Biotechnology Advances 30 (2012) 108–130 Contents lists available at SciVerse ScienceDirect Biotechnology Advances journal homepage: www.elsevier.com/locate/biotechadv Sic1 plays a role in timing and oscillatory behaviour of B-type cyclins Matteo Barberis a,⁎, Christian Linke b,c,1, Miquel À. Adrover d,1, Alberto González-Novo d, Hans Lehrach b, Sylvia Krobitsch b, Francesc Posas d, Edda Klipp a a Institute for Biology, Theoretical Biophysics, Humboldt University Berlin, Invalidenstraβe 42, 10115 Berlin, Germany b Max Planck Institute for Molecular Genetics, Ihnestraβe 73, 14195 Berlin, Germany c Department of Biology, Chemistry and Pharmacy, Free University Berlin, Takustraβe 3, 14195 Berlin, Germany d Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra, Dr. Aiguader 88, 08003 Barcelona, Spain article info abstract Available online 18 September 2011 Budding yeast cell cycle oscillates between states of low and high cyclin-dependent kinase activity, driven by association of Cdk1 with B-type (Clb) cyclins. Various Cdk1–Clb complexes are activated and inactivated in a Keywords: fixed, temporally regulated sequence, inducing the behaviour known as “waves of cyclins”. The transition Cell cycle from low to high Clb activity is triggered by degradation of Sic1, the inhibitor of Cdk1–Clb complexes, at Oscillations the entry to S phase. The G1 phase is characterized by low Clb activity and high Sic1 levels. High Clb activity Cyclin waves and Sic1 proteolysis are found from the beginning of the S phase until the end of mitosis. The mechanism reg- B-type cyclins ulating the appearance on schedule of Cdk1–Clb complexes is currently unknown. Here, we analyse oscilla- Sic1 Budding yeast tions of Clbs, focusing on the role of their inhibitor Sic1. We compare mathematical networks differing in interactions that Sic1 may establish with Cdk1–Clb complexes. Our analysis suggests that the wave-like cyclins pattern derives from the binding of Sic1 to all Clb pairs rather than from Clb degradation. These pre- dictions are experimentally validated, showing that Sic1 indeed interacts and coexists in time with Clbs. In- triguingly, a sic1Δ strain looses cell cycle-regulated periodicity of Clbs, which is observed in the wild type, whether a SIC1-0P strain delays the formation of Clb waves. Our results highlight an additional role for Sic1 in regulating Cdk1–Clb complexes, coordinating their appearance. © 2011 Elsevier Inc. All rights reserved. 1. Introduction loops – via down-regulation of Clb levels via ubiquitin/26S protea- some pathway (Amon et al., 1994; Hochstrasser, 1995; Irniger Budding yeast cell cycle is driven by periodic changes in the activ- et al., 1995; King et al., 1996; Lew and Reed, 1995; Seufert et al., ity of Cdk1 kinase, regulated by different pools of cyclins that associ- 1995; Tyers and Jorgensen, 2000). ate with Cdk1 in successive waves (Fig. 1A, reviewed in Futcher, The transcriptional regulation of Clbs is a fine-tuned mechanism 1996). B-type cyclins Clb1–6 are expressed at different times and ap- (Fig. 1B, reviewed in Bloom and Cross, 2007). CLB5,6 transcription is pear sequentially in specific cell cycle phases, resulting in a significant promoted by the Mbp1/Swi6 Binding Factor (MBF) (Schwob and divergence of function (Bloom and Cross, 2007; Cross et al., 1999). Nasmyth, 1993), and interactions of Clb5,6 with Swi4/6 Binding Fac- Clb5,6 rise at the beginning of S phase and function primarily in the tor (SBF) and MBF have been reported (Pic-Taylor et al., 2004; control of DNA replication (Schwob and Nasmyth, 1993; Schwob Simon et al., 2001). CLB1,2 transcription is controlled by the Fkh2 et al., 1994; Spellman et al., 1998). Clb3,4 increase in mid-S phase at forkhead transcription factor during G2/M phase (Kumar et al., about the same time as spindle pole bodies separate and their specific 2000; Reynolds et al., 2003) and both Cdk1–Clb5 and Cdk1–Clb2 in- function is still unclear (Fitch et al., 1992; Richardson et al., 1992). teract with, and phosphorylate, Fkh2 to control Clb1,2 accumulation Clb1,2 rise as mitotic spindle assembly progresses and are involved (Hollenhorst et al., 2000; Pic-Taylor et al., 2004; Ubersax et al., in the control of mitotic exit (Deshaies, 1997; Fitch et al., 1992; 2003; Yeong et al., 2001)(Fig. 1B, arrows C and D, respectively). No Spellman et al., 1998).TheregulationofactiveCdk1–Clb complexes information is available so far about the activation of CLB3,4 transcrip- involves a combination of positive feed-forward loops – depending tion. The only study reported to date is a genome-wide location anal- on the regulated transcription of CLB genes (Bloom and Cross, 2007; ysis to identify binding sites for transcription factors, which suggested Fitch et al., 