Abstract
The increasing number of completely sequenced genomes provides a parts list of all proteins encoded by an organism. However, analysis of the complete repertoire of proteins, the proteome, is extremely complex and technically very challenging. Not the least because important determinants for protein activity will be scored at various levels like modifications, interactions and changes in production. The last couple of years have seen some extraordinary improvements in our ability to globally analyse protein features. Some of the relevant technologies for proteome analysis, like two-dimensional gel electrophoresis and mass spectrometry, will be covered, and their use for the analysis of proteomes exemplified. However, it is argued that full level understanding of living systems will require further improvement in the way we study proteins if truly system-wide models of cellular system are to be constructed.

Abstract
The function of unknown genes can be determined from the effect that mutation or deletion of the gene of interest has on the phenotype of the organism. Since in many cases this approach to analysis of gene function, often called functional genomics, produces no, or only very subtle changes in phenotype it requires comprehensive methods for their analysis. Magnetic resonance imaging and spectroscopy methods are described which allow comprehensive analysis of the morphological and metabolic consequences respectively of gene modification. The information content of magnetic resonance imaging has been further increased with the development of novel contrast agents targeted at specific molecular targets, so called 'molecular imaging'. These can report on aspects of the underlying biochemistry and physiology of the tissue. In addition gene reporter constructs are being developed which will allow imaging of gene expression.

Abstract
In the post genomic era it has become important to develop tools that rapidly enable phenotypic characterization of an organism for which the genome has been sequenced. An often applied approach is to reconstruct the metabolic network that may operate in the sequenced organism and subsequently analyze the metabolic network. When a specific network has been reconstructed it represents all possible phenotypes a certain organism may express, and it is therefore important to analyze in further details the activity of the individual branches of the network, or in other words quantify the fluxes that are carried by the individual pathways that may operate in the studied organism. Methods for rapid quantification of metabolic fluxes have therefore become important for transferring knowledge from genotype to phenotype. Such methods are in broad terms often referred to as Metabolic Flux Analysis, and these methods are, besides their above mentioned role for analysis of all possible phenotypes a cell may express, instrumental for the understanding of what lies beneath a certain phenotype. Hereby the methods also play a very important role for the rational design of organisms with improved properties, e.g. development of improved microbial strains used for fermentation processes. Tools for Metabolic Flux Analysis are therefore also at the core of Metabolic Engineering. Basically Metabolic Flux Analysis provides the mathematical framework for calculating fluxes through the biochemical pathways of a whole network, resulting in a flux map that may be compared with others obtained for different growth conditions and/or different strains.

Abstract
The development, maintenance, and environmental responses of organisms are accompanied by the selective expression of genes. The coordinated response of genes is achieved via the action of genes on each other, often referred to as the gene regulatory network. Gaining an understanding of such networks may have a profound impact on our ability to engineer metabolism. With modern microarray technology, one can simultaneously measure the expression levels of thousands of genes. This information has long been expected to provide all the necessary means to unravel the interactions between genes. Several methods have already been developed for this purpose, but in no case was experimental design taken into consideration, and the data upon which these methods were applied was never obtained with inferring gene networks in mind. In contrast, the method described here is based on a well-defined experimental setup designed with the sole purpose of quantifying how much the expression of one gene affects other genes. This method makes use of results from Metabolic Control Analysis, namely co-control analysis and the concept of regulatory strength, and is able to use relative expression levels as measured with microarrays directly. Examples of application of the method are presented using in silico experiments.

Abstract
MetaCyc is a metabolic-pathway database (DB) that describes 484 pathways and 1470 enzymes occurring in 167 organisms. As a general reference source on metabolic pathways, MetaCyc serves several purposes in metabolic engineering. It provides an encyclopedic listing of enzymes and their properties to allow a metabolic engineer to find an enzyme whose characteristics solve an engineering problem at hand. It also serves as a reference pathway DB for predicting the metabolic pathway complement of an organism from its genome. Elucidating the pathway network of a new organism is the logical first step in metabolic engineering. MetaCyc is a review-level DB in that a given entry in MetaCyc often integrates information from multiple literature sources. The pathways in MetaCyc were determined experimentally, and are labeled with the species in which they are known to occur, based on literature references examined to date. MetaCyc contains extensive commentary and literature citations. MetaCyc is available through the WWW at MetaCyc.org, and is available for local installation as a binary program for Linux, the PC, and the Sun workstation, and as a set of flat files. Contact metacyc-info@ai.sri.com for information on obtaining a local copy of MetaCyc.

