Introduction

Reproduction is critical for evolutionary success, and animals display great diversity in their reproductive life histories. The structures and strategies employed for reproduction have been relatively well-studied in some animal lineages, such as arthropods and vertebrates (Davies et al., 2013). However, a full understanding of reproductive biology is lacking in many animal phyla. For example, Phylum Xenacoelomorpha, which includes acoels, nemertodermatids, and xenoturbellids, is an early branching bilaterian lineage of primarily marine worms whose life histories are largely unknown (Cannon et al., 2016; Kapli and Telford, 2020; Philippe et al., 2011) (Fig. 1A).

Figure 1:

Roughly 400 species of acoel worms have been described (Achatz et al., 2013; Jondelius et al., 2011; Tyler, S., Schilling, S., Hooge, M., & Bush, L. F., 2012; Wallberg, 2012). Analyses of histological sections of specimens collected in the field suggest that, with few exceptions (Crezée, 1975; Mamkaev, 1965; Raikova et al., 1995), acoels are simultaneous hermaphrodites. Unlike many marine animals that reproduce sexually through spawning and external fertilization, e.g., cnidarians, poriferans, echinoderms, etc. (Giribet and Edgecombe, 2020), acoels have been reported to reproduce sexually through internal fertilization (Achatz et al., 2013). They exhibit a striking diversity in their reproductive anatomy (Table S1). This anatomical variation likely accompanies variation in reproductive physiology and behavior. While a few species have been cultured in the lab (Bailly et al., 2014; De Mulder et al., 2009; Shannon and Achatz, 2007; Zauchner et al., 2015), little is known about how acoels develop their reproductive organs, mate, or lay eggs.

To improve our understanding of acoel life histories, we studied reproduction in a lab tractable species, the three-banded panther worm Hofstenia miamia (Fig. 1B,C). H. miamia is a new research organism with features that allow us to study it in controlled conditions: its life cycle can be closed in the lab (worms develop from embryo to gravid adult in 2-3 months), and there is a growing array of experimental tools to observe and manipulate the worms (Gehrke et al., 2019; Hulett et al., 2023; Kimura et al., 2022, 2021; Ricci and Srivastava, 2021; Srivastava, 2022). Histology-based work has described the coarse anatomy of reproductive organs in animals in the genus Hofstenia (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Hookabe et al., 2024; Steinböck, 1966). However, the fine structure of these organs, how they grow and develop, and how they function in the course of reproductive behavior remains unknown. To address these questions about the reproductive life history of H. miamia, we use a combination of light and confocal microscopy, fluorescence in situ hybridization, immunofluorescence, and histology, as well as observations of reproductive development and behavior in controlled conditions. We reveal new facets of acoel reproductive biology, show that active processes of growth and destruction ensure coordinated organ development and regeneration, and establish a foundation for the mechanistic study of reproduction in acoels.

Results

An overview of Hofstenia miamia’s reproductive anatomy

The genus Hofstenia (Bock, 1923) lies within the family Hofsteniidae. The family currently contains four genera (Ahyong et al., 2024; Hookabe et al., 2024) and is an early-branching lineage within acoels (Abalde and Jondelius, 2024; Jondelius et al., 2011). Three species are currently recognized within the genus Hofstenia: H. atroviridis, H. miamia, and H. arabiensis (Ahyong et al., 2024; Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007). Like other acoels, Hofstenia are considered to be simultaneous hermaphrodites (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966).. Hofstenia miamia is likely the most common and widespread species within the genus, with a geographic distribution in the Caribbean and the North Atlantic, including the Bahamas and the Florida coast (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966). All three Hofstenia species are described as having an anterior male reproductive system including a seminal vesicle, prostatic vesicle, a penis with sclerotized needles, ‘diffuse’ testes, and a ventral male gonopore near the mouth (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966). The female reproductive system, as described, includes two ovaries spanning the posterior two-thirds of the worm. However, these descriptions were all based on worms roughly 5-8mm in length (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966). In our lab cultures, fully grown adult H. miamia can be over 1.4cm in length. We noticed that these large adults were more likely to reproduce than smaller worms, and we reasoned that their anatomy could also be different. We therefore began by re-examining the coarse anatomy of Hofstenia miamia.

We anesthetized adult worms, mounted them ventral side up on a slide, and imaged them through a dissecting microscope and found, as expected, that H. miamia’s male copulatory structures are located in the anterior of the animal, just posterior to the mouth (Fig. 1C,D; Fig. S1A). These male structures are located within a translucent cylindrical region close to the ventral surface of the worm and immediately ventral to the pharynx. We observed two opaque, white regions within the male copulatory apparatus, located towards its posterior (Fig. 1C; Fig. S1A). The larger, posterior oval structure corresponds to the seminal vesicle, while the smaller, teardrop-shaped anterior structure corresponds to the prostatic vesicle (Fig. 1C,D; Fig. S1A) (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966). Previous work reported that H. miamia has a penis (Fig. S1A): a sperm-delivery organ containing rigid needles known as ‘stylets’ (Hooge et al., 2007; Steinböck, 1966). Although stylets are difficult to see in this imaging preparation, we identified a small aperture immediately posterior and ventral to the mouth (Fig. S1A), and we observed that the worm extended its penis through it, showing that this aperture is the male gonopore and confirming the location of the penis in the anterior of the worm.

As previously described (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966), we found that the female reproductive system occupies the posterior two-thirds of the worm (Fig. 1C,D; Fig. S1B). This system consists of numerous eggs grouped into three distinct clusters: one lateral group on each side of the body extends along the anterior-posterior axis (Fig. S1B) and a medially-located cluster present just posterior to the pharynx (Fig. S1C). The lateral clusters contain oocytes of various sizes, likely corresponding to different stages of maturation. These oocyte-laden regions in the parenchyma have been identified as the ovaries in historical reports (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966). Within the ovaries, there is no obvious organization of oocytes (by size or shape) along any of the worm’s body axes. All oocytes within the ovaries have a germinal vesicle, a nucleus arrested in prophase I of meiosis that is a common feature of oocytes of nearly all animals (Grossman et al., 2017; Herlands and Maul, 1994; Liu et al., 2006; Munro et al., 2023). This nucleus breaks down during early oogenesis or after fertilization (Falleni and Gremigni, 1990; Sagata, 1996), suggesting that Hofstenia’s lateral oocytes are likely unfertilized.

