An energetic profile of Corynebacterium glutamicum underpinned by measured biomass yield on ATP.

The supply and usage of energetic cofactors in metabolism is a central concern for systems metabolic engineering, particularly in case of energy intensive products. One of the most important parameters for systems wide balancing of energetic cofactors is the ATP requirement for biomass formation YATP/Biomass. Despite its fundamental importance, YATP/Biomass values for non-fermentative organisms are still rough estimates deduced from theoretical considerations. For the first time, we present an approach for the experimental determination of YATP/Biomass using comparative 13C metabolic flux analysis (13C MFA) of a wild type strain and an ATP synthase knockout mutant. We show that the energetic profile of a cell can then be deduced from a genome wide stoichiometric model and experimental maintenance data. Particularly, the contributions of substrate level phosphorylation (SLP) and electron transport phosphorylation (ETP) to ATP generation become available which enables the overall energetic efficiency of a cell to be characterized. As a model organism, the industrial platform organism Corynebacterium glutamicum is used. C. glutamicum uses a respiratory type of energy metabolism, implying that ATP can be synthesized either by SLP or by ETP with the membrane-bound F1FO-ATP synthase using the proton motive force (pmf) as driving force. The presence of two terminal oxidases, which differ in their proton translocation efficiency by a factor of three, further complicates energy balancing for this organism. By integration of experimental data and network models, we show that in the wild type SLP and ETP contribute equally to ATP generation. Thus, the role of ETP in respiring bacteria may have been overrated in the past. Remarkably, in the genome wide setting 65% of the pmf is actually not used for ATP synthesis. However, it turns out that, compared to other organisms C. glutamicum still uses its energy budget rather efficiently.

Tailoring the properties of (catalytically)-active inclusion bodies.

BACKGROUND: Immobilization is an appropriate tool to ease the handling and recycling of enzymes in biocatalytic processes and to increase their stability. Most of the established immobilization methods require case-to-case optimization, which is laborious and time-consuming. Often, (chromatographic) enzyme purification is required and stable immobilization usually includes additional cross-linking or adsorption steps. We have previously shown in a few case studies that the molecular biological fusion of an aggregation-inducing tag to a target protein induces the intracellular formation of protein aggregates, so called inclusion bodies (IBs), which to a certain degree retain their (catalytic) function. This enables the combination of protein production and immobilization in one step. Hence, those biologically-produced immobilizates were named catalytically-active inclusion bodies (CatIBs) or, in case of proteins without catalytic activity, functional IBs (FIBs). While this strategy has been proven successful, the efficiency, the potential for optimization and important CatIB/FIB properties like yield, activity and morphology have not been investigated systematically. RESULTS: We here evaluated a CatIB/FIB toolbox of different enzymes and proteins. Different optimization strategies, like linker deletion, C- versus N-terminal fusion and the fusion of alternative aggregation-inducing tags were evaluated. The obtained CatIBs/FIBs varied with respect to formation efficiency, yield, composition and residual activity, which could be correlated to differences in their morphology; as revealed by (electron) microscopy. Last but not least, we demonstrate that the CatIB/FIB formation efficiency appears to be correlated to the solvent-accessible hydrophobic surface area of the target protein, providing a structure-based rationale for our strategy and opening up the possibility to predict its efficiency for any given target protein. CONCLUSION: We here provide evidence for the general applicability, predictability and flexibility of the CatIB/FIB immobilization strategy, highlighting the application potential of CatIB-based enzyme immobilizates for synthetic chemistry, biocatalysis and industry.

Visual exploration of isotope labeling networks in 3D.

Isotope labeling networks (ILNs) are graphs explaining the flow of isotope labeled molecules in a metabolic network. Moreover, they are the structural backbone of metabolic flux analysis (MFA) by isotopic tracers which has been established as a standard experimental tool in fluxomics. To configure an isotope labeling experiment (ILE) for MFA, the structure of the corresponding ILN must be understood to a certain extent even by a practitioner. Graph algorithms help to analyze the network structure but produce rather abstract results. Here, the major obstruction is the high dimension of these networks and the large number of network components which, consequently, are hard to figure out manually. At the interface between theory and experiment, the three-dimensional interactive visualization tool CumoVis has been developed for exploring the network structure in a step by step manner. Navigation and orientation within ILNs are supported by exploiting the natural 3D structure of an underlying metabolite network with stacked labeled particles on top of each metabolite node. Network exploration is facilitated by rotating, zooming, forward and backward path tracing and, most important, network component reduction. All features of CumoVis are explained with an educational example and a realistic network describing carbon flow in the citric acid cycle.