1992; Koch and Nasmyth, 1994) – and negative feedback that the Fkh1 forkhead transcription factor binds to Clb4 (Simon et al., 2001). Moreover, Fkh1 binds to the CLB4 promoter (CCDB database, Alfieri et al., 2007). ⁎ Corresponding author. Tel.: +49 30 2093 8383; fax: +49 30 2093 8813. The degradation of Clbs by proteolysis is a highly specific mecha- E-mail addresses: [email protected] (M. Barberis), [email protected] (E. Klipp). nism (Fig. 1C), with the consequence that Cdk1 is inactivated. Clb6 1 These two authors contributed equally to the present study. is the only B-type cyclin to be directed to degradation by the ubiquitin 0734-9750/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.biotechadv.2011.09.004 M. Barberis et al. / Biotechnology Advances 30 (2012) 108–130 109 A t5,6 t3,4 t1,2 Clb5,6 Clb3,4 Clb1,2 B-type cyclin levels B-type cyclin G1 SG2 M Cell cycle progression B Clb5 Clb6 MBF CLB5,6 A Clb3 Clb4 C ? CLB3,4 B Clb1 Clb2 D Fkh2 CLB1,2 C (b) Shirayama et al. 1999, Zachariae et al. 1998 (c) Wäsch & Cross 2002 (a) Rudner & Murray 2000 S-Cdk M-Cdk Clb5 Clb3 Clb4 Clb2 APCCdc20 APCCdh1 (d) Zachariae et al. 1998 (f) Yeong et al. 2001 (e) Yeonget al. 2001,Zachariaeet al. 1998 G1-S G2-M M (meta-ana) M-G1 Fig. 1. Mechanisms of regulation of B-type cyclins. (A) Qualitative description of waves of Clbs during the cell cycle. Time distances between the peaks of Clbs, mentioned in the text, are also visualized: t3,4−t5,6 represents the time delay between Clb5,6 and Clb3,4, t1,2−t5,6 the time delay between Clb5,6 and Clb1,2 and t1,2−t3,4 the time delay between Clb3,4 and Clb1,2. (B) Transcriptional regulation of Clbs. Cdk1–Clb5,6 promote CLB3,4 transcription (A), Cdk1–Clb3,4 promote CLB1,2 transcription (B) together with Cdk1–Clb5,6 (C), and Cdk1–Clb1,2 promote CLB1,2 transcription by stimulating its own production (D). For simplicity, Cdk1 subunit has been omitted. See text for details. (C) Clb-regulated degradation. Phosphorylation of Cdc20 by Cdk1–Clb2 activates APCCdc20 (a). APCCdc20 targets Clb5 for degradation (b) and degrades also mitotic cyclins (c). APCCdh1 degrades Clb2 further during Cdh1 mitotic exit and in the following G1 phase (d). Phosphorylation of Cdh1 by Cdk1–Clb5 (e) and by Cdk1-Clb3,4 (f) inactivates APC , thereby permitting the subsequent accumu- lation of Clb2. Bibliographic references are also indicated. 110 M. Barberis et al. / Biotechnology Advances 30 (2012) 108–130 ligase complex, SCF (Skp1/Cullin/F-box), before the degradation of described. Importantly, all parameters used for the simulations have Clb5 (Jackson et al., 2006). The other B-type cyclins are ubiquitinated the same value within all networks. by the Anaphase Promoting Complex (APC) (Peters, 2006). This mechanism of degradation occurs through the mutual regulation of 2.2. Local sensitivity analysis two forms of APC, APCCdc20 and APCCdh1, which are differentially reg- – ulated by Cdk1 Clb-mediated phosphorylation (Shirayama et al., To test the impact of parameters on the dynamics, local sensitivity 1999; Wasch and Cross, 2002). analysis is performed. Classical sensitivity analysis defines the sen- The sequential activation and degradation of Clbs gives direction- sitivity S as the change of a model output quantity O caused by fi ality to cell cycle events. In addition, Sic1, a speci c stoichiometric in- the change of a parameter value p, i.e. S=(ΔO/O)/(Δp/p). Since we – hibitor of Cdk1 Clb complexes and structural and functional homolog study the impact of parameter changes on concentration time courses, Kip1 to 27 in mammalian cells (Barberis et al., 2005a), contributes to we performed sensitivity analysis by calculating the time-dependent their regulation (Schwob et al., 1994). Sic1 is synthesized at the response coefficients R=(∂ci(t)/ci(t))/(∂p/p)(Ingalls and Sauro, end of mitosis and largely degraded at the onset of S phase in an 2003). These coefficients indicate the direction and amount of change SCF-dependent manner (Schwob et al., 1994; Verma et al., 1997b). of the time course for the concentration ci(t) upon an infinitesimal – – Cdk1 Cln1,2 and Cdk1 Clbs phosphorylate Sic1 (Nash et al., 2001; change of the parameter (or initial concentration) p. They allow trac- Nasmyth, 1996; Verma et al., 1997b), which results in Sic1 being rec- ing the effect of a parameter change on a concentration during the ognized by the protein degradation machinery. In early G1 phase, whole simulation period. The analysis focused on Cdk1–Clb com- Cdk1 activity is abolished by low levels of CLB expression, by plexes. Calculation of time-dependent response coefficients serve to Cdh1 high levels of Sic1 and by the APC -mediated ubiquitination of test the influence of parameters on the timing and strength of re- Clbs.