Dissecting the puzzle of life: modularization of signal transduction networks

Abstract

Cells have developed complex control networks which allow them to sense and response to changes in their environment. Although they have different underlying biochemical mechanisms, signal transduction units in prokaryotes and eukaryotes fulfill similar tasks, such as switching on or off a required process or amplifying a certain signal. The growing amount of data available allows the development of increasingly complex models which offer a detailed picture of signaling networks, but the properties of these systems as a whole become difficult to grasp. A sound strategy to untangle this complexity is a decomposition into smaller units or modules. How modules should be delimited, however, remains an unanswered question. We propose that units without retroactive effects might be an interesting criterion. In this contribution, this issue will be explored through several examples, starting with a simple two-component system in Escherichia coli up to the complex epidermal growth factor signaling pathway in human cells.

Introduction

Cells are equipped with exquisite sensing systems which allow them to be continously aware of the conditions in their environment and react appropriately to these conditions. The basic elements of a cellular signaling system are a sensor protein, made of a receptor domain and a transmitter domain, and a response regulator, consisting of a receiver domain and a regulator domain (Lengeler, 2000). Stimulation of the sensor (normally bound to the cell membrane) leads to activation of the transmitter, which produces an intracellular signal. This signal is processed by a cascade of molecules and finally arrives at the receiver, which in turn activates the regulator (see Fig. 1). Regulators produce a response by modulating gene expression or enzyme activities. The key components in this transfer of information are proteins, which form networks and are able to perform computational tasks. Proteins can change their state by interaction with other proteins or by biochemical modifications (such as phosphorylations) catalyzed by other proteins (Bray, 1995). Another common mechanism is the release of small molecules called second messengers, which diffuse in the cell and activate other proteins (Krauss, 2001). Interestingly, although eukaryotes systems are generally more complex, both prokaryotes and eukaryotes follow the same logic. Especially in eukaryotes, enhanced computation possibilities are achieved by inserting elements between the basic elements described above (Lengeler, 2000).

Bacteria, for example, have the capability to use a broad range of nutrient sources for life. Furthermore, they are also able to synthesize a number of monomers like amino acids if these are not provided in the medium. To sense their external environment, bacteria often use rather simple signal transduction systems. A paradigm of bacterial signal transduction is the two-component system that consists just of two elements, the sensor kinase and the response regulator (Hoch & Silhavy, 1995). Bacteria are also able to sense intracellular conditions. One representative is the phosphotransferase system (PTS) (Postma, Lengeler, & Jacobson, 1993). The PTS is an uptake system for several carbohydrates in Escherichia coli. In addition, it acts as a sensor and is involved in the control of uptake of a number of carbohydrates.

Human cells also posses a complex signaling system which allows them to exchange information and thus coordinate themselves. In most cases, the signaling processes follow the general schema described above. The binding of extracellular signals such as hormones or growth factors to receptors results in changes in the intracellular part (transmitter) of the receptor. The thus activated receptor transmits the signal to intracellular signaling intermediates, triggering signaling cascades which finally activate transcription factors (regulators) which move into the nucleus, changing the gene expression of the cell (Downward, 2001). Essential processes like proliferation, cell development or even the suicide of the cell are controlled by cell signaling. Since it is related to such basic properties of the cells, defective signal processing can lead to important diseases such as cancer or diabetes, and thus signaling pathways are important targets for disease therapy (Levitzki, 1996).

The high number of components involved, the complex crosstalk phenomena among the different pathways and the biophysical regulation set up a picture difficult to grasp (Asthagiri & Lauffenburger, 2000). A useful tool to untangle this complexity might be mathematical modeling. The knowledge and amount of data available about signaling networks grows steadily, boosting the development of increasingly complex models. These models offer a highly detailed picture of signaling pathways, but the properties of these systems as a whole become difficult to understand. This holistic understanding is the target of the emerging discipline of systems biology (Kitano, 2002). Engineers usually face synthesis problems: design a system with certain characteristics using a set of well-characterized elements. A system-biologist has to face an analysis problem: understand the properties of a complex network. Therefore, the definition of functional units, i.e. entities whose function is separable from those of other units, might help to analyze biological systems (Hartwell, Hopfield, Leibler, & Murray, 1999) since, once modules are defined, they could be systematically analyzed regarding properties such as stability, robustness and dynamic behavior and classified, creating a library of reusable units. Once these relatively simple units are well understood, they can be re-assembled in order to analyze the emergent properties of the resulting systems, as engineers do. Furthermore, this set of reusable elements would simplify the set-up of models, since many parts of biological networks are found in several signal transduction pathways.

How biochemical modules should be delimited still remains an unanswered question, and is a topic under intense investigation. While several approaches are based on network-clustering methods applied to experimental data (e.g. Rives & Galitski, 2003), others try to develop a suitable theoretical framework for the analysis of modular networks (e.g. Bruggeman, Westerhoff, Hoek, & Kholodenko, 2002). We have previously introduced three biologically motivated criteria for defining functional units: (1) common physiological task (all the elements of a functional unit perform the same task, e.g. the specific catabolic pathways for individual carbohydrates), (2) common genetic units (the genes for all enzymes of a functional unit are organized in genetical units, e.g. operons and modulons in bacteria) and (3) common signal transduction network (all elements of a functional unit are interconnected within a common signal transduction system and the signal flow over the unit border is small compared to the information exchange within the unit (Kremling, Jahreis, Lengeler, & Gilles, 2000).

In this contribution, a novel criterion for the definition of modules, namely the absence of retroactivity in the connections between the modules, will be proposed. The different situations that can lead to a retroactivity-free connection will be first examined by means of the network theory, and later applied to two signaling systems in prokaryotes (the two-component system and the control of carbohydrate uptake) and two in eukaryotes (the MAPK Cascade and the EGF signaling network). Our approach does not intend to provide an algorithm to find modules from a set of experimental data (using e.g. network-clustering methods), but rather a theoretical framework to analyze signaling networks in a modular and systems-theoretical manner.