Unravelling the regulatory structure of biochemical networks using stimulus response experiments and large-scale model selection.

To unravel the complex in vivo regulatory interdependences of biochemical networks, experiments with the living organism are absolutely necessary. Stimulus response experiments (SREs) have become increasingly popular in recent years. The response of metabolite concentrations from all major parts of the central metabolism is monitored over time by modem analytical methods, producing several thousand data points. SREs are applied to determine enzyme kinetic parameters and to find unknown enzyme regulatory mechanisms. Owing to the complex regulatory structure of metabolic networks and the amount of measured data, the evaluation of an SRE has to be extensively supported by modelling. If the enzyme regulatory mechanisms are part of the investigation, a large number of models with different enzyme kinetics have to be tested for their ability to reproduce the observed behaviour. In this contribution, a systematic model-building process for data-driven exploratory modelling is introduced with the aim of discovering essential features of the biological system. The process is based on data pre-processing, correlation-based hypothesis generation, automatic model family generation, large-scale model selection and statistical analysis of the best-fitting models followed by an extraction of common features. It is illustrated by the example of the aromatic amino acid synthesis pathway in Escherichia coli.

Serial flux mapping of Corynebacterium glutamicum during fed-batch L-lysine production using the sensor reactor approach.

Using our recently developed sensor reactor approach, lysine-producing, nongrowing Corynebacterium glutamicum MH20-22B cells were subjected to serial (13)C-labeling experiments for flux analysis during the leucine-limited fed-batch production phase in a 300-L bioreactor. Based on two-dimensional (2D) nuclear magnetic resonance (NMR) measurements of (13)C-labeling patterns of cytoplasmic free metabolites, metabolic flux distributions in the central metabolism were successfully determined. Focusing on the highly concentrated metabolite L-glutamate, the working hypothesis was validated that the equilibration of labeling patterns in intracellular pools was much faster (up to 9.45 min) than the labeling period (3 h) used in the experiments. Analysis of anaplerotic reactions revealed that highly selective lysine production was accompanied by a significant reduction of decarboxylating reactions from 10 mol% to only 2 mol%, whereas PEP/pyruvate-carboxylating fluxes remained constant at about 40 mol% of consumed glucose. These results support the conclusion that an optimized C. glutamicum L-lysine producer should possess increased PEP carboxylase and/or pyruvate carboxylase activity combined with downregulated, decarboxylating fluxes consuming oxaloacetate/malate. The findings also illustrate the usefulness of the sensor reactor approach in the study of industrial fermentations.

13C metabolic flux analysis.

Metabolic flux analysis using 13C-labeled substrates has become an important tool in metabolic engineering. It allows the detailed quantification of all intracellular fluxes in the central metabolism of a microorganism. The method has strongly evolved in recent years by the introduction of new experimental procedures, measurement techniques, and mathematical data evaluation methods. Many of these improvements require advanced skills in the application of nuclear magnetic resonance and mass spectrometry techniques on the one hand and computational and statistical experience on the other hand. This minireview summarizes these recent developments and sketches the major practical problems. An outlook to possible future developments concludes the text.

A universal framework for 13C metabolic flux analysis.

A general methodology is presented for the modeling, simulation, design, evaluation, and statistical analysis of (13)C-labeling experiments for metabolic flux analysis. The universal software framework 13C-FLUX was implemented to support all steps of this process. Guided by the example of anaplerotic flux determination in Corynebacterium glutamicum, the technical details of the model setup, experimental design, and data evaluation are discussed. It is shown how the network structure, the input substrate composition, the assumptions about fluxes, and the measurement configuration are specified within 13C-FLUX. Based on the network model, different experimental designs are computed depending on the goal of the investigations. Finally, a specific experiment is evaluated and the various statistical methods used to analyze the results are briefly explained. The appendix gives some details about the software implementation and availability.

Metabolic isotopomer labeling systems. Part I: global dynamic behavior.

