Cytokinesis review

Early cytokinesis; from anaphase to midbody formation cytokinesis is the physical separation of a dividing cell into two identical daughter cells [Eggert et al., 2006]. Cytokinesis begins following the separation of DNA during anaphase, and the consequent APC-mediated destruction of the Cdk1-cyclin B complex, which negatively regulates many proteins essential for cytokinesis [Hummer and Mayer, 2009; Jiang et al., 1998; Mishima et al., 2004].

During early anaphase microtubules are rearranged to form the central spindle, a bipolar structure composed of microtubules with overlapping plus ends [Mastronarde et al., 1993]. The microtubules of the central spindle are bundled together by the activity of several proteins essential for cytokinesis such as Protein regulator of cytokinesis 1 (Prc1) [Zhu and Jiang, 2005] and members of the centraspindlin complex, Mitotic kinesin-like protein 1 (Mklp1) and the Rho family GTPase activating protein, CYK4 (also known as MgcRacGAP) [Mishima et al., 2002]. The activity of the centraspindlin complex also requires the localisation of the CPC to the midbody, and Aurora B is known to phosphorylate CYK4 [Ban et al., 2004]. Prc1,centraspindlin and the CPC are transported to the spindle midzone along microtubules and their transport to the midzone can be diminished by inhibitory phosphorylation by Cdk1 [Hummer and Mayer, 2009; Jiang et al., 1998; Mishima et al., 2004].

The central spindle is required for the correct positioning of the actomyosin contractile ring, which forms inside the plasma membrane at a cleavage plane designated by microtubule asters and the central spindle [Bringmann and Hyman,2005]. Contractile ring construction involves the recruitment and activation of RhoA to the cleavage plane. RhoA is activated by the Rho Guanine nucleotide Exchange Factor (GEF), Ect2, and RhoA localisation on the cleavage plane is dependent on Ect2 activity [Yuce et al., 2005]. Ect2 is recruited to the midzone by Plk1 phosphorylation of CYK4 in a Prc1 dependent manner [Wolfe et al., 2009]. RhoA activation causes reorganisation of the cytoskeleton through its effectors, such as the formins [Alberts, 2001], which cause actin polymerisation, and Rho-associated coiled-coil-forming kinase [Heselmeyer et al.], which phosphorylates regulatory myosin light chain (rMlc) and causes nonmuscle myosin II activation [Amano et al.,1996].

Active non-muscle myosin II causes the translocation of actin filaments and constriction of the contractile ring around the compacted microtubules of the central spindle, forming the cleavage furrow [Straight et al., 2003]. Cleavage furrow ingression results in the formation of a thin intercellular bridge between the two incipient daughter cells. The centre of which is the midbody. The midbody, or Flemming body, which was initially described by Flemming in 1891 [Paweletz, 2001]. As the name suggests its a ring like structure located at the centre of the intercellular bridge. It acts as a platform for recruitment of several proteins that are essential for the final stage of cytokinesis, known as abscission.

Following cleavage furrow ingression and the formation of the midbody the final step required for completion of cytokinesis is abscission of the intercellular bridge connecting the two daughter cells. The exact nature of the mechanisms required for the separation of daughter cells is not fully understood. However, several key components that are recruited to the midbody prior to abscission have been identified.

The centraspindlin complex, specifically Mklp1, is essential for recruitment of centriolin and centrosomal protein of 55kDa (cep55) [Gromley et al., 2005; Zhao et al., 2006]. Centriolin is a centrosomal protein recruited to the midbody following furrow ingression [Gromley et al., 2003] an it recruits proteins of the vesicle-targeting exocyst complex and vesicle-fusion soluble N-ethylmaleimidesensitive factor attachment protein receptor (SNARE) components, such as endobrevin and SNARE-associated protein snapin to the midbody [Gromley et al.,2005]. The exocyst and SNAREs are responsible for the fusion of vesicles delivered to the midbody prior to abscission and can generate membrane stalks that facilitate the final fusion event [Gromley et al., 2005]. Furrow ingression and abscission cause an increase in plasma membrane surface area. This is predominantly facilitated by membrane vesicle trafficking from recycling endosomes and the Golgi apparatus [Boucrot and Kirchhausen, 2007; Schweitzer et al., 2005].

GTPases involved in endocytosis, such as Rab11, Arf6 and Dynamin have been shown to play a role in cytokinesis [Fielding et al., 2005; Thompson et al., 2002; Wilson et al.,2005]. Rab11 is a recycling endosome protein that is required for recycling of vesicles to the plasma membrane [Ullrich et al., 1996]. Arf6 can interact with the exocyst complex subunit Exo7p and can bind Rab11-positive membrane vesicles through interaction with the family of Rab11 interacting proteins (FIPs) 3 and 4 [Fielding et al., 2005].

The delivery and fusion of membrane vesicles at the midbody is required for abscission, however the activity of vesicle trafficking and fusion proteins at the midbody occurs prior to membrane fission [Goss and Toomre, 2008;Gromley et al., 2005], which is thought to be mediated by the endosomal complex required for transport (ESCRT) machinery.

Cep55 recruits the ESCRT-I subunit,Tsg101, and the ESCRT associated protein, ALIX, to the midbody through direct interaction with its hinge region [Iwamori et al., 2010; Morita et al., 2007]. Cep55 binding of ALIX and Tsg101 results in the midbody recruitment of ESCRT-III subunits, Charged Multivesicular Body Proteins (CHMPs); CHMP4 and CHMP6,which are responsible for membrane fission [Carlton et al., 2008]. The ESCRT-III subunit CHMP1B and vacuolar protein sorting 4 (VPS4) are also recruited to the midbody and are required for abscission [Morita et al., 2010].

VPS4 ATPase activity removes the ESCRT proteins from the membrane allowing ESCRT-III to assemble at a new site on the membrane and carry out multiple rounds of membrane sorting [Babst et al., 1998]. CHMP1B has been shown to bind the microtubule severing ATPase spastin and recruit it to the midbody [Yang et al., 2008] and the interaction between spastin and CHMP1B is thought to co-ordinate microtubule severing, which is essential for membrane abscission [Connell et al., 2009]. Overall, while many of the components required for completion of cytokinesis have been identified, the exact mechanism for abscission remains poorly understood. 

Normal somatic cells are diploid, containing two identical sets of chromosomes, and they undergo DNA synthesis to become tetraploid. The 4N tetraploid DNA is then divided during mitosis to produce two diploid daughter cells. However, failure of cytokinesis can produce a single tetraploid cell [Ganem et al., 2007]. The oncogenic potential of tetraploid cells formed from failed cytokinesis lies in the fact that they reender the cell cycle with two centrosomes as opposed to one. Increased centrosome number can lead to multipolar mitoses and aneuploidy if the tetraploid cells are not prevented from further cycling [Margolis et al., 2003].

It was proposed that a checkpoint acts to prevent cycling of tetraploid cells in a p53 and pRb dependent manner [Andreassen et al., 2001]. This tetraploidy checkpoint reportedly prevents S phase entry in cells that are tetraploid leading to cell cycle arrest in a tetraploid G1state. The inactivation of p53 allows tetraploid cells to re-enter the cell cycle, and double their DNA content and centrosome number [Borel et al., 2002], which can give rise to a multipolar mitosis and the generation of aneuploid daughter cells that ultimately leads to tumourigenesis. Thus, it was proposed that tetraploidy acts as an important intermediate that can lead to aneuploidy and the onset of oncogenesis [Weaver and Cleveland, 2007].

Consistently, the failure to complete cytokinesis resulting in the formation of tetraploid cells has been reported to promote tumourigenesis [Fujiwara et al., 2005; Steigemann et al., 2009]. The existence of the tetraploidy checkpoint has been disputed [Uetake and Sluder, 2004; Wong and Stearns, 2005], however active p53 and pRb have been shown to contribute to the mechanism that arrests tetraploid cells in a G1-like state [Heilman et al., 2009; Liu et al., 2007].

Aneuploid cells contain an aberrant chromosome number. Aneuploidy causes genome instability, which is a hallmark of cancer [Hanahan and Weinberg, 2011]. There has been much debate as to whether aneuploidy is a cause or consequence of tumourigenesis [Hede, 2005]. Recently it has been proposed that aneuploidy can act in both an oncogenic and tumour suppressive manner [Weaver and Cleveland, 2009], whereby aneuploidy resulting in moderate genetic instability can promote tumour in genesis whereas aneuploidy causing extensive genetic instability results in tumour suppression and cell death. For example, aneuploidy induced due to CENP-E heterozygosity causes an increase in lymphomas of the

spleen and adenomas of the lung in vivo and increased transformation in vitro [Weaver et al., 2003; Weaver et al., 2007].

However, CENP-E-induced aneuploidy can inhibit the progression of tumours induced by the carcinogen 7,12-dimethylbenz(a)anthracene (DMBA) [Weaver et al., 2007]. As such, the generation of tetraploid cells as a result of failed cytokinesis has the potential to cause an oncogenic or tumour suppressive outcome depending on the extent of aneuploidy caused in the next cell division.