Posterior to the pharynx, there is often a cluster of uniformly sized, spherical eggs that do not have a germinal vesicle (Fig. 1C,D; Fig. S1C). The number of eggs at this position is highly variable, ranging from zero to over a dozen eggs. In H. atroviridis, Bock (Bock, 1923) speculated that these eggs are mature, fertilized, and ready to be laid: i.e., zygotes. To test whether these eggs were indeed zygotes, we dissected out eggs from this midline cluster, as well as from the lateral ovaries. We found that 46/56 eggs (removed from the midline clusters of 9 worms) developed and hatched into juvenile worms, in line with previously established developmental timing (Fig. S1D) (Kimura et al., 2021). Meanwhile, all eggs dissected from the lateral ovaries disintegrated within 48 hours. The eggs in this medial cluster are therefore mature, likely fertilized, and ready for laying, while eggs in the ovaries are immature or unfertilized. Overall, these observations confirm and extend previous descriptions of H. miamia, and provide a high-level overview of its reproductive anatomy. Next, taking advantage of H. miamia’s tractability, we sought to obtain a finer understanding of its reproductive structures and their development.

Dynamics of reproductive organ development

How reproductive structures develop as animals reach sexual maturity is not well understood in the Xenacoelomorpha. As in the acoel Aphanostoma pulchra (previously Isodiametra pulchra) (Chiodin et al., 2013), hatchling H. miamia have no visible reproductive structures, and previous observations suggested that these develop over a month after hatching as the worms grow toward adulthood (Kimura et al., 2021; Srivastava, 2022). To determine the timeline and dynamics of reproductive development, we reared 42 zygotes in isolation and monitored their growth over time (Fig. 1E,F), precisely controlling their rearing environment such that each animal had access to controlled amounts of seawater and food (see Methods). To assess whether nutrition affects reproductive development, we split the cohort of hatchling worms into two groups that were fed differently. For each group, we gradually increased the amount of food they were given as they grew larger, to ensure that worms always had ad libitum food access while avoiding unnecessary overfeeding. These food increases occurred periodically, but we delayed them by two weeks for one feeding group (Fig. S2A,B). We reasoned that this treatment would delay growth in one group, allowing us to decouple biological age from organ and body growth. Twice weekly, we captured images of the ventral surface of each worm and quantified the size of each visible reproductive structure.

We found that worms maintained their aspect ratio from hatchling to adult, growing proportionally over time (Fig. 1E-H; Fig. S2C). At every stage of development, the size of each reproductive organ scaled with body size once that organ appeared (Fig. 1I-K). Worms with delayed increases in feeding displayed corresponding delays in their growth, both in their body size (Fig. S2D) and in the timing of appearance of their ovary and seminal vesicle (Fig. 1L). The penis and seminal vesicle developed at a smaller size in these worms compared to worms without delayed feeding increases (Fig. 1M), suggesting that resource-limited worms may invest disproportionately in male development. Multiple regression found that organ size was explained by both age and body size, but that body size was a better predictor of the size of each reproductive organ (Table S2). This suggests that the differentiation and growth of reproductive structures is coupled to the worm’s body size, rather than to its age.

We found that reproductive development generally occurs in a stepwise fashion, where organs appear in a consistent sequence at serially increasing body sizes. Newly hatched animals possess all major somatic organs (such as a brain, muscle system, pharynx, gut, etc. (Kimura et al., 2022, 2021)), but lack all reproductive structures (Fig. 1E,F). 4-6 weeks after egg laying, when the worm is about 3 ± 0.6 mm (mean ± SD) in length, the anterior-ventral tissue housing the future male copulatory apparatus becomes visibly less dense, and a groove becomes apparent in the ventral surface. Soon after, the male gonopore becomes detectable posterior to the mouth. Sperm begin to accumulate in the seminal vesicle around 35 days post laying, once a worm is about 4.4 ± 0.9 mm (mean ± SD) long. Shortly after, when a worm is 5.6 ± 0.9 mm (mean ± SD), the first oocytes become visible in nascent ovaries. These ovaries grow with the worm over time, with existing oocytes maturing and new ones continuously added. Around two months after being laid, typically after a worm is 8.7 ± 1.2 mm in length, eggs appear near the base of the pharynx (Fig. 1E,F,L,M). These eggs are laid within hours or days, and they develop and hatch into juvenile worms. Given that these worms were isolated since birth, this suggests that H. miamia can reproduce through self-fertilization.

To visualize the extent to which reproductive structures develop in sequence, we ranked each structure in each worm by its date of appearance (Fig. 1N). All worms, without exception, first developed a penis. A few worms simultaneously developed a seminal vesicle; in most other worms, this structure appeared second, with a few developing their seminal vesicle concurrent with nascent ovaries. Ovaries always appeared before the first fertilized eggs were seen in the central region. Next, we asked whether the left and right ovaries of each worm developed in synchrony. We found that on the first day that an ovary was visible, 34/34 worms had both ovaries visible. Partial regression analysis revealed that the sizes of the two ovaries within a worm are strongly correlated, even when correcting for body size (Fig. S2E). Consistent with findings from many other bilaterally-symmetric organs (Allard and Tabin, 2009; Boulan and Léopold, 2021; Harris et al., 2021; Vallejo et al., 2015; Wolpert, 2010), this suggests that an active process may synchronize ovarian growth within individual worms.

De-growth and regeneration of reproductive organs

Next, we asked whether these patterns of growth persisted in different contexts. H. miamia, like other studied acoels, is an excellent regenerator and can regenerate all tissues from a wide range of initial tissue configurations (Srivastava, 2022; Srivastava et al., 2014; Steinböck, 1966). We amputated and followed adult worms in three ways: by isolating tail tips (which contain no reproductive structures) and studying how they regenerated the majority of their organs, by cutting worms sagittally and studying how they regenerated their missing half, and by isolating heads from tails and studying how these head fragments regenerated their tails and the posterior region of their ovaries (Fig. 2A). Head and tail fragments both gradually increased in size, with tail fragments growing much faster in size (Fig. S3A-G). Within two weeks, most of these fragments had regained the characteristic shape of intact worms. Sagittally-cut fragments first shrank in size before subsequently growing (Fig. S3G), and appeared to take longer to fully regain normal worm-like appearance (Fig. S3C,D). All reproductive organs gradually regenerated, with growth dynamics contingent on the nature of the injury (Fig. 2B-J; Fig. S3A-F,H,I). Additionally, we asked how reproductive organs change in another context: degrowth. Anecdotal observations suggested H. miamia tolerates long periods of starvation but gradually shrinks in size when deprived of food. We used this starvation-induced de-growth to ask whether reproductive organs scale with decreasing body size. We starved a cohort of adult worms and quantified the de-growth of their reproductive organs over time (Fig. 2K,L). We found that all worms survived over three months of continuous starvation, gradually shrinking over time (Fig. 2M). Worms maintained their aspect ratios as they shrank (Fig. 2K,L,N), and their reproductive organs shrank correspondingly (Fig. S4A-E). Partial regression analysis of shrinking ovaries showed that they shrink synchronously (Fig. S4F), suggesting active coordination of de-growth across the left-right axis.

