Reproductive biology and energetics of the brooding sea star Anasterias antarctica ( Echinodermata : Asteroidea ) in the Beagle Channel , Tierra del Fuego , Argentina

The brooding sea star Anasterias antarctica is distributed from the coast of Patagonia to the northern Peninsula of Antarctica. In the Beagle Channel, the females of A. antarctica brood their eggs for seven months and do not feed during this period. The endoparasite Dendrogaster argentinensis (Crustacea: Ascothoracica) causes castration in several species of Anasterias. We randomly collected four samplings of adults in May, August and October (brooding period) and January (non-brooding period). The gonad (GI) and pyloric caeca index (PCI) were calculated as organ wet weight (g) x 100/total wet weight (g). Each individual was sexed by microscopic examination of the gonads. Sex ratio, brooding females/non-brooding females and mature females/ non-mature females ratios was 1:1. The male GI reached maximum values in January, when most individuals were sexually mature. The GI of non-brooding females reached its maximum during October when it was significantly higher than those from brooding females. The PCI was minimum in October, being lower in brooding females (August and October). During the non-brooding period, mature females had a significantly higher GI than non-mature females. The PCI did not vary neither between males, nor between mature and non mature females. By the end of the brooding period, non-brooding females showed a higher GI than the brooding females. This is explained by proliferation and increase of the oocytes size of non-brooding females. Mature females showed an incremented GI with presence of mature oocytes, while non-mature females exhibited more abundance of previtelogenic oocytes. Males showed synchronicity in reproductive condition. The females that have not brooded presented a process of active gametogenesis, reaching the summer with a high GI, therefore becoming mature females. Females that had brooded were probably lacking energy for new gonadal maturation. The pyloric caeca would be performing the role of a reserve organ in the brooding females, decreasing its size during the brooding period. Prevalence of D. argentinensis in A. antarctica was 11.06%. As this parasite was recorded in sea stars lacking gonads, these infected hosts could have been castrated. Rev. Biol. Trop. 65(Suppl. 1): S221-S232. Epub 2017 November 01.

Moreover, reproductive effort varies with the mode of development (Pérez et al., 2015).Generally, species with indirect development (with larvae), reproductive investment is usually equal between sexes (Raymond, Himmelman & Guderley, 2007;Mariante, Lemos, Eutrópio, Castro & Gomes, 2010).In species with direct development, reproductive effort differs between sexes.Normally, males invest more energy to gametes production than females, but females face the additional cost of incubating embryos (Raymond, Himmelman & Guderley, 2004;Gillespie & McClintock, 2007;Pérez et al., 2015).The pyloric caeca work as a reserve of energy, located inside the arms and the size and concentration of lipid are indicative of the energy reserves of the individual (Lawrence, 1987a).Some asteroids show an inverse relationship between gonadal and pyloric caeca mass variation cycles due to the transference of material towards the developing gonads (Chia & Walker, 1991), which changes the proximal composition of both organs during the reproductive cycle (Barker & Xu, 1991).
The oral-brooding sea star Anasterias antarctica (Lütken, 1857) is widely distributed on the coasts of South Patagonia and northern Peninsula Antarctica, and occurs from the intertidal zone to 150m depth (Bernasconi, 1964).This sea star is a common predator of intertidal and shallow sub littoral communities in Tierra del Fuego, feeding mainly on the mussel Mytilus chilensis and the gastropods Pareuthria plumbea and Trophon geversianus (Curelovich, 2012) and is commonly observed in the intertidal zone during low tide (Pérez et al., 2015).The females of A. antarctica of Beagle Channel, brood their eggs in the oral surface for seven months (Pérez et al., 2015) and do not feed during this period (Pérez personal obsevation), as it was also documented for Golfo San Jorge and Puerto Deseado (Patagonia, Argentina) (Gil & Zaixso, 2008;Gil, Escudera, & Zaixso 2011) and Malvinas Islands (Laptikhovsky, Brickle, Söffker, Davidson, Roux, Rexer-Huber, Brewin, Kälkvist, Brown, Brown, Black, Anders, Cartwright, Poncet & Parker, 2015).Other species of sea stars such as Anasterias rupicola, from Marion Island (Southern Ocean), feed during brooding (Blankley & Branch, 1984).In Pérez et al. (2015) males reproductive effort was found to be 25% lower than that of females, probably due to brooding costs.Gonadal maturation was recorded in summer and gametogenesis in winter in both sexes (Pérez et al., 2015).
The aim of this study was to expand knowledge on the reproductive biology and energetic of the brooding sea star Anasterias antarctica along its reproductive cycle.A second aim was to study the prevalence presence parasite D. argentinensis on A. antarctica.

Sampling and processing:
Anasterias antarctica measuring ˃ 20 mm in radius were randomly collected every two meters along the parallel transect in the intertidal zone of Ensenada Zaratiegui, Tierra del Fuego National Park, Beagle Channel (54°51'00,14'' S, 68°29'38,79'' W; Fig. 1).Considering the brooding period of A. antarctica was established four collections were made during 2009 and 2010: three of them during the brooding period: (1) on 28 May 2009, at the beginning of brooding, when oocytes are on early stages of maturity; (2) on 21 August 2009, in the middle of the brooding period, when embryo are on advanced maturity; and (3) on 19 October 2009, at the end of brooding (Fig. 2).The fourth collection corresponded with the nonbrooding period (28 January 2010).Seasonal samples of 30 individuals were taken.January sampling was used as an indicator of the nonbrooding period.The specimens collected were transported to the Laboratorio de Ecología, Fisiología y Evolución de Organismos Acuáticos from the Centro Austral de Investigaciones Científicas-CONICET (Ushuaia, Argentina) for subsequent processing.Prior to dissection, each individual was superficially dried with absorbent paper, weighed (total weight, ±0.01 g), and the distance from the tip of the longest arm to the opposite interradius (length) was measured using an electronic caliper (±0.1 mm).Animals were then dissected and gonads and pyloric caeca (digestive glands) were weighed separately (±0.01 g).The indexes of different body components (GI: gonad index and PCI: pyloric caeca index) were calculated as organ wet weight (g) x 100/total wet weight (g).GI values were partially published in Pérez et al. (2015).Gonad and pyloric caeca indexes were computed as an indicator of energy reserves stored in these organs (Lloret & Planes, 2003).

Histological determinations:
The slides were prepared from a one gonad of each specimen was fixed in Bouin´s solution for 12 h, then washed in water prior to transfer to 70% alcohol and then dehydrated in a graded ethanol series.Gonads was embedded in liquid paraffin (Paraplast ® ), sectioned at 5-7 µm (microtome Leica RM2125RT) and stained with Carazzi´s hematoxylin and eosin (Pérez et al., 2008).Sections were examined microscopically and each individual was sexed.Individuals sampled during the brooding period were assigned to one of three groups: brooding females, non-brooding females, and males.Individuals sampled during the non-brooding period were assigned to one of three groups: sexually mature females, non-mature females and males.The individuals without gonads were excluded from the analyses.

Occurrence of Dendrogaster argentinensis:
In parasitized specimens with D. argentinensis, the parasite was carefully extracted under stereoscopic magnifying glass (10x).The prevalence of infection was estimated according to Bush, Lafferty, Lotz & Shostak (1997) as the percentage of parasitized specimens by total individuals, sex and female gonadal stages.
The sex ratio and reproductive condition were determined by examining histological slides in all individuals collected throughout the study.The sex ratio was tested using a goodness of fit (G) test, with equal sex ratios as null hypothesis (Sokal & Rohlf, 1995).The individuals without gonads were exclude from the analyses.
In males, seasonal variations in GI were analyzed using a Kruskal-Wallis test followed by unplanned Dunn´s multiple comparisons test.The variations in PCI were analyzed using one-way ANOVA, followed by Tukey HSD comparisons.
The GI and PCI of females sampled during the brooding period were analyzed using a twoway ANOVA to test for differences between period (May, August and October) and reproductive condition (brooding and non-brooding females).Pair-wise differences between periods were analyzed by unplanned Tukey HSD.
In females sampled during the non-brooding period (January), differences in GI and PCI between mature and non-mature females were tested using the unpaired t test.
Assumptions of normality (Kolmogorov-Smirnov test) and homogeneity of variances (Bartlett test) were previously verified for PCI of males, and GI and PCI of females during the brooding and non-brooding period.The log transformation was applied for GI of females during non-brooding period.In case of the assumptions were not validated, a nonparametric test was used (Zar, 1984).
Correlation between GI and PCI were studied through Pearson correlations (Zar, 1984).

