Sexual reproduction is the main reproductive strategy of the overwhelming majority of eukaryotes. This suggests that the last eukaryotic common ancestor was able to reproduce sexually. Sexual reproduction reflects the ability to perform meiosis, and ultimately generating gametes, which are cells that carry recombined half sets of the parental genome and are able to fertilize. These functions have been allocated to a highly specialized cell lineage: the germline. Given its significant evolutionary conservation, it is to be expected that the germline programme shares common molecular bases across extremely divergent eukaryotic species. In the present review, we aim to identify the unifying principles of male germline establishment and development by comparing two very disparate kingdoms: plants and animals. We argue that male meiosis defines two temporally regulated gene expression programmes: the first is required for meiotic commitment, and the second is required for the acquisition of fertilizing ability. Small RNA pathways are a further key communality, ultimately ensuring the epigenetic stability of the information conveyed by the male germline.
Male germline establishment and development
The life cycle of plants alternates between two morphologically distinct generations: the sporophyte and the gametophytes. During the course of evolution, the sporophytic phase became dominant, with the short-lived gametophytes arising, at the end of ontogeny, within structures on the sporophyte. It is during the development of the gametophytes that the plant germline is specified. Animals also display considerable variation in the timing of germline specification; this lineage can be established during early or late embryogenesis or even throughout adult life, as seen in plants (for a review, see ). Interestingly, in the basal animal lineages, germ cells arise from precursors also capable of generating somatic cell types. This fluid boundary between germline and soma finds a parallel with what is observed in plants, suggesting the conservation of key developmental mechanisms.
In flowering plants, male gametophytes are formed in specialized reproductive organs of the flower: the anthers. Archesporial cells in the stamen primordium divide periclinally to produce an outer PPC (primary parietal cell) and an inner PSC (primary sporogenous cell) (Figure 1). The PPC differentiates into the tapetum, a sterile nourishing tissue that is essential for the early stages of pollen development. Consecutive divisions of the PSC generate a central mass of microsporocytes, the PMCs (pollen mother cells). PMCs remain interconnected by cytoplasmic bridges, the cytomictic channels. These channels result from the continuity of the inner side of the cytoplasmic membrane (symplast) of neighbouring cells and are believed to be necessary for the subsequent synchronization of PMC development. Such synchronization will regulate meiosis, a process that will ultimately result in the formation of four microspores (for a review, see ). Each microspore will undergo an asymmetric mitotic division, PM I (pollen mitosis I), to form a bicellular structure in which the generative cell (germ cell) is embedded within the cytoplasm of the vegetative cell (pollen grain) via the formation of an endomembrane. This asymmetric division is determinative for the establishment of the plant germline since it results in two daughter cells with distinct fates. More specifically, the generative cell gives rise to the male germline, whereas the vegetative cell (somatic) serves a supporting role in germline development and fertilization. Some isolated mutants show the importance of the asymmetric cell division for proper germline development. In scp (sidecar pollen) mutants, microspores divide symmetrically followed by a second asymmetric division of only one of the daughter cells, thus producing mature pollen with an extra vegetative cell . In the case of gem1 (gemini pollen1) mutants, phenotypes ranging from asymmetric to symmetric microspore division can be observed, leading to a variable acquisition of generative cell fate depending on the degree of division asymmetry . Germ cell fate therefore appears to be dependent on the regulated partitioning of cytoplasmic germline determinants resulting from the asymmetric PM I division . This process finds clear parallels in animals that specify their germline via a pre-formation mechanism. In such cases, the acqu-isition of germ cell fate in the developing embryo is the result of an asymmetric distribution of maternally inherited determinants (germ plasm), either before or immediately after fertilization . In angiosperms, after PM I, the generative cell undergoes PM II (pollen mitosis II), generating two identical male gametes, the SCs (sperm cells). Structurally, a tricellular pollen grain enfolds the two non-motile SCs in close association with the nucleus of the vegetative cell. This structural assembly is defined as the MGU (male germ unit) . During pollen tube growth, the VN (vegetative nucleus) is positioned close to the tube tip, with the SCs trailing behind. Seeing that defects in MGU organization seem to affect its migration efficiency through the pollen tube , the specific intercellular association between the vegetative cells and SCs is considered important for male germline function. In this process, the cytoplasmic projection between the VN and SCs may represent a preferential communication route for possible metabolic exchanges . Tellingly, membrane proteins such as TETRASPANINS 11 and 12 have been shown to outline this cytoplasmic projection and also the adhesion microdomain between the contiguous SCs . After delivery to the female reproductive tissue, one of the SCs will fuse with the egg cell, giving rise to the zygote, whereas the second SC fuses with the central cell, thus initiating development of the endosperm, a nourishing tissue equivalent to the mammalian placenta.