Figure 2:

The stepwise growth relationships we found between reproductive organs and body size during post-embryonic development largely generalized to both regeneration and de-growth (Fig. S4G-I). Starving worms lost most reproductive organs in the opposite order to which they developed them, first losing fertilized eggs, then ovaries, and eventually losing their seminal vesicle (Fig. 2O,P). However, after three months of starvation, most worms continued to retain their penes despite being below the threshold size at which we would expect to see loss of this organ (Fig. 2O), perhaps because rigid, sclerotized structures degrade differently from soft tissue. In regenerating tail fragments, the male reproductive structures generally developed before the female reproductive structures, following similar stepwise patterns to developing juvenile worms (Fig. 2Q). Scaling coefficients for each organ were statistically distinguishable across development, regeneration, and de-growth, but the effect sizes of the differences between coefficients were generally small (Fig. S4G-I, Table S3). Analyzing this effect in more detail, we found that unsurprisingly, amputated worms initially display aberrant organ scaling but recover typical scaling as they regenerate (Fig. S4J-L). Qualitatively, across all of these growth contexts, reproductive organs scaled with body size in similar ways (Fig. S4G-I, Table S3).

In sagittally-cut worms, each worm fragment retained one of its two ovaries immediately after amputation. These ovaries first partially degenerated before growing back (Fig. 2G; Fig. S3C,D), consistent with the reported loss of germline tissue during early regeneration in heads regrowing tails (Hulett et al., 2023). Body size also decreased after sagittal amputation (Fig. S3G), but ovarian degrowth was disproportionate (Fig. S4M). The rate of ovarian degrowth (normalized to body size) was significantly greater than the rate of normalized ovarian degrowth during starvation (Fig. S4M). Consistent with studies showing extensive cell death during early regeneration in other systems (Ballarin et al., 2008; Pellettieri et al., 2010; Rychel and Swalla, 2008), this suggests that an active destructive process is responsible for degrowth during ovary regeneration in acoels. The missing ovary began to grow back roughly one month after amputation. We found that the ovaries grew asymmetrically in these worms, with the new ovary growing faster (Fig. 2G). This suggests the existence of an active growth mechanism to ensure that both ovaries reach a symmetric target size.

Together, our results suggest the existence of a size-associated program that regulates the development of reproductive organs, as well as active tissue growth and destruction mechanisms to achieve organ symmetry. It is likely that this program regulates reproductive organ growth in a variety of different developmental contexts, including the transition from juvenile to adult worm, regeneration after injury and tissue loss, and during starvation-induced de-growth. With this high-level understanding of male and female reproductive system anatomy and the scaling relationships that govern their formation in hand, we next sought to understand these systems at higher resolution in order to decipher their functional morphology.

Fine structure of the male reproductive system

Next, to understand the organization of male reproductive structures at high resolution, we visualized them using a combination of histology, immunofluorescence, staining with live dyes, and fluorescence in situ hybridization (FISH) (Fig. 3; Fig. S5A). The most anterior part of the male copulatory apparatus is the penis. This structure is difficult to visualize under reflected white illumination (Fig. 1C; Fig. S1A), but can be seen when worms extend it, which we have observed them do both spontaneously and during mating (data not shown). Staining fixed worms with the dye SiR-actin labels the structure clearly, and can be visualized well in smaller adult worms more amenable to confocal microscopy, enabling a high-resolution view of the structure (Fig. 3A-C). Components of the penis can also be seen clearly in histological sections of adult worms (Fig. 3D). These show that the penis contains a bundle of rigid needle-like structures (referred to as stylets) (Fig. 3E-G, Fig. S5B,C). Our observations of the stylets match the description of sclerotized needles described in other species in the genus and in other studies of Hofstenia miamia (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966). The stylets are situated immediately posterior to the base of the penis sheath – a long, conical structure (Fig. 3C; Fig. S5E,F; Video 1). The cavity inside this sheath has previously been referred to as the ‘male antrum’ (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966). The base of the sheath is cup-shaped, and is lined with posterior-facing hair-like extensions (Fig. 3G). At its anterior end, it connects to the male gonopore (Fig. 3C). Both the gonopore and the sheath are lined with cilia (Video 1,2), consistent with reports in other hofsteniid species (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966). During mating, the penis stylets are likely pushed anteriorly, through the sheath and out of the male gonopore. This could evert the sheath, resulting in a penis-extension state in which the sheath emerges through the gonopore around the stylets, with the ring of protrusions now near the anterior tip of the penis. It is possible that these protrusions have a sensory function, and could inform fine penis movements during mating. Immunofluorescence with an antibody against FMRFamide (see Methods) revealed a set of cells that resemble neurons encircling the base of the sheath (Fig. S5E). Further studies will be needed to understand how these components regulate copulatory behavior.

Figure 3:

Next, we focused on the teardrop-shaped prostatic vesicle, which is situated anterior to the seminal vesicle (Fig. 1C,D) and immediately posterior to the penis stylets (Fig. 3H,I). Our histological studies show that the prostatic vesicle (Fig. S6C-E) is surrounded by glands (Fig. S6F-H). The secretions of this organ are thought to mix with the sperm as it passes through (Faubel, 1983; Hyman, 1951). Unexpectedly, our immunofluorescence experiments also revealed a layer of cells enclosing the prostatic vesicle (Fig. 3I). This layer may be an epithelial tissue most likely associated with the prostatic glands. The interior of the prostatic vesicle contains densely-packed sperm cells (Fig. 3D,I; Fig. S6C-G).

The seminal vesicle is the posterior-most organ of the male copulatory apparatus (Fig. 3J). Histological staining as well as immunofluorescence staining of the muscle marker Tropomyosin, showed that the seminal vesicle and prostatic vesicle are both encircled by layers of muscle (Fig. S5F; Fig. S6C-H), consistent with previous descriptions (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Steinböck, 1966). Histological and nuclear staining of the seminal vesicle in adult worms (Fig. 3J-L; Fig. S5F; Fig. S6G,H) reveal that it consists primarily of sperm cells. To confirm this, we dissected the seminal vesicle out of adult worms, and stained it with a nuclear dye. We found that individual cells emerged from the structure over time, and high-magnification imaging showed that they were indeed sperm cells, each with a ∼23μm-long nucleus and a ∼22μm-long flagellum (Fig. 3M). During one mating event, we observed a failed attempt at insemination by a worm that resulted in the release of a packet of sperm outside its partner. We recovered this 0.6 mm-long packet of sperm (Fig. S5G) and found that it consisted of cells with the same morphology as those in the seminal vesicle (Fig. S5H), confirming that these are indeed sperm. As an aside, acoel sperm has historically been described as “biflagellate”, intended to mean that sperm cells each contain two axonemes, and this property is thought to be a defining synapomorphic character of acoels (Achatz et al., 2013; Petrov et al., 2004). We show that H. miamia sperm have a single flagellum, and suggest that the axonemal properties of the flagellum may be a more accurate taxonomic character than the external flagellar morphology.

