Some Like It Fat: Comparative Ultrastructure of the Embryo in Two Demosponges of the Genus Mycale (Order Poecilosclerida) from Antarctica and the Caribbean

During embryogenesis, organisms with lecithotrophic indirect development usually accumulate large quantities of energetic reserves in the form of yolk that are necessary for larval survival. Since all sponges have lecithotrophic development, yolk formation is an ineludible step of their embryogenesis. Sponge yolk platelets have a wide range of morphological forms, from entirely lipid or protein platelets to a combined platelet showing both lipids and proteins and even glycogen. So far, there are no comparative studies on the nature and content of yolk in congeneric species of sponges inhabiting contrasting environments, which could have putative effects on the larval adaptation to environmental conditions. Here, we have taken advantage of the worldwide distribution of the sponge genus Mycale, in order to compare the embryogenesis and yolk formation in two species inhabiting contrasting latitudinal areas: M. acerata from Antarctic waters and M. laevis from the Caribbean. We have compared their brooded embryos and larvae using scanning and transmission electron microscopy, and calculated their energetic signatures based on the nature of their yolk. While the general morphological feature of embryos and larvae of both species were very similar, the main difference resided in the yolk nature. The Antarctic species, M. acerata, showed exclusively lipid yolk, whereas the Caribbean species, M. laevis, showed combined platelets of lipids and proteins and less frequently protein yolk platelets. The larvae of M. acerata were estimated to possess a two-fold energetic signature compared to that of M. laevis, which may have important ecological implications for their survival and for maintaining large population densities in the cold waters of the Southern Ocean.


Introduction
Embryogenesis in sponges is a highly diverse process that can occur in the water column in oviparous species or within the sponge body in viviparous species (see reviews by [1,2,3,4]), larvae of both species using a comparative approach, focusing on species-specific adaptations in terms of yolk types and abundances, with possible relation to life in cold waters.

Light and electron microscopy
Samples were either preserved in 4% formalin for light microscopy or preserved in 2.5% glutaraldehyde in PBS for both scanning and transmission electron microscopy following protocols detailed in [21]. Formalin-preserved samples were later rinsed for 2 h in distilled water and dehydrated through an ascending series of ethanol (70%, 96%, 100%) and xylene. Then, samples were embedded in paraffin at 60°C overnight, and cut with a Microtome Micron HM325 to 5 μm. Staining was performed using Methylene blue and Hematoxilin-Eosin standard protocols. Samples for transmission electron microscopy preserved in 2.5% glutaraldehyde in PBS were then rinsed using a solution of 0.6M NaCl and PBS and fixed for 1h at 4°C in 1% osmium tetroxide-potassium ferrocyanide. Later, samples were rinsed in PBS and distilled water several times, dehydrated though a series of ethanol and propylene oxide, and embedded in Spurr resin for 3 days. Sections of resin blocks were performed at 64 nm using an ULTRACUT ultramicrotome, stained with lead citrate and uranyl acetate and observed with a JEOL 1010 electron microscope with a Gatan module for image digitalization at the Microscopy Unit at the Scientific and Technological Centers (CCiT), Universitat de Barcelona. Samples preserved in 2.5% glutaraldehyde in PBS for scanning electron microscopy were dehydrated through an increasing ethanol series and critically-point dried. After that, samples were coated with carbon, mounted in stabs, and observed with a JEOL 7100F Field Emission scanning electron microscope with a Gatan module for image digitalization at the Microscopy Unit of the CCiT, Universitat de Barcelona.

