1 Introduction
During cell division, duplicated genetic material and other cellular components must be segregated before partitioning into two daughter cells [
1–
3]. The inheritance of chromosomes is under strict regulation and executed with high spatial and temporal control. However, how cell organelles respond to cell cycle signals and are dynamically orchestrated to ensure proper cell division remains unclear. Recently, with the help of technological advances, intensive studies found that various cell organelles actively participate in cell division through rapid changes in morphology and function to facilitate this process or at least not interfere with chromosome segregation. A plethora of proteins and dozens of distinct pathways are discovered to orchestrate organelle dynamics with cell division. Notably, small GTPases are found heavily involved in this progress (Tab.1). Whether there lies a common mechanism that orchestrates organelle dynamics with cell division and whether small GTPases are uniform mediators that execute this sophisticated process are worthy of investigation.
Small GTPases, a family of conserved proteins, are widely discovered to function as nodal points that integrate broad upstream regulatory inputs and disseminate broad effector outputs in distinct biological processes. Mostly residing in the cytosol, small GTPases (20–30 kDa) are homologous to the alpha subunit of heterotrimeric G-proteins (large G-proteins) [
4]. But unlike the alpha subunit of heterotrimeric G-protein, a small GTPase can independently hydrolyze GTP to form guanosine diphosphate (GDP) [
5]. Around 167 Ras superfamily proteins in human are divided into six prominent families on the basis of sequence identity and function: Ras, Rho, Rab, Arf, Ran, and Rag [
6–
8]. Small GTPases share a conserved conformational switch mechanism coordinating with guanine nucleotide exchange factors (GEFs) and GTPase-activating proteins (GAPs), which work in opposite ways to govern the cycle between two conformations of GDP-bound inactive and GTP-bound active states [
9–
12]. These mechanisms give small GTPases rapid on–off switches and subcellular location transfer characteristics. With abundant family members and various regulators and effectors, small GTPases are the perfect executor and adaptor for rapid structure reformation in coordination with cell division.
The potential mediator role of small GTPases that link cell division with organelle dynamics is further highlighted by its ubiquitous but distinct organelle regulation functions. RanGTP concentration gradient restores the subtle control of spindle formation [
13]. Rab and Arf families profoundly undertake intracellular vesicular transportation and the vesicle transportation processes continuously shaping ER and Golgi apparatus structures [
14–
20]. Recently, small organelles, which have long been considered disconnected from the genetic material division cycle but undergo stochastic partition, such as mitochondria, are found to coordinate with canonical cell division, and specific small GTPases emerge to be a potential mediator [
21].
In this review, we compare the role of small GTPases in cell shaping, spindle formation, membrane-bound organelle (MBO), and membrane-less organelle (MLO) structuring and highlight the mediator role of small GTPase in transducing cell cycle signaling to organelle dynamics within the cell cycle. Given that small GTPases are conserved in sequences and have related functions among subfamilies, we focus on events during mammalian cell division and use examples from other eukaryotic organisms.
2 Regulation of small GTPases during cell division
Eukaryotic cells undergo extensive and reversible cellular reorganization covering multiple dimensions among cell division. Given the dynamic roles of small GTPases during different stages of cell division, mechanisms for their spatial and temporal regulation are crucial to ensure orderly conduct. The regulation of small GTPase activity, localization, and function has been found among interaction with GAPs and GEFs. Small GTPases show high affinity toward GTP and GDP, and their GDP/GTP alternation actuates a variety of functional switches that are pivotal for cell dynamics (Fig.1 and 1B). GEFs and GAPs accelerate intrinsically sluggish GDP/GTP exchange and GTP hydrolysis speed by 10
5. For Ras, Rho, and Rab GTPases, this switch incorporates a membrane/cytosol alternation regulated by GDIs and GDI-like proteins. The post-translational modifications of small GTPases and their cofactors are also vital for their activation and function [
9,
22,
23].
Phosphorylation is a universal protein modification in eukaryotic cells and plays critical roles in every aspect of cellular life. The cell cycle is a leading example as protein phosphorylation peaks during mitosis. Mitotic kinases cyclin-dependent kinase (CDK, CDK1 and CDK2), Aurora (AURKA and AURKB), and PLK (Polo-like kinase) families and Bub1, Haspin, and Mps1 drive the entry and progression through the division phase of the cell cycle [
24]. Notably, the reversible phosphorylation of small GTPases, GEFs, and their GPAs is an essential regulatory mechanism that occurs through cell division. At the beginning, a rise in CDK1 activity activates RhoA and cell rounding by phosphorylating p190RhoGAP and Ect2 for mitotic entry. The bridge scission at the last step of cell division requires AURKB-mediated MgcRacGAP phosphorylation. In intermediate steps, the phosphorylation of small GTPases facilitates spindle formation and organelle dynamics. The important new findings of small GTPases related to spindle formation and organelle dynamics are described in the following sections.
Besides phosphor regulation, ubiquitination-directed proteolysis ensures the unidirectional progression of the cell cycle [
25–
27]. Interestingly, the ubiquitination of multiple small GTPases modifies their activities, relocalization, and degradation [
22,
27–
29]. For example, a component of the E3 ubiquitin ligase parkin-dependent ubiquitination on conserved K38 residue enhances RILP–Rab7 binding (RILP is a Rab7-interacting lysosomal protein), thus regulating Rab7 activity [
30]. HECT type E3 ubiquitin ligase Smurf1 regulates cell polarity and motility through targeting RhoA for ubiquitination and degradation [
31]. Besides small GTPase itself, GAPs, GEFs, and GDIs are frequent targets of the ubiquitination machinery [
28]. Further studies will be necessary to test the significance of small GTPase ubiquitination during mitosis.