Next, we sought to identify the testes (the organs of sperm production). Consistent with previous observations (Beltagi and Mandura, 1991; Bock, 1923; Corrêa, 1960; Hooge et al., 2007; Steinböck, 1966), nuclear staining and histological studies of adult worms revealed dense clusters of sperm in the anterior parenchyma, wrapping around the worm’s pharynx (Fig. 3N-U). Tropomyosin labeling showed that the testes are embedded within layers of muscle (Fig. S5I,J). Next, inspecting germline markers identified in previous single-cell RNA sequencing data (Hulett et al., 2023), we identified one likely to label testes: the gene pa1b3-2. FISH for pa1b3-2 reveals that this gene does indeed specifically label testes, and that its expression coincides well with the dense cellular clusters visible through nuclear staining (Fig. S7A,B).

Finally, we studied the development of the male reproductive system, using confocal microscopy to visualize the maturation of its components at high resolution. Hatchling worms do not possess any visible reproductive structures. Actin staining reveals that, within the copulatory apparatus, the sheath and stylets emerge first, followed by a gradually-inflating prostatic vesicle (Fig. 4A). The hair-like extensions of the sheath grow and become more conspicuous over time (Fig. 4A). Next, we used two methods to visualize testes maturation. First, reasoning that nuclear staining may not allow us to visualize testes at the earliest developmental stages, we performed FISH to detect mRNA for pa1b3-2 in juvenile worms of various sizes. In the smallest worms (i.e. ∼<2 mm), there was no expression (Fig. 4B-C). In larger, reproductively immature worms, we detected expression of pa1b3-2 in two lateral clusters that expanded slightly along the anterior-posterior axis as development progressed (Fig. 4B-C). Since adult testes are a single cylindrical structure that wraps around the pharynx, this early developmental pattern suggests that the two lateral structures must grow along the dorso-ventral and medio-lateral axes until they meet in the middle. Second, to test this hypothesis, we performed nuclear staining on transverse sections of worms across a broader section of development (Fig. 4D; Fig. S7C). In small worms, the testes are lateralized and do not meet on the dorsal surface. In larger worms, these lateral regions have expanded across the dorsal surface of the worm to form a single apparently-continuous region (Fig. 4E). Many acoel species are thought to have paired, lateral testes (Table S1), a feature of taxonomic importance (Jondelius et al., 2011). While juvenile H. miamia have paired testes, these organs morph into a single cylindrical structure in the transition to adulthood. As lab-reared worms can be substantially larger than those found in the wild, these results raise the possibility that other acoels may undergo similar morphological changes during development.

Figure 4:

Together, our data provide a high-resolution view of H. miamia’s elaborate male reproductive morphology. Consistent with our quantitative data, we find that this morphology emerges in a stereotyped developmental sequence during the transition to adulthood in which the penis sheath develops first, followed by nascent testes, the stylets, and the membrane-like structure enveloping the prostatic vesicle. Sperm presumably from the testes then travels to the prostatic vesicle. Over time, as all structures grow, the hair-like projections on the back of the sheath become more prominent, the testes extend both along the anterior-posterior axis and dorsally to wrap around the head, and sperm cells fill the seminal vesicle. The prostatic and seminal vesicles appear to be surrounded by muscle, epithelia, and gland cells. Our histological sections suggest that at least in adult worms, sperm cells may enter through the posterior of the seminal vesicle and travel forwards into the prostatic vesicle. We do not yet understand how this occurs in juvenile worms, how sperm cells navigate to these vesicles from their varied points of origin, or how they are hypodermically injected during mating.

Fine structure of the female reproductive system

As observed through imaging of adult worms using a stereo microscope, the female reproductive system in H. miamia consists of two lateral ovaries running longitudinally along both sides of the body, and one medial cluster of mature eggs located posterior to the pharynx (Fig. 5A-C). To visualize the structure of the ovaries, we used FISH to label mRNA of a previously-identified germline marker: cgnl1-2 (Hulett et al., 2023). We found that this gene specifically labeled the oocytes (Fig. 5D). Moreover, immunostaining with a custom antibody against Piwi-1, a stem cell and germline marker, also labeled the ovaries (Fig. 5E). We were also able to visualize the ovaries in histological sections (Fig. 5C). These methods, together with our earlier images of ovaries at different worm ages, showed that oocytes were not organized by maturity within the ovary (Fig. S8A). We did not detect a membrane or lining that envelopes the ovaries. This is consistent with previous work that suggests that acoels lack true ovaries sensu stricto: oocytes appear and develop within the parenchymal tissue without a specialized membrane that forms a discrete organ (Eckelbarger and Hodgson, 2021; Rieger et al., 1991; Schmidt-Rhaesa, 2007).

Figure 5:

Oocytes in Hofstenia are surrounded by follicular cells, cells lining developing oocytes that provide nutrition and secrete the eggshell (Bock, 1923; Rieger et al., 1991). We visualized this cell layer with immunofluorescence for Piwi and histology and found that it was consistently present around every oocyte in the ovaries irrespective of their developmental stage (Fig. 5F-G; Fig. S8B-D). We observed that many follicle cells have very large nuclei, perhaps the result of the fusion of neighboring cells (Fig. 5G; Fig. S8B-C). We also found clusters of sperm in the layer of follicular cells surrounding maturing oocytes of a variety of developmental stages (Fig. 5F,G, insets), consistent with Bock’s (Bock, 1923) proposal that the follicle may ‘trap’ sperm and control the fertilization of maturing oocytes. We observed significant morphological variation in the cellular components of the oocytes within the ovaries. Early-stage oocytes appeared small, round or oval, although others showed irregular shapes. All oocytes contained a granular substance primarily distributed peripherally and a germinal vesicle occupying a substantial portion of the oocyte (Fig. 5H-J). The presence of a germinal vesicle in the oocytes in H. miamia’s ovaries suggests that these oocytes are unfertilized. It is likely that oocytes mature in the ovaries, the germinal vesicle breaks down, and then oocytes are fertilized. These mature, fertilized eggs then travel from the ovary to the central cavity prior to being laid.

We could not determine conclusively where or when fertilization happens, or where the egg capsule (chorion) is produced. However, the spatial organization of the female reproductive system suggests that an oocyte must mature, become fertilized, and then be transported to the central cavity behind the pharynx. Across our internal examination, we did not identify any oviducts, canals or additional structures that could facilitate this migration from the ovaries to the central cluster. How eggs are transported thus remains unknown.

