Systems biological analyses of intracellular signal transduction D i s s e r t a t i o n zur Erlangung des akademischen Grades d o c t o r r e r u m n a t u r a l i u m ( Dr. rer. nat.) im Fach Biophysik eingereicht an der Mathematisch-Naturwissenschaftlichen Fakultät I der Humboldt-Universität zu Berlin von Herrn Diplom-Biochemiker Stefan Legewie geboren am 31.10.1977 in Aachen Präsident/Präsidentin der Humboldt-Universität zu Berlin Prof. Dr. Christoph Markschies Dekan/Dekanin der Mathematisch-Naturwissenschaftlichen Fakultät I Prof. Dr. Christian Limberg Gutachter: Prof. Dr. Hanspeter Herzel Prof. Dr. Jens Timmer Prof. Dr. Olaf Wolkenhauer Tag der mündlichen Prüfung: 31. 10. 2008 ii für meine Eltern iii iv Summary Intracellular regulatory networks involved in the sensing of extracellular cues are crucial to all living organisms. Signal transduction networks allow unicellular organisms sensing nutrient availability, finding mating partners and responding to stress. Moreover, intercellular communication is the fundamental basis for the functioning and homeostasis of multicellular organisms. Accordingly, many diseases including cancer are caused by deregulation of signal transduction networks. Extracellular signals are typically transmitted rapidly from the cell membrane to the nucleus by activation of multi-level enzymatic cascades which ultimately elicit slow changes in gene expression, and thereby affect the cell fate. These signalling cascades are highly interconnected, thus giving rise to complex networks, that are hard to understand intuitively. In this thesis, a combination of kinetic modeling and analysis of quantitative experiments is applied to get insights into the principles of intracellular signalling. In the first part, the dynamics of enzymatic signalling cascades involved in transducing signals from the cell membrane to the nucleus are investigated. The Ras-MAPK cascade plays a central role in various physiological processes such as cell cycle progression, cell differentiation and cell death, and mutational cascade activation appears to be crucial for cancer development. Overexpression of wildtype Ras is also frequently observed in tumours, but its functional relevance remains unclear. By analysis of an experimentally validated kinetic model of Ras signalling, it is shown in Chapter 2 that the basal state MAPK signalling can be completely insensitive towards overexpression of the uppermost cascade member Ras. Thus, the simulations reveal a “kinetic tumour suppression effect” inherent to the Ras (de)activation cycle, and also explain experimental studies showing that overexpression events within the MAPK cascade, though phenotypically silent in isolation, frequently cooperate to bring about strong cellular deregulation (“oncogene cooperation”). In Chapter 3, it is analysed how an experimentally validated MAPK cascade model responds to more physiological, transient inputs and converts them into an all-or-none, irreversible cell fate decision. More specifically, it is shown that bistability arises in the core MAPK cascade by a previously unrecognised enzyme sequestration effect that establishes a hidden positive feedback loop. Chapter 4 is focussed on the proteolytic caspase cascades controlling apoptosis, a form of cell suicide activated in response to extracellular stress. The simulations suggest an unanticipated role for inhibitors of apoptosis proteins (IAPs): Simultaneous inhibition of multiple caspases by IAPs can result in strong positive feedback regulation, and may thus be essential to establish all-or-none and irreversible initiation of cell death. Cellular commitment to a new fate typically requires ongoing extracellular stimulation and/or intracellular signalling for several hours, so that the long-term dynamics of signalling cascades are important for cellular responses. In the second part of the thesis, it is investigated how slow signal-induced changes in gene expression feed back into the signalling network and modulate its dynamical activation pattern. In Chapter 5, the general design principles underlying transcriptional feedback regulation of mammalian signalling pathways are investigated by analysing the stimulus-induced gene expression profiles of 134 intracellular signalling proteins. It turns out that transcriptional feedback regulation occurs in each of the five signalling cascades considered, and that negative feedback strongly dominates over positive feedback. Moreover, negative feedback exclusively occurs by transcriptional induction of a subgroup of signal inhibitors, termed rapid feedback inhibitors (RFIs), while downregulation of signal transducers plays no role. Systematic analysis of mRNA and protein half-lives reveals a remarkable separation of the signalling network into flexible and static parts: transcriptionally regulated RFIs are unstable at the mRNA and protein level, while other signalling proteins are generally stable. Kinetic modelling, also presented in Chapter 5, is employed to get insights into the functional implications of RFI- mediated transcriptional feedback regulation. In Chapter 6, transcriptional feedback regulation of TGFβ signalling via Smad transcription factors is analysed in more detail in primary hepatocytes to confirm the physiological relevance of transcriptional feedback v regulation at the protein level. The TGFβ family of cytokines constitute major inhibitors of cell growth, and accordingly they play important roles in various physiological and pathological processes including development, tissue homeostasis, tissue regeneration, and cancer. Genome-wide microarray analyses and protein measurements in response to TGFβ stimulation (presented in Chapter 6) suggest that the SnoN oncoprotein is the central transcriptional feedback regulator in primary mouse hepatocytes. A mathematical model including TGFβ-induced Smad signalling and SnoN-mediated feedback is fitted to experimental data obtained under various stimulation conditions, and predictions derived from the model are then quantitatively confirmed in primary hepatocytes isolated from SnoN knock-out mice. The modelling results in Chapter 6 mechanistically explain how a small pool of SnoN proteins can efficiently regulate a much larger pool of Smad proteins, and further support the relevance of transcriptional negative feedback regulation in signal transduction. Cells face a specificity problem as different extracellular stimuli frequently engage the same set of intracellular signalling pathways even though they elicit completely different biological responses. Experimental evidence suggests that stimulus-specific biological information is frequently encoded in the quantitative aspects of stimulus-specific activation kinetics (e.g., signal amplitude and/or duration). If biological information is encoded in the quantitative characteristics of intracellular signals, proper cell fate decisions require that the downstream gene expression machinery is able to accurately decode signal amplitude and/or duration. Part III of this thesis deals with such decoding of upstream signals by the gene expression machinery, and thus represents a first step towards more integrated systems biological models that include both, upstream signal transduction and downstream phenotypic responses such as cell growth. The results presented in Section 7 identify competitive inhibition and regulated degradation as mechanisms that allow intracellular regulatory networks to efficiently discriminate transient vs. sustained signals. More specifically, a combination of mathematical modelling and quantitative experimental analyses reveals that a recently discovered small non-coding RNA, IsrR, establishes a pronounced delay and duration decoding in the cyanobacterial gene expression response towards iron stress. In other words, it is shown that the small non-coding RNA, IsrR, restricts the potentially harmful and costly expression of late-phase stress proteins to severe, prolonged and ongoing stress conditions. Many of the downstream target genes induced by signalling pathways are transcription factors, thus giving rise to a complex transcriptional regulatory network. Therefore, signal decoding at the level of gene expression cannot be fully understood by insights into the functioning of small transcriptional regulatory motifs, but additionally requires integrated analyses of multiple transcription factors. In Chapter 8, a recently proposed reverse engineering approach, called modular response analysis (MRA), is applied to derive the topology of an oncogenic transcription factor network from high-throughput and knock- down data. Statistical analyses of the MRA results are used to derive predictions that can be verified experimentally, and also identify a key transcription factor cascade whose existence is supported by the published literature. In conclusion, this thesis shows how systems biological analyses can enhance our understanding of intracellular signalling networks. The approaches presented here include quantitative analyses of small regulatory motifs (Chapters 2, 3, 4 and 7), systematic investigation of high-throughput data (Chapters 5 and 6), kinetic modelling and thus integration of multiple time course experiments (Chapter 6), as well as reverse engineering of regulatory network topologies (Chapter 8). vi Zusammenfassung Intrazelluläre Regulationsnetzwerke, die an der Übertragung extrazellulärer Signale beteiligt sind, haben eine zentrale Bedeutung für alle Organismen. In Einzellern wird Signalübertragung zum Beispiel