Unifying principles of male germline establishment and development
When comparing male gametogenesis in animals with that of plants, the conservation of several major developmental hallmarks becomes clear. First, the conservation of mitosis and meiosis as a means of ensuring germ cell proliferation and reductive chromosome segregation respectively. More specifically, before meiotic entry, animal male germ cells engage in active mitotic expansion starting from a germline stem cell population. But, in contrast with plants, animal post-meiotic cells are unable to proliferate and instead acquire motility via the differentiation of a flagellum or an equivalent structure . Male gamete motility, although present in lower plants, was subsequently lost in most gymnosperms and in all angiosperms. Cytoplasmic continuity between differentiating male germ cells (or their precursors) is another conserved feature when comparing animal with plant male gametogenesis . In animals, such cytoplasmic sharing may be restricted to the initial mitotic divisions or also include the meiotic products . In plants, the symplastic continuity between PMCs regulates meiotic entry, and it can be speculated whether the physical contact between the two SCs in the MGU can actually extend to some degree of cytoplasmic sharing. In both plants and animals, this continuity may be related to a developmentally regulated allocation of limited cytoplasmic resources and/or represent a means of ensuring concerted cell development. Finally, the need for a structural association between the somatic and germ cell lineages is an equally conserved feature of plant and animal gametogenesis. The conservation of such support might stem from the probable existence of convergent evolution seeing that the last common ancestor to both plants and animals was most likely to be unicellular. The soma–germline associations may only be patent during the initial regulation of the germline stem cell population or also serve a supporting role throughout all of gametogenesis, as in our species . More specifically, in mammals, the somatic Sertoli cells support gametogenic progression by a continual direct contact with developing germ cells. In higher plants, the somatic VN supports the fertilizing ability of the enclosed SCs (via pollen tube formation) and is also involved in transposon silencing in the germline (see below). Irrespective of the system, the conservation of the soma–germline associations is likely to be indicative of a common need to establish a precise cellular environment for successful male gametogenesis.
Transcriptional programme of male germline development
Whereas the cytology of male germline development has been well characterized in both plants and animals, the actual molecular mechanisms driving this differentiation are less clear. In fact, only recently has the male germline gene expression programme started to be unveiled. In maize, gene expression profiling of different anther developmental stages identified three distinct gene repertoires. These consisted of genes expressed either before, during or after meiosis, illustrating the specificity of the gene expression programme in relation to meiotic progression . In Arabidopsis, pollen development is also transcriptionally modulated and the progression from microspores (post-meiotic cells) to mature pollen is characterized by a large-scale repression of early programme genes and the activation of a late gene expression programme . The male meiotic cells, the PMCs, share the expression of more than 1000 orthologous genes with equivalent cells in the mouse and fission yeast . Functional classes such as DNA repair and replication were significantly enriched among these conserved genes. This observation shows that some regulatory mechanisms have been evolutionarily conserved and that meiosis defines a change in the transcriptional profile of germline development. In fact, in animal spermatogenesis, major transcriptional changes also occur before meiotic entry. Studies in mammals and insects indicate that the pre-meiotic male germline cells activate a specialized transcriptional programme [16,17]. These cells produce transcripts that will be required later on (even post-meiotically) for normal cellular development, coding for genes belonging to three major functional classes: metabolism, cytoskeleton and chromosome organization.
Arabidopsis mature pollen grains show a unique transcriptome consisting of 6587 genes . The enriched mRNAs are related to signalling, cell wall metabolism and cytoskeleton dynamics, reflecting the pollen grain's commitment to germination and tube growth [18,19]. This trend is also patent in tobacco, where gene expression profiles during different stages of germination indicate a gradual increase in gametophyte-specific genes, such as regulators of cell tip expansion . Overall, data on pollen transcriptomes have clearly provided important insights into the molecular pathways involved in germ cell delivery. Yet, owing to the profound structural association between germ cells and the somatic vegetative cell, pollen analysis is unsuited for identifying the specificities of the germline transcriptome. This issue can only be satisfactorily addressed via the successful isolation of SCs from the remaining gametophyte. Not surprisingly, the first studies on SC-specific gene expression were performed in plant species where gametes are more accessible for isolation in sufficient amounts, as in the case of Zea mays and Lilium longiflorum [21,22]. These studies provided initial evidence that, unlike previously thought, mature male gametes carry a diverse repertoire of transcripts. Transcriptome studies on growing pollen tubes argue for the existence of de novo transcription at this stage [23,24], but whether this capability is restricted to the VN or is also patent in the male gametes remains to be analysed. In this regard, animal data are less ambiguous, with mature sperm in these species being largely considered transcriptionally silent . The development of a method for SC isolation in Arabidopsis allowed us to perform the most comprehensive transcriptional study of mature male gametes in this species to date. This led us to the identification of 5829 sperm-cell-expressed genes [26,27]. Of these, approximately 2400 showed enriched expression in this particular cell type. The enriched transcripts coded for products involved in DNA repair, ubiquitin-mediated proteolysis and cell cycle progression. These classes are also detected in rice SCs [28,29] and even in human male gametes , pointing to the evolutionary conservation of specific molecular processes in the male germline. In Arabidopsis, several male-germline-specific transcripts were also identified, including MGH3 (MALE GAMETE-SPECIFIC HISTONE H3) and HAP2 (HAPLESS2) [also known as GCS1 (GENERATIVE CELL-SPECIFIC PROTEIN 1)]. MGH3 is specifically expressed in both the generative cell and the SCs and has a major role in chromatin remodelling . Indeed, in both plants and animals, the chromatin of mature male germ cells is extensively remodelled via the incorporation of germline-specific histone variants (or other chromatin-bound proteins) . Such mechanisms are likely to be associated with precise structural requirements of chromatin architecture that may or may not associate with gene expression regulation. HAP2 (GCS1) encodes a sperm-cell-specific surface protein required for pollen tube guidance and gamete fusion . Loss of function of HAP2 (GCS1) orthologues in Chlamydomonas and Plasmodium also block gamete fusion . Its evolutionary conservation suggests that the functions of HAP2 (GCS1) in gamete fusion were established in the ancestor of modern plants, protists and animals . Finally, DUO1 (DUO POLLEN1) is specifically expressed in the male germline soon after PM I and encodes a R2R3 MYB transcription factor . DUO1 controls the expression of genes with broad cellular functions such as transcription regulation, protein metabolism and transport [37,38]. Interestingly, among the DUO1-regulated genes we can find several male germline determinants such as the aforementioned MGH3 and HAP2 (GCS1), and also GEX2 (GAMETE-EXPRESSED PROTEIN2), TIP5;1 (TONOPLAST INTRINSIC PROTEIN 5;1) and DAZ1–3 (DUO1-ACTIVATED ZINC FINGER1–3). Thus DUO1, by promoting the acquisition of a post-PM I germline transcriptional programme, can be considered a major regulator of SC differentiation in plants.
It has been shown in plants that transcripts stored in SCs can potentially be delivered upon fertilization and have a role in early embryo development. Such evidence came from the identification of the SSP (SHORT SUSPENSOR) transcripts which accumulate in Arabidopsis SCs, but whose encoded protein is only detected in the zygote (hence after fertilization) . SSP is a protein kinase known to activate the YODA mitogen-activated protein kinase pathway that regulates the first (asymmetric) division of the embryo. In animals, it is also a sperm-borne factor that triggers the onset of embryogenesis. This activation factor has already been identified in mammals and corresponds to PLCζ (phospholipase Cζ) (for a review, see ). Nevertheless, there is no evidence supporting the need for an egg-specific translation of this activating factor. PLCζ is responsible for the generation of intracellular Ca2+ elevations that ultimately result in egg activation. In vitro fertilization assays in plants have shown that, similarly to animals, the fusion between sperm and egg cell also results in changes in intracellular Ca2+ levels. However, in plants, this mechanism is insufficient for egg activation [41,42], suggesting that the onset of embryogenesis is dependent on additional factors. RNAs that accumulate in SCs (such as those for SSP) may be one of these factors. Could animal sperm-derived RNAs play a role in embryo development, as speculated in plants? Certainly, there is ample opportunity for such a mechanism: mature mammalian sperm contain a diverse RNA population, comprising not only mRNAs, but also different classes of small RNAs such as miRNAs and piRNAs (Piwi-interacting RNAs) [43,44]. Nevertheless, the successful cloning of animals using somatic cell nuclear transfer provides proof of concept that these sperm molecules are not essential for development to term. It should be noted that the low success rate of animal cloning procedures does not preclude a modulating role for sperm-derived RNAs during the initial stages of embryo development, as suggested previously .