In the last few years metabolic flux analysis (MFA) using carbon labeling experiments (CLE) has become a major diagnostic tool in metabolic engineering. The mathematical centerpiece of MFA is the solution of isotopomer labeling systems (ILS). An ILS is a high-dimensional nonlinear differential equation system that describes the distribution of isotopomers over a metabolic network during a carbon labeling experiment. This contribution presents a global analysis of the dynamic behavior of general ILSs. It is proven that an ILS is globally stable under very weak conditions that are always satisfied in practice. In particular it is shown that in some sense ILSs are a nonlinear extension to the classical theory of compartmental systems. The central stability condition for compartmental systems, i.e., the non-existence of traps in linear compartmental networks, is also the major stability condition for ILSs. As an important side result of the proof, it is shown that ILSs can be transformed to a cascade of linear systems with time-dependent inhomogeneous terms. This cascade structure has considerable consequences for the development of efficient numerical algorithms for the solution of ILSs and thus for MFA.

In vivo quantification of parallel and bidirectional fluxes in the anaplerosis of Corynebacterium glutamicum.

The C(3)-C(4) metabolite interconversion at the anaplerotic node in many microorganisms involves a complex set of reactions. C(3) carboxylation to oxaloacetate can originate from phosphoenolpyruvate and pyruvate, and at the same time multiple C(4)-decarboxylating enzymes may be present. The functions of such parallel reactions are not yet fully understood. Using a (13)C NMR-based strategy, we here quantify the individual fluxes at the anaplerotic node of Corynebacterium glutamicum, which is an example of a bacterium possessing multiple carboxylation and decarboxylation reactions. C. glutamicum was grown with a (13)C-labeled glucose isotopomer mixture as the main carbon source and (13)C-labeled lactate as a cosubstrate. 58 isotopomers as well as 15 positional labels of biomass compounds were quantified. Applying a generally applicable mathematical model to include metabolite mass and carbon labeling balances, it is shown that pyruvate carboxylase contributed 91 +/- 7% to C(3) carboxylation. The total in vivo carboxylation rate of 1.28 +/- 0.14 mmol/g dry weight/h exceeds the demand of carboxylated metabolites for biosyntheses 3-fold. Excess oxaloacetate was recycled to phosphoenolpyruvate by phosphoenolpyruvate carboxykinase. This shows that the reactions at the anaplerotic node might serve additional purposes other than only providing C(4) metabolites for biosynthesis.

Determination of full 13C isotopomer distributions for metabolic flux analysis using heteronuclear spin echo difference NMR spectroscopy.

13C-isotopomer labeling experiments play an increasingly important role in the analysis of intracellular metabolic fluxes for genetic engineering purposes. 13C NMR spectroscopy is a key technique in the experimental determination of isotopomer distributions. However, only subsets of isotopomers can be quantitated using this technique due to redundancies in the scalar coupling patterns and due to invisibility of the 12C isotope in NMR. Therefore, we developed and describe in this paper a 1H NMR spectroscopy method that allows to determine the complete isotopomer distribution in metabolites having a backbone consisting of up to at least four carbons. The proposed pulse sequences employ up to three alternately applied frequency-selective inversion pulses in the 13C channel. In a first application study, the complete isotopomer distribution of aspartate isolated from [1-13C]ethanol-grown Ashbya gossypii was determined. A tentative model of the central metabolism of this organism was constructed and used for metabolic flux analysis. The aspartate isotopomer NMR data played a key role in the successful determination of the flux through the peroxisomal glyoxylate pathway. The new NMR method can be highly instrumental in generating the data upon which isotopomer labeling experiments for flux analysis, that are becoming increasingly important, are based.

Bidirectional reaction steps in metabolic networks: III. Explicit solution and analysis of isotopomer labeling systems.

The last few years have brought tremendous progress in experimental methods for metabolic flux determination by carbon-labeling experiments. A significant enlargement of the available measurement data set has been achieved, especially when isotopomer fractions within intracellular metabolite pools are quantitated. This information can be used to improve the statistical quality of flux estimates. Furthermore, several assumptions on bidirectional intracellular reaction steps that were hitherto indispensable may now become obsolete. To make full use of the complete measurement information a general mathematical model for isotopomer systems is established in this contribution. Then, by introducing the important new concept of cumomers and cumomer fractions, it is shown that the arising nonlinear isotopomer balance equations can be solved analytically in all cases. In particular, the solution of the metabolite flux balances and the positional carbon-labeling balances presented in part I of this series turn out to be just the first two steps of the general solution procedure for isotopomer balances. A detailed analysis of the isotopomer network structure then opens up new insights into the intrinsic structure of isotopomer systems. In particular, it turns out that isotopomer systems are not as complex as they appear at first glance. This enables some far-reaching conclusions to be drawn on the information potential of isotopomer experiments with respect to flux identification. Finally, some illustrative examples are examined to show that an information increase is not guaranteed when isotopomer measurements are used in addition to positional enrichment data.

