Germline stem cells (GSCs) are unique in that they alone among adult stem cells pass information to the next generation. We here discuss the role of the recently discovered short RNAs called microRNAs in the germline. MicroRNAs are known to function as post-trascriptional regulators of gene expression and in that role to be essential both for development and homeostasis in many, if not all, higher organisms. MicroRNAs are needed for the creation of GSCs during embryogenesis, for the regulation of GSC function in sexually mature adults, and in the developmental programs that lead the daughter cells of GSCs to mature gametes in the course of gametogenesis. We also consider the tantalizing connection between microRNAs and a distinctive organelle of the germline, nuage. This connection is supported both by the presence of microRNA pathway components in nuage, and by the disruption of nuage and germline function by the ablation of microRNA pathway components. This suggests a role for microRNA in regulating the presumptive function of nuage: serving as a clearing house for parental mRNAs that must be either degraded, appropriately localized and stored in an untranslated state for future use, or translated for immediate use.
Germline Stem Cells (GSC), like all Adult Stem Cells (ASC), are distinguished by two essential functions: first, they must produce differentiating cells that serve to maintain adult tissues, and second they must replenish themselves in order to ensure the presence of ASCs for the future. Sexually reproducing animals possess a minimum of one type of adult stem cell: the germline stem cell (GSC), the precursor to haploid gametes that allow for sexual reproduction. These GSCs are part of an unbroken lineage of cells extending from the origin of sexual reproduction to the present.
In addition to producing differentiating cells that will become gametes and maintaining the GSC population, GSCs must protect the genome from damage in order to ensure its faithful transmission to the next generation. Recently discovered microRNAs, which are small, regulatory molecules governing gene expression, or closely related siRNAs (short interfering RNA) have been implicated in each of these essential processes in GSCs. Furthermore, germline cells in all examined animals are distinguished by the presence of morphologically distinct, electron dense bodies termed nuage, or in the male germline, chromatoid bodies (Extavour and Akam, 2003). These unique structures have been shown to both control the expression of transposable elements and to be rich in RNAi (RNA interference) components including microRNA (Findley et al., 2003). Nuage and chromatoid bodies are distinguished from the P-bodies of somatic cells in being perinuclear and rich in ribosomes (Extavour and Akam, 2003), while the P-bodies lack large ribosomal subunits, are spherical and can be found throughout the cytoplasm (Reviewed (Parker and Sheth, 2007)). They are also distinct from the germline specific yolk nuclei or Balbiani bodies which surround a core of Golgi elements, are rich in lipoproteins, and in yolk forming species, induce the formation of yolk (Reviewed (Guraya, 1979)). This review considers both the roles of microRNA/RNAi in the germline and the intriguing connection between microRNA and nuage. There are at least three closely related RNAi pathways in animals: miRNA, siRNA, and the most recently discovered germline specific pathway, piRNA (Piwi-interacting RNA) (Reviewed by Theurkauf in this volume). Three considerations complicate our understanding of the role of these short RNA mediated pathways in the germline: First, they share proteinous components or have extremely similar proteinous components, second, the short RNAs mediating these pathways are of similar length, and third, the components of all three of these pathways are found in the nuage.
2. Origin of the germline
Germline stem cells migrate to the gonad during embryogenesis where they establish a life long residence in a niche, a micro-environment where mutual communication between the GSCs and supporting stromal cells maintain the function and identity of the GSCs. GSC function and identity is also supported by specific intrinsic factors (Ward et al., 2006; Wong et al., 2005; Morrison and Spradling, 2008). Primordial Germ Cells (PGCs), the precursors of the GSCs, are specified early in embryogenesis or before by one of two separate mechanisms, either by preformation or by local signaling from the cellular environment. Preformation of the germline in many animals, including fly, is achieved by the maternal contribution of germ cell determining factors in the polar granule containing pole plasm of the oocyte. Preformation of the germline in worm, on the other hand, is the result of asymmetric cell division during early embryogenesis that serves to segregate the polar granules to a single cell which will produce the germline. Rather than preformation, mammalian PGCs are specified by cell to cell signaling from their local environment during the onset of gastrulation (Reviewed (Extavour and Akam, 2003; Saffman and Lasko, 1999)). In either case, whether the germline is specified by preformation or by cell signaling during embyogenesis, it is characterized by nuage, chromatoid bodies, or polar granules.
3. microRNA biogenesis and function
The production of microRNAs effector complexes can be separated into two parts: first, the nuclear production of pre-microRNA and second, the cytoplasmic production of microRNA and its assembly into Argonaute in the effector complex RISC (RNA Induced Silencing Complex). The precursors of microRNAs are usually expressed in the same way as protein coding genes. First they are transcribed by RNA polymerase II, then they are polyadenylated and capped to form pri-miRNA (Cai et al., 2004). Exceptions to this rule are: that microRNAs can be transcribed by pol III from within repetitive elements (Borchert et al., 2006), or transcribed from intronic regions of coding genes (Rodriguez et al., 2004), and they can be transcribed as polycistrons (Lee et al., 2002). These pri-microRNA transcripts fold into long hairpin structures as a result of their partial self complementarity and are then cleaved in the nucleus by the microprocessor complex which contains both the RNAse III enzyme Drosha and its essential cofactor Pasha/DGCR8. The resultant short hairpin shaped loops with imperfect self complementarity between the two arms of the hairpin are called “pre-miRNA” (Denli et al., 2004; Gregory et al., 2004). Some microRNAs, termed mirtrons, bypass Drosha processing by spanning entire introns which are spliced to their pre-miRNA form (Ruby et al., 2007; Okamura et al., 2007; Berezikov et al., 2007). Pre-miRNA are exported from the nucleus by Exportin 5 in a Ran/GTP dependent fashion (Lund et al., 2004).
In the cytoplasm pre-microRNA is cleaved by another RNAse III class enzyme, Dicer (Hutvagner et al., 2001). The resulting RNA duplex, miRNA/miRNA* (star strand), has 3’ two nucleotide over-hangs and 5’ phosphate moieties characteristic of RNAse III cleavage. The miRNA strand is then loaded into an Argonaute protein resident in RISC where it serves to identify partially complimentary sequences in the 3’UTR of mRNA, thereby enabling the RISC complex to repress translation of cognate messages (Forstemann et al., 2007; Grishok et al., 2001) (Reviewed (Peters and Meister, 2007)). Nuage is usually perinuclear, is often associated with nuclear pores, and it is enriched for components of the cytoplasmic processing and silencing pathway as well as maternal mRNAs. This supports the notion that nuage is a germline specific processing center for parental mRNA (Kotaja et al., 2006). The means of microRNA mediated translational repression are various: by block of translation (Wightman et al., 1993), either post-initiation (Olsen and Ambros, 1999) or pre-initiation (Pillai et al., 2005), degradation via de-adenylation (Bagga et al., 2005; Wu et al., 2006), or by sequestration in specialized structures called P-bodies (Reviewed (Liu, 2008)). In some cases the translational repression is reversible (Bhattacharyya et al., 2006). Paradoxically, microRNAs can also up-regulate the expression of their targets in mouse (Vasudevan et al., 2007).
4. The three ages of the germline
It will be convenient for us to consider the role of microRNAs in three distinct developmental stages of the germline: as PGCs en route to becoming GSCs, as GSCs engaged in the ASC defining functions of tissue production and stem cell self renewal, and as the differentiating products of the GSCs undergoing gametogenesis. We will also discuss the potential connection between nuage and short RNAs during each of these three stages of germline development.