Yolk/lipid ratio estimations
Blastomeres and larval cells were measured in 43 embryos of Mycale acerata and 17 embryos and 65 larvae of M. laevis using the software ImageJ 1.48 (http://imagej.nih.gov/ij/). Transmission electron micrographs from 6 embryos of M. acerata and 6 embryos and 4 larvae of M. laevis were used to estimate the protein/lipid ratios using the software ImageJ. The specific measurements obtained to be used as a proxy to estimate total volume of lipids and proteins in blastomeres and larval cells were as follows: average volume of embryos and larvae, number of blastomeres or larval cells containing yolk in the embryo/larva, average volume of embryo/larval cells, average number of heterogeneous platelet/homogeneous (lipid or protein) platelet per cell (measured in 10-15 cells per embryo/larvae, n = 90 in M. acerata, n = 150 in M. laevis), and average volume of heterogeneous platelet/homogeneous (lipid or protein) platelet (measured in 15 cells per embryo/larvae, n = 90 in M. acerata, n = 150 in M. laevis).
The estimation of energetic signatures was performed on total volume of lipids and proteins in blastomere and larval cells using the Energy Conversion Factors (ECF, for lipids is 7.9 kcal/g and for proteins is 4.3 Kcal/g) calculated in [23].   In M. acerata, given that the average radius of cells in the late-stage embryo was 5,2 ± 0,9 μm and the average number of cells in the embryo was 2,51 x 10 05 ± 1,7 x 10 04 the amount of yolk contained in the embryos was estimated in 10,42 x 10 06 ± 1,7 x 10 6 μm 3 of yolk, being the lipid yolk the only platelet (Table 2). Approximately 8,23 x 10 −04 ± 2,8 x 10 −05 Kcal were contained in each late-stage embryo (given that the density of lipids is ca. 1 mg/ml and its ECF of 7,9 kcal/g; [23]). Assuming that yolk is consumed during larval formation, around 1/3 of the total yolk would remain in the larva (following M. laevis observations, see below), approximately 2,74 x 10 −04 ± 4,5 x 10 −05 Kcal would be the energy contained in the larvae of M. acerata.

Mycale (Mycale) laevis
Two different stages of the embryogenesis were found within the tissue of M. laevis: late stage embryos and early parenchymella larvae ( Fig. 2C-D, 3D). Embryos were round and bluish green in color, 500 μm in longest diameter, while early larvae were also green, bullet-shaped, and slightly larger, ca. 600-700 μm ( Table 1  collagen of 5 μm (Fig. 3E, 6A-B) with relatively few spirochaete-like bacteria lying on it (Fig. 6A). Follicle cells did not contain lipid inclusions but few protein inclusions and permitted the transit of nurse cells (Fig. 6A-B) and spherulous cells (Fig. 3E, 6B) to the lumen of the embryonic follicle.
Late-stage embryos were composed of round undifferentiated blastomeres divided into macromeres and micromeres (Fig. 2C). Early larvae (Fig. 3D-F) presented an epithelial layer of elongated ciliated cells (Fig. 3F) of ca. 20 μm of largest diameter and a ring of smaller posterior   (Figs. 6C-D, 7D) were only found in blastomeres and larval cells in anterior and mid parts, and appeared to be membrane-bounded (Fig. 7C). In turn, non-membrane-bounded homogeneous yolk platelets (protein nature) were also found in nurse cells surrounding embryos and larvae (Fig. 6E). Posterior cells of larvae were ciliated and contained large lipid droplets (Fig. 6F). Larval cells contained from half to 1/3 of the yolk contained in the late-stage embryos (approximately 5,6 x 10 06 ± 1,34 x 10 06 μm 3 in embryos and 2,2 x 10 06 ± 0,5 x 10 06 μm 3 in larvae), indicating that yolk was consumed during larval formation. Larval sclerocytes were round cells in cross-section of 1,5 μm that enveloped silica spicules of ca. 1 μm in largest diameter and no inner filament (Fig. 7E). Tylostyles were also observed in the anterior and posterior parts of the larva (not shown). Few collagen fibers found within the medial part of the larvae (Fig. 7A, F). Rod-like bacteria of 5 μm in largest diameter among the cells in the mid-part of the larva (Fig. 7A-B), but not detected in the anterior or posterior parts of the larva.
Given that the average radius of larval cells in M. laevis was 4,59 ± 0,4 μm and the average number of cells in the larvae was 2,9 x 10 05 ± 8,1 x 10 04 the amount of yolk contained in early larvae in M. laevis was estimated in 2,2 x 10 06 ± 0,5 x 10 06 μm 3 of yolk, being the heterogeneous yolk the dominant platelet (Table 2). Approximately 1,2 x 10 −04 ± 2,8 x 10 −05 Kcal were contained in each embryo/larva (given that 1,21 g/ml was the average density of lipoproteins and 4,3 Kcal/g the ECF, and the density of lipids is ca. 1 mg/ml and its ECF of 7,9 kcal/g; [23]).