Bidirectional reaction steps in metabolic networks: IV. Optimal design of isotopomer labeling experiments.

This article generalizes the statistical tools for the evaluation of carbon-labeling experiments that have been developed for the case of positional enrichment systems in part II of this series to the general case of isotopomer systems. For this purpose, a new generalized measurement equation is introduced that can describe all kinds of measured data, like positional enrichments, relative (13)C nuclear magnetic resonance ((13)C NMR) multiplet intensities, or mass isotopomer fractions produced with mass spectroscopy (MS) instruments. Then, to facilitate the specification of the various measurement procedures available, a new flexible textual notation is introduced from which the complicated generalized measurement equations are generated automatically. Based on these measurement equations, a statistically optimal flux estimator is established and parameter covariance matrices for the flux estimation are computed. Having implemented these tools, different kinds of labeling experiments can be compared by using statistical quality measures. A general framework for the optimal design of carbon-labeling experiments is established on the basis of this method. As an example it is applied to the Corynebacterium network from part II extended by various NMR and MS measurements. In particular, the positional enrichment, multiplet, or mass isotopomer measurements with the greatest information content for flux estimation are computed (measurement design) and various differently labeled input substrates are compared with respect to flux estimation (input design). It is examined in detail how the measurement procedure influences the estimation quality of specific fluxes like the pentose phosphate pathway influx.

Metabolic state of Zymomonas mobilis in glucose-, fructose-, and xylose-fed continuous cultures as analysed by 13C- and 31P-NMR spectroscopy.

The reasons for the well-known significantly different behaviour of the anaerobic, gram-negative, ethanologenic bacterium Zymomonas mobilis during growth on fructose (i.e. decreased growth and ethanol yields, increased by-product formation) as compared to that on its second natural substrate, glucose, have remained unexplained. A xylose-fermenting recombinant strain of Z. mobilis that was recently constructed in our laboratory also unexpectedly displayed an increased formation of by-products and a strongly reduced growth rate as compared to the parent strain. Therefore, a comprehensive study employing recently developed NMR-based methods for the in vivo analysis of intracellular phosphorylated pool sizes and metabolic fluxes was undertaken to enable a global characterization of the intracellular metabolic state of Z. mobilis during growth on 13C-labelled glucose, fructose and xylose in defined continuous cultures. The 13C-NMR flux analysis indicated that ribose 5-phosphate is synthesized via the nonoxidative pentose phosphate pathway in Z. mobilis, and it identified a metabolic bottleneck in the recombinant xylose-fermenting Z. mobilis strain at the level of heterologous xylulokinase. The 31P-NMR analyses revealed a global alteration of the levels of intracellular phosphorylated metabolites during growth on fructose as compared to that on glucose. The results suggest that this is primarily caused by an elevated concentration of intracellular fructose 6-phosphate.

13C tracer experiments and metabolite balancing for metabolic flux analysis: comparing two approaches

Conventional metabolic flux analysis uses the information gained from determination of measurable fluxes and a steady-state assumption for intracellular metabolites to calculate the metabolic fluxes in a given metabolic network. The determination of intracellular fluxes depends heavily on the correctness of the assumed stoichiometry including the presence of all reactions with a noticeable impact on the model metabolite balances. Determination of fluxes in complex metabolic networks often requires the inclusion of NADH and NADPH balances, which are subject to controversial debate. Transhydrogenation reactions that transfer reduction equivalents from NADH to NADPH or vice versa can usually not be included in the stoichiometric model, because they result in singularities in the stoichiometric matrix. However, it is the NADPH balance that, to a large extent, determines the calculated flux through the pentose phosphate pathway. Hence, wrong assumptions on the presence or activity of transhydrogenation reactions will result in wrong estimations of the intracellular flux distribution. Using 13C tracer experiments and NMR analysis, flux analysis can be performed on the basis of only well established stoichiometric equations and measurements of the labeling state of intracellular metabolites. Neither NADH/NADPH balancing nor assumptions on energy yields need to be included to determine the intracellular fluxes. Because metabolite balancing methods and the use of 13C labeling measurements are two different approaches to the determination of intracellular fluxes, both methods can be used to verify each other or to discuss the origin and significance of deviations in the results. Flux analysis based entirely on metabolite balancing and flux analysis, including labeling information, have been performed independently for a wild-type strain of Aspergillus oryzae producing alpha-amylase. Two different nitrogen sources, NH4+ and NO3-, have been used to investigate the influence of the NADPH requirements on the intracellular flux distribution. The two different approaches to the calculation of fluxes are compared and deviations in the results are discussed. Copyright 1998 John Wiley & Sons, Inc.

Bidirectional reaction steps in metabolic networks: I. Modeling and simulation of carbon isotope labeling experiments.

The extension of metabolite balancing with carbon labeling experiments, as described by Marx et al. (Biotechnol. Bioeng. 49: 11-29), results in a much more detailed stationary metabolic flux analysis. As opposed to basic metabolite flux balancing alone, this method enables both flux directions of bidirectional reaction steps to be quantitated. However, the mathematical treatment of carbon labeling systems is much more complicated, because it requires the solution of numerous balance equations that are bilinear with respect to fluxes and fractional labeling. In this study, a universal modeling framework is presented for describing the metabolite and carbon atom flux in a metabolic network. Bidirectional reaction steps are extensively treated and their impact on the system's labeling state is investigated. Various kinds of modeling assumptions, as usually made for metabolic fluxes, are expressed by linear constraint equations. A numerical algorithm for the solution of the resulting linear constrained set of nonlinear equations is developed. The numerical stability problems caused by large bidirectional fluxes are solved by a specially developed transformation method. Finally, the simulation of carbon labeling experiments is facilitated by a flexible software tool for network synthesis. An illustrative simulation study on flux identifiability from available flux and labeling measurements in the cyclic pentose phosphate pathway of a recombinant strain of Zymomonas mobilis concludes this contribution.

Bidirectional reaction steps in metabolic networks: II. Flux estimation and statistical analysis.

Metabolic carbon labelling experiments enable a large amount of extracellular fluxes and intracellular carbon isotope enrichments to be measured. Since the relation between the measured quantities and the unknown intracellular metabolic fluxes is given by bilinear balance equations, flux determination from this data set requires the numerical solution of a nonlinear inverse problem. To this end, a general algorithm for flux estimation from metabolic carbon labelling experiments based on the least squares approach is developed in this contribution and complemented by appropriate tools for statistical analysis. The linearization technique usually applied for the computation of nonlinear confidence regions is shown to be inappropriate in the case of large exchange fluxes. For this reason a sophisticated compactification transformation technique for nonlinear statistical analysis is developed. Statistical analysis is then performed by computing appropriate statistical quality measures like output sensitivities, parameter sensitivities and the parameter covariance matrix. This allows one to determine the order of magnitude of exchange fluxes in most practical situations. An application study with a large data set from lysine-producing Corynebacterium glutamicum demonstrates the power and limitations of the carbon-labelling technique. It is shown that all intracellular fluxes in central metabolism can be quantitated without assumptions on intracellular energy yields. At the same time several exchange fluxes are determined which is invaluable information for metabolic engineering.

Experimental design for the identification of macrokinetic models and model discrimination.

An experimental design method for the identification of macrokinetic models was developed applying an extended D-optimal design criterion. The D-optimal design criterion was modified to consider variable measurement variances as well as multivariate macrokinetic models. The macrokinetics of formate dehydrogenase (FDH) production with Candida boidinii were thus identified within 10 steady state experiments in a labscale continuous stirred tank reactor (10 model parameters). Closed loop control (nutristat) was applied to set-up the operating states suggested by this experimental design method. After each set of steady state experiments the quality of macrokinetic parameters was characterized statistically. For model discrimination a parameter discrimination algorithm based on entropy formulations was adapted. Again a multivariate criterion considering variable measurement variances was developed. This discrimination algorithm was applied to discriminate the macrokinetic model of FDH production with Candida boidinii out of 10 different macrokinetic approaches. An unequivocal discrimination result could be obtained calculating model specific probabilities. These were compared with commonly used sum of squares values. (c) 1997 John Wiley & Sons, Inc. Biotechnol Bioeng 56: 564-576, 1997.

Determination of the fluxes in the central metabolism of Corynebacterium glutamicum by nuclear magnetic resonance spectroscopy combined with metabolite balancing.

To determine the in vivo fluxes of the central metabolism we have developed a comprehensive approach exclusively based on the fundamental enzyme reactions known to be present, the fate of the carbon atoms of individual reactions, and the metabolite balance of the culture. No information on the energy balance is required, nor information on enzyme activities, or the directionalities of reactions. Our approach combines the power of (1)H-detected (13)C nuclear magnetic resonance spectroscopy to follow individual carbons with the simplicity of establishing carbon balances of bacterial cultures. We grew a lysine-producing strain of Corynebacterium glutamicum to the metabolic and isotopic steady state with [1-(13)C]glucose and determined the fractional enrichments in 27 carbon atoms of 11 amino acids isolated from the cell. Since precursor metabolites of the central metabolism are incorporated in an exactly defined manner in the carbon skeleton of amino acids, the fractional enrichments in carbons of precursor metabolites (oxaloacetate, glyceraldehyde 3-phosphate, erythrose 4-phosphate, etc.) became directly accessible. A concise and generally applicable mathematical model was established using matrix calculus to express all metabolite mass and carbon labeling balances. An appropriate all-purpose software for the iterative solution of the equations is supplied. Applying this comprehensive methodology to C. glutamicum, all major fluxes within the central metabolism were determined. The result is that the flux through the pentose phosphate pathway is 66.4% (relative to the glucose input flux of 1.49 mmol/g dry weight h), that of entry into the tricarboxylic acid cycle 62.2%, and the contribution of the succinylase pathway of lysine synthesis 13.7%. Due to the large amount and high quality of measured data in vivo exchange reactions could also be quantitated with particularly high exchange rates within the pentose phosphate pathway for the ribose 5-phosphate transketolase reaction. Moreover, the total net flux of the anaplerotic reactions was quantitated as 38.0%. Most importantly, we found that in vivo one component within these anaplerotic reactions is a back flux from the carbon 4 units of the tricarboxylic acid cycle to the carbon 3 units of glycolysis of 30.6%. (c) 1996 John Wiley & Sons, Inc.

In vivo stationary flux analysis by 13C labeling experiments.

Stationary flux analysis is an invaluable tool for metabolic engineering. In the last years the metabolite balancing technique has become well established in the bioengineering community. On the other hand metabolic tracer experiments using 13C isotopes have long been used for intracellular flux determination. Only recently have both techniques been fully combined to form a considerably more powerful flux analysis method. This paper concentrates on modeling and data analysis for the evaluation of such stationary 13C labeling experiments. After reviewing recent experimental developments, the basic equations for modeling carbon labeling in metabolic systems, i.e. metabolite, carbon label and isotopomer balances, are introduced and discussed in some detail. Then the basics of flux estimation from measured extracellular fluxes combined with carbon labeling data are presented and, finally, this method is illustrated by using an example from C. glutamicum. The main emphasis is on the investigation of the extra information that can be obtained with tracer experiments compared with the metabolite balancing technique alone. As a principal result it is shown that the combined flux analysis method can dispense with some rather doubtful assumptions on energy balancing and that the forward and backward flux rates of bidirectional reaction steps can be simultaneously determined in certain situations. Finally, it is demonstrated that the variant of fractional isotopomer measurement is even more powerful than fractional labeling measurement but requires much higher numerical effort to solve the balance equations.

Object-oriented programming for the biosciences.

The development of software systems for the biosciences is always closely connected to experimental practice. Programs must be able to handle the inherent complexity and heterogeneous structure of biological systems in combination with the measuring equipment. Moreover, a high degree of flexibility is required to treat rapidly changing experimental conditions. Object-oriented methodology seems to be well suited for this purpose. It enables an evolutionary approach to software development that still maintains a high degree of modularity. This paper presents experience with object-oriented technology gathered during several years of programming in the fields of bioprocess development and metabolic engineering. It concentrates on the aspects of experimental support, data analysis, interaction and visualization. Several examples are presented and discussed in the general context of the experimental cycle of knowledge acquisition, thus pointing out the benefits and problems of object-oriented technology in the specific application field of the biosciences. Finally, some strategies for future development are described.