CELL CYCLE G2/M CONTROLS AND CHECK-POINTS

To insure the accurate transmission of genetic information, eukaryotic cells have developed an elaborate network of checkpoints to monitor the successful completion of every cell cycle step and to respond to cellular abnormalities such as DNA damage and replication inhibition as they arise during cell proliferation. The two best characterized checkpoints are those for DNA damage and DNA replication (Fig. 1A, top. For reviews, see ref 1, 2, 3, 4. Activation of either of these two checkpoints causes a cell cycle arrest, and they were first characterized in detail by genetic analysis in fission yeast (Schizosaccharomyces pombe). The G2 to M transition is controlled in fission yeast by the phosphorylation status of Tyr15 on Cdc2, the cyclin-dependent kinase which regulates the cell cycle in all eukaryotic cells 5. In fission yeast, Tyr15 is phosphorylated by the Wee1 and Mik1 kinases to hold the cell in G2, and rapid dephosphorylation by the Cdc25 phosphatase triggers the G2 to M transition 5, 6, 7, 8.

Figure 1
figure 1

Models for the classic cell cycle G2/M checkpoints and HIV-1 Vpr-induced cell cycle G2/M regulation. (A) Pathway for DNA damage and replication checkpoints in fission yeast. (B) Current model for Vpr-induced G2 arrest. The X indicates that Vpr may bind to cellular protein(s) and regulates the PP2A holozyme. →, activation; , inhibition; ?, unknown.

Full size image

The DNA damage checkpoint (Fig. 1A, top) is activated by ionizing radiation or ultraviolet light, and activation of this checkpoint leads to inhibitory Tyr15 phosphorylation of Cdc2 by a multistep pathway 9, 10. The early genes in the pathway, which include Rad1, Rad3, Rad9, Rad17, Rad26 and Hus1, are thought to sense the DNA damage and lead to phosphorylation of the Chk1 protein 11. In response to double strand DNA breaks (DSBs) induced by ionizing radiation, for example, Rad17 acts as a checkpoint-specific loading factor, which responds to the DNA damage by loading a 9-1-1 protein complex onto the sites where DNA is damaged 12, 13. The 9-1-1 protein complex is also known as the checkpoint clamp complex (CCC), which is composed of Rad1, Rad9 and Hus1 13. In addition, the Rad3-Rad26 protein complex also binds to sites of DNA damage independently of the 9-1-1 protein complex. The independent binding of two protein complexes to DNA damage to initiate the DNA structure checkpoint is believed to protect the cell against inappropriate checkpoint activation 2, 13, 14. Activation of Chk1 is mediated by Crb2, which may bridge Rad3 and Chk1 15, 16, 17. The activated Chk1 kinase then directly phosphorylates the Cdc25 phosphatase 18. The phosphorylated Cdc25 binds Rad24/25 (14-3-3) protein, and this complex is transported out of the nucleus to render Cdc25 inactive 19. The activated Chk1 also regulates the Mik1 kinase to inhibit Cdc2 20. DNA damage thus initiates a Chk1-mediated protein phosphorylation cascade ending in the inactivation of Cdc25 phosphorylase and activation of Mik1 kinase to increase inhibitory phosphorylation of Tyr15 on Cdc2.

The DNA replication checkpoint (Fig. 1A, bottom) is activated by treatment with hydroxyurea, which inhibits DNA replication, and this checkpoint also controls the G2 to M transition through inhibitory phosphorylation of Cdc2 10. Parts of this DNA replication checkpoint are shared with the DNA damage checkpoint as Rad1, Rad3, Rad9, Rad17, Rad26 and Hus1 are required for both checkpoints in fission yeast 21. The same 9-1-1 and Rad3-Rad26 checkpoint protein complexes may associate with the DNA replication complex 13. However, the DNA replication checkpoint acts primarily through phosphorylation of Cds1 kinase, which is mediated by another protein Mrc1 22, 23, with minor participation of Chk1 kinase, and either kinase is sufficient by itself to give cell cycle arrest when DNA synthesis is inhibited 24. Activated Cds1 kinase inactivates Cdc25 through a similar mechanism as Chk1 1, 24. Cds1 also activates the Mik1 kinase, which phosphorylate Tyr15 of Cdc2 25.

Cell cycle G2/M controls, which were often initially defined in fission yeast, are highly conserved, and most of the genes required for the checkpoints have human homologues (Tab. 1). In general, these homologues have similar, although not always identical, roles in the control of the human cell cycle. There is a tendency for multiple, partially redundant checkpoints in human cells compared to simpler checkpoints in yeast probably reflecting the more complex requirements for cell cycle control in multicellular eukaryotes. For example, the single rad3 gene in fission yeast is required for both the DNA damage and replication checkpoints and activation of the chk1 and cds1 checkpoint kinases 2, 13, 14. In human cells there are two homologues of rad3, ATM and ATR. The primary role of ATM is in the DNA damage checkpoint initiated by double strand breaks and activation of CHK2, the human homologue of cds1, whereas the primary role of the essential ATR gene is in the DNA replication checkpoint or responses to many forms of DNA damage and activation of CHK1 26, 27. Similarly, there is only one tyrosine phosphatase Cdc25 that dephosphorylates Cdc2 in fission yeast. In human cells, there are three CDC25 homologues, CDC25A, CDC25B and CDC25C, and each of them can be phosphorylated by CHK1 28. All three of these phosphatases have been shown to be involved in the control of the G2/M transition, even though their specific roles in this process have not yet been well characterized 29, 30, 31. The p53 gene is an example of an additional level of cell cycle control in human cells.The p53 transcription factor, which has no homologues in yeast, has multiple roles including regulating the cell cycle and apoptosis with an essential role in the G1 DNA damage checkpoint 32. It also has important roles in the G2 damage checkpoint. It inhibits Cdc2 through activation of Gadd45, p21, and 14-3-3σ. In addition, it is involved in regulatory feedback loops with ATM/ATR and CHK1 33. The conservation of checkpoints even extends to the regulatory mechanisms as illustrated by the negative regulation of CDC25 by relocation to the cytoplasm from the nucleus in both fission yeast and human cells. This relocation in both organisms is dependent on 14-3-3 proteins 19, 34.

Table 1 Functional conservation of cell cycle G2/M machinery between human and fission yeast
Full size table

CELL CYCLE G2/M ARREST INDUCED BY HIV-1 VPR

The HIV-1 Vpr protein induces cell cycle G2 arrest through inhibitory phosphorylation of Cdc2 both in fission yeast and human cells, suggesting that Vpr affects a conserved cellular process. Specifically Vpr induces hyper-phosphorylation of fission yeast Cdc2 or human CDK1, the human homologue of Cdc2 35, 36, 37. It exerts its inhibitory effect through T14A and Y15F of CDK1 and Y15F of Cdc2, as expression of nonphosphorylatable Cdc2 mutants, T14A Y15F of CDK1 and Y15F of Cdc2, prevents Vpr-induced G2 arrest 36, 38. Furthermore, Vpr inhibits Cdc25 phosphatase 39, 40 and activates Wee1 kinase 39, 41 to promote phosphorylation of Cdc2/CDK1 during induction of G2 arrest. Consistent with the roles of Wee1 and Cdc25 in Vpr-induced G2 arrest, proteins which regulate Cdc25 or Wee1 have been shown to either augment or alleviate Vpr-induced G2 arrest. For example, fission yeast Wos2, which is a homologue of human p23 and a Wee1 inhibitor 42, is a Vpr suppressor when overproduced 39. The Cdc25 inhibitor rad25 19, which is a homologue of human 14-3-3 proteins, enhances Vpr-induced G2 arrest when overproduced in fission yeast 39. Recent studies further show that Vpr binds to Cdc25C and to 14-3-3 in human cells 43, 44, 45.

Given that the DNA checkpoints and Vpr both induce G2 arrest through inhibitory phosphorylation of Cdc2 which is regulated by Wee1 and Cdc25, Vpr might induce G2 arrest through a checkpoint pathway. This possibility has been evaluated in fission yeast by expressing vpr in mutant fission yeast strains defective in early and late steps of the checkpoint pathways. None of the early checkpoint-specific mutants (rad1, rad3, rad9 and rad17) showed a significant effect on the induction of G2 arrest by Vpr 38, 39, 46. Furthermore, mutations in both chk1 and cds1, which are thought to be the last steps specific for the checkpoint 1, 18, 24, also do not block Vpr-induced G2 arrest 39, 46. Therefore, Vpr does not appear to use the DNA-damage or DNA-replication checkpoint pathways to induce G2 arrest in fission yeast.