Comparative morphology of reproductive elements in Mycale species
Both Mycale acerata and M. laevis are brooding species, with several embryonic stages within the same individual, which concurs with the observations in other Mycale species [38,47]. Embryos and larvae of the genus Mycale are usually highly pigmented [39,41,44,47], being green or blue in M. laevis with a lighter posterior pole (this study and [50]) and yellow in M. acerata (this study). The embryos of M. acerata and M. laevis were highly similar in size (ranging from 500 to 750 μm) and shape, being also strongly similar to the embryos and larvae of other  Mycale species (see Table 1; [38,39,42,45,46,47,49,50,51]. Embryos of M. acerata and embryos and larvae of M. laevis appeared surrounded by a follicle comprised of collagen (thicker in M. laevis than in M. acerata) and a pinacocyte-like cell layer, while in M. fistulifera also spicules appear surrounding embryos [47]. Comparative Ultrastructure of the Embryo in the Genus Mycale Most of the larvae of the Mycale species studied so far possess spicules exclusively in the posterior part [37,40,41,45,46,47]: all spicule types present in adults except for sigmas in M. hentscheli (previously known as M. macilenta var. australis) [45,46], tylostyles and sigmas in M. fistulifera [47], oxytylotes and anisochelae in M. fibrexilis [41], and subtylostyles, sigmas, toxas, and anisochelae in M. syrinx [42]. In M. laevis, we found tylostyles in both the anterior and posterior part of the larvae, and also sclerocytes containing spicules, similar to that found in Mycale sp. [49]. Given that no mature larvae were found in M. acerata, we cannot discard the occurrence of spicules in the larvae of the Antarctic species. Interestingly, it has been suggested that the presence of spicules in larvae may help larval sinking when yolk is depleted and therefore enhance the changes of larvae settlement by contacting the substratum (e.g., [52]).
The larval ciliation pattern in the genus Mycale seems to differ slightly from species to species. Among the non-tufted parenchymella larvae described in the different species in the genus, there are examples with entirely ciliated larvae [38], larvae ciliated except for the anterior and posterior part [47,49], larvae only ciliated in a posterior ring [50], or larvae entirely ciliated except for the posterior pole [37,39,40,42,45,53]. In M. laevis, the ciliation appeared in the entire larva, except for a small bare area in the posterior pole, which contrasts with that observed previously for M. laevis larvae (orange morph) where ciliation appeared only in a posterior ring [50]. No data on the larval ciliation could be described for M. acerata since no larvae were found in the samples of this species.
In M. laevis, spherulous cells appeared in the lumen of the embryonic follicle and close to the collagenous layer of the follicle. In Mycale sp. [48,49], spherulous cells were detected intermingled among the ciliated epithelial cells of the larva. The nature of the spherules in the spherulous cells is unknown in M. laevis and Mycale sp., but given the strong pigmentation of the larvae in the genus Mycale, and the observation of pigment granules in the larval epithelium of M. fibrexilis [41], it could well be that the spherules are pigment accumulations. However, there is an alternative explanation for the presence of spherulous cells in the larvae of M. laevis. Spherulous cells containing secondary metabolites in Aplysina fistularis [54] bear a strong resemblance with the spherulous cells of Mycale laevis. Interestingly, unpalatability of larvae of Mycale laxissima has been described before related to the use of secondary metabolites against sympatric predators [55]. However, it appears that M. laevis is not deterrent of sympatric predators [56,57]; similarly, M. acerata is non-deterrent against the sympatric predator Odontaster validus [58]. Thus, we hypothesize that spherulous cells in M. laevis might not be involved in defensive strategies but rather in other biological processes.
Only in M. laevis, bacteria were present, being in the inner part of the larva, suggesting a mechanism of vertical transmission for bacterial symbionts in this species. This is the first time that such a mechanism is reported for any Mycale species, a mechanism otherwise very common in other sponges (e.g., [59,60]).
One of the most striking differences between the embryos of M. laevis and M. acerata was the nutrient reserve nature and content. While in M. laevis the yolk nutrient reserves were comprised of homogeneous protein yolk, lipid droplets, and heterogeneous yolk (mixture of protein, lipid, and glycogen), in M. acerata embryos appeared to rely completely upon lipid droplets for their further development, survival, and subsequent settlement. Usually, sponges possess different degrees of abundance of protein and lipid yolk and glycogen (e.g., [21,35,61,62]). However, the discovery of embryos entirely containing lipid yolk occurring in M. acerata is highly remarkable since it is the first time that such a feature is reported for any sponge. However, it is important to note here that protein yolk could perhaps be formed in the latest stage of the embryonic development of M. acerata, even though this possibility has never been observed in any other sponge before.