Egg-laying behavior

Egg laying has only been directly observed a few times in acoels, mostly in species within the Convolutidae (Costello and Costello, 1939; Gardiner, 1898). From these observations, acoels are known to lay eggs through the female gonopore (if present) (Gardiner, 1898), a mode of egg laying used by many flatworms (Tong and Ong, 2020), or through breaks in the body wall (Apelt, 1969; Costello and Costello, 1939). It has also been suggested that acoels could lay eggs through the mouth; however, direct observation of such egg laying has been challenging (Apelt, 1969; Costello and Costello, 1939; Watzin, 1984).

Previous authors have speculated that H. miamia lay eggs through breaks in the body wall, or perhaps partially through the mouth, but egg laying was not observed (Bock, 1923; Steinböck, 1966). To observe how worms lay eggs, we isolated gravid adult worms and filmed them from underneath for a 24 hour period. 3 worms laid a total of 25 eggs in these conditions. We found that worms exclusively laid eggs through their mouths in events typically lasting less than 2 minutes (Fig. 6A,B; Video 3; Fig. S9A). During this event, the worm performs a series of muscle contractions to transfer a single egg from the ventral pocket of fertilized eggs into the pharynx, and then applies further muscle contractions from posterior to anterior to move the egg to the mouth. The worm then places each egg on the substrate with its mouth, likely secreting mucus to attach the egg to the substrate (Fig. 6A,B; Video 3). Observations from these videos, as well as of egg clutches in our culturing tanks, show that worms can either deposit a single egg in one location or lay eggs in one or more clutches.

Figure 6:

To understand the temporal dynamics of egg-laying in H. miamia, we then quantified the time course of egg-laying in several contexts. We found that adult worms with previous access to mates, once isolated, continued laying eggs for over a month (Fig. 6C). Juvenile worms reared to adulthood in isolation also laid a single burst of eggs (Fig. 6D), multiple months after first being isolated (although the timing of this burst may depend on rearing conditions and the health of the worms). Consistent with our finding that worms reared entirely in isolation can lay eggs (Fig. 1L,M), this burst of egg laying is also likely the result of a selfing event in which the worms fertilize their own eggs, or possibly the result of a form of parthenogenesis allowing the activation of unfertilized eggs. We also allowed virgin worms to mate once, in controlled conditions, and subsequently isolated them. These mated and inseminated worms laid eggs for over three months after a single mating, suggesting that received sperm may be stored in follicular cells for several months (Fig. 6E).

Next, we asked how much control the worms had over their egg laying. To test whether worms choose specific locations to lay eggs, we quantified the spatial positions of eggs laid in their culturing tanks. We found that worms have a strong preference for laying eggs on the walls of their containers, rather than on the floor (Fig. 6F). Eggs laid on walls are preferentially laid close to the water line (Fig. 6G, Fig. S9B). This preference is not simply because the worms lay eggs where they are: we observed that worms spend the majority of their time on the floors of their containers, and seem to glide up the walls specifically to lay eggs (Fig. S9C). These spatial preferences do not seem affected by food availability: when deprived of food, egg laying still primarily occurs on the walls of culture boxes (Fig. S9D). These data suggest that worms make active substrate choices for egg laying.

We also observed that worms lay 49% of their eggs in clutches. Individual worms can produce these clutches (Video 3). Clutch sizes are often small (Fig. S9E), but the largest can contain over 30 eggs deposited within a 3-4 day window in culture boxes containing 20-50 worms (Fig. 6H, Fig. S9E). The vast majority of individual worms lay fewer than 10 eggs in a 3-4 day period (Fig. S9F), suggesting that some clutches may be communal. Indeed, in other culturing conditions with large groups of worms, clutch size often exceeds 140 eggs (Fig. 6I). To test whether worms lay eggs in communal clutches, we allowed worms to lay eggs in containers for 3 days. We then swapped worms between containers (or, for control containers, we swapped worms and removed old eggs). We asked how worms interacted with egg clutches laid in the first 3-4 days, and found that many new eggs were laid in pre-existing clutches (Fig. 6J, Fig. S9G). Worms added new eggs to 42 percent of old egg clutches. This shows that worms frequently lay eggs in communal egg clutches.

Finally, we asked whether worms withhold eggs in suboptimal environments. Our observations (see Fig. 2K-L) suggested that worms are physiologically resilient to food stress. We therefore conducted an experiment with a set of worms, half of which were deprived of food for one feeding period (3-4 days). The other set was fed normally during this period. We found that food-deprived worms laid significantly fewer eggs during this period (Fig. 6K). We hypothesized that this reduction in egg-laying was because the worms were actively withholding their eggs in the absence of food. Next, we repeated this experiment, first depriving some worms of food and subsequently feeding them. After food-deprived worms were given food, they laid an excess of eggs (Fig. 6L), showing that the worms withhold egg-laying in food-limited environments. More generally, these data show that H. miamia assess their environments to decide when and where to lay eggs.

Discussion

Our work describes the reproductive life history of an acoel across its life cycle (Fig. 7A). We reveal many new facets of acoel biology, describe the structure and dynamics of growth, regeneration, and egg-laying behavior, and establish a foundation for the experimental study of reproductive biology in acoels. Below, we discuss the significance of our findings for reproductive development, behavior, and the evolution of life history strategies.

Figure 7:

Reproductive development and regeneration

Little is known about the development or regeneration of reproductive structures in acoels. Our data show that in H. miamia, reproductive structures develop and regenerate in a stereotyped sequence, with male organs appearing before female ones. During starvation-induced de-growth, these organs degenerate in the opposite sequence. This stereotypy is unexpected and may not be a universal feature of acoels: sparse data from three other acoel species, Solenofilomorpha funilis (Crezée, 1975), Otocelis luteola (Kozloff, 2000a) and Aphanostoma pulchra (Perea-Atienza et al., 2013), indicate that these species do not exhibit identical sequences. The stereotypy we observe is consistent with the idea that a single, size-associated program regulates reproductive organ development in H. miamia, and that it may be deployed in a variety of growth contexts.

Our results also show that H. miamia is capable of regenerating all reproductive structures from a variety of initial tissue configurations. The trajectories toward organ replacement vary based on this initial state, and reveal important features of the regenerative process. For instance, we found that worms missing their sagittal halves or their tails first shrink in size before re-growing. This may be in part because the early stages of wound closure and regeneration require the reorganization of existing tissues. In any case, this reduction in body size is associated with disproportionate reductions in reproductive organ size, demonstrating active tissue destruction mechanisms, possibly mediated by apoptotic processes similar to those reported in other regenerative species (Ballarin et al., 2008; Pellettieri et al., 2010; Rychel and Swalla, 2008). In addition, we found that the two ovaries within a worm grow in a significantly correlated manner, and after sagittal amputation, the missing ovary grows disproportionately to regenerate symmetry. In principle, such bilateral symmetry could be achieved passively, without any feedback (Reddien, 2018). However, the strong within-worm correlation, even after controlling for body size, suggests that there is either organism-specific coordination of growth or death rates, or more likely a feedback process that ensures bilateral symmetry.

In many animal lineages such as arthropods, vertebrates, planarians, and nematodes, organs typically grow symmetrically, and scale with body size (Boulan and Léopold, 2021; Ko et al., 2024; Uppaluri and Brangwynne, 2015; Wolpert, 2010). The mechanisms underlying this coordinated, scaled growth are not fully understood, and may vary across tissues and species. For instance, some vertebrate tissues appear to grow using cell-intrinsic programs; others - and many insect tissues - appear to rely on central coordinating mechanisms that read signals of growth from the circulation or from unknown sources (Allard and Tabin, 2009; Boulan and Léopold, 2021; Harris et al., 2021; Trible and Kronauer, 2017; Vallejo et al., 2015; Wolpert, 2010). How different organs ensure appropriate, coordinated scaling remains poorly understood. We find that H. miamia must also solve this problem: their reproductive organs scale with body size during development, regeneration, and de-growth, and an active coordination mechanism ensures that the ovaries return to symmetry after perturbations. How do tissues across the body ‘read’ the same indicators of size, and how is their growth coordinated? Our work establishes a foundation for the experimental study of these and other developmental phenomena.

Egg-laying physiology and behavior

Despite speculation, few acoels have been directly observed laying eggs (Costello and Costello, 1939; Gardiner, 1895; Watzin, 1984). We find that H. miamia lays fertilized eggs through its mouth: a mode of egg-laying not found in other animals. This unusual behavior raises multiple questions. First, where in the animal does fertilization occur? Our observations suggest that follicular cells surrounding immature oocytes may act as a selective barrier to sperm, preventing many sperm cells from reaching the oocyte, a function suggested by Bock (1923). Additionally, given that H. miamia can lay eggs for months after a single mating despite lacking a seminal bursa, it is likely that the follicular cells function as a sperm storage organ. Second, we find that H. miamia appears capable of self-fertilization, as it can lay eggs without mating. While some acoels are capable of asexual reproduction through fissioning and budding (Åkesson et al., 2001; Ax and Schulz, 1959; du Bois-Reymond Marcus, 1955; Hanson, 1960; Ishikawa and Yamasu, 1992; Zabotin and Evtugyn, 2021), self-fertilization has not been reported previously. How does sperm travel to the ovaries? It is likely that sperm cells migrate posteriorly from the testes towards the ovaries in isolated worms, and from arbitrary body regions towards the ovaries after mating. In addition, sperm must travel from all regions of the testes to the seminal and prostatic vesicles. Whether they use chemical cues to navigate towards oocytes and these vesicles, and the mechanics of this process, remain unknown. Third, how do mature eggs travel from the lateral ovaries to the central cavity, and how are the eggs in this cavity physically loaded into the pharynx? Our anatomical observations have so far failed to reveal obvious oviducts, or other tube-like structures that could facilitate movement of eggs into the pharynx. The only known opening at the posterior of the pharynx is the pharyngeal sphincter that allows food to pass into the gut. We therefore speculate that eggs travel through the gut to the central cavity, which may simply be a pocket within the gut. Eggs may then be moved into the pharynx through a form of reverse peristalsis (and observation of the worm prior to egg-laying shows waves of muscle contraction proceeding posterior-to-anterior as each egg is loaded into the pharynx).

We also find that H. miamia appears to make active choices about when and where to lay fertilized eggs. H. miamia has clear spatial preferences for egg-laying in our culture chambers. Worms often lay eggs in communal clutches, sometimes adding to another worm’s clutch. They also have environmental preferences, withholding egg-laying when deprived of food and then quickly laying eggs once food is present again. Together, these results show that H. miamia surveys its environment and then integrates this assessment into whether - and where - it is suitable to lay an egg. More generally, the existence of communal egg clutches raises the possibility that acoels may have a rich and unexplored social repertoire.

Evolution of reproductive strategies within acoels

Placed in a comparative context, our description of H. miamia’s reproductive life history reveals that acoels and their sister lineage, the nemertodermatids, display a striking diversity of reproductive morphologies, and likely a corresponding diversity in reproductive behavior (Fig. 7B). This diversity reveals correlated suites of reproductive traits that suggest a small number of life history strategies. For instance, the presence of a bursa and female gonopore are associated with a muscular penis, suggestive of cooperative, genital-handshake style mating. The absence of a bursa and female gonopore are associated with a needle-laden penis, suggestive of competitive, hypodermic-insemination style mating. Comparative work in the distantly-related flatworms shows that these anatomical and behavioral traits indeed coevolve similarly (Brand et al., 2022a, 2022b), in accordance with social evolutionary theory on sexual conflict (Charnov, 1979; Michiels and Newman, 1998; Ramm et al., 2015). Within the genus Macrostomum, a competitive hypodermic insemination syndrome has evolved at least 14 times from ancestral cooperative, reciprocal insemination (Brand et al., 2022a, 2022b). This syndrome involves the correlated evolution of a sharpened penis, a simplified female sperm-receiving organ, sperm cells lacking bristles and other adaptations for post-copulatory sexual conflict, and associated behavioral changes (Brand et al., 2022a, 2022b, 2022c; Schärer et al., 2014, 2011). Whether similar anatomical associations truly predict mating strategies in acoels remains unknown. In addition, the variability of some morphological features does not obviously fit this pattern of cooperative vs competitive reproductive strategies. For example, the relative locations of the testes and ovaries is highly variable, as is the number of bursal nozzles, suggesting that there may be further evolutionary patterns awaiting explanation. Our work establishes an approach to study reproductive anatomy and behavior in a model acoel. This approach can be applied to many other acoel species. Given their rich diversity, acoels are a promising clade in which to study the evolution of reproductive strategies, and in which to test the generality of theories of sexual conflict.

Conclusion

These findings establish foundational knowledge of anatomical, physiological, and behavioral elements of H. miamia’s reproductive life history. This enables future work on the molecular genetics of reproductive organ development and regeneration, on the physiological processes involved in egg maturation and oviposition, and on the neuroscience of reproductive behavior.

Methods

Animal husbandry

We reared gravid H. miamia in communal plastic 2.25 L boxes with approximately 1.25 L of artificial seawater (37 ppt, pH 7.8-8.2). These communal tanks were kept in incubators held at 21°C. Twice weekly, we collected embryos, changed their seawater, and fed them with Artemia sp. (brine shrimp). The embryos were raised in petri dishes. Once they hatched, we transferred juvenile worms to tanks maintained at room temperature and fed them with marine rotifers (Brachionus plicatilis). When reproductive organs began to develop, we transferred worms to larger tanks, each typically housing 30-50 worms. To pair animals for mating, we transferred juvenile worms from tanks to 24-well plates and reared them in isolation following similar feeding and cleaning routines as listed above. In order to study the regeneration of reproductive systems, we isolated adult worms from communal tanks, anesthetized them in 15% tricaine (ethyl 3-aminobenzoate methanesulfonic acid), and then amputated them with micro knives (Fine Science Tools #10316-14). We maintained regenerating worm fragments in 6-well plates and cleaned plates twice weekly. Five days after amputation, we fed these worms Artemia sp. (brine shrimp) or rotifers. We also maintained starving animals in 6-well plates that we cleaned twice weekly but that were not fed. All animals are derived from an inbreeding population of worms collected in 2010 from Bermuda.