4.1. microRNAs in PGCs
PGCs are characterized by polar granules or nuage, which contain many of the same factors including members of the Argonaute family (Harris and Macdonald, 2001). This raises the question of whether nuage might be the precursor of polar granules, but live imaging studies have failed to demonstrate such a relationship (Snee and Macdonald, 2004).
In mouse, it has recently been reported that ablating Dicer specifically in the PGCs shows the microRNA pathway to be essential for the proliferation of PGCs, and further that the PGCs have a distinctive microRNA expression profile with conserved microRNA clusters known to promote cell cycling (miR-17–92 cluster, and miR-290–295 cluster) being highly expressed in all PGCs while expression of other miRNAs (miR-141, −200a, −200c and −323) decreased progressively as the PGCs developed in embryo. Other microRNAs, (let-7a, d, e, f and g), as well as miR-125a and miR-9 increased in male but not in female PGCs over the course of embryonic development (Hayashi et al., 2008).
In fly, the amount of germline determining pole plasm, hence the number of pole cells which give rise to the PGCs, can be increased by increasing the expression of PIWI while the number of pole cells, precursors of the PGCs, can be reduced by ablating the amount of either Dicer or dFMRP, the fly's homolog of the Fragile X Mental Retardation Protein (Megosh et al., 2006). dFMRP has been biochemically associated both with Ago1 and with specific microRNAs, some of which have been shown to be essential for the assembly of FMRP on target mRNA in mammals (Plante et al., 2006). Furthermore, PIWI, dFMRP, and Dicer coimmunoprecipitate from polar granule enriched fractions (Megosh et al., 2006). Here again, the close association of factors instrumental in the piRNA and microRNA pathways makes their similar mechanisms difficult to deconvolve.
4.1.1. Inhibition of miRNA activity
One peculiarity of microRNA gene expression worth mentioning is that it has been observed that protein coding genes near microRNA coding genes are enriched for two properties: predicted targets for the local microRNA and low levels of expression (Stark et al., 2005). This seems at odds with the more general observation that conserved target/microRNA pairs are rarely co-expressed (Stark et al., 2005; Farh et al., 2005) (Reviewed (Massirer and Pasquinelli, 2006)). The PGCs offer one exciting exception to the rule that targets and microRNAs are not co-expressed: during fish embryogenesis, in PGCs, essential germline genes nanos-1 and TDRL7 are co-expressed in the presence of cognate miR-430 thanks to the microRNA abrogating influence of Dnd-1(Dead eND-1) that binds sequence specific sites in the 3’UTR of nanos-1 and TDRL-7. Elsewhere in the developing embryo, miR-430 is repressive and is responsible for clearing the embryo of maternal transcripts during the maternal/zygotic transition. Similarly, the human gene LATS2 is protected from miR-372 by the presence of Dnd-1 in a germline derived cancer cell line (Giraldez et al., 2006; Kedde et al., 2007). A similar role in clearing maternal mRNA from the embryo has recently been reported for the miR-309 cluster in fly (Bushati et al., 2008). Other means of repression of the activities of miRNA in another stem cell type has recently been elucidated. Lin-28 has been shown to selectively block the processing of let-7 in ESC at the level of Drosha processing of pri-miRNA (Viswanathan et al., 2008). It remains to be seen whether similar post transcriptional but pre-effector complex regulatory mechanisms exist in the germline.
4.2. microRNA in GSC function
After the GSCs have made themselves at home in their niche, the microRNA pathway continues to play a vital role. Here, as in PGCs, nuage is present. In flies, microRNA has been shown to be needed for two functions: to promote GSC cycling (Hatfield et al., 2005) and for maintenance of the GSC in their niche (Park et al., 2007; Shcherbata et al., 2007; Yang et al., 2007; Jin and Xie, 2007). The compromised ability of flies to retain GSCs in their niche upon loss of the microRNA pathway is conditional and reveals a mysterious plasticity during late development (pupal) which adult flies lack. When Dicer-1 and hence miRNA biogenesis, is eliminated during development, the GSC remain in their niche, albeit with a greatly slowed cell cycle. However, when Dicer-1 is lost during adulthood, the GSCs are lost (Shcherbata et al., 2007; Jin and Xie, 2007). Intriguingly, eliminating the microRNA bantam phenocopies the Dicer-1 knockout making bantam, a microRNA found in flies but not mammals, the only single microRNA known so far to be essential for GSC function (Shcherbata et al., 2007). Further studies will be needed to reveal the cause of this youthful plasticity. It is suggestive that loss of an essential component of the TGF-β pathway, Mad, exhibits the same dual phenotypes and genetically interacts with the microRNA pathway. The youthful ability of developing flies to overcome the loss of Dicer-1 is absent in a Mad± background. Similarly, Dicer-1± flies are unable to overcome the loss of Mad during development (Shcherbata et al., 2007). Further evidence of the need for the microRNA pathway in GSCs comes from the observation that loss of Ago1, a microRNA binding protein and an essential factor for microRNA induced gene silencing in fly, results in loss of GSCs, while over expression of Ago1 results in supernumerary GSCs, suggesting that the effector stage and not microRNA biogenesis is the limiting step in microRNA dependent regulation (Yang et al., 2007).
Regardless of whether the microRNA pathway in fly's GSCs is ablated in adults or during development, the cell cycle is disturbed. Female dicer-1 GSC clones exhibit a low cell cycle index, an elevated frequency of Dacapo/p21/p27 and CycE, and reduced frequency of cycB expression, suggestive of defects in G1/S transition. Furthermore, reducing the level of the CKI, dacapo rescues this cell cycle defect, supporting a model in which dacapo expression is suppressed by the presence of microRNAs. Consistent with this model, a dacapo gene bearing a truncated 3’UTR is insensitive to the presence of Dicer-1. Furthermore, this shows that the promoter of dacapo is not instrumental in it’s microRNA dependent regulation.
4.2.1. microRNA and GSC chromatin states
Chromatin modifications unique to stem cells have been observed in various stem cell types (Reviewed (Spivakov and Fisher, 2007)) including GSCs in flies (Maines et al., 2007) and mammals (Sasaki and Matsui, 2008). Importantly, the microRNA pathway has been shown to play a part in defining chromatin state. De novo methylation in mouse embryonic stem cells was shown to be regulated by the transcriptional repressor Rbl-2 which is in turn targeted by the conserved miR-290 microRNA cluster. Strikingly, the same miRNA cluster was shown to be highly expressed in spermatagonia and proliferating PGCs (Benetti et al., 2008) suggesting a role specific to pre-differentiated states. Dicer has also been shown to be essential for the maintenance of chromatin state in a hybrid cell line (Fukagawa et al., 2004). Another intensively studied epigenetic mechanism with GSC specific roles is the silencing of developmentally important loci by the Polycomb Group (PcG) protein complex. Among the components of the RNAi pathway which are required for the nuclear co-localization of the sequences on which PcG assembles, Polycomb Response Elements (PREs) in fly is Ago1 (Grimaud et al., 2006), the silencing effector of the microRNA pathway. Notably, components of the PcG are up-regulated in the GSCs of fly compared with somatic Kc cells (Kai et al., 2005).
