BIOLOGICAL OVERVIEW

Chiffon is the Drosophila homolog of yeast Dbf4; its function is associated with the origin recognition complex (ORC), which is a multisubunit protein complex that binds to chromosomal origins of replication and is required for the initiation of cellular DNA replication. In yeast, one of the last steps in activation of the origin of replication appears to be phosphorylation of one or more origin-associated proteins by the Cdc7-Dbf4 protein kinase (Jackson, 1993; Lei, 1997; Owens, 1997). Cdc7 (Drosophila homolog: Cyclin-dependent kinase 7) is a cdk-like serine/threonine protein kinase whose activity is regulated in a cell-cycle-dependent manner by association with its regulatory subunit Dbf4. Dbf4 binds to origin-associated proteins, thus providing a likely mechanism for recruitment of Cdc7 to target(s) at the origin (Dowell, 1994). As the regulatory complex of the Cdc7-Chiffon dimer, Chiffon interacts with unknown regulators and activates the kinase function of its dimerization partner, thus triggering the initiation of cellular DNA replication. A better understanding of Chiffon should lead to an enhanced understanding of the regulation of DNA replication, a central feature of the biology of the cell (Landis, 1999 and references therein).

To meet demand for rapid synthesis of chorion (eggshell) proteins, Drosophila ovary follicle cells amplify the two chromosomal clusters of major chorion genes ~80-fold. Amplification occurs through repeated firing of one or more DNA replication origins located near the center of the gene clusters. Chorion gene amplification is amenable to both molecular and genetic analysis, making it an ideal model system for the study of metazoan chromosomal DNA replication origins and their regulation. Constructs derived from the third chromosome chorion gene cluster are able to amplify when reinserted into the Drosophila genome using P element-mediated transformation. This assay has allowed mapping of a cis-acting control element required for high levels of amplification (termed ACE3 for Amplification Control Element, 3rd chromosome), as well as several stimulatory regions. Two-dimensional gel analysis of DNA replication intermediates originating from the endogenous locus and from transgenic constructs has allowed mapping of the major origin of replication, called Ori-beta, as well as one or two more minor origin regions (Landis, 1999 and references).

Mutations that disrupt genes required in trans for amplification cause female sterility. The sterility is due, at least in part, to underproduction of chorion proteins and synthesis of thin, fragile chorions. Twelve genes have been identified that are required for amplification (Calvi, 1998; Royzman, 1999). The first of these to be cloned (k43), was found to encode the Drosophila homolog of the S. cerevisiae Origin Recognition Complex subunit 2 (ORC2) (Gossen, 1995; Landis, 1997). The ORC is a complex of six proteins (Orc1p-Orc6p) that binds to S. cerevisiae chromosomal DNA replication origins in vivo and in vitro. The phenotypes of mutations in yeast ORC subunit genes demonstrate that ORC is required for origin function, as well as for transcriptional silencing of the yeast mating-type loci. ORC is associated with the origin throughout the cell cycle and activation of the origin appears to result from association of additional proteins with the ORC. The identification of Drosophila k43 as the homolog of S. cerevisiae ORC2 (see Drosophila Origin recognition complex 2) suggests that all or part of the mechanism of origin regulation may be conserved between yeast chromosomal origins and the chorion gene origins. The chiffon gene was originally identified in female sterile mutants, which lay eggs with thin eggshells (Schüpbach, 1991). Experiments have shown that chiffon is required in trans for chorion gene amplification. Cloning and characterization of the Drosophila chiffon gene have revealed that the predicted Chiffon protein contains two domains (designated CDDN for Chiffon, Dbf4, Dfp1 and NimO); these four proteins are related to regulators of DNA replication and cell cycle in lower eukaryotes (Landis, 1999).