Abstract
The ability to change enzyme activities by changing the expression of the corresponding gene(s) is an important tool for Metabolic Control Analysis, Metabolic Engineering and Metabolic Optimization. We here review a selection of cases where promoters have been used over the years to change enzyme activities and look at some of the advantages and disadvantages of the different methods employed. We present some of the new promising tools that have become available more recently, particularly the synthetic promoter libraries (SPL) technology that allows for fine tuning of the steady state expression of several genes individually and which appears to be universally applicable.

Abstract
This chapter gives an overview of Metabolic Pathway Analysis and its historical roots. The basic concepts of the field are outlined and their relevance for biotechnology are discussed. In particular, the concept of elementary flux mode is explained and compared with that of extreme pathway. It is shown that elementary-mode analysis is a useful tool for better understanding the complex architecture of metabolism and for determining those metabolic routes on which the molar conversion yield of a substrate-product pair under consideration is maximum. The presentation is illustrated by a simple example from monosaccharide metabolism and a system describing lysine production in Escherichia coli, which is kept relatively simple for didactic reasons. The latter system gives rise to 14 elementary modes, four of which produce lysine with different molar yields. Furthermore, special problems arising in the modeling of large networks are discussed. An overview of the literature on theoretical and applied problems of metabolic pathway analysis is given. An outlook on future prospects in the field concludes the chapter.

Abstract
It is now possible to reconstruct genome-scale models of cellular metabolism. Our research group has reconstructed such models for four organisms: Escherichia coli, Haemophilus influenzae, Helicobacter pylori, and Saccharomyces cerevisiae. Constraint-based analysis methods, such as Flux Balance Analysis, Extreme Pathway Analysis, and Phenotypic Phase Plane Analysis, allow for calculations of various cellular phenotypes. Predictions can be made about the essentiality or non-essentiality of a gene, about the optimal growth rate of an organism, and about substrate uptake and by-product secretion rates. These constraint-based in silico models of metabolism provide a framework upon which other cellular functions can also be modeled, beginning with transcriptional regulation. A constraints-based modeling approach has the advantages of being data-driven and scalable to whole genome-scale analyses. The capabilities of constraint-based modeling are illustrated through a detailed description of what has been learned from studying the in silico models of the four organisms mentioned above.

Abstract
Three types of measurement data are closely related to metabolic activity: (i) metabolic fluxes which can be determined by ¹³C labeling experiments; (ii) intracellular metabolite concentrations which can be measured by HPLC, enzymatic assays or LC-MS/MS; (iii) enzyme activities which can be determined from cell extracts. All these data can be measured in different stationary states of one or several strains of the microorganism under investigation. Clearly, the sole knowledge of fluxes is not sufficient to understand the regulation of the metabolic system. Up to now, only few attempts have been made to use the combined information on fluxes, pool sizes and enzyme activities as a data base for building models of metabolic regulation. This integrative approach is a big challenge for systems biology. A general modeling strategy is presented that enables regulatory models to be con-structed from the available stationary data. The corresponding mathematical framework is explained and the arising methodological problems are illustrated. The method is then exemplified by the anaplerotic pathways of Corynebacterium glutamicum where several stationary states of different strains have been completely characterized. The resulting regulatory model has good predictive properties with respect to genetic modifications of the anaplerotic enzyme levels.

Abstract
Metabolic modeling aims at the quantitative understanding of complex intracellular reaction networks. The most important requirement for this task is the validity of a model, i.e. its ability to predict the intracellular dynamic behavior with reasonable precision. For validation purposes a large amount of quantitative intracellular data about metabolite concentrations, metabolic fluxes, reaction kinetics and enzyme activities is becoming available. However, this contribution shows that the concept of model validity involves some intrinsic problems. In this context mathematical methods are required that assist in iterative model refinement and simplification, sensitivity analysis, parameter estimation, prediction computations, model selection and experimental design. In the application of model validation methods several misinterpretations, misconceptions and pitfalls must be avoided. In this contribution the available methods are critically discussed and demonstrated by using two simple illustrative networks. It is shown that the idea of a "true" model is a misconception, that model simplification does not always make sense, that well-determined reaction parameters are not in general a good criterion for model selection and that models with precise predictions need not be valid. Finally, an outlook is given concerning the problem of large-scale model validation and the tools that must be developed for that purpose.

