Morphological description of the first protozoeal stage of the deep-sea shrimps Aristeus antennatus and Gennadas elegans, with a key

Accurate information on commercial marine species larvae is key to fisheries science, as their correct identification is the first step towards studying the species’ connectivity patterns. In this study, we provide a complete morphological description of the first protozoeal stage of the valued deep-sea blue and red shrimp Aristeus antennatus and of the small mesopelagic shrimp Gennadas elegans. These two larval morphologies previously posed a risk of misidentification, thus hindering the study of A. antennatus larval ecology and dynamics in the context of fisheries science. Using specimens caught in the plankton at various locations in the Northwestern Mediterranean Sea and identification confirmed by molecular methods, the larvae of A. antennatus and G. elegans are distinguished from each other by the ornamentation of the antennula. A possible confusion in previous descriptions of Aristeidae larvae is addressed and a new key for the identification of Dendrobranchiata larvae provided.

Fisheries science depends on reliable and sufficient data about exploited species to build efficient strategies that ensure the durability of marine resources. One fundamental aspect of fisheries science is the study of species connectivity, as this information can shape the definition of stocks and set the range and scope of management instruments. Regardless of their adult habitat, many species have planktonic larvae. During this life phase, organisms are easily transported by currents; this plays a key role in terms of dispersal strongly influencing species' connectivity and recruitment patterns [1][2][3] . For crustacean decapods, there is a well-documented body of knowledge about the larval stages of some exploited species [4][5][6] . However, this is not the case for deep-sea Dendrobranchiata, for which information is lacking despite the economic relevance in fisheries of some species. The scarcity of these larvae in plankton samples and the challenges of rearing these species in the laboratory are probably one of the main causes of the limited number of descriptive studies on the subject. As a result, observed data on deep-sea Dendrobranchiata larval abundance and distribution are scarce, and many of their larval stages are still undescribed 7 .
The deep-sea blue and red shrimp Aristeus antennatus (Risso 1816) is targeted by bottom trawlers in the entire Mediterranean Sea and the Northwestern coast of Africa. Its global catch reached 2,988 tonnes in 2016 8 and in some areas like the Spanish Mediterranean coast, this species alone can represent up to 50% of fishermen associations' yearly revenues 9,10 . Its adult biology has been thoroughly studied [11][12][13] , particularly in the Northwestern Mediterranean Sea, where it has been subject to a long-term co-management plan at a local scale 14 . The reproductive period of A. antennatus spans from May to September, with a peak in July and August, when females aggregate at the continental shelf break 15,16 . As for the mesopelagic shrimp Gennadas elegans, its distribution englobes the whole Atlantic Ocean and the Mediterranean Sea. It has no commercial interest but it is often caught open 1 Institut de Ciències del Mar (ICM-CSIC), Passeig Marítim de la Barceloneta, 37- www.nature.com/scientificreports/ accidentally by bottom trawlers targeting A. antennatus. The reproductive cycle of G. elegans has not yet been studied, but larvae of the species have been caught in the plankton all year round (e.g. 17 ).
According to general knowledge about dendrobranchiate shrimps, the females spawn their eggs into the water column. The eggs then hatch into a nauplius, the first free-living larval phase which metamorphoses into a series of zoeal stages, often referred to as protozoea in their early stages and mysis during the late stages. The last mysis moults into a decapodid, which after a series of moults becomes a juvenile and begins searching for settlement in the adult habitat 18 . For A. antennatus, only 3 protozoeas and 2 mysis stages have been identified and described from plankton samples [20][21][22] . In 1955, Heldt 20 described two larval series obtained from plankton samples in the Balearic Sea (Northwestern Mediterranean) and reared in laboratory conditions that she attributed to Aristeus antennatus and Aristaeomorpha foliacea. For A. antennatus, the publication presented the morphological description for the three protozoea stages and the first mysis stage; for A. foliacea, it described the last naupliar stage, the protozoea II and III and the first mysis stage. In particular, the first protozoea (PZ I) of A. antennatus was described from a single individual, whereas the PZ I of A. foliacea remained undescribed since, as mentioned by the author 20 , the single available specimen was lost. Occurrence of A. antennatus larvae in the plankton have been reportedly scarce 21,23-26 until a recent study reported findings of all known larval stages of the species, with a particular high abundance of the PZ I 22 . For G. elegans, the only available description features only the PZ II and older stages 5 , while the description of the PZ I is included in a previous, more general study on the genus Gennadas 27 . Occurrence of Gennadas spp. PZ I has been widely reported in zooplankton studies (e.g. 21,[27][28][29].
Knowledge about Dendrobranchiata PZ I is particularly useful for fisheries science as this stage generally occurs from a few hours to a few days after hatching and can provide information on the spawning areas of the species 18 . Furthermore, information on larval behavior and distribution is essential to determine the connectivity patterns of commercial species and establish effective management strategies 30 . In this context, accurate identification of the larvae is key. The objective of this study was to accurately and comprehensively describe the first protozoeal stage of the deep-sea shrimps A. antennatus and G. elegans, to compare them in search for morphological distinguishing characters, and how the findings relate to previously available information.

Results
The main differential morphological characters between the first protozoea stage of the two species are summarized in Table 1. Also, we propose an identification key to distinguish the first protozoeal stage of Dendrobranchiata larvae of species occurring in the Northeastern Atlantic ocean and Mediterranean Sea, gathering information from our own observations and from available literature [31][32][33][34][35][36][37][38] . The general body morphology description of the Dendrobranchiata first protozoea stage can be found in some recent references 18,19 . The first protozoea (PZ I) of Dendrobranchiata larvae has a carapace covering part of the cephalotorax, followed by an unsegmented pleon and finishing in a large bilobed telson ( Fig. 3E), and the naupliar eye is still visible (Fig. 3C). These larvae have two pairs of antennae in the anterior part of the carapace: the first pair (antennula) is uniramous and the second one (antenna) is biramous. In the antennae Table 1. Summary of most relevant differential morphological characters between Aristeus antennatus and Gennadas elegans protozoea I larvae and the previous morphological description of the same larval stage attributed to A. antennatus. a: aesthetascs, s: setae.

Features
Gennadas elegans (this study) Aristeus antennatus (this study)  www.nature.com/scientificreports/  36 ; B and I. 27 ; C. 37 ; D, E, G and H. 38 ; F. 39 ; J, L and M. 5 ; K. 40 ; N. 20 . Drawings not to scale. www.nature.com/scientificreports/ (e.g. Fig. 3K, J, M), the exopod is composed by a long plumose outer ramus with several ringlets throughout its length, and the endopod is the inner ramus. The mouth appendices are composed by a pair of mandibles, with incisor and molar processes, and two pairs of maxillae. The larvae also present 2 pairs of biramous maxillipeds where the outer ramus is the exopod and the inner ramus is the endopod. The third pair of the maxilliped, when present, is still rudimentary. (fig. 1). Size: TL (total length) = 1.12-1.25 mm; CL (carapace length) = 0.37-0.49 mm; N (number of protozoea examined) = 13. Carapace (Fig. 1A): carapace almost rounded, longer than wider, reaching the level of the second maxilliped, with frontal organs visible at the anterior part; naupliar eye present flanked by a pair of compound eyes that are already visible through the carapace; 6 thoracic somites visible. Antennula (Fig. 1A,B): first paired uniramous appendage in the cephalotorax, consisting of 3 articles: proximal article subdivided in 5 ringlets, bearing 1 short serrulate seta on the posterior end; second article with 1 positioned at mid-length of article and 3 serrulate setae distally; distal article with 3 aesthetascs subterminally and 3 long sparsely plumose setae on the posterior end.
Third maxilliped (Fig. 1A,I): biramous paired appendage placed in the first thoracic somite not covered by the carapace, consisting of an endopod and an exopod. Endopod represented by a small bud tapered at the end; exopod unsegmented with 2 long plumose setae distally.
Third maxilliped ( Fig. 2A,I): biramous paired appendage placed in the first thoracic somite not covered by the carapace, consisting of an endopod and an exopod. Endopod represented by a small bud rounded at the end; exopod unsegmented with 2 long plumose setae distally.
Identification key for the first protozoeal stage of Dendrobranchiata larvae of the Northeastern Atlantic and Mediterranean Sea. Median branch of the anterior process of pereion with denticles only (Fig. 3D)

Deosergestes corniculum
Posterior process of pereion not swollen at base 9 9 Lateral process with 7 long spines at the base (Fig. 3H)

Sicyonia carinata
Length of antennula approximately equal to that of antenna Parapenaeus longirostris 12 Exopod of the third maxilliped with 3 setae (Fig. 3N)