Ecological implications of yolk composition in Mycale species
Antarctic marine species face a range of unique environmental challenges like extreme low and stable temperatures, typically between 0 to −1.8°C for most of the year, combined with the most intense seasonality in primary production in the world's oceans and highly seasonal ice scouring [63,64,65,66]. Marine animals are found abundantly at all low temperatures which per se do not limit life. There are many different adaptations to enable life in cold ecosystems, like slow growth rates, antifreezing proteins, or psychrophilic enzymes [64,67]. In relation to reproductive processes, there is a tendency towards long development periods, brooding, producing large eggs, and lecithotrophic strategies, especially in molluscs [64,68,69,70], although there are also many examples of highly abundant marine invertebrates in Antarctica with planktotrophic larvae [69,71].
Given that the Antarctic ecosystem is highly controlled by its strongly seasonal primary productivity, marine invertebrates have to cope with food scarcity during most of the year [65,72]. In this sense, one of the most abundant marine invertebrate in Antarctica, the krill Euphasia superba, depends entirely on lipid reserves to survive the winter [73,74]. Also, in Antarctic phytoplankton communities, up to 80% of their fixed carbon is transformed to lipids, compared to the 20% that has been observed in their counterparts of temperate communities [75].
Lipids are the major metabolic energy reserve in the larvae of marine animals [76]. In our study, the carbon stored in yolk reserves in the embryos of the Antarctic M. acerata was entirely fixed in lipids, while in the Caribbean M. laevis, lipid yolk was only the 30% of the total nutrient reserves of the embryos and larvae. Fat is an ideal storage material because it liberates twice as much energy as it is liberated by an equal weight of carbohydrate or protein [77]. Fat also serves to buoy floating animals, since it has lower specific gravity than water [77]. In low productivity environments such as the Southern Ocean, larval survival depends upon its energetic content and the metabolic rate at which those reserves are consumed during development [64,72]. In the Southern Ocean, the importance of low food availability has been raised as the major factor limiting developmental rates in marine invertebrates, being a greater constraint than the influence caused by low temperatures [64,72]. For instance, during the embryonic development of the sea urchin Sterechinus neumayeri, the metabolic rates increase largely during late stages of embryogenesis, when morphogenetic movements occur [78]. Likewise, in fish development, the embryo undergoes intensive cellular movements (gastrulation and epiboly), which require great quantities of energy rich molecules [79]. In the case of M. acerata, the large storage of lipid droplets could favor rapid morphogenetic movements and therefore quick metamorphosis, as well as a large energetic reserve storage that may enable massive recruitments of larvae. In addition, the large storage of lipids that is estimated to occur in larvae of M. acerata could reflect a higher buoyancy of the larvae. This higher buoyancy could enable drifting during longer periods in the currents [49] and therefore facilitate dispersal to more distant areas. In summary, large lipid storage in M. acerata embryonic elements could be behind the remarkable ecological success reported for this species in the Ross Sea and the South Shetland Islands, where these sponges form large and massive populations [28,29,30]. This is the first study comparing the embryogenesis in two sponges inhabiting contrasting habitats at an ultrastructural level, which allowed the discovery of divergent strategies in the yolk formation with presumable ecological implications. In particular, the observation of an entirely lipid yolk content in the sponge M. acerata is unique among the phylum Porifera, and encourages further studies on other Antarctic sponges to establish whether this is a general trend in cold environments.