Other modifications include prenylation by a prenyltransferase on the cysteine, and the methylation of the C-terminal cysteine by isoprenylcysteine carboxyl methyltransferase. Prenylation isoprenylcysteine is required for membrane targeting and localization of small GTPase [
22,
32,
33]. These modifications may not be as easy or fast as phosphor or ubiquitin regulation in shifting small GTPase activity when confronting rapid cell division events. Still, their roles in cell division are also worth exploration.
3 Spindle apparatus
The most significant phenomenon of eukaryotic cell division is the establishment of microtubule (MT)-based bipolar spindle apparatus. The cellular spindle apparatus includes three main components, i.e., spindle MTs, condensed chromosomes, and centrosomes. These three parts coordinate with associated proteins to guarantee chromosome inheritance [
34,
35].
Dozens of small GTPases covering all five subfamilies (except Rag) localize along spindle MTs, chromosomes, around centrioles, and between kinetochore–MT contacting sites and directly mediate spindle formation, position, orientation, and proper function.
Ran, as the single member of its subfamily, functions in centrosome duplication, MT dynamics, chromosome alignment, kinetochore attachment, and nuclear envelope dynamic regulation [
13]. As for the spindle apparatus, Ran regulates the life cycle of centrosomes, and the Ran gradient directs spindle MT organization and assembly [
36]. In mitotic cells, PLK1 and PAK4 phosphorylate Ran protein on S135 [
37,
38], and PLK1 phosphorylates Ran-binding protein 1 (RANBP1). Phosphor-activated Ran and RANBP1 modulate Cyclin B1’s function at the centrosome to facilitate MT nucleation and stabilization [
36,
39].
For the Rho subfamily, Rho GTPases are central modulators of cell morphogenesis through the remodeling of the cytoskeleton [
40,
41]. The role of Rho GTPases that marks cytokinesis in animal cell division has now extended to the metaphase in coordinating mitotic spindle formation and the bi-orient attachment of spindle MTs to the kinetochore. This subject has been extensively reviewed [
42]. Cdc42 is a master regulator of cell polarity. During metaphase, Cdc42 localizes to the spindle and centrosomes, facilitating spindle orientation and formation for chromosome alignment [
43]. Cell division signals are transmitted to Rho GTPase and cofactors by phosphorylation modification via AURKA/B and CDK1 kinases [
42].
Rab and Arf proteins link membrane transport and cytoskeleton dynamics [
17]. During spindle formation, endosomal transport facilitates centrosome maturation, bipolar spindle formation, orientation, positioning, and function. The spindle-associated Rabs have been extensively reviewed [
44]. Briefly, several Rabs and interacting proteins associate with spindle apparatus during cell division [
45]. Rab5- and Rab11-positive endosomes localize on the mitotic spindle and at mitotic spindle poles (Fig.2). Rab5, in association with dynein, regulates astral MT size and spindle alignment. Rab5/CENP-F form complexes adjacent to kinetochores, and the silencing of Rab5 increases aberrant and unattached kinetochores, which may be due to reduced CENP-F at kinetochores. Mitotic spindle poles are active sites in vesicular transportation. Rab11-positive endosomes are reported to be involved in spindle pole organization and orientation by bringing MT-nucleating factors and spindle pole proteins to poles [
46,
47]. For the Arf family, Arl8A and Arl8B bind to tubulin and localize with MTs on the spindle midzone in late mitosis and may function in maintaining centrosome integrity [
48,
49].
Ras GTPases are known as signaling nodes but do not directly participate in spindle structuring. However, a Ras subfamily member RalA is proven to have functional interaction with a chromosomal passenger complex (CPC) [
50]. RalA and its CPC-localized GEF TD-60 contribute to the regulation of kinetochore–MT interaction in early mitosis. The loss of RalA causes AURKB activity loss at the centrosome and increases the interkinetochore stretch with perturbed spindle structures [
50].
4 MLOs
MLOs belong to an unconventional class of organelles formed by phase separation but are not sealed with a phospholipid membrane [
51–
54]. Recent discoveries on phase separation revolutionized our understanding of MLO organization, molecular properties, physical features, and biological functions. DYRK3 kinase is found to diffuse MLOs at mitotic onset [
55], extending the phase separation concept to cell division [
52]. However, the dynamic regulation of most MLOs during mitosis except pericentriolar material (PCM) is largely unknown. PCM is a typical example of MLOs surrounding centrioles, which nucleate MTs for mitotic apparatus establishment [
56,
57]. Previous work demonstrated that Ran regulates multiple steps of the centrosome life cycle [
36]. Briefly, Ran interacts with the PCM scaffolding protein AKAP450 at centrosomes and regulates mitotic MT nucleation and stabilization [
58,
59]. Mitotic kinases PLK1 and PAK4 phosphorylate Ran at S135 and switch Ran into Ran–GTP and RanBP1-binding state to promote MT nucleation [
37,
60].
Although MLOs are free of lipid membrane, membrane and factors regulating membrane trafficking steps are emerging as important components of MLO formation and function [
52]. In yeast, small GTPase Arf1 mutation [
19] and other secretory mutants promote stress granule (P-body) accumulation under stress, which are related to P-body secretion malfunction [
61]. Membrane traffic-associated small GTPases Rab5, Rab6′, and Rab11 are found on endosomes around spindle poles and contribute to spindle pole maturation [
46,
62–
65]. These findings suggest a major role for the endocytic pathway and small GTPases in MLO regulation.
5 MBOs
Eukaryotic cells organize their cytoplasm into multiple compartments to accommodate increased cell size and life process complexity. The compartments that are surrounded by biological membranes are MBOs. In terms of cell division, organelles may utilize different strategies for reorganization and partition [
1,
2], and small GTPases are widely found to be involved in this process.
6 Plasma membrane
Cells undergo dramatic shape changes to facilitate genetic material partition and the formation of daughter cells [
2,
66,
67]. The plasma membrane and its associated cytoskeleton underneath support massive shape changes from cell rounding to abscission at the end of cytokinesis. The plasma membrane lipid composition dramatically changes to facilitate cell division especially during cytokinesis [
68,
69]. Small GTPases especially Rho GTPases control cell shaping through their ability to modulate MT dynamics and the actin–myosin cytoskeleton.
6.1 Cell rounding
To enter mitosis, eukaryotes use a rounding process to meet geometric requirements and facilitate chromosome capture and spindle function [
70]. Focal adhesions (FAs) that link the extracellular matrix with cell skeleton must dismantle to ensure cell rounding. A key regulator in FA dismantling is the Ras subfamily small GTPase Rap1 [
25]. Rap1 inactivation facilitates FA disassembly by preventing the Rap1 effector RIAM complex formation and the integrin activator talin interaction (Fig.2) [
71,
72]. FA dismantling allows G2 cells to pass through the “adhesion-dependent checkpoint” and entry to mitosis. FA dismantling inactivates the “adhesion-dependent checkpoint” through the DEPDC1B/RhoA/PTPRF axis. Briefly, DEPDC1B accumulates at the G2 phase, and by competing with RhoA for the interaction with PTPRF, DEPDC1B frees RhoA from PTPRF and RhoA nucleotide exchange factor GEF-H1 [
73]. This mechanism promotes FA disassembly as cells enter prophase. The released RhoA then becomes available for ECT2-dependent activation at other sites of the PM and promotes actomyosin cortex formation in prophase. However, during this process, how the cell cycle signal is transduced to Rap1 for FA dismantling is still unknown.
Cell rounding in a confined environment
in vitro needs cytoskeleton disassembly and reorganization. The cytoskeleton forms a mitotic actomyosin network underneath the plasma membrane, increasing cortical rigidity and providing attachment sites for astral MTs to facilitate spindle position and function [
70,
74,
75].
RhoA is a key element in the formation of the actomyosin cortex. At mitotic entry, ECT2 released from the nucleus activates RhoA. CDK1 and PLK1 phosphor activate ECT2 and increase its GEF function toward RhoA. Activated RhoA induces the retraction of the cell margin, generates nonbranched actin filaments through DIA, and activates myosin via ROCK [
76,
77].
6.2 Cell elongation and cytokinesis
After sister chromatins segregate, cells elongate and form a cleavage ingression furrow to finalize cell division. This stage of cell shape change uses the same set of proteins as in cytokinesis, ECT2, RhoA, DIA, actin, myosin, and anillin [
78,
79]. The drop of CDK1–Cyclin B1 activity at this stage of mitosis shifts RhoA localization, thus starting repolarization via GAP and GEF MgcRacGAP midzone localization. RhoA and Rac1, another Rho subfamily of small GTPase, establish a gradient for actomyosin cortex polarization.
Rho and Rab GTPases have well-established functions in cytokinesis, share lipid PTMs, and show specific lipid preferences. As the plasma membrane lipid composition dramatically changes in the cell cycle [
68], their localization and interaction with regulators and effectors may also be affected. During anaphase, PtdIns(4,5)P2 and PtdIns(3,4,5)P3 enrich in the furrow and poles separately, and form a lipid gradient, leading to RhoA/Rac activity gradient for the repolarization of cortical actomyosin establishment [
69,
80]. The ingression process transforms the cell into a dumbbell shape with apparent surface area expansion.
6.3 Polar relaxation
A stable ingression furrow’s position and function also require a reduction of contractility at polar cortex regions. Several signals may contribute to the relaxation of the cortex at opposing cell poles [
67]. Among these regulating pathways, the establishment of a high Ran–GTP gradient at cell poles may be associated with polar relaxation. As chromosomes approach cell poles, the associated RCC1 activates Ran and then functions via importin β to regulate the activity and localization of a wide range of actin regulators [
81–
83].
6.4 Mitosis to interphase transition
Finally, cells process mitosis-to-interphase transition and reorganize interphase structures in new cells. With the degradation of Cyclin B1, protein phosphatases PP1 and PP2A direct lamin and/or nucleoporin dephosphorylation to re-establish the nuclear envelope (NE) boundary [
84]. Furthermore, proteins, such as anillin, RacGAP1, and ECT2, released from the nucleus to facilitate mitotic cell structure are again isolated from the cytoplasm to re-establish interphase cell organization. The cell–matrix interaction is acquired through Rap1 reactivation, promoting the assembly of cell–substrate adhesion [
25].
