A comparison of spermatogenesis between flies and men—conserved processes of male gamete production

Abstract

BACKGROUND

Spermatogenesis is a dynamic process that involves the co-ordinated development of millions of cells, from stem cells to highly polarized sperm capable of motility and fertility. It is, therefore, not surprising that many thousand genes are required for male fertility. Mutant mouse models are routinely employed to test the function of these genes as well as to validate genetic variants that may be causing human male infertility. The use of mice and other animal models has led to significant knowledge gain regarding the genetic regulation of mammalian male fertility. However, due to the sheer number of genes and genetic variants to be tested these approaches are expensive and time-consuming. We and others have investigated the use of alternate model organisms to expedite validation approaches, including the utility of the fruit fly Drosophila melanogaster.

OBJECTIVE AND RATIONALE

This review explores the conserved mechanisms of sperm production between mammals and flies, with a focus on the human setting where possible.

SEARCH METHODS

Studies were identified via PubMed using searches including keywords related to the focus of this review, including human, mammalian, and fly or Drosophila spermatogenesis and male fertility. Follow-up searches including using search terms for specific structures and processes for comparison between species included, but were not limited to, male reproductive tract, spermatogenesis, spermatogonia and stem cell niche, meiosis, spermiogenesis and its sub-processes, and sperm/spermatozoa. No time frame or species restrictions were placed on searches.

OUTCOMES

We identify key phases of spermatogenesis that are highly conserved between humans and flies, including the early germ cell divisions and the ratio of haploid germ cells generated for each spermatogonial stem cell, allowing their use as a model organism to explore such processes. Some processes are moderately well conserved between mammals and flies, including meiosis with the notable absence of ‘crossing over’ in flies. We also identify some processes that are poorly conserved, such as a divergence in sperm tail accessory structures, for which flies are not likely a suitable model organism to decipher human biology or for mammals broadly. Examples of where the fly has been or could be useful to study mammalian gene function in male fertility have also been described.

WIDER IMPLICATIONS

Drosophila melanogaster is undoubtedly a useful model organism for studying a wide range of human diseases with genetic origins, including male infertility. Both humans and flies possess a pair of testes with the primary role of generating sperm. The formation of cysts in Drosophila testes allows germ cells to constantly proliferate and stay synchronized at the respective maturation phase, as is the case for humans. While both organisms use a method of sperm storage, mammalian sperm undergo post-testicular modifications and are stored in the epididymis. In Drosophila, sperm are stored in the seminal vesicle, and do not appear to undergo any overt post-testicular modifications in this epididymis-like structure. The seminal vesicle is a separate organ in mammals that is responsible for generation of the seminal fluid. It is important to note that male fertility and thus spermatogenesis are subject to significant evolutionary pressure, and there is a degree of variation in its processes between all species. As such, the absence of a phenotype in mutants would not determine that the gene is dispensable for fertility in humans. While flies are useful for genetic studies to confirm human disease causality, we propose they should be used primarily to pre-screen and select strong candidates for further interrogation in mammalian species for translational pathways in the context of human fertility.

REGISTRATION NUMBER

N/A.

Graphical abstract

Drosophila spermatogenesis shows significant similarities in spermatogenesis to humans, with identical germ cell ratios and function but some differences in meiosis and sperm structure.

Introduction

Male infertility is a complex disorder that affects at least 7% of men, and in the majority of cases, the exact aetiology is unknown (Kimmins et al., 2024). Estimates indicate that a significant portion of male infertility is caused by genetic factors (Krausz and Riera-Escamilla, 2018; Houston et al., 2021b; Xavier et al., 2021), alongside other potential causes including environmental and lifestyle factors, infections, and physical trauma (Kimmins et al., 2024). Equally, male infertility is likely often multifactorial, i.e. it may involve interactions between multiple genetic variants, or between genetic variants and environmental factors. Male infertility has a highly heterogeneous presentation, ranging from a complete absence of sperm in the ejaculate (azoospermia), through to reduced sperm output (oligozoospermia), abnormal sperm motility (asthenozoospermia), or structure (teratozoospermia), or a combination of multiple defects (e.g. asthenoteratozoospermia). It may also involve changes to the quality of seminal fluid (Schjenken et al., 2021).

Transcriptome analyses have revealed that more than 20 000 genes are expressed during spermatogenesis (Soumillon et al., 2013), with thousands of these genes enriched in the testis compared with other tissues (Djureinovic et al., 2014; Fagerberg et al., 2014; Uhlen et al., 2015). While numerous factors delay efforts to elucidate the complement of infertility-causing genetic factors (reviewed in Houston et al., 2021a; Xavier et al., 2021; Oud et al., 2022), the uptake of next generation sequencing in the field is leading to the discovery of hundreds of potential genetic causes of human male infertility (Houston et al., 2021b). Conversely, given the rapid and expanding adoption of medically assisted reproduction, there is a dire need for increased research into the aetiology of male infertility and the consequences of using suboptimal sperm on offspring health and fertility, to allow for appropriate genetic counselling.

While male infertility is common at a population scale, the incidence of each individual infertility-causing variant is relatively low and, as indicated above, is predicted to span thousands of gene targets (Kimmins et al., 2024). Efforts to define genetic variants causing infertility are often hamstrung by the inability of affected men to generate the large families required for traditional linkage studies. In addition, replication studies are difficult and often require very large cohorts (multiple thousand men) before several different variants are identified in the same gene (Nagirnaja et al., 2022). Notably, any dominant de novo infertility-causing mutations will not be passed on to future generations naturally, as they cause infertility (Oud et al., 2022). In fact, the advent of medically assisted reproduction, including in vitro fertilization and intracytoplasmic sperm injection, has enhanced the need for further research into the aetiology of male infertility. If the cause of infertility is genetic, the natural barriers preventing infertile men from having children are often bypassed when assisted reproduction is used, and thus these variants can be passed on to future generations. This highlights the need to define a comprehensive list of essential fertility genes and known pathogenic variants, for screening purposes and to aid in providing better genetic counselling.

Model organisms are an important tool to validate and test causality. Mice are commonly used as models for studying human spermatogenesis, as the processes required for sperm production are highly similar between mammals (Bonilla and Xu, 2008) and mice are the gold standard model organism for studying male infertility (reviewed in Borg et al., 2010; Jamsai and O’Bryan, 2011; Houston et al., 2021a). Hundreds of genes required for mammalian male fertility have been discovered or validated using mouse models, including many knockout models based on genes containing variants identified in infertile men (reviewed in Jiao et al., 2021). Mouse models do, however, present logistical challenges compared to lower order model organisms: male mice do not become fertile until ∼7 weeks of age, their maintenance can be expensive, and the workload required to comprehensively study even one gene knockout model is often high (Houston et al., 2021a). In this review, we discuss the applicability of Drosophila (flies) for genetic screens and as tools to understand human male infertility. Specifically, we focus on Drosophila melanogaster and note that variation exists between different Drosophila species (Scharer et al., 2008; Ramm et al., 2014).

Drosophila melanogaster (the fruit fly) presents as an alternative model organism wherein at least some processes are clearly similar to human spermatogenesis (White-Cooper and Bausek, 2010). However, the processes that best model the human situation and those that are not comparable are not well detailed. Looking broadly, Drosophila are undoubtedly useful in studying human genetics, as over 75% of human ‘disease genes’ have Drosophila orthologues (Pandey and Nichols, 2011). Recent data also highlight a conserved network of at least 104 genes in metazoans required for male fertility (Brattig-Correia et al., 2024). Flies are cheaper to maintain, have a short reproductive cycle (∼2 weeks), and have a wide variety of well-defined genetic tools available to facilitate gene analysis (reviewed in Siddall and Hime, 2017). For example, Yu et al. (2015) used a fly RNA interference (RNAi) screen to test the role of 22 fly orthologues of genes affected by genetic variants identified in a non-obstructive azoospermia GWAS. Through this approach, seven essential male fertility genes were discovered and validated (Yu et al., 2015). As another example, sophisticated molecular genetic tools allow the ‘humanizing’ of the Drosophila genome, whereby a fly gene can be replaced or rescued by its human orthologue (Ugur et al., 2016; Wangler et al., 2017; Bellen et al., 2019; Houston et al., 2025). Similarly, the human genetic variant can be recapitulated in the fly in vivo to directly test variant pathogenicity to gene function. Comparing in vivo function of the wild-type human gene/protein with that of a candidate infertility-causing genetic variant is the most direct approach to determine if the genetic variant is the underlying cause of the infertility in the original patient(s).

Functional conservation between human and Drosophila orthologues for ‘disease-causing genes’ has been well established, with several examples showing consistent phenotypes between species, particularly for brain-related diseases (Ecovoiu et al., 2022). For example, loss of Rbf expression in Drosophila induces developmental delay as well as altered brain and behavioural phenotypes that mirror findings in humans resulting from damaging variants in the orthologue RBL2 (Aughey et al., 2024). A separate study validated a human variant in VCP as disease causing for human frontotemporal lobar degeneration using a fly model. While co-expression of the fly orthologue TER94 could rescue loss of brain volume, mutation of the conserved amino acid in TER94 failed to rescue the phenotype (Tsumaki et al., 2024). Similar results have been observed for variants in spliceosome genes U2AF2 and PRPF19 when modelled in flies (Li et al., 2024b). A thorough review of the capacity of Drosophila to model genetic human disease (with a focus on brain disease) and additional evidence of human gene orthologue rescues for fly disease also exists (Ecovoiu et al., 2022). In the context of fertility, fly gene AK-3 is essential for axoneme formation and sperm release. Loss of AK-3 function can be rescued by co-expression of human orthologue CKMT1B (Chen et al., 2025). We have also shown that the human HSPA4 and HSPA4L genes can partially rescue spermatogenesis in a Hsc70Cb mutant fly model wherein all germ cells are lost (Houston et al., 2025). Thus, the utilization of flies to validate genetic causes of human disease, including male fertility, holds great promise.

While functional validation in higher organisms such as mice or gene replication in humans is ultimately required, fly-based studies provide the opportunity to rapidly pre-screen the effects of the hundreds of potential loss-of-function variants being identified though the genome/exome sequencing of infertile men. For example, two studies (Riera-Escamilla et al., 2022; Sieper et al., 2024) recently combined exome sequencing with gene knockdown approaches in flies to test the role of conserved RBBP7 and TEX genes in male fertility. While the benefits of flies as a model are clear, there are some distinct differences between fly and mammalian reproduction to be considered. This review aims to directly compare and contrast the cell biology of Drosophila and mammalian spermatogenesis at each developmental stage. Where limited human data are available, insights gathered from other mammalian species are discussed, largely from mice and rats, where spermatogenesis is highly similar. We highlight the benefits of using Drosophila melanogaster as a model species for human male infertility research and outline the potential caveats and limitations of using this invertebrate model. This review is largely focused on spermatogenesis in sexually mature/adult men and flies. Differences in hormonal regulation of spermatogenesis between species will not be discussed within this review but are reviewed in Sofikitis et al. (2008), O’Shaughnessy (2014), Smith and Walker (2014), and Meiselman et al. (2017).

Methods

Studies were identified via PubMed using searches including keywords related to the focus of this review, including human, mammalian, and fly or Drosophila spermatogenesis and male fertility. Follow-up searches including using search terms for specific structures and processes for comparison between species included, but were not limited to, male reproductive tract, spermatogenesis, spermatogonia and stem cell niche, meiosis, spermiogenesis and its sub-processes, and sperm/spermatozoa. No time frame or species restrictions were placed on searches.