4.3. microRNA in GSC differentiation
The differentiating germline produced by the asymmetrical division of GSCs is also dependent on the microRNA pathway. In mouse, Dicer is essential for oocytes as germline specific knockouts of Dicer arrest in meiosis I and exhibit both chromosome segregation defects and multiple disorganized spindles. It is not clear what short RNAs might be responsible for this phenotype as both siRNA and miRNA are processed by the single Dicer found in mouse and human (Murchison et al., 2007). The authors of this study posit two hypothesis: first, that loss of centromeric repeat derived siRNAs could prevent the establishment of appropriate centromeric chromatin structure for assembly of the kinetochore, and second, that the loss of microRNAs could deregulate the levels of gene products needed for successful meiosis. In support of the second hypothesis, they point out that analysis of the transcripts found in developing oocytes lacking Dicer are enriched for genes implicated in microtubule related processes including twelve genes with predicted microRNA targets, notable among them being the GTPase Ran whose activity is known to be directly regulated by the condensation state of chromosomes (Murchison et al., 2007). In counter point, it has more recently been shown that endogeneous siRNAs (endo-siRNA) derived from the annealing of pseudogene transcripts with their coding paralog transcripts regulates gene expression in mouse oocytes (Tam et al., 2008). Drosophila however seem to confine the role of endo-siRNA to somatic tissue (Czech et al., 2008; Okamura et al., 2008; Ghildiyal et al., 2008; Kawamura et al., 2008). In the mouse's male germ line it has recently been shown that in germline specific Dicer knockouts spermatogonia exhibited poor proliferation. Interestingly, many of the same cell cycle promoting microRNAs that are up regulated in PGC (see above) are also up regulated in developing spermatagonia (Hayashi et al., 2008).
Among the jobs that must be completed during oogenesis is the maternal deposition of mRNAs which will be translated before the onset of zygotic transcription (Reviewed (Saffman and Lasko, 1999)). Maternally deposited microRNAs have also been found in fly (Liu et al., 2007) and mammals (Tang et al., 2007) but not in fish (Chen et al., 2005). One study suggests that maternally deposited microRNAs are needed for zygotic function in mouse. When oocytes were depleted of microRNA by ablating Dicer in the germline, the resultant zygotes failed to complete the first cell division. It is, however, unclear as to whether this defect in the zygote is a consequence of lack of microRNA function in the zygote or a result of disorganization of the oocyte (see above) caused by lack of Dicer during oogenesis. This same study detected the presence of many microRNAs in the mature Dicer+ oocytes (Tang et al., 2007). MicroRNA is also plentiful in mouse sperm; but these miRNAs were detected only at very low levels in newly fertilized egg and furthermore, failed to down regulate maternal mRNAs with predicted target sites (Amanai et al., 2006). Evidently, paternally contributed microRNA plays no major role in the zygote.
In fly, mutations in many components of the RNAi pathway, which as we have noted shares components among its various branches, result in similar axial patterning defects in oocytes (Findley et al., 2003; Cook et al., 2004; Lim and Kai, 2007). Interestingly, miR-280 is predicted to regulate oskar and kinesin heavy chain, both of which are essential to microtubule organization and axis specification (Stark et al., 2003). One possibility is that the special organization needed for nuage to perform its putative function in authorizing and directing the localization of mRNAs is compromised by this disruption of the cytoskeleton.
Three observations: first that nuage-like electron dense bodies are ubiquitous in germ cells (Extavour and Akam, 2003), second, that many components of RNA silencing pathways are found in these electron dense bodies (Findley et al., 2003; Lim and Kai, 2007), and third that nuage is perinuclear and often associated with nuclear pores (Eddy, 1975), have together lead to the working hypothesis that nuage is serving as a clearing house for RNA being exported from the nucleus in the germline. The decisions that would need to be made in order to coordinately silence, store, or express the separate messages in the transcriptome might be made in this mysterious germline defining structure (Findley et al., 2003).
Thus GSCs require the ancient and conserved microRNA pathway in order to establish their identity during embryogenesis, again require the microRNA pathway in order to function in the sexually mature adult, and yet again so that they may differentiate into the gametes destined to seed the next generation. In all three of these processes, microRNAs are needed so that GSCs may serve their function as the Fountain of Youth. We anticipate rapid discovery of the specifics of microRNA function in GSCs and as well as what role nuage plays in those functions.
We thank members of Ruohola-Baker laboratory for useful discussions. This work was supported by AHA fellowships for S. H.R., and MOD and NIH grants and the Tietze award for HR-B.
This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.