Chorion gene amplification occurs through the repeated firing of one or a small number of origins located near the center of each of the chorion gene clusters. Amplification is expected to utilize all or part of the cell’s general DNA replication machinery. In addition, there must be some mechanism making this process specific to the follicle cells and to only the origins associated with the chorion gene clusters. In S. cerevisiae, ORC is required for origin activity; however, activation of the origin requires binding of additional proteins prior to S phase. Firing of the S. cerevisiae origin in S phase also requires the activity of the Cdc7-Dbf4 protein kinase. Cdc7 is a cdk-like serine/threonine protein kinase whose activity requires binding of the regulatory subunit Dbf4 (Jackson, 1993). Dbf4 binds to proteins associated with the origin, thus providing a likely mechanism for recruitment of Cdc7 to the origin (Dowell, 1994). The Mcm2-7 proteins comprise a family of six highly conserved proteins that associate with the ORC origin complex prior to S phase; they are required for initiation of DNA synthesis. Cdc7-Dbf4 phosphorylates a subset of the Mcm proteins in vivo and in vitro suggesting that this may be part of the mechanism by which Cdc7-Dbf4 activates the origin (Lei, 1997). Cdc7 homologs have now been characterized and cloned from S. pombe, Drosophila, Xenopus laevis, mouse and human. This conservation suggests that regulation of DNA replication origin firing by Cdc7-family protein kinases is conserved through eukaryotic evolution. The cloning of the chiffon gene suggests that function of the regulatory subunit of the Cdc7 protein kinase may also be conserved (Landis, 1999).

The predicted Chiffon protein was found to contain two domains, designated CDDN1 and CDDN2, which are also found in S. cerevisiae Dbf4. The CDDN domains are also found in S. pombe Dfp1 protein, which is the S. pombe homolog of Dbf4 (Brown, 1998), as well as in an S. pombe gene highly related to Dfp1 (called 'similar to rad35/dfp1'; K. Oliver, direct submission to GenBank). The CDDN domains are also found in the predicted translation product of the Aspergillus nimO gene. The nimO gene is required for DNA synthesis and mitotic checkpoint control in Aspergillus (James, 1999). That study points out the homology between nimO protein and Dbf4 in the CDDN1 region, and makes the interesting observation that the structure of the CDDN1 region is consistent with a single Cys2-His2 zinc finger-like motif with a short central loop of 9 bp. Finally, the CDDN domains have also been found in a predicted human protein of unknown function and in several mouse cDNAs of unknown function. The data suggest a family of eukaryotic proteins related to Dbf4 and involved in initiation of DNA replication (Landis, 1999).

The similarity among Chiffon, Dbf4, and Dfp1 suggest a model for Chiffon’s role in amplification: Chiffon may function in the activation of the chorion gene origins as the regulatory subunit of a kinase involved in origin firing, most likely the Drosophila homolog of Cdc7. S. cerevisiae Dbf4 contacts Cdc7 through the carboxyl terminus, where the CDDN1 domain is located. Thus it is hypothesized that, analogous to Dbf4 function in S. cerevisiae, Chiffon may contact the Drosophila Cdc7 homolog through the conserved CDDN1 domain and recruit it to the ORC via conserved ORC contact sites in the Chiffon amino terminus, perhaps the CDDN2 domain. Chiffon could also be hypothesized to recruit other, as yet unidentified proteins to the origin. Alternative and less direct models for Chiffon function during amplification cannot be ruled out. In addition to the defect in chorion gene amplification, the chiffon null phenotype also includes rough eyes and thin thoracic bristles. While there are several possibilities for how chiffon might be required for normal eye and bristle development, these phenotypes are consistent with a defect in DNA replication and/or S phase control in the cells forming these structures. For example, roughex regulates cyclin levels and entry into S phase, and roughex mutants are viable with rough eyes similar to chiffon nulls (Thomas, 1997). Morula is a regulator of mitotic- and endo-cell cycles; hypomorphic morula mutants have rough eyes and thin thoracic bristles similar to chiffon null mutants (Reed, 1997). Finally, specific hypomorphic mutations in either the dDP or dE2F subunits of the Drosophila cell cycle regulator E2F cause rough eyes, thin thoracic bristles and defective chorion gene amplification nearly identical to chiffon nulls (Royzman, 1997 and 1999). Thus, Drosophila chorion gene amplification, eye development and thoracic bristle development appear to be processes that are particularly sensitive to defects in the cell cycle/DNA replication machinery (Landis, 1999).