Male reproductive tract structure

Humans

In humans (and all mammals), the process of spermatogenesis occurs in the seminiferous tubules contained within a pair of testes (Fig. 1). These tubules are composed of a stratified epithelium containing somatic Sertoli cells, which sit on the basement membrane and form the structural basis of the epithelium, and embedded germ cells (Fig. 2) (Adham et al., 2001; Zhao et al., 2004; Tseden et al., 2007; Sharma et al., 2018). Germ cells are continuously generated by spermatogonial stem cells (SSCs) that sit on the basement membrane in specialized niches created by the Sertoli cells (reviewed in Oatley and Brinster, 2012). As germ cells progress through development, they broadly move through the epithelium towards the tubule lumen. A series of junctions between adjacent Sertoli cells, collectively known as the blood–testis barrier, segregates the epithelium into a basal compartment (Koskimies et al., 1973; Cameron and Snydle, 1980; Cheng and Mruk, 2009). In addition, Sertoli cells provide nutritional and structural support to the developing germ cells throughout spermatogenesis (Rato et al., 2012).

Figure 1.

Male reproductive organs and tracts in humans and flies. Major male reproductive organs are shown for (A) humans and (B) flies. Organs of the same colour are analogous between species. The seminal vesicles and prostate are presumed to be collectively analogous to the accessory glands and ejaculatory bulb. Not to scale.

Figure 2.

A comparison of human and fly spermatogenesis. (A) In humans, multiple seminiferous tubules exist within the testis, each harbouring millions of germ cells along their length through the depth of their epithelium. A longitudinal section of a testis is shown, highlighting the seminiferous tubules that connect to the rete testis, as well as a cross-section of a single seminiferous tubule to highlight germ cell development. Sertoli cells and SSCs line the basement membrane of seminiferous tubules, which are surrounded by peritubular myoid cells. The Sertoli cell cytoplasm extends deep into the centre of tubules to provide the nutritional and structural support of male germ cells that remain in contact with this cytoplasm throughout their entire development from spermatogonia to spermatozoa. Spermatogonia divide and commit to spermatogenesis, then as spermatocytes traverse the blood–testis barrier, a collection of specialized junctions between neighbouring Sertoli cells to sequester immunologically foreign germ cells from the immune system. As germ cells divide, clones remain in contact with each other by intercellular germ cell bridges that allow the sharing of protein and RNA. Spermatocytes undergo meiosis to form round spermatids, which undergo a dramatic transformation as they are remodelled into highly specialized spermatozoa and then individualized for release via the processes of spermiation. PM, peritubular myoid cell; BM, basement membrane; SSC, spermatogonial stem cell; Sg, spermatogonia; BTB, blood–testis barrier; Sc, spermatocyte; SC, Sertoli cell; RS, round spermatid; ES, elongating spermatid; Sp, spermatozoa. (B) In flies, each testis comprised a single blunt-ended tubule in which continuous germ cell proliferation occurs in isolated cysts. Each cyst is originally populated by two cyst cells and one gonialblast (akin to spermatogonia), each arising from their own stem cell population located at the hub. The hub is a small mass of somatic cells that co-ordinates stem cell maintenance and division. As gonialblasts (and all male germ cells) mature, they remain encased with the same two cyst cells, which allows the germ cell cyst to migrate as a discrete unit along the length of the testis. At the 16-cell stage, spermatogonia mature into spermatocytes and undergo meiosis to produce round spermatids. Spermiogenesis permits the remodelling of round into elongating spermatids, which then undergo individualization and a process of coiling due to the immense length of fly sperm tails. CySC, cyst stem cell; GSC, germline stem cell; CC, cyst cell; Gb, gonialblast; HC, head cyst cell; TC, tail cyst cell; iSp, individualized spermatid. Panel (A) is adapted from Houston et al. (2021a).

One of the major differences between humans and flies is the level of post-testicular sperm modification and the location of sperm storage in the male (and female reproductive tracts). Post-testicular sperm maturation in humans (and other mammals) occurs primarily in the epididymis, a highly coiled tube attached to the testis via efferent ducts (Nixon et al., 2015; Barrachina et al., 2022). Mammalian sperm are transported sequentially through the four major regions of the epididymis: the initial section and caput (or head), the corpus (or body), and the cauda (or tail). The caput and corpus epididymis are primarily responsible for sperm maturation, while the cauda epididymis acts as a storage vessel for functionally immature sperm prior to ejaculation (reviewed in Cornwall, 2009). Epididymal sperm maturation involves changes in sperm protein and RNA content, and plasma membrane composition, conferring the potential for progressive sperm motility and capacitation, which are essential for sperm to reach, bind, and fertilize an oocyte (Hinrichsen and Blaquier, 1980; Nixon et al., 2015; Gervasi and Visconti, 2017; Barrachina et al., 2022).

During ejaculation, human sperm leave the epididymis via the vas deferens and eventually mix with seminal fluid secreted from accessory glands, including the seminal vesicles and the prostate, and fluids within the female reproductive tract. Seminal fluid provides metabolites, and immune and functional support to sperm as they manifest progressive motility and travel through the female reproductive tract (Drabovich et al., 2014; Schjenken and Robertson, 2020; Schjenken et al., 2021). Within the female reproductive tract, sperm undergo the process of capacitation, and, in parallel, ultimately manifest a state of hyperactivated motility. Collectively, these processes enable sperm to reach, recognize, and bind an oocyte and to undergo the acrosome reaction (Chang, 1951; Austin, 1952; van Duin et al., 1994; Aitken et al., 1995; Asquith et al., 2004).

Drosophila

The Drosophila testis is a small, blunt-ended, coiled tube (Figs 1 and 2). Like in humans, flies have two testes, but spermatogenesis begins at the apical tip of the testis (adjacent to a SSC niche—‘the hub’). As with humans, germ cell development progresses in ‘cysts’ of interconnected cells along the apical–basolateral axis (Fig. 2). All phases of spermatogenesis can be viewed concurrently in a single fly testis (reviewed in Siddall and Hime, 2017). Divergent from mammalian systems, both germ cells and cyst cells (analogous to Sertoli cells) each originate from a stem cell population that sits adjacent to the hub. Upon commitment to spermatogenesis, each of the stem cell types (Fig. 1) divides to form a cyst that contains two somatic cyst cells surrounding a gonialblast. In flies, the two cyst cells remain associated with the same population of germ cells throughout all germline divisions and move with the cyst as it migrates along the testis. In turn, they engulf and protect the developing spermatogonia (gonialblast), spermatocytes, and spermatids, while also providing signals to stimulate differentiation (Hardy et al., 1979; Cheng et al., 2011). In mammals, Sertoli cells remain in place and connected to the basement membrane, and germ cells migrate through the depth of the Sertoli cells as they develop, meaning each mammalian Sertoli cell is in contact with up to five to six germ cell sub-types at any one time (SSCs, spermatogonia, primary and secondary spermatocytes, round and elongating spermatids). By contrast, fly cyst cells are only in contact with a single germ cell type at a time.

Drosophila sperm are also immature upon exiting the testis. Flies do not, however, have an epididymis analogous to that seen in mammals. Instead, sperm enter, and are stored in, a structure, rather confusingly known as the seminal vesicle (Yuan et al., 2019) (Fig. 1). This structure is not equivalent to the seminal vesicles in mammals and is rather a simplified epididymis. Once released from the seminal vesicles, sperm transit through the ejaculatory duct (Fig. 1), where they encounter seminal fluid proteins that are released from the accessory glands. The combined sperm and seminal fluid mix is then pushed through the ejaculatory bulb into the female reproductive tract (Avila et al., 2011, 2015). Drosophila accessory glands (Fig. 1) bear functional homology to the mammalian seminal vesicles and prostate, and potentially the bulbourethral gland, which secrete seminal fluid proteins (Wilson et al., 2017; Cohen and Wolfner, 2018). While the full composition of Drosophila seminal plasma is yet to be determined (Majewska et al., 2014), at least 100 proteins are produced by the accessory glands, ejaculatory duct, and ejaculatory bulb, which assist in sperm function, proteolysis, and cell signalling, among other roles that influence fertilization (Takemori and Yamamoto, 2009; Ruhmann et al., 2016). A comparison of seminal fluid proteins between humans and flies has been undertaken (Mueller et al., 2004; Gilany et al., 2015; Wigby et al., 2020). This revealed similarity in protein classes and function between species, with 30% of Drosophila seminal fluid proteins exhibiting moderate or high homology to those found in human seminal plasma (Wigby et al., 2020). One of these proteins, Sex Peptide, can elicit physiological and behavioural changes in mated females to facilitate egg-laying and promote inheritance of the male’s genes by inhibiting remating of females (Wolfner, 1997).

In summary, the basic male reproductive tract anatomy is highly comparable from flies to men (Table 1), with some distinct differences arising in specific organs and the level of sperm post-testicular maturation. These differences will be explored below.

Table 1.

Equivalent sex organs and reproductive cell types between humans and Drosophila.

Sex organ (mammal)	Sex organ (fly)
Testis	Testis
Epididymis	Seminal vesicle
Prostate	Accessory gland+ejaculatory bulb
Seminal vesicle
Cell type (mammal)	Cell type (fly)
Spermatogonial stem cell	Germline stem cell
Spermatogonium	Gonialblast
Spermatocyte	Spermatocyte
Round spermatid	Round spermatid
Elongating spermatid	Elongating spermatid
Spermatozoon	Spermatozoon
Sertoli cell	Cyst cell

Somatic cells

Mammals

As defined above, Sertoli, or nurse, cells play numerous roles in spermatogenesis and via their supportive functions define the capacity for sperm production (Russell and Peterson, 1984; Orth et al., 1988; Meroni et al., 2019). The Sertoli cell cytoplasm has a filigree-like arrangement, which provides a large surface area to remain in contact with a number of germ cell subtypes, as well as other Sertoli cells (Russell et al., 1983; Weber et al., 1983). The junctions between adjacent Sertoli cells form the blood–testis barrier (reviewed in Cheng and Mruk, 2012), which provides a physical barrier to regulate the entry of nutrients and endocrine molecules into, and exclude harmful toxins from, the adluminal region of the seminiferous epithelium. Specifically, the blood–testis barrier provides a barricade to sequester the ‘foreign’ and highly immunogenic spermatocytes and spermatids from immune-mediated attack (Fijak et al., 2011). In humans and mice, Sertoli cells undergo a major proliferative cycle during puberty to define their maximal number, establishing the spermatogenic potential of the testis (Griswold, 1998; Meroni et al., 2019). A tight cross-talk exists between the Sertoli cells and germ cells (Steger et al., 1996). This is further evidenced by an abnormal human Sertoli cell expression profile in conditions such as Klinefelter syndrome and non-obstructive azoospermia (Zhao et al., 2020).

Sertoli cells also secrete substances that are crucial for spermatogenesis such as energy sources for the developing germ cells. One example of this is the production of lactate for pachytene spermatocytes and round spermatids, which require significant energy production to power meiosis and spermiogenesis (Jutte et al., 1981). Other critical energy sources such as glucose and pyruvate are also secreted by Sertoli cells (reviewed in Alves et al., 2013).

While not a focus of this review, human (and mammalian) testes also contain Leydig cells in their inter-tubular compartment, which are crucial for the hormonal regulation of spermatogenesis via roles in androgen production—primarily testosterone (reviewed in Smith and Walker, 2014).

