|
Gene name - germ cell-less
Synonyms - Cytological map position - 44E3 Function - chromatin component Keywords - gonadogenesis, transcriptional repression, nuclear envelope |
Symbol - gcl FlyBase ID: FBgn0005695 Genetic map position - 2R Classification - BTB/POZ domain protein Cellular location - nucleus |
During early Drosophila and C. elegans development, the germ cell precursors undergo a period of transcriptional quiescence. Germ cell-less (Gcl), a germ plasm component necessary for the proper formation of 'pole cells', the germ cell precursors in Drosophila, is required for the establishment of this transcriptional quiescence. While control embryos silence transcription prior to pole cell formation in the pole cell-destined nuclei, this silencing does not occur in embryos that lack Gcl activity. The failure to establish quiescence is tightly correlated with failure to form the pole cells. Furthermore, Gcl can repress transcription of at least a subset of genes in an ectopic context, independent of other germ plasm components. These results place Gcl as the earliest gene known to act in the transcriptional repression of the germline. Gcl's subcellular distribution on the nucleoplasmic surface of the nuclear envelope (Jongens, 1994) and its effect on transcription suggest that it may act to repress transcription in a manner similar to that proposed for transcriptional silencing of telomeric regions (Leatherman, 2002).
A period of transcriptional quiescence in the early germ cells is not only a conserved feature in Drosophila and C. elegans, but it also appears to be important for their development, since mutations that disrupt this quiescence affect the formation of the germline. The pie1 gene encodes a protein that acts as a transcriptional repressor in the early germ cell precursors of C. elegans. Embryos that lack pie1 activity fail to repress transcription in the germ cell precursors and also fail to form a proper germline. In Drosophila, the nanos and pumilio genes are required for transcriptional quiescence of the pole cells. In embryos that lack nanos or pumilio activity, the pole cells prematurely or inappropriately express genes and fail to develop into functional germ cells (Leatherman, 2002).
The Germ cell-less protein is a germ plasm component and specifically associates with those nuclei that enter the germ plasm, induce the formation of pole buds, and then are incorporated into the resulting pole cells (Jongens, 1992). In these nuclei, Gcl is localized to the nucleoplasmic surface of the nuclear envelope (Jongens, 1994), a localization necessary for its function (Robertson, 1999). Previously, examination of the germ cell-less null phenotype revealed that most embryos lacking maternally contributed gcl (hereafter called Δgcl embryos) form no pole cells; the remaining embryos form a very small number of pole cells (Robertson, 1999). This failure to form pole cells does not appear to be due to a defect in germ plasm formation, maintenance, or levels. Analysis of known germ plasm components in Δgcl embryos has failed to reveal any defects, suggesting that the failure to form pole cells in these embryos is due to a direct requirement for Gcl in this process (Robertson, 1999; Leatherman, 2002).
To investigate whether loss of gcl activity results in a failure to establish or maintain the state of transcriptional quiescence necessary for proper germ cell development, both Δgcl and control embryos were stained with the H5 antibody. This monoclonal antibody recognizes a phosphorylated form of RNA polymerase II that is associated with active transcription, and it has been used as a reliable marker for the transcriptional quiescence of the germ cell precursors of Drosophila and C. elegans (Leatherman, 2002).
Interestingly, in the control embryos, a difference in H5 staining is observed in pole bud nuclei, and this occurs at an earlier stage than what was previously reported (Seydoux, 1997). Pole buds first appear during nuclear cycle 9 and pinch off at the end of nuclear cycle 10 to form the pole cells. In many cycle 9 embryos, a slight decrease in H5 staining was observed in the pole bud nuclei (identified by their association with Vasa staining) compared to the somatic nuclei, and by cycle 10, the pole bud nuclei display a dramatic reduction in H5 staining, indicating that a state of transcriptional quiescence is being established prior to the formation of the pole cells (Leatherman, 2002).
In Δgcl embryos, the pole bud nuclei fail to become transcriptionally quiescent. When stained with the H5 antibody, most pole bud nuclei in the Δgcl embryos stained at levels comparable to the somatic nuclei. However, in a few scattered Δgcl pole bud nuclei, a reduction was observed in H5 staining similar to that seen in wild-type. Since the few successfully formed pole cells in Δgcl embryos have dramatically reduced H5 staining similar to control embryo pole cells (Robertson, 1999), it is speculated that the few silenced pole bud nuclei observed might be those that will become part of the few pole cells seen in Δgcl embryos (Leatherman, 2002).
To further explore this possibility, this loss of transcriptional quiescence was quantitated by counting the pole bud nuclei in control and Δgcl embryos at the pole bud stage (nuclear cycle 10) and noting whether they had reduced H5 staining. Between 7 and 14 pole bud nuclei per embryo were found in both control and Δgcl embryos, as is expected. In control embryos, nearly all nuclei had reduced H5 staining. However, in 50% of the Δgcl embryos, none of the pole bud nuclei displayed reduced H5 staining. This correlates well with the observed 48% of Δgcl embryos with no pole cells at the blastoderm stage (Robertson, 1999). Of the total number of Δgcl pole bud nuclei counted, only 11.9% had reduced H5 staining, indicating transcriptionally silenced nuclei. After pole cell formation, each pole cell then divides between 0 and 2 times, so the total number of transcriptionally silenced pole bud nuclei cannot be compared directly to the number of pole cells observed in Δgcl embryos. However, since 11.9% of the pole bud nuclei are silenced, a reduction to a similar percent would be expected in the number of pole cells in Δgcl embryos compared to control embryos. Since control embryos have an average of 23.4 pole cells at the blastoderm stage (Robertson, 1999), an average of (23.4 × 0.119), or 2.8 pole cells, would be expected to be present in blastoderm-stage Δgcl embryos based on the number of silenced pole bud nuclei. This number is identical to the observed (Robertson, 1999) average of 2.8 pole cells in Δgcl embryos at the blastoderm stage (Leatherman, 2002).
This strong quantitative correlation, in addition to the failure to find any other defects, including defects in the germ plasm, suggests that those few nuclei that successfully silence transcription in the pole buds of Δgcl embryos (as shown by loss of H5 staining) are those that will form the few functional pole cells. This implies that establishing transcriptional quiescence is a necessary step for pole cell formation; however, further experiments will be necessary to prove this type of causal relationship between transcriptional silencing and pole cell formation (Leatherman, 2002).
Since the H5 stainings indicate that pole bud nuclei in Δgcl embryos fail to become transcriptionally silent, it was speculated that it should therefore be possible to see misexpression of specific gene transcripts in these nuclei. The expression of two genes that are transcribed at this time, sisterless A (sisA) and sisterless B (sisB), was examined. These transcripts are ubiquitiously expressed in nuclei as early as nuclear cycle 8 but are repressed in pole bud nuclei. By using whole-mount in situ hybridization, it was found that sisA and sisB transcripts are present not only in somatic nuclei, but also in the majority of pole bud nuclei in Δgcl embryos, and this finding independently verifies that Δgcl embryos are deficient in transcriptional silencing in the pole bud nuclei (Leatherman, 2002).
The results described above indicate that gcl is required to repress transcription during the establishment of the germ cell lineage. To determine if this activity is dependent or independent of other germ plasm components, the effect of ectopically localizing Gcl on transcription was examined. Replacement of the 3'UTR of the gcl transcript with the 3'UTR of bicoid results in the anterior localization of gcl mRNA and protein to the anterior pole of the embryo (Jongens, 1994). In these 'hgb' embryos, a slightly variable but consistent decrease was found in the intensity of H5 staining in the anterior nuclei compared to control embryos throughout the syncytial blastoderm stage, and this decrease indicates that Gcl is sufficient to repress transcription ectopically. However, the anterior expression of Gcl clearly does not lead to complete silencing of the anterior nuclei, since some H5 staining persists (Leatherman, 2002).
The reduced H5 staining observed in the anterior of the hgb embryos could be due to global partial repression of all genes, or it could result from a specific subset of genes being severely repressed while others are unaffected. To distinguish between these possibilities, the expression was examined of specific genes whose expression pattern includes the anterior of the embryo, including sisA, sisB, tailless, huckebein, hunchback, and knirps. These genes are all independently activated by maternally contributed factors, so any effects on their transcription are likely to be direct rather than a consequence of an earlier defect. By using in situ hybridization, it was found that the early anterior expression domains of sisA, sisB, tailless, and huckebein are severely repressed in all of the hgb embryos examined, but no effect was seen on hunchback and knirps expression. These data suggest that the transcriptionally repressive effect of Gcl is not global, but rather specific to a subset of genes. Gcl is also present in a variety of tissues later in development (Jongens, 1992), at times when transcription is active, which further suggests a non-global mode of silencing (Leatherman, 2002).
If Gcl silences only selected genes, then the question arises as to how the pole cells accomplish what appears to be complete repression of mRNA transcription. Reporter genes driven by the strong Gal4-VP16 activator have been shown to be unable to be activated in the early germ cells. If only specific genes are repressed, it is unlikely that a novel transgene would be repressed by the same silencing mechanism, thereby arguing that there is widespread repression of transcription in early pole cells. Furthermore, observations of the compaction of pole cell nuclei, that is mediated by Gcl, suggest a more global mode of silencing than current studies. Pole cell nuclei are normally round in shape and more compact than somatic nuclei, which is consistent with chromatin silencing. Compaction increases when Gcl is overexpressed, and ectopic compaction occurs when Gcl is ectopically localized (Jongens, 1994). It seems unlikely that Gcl could accomplish such a global nuclear structure change just through the repression of transcription of a few genes (Leatherman, 2002).
One possible explanation for this discrepancy is the existence of other as yet unidentified factors that act in transcriptional silencing during pole cell formation which allow the formation of the few pole cells that form in Δgcl embryos. These other factors, while less effective initially than Gcl, might have a more global mode of action. This argument is supported by experiments that show that, when Oskar is ectopically localized to the anterior, which causes ectopic pole cell formation by ectopic localization of essential germ plasm components, hunchback transcription is repressed. Since Gcl does not repress hunchback transcription, there must be other factors that accomplish this. Another possibility is that general repression can be accomplished by silencing only a few genes. At the time of pole cell formation, there are very few genes known to be transcribed, so it is conceivable that specific mechanisms could exist for silencing all of them. Repression of these few genes could alter the developmental program in the pole bud nuclei, thereby resulting in more widespread silencing in pole cells (Leatherman, 2002).
Data on Nanos and its binding partner Pumilio support this view. These factors are also required for transcriptional silencing in Drosophila pole cells, although at a later stage than that at which Gcl is required. The pole cell migration defect in nanos mutant embryos is partially alleviated when sex-lethal, one improperly expressed transcript, is removed. This suggests that the number of genes that nanos causes to be transcriptionally repressed is small, since removal of only one gene can cause rescue. Furthermore, transcriptional activation of sex-lethal is dependent on the two genes shown to be repressed in the germ plasm by Gcl, sisA and sisB; this finding hints that sex-lethal is a key gene that must be silenced in the early pole cells in order for them to achieve their proper developmental fate (Leatherman, 2002).
Work on a mouse homolog of Drosophila gcl further suggests a specific, rather than global, mode of repression for Gcl. mgcl-1, a functional homolog of gcl, is highly expressed in spermatocytes ( Leatherman, 2000) at a time when transcriptional activity in these cells is high. mGcl physically interacts with the DP3α subunit of E2F (de la Luna, 1999). mGcl can inhibit progression through the cell cycle by repressing the E2F transcription factor, possibly due to its sequestration at the nuclear envelope. This work demonstrates a mechanism of transcriptional repression for a specific set of genes -- those that are transcriptionally controlled by E2F. However, it is unclear at this point whether this specific role for Gcl occurs in Drosophila, since Gcl does not bind to Drosophila DP, the only apparent isoform of this protein in the Drosophila genome (Leatherman, 2002).
While the mechanism by which Gcl accomplishes transcriptional silencing is not known, interesting parallels in budding yeast, S. cerevisiae, suggest a possible mechanism whereby Gcl could be working. In yeast, the nuclear periphery has been linked to transcriptional silencing activity, largely through studies of telomeric chromatin and the subtelomeric genes that are affected. MLP1 and MLP2, which are tethered to the nuclear envelope, provide the anchor by which the Yku70/Yku80 heterodimer, which binds to chromatin, brings telomeric chromatin into the perinuclear 'silent domain'. Once chromatin is at the nuclear periphery, silent information regulators 3 and 4 (Sir3 and Sir4) can gain access to the DNA and cause transcriptional silencing. Since Gcl localizes to the nuclear envelope (Jongens, 1992), it is speculated that it could be accomplishing transcriptional repression similarly to MLP 1 and MLP2 by anchoring chromatin to the nuclear periphery through protein binding partners. This model is currently being tested by looking at the function of binding partners of Gcl (Leatherman, 2002 and references therein).
A mouse homolog of gcl, named mouse germ cell-less-1 (mgcl-1) codes for a protein (mGcl-1) that has 36% identity and 56% similarity with the Drosophila Gcl, and both the Drosophila and mouse gcl gene products contained the BTB/POZ domain, which is a conserved protein-protein interaction domain. Genetic analysis clearly reveals that mgcl-1 can rescue the Drosophila gcl mutation (Kimura, 2003).
date revised: 25 September 2003
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.