There are a number of possibilities for how chiffon might function in S phase regulatory pathways with the above-mentioned and/or other cell cycle and S phase regulators. Identification of proteins that interact with chiffon in vivo will begin to test these models and should facilitate the dissection of the mechanism(s) regulating chorion gene amplification. Identification of proteins that interact with the evolutionarily conserved CDDN domains of Chiffon, using techniques such as yeast two-hybrid system, should prove particularly informative. The chiffon null phenotype demonstrates that Chiffon is required for chorion gene amplification and normal eye and bristle development. Chiffon might therefore be a tissue-specific regulator of DNA replication and/or cell cycle. However, it remains possible that Chiffon might be required in additional tissues, or even in every cell for DNA replication and/or cell cycle control. Because wild-type Chiffon mRNA is maternally supplied to chiffon mutant embryos, wild-type Chiffon protein might persist through embryonic and larval development and mask a more general requirement for chiffon function. Additional experiments, such as analysis of chiffon mutant germline clones will be required to address these questions (Landis, 1999).

Temporal control of late replication and coordination of origin firing by self-stabilizing Rif1-PP1 hubs in Drosophila

In the metazoan S phase, coordinated firing of clusters of origins replicates different parts of the genome in a temporal program. Despite advances, neither the mechanism controlling timing nor that coordinating firing of multiple origins is fully understood. Rif1, an evolutionarily conserved inhibitor of DNA replication, recruits protein phosphatase 1 (PP1) and counteracts firing of origins by S-phase kinases. During the midblastula transition (MBT) in Drosophila embryos, Rif1 forms subnuclear hubs at each of the large blocks of satellite sequences and delays their replication. Each Rif1 hub disperses abruptly just prior to the replication of the associated satellite sequences. This study shows that the level of activity of the S-phase kinase, DDK, accelerated this dispersal program, and that the level of Rif1-recruited PP1 retarded it. Further, Rif1-recruited PP1 supported chromatin association of nearby Rif1. This influence of nearby Rif1 can create a “community effect” counteracting kinase-induced dissociation such that an entire hub of Rif1 undergoes switch-like dispersal at characteristic times that shift in response to the balance of Rif1-PP1 and DDK activities. A model is proposed in which the spatiotemporal program of late replication in the MBT embryo is controlled by self-stabilizing Rif1-PP1 hubs, whose abrupt dispersal synchronizes firing of associated late origins (Cho, 2022).

During a typical metazoan cell cycle, large genomic domains initiate their replication at distinct times in S phase. Cytological studies over 60 y ago revealed that DNA sequences in the compacted heterochromatin replicate later in S phase compared to euchromatin. These early studies and recent detailed analyses revealed a complex program among late replicating domains, in which different domains initiate replication with a specific delay. Execution of this stereotyped schedule occupies much of the S phase and must finish before mitosis. Despite recent advances in genomic methods for profiling global replication timing, the basis of the timing control is not yet solved, and how multiple origins are coordinated to fire together especially within repetitive DNA sequences is not known (Cho, 2022).

The Drosophila embryo offers a unique setting in which to examine the control of temporal programing of replication. In the earliest nuclear division cycles, there is no late replication, closely spaced origins throughout the genome initiate replication rapidly at the beginning of interphase, and their simultaneous action results in an extraordinarily short S phase (3.5 min). Late replication is developmentally introduced during the synchronous blastoderm nuclear division cycles, first influencing pericentric satellite sequences that form a major part of metazoan genomes (over 30% in Drosophila). Individual blocks of satellite DNA are typically several megabase pairs in length, each composed of a different simple repetitive sequence. During the 14th cell cycle at the midblastula transition (MBT), the ∼6,000 cells of the entire embryo progress synchronously through a temporal program in which the different satellites are replicated with distinctive delays (4), dramatically extending the duration of S phase (Cho, 2022).

The initial onset of late replication during development provides a simplified context in which to define its mechanism, because numerous complex features associated with replication timing have not yet been introduced. For example, chromatin states can have major impacts on replication timing. Consistent with this, late-replicating satellite sequences are usually heterochromatic, carrying the canonical molecular marks of constitutive heterochromatin (histone H3 lysine 9 methylation and HP1). During initial Drosophila embryogenesis, the satellites lack significant levels of these marks, and they replicate in sync with the rest of the genome. Surprisingly, the introduction of the delays in replication to the satellite sequences precedes a major wave of heterochromatin maturation in the blastoderm embryo. Furthermore, in a Rif1 null mutant (Rif1KO), the S phase of cycle 14 is significantly shorter, and the late replication of satellite sequences is largely absent even though HP1 recruitment appears normal. Thus, a Rif1-dependent program bears virtually full responsibility for the S-phase program at the MBT (Cho, 2022).

Rif1 is a multifunctional protein with an evolutionarily conserved role in regulating global replication timing. In species from yeast to mammals, mutation or depletion of Rif1 disrupts genome-wide replication timing. Studies in a variety of systems revealed several aspects of Rif1 function. Yeast Rif1 associates with late origins, while the Rif1 of both Drosophila and mammals binds broadly within large late-replicating domains. Rif1 has a conserved motif for interacting with protein phosphatase 1 (PP1), and mutations in the PP1-interacting motifs lead to hyperphosphorylation of MCM helicase in the prereplicative complex (pre-RC) and the disruption of global replication timing. Rif1 itself also harbors many sites recognized by S-phase kinases, including CDK and DDK, near its PP1-interacting motifs. In yeast, both a Rif1 mutant with phosphomimetic changes at these phosphorylation sites and a null mutation of Rif1 partially restore the growth defect of DDK mutants. These data suggest an interplay of Rif1 and DDK, wherein DDK acts first upstream of Rif1 phosphorylating it to disrupt its interaction with PP1, thus lowering the threshold of S-phase kinase activities required for origin firing. Second, DDK acts downstream to directly phosphorylate pre-RC and trigger origin firing. However, how these various features of Rif1 and DDK functions are integrated over large genomic regions to provide a domain-level control of replication timing remains elusive (Cho, 2022).

Studies in flies indicate that Rif1 has adopted a developmental role in governing the onset of the late replication program described above. During the early embryonic cell cycles, high Cdk1 and DDK activities jointly inhibit maternally deposited Rif1, promoting synchronous firing of origins throughout the whole genome to ensure completion of DNA replication during the short interphases. As the cell cycle begins to slow and oscillations in Cdk1 activity emerge, a transient Rif1-dependent delay in the replication of satellite sequences slightly prolongs S phase. When the embryo enters the MBT in cycle 14, abrupt down-regulation of Cdk1 more fully derepresses Rif1, which accumulates in semistable foci (hubs) at satellite DNA loci. High-resolution live microscopy reveals that different Rif1 hubs disperse abruptly at distinct times, followed by proliferating cell nuclear antigen (PCNA) recruitment as the underlying sequences replicate. Mutated Rif1 that is nonphosphorylatable at a cluster of CDK/DDK sites fails to dissociate from satellite DNA and dominantly blocks the completion of satellite DNA replication before mitosis. Conversely, ectopically increasing CDK activity in cycle 14 shortens the persistence of endogenous Rif1 foci and advances the replication program. These findings suggest that each Rif1 hub maintains a local nuclear microenvironment high in Rif1-recruited PP1 that inhibits DNA replication, and that kinase-dependent dispersal of Rif1 hubs is required to initiate the replication of satellite sequences. If it were understood what coordinates Rif1 dispersal throughout the large Rif1 hubs, this model could explain how firing of clusters of the underlying origins is coordinated and how replication of different satellites occurs at distinct times. However, the precise mechanisms controlling the dynamics of Rif1 hubs remain unclear (Cho, 2022).

Since Rif1 can recruit PP1 and form phosphatase-rich domains in the nucleus, it was hypothesized that localized PP1 counteracts kinase-induced Rif1 dissociation so that the Rif1 hubs are self-stabilizing. If this self-stabilization is communicated within each hub, a breakdown in self-stabilization would lead to a concerted collapse of the entire hub and allow origin firing throughout the associated satellite sequence. The current findings indicate that the opposing actions of phosphatase and kinase combined with communication within the hubs create a switch in which a large phosphatase-rich domain is stable until kinase activity overwhelms the phosphatase. It is proposed that for large late-replicating regions of the genome, recruitment of Rif1-PP1 creates a new upstream point of DDK-dependent regulation in which DDK triggers the collapse of the phosphatase-rich domain to create a permissive environment for kinase-induced firing of all previously repressed origins (Cho, 2022).