Finally, mammalian spermatogenesis is aided by the activity of both peritubular cells that surround the basement membrane on the basal (inter-tubular) side of the basement membrane and immune cells within the interstitium. Peritubular cells control the peristaltic movements that aid in transiting sperm that have undergone release into the rete testis and onward to the epididymis in humans (reviewed in Albrecht, 2009). Peritubular cells also secrete paracrine factors that modulate spermatogenesis. A host of immune cell types exist within the testis, including macrophages, dendritic, mast, and T cells (reviewed in Bhushan et al., 2020). The main function of these immune cells is to regulate the immune response and promote immune privilege to prevent inflammation in the testis.

Drosophila

As with germline stem cells (GSCs) (discussed below), Drosophila somatic cyst stem cells (CySCs) undergo continuous divisions to generate somatic cyst cells. This represents a major difference to mammalian spermatogenesis, where the Sertoli cell number is determined during or prior to puberty (Orth et al., 1988; Sharpe et al., 2003). CySCs cells are flat and irregularly shaped, and encapsulate the GSCs to ensure constant cross-talk (Cheng et al., 2011). Like GSCs, when CySCs divide, one of the daughter cells continues as a CySC while the other differentiates into a cyst cell (Hardy et al., 1979). A complete loss of GSCs causes the normally quiescent hub cells to re-enter the cell cycle, migrate away from the niche, and convert into CySCs (Hetie et al., 2014), reinforcing that reciprocal signalling and cross-talk between the GSCs and the CySCs assist in the regulation of the stem cell proliferation and commitment to spermatogenesis. A loss of hub cells cannot be compensated for, either from other hub cells or from cells outside of the hub, indicating the static nature of the niche after it has been established (Hétié et al., 2023).

After differentiation of the germline from GSCs into gonialblasts, two cyst cells (each generated by a separate CySC) surround each gonialblast, which seeds the formation of a cyst. Cyst cells isolate strings of developing sister germline cells from each other in individual cysts. At the same time as the expansion of paired cyst cells, incomplete cytokinesis results in cytoplasmic bridges between germ cells (also called ring canals in flies), within the same cyst (Hime et al., 1996). Throughout spermatogenesis, cyst cells expand in size to accommodate the increase in size and number of the germ cells they encapsulate (Zoller and Schulz, 2012) (Fig. 1). For example, when the sperm head and tail are formed, the cyst cells lengthen, one growing much more extensively to become the ‘tail cyst cell’, while the rostral cyst cell is termed the ‘head cyst cell’ (Zoller and Schulz, 2012). By the late elongating spermatid period, the spermatid nuclei are oriented at the head cyst cell end (Fabian and Brill, 2012).

In contrast to the mammalian setting, there are no Leydig analogous cells in the fly testes and the hormonal regulation of spermatogenesis is regulated by a network of ecdysone, ecdysis-triggering hormone, and juvenile hormone (Meiselman et al., 2017). Similar to the peritubular cells in mammals, in flies, a muscular sheath encases each testis. This sheath plays roles in both maintaining the coiled structure of testis and, like peritubular cells in mammals, is involved in cell signalling to influence spermatogenesis (Rothenbusch-Fender et al., 2017; Chang et al., 2019). The contribution of this sheath to germ cell migration within the testis is currently unclear.

In summary, there are differences in the structure and arrangement of the Sertoli cells in human spermatogenesis and the cyst cells in fly spermatogenesis. In humans, the number of Sertoli cells is fixed and does not change significantly throughout life. Drosophila cyst cells are, however, constantly generated from CySCs and develop alongside the germ cells they enclose. Common between species is that as germ cells mature, they remain in contact with the same Sertoli/cyst cells. Sertoli cell function is also broadly conserved between species.

Male germ cells—spermatogonia and cell divisions

Humans

In humans and all mammals, spermatogenesis originates from a population of cells called SSCs, which reside in a stem cell niche. The SSCs renew themselves and generate daughter cells that commit to spermatogenesis and develop into transit amplifying type A spermatogonia before maturing into type B spermatogonia (Fig. 3). Somewhat confusingly, the nomenclature for different types of spermatogonia differs between mammalian species, including between mice and humans (Guo et al., 2014).

Figure 3.

Male germ cell divisions in humans and flies. Cell types are shown above the respective clusters, while total numbers of sister germ cells at each stage are shown underneath. In both humans and flies, a single spermatogonial/germline stem cell generates 64 spermatids from each division. Red insets indicate the cell divisions of each of the eight committing spermatogonia prior to meiosis—i.e. once for each of the eight cells. (A) Human spermatogenesis begins with spermatogonial stem cells (A_dark and/or A_pale), which undergo asymmetric cell division to replenish the stem cell population (self-renewal) as well as to generate spermatogonia that undergo further mitotic divisions to commit to spermatogenesis. Type B spermatogonia mature to primary spermatocytes that undergo meiosis to form spermatids. Spg, spermatogonia; 1° spc, primary spermatocyte; 2°, secondary spermatocyte; sptd, spermatid. (B) In flies, spermatogenesis similarly begins with germline stem cells that undergo asymmetric division to self-renew and generate a gonialblast that commits to spermatogenesis. Subsequent cell divisions are highly similar to the human setting and result in identical germ cell numbers at each stage. * at the 16-cell stage, primary spermatocytes are formed after the cyst of spermatogonia undergoes S-phase prior to meiosis. GSC, germline stem cell; GB, gonialblast. Panel (A) is inspired by Fayomi and Orwig (2018).

The model for categorizing human spermatogonia was first defined by Clermont (1966), who separated the spermatogonia based on their chromatin structure and staining intensity when stained with haematoxylin. A_dark spermatogonia are small, spherical cells that reside on the basement membrane of the seminiferous tubules (Clermont, 1966). They have a uniformly stained nucleus and dense chromatin that is most easily viewed by a strong haematoxylin stain and a bright ‘halo’ surrounding the nucleus. A_pale spermatogonia stain lightly with haematoxylin (hence the name, pale). Type B spermatogonia, also have pale haematoxylin-stained chromatin but can be distinguished from A_pale spermatogonia by their presence of heterochromatin (De Rooij and Russell, 2000).

While both A_dark and A_pale spermatogonia can be defined as stem cells, they perform very different functions in maintaining the human germline. A_dark spermatogonia serve as the more traditional stem cells that very rarely divide and A_pale spermatogonia are the stem cell population that divide regularly to both self-replicate and maintain spermatogenesis by producing type B spermatogonia that commit to sperm production (Waheeb and Hofmann, 2011; Fayomi and Orwig, 2018). Meanwhile, A_dark and A_pale spermatogonia may be the same cell type at alternate stages of mitosis (Hermann et al., 2010; Fayomi and Orwig, 2018). In humans, one SSC (A_dark/A_pale) gives rise to eight type B spermatogonia connected as a chain of germ cells, of which each of the eight type B spermatogonia undergo meiosis to generate a total of 64 haploid spermatids (Fig. 3). Comparatively, in mice, additional transit amplifying divisions in undifferentiated and then differentiating spermatogonia allows a single stem cell to give rise to 4096 spermatids (Fayomi and Orwig, 2018).

Drosophila

In Drosophila, the male germline is an excellent model for examining stem cell self-renewal and differentiation biology (Davies and Fuller, 2008). Like in mammals, Drosophila GSCs reside within a ‘stem cell niche’, a microenvironment that assists the stem cells to either commit to spermatogenesis or undergo self-renewal (Spradling et al., 2011). These processes are tightly controlled by the niche, as excessive differentiation or commitment leads to a depletion of the GSCs, while imbalanced self-renewal can lead to an overpopulation of replicating cells, and eventually tumorigenesis, as can also occur in mammals (Yamashita et al., 2005; Monk et al., 2010; Dominado et al., 2016; Yang et al., 2021). The stem cell niche resides at the apical tip of the Drosophila testis (de Cuevas and Matunis, 2011), with 8–12 GSCs surrounding the 16–24 somatic cells of the hub (Siddall and Hime, 2017).

At the 16-cell stage, spermatogonia differentiate into spermatocytes by undergoing pre-meiotic DNA synthesis (S-phase) (Giansanti and Fuller, 2012), followed by meiosis to form 64 spermatids contained within a single cyst (Fig. 3). In this regard, flies undergo the same number of divisions and develop the same ratio of haploid germ cells formed per stem cell division. When the GSCs divide, the mitotic spindle is oriented perpendicular to the hub such that one daughter cell remains attached to the hub in order to remain a GSC and the other is moved away from the niche and forms a gonialblast, committed to spermatogenesis (de Cuevas and Matunis, 2011). As described above, the gonialblast itself is surrounded by two somatic cyst cells. After four rounds of transit-amplifying divisions, a group of 16 interconnected spermatogonia are generated from a single gonialblast and remain encased within the original two cyst cells (Fig. 3). This framework continues through mitosis into spermatocytes that undergo meiosis, where cytoplasmic bridges are co-ordinated by a germ cell-specific organelle called the fusome (Lin et al., 1994; Hime et al., 1996). This organelle is rich in membrane skeletal proteins spectrin and adducin, and, during early spermiogenesis, actin (Lin et al., 1994; Hime et al., 1996). The fusome enlarges during male meiosis to aid in the coordination of meiotic divisions and persists into spermatids, where it eventually disassembles during late spermiogenesis. While the fusome is common among the male and female germlines in flies, and a similar structure has been noted in Xenopus female germline cysts (Kloc et al., 2004), it is unclear if there is an analogous structure in the mammalian germline.

In summary, the pre-meiotic phase of spermatogenesis is highly similar between flies and humans, where for example flies have long been used to understand the role of the testicular stem cell niche and of transit amplification (Monk et al., 2010; Quinn et al., 2013; Voog et al., 2014). The underlying processes and number of mitotic divisions are comparable between humans and flies, with a single SSC/GSC in both species contributing a total of 64 spermatids.

Meiosis—spermatocytes

Precise chromosome segregation during meiosis is critical to ensure that each germ cell receives a single copy of each chromatid, where errors in chromosome segregation during meiosis lead to aneuploidy (Lane and Kauppi, 2019). Mammalian meiosis achieves an increase in genetic diversity through the processes of crossing over and ultimately results in the production of four haploid spermatids from a single diploid spermatocyte. This section of the review will focus on key mechanisms of chromosome dynamics during meiosis, including cohesion and synapsis of homologous chromosomes, recombination, and chromosome segregation. Male meiosis in both mammals and flies begins with a single round of DNA replication in primary spermatocytes. Once this replication is complete, several differences exist between the spermatocytes of flies and mammals.

Mammals

Meiosis I begins with prophase I, where homologous chromosomes pair and initiate the process of synapsis. Homologous chromosomes are composed of two pairs of sister chromatids that are generated during the meiotic S phase. While each pair of chromatids is bound together by cohesion forces generated by cohesin complexes, the interactions between homologous chromosomes are stabilized via development of a zipper-like structure called the synaptonemal complex in human (and mammalian) spermatocytes (Lee et al., 2003; Jan et al., 2018). The protein composition of cohesin complexes is highly conserved from yeast to humans but differs minorly between mitosis and meiosis by the incorporation of germ cell-specific proteins such as REC8 and STAG3 (Lee et al., 2003; Brooker and Berkowitz, 2014; Fukuda et al., 2014; van der Bijl et al., 2019; Krausz et al., 2020). The synaptonemal complex is composed of two lateral elements, each containing SYCP2 and SYCP3 proteins attached to one of the homologous chromosomes; a central element composed of SIX60S1, SYCE1-3, and TEX12; and several transverse filaments of SYCP1 that run perpendicularly and connect the central and lateral elements together (Gao and Colaiacovo, 2018; Fan et al., 2021).