Early data in human cells tended to support the conclusion that Vpr does not induce G2 arrest through the DNA damage checkpoint pathways. Vpr still induced G2 arrest in cells from patients with ataxia telangiectasia (AT) 40. These AT cells are mutant for the ATM gene, which is a human homologue of fission yeast Rad3, and they do not arrest in G2 in response to DNA damage caused by ionizing radiation 47, 48, 49. However, recent reports show that Vpr activates ATR, the second human homologue of fission yeast Rad3, and other steps in this checkpoint pathway such as RAD17, HUS1, BRCA1 and γ-H2AX 50, 51. Down-regulation of ATR or CHK1 by RNAi partially attenuated Vpr-induced G2 arrest 50, 51. These studies suggest that Vpr may induce G2 arrest through a cellular response to DNA replication stress or to a signal that “mimics” DNA damage. However, expression of vpr does not increase gene mutation frequency 52, which argues against the possibility that Vpr actually causes DNA damage. It is thus reasonable to think that other signals other than actual DNA damage triggers DNA damage-like cellular responses. These cellular responses could include the nuclear herniation caused by Vpr 53 or cellular stress responses to vpr gene expression 54, 55, 56. Since ATR and CHK1 have primary roles in the responses to changes in DNA replication, an alternative possibility is that Vpr may interfere with DNA replication. This possibility is certainly supported by a number of reports showing that Vpr induces genomic instability, formation of micronuclei and aneuploidy 57, 58. All of these changes in DNA structures could be perceived as replication stresses, which would trigger cell cycle arrest.

Considering that G2/M DNA checkpoints are highly conserved between mammalian and fission yeast cells (Tab. 1), it is unclear at the moment why human ATR and CHK1 are activated by Vpr but rad3 (the fission yeast homologue of ATR/ATM) or chk1/cds1 (CHK1/CHK2) double deletion in fission yeast does not block Vpr-induced G2 arrest 38, 39. One factor possibly contributing to the observed activation of ATR in mammalian cells is that retroviral integration appears to activate ATR 59, and the experiments showing activation of ATR by Vpr were done with lentiviral vectors which might therefore activate ATR to some extent independently of Vpr. In addition, it was noticed that activation of ATR and CHK1 only accounts for part of the G2 arrest induced by Vpr 50. Other as yet unidentified molecular mechanism(s) may explain at least half of the G2 cell population induced by Vpr. Interestingly, Roshal et al 60 showed that treatment of Vpr-producing mammalian cells with caffeine completely blocks Vpr-induced G2 arrest. Caffeine is part of the methylxanthine family, and similar to the caffeine effect, another methylxanthine pentoxifylline (PTX) also inhibits Vpr-induced G2 arrest in mammalian cells 61. Similarly, both PTX and caffeine suppress Vpr-induced G2 arrest in fission yeast 39, 58 [our unpublished data]. Since PTX or caffeine inhibits Vpr-induced G2 arrest in fission yeast where the classic DNA checkpoints apparently play no role, these observations suggest molecular mechanisms other than the classic DNA checkpoints may be involved in activation of ATR and regulation of Cdc25 and Wee1.

The additional molecular mechanism might involve protein phosphatase 2A (PP2A). Although this protein phosphatase has no known role in the DNA checkpoints, it has an important role in Vpr-induced G2 arrest. Okadaic acid is a specific inhibitor of PP2A, and okadaic acid was shown to inhibit Vpr-induced G2 arrest both in human 35 and fission yeast cells 37. Further evidence for an important role of PP2A comes from PP2A mutant strains. PP2A is composed of three subunits, one catalytic (C) and two regulatory (A and B) subunits. When vpr was expressed in a strain with a deletion for a catalytic subunit (ppa2) or a regulatory subunit (pab1) of PP2A, Vpr-induced G2 arrest was reduced 39, 46. Taken together, it is possible that a concerted cellular mechanism interlinks PP2A and possibly ATR/CHK1 in the cellular response to vpr gene expression during the induction of G2 arrest.

