BIOLOGICAL OVERVIEW

The proto-oncogene c-myb was first encountered in chickens as a transduced retroviral oncogene, v-myb, which causes a myeloid leukemia and transforms myeloid cells in culture. Mutations affecting c-myb have since been implicated in the tumors of both humans and mice. The product of c-myb (Myb) is a transcription factor that binds a specific DNA sequence. The protein is divided into at least three discrete functional domains: one for binding to DNA, a second for activation of transcription, and a third domain that governs the biochemical activity of the protein. It is generally believed that Myb proteins, including MybA and MybB, two additional vertebrate Myb proteins that are related to c-myb, play roles in the cell division cycle. c-myb and MybB have been implicated in the G1/S transition, whereas MybA is more likely to be involved in cellular differentiaton. Myb-like proteins with related DNA-binding domains have also been found in yeast and plants (Katzen, 1998 and references).

Drosophila Myb oncogene-like is required for both mitosis and prevention of endoreduplication in wing cells. Drosophila Myb apparently acts at or near the time of the G2/M transition. The two mutant alleles of Drosophila Myb produce the same morphologically altered phenotype, although the responsible mutations are located in different functional domains of the gene product. Mutant wings have approximately half the number of hairs as wild-type wings, and the mutant hairs are considerably larger than normal. Mutant hairs are less regularly spaced, less uniform in orientation, and occasionally grouped in small clusters, indicating a disturbance of tissue polarity. The mutant phenotype can be partially suppressed by ectopic expression of either cdc2 or string, two genes that are known to promote the transition from G2 to M. Drosophila Myb is thought to be required for completion of cell division and may serve two independent functions: (1) promotion of mitosis and (2) prevention of endoreduplication when cells are arrested in G2 (Katzen, 1998).

In wild-type wings, each cell that is not specialized for another purpose is represented by a single hair. At 3 hr after puparium formation (APF) all wing cells except those at the anterior and posterior margins arrest synchronously in G2. The arrest persists until 12 hr APF, when mitosis occurs, followed by an additional cell division cycle that is completed by 24 hr APF. Thereafter, the cells remain in G0. For wings dissected out of mutant pupae between 24 and 36 hr APF, the density of nuclei in any given region of the wing is approximately half of that found for the same region of a wild-type wing, and the mutant nuclei are larger than their wild-type counterparts. Since no cell death is observed in mutant wings, it is concluded that mutant wings have fewer but larger cells, each of which produces a hair. This result points to a defect in cellular proliferation during the final stages of wing development in the Drosophila Myb mutants. When examined during the early G2 arrest (at 6 hr APF), the number of wing cells in myb1 mutants is normal, but in postmitotic wings (after 24 hr APF), the number of wing cells in myb1 mutants is half that of wild type. It is not possible to examine the wings between 7 and 24 hr APF, the period during which the final cell cycles occur. If both of the cell cycles that normally ensue are defective, the number of cells should be reduced by a factor of four, not just two. Thus, it is likely that only the final cell division fails to occur in the mutant wings (Katzen, 1998).

BrdU incorporation studies show that mutant as well as wild-type wings actively replicating DNA. The patterns of BrdU incorporation for the myb1 wings indicate that they are less mature than the wild-type wing, a finding that is consistent with observations that myb1 mutants develop more slowly than the parental flies. It is concluded that myb1 mutant wing cells enter into their final S phase but do not undergo a final division, presumably leaving them with a 4C content of DNA. To assess further the cell cycle state of the mutant Myb wing cells, a test was performed to determine whether the mutant phenotype can be suppressed by overexpression of either wild-type or activated alleles of cdc2 (the cyclin-dependent kinase that regulates the G2/M transition) or string (the Drosophila homolog of cdc25, the protein tyrosine/threonine phosphatase that regulates cdc2 activity). Overexpression of either cdc2 or string is able to partially suppress the myb2 wing phenotype, increasing the number of hairs by 30% to 35% and correcting the orientation of the hairs. These results support the conclusion that the mutant Myb wing cells are arrested at the G2/M transition (Katzen, 1998).

High-resolution, three-dimensional wide-field fluorescence microscopy was used to examine the amount of DNA contained in nuclei of individual wild-type and mutant cells. The average DNA content of mutant myb1 nuclei proves to be double that of wild-type nuclei, whereas the average for myb2 nuclei is only slightly elevated. Surprisingly, the average DNA content of mutant myb1 nuclei is even greater at 25°C than at 18°C, and the average for myb2 nuclei increases to approximately double that of wild-type. The mutant cells generally contain more DNA than wild-type cells. But there is substantial heterogeneity in the data for mutant cells, first apparent as relatively large standard deviations for the measurements of DNA content. Analysis of the data for individual cells provides an explanation: at the semi-permissive temperature of 18°C, the amount of DNA in myb1 nuclei varies from 2C to well in excess of 4C. At 25°C, the majority of nuclei contain more than 4C DNA. It is concluded that the mutation causes endoreduplication in a fraction of the arrested wing cells, the severity of which varies from one cell to another (Katzen, 1998).

These studies conclude that two cell cycle checkpoints appear to be disturbed in the wing cells of Myb mutants: regulation of the G2/M transition and prevention of re-entry into S phase before M phase occurs. Is one of these defects a consequence of the other or are they independent of each other? In string mutant embryos, cells arrest at the G2/M boundary and do not enter S phase, indicating that the mechanism for preventing reinitiation of DNA replication before mitosis can remain intact when cells are prematurely arrested in G2. Therefore, it is unlikely that the endoreduplication in mutant Myb cells is simply a consequence of the abnormal arrest in G2. Because overexpression of either cdc2 or string can partially suppress the Myb phenotype, at least a proportion of the mutant cells must still be competent for mitosis, suggesting that the block in G2 is not just a consequence of the mutant cells entering into an endocycle and losing the ability to divide. It is concluded that Myb may play active and independent roles in both the G2/M checkpoint and the S phase reduplication checkpoint. Final determination of whether Myb function is required in all cell cycles will have to await the generation of stronger alleles of Myb (Katzen, 1998).

Drosophila lin-52 acts in opposition to repressive components of the Myb-MuvB/dREAM complex

The Drosophila Myb-MuvB/dREAM complex (MMB/dREAM) participates in both the activation and repression of developmentally regulated genes and origins of DNA replication. Mutants in MMB subunits exhibit diverse phenotypes, including lethality, eye defects, reduced fecundity, and sterility. This study used P-element excision to generate mutations in lin-52, which encodes the smallest subunit of the MMB/dREAM complex. lin-52 is required for viability, as null mutants die prior to pupariation. The generation of somatic and germ line mutant clones indicates that lin-52 is required for adult eye development and for early embryogenesis via maternal effects. Interestingly, the maternal-effect embryonic lethality, larval lethality, and adult eye defects could be suppressed by mutations in other subunits of the MMB/dREAM complex. These results suggest that a partial MMB/dREAM complex is responsible for the lethality and eye defects of lin-52 mutants. Furthermore, these findings support a model in which the Lin-52 and Myb proteins counteract the repressive activities of the other members of the MMB/dREAM complex at specific genomic loci in a developmentally controlled manner (Lewis, 2012).

The primary objective of this study was to explore the function of Lin-52, the smallest subunit of the MMB/dREAM complex, which was previously defined by biochemical purification. A lin-52 null mutant was generated, and a regulatory interplay was identified between different subunits of the MMB/dREAM complex. Lin-52 functions, in part, to inhibit the repressive activities of the other MMB/dREAM subunits, including Mip130, Mip120, and Mip40. Several phenotypes observed in lin-52 mutants (larval lethality, adult eye defects, and maternal-effect embryonic lethality) were suppressed by mutations in these core subunits of the MMB/dREAM complex. Similar genetic interactions have been between myb mutants and the same core MMB/dREAM complex subunits. However, lin-52 is essential for some biological processes that do not require myb, including embryonic development and adult eye development. Consistent with these observations in vivo, previous studies have shown that in cell culture the number of genes regulated by Lin-52 alone far exceeds the number of genes regulated by Myb alone or by both Lin-52 and Myb together (Lewis, 2012).

It was hypothesized that in some contexts, Myb inhibits the repressive activities of Mip130, Mip120, and Mip40 and thereby aids in the activation of genes required for normal cell cycle progression and organismal development. This model, originally based on evidence from unexpected genetic interactions, was supported by subsequent RNAi studies in cultured cells and by gene expression studies in vivo. Gene expression analysis after RNAi depletion in Drosophila Kc cells revealed a combinatorial requirement of MMB/dREAM subunits for either gene activation or repression. These studies showed that two-thirds of the genes repressed by MMB/dREAM did not require Myb protein, whereas most genes activated by MMB/dREAM required Myb protein (Lewis, 2012).

The gene networks regulated by MMB/dREAM can be largely categorized into three main classes of repressed genes, namely, Myb-, E2F2-, or Mip120/Mip130-dependent repression, and a single major class of Myb-activated genes. Lin-52 is required for 76% of Myb-activated genes (for comparison, 56% of genes were dependent on Mip40), supporting a model in which Lin-52 plays an important role in gene activation. The current thinking is that regulation is hierarchical, such that specific loci may be repressed and then activated (or vice versa). The members of the MMB complex would help to establish one level of regulation at a primary stage of determination, while other cis-acting factors would determine the normal basal and induced levels of expression at another level of transcriptional control. The mechanistic basis of this process remains to be elucidated. However, it is clear that repression of many genes requires E2F2 and associated Rbf proteins and often the recruitment of L(3)MBT. Activation at this first level may require Myb, and as suggested in this, Lin-52 may be key as an activator of transcription early in embryogenesis and in the developing eye (Lewis, 2012).

The finding that Drosophila lin-52 appears to counteract the activities of other MMB/dREAM complex subunits may extend to paralogous genes. The Drosophila gene wake-up-call (wuc) was recently identified as a lin-52 paralog that is exclusively expressed in testes (Doggett, 2011). Wuc protein directly interacts with Aly, a Mip130 paralog that is also a subunit of the testis meiotic arrest complex (tMAC). Three subunits of tMAC are testis-specific paralogs of MMB/dREAM subunits: Aly (Mip130), Tomb (Mip120), and Wuc (Lin-52). The other two subunits of tMAC are also present in MMB/dREAM: Mip40 and p55/Caf1. The loss of tMAC function results in a decrease in testis-specific gene expression and meiotic arrest in spermatocytes. Interestingly, wuc mutants could partially rescue expression of several putative tMAC target genes in aly mutant testes. These data suggest that wuc opposes aly function in a manner analogous to the opposition of lin-52 and mip130 shown in this study (Lewis, 2012).

Recent work in mammalian systems has demonstrated an important function for Lin-52 homologs in the DREAM complex. Phosphorylation of Lin-52 by the DYRK1A kinase was required for full assembly of the human DREAM complex. These results indicate that Lin-52 plays an important role in mediating protein-protein interactions among DREAM complex subunits (Lewis, 2012).

Unlike Drosophila, C. elegans lacking lin-52 is viable. However, the mutant worms are sterile. Genome-wide ChIP and gene expression analyses indicate that DRM directly represses 119 genes, primarily those involved in reproduction and development. Expression of these genes in specific developmental programs or tissues presumably requires the inhibition of the repressive DRM complex. C. elegans appears to have lost an animal-type myb gene during the course of evolution. It is not currently known if the DRM complex contains another transcription factor that may function to inhibit the repressive activities of the DRM complex. In the absence of a dedicated activator subunit, Lin-52 may serve in both the activation and repression of DRM-regulated genes. Perhaps phosphorylation by a DYRK1A-like kinase or some other posttranslational modification alters Lin-52 function in order to inactivate DRM activity. Additional studies will be required to determine whether such posttranslational modifications of Drosophila Lin-52 are required to regulate MMB/dREAM complex stability and function (Lewis, 2012).

Cyclin A-Myb-MuvB-Aurora B network regulates the choice between mitotic cycles and polyploid endoreplication cycles

Endoreplication is a cell cycle variant that entails cell growth and periodic genome duplication without cell division, and results in large, polyploid cells. Cells switch from mitotic cycles to endoreplication cycles during development, and also in response to conditional stimuli during wound healing, regeneration, aging, and cancer. This study used integrated approaches in Drosophila to determine how mitotic cycles are remodeled into endoreplication cycles, and how similar this remodeling is between induced and developmental endoreplicating cells (iECs and devECs). The evidence suggests that Cyclin A / CDK directly activates the Myb-MuvB (MMB) complex to induce transcription of a battery of genes required for mitosis, and that repression of CDK activity dampens this MMB mitotic transcriptome to promote endoreplication in both iECs and devECs. iECs and devECs differed, however, in that devECs had reduced expression of E2f1-dependent genes that function in S phase, whereas repression of the MMB transcriptome in iECs was sufficient to induce endoreplication without a reduction in S phase gene expression. Among the MMB regulated genes, knockdown of AurB protein and other subunits of the chromosomal passenger complex (CPC) induced endoreplication, as did knockdown of CPC-regulated cytokinetic, but not kinetochore, proteins. Together, these results indicate that the status of a CycA-Myb-MuvB-AurB network determines the decision to commit to mitosis or switch to endoreplication in both iECs and devECs, and suggest that regulation of different steps of this network may explain the known diversity of polyploid cycle types in development and disease (Rotelli, 2019).

Endoreplication is a common cell cycle variant that entails periodic genome duplication without cell division and results in large polyploid cells. Two variations on endoreplication are the endocycle, a repeated G/S cycle that completely skips mitosis, and endomitosis, wherein cells enter but do not complete mitosis and / or cytokinesis before duplicating their genome again. In a wide array of organisms, specific cell types switch from mitotic cycles to endoreplication cycles as part of normal tissue growth during development. Cells also can switch to endoreplication in response to conditional inputs, for example during wound healing, tissue regeneration, aging, and cancer. It is still not fully understood, however, how the cell cycle is remodeled when cells switch from mitotic cycles to endoreplication (Rotelli, 2019).

There are common themes across plants and animals for how cells switch to endoreplication during development. One common theme is that developmental signaling pathways induce endoreplication by inhibiting the mitotic cyclin dependent kinase 1 (CDK1). After CDK1 activity is repressed, repeated G / S cell cycle phases are controlled by alternating activity of the ubiquitin ligase APC/CCDH1 and Cyclin E / CDK2. Work in Drosophila has defined mechanisms by which APC/CCDH1 and CycE / Cdk2 regulate G / S progression, and ensure that the genome is duplicated only once per cycle. Despite these conserved themes, how endoreplication is regulated can vary among organisms, as well as tissues within an organism. These variations include the identity of the signaling pathways that induce endoreplication, the mechanism of CDK1 inhibition, and the downstream effects on cell cycle remodeling into either an endomitotic cycle (partial mitosis) or endocycle (skip mitosis). In many cases, however, the identity of the developmental signals and the molecular mechanisms of cell cycle remodeling are not known (Rotelli, 2019).

To gain insight into the regulation of variant polyploid cell cycles, two-color microarrays have been used to compare the transcriptomes of endocycling and mitotic cycling cells in Drosophila tissues (Maqbool, 2010). Endocycling cells of larval fat body and salivary gland have been shown to have dampened expression of genes that are normally induced by E2F1, a surprising result for these highly polyploid cells given that many of these genes are required for DNA synthesis. Nonetheless, subsequent studies showed that the expression of the E2F-regulated mouse orthologs of these genes is reduced in endoreplicating cells of mouse liver, megakaryocytes, and trophoblast giant cells. Drosophila endocycling cells also had dampened expression of genes regulated by the Myb transcription factor, the ortholog of the human B-Myb oncogene (MYBL2). Myb is part of a larger complex called Myb-MuvB (MMB), whose subunit composition and functions are mostly conserved from flies to humans. One conserved function of the MMB is the induction of periodic transcription of genes that are required for mitosis and cytokinesis. It was these mitotic and cytokinetic genes whose expression was dampened in Drosophila endocycles, suggesting that this repressed MMB transcriptome may promote the switch to endocycles that skip mitosis. It is not known, however, how E2F1 and Myb activity are repressed during endocycles, nor which of the downregulated genes are key for the remodeling of mitotic cycles into endocycles (Rotelli, 2019).

In addition to endoreplication during development, there are a growing number of examples of cells switching to endoreplication cycles in response to conditional stresses and environmental inputs. These cells will be called induced endoreplicating cells (iECs) to distinguish them from developmental endoreplicating cells (devECs). For example, iECs contribute to tissue regeneration after injury in flies, mice, humans, and the zebrafish heart, and evidence suggests that a transient switch to endoreplication contributes to genome instability in cancer. Cardiovascular hypertension stress also promotes an endoreplication that increases the size and ploidy of heart muscle cells, and this hypertrophy contributes to cardiac disease. It remains little understood how similar the cell cycles of iECs are to devECs (Rotelli, 2019).

Similar to the developmental repression of CDK1 activity to promote endocycles, it has been shown that experimental inhibition of CDK1 activity is sufficient to induce endoreplication in flies, mouse, and human cells. These experimental iECs in Drosophila are similar to devECs in that they skip mitosis, have oscillating CycE / Cdk2 activity, periodically duplicate their genome during G / S cycles, and repress the apoptotic response to genotoxic stress. This study uses these experimental iECs to determine how the cell cycle is remodeled when cells switch from mitotic cycles to endoreplication cycles, and how similar this remodeling is between iECs and devECs. The findings indicate that the status of a CycA-Myb-AurB network determines the choice between mitotic cycles and endoreplication cycles in both iECs and devECs (Rotelli, 2019).

This study has investigated how the cell cycle is remodeled when mitotic cycling cells switch into endoreplication cycles, and how similar this remodeling is between devECs and experimental iECs. Repression of a CycA-Myb-AurB mitotic network promotes a switch to endoreplication in both devECs and iECs. Although a dampened E2F1 transcriptome of S phase genes is a common property of devECs in flies and mice, this study found that repression of the Myb transcriptome is sufficient to induce endoreplication in the absence of reduced expression of the E2F1 transcriptome. Knockdown of different components of the CycA-Myb-AurB network resulted in endoreplication cycles that repressed mitosis to different extents, which suggests that regulation of different steps of this pathway may explain the known diversity of endoreplication cycles in vivo. Overall, these findings define how cells either commit to mitosis or switch to different types of endoreplication cycles, with broader relevance to understanding the regulation of these variant cell cycles and their contribution to development, tissue regeneration, and cancer (Rotelli, 2019).

The findings indicate that the status of the CycA-Myb-AurB network determines the choice between mitotic or endoreplication cycles (The CycA-Myb-AurB network regulates the choice between cell cycle programs). These proteins are essential for the function of their respective protein complexes: CycA activates CDK1 to regulate mitotic entry, Myb is required for transcriptional activation of mitotic genes by the MMB transcription factor complex, and AurB is the kinase subunit of the four-subunit CPC. While each of these complexes were previously known to have important mitotic functions, the data indicate that they are key nodes of a network whose activity level determines whether cells switch to the alternative growth program of endoreplication. The results are consistent with previous evidence in several organisms that lower activity of the Myb transcription factor results in polyploidization, and further shows that repressing the function of the CPC and cytokinetic proteins downstream of Myb also promotes endoreplication. Importantly, genetic evidence indicates that not all types of mitotic inhibition result in a switch to endoreplication. For example, knockdown of the Spc25 and Spc105R kinetochore proteins or the Polo kinase resulted in a mitotic arrest, not a switch to repeated endoreplication cycles. These observations are consistent with CycA / CDK, MMB, and the CPC playing principal roles in the mitotic network hierarchy and the decision to either commit to mitosis or switch to endoreplication cycles (Rotelli, 2019).

While knockdown of different proteins in the CycA-Myb-AurB network were each sufficient to induce endoreplication cycles, these iEC populations had different fractions of cells with multiple nuclei diagnostic of an endomitotic cycle. Knockdown of cytokinetic genes pav and tum resulted in the highest fraction of endomitotic cells, followed by the CPC subunits, then Myb, and finally CycA, with knockdown of this cyclin resulting in the fewest endomitotic cells. These results suggest that knocking down genes higher in this branching mitotic network (e.g. CycA) inhibits more mitotic functions and preferentially promotes G / S endocycles that skip mitosis, whereas inhibition of functions further downstream in the network promote endomitosis. Moreover, different levels of CPC function also resulted in different subtypes of endoreplication. Strong knockdown of AurB inhibited chromosome segregation and cytokinesis resulting in cells with a single polyploid nucleus, whereas a mild knockdown resulted in successful chromosome segregation but failed cytokinesis, suggesting that cytokinesis requires more CPC function than chromosome segregation. It thus appears that different thresholds of mitotic function result in different types of endoreplication cycles. This idea that endomitosis and endocycles are points on an endoreplication continuum is consistent with evidence that treatment of human cells with low concentrations of CDK1 or AurB inhibitors induces endomitosis, whereas higher concentrations induce endocycles. The results raise the possibility that in tissues of flies and mammals both conditional and developmental inputs may repress different steps of the CycA-Myb-AurB network to induce slightly different types of endoreplication cycles that partially or completely skip mitosis. Together, these findings show that there are different paths to polyploidy depending on both the types and degree to which different mitotic functions are repressed (Rotelli, 2019).

The findings are relevant to the regulation of periodic MMB transcription factor activity during the canonical mitotic cycle. Knockdown of CycA compromised MMB transcriptional activation of mitotic gene expression, and their physical association suggests that the activation of the MMB by CycA may be direct. The MMB-regulated mitotic genes were expressed at lower levels in CycA iECs, even though Myb protein levels were not reduced. This result is consistent with the hypothesis that CycA / CDK phosphorylation of the MMB is required for its induction of mitotic gene expression. Moreover, misexpression of Myb in CycA knockdown follicle cells did not prevent the switch to endoreplication, further evidence that CycA / CDK is required for MMB activity and mitotic cycles. While the dependency of the MMB on CycA was not previously known in Drosophila, it was previously reported that in human cells CycA / CDK2 phosphorylates and activates human B-Myb in late S phase, and also triggers its degradation. While further experiments are needed to prove that CycA / CDK regulation of the MMB is direct, interrogation of the results of multiple phosphoproteome studies using iProteinDB indicated that Drosophila Myb protein is phosphorylated at three CDK consensus sites including one, S381 that is of a similar sequence and position to a CDK phosphorylated site on human B-Myb (T447). The hypothesis is favored that it is CycA complexed to CDK1 that regulates the MMB because, unlike human cells, in Drosophila CycA / CDK2 is not required for S phase, and Myb is degraded later in the cell cycle during mitosis. Moreover, it is known that mutations in CDK1, but not CDK2, induce endocycles in Drosophila, mouse, and other organisms. A cogent hypothesis is that CycA / CDK1 phosphorylates Myb, and perhaps other MMB subunits, to stimulate MMB activity as a transcriptional activator of mitotic genes, explaining how pulses of mitotic gene expression are integrated with the master cell cycle control machinery. It remains formally possible, however, that both CycA / CDK2 and CycA / CDK1 activate the MMB in Drosophila. The early reports that CycA / CDK2 activates B-Myb in human cells were before the discovery that it functions as part of the MMB and the identification of many MMB target genes, and further experiments are needed to fully define how MMB activity is coordinated with the central cell cycle oscillator in fly and human cells (Rotelli, 2019).

Endocycles were experimentally induced by knockdown of CycA to mimic the repression of CDK1 that occurs in devECs. The data revealed both similarities and differences between these experimental iECs and devECs. Both iECs and SG devECs had a repressed CycA-Myb-AurB network of mitotic genes. In contrast, only devECs had reduced expression of large numbers of E2F1-dependent S phase genes, a conserved property of devECs in fly and mouse. In CycA iECs, many of these key S phase genes were not downregulated, including Cyclin E, PCNA, and subunits of the pre-Replicative complex, among others. This difference between CycA dsRNA iECs and SG devECs indicates that repression of these S phase genes is not essential for endoreplication. In fact, none of the E2F1 -dependent S phase genes were downregulated in Myb dsRNA iEC. Instead, the 12 E2F1-dependent genes that were commonly downregulated in Myb dsRNA iEC, CycA dsRNA iEC, and SG devEC all have functions in mitosis. These 12 mitotic genes are, therefore, dependent on both Myb and E2F1 for their expression, including the cytokinetic gene tum whose knockdown induced endomitotic cycles. This observation leads to the hypothesis that downregulation of the E2F transcriptome in fly and mouse devECs may serve to repress the expression of these mitotic genes, and that the repression of S phase genes is a secondary consequence of this regulation. These genomic data, together with the genetic evidence in S2 cells and tissues, indicates that in Drosophila the repression of the Myb transcriptome is sufficient to induce endoreplication without repression of the E2F1 transcriptome. The observation that both CycAdsRNA iECs and devECs both have lower CycA / CDK activity, but differ in expression of E2F1 regulated S phase genes, also implies that there are CDK-independent mechanisms by which developmental signals repress the E2F1 transcriptome in devECs (Rotelli, 2019).

The results have broader relevance to the growing number of biological contexts that induce endoreplication. Endoreplicating cells are induced and contribute to wound healing and regeneration in a number of tissues in fly and mouse, and, depending on cell type, can either inhibit or promote regeneration of the zebrafish heart. An important remaining question is whether these iECs, like experimental iECs and devECs, have a repressed CycA-Myb-AurB network. If so, manipulation of this network may improve regenerative therapies. In the cancer cell, evidence suggests that DNA damage and mitotic stress, including that induced by cancer therapies, can switch cells into an endoreplication cycle. These therapies include CDK and AurB inhibitors, which induce human cells to polyploidize, consistent with the fly data that CycA / CDK and the CPC are key network nodes whose repression promotes the switch to endoreplication. Upon withdrawal of these inhibitors, transient cancer iECs return to an error-prone mitosis that generates aneuploid cells, which have the potential to contribute to therapy resistance and more aggressive cancer progression. The finding that the Myb transcriptome is repressed in iECs opens the possibility that these mitotic errors may be due in part to a failure to properly orchestrate a return of mitotic gene expression. Understanding how this and other networks are remodeled in polyploid cancer cells will empower development of new approaches to prevent cancer progression (Rotelli, 2019).

The Dm-Myb oncoprotein contributes to insulator function and stabilizes repressive H3K27me3 PcG domains

Drosophila Myb (Dm-Myb) encodes a protein that plays a key role in regulation of mitotic phase genes. This study further refine its role in the context of a developing tissue as a potentiator of gene expression required for proper RNA polymerase II (RNA Pol II) function and efficient H3K4 methylation at promoters. In contrast to its role in gene activation, Myb is also required for repression of many genes, although no specific mechanism for this role has been proposed. This study now reveals a critical role for Myb in contributing to insulator function, in part by promoting binding of insulator proteins BEAF-32 and CP190 and stabilizing H3K27me3 Polycomb-group (PcG) domains. In the absence of Myb, H3K27me3 is markedly reduced throughout the genome, leading to H3K4me3 spreading and gene derepression. Finally, Myb is enriched at boundaries that demarcate chromatin environments, including chromatin loop anchors. These results reveal functions of Myb that extend beyond transcriptional regulation (Santana, 2020).

This study identified four novel roles for the Drosophila Myb oncoprotein. First, it was shown that Myb acts as a potentiator of target gene transcription, not merely as an epigenetic factor that maintains activated expression. Potentiation is associated with increased H3K4me3 as well as recruitment and pause release/elongation of RNA Pol II. Next, it was shown that in addition to binding promoter potentiator sites, Myb binds to insulator sites genome wide, is necessary to promote the binding of BEAF-32 and CP190 at Myb-shared sites, and is required for the function of at least a subset of insulators. Third, it was shown that Myb plays a critical role in formation and/or stabilization of H3K27me3 domains, preventing silent genes within these domains from becoming derepressed, in stark contrast to Myb's well-known role as an activator of gene expression. Finally, this study showed that Myb is enriched at TAD boundaries and chromatin loop anchors (Santana, 2020).

Similar to what has previously been reported, an enrichment was identified of cell cycle genes directly regulated by Myb, with the majority being G2/M but also several S phase genes (consistent with studies on B-Myb, supporting its role in S phase). Notably, a recent study reported that Cyclin A (the cyclin that regulates aspects of both S phase and G2/M) directly binds to Myb and is required for expression of many of its targets (Rotelli, 2019). Further supporting a role for Myb in transcriptional regulation, this study found that Myb is required for optimal levels of H3K4 promoter methylation, with Myb playing a key role in RNA Pol II dynamics. Analysis of the promoter regions of Myb-potentiated targets in Myb mutants revealed two distinct classes of targets: one that requires Myb for efficient recruitment of RNA Pol II and both that use Myb for efficient RNA Pol II pause release/elongation. Myb may use NURF to establish a nucleosomal-free region that allows recruitment of RNA Pol II, similar to what has been observed with GAF at several targets. Further suggesting a critical role for Myb in transcriptional regulation, the failure to potentiate key targets in Myb mutants (e.g., okr and Rad9 for S phase; Cap-D2 for chromosome condensation; mad2, Mps1, Pen, dgt6, and msd5 for M phase) likely explains the varied phenotypes observed in Myb mutants (including increased mitotic index, partially condensed chromosomes, aneuploidy, and S phase defects) (Santana, 2020).

Recent studies have shown that several components of dREAM can bind insulator sites, and RNAi depletion of Mip40, Mip130, and E2f2 resulted in impairment of enhancer-blocking function at specific sites. Furthermore, components of dREAM (Myb, Mip120, Mip130, Rbf1, E2f2, and DP) were shown to directly interact with dCTCF and/or CP190. Interestingly, double knockdown of dCTCF and CP190 resulted in loss of dREAM (Mip40, Mip120, Mip130, and E2f2) at some shared sites. The results presented in this study further extend these studies and demonstrate that Myb directly contributes to boundary insulator function (including sites located at divergently paired genes). Intriguingly, in many cases where one gene of a divergently paired genes (DPGs) is transcribed and the other is silent, it was found that Myb has no role in expression of the activated gene; rather, Myb prevents inappropriate activation of the silent gene. This cannot be explained by loss of the primary repressive arm of the Myb complex (E2f2/Rbfs) because knockdown of E2f2 or Rbf1/2 in Kc167 cells led to derepression of only 17 out of the 96 genes upregulated when Myb is knocked down. Furthermore, out of the 46 genes upregulated in SL2 cells upon knockdown of E2f2 or Rbf1/2, only 3 are upregulated in Myb mutant wing discs. Finally, the absence of E2f2 in 3rd instar larvae led to upregulation of genes from four DPGs; yet, these genes are not upregulated in Myb mutant wing discs. Of the insulator sites shown in this study to require Myb, the vast majority (89%) contain Myb but not E2f2 consensus sites, similar to what was observed in Kc167 cells for genes requiring Myb for repression (Santana, 2020).

Consistent with Myb playing a role in insulators that separate different chromatin neighborhoods, this study found a striking correlation between Myb occupancy and active, null, PcG H3K27me3 (all 99%), and APBS TAD boundaries (>93%). TAD boundaries have been associated with chromatin looping, and studies on CTCF and CP190 in vertebrates and invertebrates, respectively, have proposed or confirmed a role for these insulator factors in tethering the chromatin loops. Further work has shown that chromatin loop anchors can also exist within TADs. It is particularly noteworthy that Myb occupies the majority of chromatin loop anchor points identified in another study Eagen (2017), in addition to significant overlaps with several other chromatin factors (including dCTCF, Rad21, and Pc) at anchor points. Further work is needed to determine whether Myb plays a role in tethering chromatin loops, although it is intriguing to speculate that the loss of the H3K27me3 mark in Myb or dCTCF mutants could result from disrupting H3K27me3 loops, which, in turn, might diminish that region's ability to associate with an H3K27me3 nuclear subcompartment. The observation that dCTCF has been shown to directly interact with Myb further suggests that these two chromatin factors can collaborate, and a remarkable 73% of the anchors bound by dCTCF are also bound by Myb (similar to the Rad21-bound anchors, for which 75% are also occupied by Myb). It is postulated that B-Myb (the most closely related vertebrate Myb family member to Dm-Myb) and CTCF might be playing similar collaborative roles in humans (Santana, 2020).

Although a role for Myb in transcriptional activation has been recognized for some time, its role in insulator function is particularly intriguing, as this activity appears to be separate from its transcriptional role. Indeed, the chromatin changes observed at many DPGs and other insulator elements strongly support the direct role of Myb in promoting a barrier function between chromatin states (in part by promoting binding of BEAF-32/ CP190). However, unlike previously described examples of barrier insulator loss, which results in heterochromatin spreading into euchromatin, the opposite scenario were observed in Myb mutants, namely the spreading of H3K4me3, which may be enabled by the significant loss of the H3K27me3 mark. This is the very definition of barrier insulator function, namely, maintaining a discrete separation of chromatin environments. Whether Myb is playing a role in chromatin loop formation is unclear at this point, but its overlap with the majority of loop anchor points identified in at least one study in flies is intriguing. Further work will be needed to address Myb's role at the loop anchors and whether this role relates to stabilization of H3K27me3 domains and/or barrier insulator function. For example, directed mutagenesis of Myb-dependent insulator sites by using tools, such as CRISPR-Cas9, to specifically target Myb, dCTCF, or BEAF-32 sites can provide additional insight into the respective roles of these insulator factors. In addition, assay for transposase-accessible chromatin using sequencing (ATAC-seq) or chromosome conformation capture (3C)-type techniques performed in both controls and Myb mutants can be used to determine whether important chromatin/ 3D structure changes might underlie the altered transcriptional (Santana, 2020).

PROTEIN STRUCTURE

Amino Acids - 657

Structural Domains

The four regions of conservation shared between vertebrate and Drosophila Myb proteins consist of three imperfect tandem repeats (R1, R2, and R3) that comprise the DNA-binding domain (region I), a transcriptional activator domain (TA - region II), a leucine zipper (LZ - region III), and a negative regulatory domain (NR - region IV). An additional region of the mouse protein, encoded by an alternatively spliced exon, contains the majority of conserved region II. The DNA-binding domain contains the Drosophila Myb2 mutation. In this region, chicken and human sequences are identical to the mouse sequence. Region IV includes the Drosophila Myb1 mutation (Katzen, 1998)

The proto-oncogene c-Myb from Drosophila melanogaster is represented by a single locus at position 13E-F on the X chromosome, and is expressed in early embryos by transcription into two polyadenylated RNAs with lengths of approximately 3.0 and 3.8 kb. The gene may encode a protein with a molecular weight of at least 55,000 that shares a domain with c-myb (chicken) in which 91 of 125 (or 73%) of the amino acids are identical in the Drosophila and chicken genes. These findings represent the first rigorous identification of a Drosophila proto-oncogene that can encode what may be a nuclear protein; they set the stage for a genetic analysis of how Myb serves the normal organism (Katzen, 1985).

The Drosophila melanogaster homolog of c-myb contains two clusters of sequences homologous to vertebrate myb proteins, surrounded by sequences lacking homology. These results extend previous evidence for the existence of a D. melanogaster homolog of c-myb and identify two highly conserved and therefore presumably functionally important domains of c-myb proteins. DNA-binding experiments indicate that the NH2-proximal of the two homology regions functions as a DNA-binding domain. Based on the absence of the COOH-proximal homology region in truncated oncogenic derivatives of c-myb, it is likely that this homology region encodes a function whose loss is involved in activating the oncogenic potential of c-myb (Peters, 1987).

date revised: 22 May 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.