Synapsis in mammals allows for the physical exchange of double-stranded DNA between two non-sister chromatids via the process of recombination and in doing so contributes to increasing the genetic diversity of gametes (Zickler and Kleckner, 2015). During this process, the abundant double-strand breaks that are programmed early in meiosis are repaired, following formation of the synaptonemal complex. Repair of around 1 in 10 of these breaks results in crossing over—exchange of DNA between non-sister chromatids, which is controlled by members of the MLH protein family (reviewed in Pannafino and Alani, 2021). At least one crossover event occurs per pair of homologous chromosomes, which aids in ensuring homologous chromosomes are correctly paired (Barchi et al., 2008). Following crossover events, a structure called the chiasma forms between homologous chromosomes, which is fortified by cohesion forces and ensures that the homologs remain attached once assembled along the metaphase plate (Buonomo et al., 2000). The generation of chiasmata during meiosis I is essential for the correct segregation of homologous chromosomes into separate daughter cells, as chromosome pairs lacking chiasmata often fail to separate, resulting in aneuploid gametes (Carpenter, 1994; Hirose et al., 2011).

During the meiosis I prophase to metaphase transition, microtubule spindle bundles originating from centrioles and microtubule organizing centres at opposite sides of the cell attach to the kinetochores of each homologous chromosome (reviewed in McKee et al., 2012). Through the tension created by microtubules, this event leads to the lining up of homologs along the metaphase plate. To achieve segregation, the highly conserved enzyme separase is activated, which targets and cleaves REC8 in the cohesin complexes surrounding the homologous chromosomes, to begin resolving chiasmata (Buonomo et al., 2000; Kudo et al., 2009). During meiosis I, however, the combined efforts of a number of conserved proteins, including shugoshin and meikin, maintain sister chromatid cohesion by protecting REC8 at the centromeres from cleavage by separase (McGuinness et al., 2005; Kitajima et al., 2006; Clift and Marston, 2011; Kim et al., 2015). Next, the synaptonemal complex, which holds homologous chromosomes together, disassembles through the breakdown of the individual sub-complexes (reviewed in Cahoon and Hawley, 2016), and after disassembly, the homologous chromosomes are weakly held together by chiasmata. Kinetochores between sister chromatids are fused to ensure they do not separate during segregation (Kim et al., 2015). Thus, when microtubules attached to spindles pull the homologous chromosomes from the two opposing poles, they separate into two pairs of sister chromatids. The cell proceeds to divide, although cytokinesis is incomplete, to form two diploid spermatocytes, each containing one set of sister chromatids and connected by intercellular bridges. These bridges, which are also present between spermatogonia (Haglund et al., 2011) persist throughout germ cell development, and allow the sharing of protein and RNA across cells, maintaining the synchronized development of connected germ cells (Ventela et al., 2003; Haglund et al., 2011). Cytoplasmic bridges may play additional roles in structural support during germ cell development.

At the beginning of meiosis II, sister chromatids are still held together by the centromere, as well as by cohesion. Much like the segregation of homologous chromosomes during meiosis I, microtubule spindle bundles emanating from the spindle poles attach to the kinetochores of the sister chromatids. However, at this point, the kinetochores are not fused between sister chromatids, allowing them to be separated. The pulling strength of these spindles is so great that shugoshin is inactivated, REC8 is cleaved by separase, and the sister chromatids separate to opposite poles of the dividing cell (Clift and Marston, 2011; McKee et al., 2012). This results in the final product of meiosis: four haploid spermatids connected by intercellular bridges, which then enter spermiogenesis (haploid germ cell development).

Drosophila

Much like in mammalian spermatocytes, meiosis I in Drosophila spermatocytes commences at the 16-cell stage of spermatogenesis with a round of DNA replication. At this point, the germ cells are still located near the blunt-ended apical tip of the testis, within a somatic cell cyst. Spermatocytes then undergo a phase of active growth and increase in volume some 25-fold in size, making them the largest cellular fraction of the testis. A guide for staging of spermatocytes during male meiosis can be found in Cenci et al. (1994). A significant difference from most eukaryotic systems, including humans and mice, is that meiosis in male flies (as distinct from female meiosis) proceeds without formation of synaptonemal complexes, meaning that homologous chromosomes in Drosophila males neither form chiasmata nor undergo recombination (Cooper, 1949; Orr-Weaver, 1995; Kabakci et al., 2022b). Instead, homologous chromosomes in Drosophila spermatocytes are paired via an alternate conjunction complex, of which proteins modifier of stromalin in meiosis (Stromalin-2/SNM), univalent only (UNO), and mdg4 (MNM) (collectively, SUM proteins) have been identified (Weber et al., 2020; Kabakci et al., 2022a). It has also been recognized that a fourth protein, teflon, is required to recruit the three SUM proteins for conjunction complex formation but does not actively participate in the function of the complex (Tomkiel et al., 2001; Kabakci et al., 2022b). Collectively, the conjunction complex adheres homologs together like glue rather than surrounding and holding them in place like cohesion complexes (Kabakci et al., 2022b). As defined by loss-of-function experiments, absence of the conjunction complex members leads to aberrant segregation of chromosomes (Thomas et al., 2005; Sun et al., 2019; Kabakci et al., 2022b). Meanwhile, in common with meiosis in other eukaryotes, Drosophila spermatocytes also generate cohesin complexes to bind sister chromatids together, but flies do not contain REC8. These cohesin complexes, however, are regulated by two proteins specific to flies, ORD and SOLO (weak homology to DDX3X/DDX3Y) (Yan et al., 2010). While both genes are required for normal cohesion forces between, and thus segregation of, sister chromatids, SOLO likely directly interacts with the cohesin complex (Yan et al., 2010), whereas ORD has not been shown to colocalize with this same complex (Balicky et al., 2002).

Similar to the situation in most eukaryotes, separase is conserved in flies. Here, it mediates the cleavage of UNO in the conjunction complex, which is required for segregation of homologous chromosomes during metaphase I (Weber et al., 2020). An additional key player of Drosophila meiosis is mei-S332, a member of the shugoshin family that protects sister chromatid cohesin complexes from being dismantled during meiosis I (Kerrebrock et al., 1995).

In summary, while the basic processes of chromosome segregation to generate haploid cells are highly conserved across eukaryotes, Drosophila male meiosis diverges in two major ways. First, fly spermatocytes do not generate synaptonemal complexes, meaning no crossing over or recombination occurs. Second, while the mechanisms of sister chromatid pairing are highly similar, the cohesion forces required to hold sister chromatids together are achieved by alternate cohesin proteins. Processes of kinetochore attachment appear to be similar in flies, whereby a single fused kinetochore is apparent during meiosis I followed by each sister chromatid arranging its own kinetochore in the second meiotic division (Chaurasia and Lehner, 2018). The remaining meiotic processes appear to be similar, and in both human and fly settings achieve a set of 64 interconnected haploid cells that proceeds through the process of spermiogenesis.

Spermiogenesis—haploid germ cell development

Spermiogenesis begins immediately after the second meiotic division and is the process by which round spermatids are transformed into highly polarized sperm. In developing from round to elongated spermatids, germ cells in both humans and flies achieve several major feats, including the formation of the acrosome, the condensation and shaping of the nucleus, and the development of the sperm tail (flagellum). Despite the complex formation of each of these structures, they are constructed in parallel via complex transport pathways to allow the efficient remodelling of spermatids into polarized sperm (Pleuger et al., 2020). For additional resources, the processes of spermiogenesis in flies are well reviewed in Fabian and Brill (2012).

Acrosome and acroplaxome formation

Mammals

The acrosome is a unique organelle composed of a double membrane that covers the anterior region of the sperm nucleus. In its final state, the acrosome functions to house enzymes, e.g. acrosin, that are required to digest the zona pellucida of the egg in the female reproductive tract, and is thus essential for fertility in humans and other mammals (De Jonge et al., 1989; Siegel et al., 1990; Ferrer et al., 2012). The acrosome may also contribute to the streamlining of the sperm head shape to optimize the hydrodynamics of sperm motility and varies widely in shape between species (Austin and Bishop, 1958). Acrosome formation involves the production of a large number of proteins, some of which are produced as early as the pachytene spermatocytes in mice and humans (Escalier et al., 1991; Berruti et al., 2010; Berruti and Paiardi, 2015). Acrosome components originate from the Golgi and at least the endocytic pathway from the plasma membrane (West and Willison, 1996; Paiardi et al., 2011; Berruti and Paiardi, 2015). In mammals, acrosome biogenesis is classically split into four phases: the Golgi, cap, acrosome, and maturation phase (Clermont and Leblond, 1955; Khawar et al., 2019). Consistent with the role of the acrosome in defining head shaping, defects in acrosome formation often manifest in phenotypes such as globozoospermia in humans (large and/or round sperm heads that lack an acrosome) and can contribute to poor sperm DNA condensation and thus poor DNA integrity (Yassine et al., 2015; Talebi et al., 2018; Oud et al., 2020; Moreno, 2023).

During the initial Golgi phase of acrosome development, pro-acrosomal vesicles produced from both the Golgi apparatus and the endocytic pathways attach to, and indent into, the apical pole of the spermatid nucleus, which is anchored via the acroplaxome (Berruti et al., 2010; Xin et al., 2020). Studies in humans have identified ACTL7A as a key regulator of the fusion processes required to attach the pro-acrosomal vesicles to the nucleus (Xin et al., 2020; Zhou et al., 2023). The continued recruitment of acrosomal vesicles allows formation of the acrosomal granule. During the cap phase, the acrosomal granule begins to enlarge and flatten over the nucleus, forming the acrosomal cap (reviewed in Pleuger et al., 2020; Xiong et al., 2021). Acrosome contents are surrounded by a double membrane, the membrane surrounding the acrosome proper, the plasma membrane proximally and the nuclear membrane distally via the acroplaxome (Talbot and Chacon, 1982; Buffone et al., 2008).

Another hallmark of the cap phase is the manifestation of the marginal ring of the acroplaxome, an F-actin and cytokeratin-enriched plate-like structure that is formed at the site between the inner acrosomal membrane and the nucleus (Kierszenbaum et al., 2003). The acroplaxome has two main functions—to stabilize the acrosome during development and to anchor it to the nucleus—but it likely also plays a role in providing structural support for the shaping of the sperm head (Kierszenbaum et al., 2003, 2011; Shen et al., 2014). During the acrosome phase, the acrosome continues to spread along the dorsal nuclear surface in parallel with the remodelling of the spermatid head and distal migration of the manchette (Xiong et al., 2021).

The final phase of acrosome development, maturation phase, is characterized by the complete spread of the acrosomal granule along the acrosomal membrane (Wang et al., 2021). At the completion of acrosome development, the acrosome covers a significant portion of the apical sperm head (humans), with the remaining section referred to as the post-acrosomal sheath.

Drosophila

Acrosome development also occurs in flies and appears to be comparable to humans. The details of Drosophila acrosome development are, however, poorly studied. In the fly, formation of the acrosome begins at what would be the Golgi phase. First, a unique structure composed of flattened cisternae derived from the Golgi, the acroblast, assembles at what will become the apical side of the nucleus in late spermatocytes (Anderson, 1967). Interestingly, the acroblast is unique from the Golgi stacks present in other fly cells as it is composed of Golgi ribbons, long interconnected Golgi stacks, similar to the configuration in most mammalian cells, including during acrosome formation in germ cells (Susi et al., 1971; Kondylis and Rabouille, 2009; Yasuno et al., 2013).

To date, studies have focused on the role of Golgi maturation in the development of the fly acrosome and no evidence has been found to suggest the participation of additional pathways in acrosome biogenesis, in contrast to the situation in mammals (Yasuno et al., 2013). Consistent with the mammalian setting, in Drosophila the trans-side of the Golgi binds to the acroblast, much like the binding of the trans-side of the acrosomal vesicles to the acroplaxome in mammals (Moreno et al., 2000; Fári et al., 2016). In round spermatids, the fly acrosome is horseshoe-shaped, and, similar to human round spermatids, covers the tip of the elongating nucleus (Fári et al., 2016). As the spermatid develops, the acrosomal granule is generated from the acroblast after adhering to the nucleus (Fuller, 1993; Yasuno et al., 2013). Following acrosome development, the excess acroblast is broken down, and scattered at the posterior side of the nucleus, and discarded as a part of the ‘waste bag’ (Yasuno et al., 2013), equivalent to the mammalian residual body.

One difference between fly and human sperm function is what appears to be the lack of requirement of an acrosome reaction for Drosophila sperm to achieve fertilization. Drosophila sperm enter the oocyte through a small opening in its outer casing, called the micropyle (Nonidez, 1920). Sperm enter the oocyte with their membrane (including the acrosome) intact, and it is only when inside the oocyte, that the acrosome and cell membrane breaks down (Perotti, 1975). While it is unclear if the acrosome content is different between species as no comparison has been undertaken, acrosome proteins appear to be rapidly evolving (Dorus et al., 2010). Acrosome development is essential for fly fertilization, as mutation of acrosome-specific proteins results in a failure of fertilization due to defects in membrane fusion between gametes (Wilson et al., 2006).

In summary, there are some distinct similarities between the fly and the mammalian acrosomes. Differences, however, include that the acrosome appears to be exclusively formed from Golgi-derived material in flies, while the endocytic pathways contribute to acrosome formation in mammalian sperm. In both organisms, the acrosome plays a comparable role in defining head shape. Importantly, the mode of fertilization in flies is different to that that occurs in mammals, including that in Drosophila the acrosome remains intact when sperm enter the oocyte, meaning Drosophila is likely not suitable to study the acrosome reaction. However, major similarities in acrosome biogenesis suggest the fly is a suitable model for studying Golgi maturation and vesicle transport in the process of acrosome formation.

Nuclear condensation and sperm head shaping

Mammals

A key process in spermiogenesis is the significant re-packaging, condensation, and subsequent shaping of the spermatid nucleus. These events begin from step 8 and step 9 spermatids in men and mice, respectively (Muciaccia et al., 2013; Pleuger et al., 2020). In spermatogonia and spermatocytes, the genome is packaged by histones, similar to the packaging of DNA in somatic cells (Balhorn, 2018). In elongating spermatids, these histones begin to be replaced by transition proteins (TNPs)—TNP1 and TNP2—(Zhao et al., 2004), which are, in turn, largely replaced by protamines. In addition to core histones H2A, H2B, H3, and H4, and the linker H1 class histones that sit between nucleosomes, mammalian male germ cells also express testis-specific linker histone proteins H1t and HILS (Yan et al., 2003; Mishra et al., 2018). Approximately 85% of human sperm DNA (and an estimate of up to 99% in mice) is packaged around protamines (Tanphaichitr et al., 1978; van der Heijden et al., 2005). These remodelling events allow a 20-fold reduction in DNA volume (Balhorn, 2007) and this nuclear condensation is crucial in ensuring that the sperm has a small, hydrodynamic head. Inappropriately condensed sperm chromatin can lead to DNA damage in human sperm (Yassine et al., 2015) and, as defined above, abnormal acrosome formation. Aberrant expression of TNPs and protamines leads to subfertility or infertility in humans and mice (Yu et al., 2000; Adham et al., 2001; Zhao et al., 2004; Ravel et al., 2007; Tseden et al., 2007; Sharma and Agarwal, 2018; Amjad et al., 2021; Arevalo et al., 2022; Merges et al., 2022), underscoring the importance of DNA compaction for sperm function.

Protamine 1 (P1) is the most evolutionarily conserved of the protamines and is present in all vertebrates studied but not flies (Queralt and Oliva, 1993). The protamine P2 family is conserved in only a few mammalian species. It is present in humans and mice but not flies and is derived from a P2 precursor protein that undergoes proteolysis to form the primary (most abundant) P2 protein, as well as the less abundant P3 and P4 components (Yoshii et al., 2005; Oliva, 2006). The binding of protamines to DNA neutralizes charged chromatin, which allows DNA molecules to condense into toroid structures (Hud et al., 1995; Balhorn et al., 2000). Protamines thus achieve the condensation of nuclear DNA to ensure optimal sperm head shape, while also protecting the DNA from nucleases as the sperm passes through the male and female reproductive tracts (Sotolongo et al., 2005; Luke et al., 2016).

In addition to DNA condensation, the sperm nucleus must elongate to achieve the optimal species head shape. A network of cytoskeletal structures in elongating spermatids and an F-actin meshwork from surrounding Sertoli cells work in conjunction to achieve this transformation (Sakamoto et al., 2018; Dunleavy et al., 2019). In the spermatid, a microtubule-rich skirt-like structure called the manchette assembles around the sperm nucleus to achieve the sculpting of the base of the sperm head (Russell et al., 1991). The manchette is formed through the assembly of up to a thousand microtubules, which accumulate around the nucleus at a peri-nuclear ring and project towards the developing sperm tail (Rattner and Brinkley, 1972). The origin of manchette microtubules has not been deciphered. As spermiogenesis progresses, the manchette migrates towards the base of the sperm head, while concurrently constricting, and, in some species, pivoting to sculpt the lower portion of the nucleus into its species-specific head shape (Cole et al., 1988; Lehti and Sironen, 2016; Pleuger et al., 2020). In combination, it is thought that the constriction of the F-actin network within the overlying Sertoli cells co-ordinates the elongation of the apical spermatid nucleus, and also ensures the forces generated by the constriction of the manchette are appropriately constrained. The expansion of the acroplaxome and acrosome has also been proposed to contribute to apical sperm head shaping (Kierszenbaum et al., 2003; Sakamoto et al., 2018). Defects in manchette formation or function result in hyper-constricted, long, and knobby spermatid heads, and can also impede the distal expansion of the acrosome in instances when the manchette fails to migrate caudally (Cole et al., 1988; Liu et al., 2015; Dunleavy et al., 2017, 2021).

An additional role of the manchette is as a protein highway in the spermatid for protein trafficking throughout the spermatid (intramanchette transport) (Pleuger et al., 2020). This function is well supported by the common finding of sperm tail defects in mutant mouse models with impaired manchette formation and head shaping (Kierszenbaum, 2002; Wu et al., 2021; Yu et al., 2021; Zhang et al., 2022).

Drosophila

Similar to the processes in mammals, there are two distinct mechanisms that are required for nuclear remodelling in Drosophila: (i) the remodelling of the chromatin to facilitate a highly condensed state, around a 200-fold reduction in DNA volume (Fuller, 1993); and (ii) the sculpting of the nucleus from a spherical to needle shape (Tokuyasu, 1974a). During spermiogenesis, the spermatid nucleus undergoes significant reorganization, from a round to a needle shape. The major hallmarks of this transformation have been classified into five stages: round (early round spermatids), leaf, early canoe, late canoe, then finally needle shape in sperm (Tokuyasu, 1974a; Fabian and Brill, 2012). The reorganization of chromatin within Drosophila spermatids is remarkably similar to the processes in mammalian spermatids (reviewed in Rathke et al., 2014). Like in mammals, this process begins after the nucleus has become partially elongated (the early canoe stage, described below) (Rathke et al., 2007). It has been speculated that histone modifications on H4 and H2A immediately precede the histone to protamine transition in flies by facilitating access to the DNA (Rathke et al., 2007; Awe and Renkawitz-Pohl, 2010). In flies, there is a single known TNP, the HMG-box domain protein Tpl94D (Rathke et al., 2007). An additional three HMG-box domain proteins are expressed in fly testes (tHMG-1, tHMG-2, and HMGZ), two of which are expressed most highly during the periods of histone to protamine transition. While none of these genes/proteins have been shown as essential for DNA condensation, functional redundancies between family members are hypothesized (Gartner et al., 2015). Similar to mammalian TNPs, Tpl94D acts as the placeholder for protamines during histone removal but is not a direct orthologue to the mammalian setting (Rathke et al., 2007). A recent study has shown a conserved role for testis-specific serine/threonine kinases in the histone to protamine transition from flies to humans (Zhang et al., 2023).

While human and mouse spermatids express two major types of protamines (P1 and P2), flies contain three proteins that function in a similar role in DNA condensation: protamine A (Mst35Ba), protamine B (Mst35Bb), and a histone H1-like linker protein—Mst77F (Rathke et al., 2010; Kimura and Loppin, 2016). While Mst35Ba and Mst35Bb have no direct mammalian orthologues, they both contain cysteine/arginine repeats that are also present in mammalian protamines and function in DNA compaction during spermiogenesis (Jayaramaiah Raja and Renkawitz-Pohl, 2005). Mst77F, which bears significant sequence similarity to HILS1 (Rathke et al., 2014) appears to play an additional role in nuclear maturation, as it associates with β-tubulin in perinuclear microtubules to assist in sperm head shaping (Rathke et al., 2010). Mst77F also appears to be the most essential of the three proteins for chromatin condensation as mutants present with small, round nuclei and males are sterile (Jayaramaiah Raja and Renkawitz-Pohl, 2005).

At the start of nuclear shaping, the round spermatid nucleus flattens at the side of the cell where the basal body is located, prior to the start of DNA condensation (Riparbelli et al., 2020). A network of perinuclear microtubules that forms asymmetrically along the base of the nucleus then arranges into a series of longitudinal bundles, to form a microtubule and actin-rich structure known as the dense complex (or dense body) (Anderson, 1967). The dense complex is thought to be largely analogous to the mammalian manchette but appears to associate with just one side of the elongating nucleus rather than encircling it like a skirt (Fabian and Brill, 2012; Augiere et al., 2019; Riparbelli et al., 2020). This dense complex grows along the convex side of the elongating sperm head and its biogenesis does not appear to rely on the presence of sperm centrioles (Riparbelli et al., 2020). Conflicting data exist regarding the importance of centrioles in dense complex function (reviewed in Fabian and Brill, 2012). It is thought that the dense body is responsible for the early shaping of the nucleus as it transitions from round to leaf stage as well as nuclear shaping and providing structural support in the transition from canoe to needle-shaped morphology (Tokuyasu, 1974a). Recently, a concentration of gamma-tubulin has been found to associate with the apical pole of spermatid nucleus during head remodelling, suggesting a role of the dense complex in nucleating microtubules or potentially a role in acrosome formation that influences head shaping (Riparbelli et al., 2020). At the conclusion of head shaping and elongation, fly spermatids possess a long, thin, needle-shaped sperm head. Similarly, defects in nuclear remodelling result in abnormal, often curved or bent, head shapes (Texada et al., 2008; Kracklauer et al., 2010; Augiere et al., 2019).

