible Hill equations are all independent of underlying enzyme mechanism, are of great use in computational models and should lay the groundwork for a “new” enzyme kinetics for systems biology.

SABIO-RK is a database designed to store and offer access to information about biochemical reactions and their kinetics in a comprehensive and standardized manner. It integrates information from several sources to form a backbone of information necessary to include information about the kinetics of biochemical reactions. The kinetic data itself is primarily extracted from literature along with descriptions of the experimental conditions under which they were determined. This process is supported by the use of a web-based user interface which complies with most of the recommendations of the STRENDA committee for reporting on the results of enzyme/reaction kinetics. In this paper we describe the main characteristics of the SABIO-RK and its search and input interfaces. Availability: sabio.villa-bosch.de/sabiork

Proteases are enzymes catalysing the hydrolysis of peptides or proteins. They are key players in a wide range of biological processes such as the release of peptide hormones, nutrient acquisition, cell growth, differentiation, antigen processing and protein turnover, in all living organisms. Furthermore it is becoming more and more obvious that the abnormal functioning of some proteases may lie behind several types of diseases, including inflammation, cancer and Alzheimer's disease. Therefore proteases are attracting an increasing interest.

The MEROPS database, which is specialized in proteases, lists 555 known and putative genes encoding proteases in Homo sapiens (31st of August 2006). The number of proteins acting as proteases in the human organism may even be much higher, since proteins can develop proteolytic activities although they are not assigned as proteases. The protein disulfide isomerase A3 (PDIA3, primary accession number: P30101), which main function is protein folding, is an example for the latter case since Kito, Urade and coworkers published convincing data about a protease activity of PDIA3.

The BRENDA enzyme information system is the largest publicly available enzyme information system worldwide. The major part of its content is manually extracted from primary literature. It is not restricted to specific groups of enzymes, but includes information on all identified enzymes irrespective of the source of the enzyme. The range of data encompasses functional, structural, sequence, localization, disease-related, isolation, stability information on enzyme and ligand-related data. Each single entry is linked to the enzyme source and to a literature reference. Recently the data repository was complemented by text mining data which is stored in AMENDA and FRENDA. A genome browser, membrane protein prediction and full text search capacities were added. The newly implemented web service provides instant access to the data for programmers via a SOAP interface. The BRENDA data can be downloaded in the form of a text file from the beginning of 2007.

In this contribution we report on the JWS Online project and the progress that has been made since the first ESCEC meeting. Whilst maintaining the same user interface, we have completely redesigned the server part of JWS Online, now a) using webMathematica as the interface between the HTML pages and the Mathematica Kernel and b) storing all models as Mathematica packages, and c) using a PostgresQL database to store a full description of each model. In the last few years a number of new initiatives have started, of which some fulfil comparable roles to JWS Online and with some of which we collaborate. Here we compare JWS Online to these initiatives focusing on the three aims of JWS Online: 1) to be a repository for curated kinetic models of biological systems, 2) to be an easy to use simulator that can be accessed over the internet, 3) to help in the reviewing of manuscripts containing kinetic models.

The kinetic modelling of biochemical pathways requires a consistent set of enzymatic kinetic parameters. We report results from software development to assist the user in systems biology, allowing the retrieval of heterogeneous protein sequence, structural and kinetic data. For the simulation of biological networks, missing enzymatic kinetic parameters can be calculated using a similarity analysis of the enzymes’ molecular interaction fields. The quantitative PIPSA (qPIPSA) methodology relates changes in the molecular interaction fields of the enzymes with variations in the enzymatic rate constants or binding affinities. As an illustrative example, this approach is used to predict kinetic parameters for glucokinases from Escherichia coli based on experimental values for a test set of enzymes. The best correlation of the electrostatic potentials with kinetic parameters is found for the open form of the glucokinases. The similarity analysis was extended to a large set of glucokinases from various organisms.

The present paper will focus on the characterization of enzymes from two different types of family, Short Chain Dehydrogenases/Reductases and Matrixins. The former family includes over 3000 enzymes, and I have worked mainly with different allelic variants of alcohol dehydrogenase (ADH) from the fruit fly Drosophila melanogaster and the ADH in Drosophila lebanonensis. To date, approximately 25 matrix metalloproteinases are known in humans. I will focus here on both similarities and differences in problems regarding the standardization of assay conditions and parameters that I have experienced during my work with these two different enzyme systems.