Small RNA-mediated epigenetic regulation of the male germline
Recently, a complex repertoire of small RNAs has been detected in mature male gametes of both plants and animals. Given their role in the modulation of gene expression, it is tempting to speculate that they may serve instrumental roles in gametogenesis and in early post-fertilization development. Small RNAs are abundant in the pollen of Arabidopsis and rice and can be divided into two major classes: miRNAs [46,47] and siRNAs [48–50]. Both serve as regulators of gene expression, either transcriptionally or post-transcriptionally.
miRNAs are single-stranded molecules, usually 21–22 nt long, that, although synthesized by specific genes, can regulate the translation of mRNAs derived from several others. They are initially produced in the form of larger RNA precursors, which are later processed [in plants by DCL1 (DICER-LIKE1)] to yield mature miRNAs. These molecules are then recruited by an AGO (Argonaute) protein [part of the RISC (RNA-induced silencing complex)] and are guided to their mRNA targets to initiate endonucleolytic cleavage or translational inhibition . Studies have shown not only that miRNAs are abundant in pollen, but also that their precursors and target cleavage products can be detected . In isolated Arabidopsis SCs, miRNAs belonging to 83 different families could be identified , with recent studies using artificial constructs indicating that, at least in male gametes and like what was seen in animals, these molecules may also act through translational repression . Similarly, miRNAs are abundantly expressed during animal spermatogenesis, where they play important roles in the spatiotemporal regulation of gene expression (for a review, see ). It should be noted that, unlike in mammals, where miRNAs have been shown to play key roles in early male germ cell proliferation and late spermatogenesis, in plants the putative developmental relevance of any natural miRNA–target pair predicted to exist in the male gametophyte lacks experimental validation. In this context, it is also intriguing that dcl1 heterozygous mutant plants (in which half of the pollen do not express DCL1), have no discernible male transmission deficiencies .
siRNAs occur naturally in plants and lower animals, although their occurrence in mammals is still controversial. In plants, siRNAs originate from dsRNAs before being processed into 19–25-nt-long siRNAs. These molecules are instrumental for the reprogramming of the male germline, acting in the silencing of TEs (transposable elements). In Arabidopsis, certain TEs are reactivated in the VN, coincident with the loss of DDM1 (DECREASE IN DNA METHYLATION 1) enzyme, and resulting in the production of siRNAs by the VN. In a manner still to be clarified, these siRNAs accumulate in the neighbouring SCs, thus ensuring TE silencing in the germline . An equivalent mechanism is also present in animals, where piRNAs have functions homologous with those of plant siRNAs [56,57]. More specifically, the piRNA pathway has been shown to be required to repress TEs at two developmental stages: (i) during spermatogenesis, via post-transcriptional regulation; and (ii) during germline reprogramming in embryogenesis, by purportedly directing RNA-dependent DNA methylation in developing fetal germ cells [58,59]. Overall, small RNA mechanisms seem to have been evolutionarily conserved as a means of safeguarding the (epi)genetic information contained in male gametes against possible disruptions caused by mobile genomic elements.
Conclusions and future perspectives
It can be argued that sexual reproduction is, ultimately, defined by the capability of germline cells (or their precursors) to undergo meiosis. Meiosis, besides ensuring DNA recombination and reductive chromosome segregation, dictates profound changes in the germline gene expression programme. These changes are particularly evident when comparing disparate eukaryotic species and suggest that male gametogenesis implies the successful orchestration of two evolutionarily conserved gene expression programmes: one pre-meiotic, the other post-meiotic. We can speculate that the first programme commits cells to meiosis, whereas the second ensures the acquisition of the fertilizing ability. Such ability is manifested by two equally conserved hallmarks: extensive chromatin remodelling and exposure of cell-surface proteins required for subsequent fusion with the female gamete. Regulation via small RNAs adds an extra layer of complexity to both expression programmes, with these molecules playing key roles in the silencing of mobile genomic elements and, perhaps, on the actual regulation of the differentiation programme.
In the light of the rather limited success of candidate gene-driven and high-throughput platforms in identifying novel regulators of germline function, our understanding of basic mechanisms of male germ cell development may benefit from an evolutionary developmental biology approach. The integrated outlook provided by such an approach will increase the likelihood of identifying novel evolutionarily conserved molecular mechanisms of germline development.
Regulation of Fertilization and Early Seed Development: A Biochemical Society Focused Meeting held at the University of Bath, U.K., 11–13 September 2013. Organized and Edited by James Doughty (University of Bath, U.K.) and Thomas Dresselhaus (University of Regensburg, Germany).
GENERATIVE CELL-SPECIFIC PROTEIN 1
MALE GAMETE-SPECIFIC HISTONE H3
male germ unit
pollen mother cell
primary parietal cell
primary sporogenous cell
We thank Leonor Boavida for a critical reading of the paper before submission.
This work is supported by Fundação para a Ciência e a Tecnologia (FCT), Portugal [grant numbers PTDC/BIA-BCM/103787/2008 (to J.D.B.), PTDC/SAU-BID/111796/2009 (to R.G.M.) and PTDC/BIA-BCM/111822/2009 (to R.G.M.)]. P.A.P. is supported by a FCT Ph.D. fellowship [grant number SFRH/BD/63477/2009]. P.N.C. is supported by a FCT post-doctoral fellowship [grant number SFRH/BPD/84214/2012].