7 Endoplasmic reticulum (ER) and NE
7.1 Morphometric changes to the ER during division
The ER is a single MBO involved in critical cellular processes, including protein processing, lipid synthesis, and calcium homeostasis. Interphase ER is a contiguous MBO with distinct subdomains that vary in structures and functions [
85–
89]. During cell division, ER morphology shows dramatic reorganization, and its reconstruction is required for mitotic fidelity [
90]. The maintenance and remodeling of ER morphology are dependent on a number of proteins, including membrane-bending protein reticulons (RTNs), the dynamin superfamily member atlastins, and cytoskeleton [
91–
93]. In addition, certain Rab GTPases, i.e., Rab10 and Rab18, mediate the interaction between the ER and motors, thus regulating tubular ER membrane dynamics [
94,
95]. The functional characterization of Rab GTPases in
C. elegans reveals that the depletion of Rab5 phenocopies the ER structure from the depletion of the Rtn4a and DP1/NogoA with reduced ER sheets [
96]. A recent study showed that ER-shaping proteins REEPs are required to preserve tubular ER structure during mitosis. The depletion of REEPs reduces ER curvature and increases sheet-like ER proportion [
97]. REEP3 and REEP4 are also needed to clear ER from metaphase chromatins, preventing their premature association to ensure proper division and NE architecture [
98]. How cell division signaling connects ER-shaping proteins, such as REEPs, and how these proteins respond remain an open question (Fig.2).
7.2 NE breakdown (NEBD) and assembly
The NE is a part of the ER network that is shaped into a large bimembrane cisterna consisting of outer (ONM) and an inner (INM) nuclear membranes, which together isolate and protect chromatin from the cytoplasm environment during interphase. Following the CDK1 activity cycle, NE breaks down during prometaphase, and the reverse process re-establishes this barrier during mitotic exit. Both processes require the dynamic remodeling of the ER with cytoskeleton. This subject has been reviewed extensively [
99–
102].
During interphase, the INM is widely tethered to chromatin and a meshwork of intermediate filaments [
102,
103]. CDK1-mediated phosphorylation drives NEBD via cytoskeletal contact disassociation and nuclear lamina dissolution [
1,
2,
104–
107]. Thus, NE is allowed to disengage from chromatin, and NE proteins redistribute into the mitotic ER. However, previous work suggested that ER membrane remodeling Rab5 is critical for NEBD [
108]. Rab5 is a Rab subfamily small GTPase that functions in endocytosis and early endosome fusion [
45,
109,
110]. In the
C. elegans embryo model, the depletion of Rab5 reduces ER clusters and thickens tubules similar to the depletion of YOP-1/RET-1. The Rab5 depletion inhibits NE disassembly and generates total embryonic death. A previous study showed the conservative role of Rab5 in shaping ER in HeLa cells [
108]. This study proposed a connection between ER and NEBD, and the peripheral ER structure is required to facilitate NE component diffusion into ER because the disruption of ER structure inhibits NEBD. This model on Rab5 leads to an interesting assumption that other small GTPases that regulate ER structure may also function in NEBD or NE reassembly. Another factor in NEBD is the mechanical force mediated by MTs and actin filaments that rupture membranes with lamina, thereby assisting NE disassembly [
111–
116]. This process also involves Ran activities in regulating MT dynamics during NEBD [
114]. To comply with interphase functions, cells undergo another dramatic reorganization after division. One of the most significant changes during mitotic exit is the reformation of the NE [
117]. First found in yeast, Ran has been identified as a key regulator in NE formation in several eukaryotes [
118–
122]. At telophase, Ran in its GDP-bound conformation binds with chromatin surface to recruit RCC1, which subsequently activates Ran to a GTP-binding state. The spatial organization of RanGDP/RanGTP promotes the assembly of NPCs and vesicle fusion via interactions between RanGTP and importin-β to form NE with interphase structure and functions [
123]. The COPII vesicular formation requires Sar1 and its effector Sec23/24 and Sec13/31 [
124]. The effector Sec13 is also found in Nup84p complex in organizing nuclear pore complex formation [
125]. A recent study in yeast found that Cdc42 regulates ESCRT disassembly and functions in NE sealing and ER fission after division [
126].
8 Golgi apparatus
The Golgi apparatus is a central station for post-translational modifications and secretion. The quantitative proteomics analysis in rodent and human samples identified 348 Golgi and associated proteins, in which Rab and Arf with their associates make up 17% of total candidates [
127]. Many other small GTPases from Ras and Rho subfamily show transient or permanent localization on the Golgi apparatus. Small GTPases are also actively involved in Golgi life cycles. In mammals, the Golgi apparatus is a continuous membranous system composed of stacks, and each stack has a
cis entry face and a
trans exit face connected by tubular bridges, i.e., the “Golgi ribbon” [
128]. The behavior of the Golgi apparatus during cell cycle is recorded in mammalian cells. From the G1 phase to the S phase, the Golgi mass increases significantly. During cell division, the Golgi apparatus undergoes extensive fragmentation to assist its correct partitioning and inheritance. The mitotic phosphorylation of Golgi reassembly stacking proteins (GRASPs) by mitotic kinases at the C-terminal serine/proline-rich (SPR) inhibits their oligomerization and results in the separation of the Golgi cisternae [
129]. After cell division, fragments undergo a reverse series of changes and rapidly re-establish the Golgi ribbon and polarization in the two daughter cells [
130–
132].
8.1 Golgi fragmentation and Golgi checkpoint
The Golgi apparatus fragmentation begins at the late G2/early prophase when the connected Golgi ribbon breaks into individual stacks. At metaphase, stacks further go through unstacking and vesiculation to form Golgi clusters [
1,
2]. During this process, a key step for Golgi sequential disassembly is the inactivation of small GTPases Arf1 and Sar1 that operate at the ER/Golgi interface (Fig.2) [
130]. Arf1 at the Golgi apparatus is inactivated by its GBF1, which is phosphor-inactivated by AMPK and CDK1 in mitosis [
96,
133]. The inactivation of Arf1 contributes to Golgi disassembly through the inhibition of Arf1 effector recruitment to membranes. COPI vesicles have been shown to participate in the retrieval of proteins from the Golgi back to the ER. The inactivation of Arf1 results in the release of COPI from Golgi and Golgi dispersion. The Arf1 effector Golgin160, the connector of Golgi membranes with dynein motors, is released from Golgi membranes as a direct consequence of Arf1 inactivation in mitosis, leading to loss of dynein from these membranes. This phenomenon correlates temporally with the Golgi apparatus becoming fragmented [
134]. At ER exit sites, CDK1 mediates Sar1 inactivation during prophase and results in a reduction in membrane and protein export from the ER [
135,
136]. The active Sar1 GTP–GDP binding cycle recruits COPII coat proteins and other effectors, maintaining the ER export domain and supporting Golgi dynamic structuring. When expressing Sar1 T39N GDP-restricted mutant, ER export sites disappear, Golgi integral membrane proteins are trapped in the ER, and Golgi peripheral proteins are retargeted to the ER or cytoplasm, thus leading to the disappearance of Golgi or Golgi-like structure within cells [
137–
139].
Research on the mechanisms of Golgi fragmentation led to the striking discovery that the inhibition of the Golgi disassembly process induces a potent and persistent G2 block of cell cycle progression [
130,
140]. The blockage of Golgi fragmentation reduces AURKA activation and centrosome localization, which sequentially blocks G2/M master regulator CDK1–Cyclin B1. Therefore, the inheritance of the Golgi complex is actively participated in cell cycle and acts as an organelle-based cell cycle checkpoint, i.e., the Golgi checkpoint in monitoring proper organelle partition [
141].
8.2 Reassembly of Golgi complex
In contrast to fragmentation, Golgi reformation occurs through the sequential reactivation of Sar1 and Arf1. The Sar1 reactivation re-establishes ER export domains first by recruiting Rab1, p115, SNAREs, and matrix proteins, which support its differentiation into ERGIC [
142]. The activity of Arf1 and dozens of its effectors further differentiate newly formed membrane structure to morphologically and geographically specialized Golgi membranes distinct from ER [
143]. The Golgi reconstruction after BFA treatment shows that the Rab GTPase-regulated targeting fusion pathways are required in Golgi recovery and
cis-to-
trans polarity [
18]. The Rab-mediated Golgi polarization may also apply to Golgi reassembly in the cell cycle. In late mitosis, each daughter cell reassembles a pair of twin Golgi complexes at two distinct locations: a smaller cluster near the midbody and a bigger juxtanuclear compartment. Arf-associated GEF GBF1 phosphorylation and degradation allow a midbody-located cluster to migrate along the cell skeleton to the other twin, forming a single perinuclear Golgi ribbon. Expressing nondegradable GBF1 or treating with GBF1 inhibitor BFA blocks cytokinesis and Golgi reformation [
133,
144].
Other Golgi-associated small GTPases have also been shown to have links to Golgi-based processes and mitotic progression. Rab6A′ regulates retrograde transport from late endosomes via the Golgi to the ER through its interaction with p150Glued and GAPCenA and participates in the transition from metaphase to anaphase during mitosis [
145]. Together, these findings raise the possibility that the activities of these GTPases are relevant to the sophisticated process of mitotic Golgi dynamic regulation and constitute an organelle-based checkpoint participating mitotic entry, metaphase-to-anaphase transition, and cytokinesis.
9 Mitochondria
Mitochondria are considered semiautonomous organelles, and their growth and fission–fusion cycle can occur independently of the cell cycle [
146]. However, recent studies proved that the mitochondria fission–fusion cycle and its distribution are coordinated with cell cycle (Fig.2) [
2,
147]. Mitotic kinases CDK and AURKA regulate mitochondrial fission–fusion cycles by the phosphorylation of dynamin-related GTPase proteins (DRPs) [
148,
149]. AURKA induces the phosphorylation of RalA (a Ras subfamily member), promoting its effector RALBP1 to mitochondria during mitotic entry. RALBP1 then serves as a scaffold for CDK1-dependent activation of DRP1 [
21]. Cells devoid of RalA show decreased cellular ATP levels and disrupted cytokinesis. Such an effect may be the consequence of impaired energy production as the fission blockage disrupts ATP production [
150,
151]. Recently, the last step of mitochondrial division has been shown to depend on the recruitment of the trans-Golgi network PI(4)P-containing vesicles at the ER–mitochondria contact site by Arf1 and its effector PI(4)KIIIb. The loss of Arf1 or PI(4)KIIIb prevents PI(4)P generation and leads to a hyperfused and branched mitochondrial network, which may be through disrupting the Arp2/3-dependent actin polymerization [
152].
The fragmentation of the mitochondria helps stochastic mitochondrial inheritance. However, this phenomenon may not be the only factor because artificially disrupted cytoskeleton-mediated mitochondrial distribution leads to asymmetric cell division and mitotic defects [
153,
154]. The Arf1 and Rho subfamily small GTPase mitochondrial rho GTPase (Miro) oppositely regulates mitochondrial retrograde and anterograde movement [
155]. During cytokinesis, mitochondria interact with growing MT tips and move toward the cleavage furrow. CENP-F (cytoskeletal-associated protein), which associates with MT-growing tips, is recruited to mitochondria by directly interacting with Miro1 and Miro2 via its C-terminal highly conserved domain [
154]. Miro and CENP-F interaction peak at S/G2 and cytokinesis, remain low at mitosis. This phenomenon indicates a dynamic interaction and regulated process. The knockdown of Miro decreases the spreading of the mitochondrial network and shows cytokinesis-specific defects. In the opposite direction, active Arf1 and its GEF GBF1 physically interact with Miro and block Miro-dependent retrograde mitochondrial transportation [
155]. The specific localization of the mitochondria is regulated to meet local energy demands [
156]. Miro depletion induces mitochondria misallocation and cell division failure. These results suggest that the mitochondrial distribution coordinates with cell division. Altogether, small GTPases’ function in the coordination of power supply, fission–fusion cycle control, and cytosolic distribution suggests a binary regulation connecting mitochondrial inheritance with the powering of cell chromosome partition. Cell cycle signals are predicted to implement on the Miro GTPase domain or CENP-F Miro-binding peptide, where serine and threonine residues are phosphorylated during the cell cycle [
154,
157–
159]. GBF1 is inhibited through phosphorylation by AMPK, and the degradation of GBF1 is required for the inheritance of the Golgi apparatus and cytokinesis [
133,
144]. Thus, GBF1 may coordinate the Golgi apparatus and the mitochondrial network reorganization during cell division.
10 Midbody and membrane ingression
Midbody, a transient organelle, links two nascent daughter cells that form after furrow ingression during cytokinesis [
160]. The cleavage furrow ingression compacts antiparallel midzone bundles into a 1–1.5 µm diameter structure, i.e., the midbody [
161]. The midbody is rich in MT, MT-interacting proteins, and membrane-based proteins as its formation and final abscission require the cooperation of the cytoskeleton with the membrane system [
162,
163]. The spatial and temporal coordination of the Rho family GTPase with mitotic kinase and downstream effectors ensures accurate ingression, midbody position, and final abscission [
42,
164–
167]. Recent studies highlighted the role of membrane and membrane trafficking during this process.
At the heart of cytokinesis in animal cells is the centralspindlin complex. The centralspindlin complex comprises two proteins: the kinesin-like protein Mitotic kinesin-like protein 1 and the Rho GTPase-activating protein (RhoGAP) CYK-4. Through its targeted localization to a narrow region of antiparallel MT overlap immediately following chromosome segregation, the centralspindlin initiates central spindle assembly. During cell division, the centralspindlin has several critical functions, including positioning the division plane, regulation of Rho family GTPases, and midbody assembly and abscission.
10.1 Position of ingression furrow
An ingression furrow is established immediately after chromosome segregation at the location where the chromosomes once localized during the metaphase to finalize cell division [
79,
160,
168]. How cells precisely get the ingression furrow into position has drawn wide attention for decades as its functions in symmetric and asymmetric cell division and cell fate determination. The Rho small GTPase marks the narrow zone for ingression furrow as RhoA accumulates to the equatorial cortex ahead of furrow formation [
169–
171]. The activity of the equator cortex that recruits RhoA is strictly regulated by Rho GEFs (ECT2, GEF-H1, and Myo-GEF) and GAPs (MgcRacGAP, p190RhoGAP, and MP-GAP). ECT2 localizes to spindle midzone by interacting with the centralspindle complex, which depends on MgcRacGAP phosphorylation by mitotic kinase PLK1 [
161]. Thus, Ect2 is stabilized in an active conformation to interact with RhoA. A detailed experiment in human cells shows that ECT2 membrane-binding is required and sufficient for cytokinesis [
172].
10.2 Membrane ingression
Membrane ingression and invagination of the cell membrane are driven by a contractile ring composed of paralleled actin filaments and nonmuscle myosin II [
173,
174]. RhoA regulates the contractile ring in multiple aspects. RhoA stimulates actin polymerization and regulates myosin activity and other scaffold proteins. RhoA is a key regulator that controls actin polymerization by reliving formin autoinhibition [
175,
176] at the ingression position [
169,
171]. Formin proteins harbor two proline-rich forming homology domains, i.e., FH1 and FH2 [
177], that control actin polymerization [
177–
180]. The mammalian formin mDia1 selectively binds with active RhoA and promotes actin polymerization [
176] together with another actin cytoskeletal regulator profiling [
180]. RhoA regulates myosin by activating Rho-kinase (ROCK) to phosphorylate the regulatory light chains of myosin-II. Three myosin-regulating kinases (and MLCK) are found to phosphorylate myosin at myosin regulatory light chain (MRLC) to activate myosin. ROCK and possibly Citron, but not MLKC, depend on RhoA to acquire the correct localization at the ingression site. The ROCK recruited by RhoA regulates myosin through phosphorylating MRLC at S19 [
181]. MLC phosphorylation is regulated in part by myosin phosphatase (MYPT1), which can be phosphorylated at T853 by ROCK. ROCK-directed MYPT1 phosphorylation reduces phosphatase activity, thus maintaining myosin activity [
182–
184]. As for Citron kinase, which is believed to be activated by interacting with active RhoA and promoting the constriction of the actomyosin ring just like ROCK, it is later proven to be unnecessary for ingression but acts as an organizing center for the midbody [
42,
185,
186]. Study on
Drosophila and human cells show that Citron kinase is required for correct RhoA localization at cleavage, thus indicating its function as a regulator but not effector of RhoA. RhoA also regulates scaffold proteins anillin and septins in coordination with ingression besides actin and myosin II [
42]. In contrast to RhoA, the other Rho subfamily member Rac1 acts as a negative regulator for contractile ring constriction [
42]. Rac1 inactivation by MgcRacGAP abrogates MLCK inhibition by PAK1/2 and ARHGEF, thus sustaining myosin activity. Furthermore, the Rac1 inactivation prevents branched actin filament nucleation by inactivating the Arp2/3 complex, thus confining a narrow zone for proper actin–myosin function [
187].
As mentioned before, cell surface area expansion during cytokinesis requires the transportation of membrane vesicles to be inserted at the cleavage site [
188–
190]. Likewise, the actomyosin ring function requires the timely delivery of regulatory proteins and remodeling factors required at different stages of this process. In
Drosophila spermatocytes, Rab1 (
omt) mutation disrupts cytokinesis. Rab1 organizes the Golgi-dependent secretion through the GOLPH3 complex and controls membrane trafficking to facilitate new membrane supplementation along the ingressing cleavage furrow [
191]. Rab11 mutation, which disrupts the recycling endosome, also causes furrow ingression defects in a
Drosophila model by regulating the rate of actin polymerization through the RhoGEF2–Rho1 (Rac1 in human) pathway [
192]. Arf1 localizes to the Golgi apparatus and plays a central role in regulating the Golgi apparatus material flow. In the
Drosophila model, an Asap–Arf1–Golgi pathway appears to fuel cleavage furrow ingression. Asap dominates Arf1 accumulation at Golgi for proper Golgi organization and the Golgi output further fueling ingression furrow [
193]. The widespread participation of small GTPases around ingression sites together with their functions in membrane and cytoskeleton dynamics suggest that cytokinesis relies on membrane traffic to coordinate cytoskeletal elements and lipid-based membranes [
194–
197].
10.3 Abscission
The final step of cell division is the abscission that physically separates one linking protoplast into two independent daughter cells [
78,
162]. Abscission requires the disassembly of cortical actin in the intercellular bridge [
198]. Similarly, this process is driven by RhoA inactivation, whereas Aurora B phosphorylates Rac GAP MgcRacGAP at S387 [
199]. Several Rabs and Arfs are involved in this cell cycle step as vesicle trafficking dominates cytokinesis [
200,
201]. Rab11 and Rab35 endosomes ensure low F-actin levels required for scission at the abscission site [
194]. Rab11 acts as an upstream regulator for RhoA inactivation in bridges. Rab11–FIP3 endosomes transport RhoA GAP p50RhoGAP to inhibit RhoA activities on actin polymerization. Rab11 or FIP3 depletion leads to F-actin accumulation at the bridge, delaying ESCRT-III filament assembly at the abscission site and abscission [
198]. In addition, Rab35 endosome transports OCRL to reduce PtdIns(4,5)P2, which is required for F-actin polymerization [
202]. In parallel to reducing polymerization, Rab35 also increases F-actin depolymerization via recruiting oxidoreductase MICAL1 at the intercellular bridge, which oxidizes methionine residues of F-actin and speeds up its disassembly [
203]. Rab11 and Rab35 are well-established regulators in this process, and other small GTPases, such as Arf6, RalA/RalB, and several Rabs, are also present within midbody and bridges. Briefly, Arf6 forms a complex with the MKLP1 kinesin of the centralspindlin complex and concentrates at the midbody for bridge stability and abscission [
65,
204]. RalA and RalB control exocyst localization at the furrow and midbody, respectively, and their collaboration is required for abscission completion [
205,
206]. RNAi screening found that several Rab GTPases may play a modulatory role in abscission [
207].
The cooperation between cell skeleton and vesicular traffic is a driving force behind cytokinesis. Plasma membrane lipids bring new perspectives into cytokinesis regulation. Lipid composition at furrow and bridge changes dramatically and is essential for ingression and abscission regions during cytokinesis [
68,
69,
208]. PtdIns(4,5)P2 stabilize the bridge. Thus, for proper abscission, the PI5 phosphatase OCRL-mediated hydrolysis is required at the abscission site [
209–
211]. As previously mentioned, Rab35 endosomes drive OCRL’s transportation and bridge localization [
202].
11 Autophagosomes
Autophagosomes are double-membrane vesicles functioning in the lysosome-dependent degradation process. This process intersects with continuous and dynamic vesicle transportation, requiring small GTPases. For example, Rab and Arf GTPases regulate endocytic membrane trafficking pathways. Briefly, the Rab subfamily, Rab1, Rab5, Rab7, Rab9A, Rab11, Rab23, Rab32, Rab33B, and Sec4 function in autophagosome formation [
212–
215]. Rab7, Rab8B, and Rab24 regulate maturation [
213,
214,
216–
219]. The Arf family Arf1, Arf6, Arf8, and Sar1 determine its biogenesis and cellular localization [
213,
220]. In addition, other small GTPases outside Rab or Arf family, such as RalB, are involved. RalB regulates mammalian Atg1 ortholog (ULK1)-mediated formation of autophagosome [
221], indicating that small GTPases are widely involved in autophagosome dynamic regulation. However, whether autophagy is active during division is still debated. A recent study proposed that nutrition status is uncoupled with autophagy because the interaction between lysosome resident Rag GTPase (a small group of Ras superfamily small GTPase [
222,
223] functioning in mTOR activation [
224,
225]) and RAPTOR is blocked by CDK1–Cyclin B-mediated phosphorylation [
226]. This phenomenon may extend small GTPases from morphological regulation to biochemical processes. Current evidence also suggested that the phosphoregulation of small GTPase at mitosis entry relocates small GTPase and their cofactors, which may limit their autophagic membrane trafficking ability. The exquisite regulation of degradation may be required to avoid the unplanned degradation of exposed genetic content [
2]. Further research on how these autophagosomes associated with small GTPases are functioning in mitosis and whether it is similar to that in interphase will provide knowledge regarding the mechanisms underlying autophagosome cell cycle regulation and help us understand the potential crosstalk between autophagosomes and cell division.
12 Lipid droplets
Lipid droplets are cytoplasmic lipid-rich storage organelles covered by a phospholipid monolayer [
227–
229]. Lipid droplets are important for membrane expansion and energy generation during cell division, thus its biogenesis and turnover may be carefully regulated along with the cell cycle [
68,
69,
208,
230]. In fission yeast, the number of lipid droplets increases during G2 phase and remains relatively constant after G2. In budding yeast, the collection of excess lipids into droplets is vital for successful cytokinesis [
208]. A recent study observed the rapid dispersion of lipid droplets in mammalian cell lines during S phase [
231]. However, no systematic analysis of lipid droplet distribution or functions during division is observed. Similar to autophagosomes, lipophagy and lipid droplet biogenesis together with its localization depend on Rabs and Arfs [
232–
234]. Thus, research on the mitotic regulation of lipid droplet-associated GTPases may give further insights into their binary coordination.
13 Peroxisomes
Peroxisomes are single MBOs involved in catabolism and reduction of reactive oxygen species [
235]. Peroxisomes cluster around spindle poles and partition into daughter cells in a spindle pole-dependent manner in mammalian cells [
236,
237]. The symmetric distribution of peroxisomes coordinates with spindle orientation, and thus regulates cell growth/differentiation balance during epidermal development [
238]. An important goal of future research is to identify the molecules that govern peroxisome distribution in coordination with cell division. Small GTPases, like Rho, Rab, and Arf proteins, are the main regulators connecting membranes to cytoskeleton motors or adapters, potentiating their functions in this process [
17,
239]. In interphase, Rho GTPase Miro1 functions as an adaptor for MT-based peroxisome motility [
240]. The binding of peroxisomes to MTs and actin is regulated by RhoA activities [
241]. In addition to its distribution, peroxisome biogenesis, growth, and fission are tightly associated with small GTPases [
242] (27 small GTPases identified in rat liver peroxisomes [
243]). Further research is required to determine the role of small GTPases in regulating peroxisome dynamics in cell division.
14 Conclusions and perspectives
14.1 Small GTPases coordinate organelle reorganization with cell division
Unified regulation and cooperation between heterogeneous processes is the prerequisite and foundation of exquisite life activities. Small GTPases are identified as nodal points of cell signaling pathways. When we consider the processes of organelle dynamics in cell division, small GTPases also show extraordinary molecular features and functional properties to fit such sophisticated orchestration. First, small GTPase localization is highly compartmentalized [
16,
244–
246]. Sufficient protein species enables various organelles or specific organelle domains to have a unique complement of small G proteins [
6,
247]. Second, small GTPases are liable to PTMs. Thus, they can receive cell cycle signals and transduce them to specific effectors, thus linking organelle dynamics to cell division [
22,
248]. Third, due to their tunable nature, small GTPases can shift between on and off states and translocate among locations to complex with diverse effectors that perform various functions, thus trimming organelles [
9–
11,
248,
249]. Additionally, highly related small GTPases can have subtle differences in their binding affinities for the same effector protein, thus helping us gain valuable insight into the fine-tuning of organelle structuring and membrane domain formation [
250]. Several small GTPases are found as substrates of key mitotic kinases. In this vein, the pervasive crosstalk between mitosis-specific kinases and small GTPases is speculated. This regulation network, mitotic kinase–small GTPase, may constitute a regulation mechanism of organelle dynamics in coordination with genetic material partition during mitosis.
14.2 Small GTPase family expands along with the cell evolution
The origin of membrane organelles in eukaryotes is one of the research foci of evolutionary cell biology. Being the nodal component of the membrane system, the expansion of the small GTPase family with specified interacting proteins may be the prerequisite for the generation of increased organelle complexity and regulatory control in autogenously derived eukaryotic compartments. In bacteria and archaebacteria, which lack nucleus and MBO, few, if any, small GTPases are found. By contrast, small GTPases take around 0.6% of the whole genome of
Homo sapiens. Significantly, the proportion peaks in unicellular organisms, which operate all life activities in a single cell and normally harbor a complicated and specialized cellular structure, e.g., 1.7% in
Naegleria gruberi [
251]. Naturally, specified organelles increase biological efficiency at the cost of regulation complexity especially during cell division. The analysis of small GTPase functions in organelle dynamics in cell growth and division may give directions toward organelle origination and mitosis evolution.