Histology

We fixed adult specimens of H. miamia in 4% paraformaldehyde in artificial seawater for 24 hours at room temperature. Following fixation, specimens were washed twice with 70% ethanol to remove the fixative and then preserved in a third change of 70% ethanol for histological preparation. Animals were dehydrated through a graded ethanol series (95% and 100%), cleared in Histoclear, and embedded in Leica Surgipath (Paraplast) with a melting point of 56°C. We made longitudinal, transverse, and sagittal sections at a thickness of 5 μm using a Leitz 1512 microtome. We stained these sections using a standard protocol with hematoxylin and eosin, mounted them on glass slides with Permount, and imaged them using a Zeiss Axio Scan.Z1, and an Olympus BX50 microscope.

Fluorescence in situ hybridization (FISH)

We synthesized riboprobes following protocols established in (Srivastava et al., 2014). Following 1-2 weeks of starvation, we fixed whole worms in 4% paraformaldehyde in artificial sea water for one hour at room temperature. We then washed the fixed animals with PBST (PBS + 0.1% Triton-X-100) and transferred them to 24-well plates in small baskets with a mesh bottom, with 4-6 animals in each basket. To remove pigment autofluorescence, we treated the animals with a bleach solution (containing 4% hydrogen peroxide, SSC, and formamide) and left them under a light for two hours. For all washes, 800 μL of solution was used. We first permeabilized the animals using proteinase K solution (0.1% SDS, 1 μL/10 mL proK in PBSTx). After 10 minutes, we post-fixed the worms in 4% formaldehyde in PBST, washed twice in PBST, and then washed in a 1:1 PBST:PreHyb solution for 10 minutes. We then incubated the samples in PreHyb solution (50% DI formamide, 25% 20X SSC, 0.05% 10% Tween-20, 1mg/mL yeast tRNA, 20% water) for two hours in a 56°C hybridization oven and then transferred them to a hybridization solution (50% DI formamide, 25% 20X SSC, 0.05% 10% Tween-20, 1 mg/mL yeast tRNA, 10% 50% dextran sulfate, 10% water) containing riboprobe(s) overnight. We then put the specimens through several 30-minute washes at 56°C, starting with two PreHyb washes, two 1:1 PreHyb:2XSSCT washes, two 2XSSCT (2XSSC with 0.1% TritonX-100) solution washes, two 0.2XSSCT (0.2X SSC with 0.1% TritonX-100) solution washes. We cooled the specimens to room temperature and then washed them in two ten-minute PBST washes. We then performed a 1-hour blocking at room temperature with blocking solution (5% horse serum and 5% casein in PBST). We incubated specimens with anti-Digoxigenin-POD (1:1500 dilution; Roche, 11633716001) in a blocking buffer overnight. The following day, we washed animals 6 times with PBST, incubated with a tyramide buffer (1.1688g/mL NaCl, 6.18mg/mL boric acid, filled with water and adjusted pH to 8.5) for 10 minutes, then developed with rhodamine-conjugated tyramide solution for 10 minutes. We washed animals again with PBST and incubated for 1 hour in 1:500 PBST:Hoechst to label nuclei. We mounted the animals on glass slides with VECTASHIELD® PLUS Antifade Mounting Medium (Vector Laboratories, H-1900). We estimated worm length (Fig. 4A-D) by measuring copulatory apparatus length (gonopore to prostatic vesicle) or cross-sectional worm width within the image and comparing that measurement to measurements of developing worms.

Whole-mount immunofluorescence

We fixed worms in 4% paraformaldehyde (Electron Microscopy Sciences, 15714) in artificial sea water for 1-2 hours at room temperature (juveniles were fixed for one hour, adults were fixed for two) before washing with PBST (PBS + 0.1% Triton-X-100). We anesthetized adult worms in 0.5 mg/mL tricaine for 5 minutes prior to fixation to minimize epidermal rupture. For immunofluorescence, we washed worms in PBST, blocked for one hour at room temperature in 10% goat serum in PBST, and incubated them in primary antibody for 48-72 hours at 4°C on a shaker (juveniles for 48 hours, adults for 72 hours). The following day, we washed worms thoroughly in PBST (8 x 20 min washes) on a nutator before blocking for one hour at room temperature on a shaker. We incubated worms in secondary antibody overnight at 4°C; the following day, we washed them thoroughly in PBST (8 x 20 in washes) on a shaker before adding direct conjugate dyes and mounting them on glass slides with VECTASHIELD® HardSet™ Antifade Mounting Medium (Vector Laboratories, H-1400-10). We used the following antibodies: Tropomyosin (custom) (Hulett et al., 2020), Piwi-1 (custom), Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 568 (ThermoFisher, A-11011), Hoechst (ThermoFisher, H1399), SiR-actin (Cytoskeleton, CY-SC001). We acquired all images with FISH and immunofluorescence on a Leica SP8 point-scanning confocal microscope. We generated the Piwi-1 (Rabbit polyclonal) custom antibody using GenScript as previously described for Tropomyosin (Hulett et al., 2020). Briefly, a peptide was synthesized from the H. miamia Piwi-1 protein sequence, expressed in E. coli, and used to immunize rabbits. The Piwi-1 antibody used in this paper was #SC1195 (0.845 mg/mL), used at the concentration of 2 ug/mL.

Quantitative analysis of development, regeneration, and de-growth

We collected 42 zygotes laid from wild-type worms. Two embryos did not hatch, and two worms died during the course of the experiment (likely due to handling error). Four worms were removed from the experiment at different timepoints for additional imaging, resulting in 34 worms that we followed into adulthood (Fig. S2B). We transferred each embryo to a well of a 24-well cell culture plate with artificial seawater. Plates were stored in a temperature-controlled incubator. Without removing the embryos from the plate, we imaged them twice weekly through a dissection microscope with white illumination from LEDs mounted above the sample. We also changed water on these days. We added the same volume of water to each well (2.35 mL in each 12-well). Once at least one embryo had hatched, we added rotifers to each well. On each feeding day, we calculated the concentration of rotifers by counting the number of rotifers in diluted samples. Based on this concentration, we changed the volume of rotifers added to each well to ensure a consistent number of rotifers was given to each animal. At different stages during the experiment, we increased the number of rotifers (Fig. S2A). These increases were gradual to ensure worms always had more food than they could eat, while minimizing excess food which could affect water quality. We removed hatched worms from wells for imaging twice weekly. We anesthetized worms in tricaine, and mounted them on microscope slides with a 20x20 mm sticker grid with 1mm resolution (ThomasScientific, #1207X53). Depending on the worms’ pigmentation, this grid is visible through the body of some worms, but this did not impact the visibility of reproductive organs. When worms were larger, we moved them to larger 6-well plates in 10 mL of artificial sea water and began to feed them brine shrimp in addition to rotifers (Fig. S2A). We measured the number of brine shrimp added using the same concentration procedure outlined above for rotifers. Once worms had matured, we increased the number of brine shrimp.