Aristeus antennatus
Discussion. Although morphologically quite similar in most of their characters, the first protozoeal stages of A. antennatus and G. elegans bear some differences that will allow to distinguish them, as shown in Table 1 and in the identification key proposed. The first protozoea of A. antennatus presents 1, 4, 3 setae along the segments of the antennula, whereas in the case of G. elegans, the setal formula is 0, 1, 4. These characters are relatively easy to observe at the stereomicroscope, in most cases without the need of dissecting the specimens, and should provide an easy guide to differentiating the first protozoea of these two species. The identification and morphological description of the larval series of A. antennatus found in the plankton off the Balearic archipelago by Heldt in 1955 20 has proven to be fundamentally correct, as the descriptions of the rest of known stages of the species-PZ II, PZ III and mysis I-have been recently confirmed 22 . However, when comparing the A. antennatus PZ I from the present study with the one described by Heldt 20 , we found differences in the size of the larvae-the sole specimen in the cited study measured 1.55 mm, whereas in the present study the average total length is 1.2 mm. Moreover, we found differences between the two studies in the number of aesthetascs on the antennula, and in the number of setae on the exopod of the third maxilliped. While the possibility of an error can never be excluded, Heldt's meticulous work and thorough descriptions in all her publications on Penaeid larvae make it unlikely that she would draw and describe a morphological character that she did not observe. We here expose our considerations about this contradiction.
First, Heldt's study refers that one single specimen of first protozoea stage was caught for each of the studied species, A. antennatus and A. foliacea, but that the latter was apparently lost during manipulation and could not be described. Second, as seen in Table 1, the total length of the A. antennatus PZ I specimen measured by Heldt is 1.55 mm, while the next stage, PZ II, measured 1.50-2.03 mm 20 : this would mean that the PZ II was smaller than its previous stage. Variability in total length of these larvae has not been studied and might allow for such values, but Carreton et al. 22 found an average total length of only 1.2 mm (± 0.05) for the PZ I. On the other hand, the PZ II of A. foliacea examined by Heldt measured 1.9 mm 20 which is more in agreement with the length of the PZ I larva described as A. antennatus. Finally, Heldt's description of A. antennatus PZ I accounts for 3 setae on the exopod of the third maxilliped (mxp3), whereas in our findings, all individuals presented only 2 setae. Furthermore, it seems that, in Heldt's description, A. foliacea PZ II larvae present more developed characters than A. antennatus PZ II, as the mxp3 is described in A. foliacea with 3 setae on the exopod and 2 on the endopod, while in the case of A. antennatus, it only presents setae on the exopod. It would then be possible that, in the case of the PZ I, the more setose (3-setae) third maxilliped belongs to A. foliacea and the less setose (2-setae) one belongs to A. antennatus. For these reasons, we conclude that Heldt's description of A. antennatus PZ I is probably that of A. foliacea. The PZ I of A. antennatus would then have remained undescribed until now.
The present study provides the first detailed morphological description of the protozoea I larvae of A. antennatus and G. elegans according to modern standards, made from plankton samples after identification being confirmed with molecular analysis. The protozoea I larvae of the two studied species can be morphologically distinguished from one another mainly by the setation of the antennula. An identification key is provided allowing for the morphological identification of all first protozoea larvae of Dendrobranchiata for the Mediterranean Sea and Northeast Atlantic Ocean known today.
In a context where fisheries science is increasingly drawing on marine connectivity to design regional-scale management strategies for commercial species, larval distribution studies are one of the first stepping stones to effective planning, as they broaden the knowledge on species dispersal patterns. It is then essential to ensure a correct identification of these larvae, and morphological characters provide accurate, at-hand information even when molecular methods are not applicable. Our results set a starting point for A. antennatus connectivity studies in the frame of fisheries management, and we are confident that the identification key provided will make classification of the featured early larval stages accessible to both taxonomers in the field and non-specialists.

Method
Specimen collection. For A. antennatus larvae, the sampling was carried out in August 2016 in various locations off the Spanish Mediterranean coast (Table 2). We used a neuston sledge with a 300-µm mesh net between 0.5 and 1 m depth over bottoms of 123 to 1626 m. For G. elegans larvae, we sampled 3 stations off the Catalan coast in February 2017 ( Table 2). The selection of this second sampling interval outside of the reproductive period of A. antennatus was deliberate in order to avoid collecting a mix of the two species. We used a 60-cm diameter bongo with a 300-µm mesh net in oblique tows between 500 m depth and the surface, over bottoms of 1,952 and 1,790 m. All PZ I larvae from both samplings were sorted and identified at the stereomicroscope using the available keys and descriptions 20,21,31 and stored individually in 96% ethanol.
From the total of PZ I larvae caught in each sampling (527 in the summer and 11 in the winter), Carreton et al. 22 performed extraction, amplification and sequencing of the Cytochrome Oxydase I (COI) and 16S rDNA molecular markers on randomly-selected individuals (24 in the summer and 4 in the winter). All summer individuals analysed were identified as A. antennatus and all winter individuals as G. elegans. The genetic distance values were 0.00 within each species and 0.15 between species, the latter calculated with 16S rDNA data. Carreton et al. 22 also took Scanning Electron Microscopy (SEM) images and measurements of total length and carapace length for individuals of both taxa and sampling season.
Drawings and measurements. Drawings and measurements were made following the methods and equipment presented by Bartilotti et al. 39 . Additionally, and since they are transparent, the larvae were stained Scientific RepoRtS | (2020) 10:11178 | https://doi.org/10.1038/s41598-020-68044-9 www.nature.com/scientificreports/ with Chlorazol Black and Hematoxylin before being drawn. The long aesthetascs on the antennulae as well as the long plumose setae on the distal end of the exopods and on the uropods and telson were drawn truncated; the setules from setae were omitted from drawings when necessary. The drawings were then improved and digitally organized using GIMP software 40