In summary, while the complement of proteins involved in the histone to protamine transition differs between Drosophila and mammals, the processes involved follow a similar series of events. Sperm head shape is highly variable across species, even within mammals and notably between mice and men. While human sperm exhibit a spatulate/oval shape, flies generate needle-shaped sperm. Though the processes involved in sculpting the sperm head across species are not fully understood, one conserved feature is that the manchette (mammals) and dense body (flies) are both microtubule and actin-rich structures and indispensable for normal sperm head shaping.

Sperm tail formation

The axoneme is a conserved microtubule-based structure that forms the core of the sperm tail (and somatic cilia) and drives sperm motility. One notable difference between species is that Drosophila melanogaster sperm tails (∼1.8 mm, although variable across fly species) are enormous and estimated to be around 36 times the length of human sperm (50 µm). To initiate axoneme formation in mammals, early round spermatid centrioles undergo a maturation phase to develop into a basal body, which implants into the posterior side of the nucleus at the opposite side of the forming acrosome and attaches to the nuclear membrane; reviewed in Lehti et al. (2017) and Pleuger et al. (2020). The basal body comprised the proximal (daughter) centriole, which plays a role in generating a centriolar adjunct, and the distal (mother) centriole, from which microtubules extend to form the axoneme (Fawcett and Phillips, 1969; Avidor-Reiss et al., 2019). Both centrioles are involved in forming the head–tail coupling apparatus, which fortifies the connection between the sperm head and tail (Tapia Contreras and Hoyer-Fender, 2019).

To build a sperm tail in both mammals and flies, a compartmentalized cilium is built within a cellular compartment that is isolated from the rest of the spermatid cytoplasm. To achieve compartmentalization, a complex cellular gate called the transition zone (TZ) (Vieillard et al., 2016) forms directly from the distal centriole at the start of axoneme development and remains at the interface between the cytoplasmic and ciliary compartments for the entirety of tail development. The TZ plays an essential role in governing the selective transport of proteins and cargoes into the ciliary compartment (Awata et al., 2014). This includes components of the sperm-specific accessory structures such as the outer dense fibres (ODFs) and fibrous sheath (Houston et al., 2024a)—structures unique to mammalian sperm that are required for sperm motility and are not found in insect sperm. The modes of sperm tail development and differences between species will now be discussed in more detail.

Mammals

Axoneme formation begins early during mammalian spermiogenesis, when the basal body attaches to what will become the implantation fossa, on the posterior side of the nucleus (Lehti et al., 2017). Once the basal body has bound to the nucleus and plasma membrane, the axoneme begins to extend through a version of a bidirectional process called intra-flagellar transport (IFT) (Kierszenbaum et al., 2011; San Agustin et al., 2015; Avidor-Reiss and Leroux, 2018). IFT is highly conserved in all eukaryotic cilia and flagella, and uses motor proteins to transport vesicles and protein complexes along the axoneme microtubules, resulting in the extension of the microtubule axoneme itself and the transport of key sperm cilia/flagella proteins for incorporation within the tail (reviewed in Pleuger et al., 2020).

The sperm tail axoneme is composed of ‘9 + 2’ microtubule formation wherein a ring of nine microtubule doublets surrounds a central pair of singlet microtubules in humans and other mammals (Baccetti et al., 1981; Leung et al., 2021). Each doublet is connected to adjacent doublets by thin nexin linkers, while radial spokes connect each doublet to the central pair (Bozkurt and Woolley, 1993). The radial spokes and nexin links regulate motility by optimizing the degree of axoneme bending, aiding in structural stability, and may contribute to dynein motor function (Heuser et al., 2009; Kobayashi and Takeda, 2012; Zhu et al., 2017; Morohoshi et al., 2020). One final component of the core axoneme is the dynein arms, which attach to the outer microtubule doublets and are likely loaded by a variant of IFT (Pleuger et al., 2020). To effect sperm motility, the dynein arms hydrolyse ATP provided from major metabolic pathways in the midpiece and principal piece to induce microtubule sliding.

Once the core scaffold of the mammalian sperm tail (the axoneme) is built, it is next endowed with sperm tail accessory structures. These include the fibrous sheath, which is loaded into what will become the principal piece of the sperm tail (Brown et al., 2003; Miki et al., 2004); the ODFs that run along the entire length of the sperm tail except for the end piece (Irons and Clermont, 1982a, b; Clermont et al., 1990; Zhao et al., 2018), and the mitochondrial sheath that is loaded into the midpiece compartment (Otani et al., 1988; Pleuger et al., 2020). The fibrous sheath is the first accessory structure to begin formation, though is largely loaded in parallel with ODFs (discussed below). The loading of both structures is likely aided by a mechanism of intramanchette transport (Lehti et al., 2013; Zhang et al., 2020). Intramanchette transport utilizes the microtubules of the manchette, which project from the perinuclear ring distally into the cytoplasmic lobes of the spermatid, and towards the centriolar adjunct. It thus allows transport of proteins from the apical regions of the spermatid and cytoplasmic lobes to the centriolar adjunct for entry into the ciliary compartment (reviewed in Kierszenbaum, 2002; Pleuger et al., 2020).

The fibrous sheath is composed of numerous proteins (largely A-kinase anchor proteins) required for structural support during ciliary beating, anchoring of glycolytic enzymes, and in signal transduction for capacitation (Brown et al., 2003; Eddy et al., 2003; Miki et al., 2004). Fibrous sheath components are first loaded at the distal end of the axoneme and progressively backfilled along the tail towards the nucleus (Irons and Clermont, 1982a,b).

ODFs are the second accessory structure to commence loading into the sperm tail. ODFs contribute to the elasticity and structural support of the beating sperm tail and are thought to protect against any shearing forces the sperm may be subjected to during ejaculation and in the female reproductive tract (Baltz et al., 1990; Zhao et al., 2018). They are also the site for mitochondrial attachment within the midpiece (Zhang et al., 2004; Pleuger et al., 2020). Prior to their loading into the sperm tail, ODFs are translated and believed to accumulate in the cytoplasm as granulated bodies (Clermont et al., 1990). They are then loaded in a proximal to distal manner along the tail, across what will become the mid and principal pieces. Nine ODFs surround the axoneme core in the midpiece, associating with the microtubule doublets. ODFs 3 and 8 are replaced by longitudinal columns of the fibrous sheath in the principal piece compartment (Fawcett, 1975; Ricci and Breed, 2005). As is the case for fibrous sheath formation, IMT is hypothesized to be important for the trafficking of ODF components for their entry into the sperm tail compartment (Rivkin et al., 2008; Dunleavy et al., 2017; Pleuger et al., 2020).

Unlike the fibrous sheath and the ODFs, the mitochondrial sheath of mammalian sperm is not loaded via the TZ, but directly from the cytoplasm (Otani et al., 1988). Towards the end of sperm tail development, TZ activity is lost when an overlapping and related structure called the annulus migrates distally along the axoneme (Houston et al., 2024a). The annulus will ultimately demarcate the junction between the mid and principal pieces of the sperm tail, noting that fibrous sheath formation is complete before annulus migration (Kissel et al., 2005). The movement of the annulus and the associated plasma membrane exposes a segment of the ODFs, surrounding the axoneme, to the cytoplasm. It is to this region of the ODFs that mitochondria attach to form the mitochondrial sheath (Kwitny et al., 2010). Up until this point, the entire sperm tail has been built in an isolated ciliary compartment (Kwitny et al., 2010; Toure et al., 2011). The mechanisms by which the mitochondria are recruited from the cytoplasm are poorly understood (Pleuger et al., 2020). Once built, the mitochondrial sheath houses a large number of mitochondria, which accumulate end-to-end in a helix and directly attach to the ODFs and each other via inter-mitochondrial linker complexes (Otani et al., 1988; Shimada et al., 2021).

In summary, the mammalian sperm tail axoneme is generated by a process of IFT and intramanchette transport, followed by supplementation of the fibrous sheath and ODFs. After annulus migration and exposure of the midpiece to the cytoplasm, mitochondria are loaded onto the proximal region of the ODFs (surrounding the core axoneme).

Drosophila

Like the human sperm axoneme, the Drosophila sperm axoneme is arranged in the ‘9 + 2’ microtubule structure. However, in Drosophila, an additional set of ‘accessory microtubules’ exists, surrounding the 9 + 2 scaffold and forming what is sometimes referred to as a 9 + 9 + 2 structure (Nielsen and Raff, 2002; Fatima, 2011) (Fig. 4). The accessory microtubules are initially loaded as what is described based on their shape as ‘C shaped blades’ in early tail development. As the axoneme elongates, the accessory microtubules morph into a closed, barrel shape (Gottardo et al., 2018). Further, via a poorly defined process, an electron dense material becomes deposited at the site of accessory microtubules. During elongation, the central pair of microtubule doublets is also loaded and matured (Gottardo et al., 2018). Finally, in sperm, the accessory microtubules ultimately bear resemblance to the mature central pair of microtubule doublets in mammalian species (Fig. 4). The Drosophila axoneme also includes nexin links, radial spokes, and dynein arms in a structure conserved with that of mammalian sperm (Mencarelli et al., 2008). A notable difference between human and Drosophila sperm, is that fly sperm do not contain two of the three accessory structures characteristic of mammalian sperm: the fibrous sheath and ODFs. Despite no strong evidence that true ODF-like structures exist in fly sperm, research has identified several genes expressed in fly testes (Mst98Ca, Mst98Cb, and Mst87F) that bear some sequence homology to mammalian ODFs (Burfeind and Hoyer-Fender, 1991; Schafer et al., 1993; Blumer et al., 2002). We, however, predict that the location of the accessory microtubules indicates an analogous structural role to ODFs.

Figure 4.

Sperm tail development and axoneme structure in humans and flies. (A) Sperm tail development in humans begins in round spermatids (i), with the docking of a basal body to the nuclear and plasma membranes, in order to construct the axoneme in a restricted ciliary compartment (ii), and the formation of a transition zone to regulate entry into this compartment. Following formation of the core microtubule axoneme (iii), components of the fibrous sheath and, at an overlapping time, outer dense fibres, are loaded into the sperm tail (iii–iv). Following, the annulus migrates and forms a diffusion barrier between the mid and principal pieces (iv). At the same time, the annulus drags the plasma membrane down, exposing a short region of the axoneme to the spermatid cytoplasm, which becomes the midpiece compartment (iv). Mitochondria are rapidly assembled into this compartment (iv) and remaining cytoplasm is stripped during the process of sperm individualization (v). as, acrosome; n, nucleus; bb, basal body; tz, transition zone; mc, mitochondria; an, annulus; ax, axoneme; fs, fibrous sheath; ht, head–tail coupling apparatus. (B) In flies, sperm tail development begins in earnest in spermatocytes (i) with the docking of basal bodies to the plasma membrane and the development of a cellular gate, the transition zone in a small ciliary pocket. In round spermatids (ii) the basal body docks to the nuclear membrane and the nebenkern forms as a precursor of the mitochondrial derivatives. As the axoneme is constructed, the transition zone migrates and maintains a fixed distance from the tip of the tail (ciliary compartment). At the same time, the mitochondrial derivatives unfold and aid in driving axoneme elongation within the cytoplasmic compartment. cyto, cytoplasm; ab, acroblast; nk, nebenkern; md, mitochondrial derivative. (C) While the core sperm axoneme structure is highly conserved between humans and flies, a set of accessory microtubules surrounds the ‘9 + 2’ axoneme and bear some resemblance to the central pair doublet structure. In human (and mammalian) but not fly sperm, outer dense fibres and a fibrous sheath encase the axoneme. Human and fly sperm both contain a mitochondrial sheath structure—in humans, this comprised a set of intercalated mitochondria in a left-handed helix and is restricted to the midpiece; in flies, this is represented by a major and minor mitochondrial derivative that sit adjacent to, and span the entire length of, the sperm tail. MiM, minor mitochondrial derivative. Parts of this figure are adapted from Dunleavy et al. (2019). Some structures are based on electron microscopy from Tokuyasu (1974a, b), Li et al. (1998), and Gottardo et al. (2018).