RESULTS
Sex ratio, brooding females/non-brooding females and mature females/non-mature females ratios did not differ significantly from 1:1 (Test G, P>0.05 in all cases).
Females brooding period: For GI, the interaction between period and reproductive conditions was not significant (two-way ANOVA, F 2,38 = 2.3209 , P>0.05).On the other hand, the GI of non-brooding females reached its maximum in October when it was significantly higher than those of brooding females (Simple effect, P<0.05) (Fig. 4A).For PCI, the interaction between period and reproductive conditions was not significant (two-way ANOVA, F 2,46 = 0.5765, P>0.05) (Fig. 4B).The PCI was minimum in October (Simple effect, P<0.05) and were lower in brooding females than non-brooding females (August and October).The correlation between GI and PCI was not significant for brooding (r: 0.047; P>0.05) and non-brooding females (r: 0.018, P>0.05).

Histology of gonads:
Females sampled during the non-brooding period were divided in two groups: sexually mature females and non-mature females.
The ovaries of mature females contain numerous mature ova with acidophilic cytoplasm, densely packed into the lumen and small number of previtellogenic oocytes located along the acinus wall (Fig. 6A).Non-mature females present along acinus wall oogonias, previtellogenic oocytes and small vitellogenic oocytes (Fig. 6B).Occurrence of Dendrogaster argentinensis: Prevalence of D. argentinensis in A. antarctica was 11.06%, all were found in the coelomic cavity of sea stars.Males, brooding, non-brooding and mature females had parasites.All specimens without gonads were parasitized (Table 1).
Brooding in A. antarctica from Beagle Channel occurs during the coldest months of the year (austral winter), as it was reported for other populations (Gil et al., 2011;Laptikhovsky et al., 2015) and also for other   asteroid such as Leptasterias polaris (Himmelman et al., 1982) and Leptasterias tenera (Hendler & Franz, 1982).During brooding and non-brooding period, there were females in two reproductive conditions simultaneously.Through the brooding period there were brooding and non-brooding females and which showed relatively lower GI values in relation to the average GI recorded in summer (nonbrooding period) (Fig. 4A).Towards the end of the brooding period, non-brooding females showed a significantly higher GI than the brooding females.Similar results were reported by Bosch & Slattery (1999) for the antarctic sea star Neosmilaster georgianus and by Gil & Zaixso (2007) for Anasterias antarctica.This can be explained by histological observations of A. antarctica ovaries due to the proliferation and the increase of the oocytes size (vitellogenic oocytes) of the non-brooding females, while the ovaries of brooding females either had no oocytes or had small previtellogenic oocytes.
Once the brooding period was over, there were two reproductive conditions found in females during the non-brooding period: mature females and non-mature females (Fig. 5A).Mature females 6.5% GI showed histologically mature oocytes (Fig. 6A), which could be spawned and fertilized by the end of the summer, while the GI from the non-mature females remained relatively low (approximately 0.5%) exhibiting mainly previtelogenic oocytes (Fig. 6B).Therefore the females that didn´t brood (non-brooding females) present an active gametogenesis reaching the summer with a high GI and becoming mature females, whereas the females that have brooded (brooding females) would not have enough energy to grant a new gonadal maturation during summer, and a new incubation.Though more frequent samples are necessary we suggest that would become mature (non-brooding) females during following summer taking then two years to complete the oogenesis cycle Increased GI in males and mature females during summer coincides with a maximum gonadal development and maturation of the gametes (Figs.3a, 5a and 6).This result agrees with Salvat (1985) and Gil & Zaixso (2007) where they suggest that spermatogenesis has an annual cycle while oogenesis could be a biennial cycle in A. antarctica.Unlike the observed in females, the males showed synchronicity in the reproductive condition, maintaining GI values around 1-2% during May, August and October and increasing significantly in January (Fig. 3a) when all males were sexually mature.GI maximum values reported here were lower than those reported for Neosmilaster georgianus (Bosch & Slattery, 1999) and similar to those reported for Leptasterias polaris (Raymond et al., 2004).This difference in the synchronicity between sexes would be interpreted as a consequence of highly differential energetic costs of reproduction of A. antarctica at high latitudes, which imposes only to females a biannual cycle.
The pyloric caeca work as a reserve of energy, located inside the arms of the asteroids, and an indicator of the individuals energy reserves is the size and concentration of lipid present on it (Giese, 1966;Lawrence, 1987b).Bosch & Slattery (1999), McClintock & Pearse (1987) and McClintock (1989) reported for a high accumulation of pyloric caeca lipids for the antarctic sea stars Neosmilaster georgianus, Diplasterias brucei and Notasterias armata which may provide easily metabolized energy stores for the incubation period.
The PCI in Anasterias antarctica remained constant through the study in males, nonbrooding females, mature females and nonmature females (Figs. 3b,4b and 5b), being approximately 15%.Significantly lower PCI values were recorded in brooding females from the middle to the end of the brooding period, reaching nearly 5% (Fig. 4b).This is due to the brooder's inability to feed during incubation because the embryos mass is located over the mouth.Likewise, it is important to highlight that both the population of A. antarctica here studied, and N. georgianus (Bosch & Slattery, 1999) were never observed feeding while brooding.It is possible to conclude that the pyloric caeca is performing the role of a reserve organ in the brooding females, decreasing its mass during the brooding period.Acordding to this results, Cerda, Flores, Zaixso & Córdoba, (2016) reported carbohydrates of pyloric caeca and the body wall as reserve substances to endure the periods of long-term starvation of brooding females or those of reduced consumption rates of non-brooding females and males.During summer, both mature and nonmature females showed a high PCI (aproximately 15%), leading to conclude that both groups have fed and destined energy to pyloric caeca, in accordance with the maximum feeding activity found before and after the reproductive period of A. antarctica in Malvinas Islands by Laptikhovsky et al. (2015).
Constancy in PCI values in both sexes and the different reproductive conditions (except in brooding females) suggests that the food of A. antarctica´s in the Beagle Channel could be available throughout the year that should be confirmed almost by increasing samplings.Bosch and Slattery (1999) found that N. georgianus PCI descending in brooding females while the incubation proceeds and not confirming differences between PCI values of males and non-brooding females, suggesting a similar pattern than those reported here.Hendler & Franz (1982) reported a cesation of feeding during brooding period toghether with a decline in pyloric caeca supply in Leptasterias tenera.Raymond et al. (2004) described in Leptasterias polaris that the PCI descends both in males and females; in males this decrease was during the gametogenesis while in females during the months of incubation as it was observed in A. antactica.Gil & Zaixso (2007) observed a decrease in the PCI and the feeding rate in Anasterias antarctica durig winter.
Our findings reinforce the idea that the reproductive effort differ between sexes in brooding (Chia, 1969).Males invest considerably more effort in gamete production than females, adopting a strategy characteristic of broadcast spawning species, while females assume a strategy that is more characteristic of brooders as described by other authors (Raymond et al., 2004;Mercier & Hamel, 2008).
Even though the reproductive effort of female A. antarctica is 25% higher than in males (Pérez et al., 2015), males gametes production is higher, as reflected in the IG values (Figure 3a,4a).Females reproductive effort is higher due to the long starvation they suffer during the brooding period (Pérez et al., 2015), wich is evidenced in the decrease of the energetic reserves of the pyloric caeca in the brooding females from the middle to the end of the brooding period in concordance with Gil & Zaixso 2007.The overall prevalence of D. argentinensis in A. antarctica was high (about 11 %), compared to the 1.5% of A. antarctica infested in Ría de Puerto Deseado (Santa Cruz, Argentina) (Salvat, 1985), Golfo San Jorge (Patagonia, Argentina) (Gil, pers. comm.) and 3% in Península Foca,(Santa Cruz, Argentina) (Gil, pers. Comm.).In host populations with a high prevalence of infection, parasitic castration may negatively affect their reproductive potential unless there are mechanisms that compensate for reproductive losses.For example, the prevalence of D. argentinensis was over 90% in Allostichaster insignis on Otago and the castration resulted in an increase in the incidence of asexual reproduction among larger individuals (Palmer, 1997).In this study, D. argentinensis was recorded in 100% of the sea stars lacking gonads therefore these infected hosts could have been castrated (prevalence was 11.06% and 28% of them did not presented gonads).However, comprehensive investigations are needed to understand the effect of this parasite on the reproduction of A. antarctica.
Finally, it is possible to assert the importance of the pyloric caeca as a reserve organ in the brooding females especially in species where feeding is impossible during incubation period.Moreover, due to this impossibility of feeding by females throughout this period reproductive effort is higher than in males, confirming that the reproductive effort differ between sexes in brooding species.

ACKNOWLEDGMENT
This study was supported by grants from ANPCyT, CONICET (PIP 0128 and PICT 1270), Fundación Científica Felipe Fiorellino and University of Maimónides.Authors are grateful to the members of Laboratorio de Ecología, Fisiología y Evolución de Organismos Acuáticos (CADIC-CONICET) and particularly to D. Aureliano, M. Gutierrez and S. Rimbau for technical assistance.