Abstract
E. coli is an important organism not only for experimental study, but also for the ambitious goal of constructing a mathematical model of the entire metabolic and regulatory mechanisms. Building such a model is an arduous process characterized by several steps:

(1) determining the function of the proteins and genes and their static interactions (metabolic and gene regulatory);
(2) constructing the mathematical relations of each interaction so that the corresponding kinetics is properly represented;
(3) calibrating the model against experimental kinetics data to determine rate constants and verifying the kinetic rate laws; and
(4) integrating the pieces into a coherent model.

During each of these steps, computational tools, such as graph theory or linear stability theory, are available to explore the properties of the model. Once constructed, the models, in the forms of the resulting differential equations, can be used for simulating the functioning of the cell in different situations. These simulations, in turn, can be used for a variety of scientific problems, many of which are of great interest in the pharmaceutical industry.

Abstract
Fluxes into secondary metabolism are usually much smaller than the primary metabolic flux they branch off from. Using metabolic control analysis general principles are derived that result from this feature. Control coefficients that quantify the control over this branch flux are expressed in elasticity coefficients for the branching metabolite, showing that the enzymes in the branch have a control of 1 on the branch flux (i.e. the flux is proportional to the enzyme activities in this branch). We show that, paradoxically, this does not make the branching enzymes necessarily the best target for increasing of the flux into the branch. It may be more effective to modulate enzymes in the main pathway. The control of the enzymes in the main pathway on the branch flux is dependent on the relative magnitude of the elasticity of the main pathway enzymes for the branching metabolite compared to the sensitivity of the enzymes in the branch. Typically when the enzymes in the main pathway have a low sensitivity for the branch-point metabolite it is better to manipulate outside the branch. The approach is illustrated with core and detailed kinetic models and shown experimentally for the branches around pyruvate in the lactic acid bacterium, Lactococcus lactis.

Abstract
Complex metabolic engineering, requiring the co-ordinate manipulation of multiple genes, is still in its infancy, but a variety of different methods for multi-gene manipulation have already been tested in plant and animal systems. Most conventional approaches have significant drawbacks, for instance 'iterative engineering' strategies take several generations to complete and can be slow and labour-intensive. Methods that promote the accumulation of multiple or complex transgenic loci or allow repeated use of particular regulatory sequences (such as promoters) can also suffer from transgene silencing. Recent years have seen the emergence of a number of novel methods, designed to overcome some of these problems, by using single transgenes to engineer co-ordinate suppression or over-expression of multiple target genes. This chapter reviews both conventional and novel methods for multi-gene engineering, presenting both the potential uses and problems associated with each method.

Abstract
The osmoprotectant glycine betaine is derived from choline via the intermediate betaine aldehyde in glycine betaine accumulating plants. Metabolic engineering of glycine betaine synthesis to increase plant resistance to osmotic stresses (drought and salinity) has initially focused on manipulation of genes encoding enzymes of choline and betaine aldehyde oxidation, including expression of choline monooxygenase and betaine aldehyde dehydrogenase in the chloroplast of tobacco leaves. However, these initial genetic interventions have met with only modest success, prompting detailed analysis of metabolic constraints in the choline synthesis network that limit flux to glycine betaine. Metabolic flux analysis using transient radiolabeling has shown that limitations reside at the level of choline import from the cytosol to the chloroplast, and endogenous synthesis of choline moieties in tobacco. Insights gained from metabolic modeling have helped guide subsequent rounds of engineering.

Abstract
The potential for metabolic engineering is briefly reviewed in the light of what has been added by this book. The list of frequently asked questions of Chapter 1 is answered. It is indicated that Integrative Genomics, Integrative Bioinformatics and System Biology are likely to be important new developments towards strengthening the tools for metabolic engineering.

Current Books: