Portal de Periódicos da CAPES

Mitotic-Exit Control as an Evolved Complex System

2005; Cell Press; Volume: 121; Issue: 3 Linguagem: Inglês

10.1016/j.cell.2005.04.006

1097-4172

William J. Bosl, Rong Li,

Fungal and yeast genetics research

The exit from mitosis is the last critical decision during a cell-division cycle. A complex regulatory system has evolved to evaluate the success of mitotic events and control this decision. Whereas outstanding genetic work in yeast has led to rapid discovery of a large number of interacting genes involved in the control of mitotic exit, it has also become increasingly difficult to comprehend the logic and mechanistic features embedded in the complex molecular network. Our view is that this difficulty stems in part from the attempt to explain mitotic-exit control using concepts from traditional top-down engineering design, and that exciting new results from evolutionary engineering design applied to networks and electronic circuits may lend better insights. We focus on four particularly intriguing features of the mitotic-exit control system and attempt to examine these features from the perspective of evolutionary design and complex system engineering. The exit from mitosis is the last critical decision during a cell-division cycle. A complex regulatory system has evolved to evaluate the success of mitotic events and control this decision. Whereas outstanding genetic work in yeast has led to rapid discovery of a large number of interacting genes involved in the control of mitotic exit, it has also become increasingly difficult to comprehend the logic and mechanistic features embedded in the complex molecular network. Our view is that this difficulty stems in part from the attempt to explain mitotic-exit control using concepts from traditional top-down engineering design, and that exciting new results from evolutionary engineering design applied to networks and electronic circuits may lend better insights. We focus on four particularly intriguing features of the mitotic-exit control system and attempt to examine these features from the perspective of evolutionary design and complex system engineering. Fundamentally, the mitotic-exit system, which includes the Cdc14 early anaphase release (FEAR) and mitotic-exit network (MEN) pathways, has relatively simple functionality. Yet an intricate complex control system has evolved to make the basic functions robust and precise under a variety of circumstances. From an engineering perspective, the mitotic-exit system, like many other biological pathways, appears to exhibit characteristics of a complex adaptive system. Simple systems, even complicated ones, can be decomposed into modules or pieces at all scales. An automobile or modern jet aircraft, as complicated but noncomplex systems, can be understood as the sum of their subsystems: the computers, the engine, braking systems, the flight stabilizers, and other major subsystems all have a clear function in the whole. Each of these can also be broken down and understood in terms of yet smaller components, down to the most basic mechanical and electronic parts. Viewed from the design perspective, this complicated system can be put together by many engineers, each working independently on separate components according to a master, top-down design plan. Traditional engineering design depends on this top-down, modular approach and the decomposability of the system. The system must perform precisely as the sum of all the components: it is designed under this assumption. Complex systems generally cannot be decomposed and built this way. Attempts to do so have met with spectacular failure. For example, the United States government spent billions of dollars designing a new air traffic control system that was ultimately scrapped. The system required was far too complex for the traditional design methods being used (Bar-Yam, 2003Bar-Yam, Y. (2003). When systems engineering fails—Toward complex systems engineering. International Conference on Systems, Man & Cybernetics. IEEE Press 2, 2021–2028.Google Scholar). On the other hand, the global internet, arguably the most complex human-engineered project to date, had no master blueprint but was evolved (Berners-Lee, 2000Berners-Lee T. Weaving the Web. HarperInformation, New York2000Google Scholar). In fact, it was realized early that the number of different kinds of computers, different communication standards, and the desire of programmers to do things their own way required an evolutionary design strategy. An important result of evolutionary design is that the dynamics of a complex system cannot be understood from its components and their interactions alone. The whole is more than the sum of the parts, which also imposes a natural scale on the system, below which system functions are lost. Some system functions cannot be found in any single component but exist only when components are combined in a certain configuration. However, some components may play critical roles in the system and their function is quite clear. In general, evolutionary design proceeds by allowing natural selection to manipulate components to construct a (complex) system that achieves the desired global behavior. The resulting designs often look very different from those that an engineer following traditional design principles would concoct (Antonsson and Cagan, 2001Antonsson E.K. Cagan J. Formal Engineering Design Synthesis. Cambridge University Press, Cambridge, United Kingdom2001Crossref Google Scholar). An interesting observation is that biological systems tend to defy modular design. Although attempts have been made to view biochemical networks in neat modular packages (Hartwell et al., 1999Hartwell L.H. Hopfield J.J. Leibler S. Murray A.W. From molecular to modular cell biology.Nature. 1999; 402: C47-C52Crossref PubMed Scopus (2668) Google Scholar), many interconnections between modules prohibit the black-box modularity that is a hallmark of top-down engineering design (Antonsson and Cagan, 2001Antonsson E.K. Cagan J. Formal Engineering Design Synthesis. Cambridge University Press, Cambridge, United Kingdom2001Crossref Google Scholar). Often, proteins that are key components in one biological pathway can be found performing other functions in another pathway. The term "pathway" is used to group proteins conceptually, but it is recognized that the black box is in fact rather transparent as proteins are routinely harnessed in multiple pathways. It is important, however, to emphasize that there are significant differences between evolutionary algorithms applied to engineering design and the evolutionary processes that occur in biological systems. Nevertheless, both natural and artificial evolved systems exhibit properties unlike traditionally engineered systems, which proceed from a predetermined overall plan. Evolution only tinkers (Alon, 2003Alon U. Biological networks: the tinkerer as an engineer.Science. 2003; 301: 1866-1867Crossref Scopus (443) Google Scholar) with existing parts until a working solution is found; it does not optimize or coordinate functions in advance. This perspective may help to understand large regulatory networks such as the mitotic-exit control system. The purpose of this article is not to provide a comprehensive review of mitotic-exit regulators and pathways (for that, several excellent recent reviews are available [Morgan, 1999Morgan D.O. Regulation of the APC and the exit from mitosis.Nat. Cell Biol. 1999; 1: E47-E53Crossref PubMed Scopus (297) Google Scholar, Murray, 2004Murray A.W. Recycling the cell cycle: cyclins revisited.Cell. 2004; 116: 221-234Abstract Full Text Full Text PDF PubMed Scopus (874) Google Scholar, Seshan and Amon, 2004Seshan A. Amon A. Linked for life: temporal and spatial coordination of late mitotic events.Curr. Opin. Cell Biol. 2004; 16: 41-48Crossref PubMed Scopus (36) Google Scholar, Simanis, 2003Simanis V. The mitotic exit and septation initiation networks.J. Cell Sci. 2003; 116: 4261-4262Crossref PubMed Scopus (23) Google Scholar). Instead, we focus on several important yet puzzling features of the mitotic-exit system and attempt to examine the underlying design principles from the perspective of complex systems constructed through evolutionary processes. The critical cell-cycle transition that controls the decision to physically divide a cell into two, an event known as cytokinesis, is termed "mitotic exit" because cytokinesis occurs with an interphase state of Cdk (cyclin-dependent kinase) activity. On paper, the decision to undergo mitotic exit is made based on a simple criterion: the genetic materials (chromosomes) must be segregated fully along an axis that is perpendicular to, and divided by, the plane of cleavage. The axis of cell division is often predetermined by a cell's environment, contacts, and developmental program. The spatial organization of cell division in budding yeast, like asymmetric cell divisions in many metazoan organisms, is ultimately determined by the axis of cell polarity (Figure 1) (Pruyne et al., 2004Pruyne D. Legesse-Miller A. Gao L. Dong Y. Bretscher A. Mechanisms of polarized growth and organelle segregation in yeast.Annu. Rev. Cell Dev. Biol. 2004; 20: 559-591Crossref PubMed Scopus (292) Google Scholar, Roegiers and Jan, 2004Roegiers F. Jan Y.N. Asymmetric cell division.Curr. Opin. Cell Biol. 2004; 16: 195-205Crossref PubMed Scopus (194) Google Scholar). Cell polarity directs asymmetric segregation and inheritance of proteins and organelles between the two progeny cells (called the mother and the bud). The actin cytoskeleton, established in a polarized manner early in the cell cycle, and a number of proteins localized in the bud ensure that the mitotic spindle is aligned and positioned such that elongation of the spindle in anaphase results in distribution of sister chromosomes to the two sides of the bud neck where the cytokinetic machine is assembled. The mitotic-exit control system in yeast ensures the temporal order between chromosome segregation and cytokinesis and also entails a spatial sensor to monitor the position of the elongated anaphase spindle relative to the polarity axis and the plane of cytokinesis. The output of the sensor must be able to influence the basic modules that control the timing of cell-cycle transitions, such as Cdk/cyclin complexes and the ubiquitin-mediated proteolysis system (Ingolia and Murray, 2004Ingolia N.T. Murray A.W. The ups and downs of modeling the cell cycle.Curr. Biol. 2004; 14: R771-R777Abstract Full Text Full Text PDF Scopus (56) Google Scholar, Morgan, 1999Morgan D.O. Regulation of the APC and the exit from mitosis.Nat. Cell Biol. 1999; 1: E47-E53Crossref PubMed Scopus (297) Google Scholar, Murray, 2004Murray A.W. Recycling the cell cycle: cyclins revisited.Cell. 2004; 116: 221-234Abstract Full Text Full Text PDF PubMed Scopus (874) Google Scholar). Figure 2 shows an overview of the mitotic-exit control system. The onset of anaphase is marked by the sudden separation of sister chromatids attached to opposite poles of the mitotic spindle. Sister chromatid separation is initiated by the APC/Cdc20 complex, which also triggers degradation of mitotic cyclins, and the FEAR pathway (Stegmeier et al., 2002Stegmeier F. Visintin R. Amon A. Separase, polo kinase, the kinetochore protein Slk19, and Spo12 function in a network that controls Cdc14 localization during early anaphase.Cell. 2002; 108: 207-220Abstract Full Text Full Text PDF PubMed Scopus (361) Google Scholar). Since Cdk1/cyclin is needed to sustain APC/Cdc20 activity, only partial cyclin degradation is achieved by APC/Cdc20 (Geymonat et al., 2002aGeymonat M. Jensen S. Johnston L.H. Mitotic exit: the Cdc14 double cross.Curr. Biol. 2002; 12: R482-R484Abstract Full Text Full Text PDF Scopus (26) Google Scholar). The FEAR network has a dual role: it is required for completion of chromosome separation (D'Amours et al., 2004D'Amours D. Stegmeier F. Amon A. Cdc14 and condensin control the dissolution of cohesin-independent chromosome linkages at repeated DNA.Cell. 2004; 117: 455-469Abstract Full Text Full Text PDF PubMed Scopus (231) Google Scholar) and also causes transient release of Cdc14 from its "prison"—the nucleolus (Stegmeier et al., 2002Stegmeier F. Visintin R. Amon A. Separase, polo kinase, the kinetochore protein Slk19, and Spo12 function in a network that controls Cdc14 localization during early anaphase.Cell. 2002; 108: 207-220Abstract Full Text Full Text PDF PubMed Scopus (361) Google Scholar, Azzam et al., 2004Azzam R. Chen S.L. Shou W. Mah A.S. Alexandru G. Nasmyth K. Annan R.S. Carr S.A. Deshaies R.J. Phosphorylation by cyclin B-Cdk underlies release of mitotic exit activator Cdc14 from the nucleolus.Science. 2004; 305: 516-519Crossref Scopus (139) Google Scholar, D'Amours and Amon, 2004D'Amours D. Amon A. At the interface between signaling and executing anaphase–Cdc14 and the FEAR network.Genes Dev. 2004; 18: 2581-2595Crossref PubMed Scopus (108) Google Scholar). Cdc14 is a protein phosphatase that triggers mitotic exit by dephosphorylating multiple targets (see below). Therefore, the FEAR pathway performs a control and timing function that connects chromosome separation to mitotic exit. If APC/Cdc20 and the FEAR pathway set the stage for finishing mitosis, the MEN provides the eventual trigger (McCollum and Gould, 2001McCollum D. Gould K.L. Timing is everything: regulation of mitotic exit and cytokinesis by the MEN and SIN.Trends Cell Biol. 2001; 11: 89-95Abstract Full Text Full Text PDF PubMed Scopus (219) Google Scholar). In a general sense, the MEN is a signal transduction system that monitors the position of the anaphase spindle relative to the polarity axis and the bud neck and then turns on a second wave of cyclin degradation and the cytokinetic machine (Figure 3). The MEN also provides a control device through which mitotic exit can be delayed, by returning Cdc14 to the nucleolus, if the spindle is improperly positioned (Geymonat et al., 2002aGeymonat M. Jensen S. Johnston L.H. Mitotic exit: the Cdc14 double cross.Curr. Biol. 2002; 12: R482-R484Abstract Full Text Full Text PDF Scopus (26) Google Scholar). The design of the spatial sensor in the MEN is clever: the orientation of the anaphase spindle is monitored by measuring the proximity of one of the spindle pole bodies (SPB) to the polar cortex in the bud. This proximity is only achieved if the anaphase spindle is properly aligned and one of the spindle poles successfully penetrates the bud neck. The sensor is composed of two general parts: components that mark the polar cortex in the bud and components that mark the SPB destined for the bud (Seshan and Amon, 2004Seshan A. Amon A. Linked for life: temporal and spatial coordination of late mitotic events.Curr. Opin. Cell Biol. 2004; 16: 41-48Crossref PubMed Scopus (36) Google Scholar). A central component capable of switch-like function is a small GTPase called Tem1, which localizes preferentially to the bud SPB. The full activity of Cdc14 leads to mitotic exit at least in part through dephosphorylation of Cdh1, another APC cofactor involved in cyclin degradation, and Sic1, an inhibitor of Cdk1 (Prinz and Amon, 1999Prinz S. Amon A. Dual control of mitotic exit.Nature. 1999; 402: 133-135Crossref Scopus (8) Google Scholar). APC/Cdh1 and Sic1 together eliminate mitotic Cdk1 activity, leading to mitotic exit, though additional targets of Cdc14 may yet be identified (D'Amours and Amon, 2004D'Amours D. Amon A. At the interface between signaling and executing anaphase–Cdc14 and the FEAR network.Genes Dev. 2004; 18: 2581-2595Crossref PubMed Scopus (108) Google Scholar). Cdc14 holds the key for mitotic exit. Prior to anaphase onset, Cdc14 is imprisoned in the nucleolus by binding to a nucleolar protein called Net1/Cfi1 (Torres-Rosell et al., 2005Torres-Rosell J. Machin F. Aragon L. Cdc14 and the temporal coordination between mitotic exit and chromosome segregation.Cell Cycle. 2005; 4: 109-112Crossref Scopus (17) Google Scholar). Its release and hence activation, strangely, are controlled sequentially by the FEAR network and the MEN. Cdc14 is bound in the nucleolus until early anaphase, when activation of the FEAR network initiates its release, a process that is thought to require phosphorylation of both Cdc14 and Net1/Cfi1 (Visintin et al., 2003Visintin R. Stegmeier F. Amon A. The role of the polo kinase Cdc5 in controlling Cdc14 localization.Mol. Biol. Cell. 2003; 14: 4486-4498Crossref Scopus (73) Google Scholar, Azzam et al., 2004Azzam R. Chen S.L. Shou W. Mah A.S. Alexandru G. Nasmyth K. Annan R.S. Carr S.A. Deshaies R.J. Phosphorylation by cyclin B-Cdk underlies release of mitotic exit activator Cdc14 from the nucleolus.Science. 2004; 305: 516-519Crossref Scopus (139) Google Scholar). After the initial release, Cdc14 would return to its imprisonment in the nucleolus unless the MEN is activated to sustain its release. What mechanisms could allow Cdc14 to be released in two pulses? In the first step, Cdc14's short-lived freedom could result from two negative feedback loops: (1) the released Cdc14 catalyzes dephosphorylation of itself and Net1/Cfi1, which enables their interaction, leading to resequestration of Cdc14 into the nucleolus (Jaspersen and Morgan, 2000Jaspersen S.L. Morgan D.O. Cdc14 activates cdc15 to promote mitotic exit in budding yeast.Curr. Biol. 2000; 10: 615-618Abstract Full Text Full Text PDF PubMed Scopus (136) Google Scholar); (2) the activities that promote Cdc14 release, such as Cdk1 (Azzam et al., 2004Azzam R. Chen S.L. Shou W. Mah A.S. Alexandru G. Nasmyth K. Annan R.S. Carr S.A. Deshaies R.J. Phosphorylation by cyclin B-Cdk underlies release of mitotic exit activator Cdc14 from the nucleolus.Science. 2004; 305: 516-519Crossref Scopus (139) Google Scholar), decline due to APC-mediated proteolysis that occurs downstream from Cdc14. These two negative feedbacks would ensure that Cdc14 release is not sustained without an additional activation step where the MEN comes into play. The MEN possibly sustains Cdc14 release by using a kinase Dbf2, activated downstream of Tem1 GTPase, to keep Cdc14 and Net1 in the phosphorylated state (Visintin et al., 2003Visintin R. Stegmeier F. Amon A. The role of the polo kinase Cdc5 in controlling Cdc14 localization.Mol. Biol. Cell. 2003; 14: 4486-4498Crossref Scopus (73) Google Scholar). Interestingly, the MEN-induced Cdc14 release might also be self-terminating due to a negative feedback loop: the released Cdc14 localizes to the SPB where it dephosphorylates Bfa1, a subunit of the GTPase-activating protein (GAP) for Tem1, resulting in GAP activation and Tem1 assuming the inactive GDP bound state (Pereira et al., 2002Pereira G. Manson C. Grindlay J. Schiebel E. Regulation of the Bfa1p-Bub2p complex at spindle pole bodies by the cell cycle phosphatase Cdc14p.J. Cell Biol. 2002; 157: 367-379Crossref PubMed Scopus (109) Google Scholar). To make the matter even more complicated, the two steps of Cdc14 release are connected with a positive feedback loop: Cdc14 released by the FEAR stimulates MEN activity by dephosphorylation of Cdc15 (Stegmeier et al., 2002Stegmeier F. Visintin R. Amon A. Separase, polo kinase, the kinetochore protein Slk19, and Spo12 function in a network that controls Cdc14 localization during early anaphase.Cell. 2002; 108: 207-220Abstract Full Text Full Text PDF PubMed Scopus (361) Google Scholar). Figure 4 illustrates the intricate interconnectedness of these feedback loops. The complicated interconnections that have so far been identified suggest that it may not be possible to decompose the mitotic-exit process into distinct modules, as might be required for analyzing a top-down engineering design. Why does yeast employ this two-clutch, self-limiting system to control Cdc14 release? Negative feedback loops in electronic circuits are commonly used to remove distortion from amplified signals, a way of adding robustness to the system by damping out noise. Noise is a common problem in biological control systems because biochemical interactions are often reversible and incomplete. Fluctuation in the level of the reaction components and variation in reaction rates as a function of environmental parameters all contribute to the noise. For example, as discussed below, the Tem1 GTPase, a key MEN component, can self-activate and its rate of nucleotide exchange is sensitive to temperature. Fluctuation in Tem1 activity can seriously affect timing of cell-cycle events by influencing Cdc14 release. The negative feedback loops could damp out the spontaneous fluctuation of free Cdc14 level due to noisy components of the mitotic-exit control system. Negative feedback loops in biological networks indeed have also been found as parts of an elegant structure for creating natural oscillatory or timing functions that are robust to noisy input signals (Becskei and Serrano, 2000Becskei A. Serrano L. Engineering stability in gene networks by autoregulation.Nature. 2000; 405: 590-593Crossref PubMed Scopus (1132) Google Scholar). The sequence of events from chromosome separation to segregation and cytokinesis requires strict ordering. This requirement may have caused negative-feedback structures to evolve in the mitotic-exit control apparatus. Another possible explanation for the two-clutch release of Cdc14 is that the MEN plays a surveillance role. If the FEAR network ignites the fuse that leads to mitotic exit, the MEN seems to be a separate control device inserted halfway in the fuse and allows another input into the decision to undergo mitotic exit. In this capacity, the MEN acts as a spatial sensor monitoring the orientation and position of the anaphase spindle. An interesting question is how this additional control has evolved. One possibility is that in a primitive cell, where the orientation of mitosis might be inconsequential, the FEAR network could represent the sole mechanism for Cdc14 release, and the MEN emerged later coevolving with oriented cell division. The FEAR and the MEN are structured quite different, and yet the two modules are redundant for mitotic exit: whereas the FEAR is not required for mitotic exit with normally functioning MEN, the requirement for the MEN in mitotic exit can be diminished by slight overexpression of Spo12, a FEAR network component (Toyn and Johnston, 1993Toyn J.H. Johnston L.H. Spo12 is a limiting factor that interacts with the cell cycle protein kinases Dbf2 and Dbf20, which are involved in mitotic chromatid disjunction.Genetics. 1993; 135: 963-971Crossref PubMed Google Scholar). And certainly in meiosis the MEN does not seem to play a major role. Thus, it is possible that the original FEAR control of mitotic exit became less effective during the evolution process to allow additional control by the MEN. GTPases are often used as biological switches because these proteins adopt different conformations when bound to GTP or GDP and can convert between the two nucleotide bound states through GTP hydrolysis and nucleotide exchange reactions (Bourne, 1995Bourne H.R. GTPases: a family of molecular switches and clocks.Philos. Trans. R. Soc. Lond. B Biol. Sci. 1995; 349: 283-289Crossref Scopus (36) Google Scholar). The GTP bound state is usually the "on" state where the GTPase interacts with downstream effectors and elicits specific signaling effects. For most Ras super family GTPases, the exchange of GDP to GTP is limited by the rate of GDP dissociation, a property that positions the guanine nucleotide exchange factor (GEF) as a critical regulator of the "on" switch. The reverse switch, from the GTP to GDP bound state, is catalyzed by the GTPase itself and is accelerated by the GTPase-activating proteins (GAP). Tem1, however, has an unusually high intrinsic nucleotide exchange ability under physiological conditions and thus on its own exhibits little dependence on the GEF for activation. The intrinsic nucleotide exchange reaction appears to be temperature dependent, and only at 13°C was the GDP dissociation rate slow enough to be measured in a previous study (Geymonat et al., 2002bGeymonat M. Spanos A. Smith S.J. Wheatley E. Rittinger K. Johnston L.H. Sedgwick S.G. Control of mitotic exit in budding yeast. In vitro regulation of Tem1 GTPase by Bub2 and Bfa1.J. Biol. Chem. 2002; 277: 28439-28445Crossref Scopus (82) Google Scholar). The self-activating property of Tem1 forms the basis for several important properties of the mitotic-exit network. First, the fast intrinsic GDP-to-GTP exchange introduces high flexibility to the Tem1 GTPase switch, allowing fine tuning of the relative levels of Tem1GTP and Tem1GDP by both the GEF (Lte1) and the GAP (Bub2/Bfa1 complex). For example, a high level of Tem1GTP can be achieved by either promoting GEF action or by inhibiting GAP activity. Conversely, Tem1GTP can be reduced either by restricting the interaction with the GEF or by activating the GAP. Indeed, the activity and localization of Lte1 and Bub2/Bfa1 complex are regulated in many ways (see below), allowing Tem1 to function as a dynamic switch that integrates spatial and temporal inputs through multiple pathways. Second, the flexibility in the way by which Tem1 can be activated explains the high degree of functional redundancy in mitotic-exit regulation observed in genetic experiments. Although Tem1 is required for mitotic exit, Lte1 is not required at temperatures above 30°C (Yoshida et al., 2003Yoshida S. Ichihashi R. Toh-e A. Ras recruits mitotic exit regulator Lte1 to the bud cortex in budding yeast.J. Cell Biol. 2003; 161: 889-897Crossref PubMed Scopus (51) Google Scholar). In the absence of Lte1, Tem1 can be activated through its intrinsic nucleotide exchange reaction coupled with inhibition of its GAP (Bfa1/Bub2) through phosphorylation by Cdc5, a FEAR network component (Geymonat et al., 2003Geymonat M. Spanos A. Walker P.A. Johnston L.H. Sedgwick S.G. In vitro regulation of budding yeast Bfa1/Bub2 GAP activity by Cdc5.J. Biol. Chem. 2003; 278: 14591-14594Crossref Scopus (73) Google Scholar, Hu et al., 2001Hu F. Wang Y. Liu D. Li Y. Qin J. Elledge S.J. Regulation of the Bub2/Bfa1 GAP complex by Cdc5 and cell cycle checkpoints.Cell. 2001; 107: 655-665Abstract Full Text Full Text PDF PubMed Scopus (202) Google Scholar). Lte1 and the temperature sensitivity of Tem1 self-activation through its intrinsic nucleotide exchange may have coevolved to deal with low-temperature situations, when spindle orientation and movement are slowed down due to impaired microtubule assembly (Huffaker et al., 1988Huffaker T.C. Thomas J.H. Botstein D. Diverse effects of beta-tubulin mutations on microtubule formation and function.J. Cell Biol. 1988; 106: 1997-2010Crossref PubMed Scopus (261) Google Scholar, Richards et al., 2000Richards K.L. Anders K.R. Nogales E. Schwartz K. Downing K.H. Botstein D. Structure-function relationships in yeast tubulins.Mol. Biol. Cell. 2000; 11: 1887-1903Crossref PubMed Scopus (101) Google Scholar). In this situation, failure in spindle positioning could occur frequently, thus necessitating a spatial sensor that facilitates the coupling between spindle orientation and Tem1 activation. Experiments with evolutionary algorithms demonstrate how natural selection can exploit specific or unusual properties of network components in the design process. Thompson, 1997Thompson, A. (1997). Temperature in Natural and Artificial Systems. Paper presented at 4th Eur. Conf. on Artificial Life (ECAL97) (Cambridge, MA: MIT Press). http://www.informatics.sussex.ac.uk/users/adrianth/er97/paper.ps.Google Scholar) used computational algorithms to directly manipulate a semiconductor medium, called a field programmable gate array (FPGA), to automatically construct an electronic circuit. Note that the computer algorithm manipulated the network connections between real transistors. The system evolved by this process is not an idealized mathematical system but a physical system, where the laws of physics and the subtle variability of real materials determine system behavior. The final evolved circuit operated perfectly over the 10°C temperature range that the population experienced during evolution. Some circuit elements appeared to be disconnected from the main circuit but interacted through subtle electrical coupling properties, showing that physical characteristics not included in the design algorithm were nevertheless exploited by the evolutionary process. The circuit was also much smaller—by one or two orders of magnitude—than would be expected from conventional design experiments, demonstrating a very efficient use of resources. Whether this is a general tendency of evolutionary design that would be operative in biological systems is not known and should be explored further. Additional experiments showed that greater robustness was built into the system when it was exposed to a wider range of conditions during the evolutionary process (Thompson, 1996Thompson, A. (1996). Evolutionary Techniques for Fault Tolerance. Proc. UKACC Int. Conf. on Control (CONTROL'96), IEEE Conference Publication 427, 693–698. http://www.informatics.sussex.ac.uk/users/adrianth/control96/paper.ps.Google Scholar, Thompson and Wasshuber, 2000Thompson A. Wasshuber C. Design of single-electron systems through artificial evolution.Int. J. Circ. Theor. Appl. 2000; 28: 585-599Crossref Scopus (13) Google Scholar). The surprising result of this experiment was that natural selection resulted in an efficient, robust system that incorporated unique characteristics of the components in ways that were bizarre and unlike anything an engineer would do following traditional design practices. The unusual biochemical property of Tem1 may have been incorporated into or coevolved with the MEN in a similar manner in response to a need for integrating multiple input signals and robustness to a range of external or internal variations. The Tem1 GTPase and its regulators form the core of the spatial sensor that monitors the correct orientation of the anaphase spindle. The simple model for the function of this spatial sensor has been that the correct spindle orientation is