This study has investigated the mechanisms that control the timing of Rif1 foci dispersal from satellite sequences, which dictates the onset of late replication in the MBT embryo. Rif1-recruited PP1 was demonstrated to mediate self-stabilization of Rif1 hubs, while the S-phase kinase DDK opposes PP1 action and triggers the dispersal of Rif1 hubs. A model is proposed in which the firing of late origins is primarily controlled by a de-repression step upstream of the activation of the pre-RC. In this model, hubs of Rif1 create domains of locally high PP1 that prevent kinase activation of underlying pre-RCs. However, a changing balance of local phosphatase and kinase levels leads to the abrupt destabilization of different Rif1 hubs at distinct times (see A model for the multiple actions of PP1 in stabilizing Rif1 hubs.). This alleviates PP1 inhibition of hub-associated origins at specific times to trigger replication of the different satellites at different times. While this simple model appears sufficient to explain the late replication at its initial onset in the early Drosophila embryo, numerous other factors impact the replication program at later stages when chromatin acquires more complex features. Nonetheless, as is discussed below, the simplicity of the process in this biological context offers some insights into the more enigmatic aspects of late replication, and perhaps suggests a flexible regulatory paradigm that might be used in diverse contexts (Cho, 2022).

While the mechanism is unknown, it has long been clear that large domains of the genome behave as timing units, and that the numerous origins within such domains fire coordinately if not synchronously. The hub model of late replication control in the early embryo can explain how the firing of numerous origins within megabase pairs of satellite sequences can be coordinated in late S phase. Each Rif1 hub is associated with a locus of repetitive satellite sequence (10). Coordinated dispersal of a Rif1 hub will convert the subnuclear compartment from one restricting kinase actions to a permissive one, allowing the activation of pre-RCs throughout the associated chromatin domain. It was previously unclear what leads to the coordinated dispersal of these large hubs. This study shows that a mutant Rif1 that is deficient in binding PP1 cannot form stable hubs on its own, but it joins wild-type Rif1 in semistable hubs. Importantly, the mutant and wild-type Rif1 disperse together, showing that they respond equally to the property of the domain. It is suggested that Rif1-bound PP1 can act in trans to stabilize nearby Rif1-PP1 and that the propagation of this action coordinates the behavior of Rif1 across the entire hub (Cho, 2022).

The contribution of PP1 to the self-stabilization of Rif1 hubs might be mediated by feedback at multiple levels): 1) PP1 might activate itself by removing inhibitory phosphorylation catalyzed by Cdk1; 2) It could reverse Cdk1/DDK-mediated phosphorylation of Rif1 that disrupts PP1-recruitment; 3) It could reverse phosphorylation of Rif1 that disrupts Rif1 chromatin association; or 4) In a circuitous pathway, if the firing of origins were to promote Rif1 dissociation, PP1-dependent suppression of origin firing would stabilize the hubs. Any or all the above actions could reinforce the stability of Rif1-PP1 hubs, perhaps making different contributions in different situations and different organisms. However, regardless of the feedback route, a local dominance of PP1 will stabilize the Rif1 hubs, and rising kinase activity could erode this dominance of PP1. Upon reaching a tipping point, the local PP1 would no longer successfully stabilize the Rif1 hub, and S-phase kinases would then trigger complete dispersion and allow replication of the underlying chromatin (Cho, 2022).

A potential ability of origin firing to feedback and destabilize Rif1 hubs might explain observations in other organisms suggesting that the level of a variety of replication initiation factors can influence replication timing. For example, overexpression of four replication factors including a DDK subunit in the Xenopus embryo shortens the S phase at the MBT. While this has been interpreted as evidence for governance of replication timing by limitation for these factors, the effect may be indirect if overproduction of these factors overrides Rif1 suppression of pre-RC activation to advance the replication of late replicating regions as is seen in the fly embryo (Cho, 2022).

Importantly, the replication defects resulting from Cdc7 knockdown or inhibition of Cdc7, are suppressed in a Rif1 null mutant background. This shows that the level of DDK activity required to reverse or override Rif1 suppression of pre-RC activation is greater than the level needed for direct pre-RC activation. Thus, in a scenario in which rising levels of DDK during S-phase 14 act as a timer, genomic domains associated with Rif1 hubs would fail to replicate until DDK reached the high level required to destabilize the hub. This argues that replication timing depends on the threshold for derepression of the domain rather than on distinct thresholds for firing individual pre-RCs. It is therefore suggested that the timing of late replication is governed at the level of the upstream derepression step in Drosophila embryos, in contrast to the model proposed for other organisms according to which activation of pre-RCs are directly limited by availability of DDK and other replication factors. To produce the distinct temporal program of replication of different satellites, the current model requires domain-specific distinctions in the threshold for hub dispersal. Different satellite loci that are composed of a common repeat sequence replicate at the same time, while satellites composed of different sequences replicate at distinct times. This leads to a proposel that the sequence of repeats influences, likely indirectly, the threshold for Rif1 hub dispersal and the timing of replication (Cho, 2022).

The possible generality of the circuitry this study has defined in the cycle 14 Drosophila embryo can be considered in various ways. Focusing directly on Rif1 involvement in late replication, it is clear that Rif1 does not bare full responsibility for late replication at other stages. Nonetheless, a dosage-dependent function of Rif1 in controlling replication timing is also observed in Drosophila follicle cells during their mitotic cycles. Furthermore, in mammalian cells, ChIP-seq and microscopy showed that Rif1 interacts with large late-replicating domains but, as was seen in cycle 14 embryos, is absent once onset of replication of the underlying chromatin is detected. It is suggested that the mechanism described in this study will be one of multiple contributors to replication timing control in other biological contexts, and it is likely to be the major mode of replication timing in the rapid cycles of externally developing animal embryos (Cho, 2022).

Rif1 has other regulatory roles beyond timing control of pre-RC activation. In the follicle cells of Drosophila egg chambers, Rif1 is recruited to specialized replication forks during chorion gene amplification where it suppresses fork progression. While this action of Rif1 is dependent on its ability to associate with PP1, other possible parallels to the mechanism described in this study are not evident. Rif1 also regulates biological processes beyond replication. It is recruited to regions of DNA damage in mammals as well as to the telomeres in yeast where it has regulatory roles involving distinct interactions. Thus, Rif1 recruitment appears to trigger alternative regulatory pathways in different circumstances (Cho, 2022).

Despite the evident diversity of biological regulation, the capacity of Rif1 to form local membraneless compartments dominated by phosphatase and to abruptly dissolve in response to kinase levels might be an example of a group of flexible regulatory strategies. Many important regulatory events, such as phosphorylation, acetylation, and ubiquitination, are countered by reverse reactions. Various processes, notably the formation of liquid-like condensates, promote local accumulation of proteins. Accumulations of proteins that promote or oppose regulatory modifications could control major regulatory pathways. Furthermore, since protein accumulations could be stabilized or destabilized by the modifications they regulate, a feedback mechanism could control the formation and destabilization of a compartment to give precise spatiotemporal control, as exemplified by the behavior of the Rif1 hubs in the cycle 14 Drosophila embryo (Cho, 2022).

GENE STRUCTURE

chiffon potentially encodes two alternate open reading frames, one 5.085 kb and the other 5.133 kb, encoding proteins of 1695 and 1711 amino acid residues, respectively. The two proteins differ in their 3' amino acid sequences and 3' UTRs

Bases in 5' UTR - 691

Bases in 3' UTR - 859 and 771

PROTEIN STRUCTURE

Amino Acids - 1695 and 1711

Structural Domains

The predicted 5.085 kb chiffon ORF sequence was used to search the NCBI databases for related proteins. The predicted Chiffon protein was found to be related to a group of known or predicted proteins: S. cerevisiae Dbf4; S. pombe Dfp1 (which is the S. pombe homolog of S. cerevisiae Dbf4); Aspergillus nidulans NimO protein termed SRAD35 (which stands for 'Similar to RAD35', where RAD35 is another name for Dfp1), and finally the protein predicted by translation of specific human and mouse clones. The protein sequences were aligned. The region of greatest similarity between Chiffon and the other proteins is a 44 amino acid residue domain designated CDDN1 (for chiffon, Dbf4, Dfp1 and NimO). In the CDDN1 domain, Chiffon is 43% identical to Dbf4; 41% identical to Dfp1; 36% identical to NimO, and 32% identical to the human BAC clone translation product. Chiffon is also related to the other proteins in an amino terminal region designated CDDN2, although to a lesser extent than for CDDN1 (Landis, 1999).

date revised: 2 January 2023

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.