POTENTIAL ROLE OF PP2A IN CELL CYCLE G2/M REGULATION DURING VIRAL INFECTIONS

Other evidence supporting involvement of PP2A in Vpr-induced G2 arrest comes from other viral proteins with effects on cell cycle G2/M controls. PP2A appears to be a common viral target since other viruses such as simian virus 40 (SV40), polyoma virus, human T lymphotrophic retrovirus and adenovirus affect the enzymatic activity of at least a subset of PP2A proteins (see review 62). Even though these viruses are not otherwise related, they all seem to have adapted a similar strategy to affect cellular processes by direct interaction with PP2A. Similar to the Vpr effects, both adenoviral E4orf4 63, 64, 65 and HTLV Tax protein induce cell cycle G2 arrest 59. These two viral proteins both bind to PP2A and affect its enzymatic activity 63, 66. Interestingly, similar to Vpr, Tax-induced G2 arrest is reversed by caffeine 59. Further examinations indicated that Tax binds to CHK2 in Jurkat T-cells 59 but it complexes with CHK1 in other T-cells 67. In contrast, the T-antigens of both SV40 and polyoma virus promote cell proliferation by down-regulation of PP2A through direct competition with the B regulatory subunit (Reviewed in 68). Further studies on the interactions of Vpr and these other viral products with PP2A should provide additional insight into the role of PP2A in cell cycle G2/M regulation.

A NEW MODEL FOR CELL CYCLE G2/M TRANSITION

Based on our current knowledge about the effect of Vpr on cell cycle G2/M regulation, a new model for the cell cycle G2/M transition is proposed for HIV-1 Vpr as shown in Fig. 1B. Although this new model may not be mutually exclusive from the classic G2/M checkpoint pathways (Fig. 1A), it nevertheless has the unique feature of a PP2A-mediated mechanism or mechanisms for G2/M control by Vpr. We hypothesize that Vpr induces G2 arrest at least in part by going through an alternative cellular pathway(s) in which expression of vpr activates PP2A activity either by direct association with the PP2A enzyme complex or by association with an intermediate protein(s) X. In addition, there may also be a concerted cellular mechanism which interlinks PP2A and ATR/CHK1 in the cellular response to vpr during the induction of G2 arrest. A protein phosphorylation cascade (depicted as “?”) including PP2A is probably in part responsible for activation of the mitotic inducer Wee1 by altering its phosphorylation levels, which in turn inhibits Cdc2 by Tyr15 phosphorylation. While Wee1 plays the major role in the induction of G2 arrest by Vpr, Cdc25 appears to play a minor role and to be partially inhibited by this proposed regulatory pathway or by direct interaction with Vpr. Since overexpression of Sum1, a putative Cdc25 inhibitor, shows a synergistic effect with Vpr, Sum1 may be required by Vpr to inhibit Cdc25 either from upstream or downstream of Cdc25. Wos2 is another possible step in the pathway, and Vpr most likely inhibits Wos2 during the induction of G2 arrest. Since Wos2 physically interacts with Cdc2, Wos2 may be a downstream inhibitor of Wee1 (Fig. 1B). One possible mechanism for the inhibition of Wee1 activity by Wos2 is that Wos2 physical blocks the Tyr15 phosphorylation site on Cdc2. Vpr might then inhibit Wos2 by preventing Wos2 from binding to Cdc2. Based on the demonstrated similarities for cell cycle controls between fission yeast and human cells, it seems likely that pathway for Vpr-induced G2 arrest in fission yeast will also apply in most respects to that part of Vpr-induced G2 arrest in human cells which is not solely dependent on ATR.

SUMMARY

Besides the two most well-characterized cell cycle G2/M regulation pathways, i.e., DNA damage and replication checkpoints, some viral infections appear to modulate host cell cycle machinery in a fashion that is characteristically different from the classic checkpoints. This unique host-pathogen interaction is exemplified here in the induction of cell cycle G2 arrest by HIV-1 Vpr. PP2A seems to be a common cellular target for unrelated viruses. Therefore, further studies on the interactions of these viral proteins with PP2A and other related cellular factors should provide insight into the basic biology of cell cycle G2/M regulation and the biological significance of viral-induced G2 arrest during host-pathogen interactions.