The first step in axoneme formation in Drosophila involves attachment of four basal bodies to the cell membrane in early spermatocytes, much earlier than in round spermatids in mammals (Basiri et al., 2014). This event stimulates maturation of the basal body, elongation of the centriole, and the formation of a short, compartmentalized primary cilium. As highlighted by localization of Mks1, Cep290, and other TZ proteins, the short cilium is thought to be composed primarily of TZ proteins (Basiri et al., 2014; Vieillard et al., 2016; Avidor-Reiss et al., 2017). During the meiosis period, the centrioles internalize and drag the cell membrane inwards to form an invagination that maintains the short cilium in its own ciliary compartment due to the presence of a ciliary cap (Basiri et al., 2014; Vieillard et al., 2016). It is in round spermatids, as is the case in mammals, that the basal body attaches to the posterior side of the nucleus, prior to initiating axoneme and sperm tail development (Fabian and Brill, 2012). To do so, microtubules from the basal body extend distally out of the TZ, strictly within the ciliary compartment of the tail. A key difference between species is that, in flies, spermatid tail elongation occurs in the absence of IFT pathways and appears to be driven by extension of the mitochondrial derivate (Han et al., 2003; Noguchi et al., 2011). Further, spermatid tail elongation can occur in the absence of an axoneme, resulting in a non-functional tail-like structure with a mitochondrial derivative but no core motility machinery (microtubules or dynein arms) (Basto et al., 2006). A second major difference between fly and mammalian spermatids is that the TZ continually travels away from the basal body in flies as the tail is being built (Fig. 4). As it translocates, the TZ maintains a fixed distance from the end of the sperm tail (∼2 µm) and this portion of the tail remains encased in the ciliary membrane, protected by the ciliary cap that is present from the spermatocyte stage (Basiri et al., 2014). In turn, the assembled axoneme (proximal to the TZ) is exposed to the cytoplasm of the cell (Avidor-Reiss and Leroux, 2018). Upon exposure of the nascent tail to the cell cytoplasm, additional proteins are incorporated within or around the axoneme, including the large mitochondrial derivatives, which are loaded in a proximal to distal manner (Tokuyasu, 1975). As discussed below, this difference compared with mammalian sperm tail development is likely due to the role of the mitochondrial derivative in driving elongating of the sperm tail (Noguchi et al., 2011).

Unlike in mammals, mitochondria are loaded onto the fly sperm tail concurrently with axoneme development (Vedelek et al., 2024). Specifically, at the same time as the basal body docks to the nucleus in each spermatid, multiple mitochondria begin to aggregate alongside the basal body and commence undergoing fusion that sees formation of the nebenkern (Tokuyasu, 1975; Aldridge et al., 2007; Dorogova et al., 2013). The nebenkern is a unique structure found in the spermatids of some insects and is composed of two large, intertwined mitochondrial derivates, the major and minor derivates (Tokuyasu, 1975). Transmission electron microscopy images of the mitochondrial derivative in round spermatids revealed the interleaved derivatives as a structure that appears similar to a sliced onion, hence this stage of spermatid development is also referred to as the ‘onion stage’. The fate of which derivative becomes the major and minor is determined by the protein Bb8, which also aids in development of the derivatives (Vedelek et al., 2016). Similarly, the ATP synthase subunit protein ‘Knotted Onions’ is likely required for optimal membrane curvature of the derivatives (Sawyer et al., 2017). Both derivatives are loaded onto, but adjacent to, the axoneme, rather than helically arranged around the axoneme as seen in mammalian sperm.

The mitochondrial fusion processes that mediate derivative formation are controlled by an essential protein named ‘Fuzzy Onions’. In the absence of fuzzy onions, male flies are sterile due to the formation of multiple smaller onion-shaped mitochondrial derivatives (Hales and Fuller, 1997). Similarly, processes of mitochondrial fission are essential for spermatogenesis and particularly during spermiogenesis, for the process of nebenkern/derivative unfurling. Dynamin-related protein 1, a key regulator of fission, is required for normal mitochondrial derivative unfurling from the nebenkern, as well as their elongation for sperm tail loading (Aldridge et al., 2007). Altogether, numerous proteins are involved in harmonious mitochondrial derivative formation and unfurling. A final example includes the conserved Parkin (Parkinson’s disease) protein, which is required for both mitochondrial derivates to form normally and then unfurl from the nebenkern (Riparbelli and Callaini, 2007). In its absence, just one derivative unfurls alongside the axoneme.

Mitochondrial elongation is one of the primary drivers of fly sperm tail extension and it has been suggested that the mitochondrial derivatives play a structural role for the extremely long sperm tails (Noguchi et al., 2011). The mitochondrial adapter protein Milton, which binds kinesin motor protein, facilitates anchoring of the nebenkern near the base of the spermatid nucleus (Aldridge et al., 2007). Nebenkern anchoring to the axoneme is potentially strengthened by another protein, Merlin, while Milton is also required for elongation of the mitochondrial derivatives along the sperm tail (Aldridge et al., 2007; Dorogova et al., 2008). When axoneme development is normal but mitochondrial elongation is impaired, fly sperm tails become bent, suggesting that the axoneme foundation is not strong enough to support tail elongation alone (Noguchi et al., 2011). Further, microtubules assemble alongside the derivatives independently of typical microtubule-organizing centres (Noguchi et al., 2011; Chen et al., 2017). These newly formed microtubules polymerize in the cytoplasm, on the surface of elongating mitochondrial derivatives, and are largely composed of a different beta tubulin (beta 3) than that present within the axoneme (beta 2) (Hoyle and Raff, 1990). These microtubules are crosslinked with the mitochondria, and are then able to generate a sliding force, which elongates the mitochondria (Noguchi et al., 2011). As extension progresses, additional microtubules are crosslinked to those microtubules attached to the mitochondria, in a proximal to distal wave. This event slows the lengthening process and secures the derivatives in place, forming a stable zone, while the derivatives at more distal points continue growth to be loaded onto the tail (Noguchi et al., 2011). During the processes of individualization (discussed below) a large portion of the minor mitochondrial derivative is stripped away from the axoneme (Fabrizio et al., 1998; Bazinet and Rollins, 2003; Steinhauer et al., 2019).

In summary, fly sperm tail structure and development vary significantly between mammals and flies. While both species have a ‘9 + 2’ axoneme microtubule structure, the fly has an additional ring of accessory singlet microtubules, exhibiting a ‘9 + 9 + 2’ structure. In mammals, the axoneme forms entirely as a long, compartmentalized cilium, after which a short region of the proximal axoneme, which ultimately becomes the midpiece, is exposed to the cytoplasm concurrently with annulus migration. Conversely, fly sperm exhibit a short, compartmentalized cilium that maintains a fixed distance from the distal end of the sperm tail, and virtually the entire tail is ultimately exposed to the cytoplasm. The mammalian sperm tail contains accessory structures, the ODFs, and fibrous sheath, that are not found in insect sperm. While both species contain mitochondria within the tail, fly sperm harness a large major mitochondrial derivative and a slim minor derivative that span its entire length. In contrast, mammalian sperm mitochondria are confined to the midpiece segment.

Intercellular bridges, individualization, and spermiation

Spermatogenesis concludes with the individualization of sperm and the removal of excess cytoplasm and cytoplasmic bridges that exist between germ cells (reviewed in Greenbaum et al., 2011). Sperm are released from the testis via the processes of spermiation.

Mammals

During the process of spermiation, mammalian sperm are individualized, whereby their excess cytoplasm, organelles, and plasma membrane including cytoplasmic bridges are removed as ‘residual bodies’, which ultimately become phagocytosed by Sertoli cells. While the processes of sperm individualization are better described in Drosophila, a similar process occurs in mammalian sperm. First, cytoplasmic bridges between connected spermatids are severed to allow spermatid separation (Dym and Fawcett, 1971). While proteins such as TEX14 and CEP55 have been identified as key for the formation and maintenance of these bridges (Greenbaum et al., 2006; Sinha et al., 2018), the factors involved in their removal are poorly understood. It has, however, been highlighted that mammalian sperm individualization requires members of the autophagy and neddylation pathways (Huang et al., 2019, 2021). Autophagy protein ATG5 appears to play roles in the removal of excess cytoplasm and is required for the processes of spermatid individualizing that allow normal spermiation (Huang et al., 2021). Similarly, the neddylation-regulating protein DCUN1D1 appears to play a role in cytoplasmic bridge removal and individualization (Huang et al., 2019). We have also identified that the dynein light chain-containing protein AXDND1 is required for sperm individualization during spermiogenesis (Houston et al., 2024b). During the process of individualization, the majority of the sperm cytoplasm containing excess mitochondria, the Golgi, and several other structures, is removed and displaced as the residual body (Breucker et al., 1985). The residual body develops concurrently or soon after annulus migration as the cytoplasm buds off and shifts to one side of the flagellum, where it is then separated into unequal elements. Tubulobulbar complexes (TBCs) then remove excess cytoplasm from the spermatid to allow for a streamlined cellular shape (Russell et al., 1989; Adams et al., 2018; Huang et al., 2021). The larger element of the residual body is collected and phagocytosed by Sertoli cells, while the smaller portion remains attached to near the annulus as the cytoplasmic droplet. Failure to remove excess cytoplasm can impair sperm cytoplasmic droplet formation, acrosome reaction, and sperm individualization (Li et al., 2024a).

During the process of spermiation, spermatids are individualized and then released from the Sertoli cells into the lumen of the seminiferous tubules. This process begins when structurally mature elongated spermatids line up along the edge of the lumen (O’Donnell et al., 2011). Actin-rich TBCs then develop between the spermatid head and penetrate the surrounding Sertoli cells (Russell, 1979; Russell and Malone, 1980; Upadhyay et al., 2012). In parallel to TBC development, ectoplasmic specialization complexes are cleaved to allow disengagement of spermatids from Sertoli cells (Guttman et al., 2004; Dunleavy et al., 2017). Ectoplasmic specialization complexes are junctions that appear early in spermiogenesis between the Sertoli cell and round spermatids (Grove and Vogl, 1989; O’Donnell et al., 2011) that aid in moving germ cells through the seminiferous epithelium (Wong et al., 2008). Interestingly, TBC protein localization has been found to vary between species. In humans, which have a spatulate-shaped head, TBCs are found at the apical tip of the head, while in the sperm of rodents, TBCs form in the inner curve of the sickle or falciform-shaped head (Russell, 1979; Russell and Malone, 1980; O’Donnell et al., 2011).

The final step of spermiation, termed disengagement, involves the complete detachment of spermatids from the seminiferous epithelium, allowing their release into the lumen. Numerous mutant mouse models exist where poorly formed sperm are retained by Sertoli cells, leading to what is known as a partial spermiation failure, potentially as a sperm quality control mechanism (O’Bryan et al., 2008; O’Donnell et al., 2014; Fu et al., 2018; Zhang et al., 2021; Xue et al., 2022; Houston et al., 2024a). Further, this process is highly reliant on hormone regulation, and ablating testosterone levels in humans results in a dramatic suppression of spermiation (Saito et al., 2000; McLachlan et al., 2002). Following the removal of ectoplasmic specialization complexes, spermatids remain in contact with Sertoli cells via other undefined specialized junctions (Beardsley and O’Donnell, 2003; Dunleavy et al., 2017). Finally, released sperm travel via the rete testis to the epididymis, where they undergo numerous post-testicular modifications and are stored prior to ejaculation.

Drosophila

In flies, the final phase of spermiogenesis involves the breaking of intercellular bridges and the individualization of each spermatid into its own membrane. Like in mammals, this process involves the removal of the excess cytoplasm and organelles from the elongated spermatid (Tokuyasu et al., 1972a). To facilitate this process, the individualization complex (IC) is formed, which travels from the head to the tail of the cyst. The IC is made up of 64 investment cones (one per spermatid) containing F-actin that form an arrow-shaped structure, each localized between spermatids (Fabrizio et al., 2020). The investment cones migrate distally from the nucleus to the tip of the tail (Fabrizio et al., 1998). As the IC travels along the spermatid, excess cytoplasm is removed from between the individualizing sperm. Ultimately, doing so results in a cystic bulge that comprises the degenerated acroblast and other organelles that are no longer required by the spermatids (Noguchi and Miller, 2003). When the IC reaches the tip of the tail, it (and the cystic bulge) detaches and is referred to as a ‘waste bag’ and is analogous to the residual body seen in mammalian testes (Tokuyasu et al., 1972b; Fabrizio et al., 1998). At least 70 genes have been identified to play essential roles in spermatid individualization in flies (Steinhauer, 2015), many of these involved in cytoskeletal dynamics that permit efficient membrane severing, or normal translational activation/repression (Li et al., 2019; Yuan et al., 2019; Fabrizio et al., 2020; Chen et al., 2021; Gilmutdinov et al., 2021; Vedelek et al., 2021). While none of these gene orthologues have yet been defined to be required for sperm individualization in humans or other mammals, CUL3 may be involved in the process in both species (Huang et al., 2019).

Due to their length, Drosophila spermatids undergo an additional step not seen in mammals in which the sperm tail is coiled into smaller loops. Coiling involves the looping of each of the 64 spermatid tails contained within the same cyst, via the retraction of the tail to the distal (seminal vesicle-connected) end of the testis (Tokuyasu et al., 1972b; Ghosh-Roy et al., 2004; Maaroufi et al., 2022). The coiling process is initiated at the head region of the spermatid bundle, which at this point, is located at the distal side of the testis, i.e. the opposite side to the hub. The sperm head is secured by the terminal epithelium of the testis, allowing the tail to retract towards it (Tokuyasu et al., 1972b). Specifically, the spermatid heads maintain contact with the head cyst cell via septate junctions, which are analogous to mammalian tight junctions (Dubey et al., 2019). Proteins Discs-large-1 and Neurexin-IV play essential roles in establishing these junctions and their absence results in premature release of spermatids (Dubey et al., 2019). As the sperm bundle coils, the tails are packed into a hexagonal lattice. The complexity of this process also allows any defective sperm to be removed from the bundle, via their segregation into a distinct area within the cyst lumen (Tokuyasu et al., 1972a, b). In summary, while flies undergo an additional step of coiling during their process of spermiation, the initial steps of individualization appear to be largely similar, resulting in the formation of a ‘waste bag’ analogous to the residual body, and the stripping of most of the sperm cytoplasm. As defined above, however, germ cell cysts are encapsulated by two cyst cells. The processes of spermatid disengagement are not as well studied in flies and as in mammals, but in both situations spermatids are strongly associated with supporting somatic cells via tight/septate junctions before sperm release.

Discussion

Drosophila melanogaster is undoubtedly a useful model organism for studying a wide range of human genetic diseases, including male infertility. This review has highlighted some significant differences, and the many similarities between Drosophila and human spermatogenesis (key points are summarized in Tables 1 and 2).

Table 2.

An overview of the main similarities and differences between the phases and processes of spermatogenesis in mammals (with a focus on humans) and Drosophila.

Process	Similarities between humans and flies	Differences in flies
Germ cell divisions	The ratio of germ cells from each stem cell division is identical between species.	No notable differences.
Sertoli/cyst cell divisions	Both cell types encapsulate germ cells and support their maturation in the testis.	Cyst cells are sourced from a pool of continually dividing cyst stem cells. They migrate with the germ cells in flies.
Meiosis	Kinetochore attachment and chromosome segregation are comparable.	Crossing over does not occur—no true synaptonemal complexes are formed. Alternate cohesin proteins are utilized.
Spermiogenesis – Acrosome formation – Nuclear shaping – Sperm tail formation – Spermiation	The acrosome is required at a structural level and is, in part, developed by similar Golgi-derived pathways. A microtubule-based structure shapes the sperm head in both species (manchette and dense body). DNA condensation follows a similar process, utilizing transition proteins and protamines. The core axoneme is built in the cilia compartment using a transition zone to segregate it from the cytoplasmic compartment. Sperm are individualized into separate cells via similar cytoskeletal-driven processes. The waste bad is analogous to the residual body.	The acrosome is not required for hydrolysis of barriers surrounding the egg. Endocytic pathways may not contribute to acrosome development in flies. Sperm heads are needle-shaped in flies (though sperm shape varies widely across species). The specific proteins involved in DNA condensation are not orthologues. A larger portion of the tail becomes exposed to the cytoplasm due to the continuous migration of the transition zone. Fly sperm axonemes contain and additional row ring of accessory microtubules (singlets) but no fibrous sheath. Fly sperm are not similarly segmented and contain large mitochondrial derivates along their entire length. Fly sperm undergo a process of coiling due to their immense length.

While mice may not be as reflective of human biology as some closer related species such as non-human primates, they are often deemed the gold standard for medical research due to significant gene conservation (99%) with humans and the ease for genetic manipulation and housing. Notwithstanding, mice are expensive to house and their use is highly regulated. By comparison, flies share 75% of gene content with humans for ‘disease-causing genes’ and have a short life cycle of ∼2 weeks. They can also be rapidly modified to generate mutant lines, and a wealth of commercially available mutant lines already exist. Moreover, a recent study has identified that 2343 testis-expressed genes are conserved from humans to flies (Wang et al., 2025a). By comparison, 3502 genes were conserved from humans to mice, reinforcing Drosophila as a valuable model. When considering humans, flies, and mice, 1277 genes were conserved across the three species. The benefit of using Drosophila as a model is that it provides a fast and inexpensive method to screen a large number of conserved genes. This interpretation is strengthened by a separate study which showed that at least 11% of fly genes when mutated lead to male sterility (Hackstein et al., 2000). Large domestic species such as pigs and bulls are likely unsuitable for similar studies that can be achieved in mice and flies. However, in some cases, they may offer specific advantages through forward genetic screens, i.e. linked gene function after identifying an infertility phenotype through multi-omics or genome-wide associated screens.

Underscoring their utility, fly models have provided several key insights into human gene function in spermatogenesis. Vasa is an RNA helicase essential for germ cell function and perhaps the most widely known example of an essential and conserved fertility gene first discovered in flies (Hay et al., 1988). Subsequently, an analogous function of Vasa in spermatogenesis in numerous species, including humans (DDX4), has been confirmed (Castrillon et al., 2000; Chuma et al., 2006). DDX4 has since been leveraged as a driver (e.g. Ddx4-Cre in mice) to generate germ cell-specific knockout lines and used as a specific germ cell marker in humans and mice (Guo et al., 2021). As another example, the fly gene boule was found to encode a translational regulator required for male germ cell entry into meiosis and spermatid differentiation (Eberhart et al., 1996; Cheng et al., 1998). boule was confirmed as a single gene homologue of the DAZ1 (Deleted in Azoospermia) and DAZL (DAZ-like) genes in humans. Further study of DAZ genes identified multiple DAZ family members important for spermatogenesis (Eberhart et al., 1996; Yen et al., 1997). Further, an interactor of boule, twine, was validated in flies to be essential for entry into meiosis (Alphey et al., 1992; Maines and Wasserman, 1999) and its orthologue CDC25A identified as an essential fertility in humans (Pal et al., 2025). Other key examples include doublesex (dsx), pelota, and fuzzy onions (fzo). dsx plays a role in sex determination and male fertility in flies (Colaianne and Bell, 1972), similar to the orthologous DMRT (doublesex and mab3-related-transcription factor) family in mammals, including humans (Bellefroid et al., 2013; Zhao et al., 2015; Zarkower and Murphy, 2022). pelota is required for meiotic cell division in fly spermatogenesis (Eberhart and Wasserman, 1995), and its mammalian orthologue Pelo is required for SSC maintenance (Raju et al., 2015). fzo is another classic example of an evolutionary conserved gene discovered in flies. fzo is a dynamin-type GTPase that is essential for mitochondrial fusion during spermiogenesis (Hales and Fuller, 1997). The mammalian orthologues, mitofusins 1 and 2 (MFN1/2), play a similar role in mitochondrial fusion during spermatogenesis (Santel and Fuller, 2001; Varuzhanyan et al., 2019). More recent examples of findings from flies that will inform mammalian biology include: Poldip2, which was recently shown to be essential for paternal mitochondrial elimination during fertilization (Wang et al., 2025b); AK-3, which is essential for sperm tail axoneme central pair microtubule doublet formation (Chen et al., 2025); and RNF113, which has been implicated in male germ cell transcription broadly (Brattig-Correia et al., 2024). Collectively, these examples exemplify the value of using flies as a tool to define human male germ cell function.

While throughout this review we have compared the similarity in spermatogenic processes, it is important to note that the extent of gene conservation between species is perhaps a more relevant indication of whether the gene should be studied in flies. This can be tested using tools such as DIOPT and BLAST. As outlined in recent studies, a core network of spermatogenesis genes is conserved across metazoans and notably between humans, mice, and flies (Brattig-Correia et al., 2024; Wang et al., 2025a). To improve the efficiency of Drosophila screens, a standardized pipeline for fertility testing and histological analysis, including rescue with the human orthologue where possible, should be implemented. Additional data, such as available germ cell mRNA expression profiles (Witt et al., 2019), in situ hybridization and antibody localization data could also be used to ascertain the germ cell types in which the gene of interest is expressed and whether this correlates to humans and other mammals (Jung et al., 2019; Mahyari et al., 2021). While this would allow the ranking of priority candidates, the efficiency of generating fly models allows the parallel testing of numerous genes. More than one RNAi or mutant line should be tested where possible, and the RNA level of the gene of interest validated in mutants, to ensure the phenotype (or lack thereof) is due to specific knockdown. Notwithstanding these points, in this review, we have revealed that many of the processes required for fly spermatogenesis are similar to those in humans and other mammals.

Due to some fundamental differences in spermatogenesis in flies as outlined throughout this review, we recommend that further analysis of the genes identified by fly screens as potentially disease causing in humans should be studied in other model organisms, such as mice. This approach will further ascertain function or pathogenicity. It is our thinking that the absence of a phenotype in mutant flies does not always suggest the gene is not required for human spermatogenesis, but rather the presence of a phenotype provides additional confidence and mechanistic data to a gene’s involvement in sperm production. Validation should be attempted in healthy human and/or infertile patient tissue where possible.

Data availability

No original data are presented in this manuscript.

Authors’ roles

B.J.H., L.M.C., and M.K.O.B. wrote the review. J.E.M.D., R.B., and G.R.H. edited and provided feedback.

Funding

None to declare.

Conflict of interest

None to declare.