We manually annotated images in FIJI. We measured lengths from the most posterior part of the male reproductive system (e.g., base of the penis tube before sperm production begins or base of the seminal vesicle) to the male gonopore, the length of the seminal vesicle, the length of each ovary, the length of the worm’s body, and the width of the worm’s body. We calculated penis length by subtracting the length of the seminal vesicle from the length of the male gonad. All measurements were converted to millimeters using image-specific scale information.

We repeated these imaging and measurement processing procedures with starving worms and regenerating worm fragments. We anesthetized and imaged starved and fed control worms roughly once a week for 11 weeks. 106 days after isolation, we imaged the starved worms again, at which point one starved worm had died. We imaged regenerating worms starting 2 days post-amputation and then again every 2-5 days until 25 days post-amputation at which point we took images less frequently. We did not feed regenerating fragments for the first four days after amputation, and then fed them normally. In these experiments, we used a two-pointed line or multi-jointed spline to measure lengths. We performed all analyses in python, with code assistance from GPT4.

Egg-laying behavior

We moved adult worms from a communal tank to individual wells in the center of a 24-well plate or 12-well plate (Fisher Scientific #07-200-82). We mounted the plate on an elevated platform within an enclosed behavioral rig. The plate was illuminated with uniform white light, from addressable LEDs (Adafruit #SK6812RGBW, powered by 5V DC) arranged in a ring around the plate behind a white acrylic diffuser (McMaster-Carr #8505K741). An infrared camera (Basler Ace ac4024-29um USB 3.0 monochrome) was mounted underneath the plate to allow filming of the worm’s ventral surface. A USB fan was mounted on the floor of the behavioral chamber for temperature regulation. We filmed worms at 1Hz for 8-12 hours. After filming, each worm was returned to its original communal tank.

To determine the timing of egg laying in adults with continuous access to mates, we collected adult worms from communal tanks and each was placed in an individual well of a 6-well plate. Every 3-4 days, we counted the number of eggs laid by each worm, removed the eggs, changed the water, and fed worms with brine shrimp or rotifers. To induce self-fertilization, we moved juvenile worms with immature reproductive systems from communal tanks to 24-well plates and fed them with rotifers. Once these worms had increased in size, we moved them to 6-well plates and fed with artemia. We recorded egg counts every 3-4 days when water and food were refreshed. To determine the time course of egg laying after mating once, we isolated juvenile worms. Once these worms were reproductively mature, we paired each worm with another isolated worm and allowed them to mate. After mating, we returned the worms to isolation. We recorded egg counts as previously described. If worms stopped laying eggs for more than two weeks, we allowed them to mate again before re-isolating them.

To find egg positions, we maintained communal tanks according to our husbandry protocol. Every 3-4 days, before cleaning, we recorded images of each face of the tank (2 short faces, 2 long faces, and the floor of the tank). We then scored the x- and y-coordinates of eggs on each face using FIJI. Since few eggs were laid on the floors of the tanks and it was unclear whether these eggs were attached or had been dislodged from the walls of the tanks, we did not score their locations (Fig. 6F). We also scored the x- and y-coordinates of the boundaries of the water in each image. From these image-specific boundaries, we calculated a common set of bounds for each type of face (i.e., short, long). We scaled egg coordinate positions and aligned them to common bounds for analysis of spatial egg laying preferences. To ask whether worms add to existing clutches of eggs, we cared for communal tanks as previously described but did not remove eggs from their faces. We took images of each face, and then we transferred each cohort of worms in a tank to a different tank. After 3-4 days, we took images of tank faces again, scored coordinates of each egg, and recorded whether each egg was old or newly laid.

For experiments measuring how worms altered their egg-laying behavior based on food availability, we selected 17 and 18 communal worm tanks respectively. On the first day of the experiment, half of these tanks were randomly selected to have food (brine shrimp) withheld. After four days with or without food, we collected and counted eggs from all boxes. To test whether egg laying recovers when unfed worms are fed, we randomly selected communal tanks of adult worms to have food withheld. Four days after food was added or withheld, we removed all eggs that were laid, and then fed all communal tanks. Three days later, eggs were collected and counted.

Acknowledgements

We thank Gonzalo Giribet, Matthew Hooge, James Hanken, Javier Ortega-Hernandez, Andrew Berry, and all members of the Srivastava lab for helpful discussion. We also thank James Hanken, Júlia Chaumel, and Christina Daly for histology help and infrastructure. VC and APK are Fellows of the Jane Coffin Childs Fund for Medical Research. ST acknowledges funding from the Harvard Museum of Comparative Zoology, Herchel Smith Undergraduate Science Research Program, and the Program for Research in Science and Engineering. This project was supported by grants from the National Institute of General Medical Sciences of the NIH to MS: R35GM128817 and R35GM153252.

Additional information

Author Contributions

VC, SET, and MS designed the project; SET and VC performed development, physiology and behavior experiments; VC and SET analyzed the data; APK performed and interpreted immunofluorescence and imaging experiments; DMB performed and interpreted histology experiments; VC, SET, and MS wrote the manuscript with input from all authors; and MS supervised the project.

Additional files

Table S1. Survey of reproductive traits within Acoelomorpha.

Video S1. The penis sheath is ciliated. Live-imaging of the ventral surface of a worm shows beating cilia lining the sheath. This set of cilia is distinct from the cilia lining the epidermis of the worm.

Video S2. The male gonopore is ciliated. Live imaging of the ventral surface of a worm shows that the male gonopore has beating cilia.

Video S3. Eggs are laid through the mouth. Video of an adult worm laying eggs, filmed from the ventral surface at 1Hz and sped up 2000x. The worm lowers its mouth and then contracts its body. During these contractions, a cluster of mature eggs can be seen distending the body. An individual egg is loaded into and travels through the pharynx eventually emerging through the mouth.

Supplement

Table S2:

Table S3:

Table S4:

Figure S1:

Figure S2:

Figure S3:

Figure S4:

Figure S5:

Figure S6:

Figure S7:

Figure S8:

Figure S9: