Skip to main content

REsearch to improve PROduction of SEED of established and emerging bivalve species in European hatcheries

Final Report Summary - REPROSEED (REsearch to improve PROduction of SEED of established and emerging bivalve species in European hatcheries)

Executive Summary:
Shellfish production in Europe is subject to environmental changes, notably affecting the availability of natural spatfall and favouring the emergence of novel disease outbreaks. Hatcheries offer ways to counter these threats through their potential to ensure a more stable supply of seed and to carry out selective breeding to improve disease resistance. To achieve these general objectives, the methods of rearing may be improved to make them more reliable, to decrease the seed cost and to guarantee its quality. Hatchery seed production of Pacific oyster (Crassostrea gigas) is already well established, while for other species (Mytilus edulis, M. galloprovincialis, Ruditapes decussatus, Pecten maximus) it remains limited due to market or biological bottlenecks. In this context, the strategy of REPROSEED was to combine basic research, to improve the knowledge on bivalve physiology, and applied research, to implement new techniques by applying knowledge acquired on C. gigas. The program was structured according to biological phases and successively dealt with aspects of reproduction, larval rearing, metamorphosis, and seed growth. The influence of bacterial populations as well as the physiological and immunological development of the animals were investigated using novel genomic tools over the whole process of larval and seed rearing.
The main outcomes can be summarized as follows:
(i) The methods of broodstock conditioning were improved for oyster and clam. For oyster, we demonstrated the effect of preconditioning on the quantity and homogeneity of gametes. For clam, optimal conditions of maturation and spawning were determined.
(ii) Mass sequencing of transcripts was performed at successive stages of development of the studied species, providing new data and genomic tools. They are used to identify the genes involved in the acquisition of competence, maturation, the immune response to stress and bacterial challenge. Their analysis led to better understanding of the development phases and the ontology of immune system.
(iii) Criteria of gamete quality and maturity were investigated by biochemical, proteomic and genomic approaches. Criteria were found to be species-specific. In the Pacific oyster, some proteins discriminating the bad and good quality oocytes were identified while it was not possible by screening gene expression. In male oyster and scallop, protein markers of spermatozoon maturity were identified. Conversely, expression of numerous genes differentiated bad from good oocytes in R. decussatus.
(iv) The use of RAS (recycling aquaculture systems) to rear larvae and postlarvae gave very positive results, which should encourage the replacement of traditional batch or flow-through systems with specific adaptation for each species. RAS significantly reduce needs for water, energy and labor, and provide a more stable environment.
(v) Probiotics added to scallop larvae increased their survival, settlement and the resulting seed yield. They are a promising new development for rearing such sensitive larvae.
(vi) Despite numerous trials, none of the tested inducers of metamorphosis enhanced the settlement of mussels, clams, or scallops in a reliable manner. Nevertheless, an increase of the metamorphosis rate can be reached with an appropriate substrate for settlement and a good physiological status of pediveligers.
(vii) The microbiota analysed by pyrosequencing or after culture on agar in larval and postlarval rearing were observed to be complex and niche-specific, especially in larvae and postlarvae although they ingested continuously bacteria. The bacterial communities, in each niche, changed over time but were little influenced by rearing conditions. They may influence bivalve physiology as the addition of probiotics evidenced it. They tended to become more heterogeneous in postlarvae, and vibrios were rarely detected. Finally the fingerprint of bacteria from hatcheries is unlikely to last since the microbiota was not definitively fixed.
(viii) Genetic diversity of hatchery seed can be of importance to bivalve farming, notably in restocking or selective breeding programs. Several sets of multiplex PCR microsatellite markers were developed, allowing parentage assignment. The study of genetic diversity of scallop hatchery seed was performed and compared with that observed in the wild.
(ix) Microalgae production in semi-continuous and continuous cultures in raceways yielded similar concentrations to traditional batch cultures, but showed a longer life time and increased stability of culture. After bibliographic review and some experiments, it was concluded that inert diets are not yet a realistic alternative to live microalgae for commercial hatchery seed production.
(x) A survey on the mollusc hatcheries in Europe provided new insights into this economic sector, which is still small in size and heterogeneous between species and countries. Nevertheless, the hatchery industry is increasing and innovative in many aspects. The dynamics created by the workshop organized for end-users, held at the end of the programme to present the main results, should be pursued, especially through a forum of discussion and the elaboration of programs to implement new techniques at the commercial production scale.

Project Context and Objectives:

Shellfish production in Europe is subject to environmental changes, notably affecting the availability of natural spatfall and favoring the emergence of novel disease outbreaks. Hatcheries offer ways to counter these threats through their potential to provide a more stable supply of seed and to carry out selective breeding to improve disease resistance. To achieve these general objectives, rearing methods may be improved to make them more reliable, to decrease the seed cost and to guarantee its quality. In this context, the strategy of REPROSEED was to combine basic research to improve knowledge on bivalve physiology with applied research to implement new techniques.
Worldwide aquaculture production of the Pacific oyster Crassostrea gigas reached 608 688 tons and a value of 1 303 million US dollars in 2012 (FAO). The yearly European production has fallen to 97000, of which France, as the leading country, produced 83000 tons, (491 million US dollars. Hatcheries and nurseries supply an increasing proportion of oyster seed. C. gigas is the most important species in terms of bivalve hatchery production in Europe, with 7 companies in France, 3 in Ireland, 2 in the UK and 1 in Spain. Today, 25-30% of French seed (i.e. about 1.5 109) is produced in hatcheries, resulting in the direct employment of 250 persons and an annual turnover of around 60 million euros.

The blue mussels Mytilus edulis and M. galloprovincialis are the bivalves with the highest production output in Europe (470 594 tons/year (FAO, 2012). To date, European mussel culture relies entirely on wild seed. In the Netherlands, where 40 000 tons of mussels are produced annually (Kamermans and Smaal, 2002), the supply of seed shows large fluctuations. It is difficult to collect sufficient amounts of mussel seed each year. Spain is the second world producer of mussels and the top European producer, with an output of 203 664 tons of mussels/year (Mytilus galloprovincialis), of which 99 % is produced in Galicia where mussels are a top economic and social resource. Cultures in Spain also depend on a seed supply from naturally seed settlement, which occurs on surrounding rocks. This source is becoming less adequate every year, making the development of mussel seed production in the hatchery an essential complement.

The European clam Ruditapes decussatus is native to the Mediterranean Sea and Atlantic coasts from the Iberian Peninsula up to the southern and western UK. Despite relatively low production in Europe (8200 tons/year; FAO, 2006), these clams are a high value seafood product, economically important in many European countries (mainly Portugal, Italy and Spain). Indeed, the “carpet-shell clam”, as it is also known, is the key bivalve species in Portugal: the culture of this bivalve represents the largest sector of Portuguese molluscan mariculture with over 10 000 people directly or indirectly involved in its production. Aquaculture in Portugal is greatly sustained by the culture of this species in the Ria Formosa Lagoon (South Portugal), where its production represents 34% of national aquaculture production and 80% of shellfish production (DGPA 2006). This clam is more commercially important and better appreciated by consumers than the introduced Manila clam, R. philippinarum. Culture of R. decussatus is clearly limited by the availability of seed. Its production is exclusively based on natural recruitment, which is subject to high annual fluctuations (Matias et al., 2009). Over the last few years, the productivity of the Ria Formosa Lagoon has clearly decreased due to recruitment failures. To address this situation, artificial spawning and larval rearing programs could provide an alternative source of seed.

The king scallop Pecten maximus is a native European species traditionally consumed in a number of European countries. Compared with wild scallop fisheries, aquaculture production is small (60 tons/year: FAO, 2006) and is only undertaken in France, Ireland and Norway. However, this figure underestimates the role of hatchery seed, which this is increasingly used for population enhancement programmes, notably in France and Ireland. It is a highly valued and sought-after product on the European seafood market. However, available supplies are limited due to a decline in the scallop fishery caused by recruitment failure and harmful algae. The market supply of scallops in Europe is therefore strongly dependent on imports from overseas, which presently represent approximately 50 000 tonnes of fresh and processed scallops annually. Increased European production will mainly rely on hatchery seed production. It is mostly the difficulties with larval rearing that have constrained the application of such technology to increase seed production. Some progress has recently been made by the development of flow-through systems for larval rearing but, as with most of the Pectinidae, king scallop production in the hatchery is far from being completely technically controlled and research is therefore still needed.

Compared to the collection of natural spat, which is highly seasonal, hatcheries have the advantage that they can produce seed on a year-round basis. Cultivation methodology differs according to species, practices in a particular country and the production phase. However, seed production relies on some basic principles that follow successive biological steps (Gosling 2003, Helm et al., 2004). First, broodstock is brought into the hatchery for conditioning (i.e. the activation of gametogenesis). This phase lasts approximately two months, at the end of which the bivalves are induced to spawn in a controlled environment and their gametes are mixed to perform fertilisation. Gamete quality is crucial for the success of this first step. Bivalves commonly have very high fecundity, so that a few adult individuals can generate extremely numerous offspring, although reproductive success is highly variable between parental individuals (Boudry et al., 2004). Pelagic (i.e. free swimming) veliger larvae develop within a day or two after fertilisation and are reared in static or flow-through systems for 2 to 4 weeks. During this period, larvae are fed with microalgae. Thereafter, they undergo metamorphosis and settle by attaching to a substrate. The success of this metamorphosis is the next crucial step in the hatchery. Postlarvae are then reared and fed cultured microalgae until they attain approximately 2 mm shell height (micro-nursery phase) and then this seed (also called ‘spat’) are transferred to a nursery based on a cultured phytoplankton supply for growth until they reach the size at which they may be sold to shellfish farmers. In oysters, for example, 12 mm ‘T6’ is the reference size before on-growing at sea. This final controlled phase needs to be optimised by limiting the transfer of physiologically fragile seed (which may show, for example, early gametogenesis or immune deficiency). Also, to avoid the spread of pathogens, it will be necessary to improve the sanitary quality of seed because young healthy shellfish may be carriers of pathogens, such as opportunistic vibrios, which will only infect their host if the seed are weakened.
Production of bivalve seed in hatcheries and nurseries is a relatively new industry, in which most methods have been developed using empirical approaches, adapting methods across species and measuring the resulting effect in terms of growth and survival. Applied research on biochemistry and physiology have led to a better understanding of the effect of biotic (micro-algal diet, bacterial communities, etc.) and abiotic (temperature, salinity, etc.) factors, but most of these studies have only had a limited impact on the industry. To date, a few bivalve species of major aquacultural importance in Europe have benefited from genomic approaches (transcriptome sequencing, microarrays for gene expression studies and 2D-DIGE for proteomics) and from technological innovations (flow-through rearing, automated monitoring, etc.). In the present project we intended to link together these two aspects to improve methodologies so as to develop and optimise new production processes. Studies were performed on the biological basis of key aspects of bivalve reproduction (e.g. conditioning and gamete fertilisation success, larval nutrition and ontology of the immune system, metamorphosis and settlement success, quality of seed) at the molecular level using genomics and proteomics approaches. These complemented the development of new rearing and food production methods (e.g. re-circulating systems, probiotics and low cost algal production).

Despite academic knowledge shown through scientific publications in the field, progress in bivalve aquaculture has relied mainly on empirical approaches. This is particularly true for bivalve hatcheries as this activity has developed relatively recently (» 30 years). Moreover, the factors limiting the development of hatcheries/nurseries have never been considered systematically. These hurdles are often species specific and are encountered at different stages in the biological cycle of molluscs. They mainly concern:
• Broodstock management and gamete quality
• Appropriate methods for larval rearing
• Metamorphosis synchronisation and improvement of settlement
• Quality of seed in terms of immunity, genetic diversity and sanitary status.
Improvements of knowledge in these areas may lead to better hatchery methodology, ensuring reliable seed production and limiting yield variations between months. Moreover, the interaction with the end-users through the survey, meeting and dissemination have initiated the transfer of new technology and thus contributed to the development of efficient European shellfish hatcheries.

The overall objectives of this project have been to secure and stabilise hatchery production of bivalve seed and to decrease production costs through the use of new methods. Because of their economic importance in Europe and the different challenges they present (e.g. different types of reproduction or settlement or differential sensitivity to bacteria at the larval stage), four different species were chosen for the project
- One “established species” in hatchery production
• The Pacific oyster, Crassostrea gigas, is the most important bivalve species produced in European hatcheries, notably since the development of triploids, and is the species for which scientific knowledge is the most advanced.
- Three “emerging species” in hatchery production
• The king scallop, Pecten maximus, also known as the great scallop, is hatchery-produced in France and Norway, although it still presents rearing difficulties. Because of its high susceptibility to bacteria, it is an excellent model for solving certain problems in mollusc rearing (e.g. reducing the need for antibiotics).
• The blue mussels, Mytilus edulis and the closely related M. galloprovincialis, are of major importance for European shellfish industry. There are increasing concerns about seed availability or fluctuating seed quality in the Netherlands, Spain, and France and hatcheries are expected to play a significant role in the near future.
• The European clam (grooved carpet shell), Ruditapes decussatus, is an emerging species, for which there is a demand for hatchery seed but little technological development (in contrast to the introduced Manila clam R. philippinarum).

The European flat oyster Ostrea edulis has been deliberately excluded from this project because other European projects have already focused on pathological or breeding aspects of this species.
The differing biological characteristics of these species mean that the optimisation of hatchery technologies faces different limiting factors (i.e. “bottlenecks”) in their reproduction and rearing. This project brought together major European research groups working on the physiology, eco-physiology, pathology, behaviour and genomics of bivalves and their larvae. These groups worked in collaboration, favouring a complementary approach between species, to answer the following questions:

• How can we better optimise reproductive development in bivalve broodstock?
• What are the processes affecting gamete quality? What are the mechanisms and major factors involved at each stage of their maturation and what criteria can be used to estimate their quality?
• What are the best rearing conditions to fulfil the ecological needs of larvae and seed?
• What are the microbial communities established in the rearing environment? What influence do they have on the performances and how can they be managed?
• How can stress detrimental to the immune status be assessed in larvae and seed?
• How are metamorphosis and settlement regulated and how can they be induced?
• What is the level of genetic diversity in hatchery seed?
• How can research results be transferred to end users, taking into account the barriers between producers and scientists and the limited organisation among commercial bivalve hatcheries in Europe today?

The REPROSEED program was structured according to biological phases and successively dealt with aspects of reproduction, larval rearing, metamorphosis, and seed growth. The establishment of bacterial populations as well as the physiological and immunological development of the animals were investigated using novel genomic tools over the whole process of larval and seed rearing.

Project Results:
WP2 Improvement of broodstock conditioning and gamete quality (WP leader UCBN)
Improvement of knowledge on broodstock management and gamete quality is a key step in improving production of seed of emerging bivalve species. Oocyte quality can be defined as the ability of the egg to be fertilized and subsequently develop into a normal embryo and normal larva. Similarly, spermatozoa quality can be defined as ability to successfully fertilize an egg and subsequently allow the development of a normal embryo. The quality of mollusc gametes from a single individual can be highly variable and is certainly under the influence of physiological and external factors, possibly including the influence of broodstock management practices. The reproductive cycles of two geographically distinct populations of clams, R. decussatus, were studied. Various pre-conditioning or conditioning treatments have been tested on clams, R. decussatus, and oysters, C. gigas, and the gonad development, energy parameters and the spawning inductions have been analysed. Furthermore, broodstock conditioning has also been used in the scallop P. maximus and mussel M. galloprovincialis and, for the four bivalves, fertilization and development protocols have made it possible to estimate D-larvae rates and to define different experimental groups of gamete quality.
Task 2.1. Broodstock management
Deliverables D2.2 D2.3

i) Clam, Ruditapes decussatus (IPMA)
The reproductive cycle, as well as nutrient storage and utilization were studied here over a 24-month period in two grooved carpet shell clam populations from the main production areas of Ria de Aveiro and Ria Formosa Lagoon (Portugal). The reproductive cycle of both populations followed an annual cyclicity that comprised a ripe stage in spring followed by a spawning period that began in late spring and lasted throughout summer until early autumn. This extended and continuous spawning period may be an advantageous strategy that ensures a continuous supply of gametes. Moreover, R. decussatus can adopt different reproductive strategies depending on the geographical origin. The cycles of nutrient storage and utilization showed that clams of both populations present a high reproductive effort that almost depletes its energy reserves. Nevertheless, while the Ria de Aveiro population replenishes them immediately after spawning, the same is not apparent in clams from Ria Formosa Lagoon, with their consequent exhaustion. Also, based on the glycogen pattern it was possible to infer that the Ria Aveiro population is a conservative one, with an emphasis on storage, while the Ria Formosa Lagoon population exhibited an intermediate strategy.
In a first experiment, clams were conditioned at 20±1ºC, under four nutritional regimes: unfed, two mono-specific microalgae, Isochrysis galbana clone T-ISO and Chaetoceros calcitrans and, a mixture of these microalgae. Another group of clams was conditioned at 22±1ºC and was fed the same mixture of microalgae. The gametogenesis process, energy storage and spawning success were all influenced by the nutritional value of the diet received, as shown by the differences in reproductive effort among the single and combined supplementary diets. Temperature must be carefully managed to improve the reproductive conditioning of bivalves: high temperature throughout gametogenesis shortens the time to full ripeness but does not lead to better reproductive output. Lipids are clearly of major importance in conditioning, either as energy reserves or as precursors of tissue structures, with being triacylglycerols the ones which play the most important role. The combination diet at 20ºC is the most profitable for the conditioning of R. decussatus broodstock, since the highest percentage of spawners and average number of eggs released per females was obtained with this treatment. Later, in a second experiment, clams were conditioned under four nutritional regimes, corresponding to: the “Flagellate” group (Pavlova lutheri (P. lutheri) + T-iso), the “Positive flagellate” group (P. lutheri + T-iso + C. cal), the Diatom group (C. cal + Skeletonema costatum or S. cos) and the “Positive diatom” group (P. lutheri + C. cal + S. cos). Although the histological analyses, condition index and biochemical composition showed that the clams fed with the positive diatom group diet attained the ripe stage more quickly, the clams fed with the positive flagellate diet were those that spawned.

ii) Oyster, Crassostrea gigas (UCBN)
The effect of a 5-week pre-conditioning treatment with or without algae before broodstock conditioning was investigated on various parameters. Glycogen storage, estimated on the basis of the appearance of storage tissue aspect on histological slides, was found to be more efficient for fed oysters. Moreover, the pre-conditioning treatment also impacted re-initiation of gametogenesis: fed animals presented more advanced stages than unfed ones. After broodstock conditioning, animals from both treatments were found to be sexually mature on the basis of histological analysis. The influence of diet on quality of the resulting gametes was also qualified using gametes collected from 10 individuals separately (stripping) against a pool of gametes from 10 animals of the opposite sex. For stripped spermatozoa (10 animals), fertilization rate ranged from 80 to 95% whatever the animal, but ranged from 15 to 90% for oysters that had previously been unfed. Motility and velocity of spermatozoa were not clearly different. For stripped oocytes (10 animals), the impact of the pre-conditioning step was observed mainly on D-Larvae rates, with heterologous values ranging from 15 to 90 for unfed oysters (versus 75 to 90% for fed oysters).

Task 2.2 Oocyte maturation and developmental competence
Deliverables D2.1 D 2.5 D 2.6

i) Clam, Ruditapes decussatus (IPMA, Ifremer, UNIPD)
Transcriptome comparison of stripped (less mature) and spawned (mature) oocytes made it possible to identify a set of 439 differentially-expressed genes. Among these, important transcripts encoding regulators of sex steroid synthesis and activity, such as a progestin and adipoQ receptor family member 3 (PAQR3) and steroid 17-alpha-hydroxylase/17,20lyase (CYP450-C17), were expressed at a higher level in the stripped oocytes. Gene expression of enzymes involved in the metabolism of ceramide, a signal sphingolipid thought to influence oocyte maturation and quality, showed significant variations: sphingo-myelin-phospho-diesterase and neutral ceramidase B were expressed at a higher level in the stripped oocytes, while a putative sphingomyelinase transcript was more abundant in spawned oocytes. Interestingly, also a putative vitellogenin (Vg), the major egg yolk protein precursor, was more abundant in released follicles. Moreover, two transcripts encoding proteins suggested to be involved in regulating the maternal mRNA transcription were differentially expressed. Also, the regulation of intracellular calcium levels seemed to highly differentiate the ovarian from the spawned oocytes: putative homologs of C. gigas calcium-activated chloride channel regulator 4 (two probes), regucalcin and calmodulin were more abundant in stripped oocytes, while sodium/calcium exchanger 3 had a higher expression level in spawned oocytes.
Furthermore, transcriptome comparison between two clam populations that have different responses to spawning induction elucidated a specific pathway involved in the recognition signals and binding between the oocyte and components of the sperm plasma membrane that could explain part of the differences in terms of spawning induction success. In addition, sexes and reproductive stages were compared and a correlation between mRNA levels and gonadal area was investigated. The lists of differentially expressed genes revealed that sex explains most of the variance in gonadal gene expression. Additionally, genes like Foxl2, vitellogenin, condensing 2, mitotic apparatus protein p62, Cep57, sperm associated antigens 6, 16 and 17, motile sperm domain containing protein 2, sperm surface protein Sp17, sperm flagellar proteins 1 and 2 and dpy-30, were identified as being correlated with the gonad area and therefore supposedly with the number and/or the size of the gametes produced. Then looking at oocyte quality in R. decussatus based on a microarray-based analysis aimed describing genomic features of oocytes and identifying potential markers of oocyte quality in the species. Oocytes of 25 females estimated by success of D-larval yield under controlled conditions appeared to vary from 0 to 95 %. Microarray analysis revealed 526 probes differentially expressed between good and poor quality oocytes. Some of the differentially expressed probes were identified as being involved in processes such as oocyte protection (DnaJ (Hsp40), Hsp70, Cyclophilin B, PDI), maturation (Cam-PDE1C, PRDM9, G protein), DNA repair (DNA polymerase α), sperm-egg interaction (PDI, G protein) and apoptosis (TNF). The apoptosis pathway assumed an important role in this work as the mRNA level of Caspase 8 was negatively correlated with the D-larval yield. Moreover, a G-protein transcript was identified as more abundant in poor quality oocytes, which could imply importance of prophase I block and meiotic maturation in the oocytes of this species.

ii) Oyster, Crassostrea gigas (UCBN)
Characteristics of oocytes were compared in relation to sampling location along the antero-posterior axis of gonad. When incubated in sea water, the morphology of oocytes changed (area, perimeter and circularity) without strong differences between sampling locations (except for proximal location). Transcriptomic and a proteomic approaches were undertaken to identify proteins associated with quality of oocytes. Few transcriptomic differences were observed whereas proteomic results revealed 10 distinct proteins up-accumulated in low quality oocytes, including vitellogenin-breakdown products and metabolic enzymes. Eight up-accumulated spots from high quality oocytes revealed 6 distinct proteins, including chaperone molecules and cell-cycle control proteins.

iii) Scallop, Pecten maximus. (Ifremer, UNIPD)
Spawning of oocytes may be induced by thermal shock. D-larval rate observed for spawned oocytes was clearly higher than for stripped oocytes, confirming the existence of a maturation process along the genital ducts. Transcriptome comparison of stripped (less mature) and spawned (mature) oocytes allowed to identify a set of 417 probes that were differentially expressed. Among the differentially expressed genes, important transcripts encoding proteins putatively implicated in gamete maturation, such as IQ domain-containing protein H, protein naked cuticle homolog 1 and RING finger protein 17, were expressed at a higher level in released oocytes. Moreover, gene expression of SPRY domain-containing SOCS box protein 3 (involved in ubiquitination processes), DnaJ homolog subfamily B member 13 (involved in apoptosis inhibition), U5 small nuclear ribonucleo-protein 40 kDa protein (a regulator of mRNA storing processes) and cyclin-dependent kinase regulatory subunit 1 was higher in released oocytes. Among the transcripts more abundant in stripped oocytes, putative vitellogenin (Vg), galectins, formin-2, molluscan insulin-related peptide(s) receptor and protein hu-li taishao (Adducin-like protein) were detected.

iv) Mussel, Mytilus galloprovincialis (CSIC)
Transcriptome analyses were performed in relation to oocyte quality. Sequence of mussel cyclophilin, HSP70 and PDI were identified and no differences were observed in their expression. However, when the expression of some antimicrobial peptides or the C1q gene were measured, a significantly higher expression was seen in high quality oocytes compared with bad quality ones.

Task 2.3 Spermatozoon maturation in C. gigas and P. maximus
Deliverables D 2.1 D 2.6
i) Oyster, Crassostrea gigas (UCBN)
Characteristics of spermatozoa were compared in relation to sampling location along the antero-posterior axis of gonad. No difference could be observed in spermatozoa viability, motility and velocity according to the location. A proteomic approach was undertaken to identify proteins associated with maturation of C. gigas spermatozoa. Most of the up-regulated proteins in mature spermatozoa were related to the acquisition of motility potential, energy metabolism and acrosomal reaction. This included proteins involved in flagellum formation and control (Actin-1, Tubulin beta-4B, Tektin 2, Tektin-4, Ropporin-1 and Dinein light chain), energy production (Isocitrate dehydrogenase, ATP synthase subunit D, Pyruvate dehydrogenase E1 component subunit beta, Hydroxyacyl-coenzyme A dehydrogenase, Cytochrome c oxidase subunit 6B) and acrosome reaction (Proteosome subunit beta). Some of the differentially expressed proteins do not have an identified function. These candidate proteins provide a good opportunity to understand the maturation processes of spermatozoa in C. gigas and at a larger scale in other bivalve species.

ii) Scallop, Pecten maximus (Ifremer)
Spawning induction with intragonadic injection of serotonin succeeded to induce male gamete spawning. Motility, velocity and ATP content of spawned spermatozoa induced with serotonin were highly increased in comparison with testicular spermatozoa confirming that capacitation of spermatozoa occurs in the gonoduct. Different spermatozoa sampling locations were compared along the gonoduct from the distal part to the gonopore, showing that motility and velocity increased along this axis. Moreover, incubation in seawater containing caffeine or serotonin may activate spermatozoa capacitation. In both species, differentiated proteins are identified in spawned and stripped spermatozoa. Most of them are linked with flagellum architecture and involved in sperm motility (i.e. Tektin2 in both species) or energy metabolism and mitochondrial activity. Some proteins linked with energy metabolism are under expressed in scallop stripped spermatozoa but not in oyster leading to postulate that the acquisition of energy is essential for capacitation of scallop spermatozoa. These data are in good agreement with differences in ATP content in spawned versus intratesticular spermatozoa.

Task 2.4 Control of gamete release
Deliverable D 2.4

i) Clam, Ruditapes decussatus (IPMA)
The results obtained demonstrate that chemical inducers like fluoxetina and hydrogen peroxide were not effective to induce spawning in R. decussatus. In contrast, the traditional stimulus of fast temperature change, consisting here of alternated cycles of 1 hour at 28ºC and fifty minutes at 8ºC with addition of sperm and microalgae were shown to be the most effective. However, it was possible that if some males still spawned in the tank, the percentage of spawners was higher.

ii) Scallop, Pecten maximus (Ifremer)
In males, intragonadic serotonin injection is a more efficient technique than thermal shock to induce sperm shedding without modifying sperm quality, regarding the percentage of motile spermatozoa and their velocity. In females, stripped oocytes cannot be used for further fertilization, confirming the existence of an oocyte “maturation process” along scallop female genital ducts. Contrary to the effect on males, thermal shock or serotonin injections failed to induce spawning of oocytes.
iii) Oyster, Crassostrea gigas (UCBN)
The most efficient spawning inducer (serotonin 50mM) showed a limited efficiency, with 3 to 4 time less oocytes collected compared to gonad stripping. In natural conditions, gamete release could be performed by the association of neuropeptides, gonadal peptides and polypeptides as regulatory molecules. In order to identify the regulatory peptides involved in gamete release in the oyster Crassostrea gigas, a set of 19 neuropeptides and gonadal peptides were identified by a screening performed in tandem mass spectrometry on mature and immature gonads of both sexes. From gonadal peptides, at least three polypeptides cleaved from a single vitellogenin were identified. This was complemented by a genome-wide analysis, which led to the identification of at least 46 neuropeptide-encoding precursors. An egg-laying hormone (Cg-ELH) was identified and its sequence reveals a precursor of 166 amino acids including two putative ELH mature peptides of respectively 40 and 38 amino acids.

WP3 Optimization of production of veliger larvae,

Task 3.1 Development of recirculating systems
Deliverable D3.1
i) Clam, Ruditapes decussatus (IPMA)
The effects of six nutritional regimes were evaluated on the survival, growth and biochemical composition of R. decussatus larvae, with the aim of providing crucial information on its nutritional requirements. A holistic approach incorporating all physiological responses showed that the bispecific diet I. aff galbana and C. calcitrans (60/40 cell. µL-1) was the most adapted throughout larval development. Moreover, the monospecific diet I. aff galbana provided an overall good performance. Larvae cannot use C. calcitrans at their early stages of development; however, the inclusion of these microalgae improved late larval development. The growth and survival rates of the European clam R. decussatus larvae were compared when reared in a recirculating system (RAS) and a “batch” system. The RAS system was tested with 50% and 10% of water renewal and a water flow of 240 mL min-1 and compared with the batch system. The assays were performed with an initial density of 10 larvae mL-1 and fed with 60 cell. mL-1 Isochrysis aff galbana and 40 cell. mL-1 Chaetoceros calcitrans per µL. The physical and chemical values remained stable throughout the experiment, except for dissolved oxygen, which decreased in all rearing systems and nitrates, which generally decreased in RAS but increased in “batch”. Total bacteria of larvae increased exponentially in all systems, with a higher incidence in "batch". However, Vibrionaceae were more frequent in larvae from RAS10. The larval survival at the end of the experiment was low for all rearing systems. However, the growth rate was significantly higher in RAS50 than the others. In RAS (50 and 10) at 13th day 90% of larva had developed a “foot”, however, in the “batch” culture, the larvae foot appeared but in a lower percentage and at the 17th day. These results show that it is possible to rear R. decussatus larvae in a recirculating system with better results than in traditional batch system.
ii) Scallop, Pecten maximus (UiB)
Two experiments using water recirculation technology (RAS) were performed in large scale systems (3500 L larval tanks) supplied with a continuous addition of algal feed, and 20% renewal of seawater. In the RAS, a gradual increase in CO2, decrease in pH and dissolved oxygen was observed over time. This was most obvious during the second of these experiments, when the total organic carbon content increased in both FTS and RAS. The total bacterial number was lower and more stable in FTS than in the RAS. The variations in seawater quality parameters were smaller during the first experiment compared to the second, where values of oxygen saturation were reduced to <70%, pH was 7.8 and NO3− reached 5 mg L−1. Even though these changes would seem less beneficial for survival and growth of scallop larvae, results showed that the survival at the end of the larval stage was higher in the FTS, but the yield of competent larvae ready for settlement was not significantly different (p > 0.05) due to large variations between tanks. The CV% was 28.9% in FTS, while it was 49.9% in RAS. In FTS the mean yield was 40.2%, while it was 26.5% of initial number of larvae in RAS. Large variations in survival and yield were found between the larval tanks as well as gradual reduction in pH and oxygen in RAS tanks. The results indicate that there is a large potential for reduction in water use (80%) by the use of recirculation technology.
iii) Oyster, Crassostrea gigas (Ifremer)
In C. gigas, the first trial run in 2010, based on the recirculation of 50% seawater, treated on bio-reactor and UV for C. gigas larva, was shown to be quite efficient, and similar larval development was found compared to FTS. Similar trends were confirmed in 2011 with two different seawater renewal rates 0 and 25%). When seawater recirculation was maximum (100%: for RAS 0) larvae performed as well as those reared in FTS, with similar survival, competence, growth and settlement. In 2012 at 10% RAS, at 50 larvae ml-1, no significant differences were found for larval survival (85% vs 95% in FTS The growth rate in RAS was slowed down (13.2 µm d-1 vs 15.3 µm d-1) leading to a lower number of competent larvae on day 16 (36% vs 58% in FTS) with an algal consumption correlated with the growth rate. This lengthened the rearing by 2 days to reach similar percentage of eyed larvae (60%) and, despite similar larval competence, metamorphosis occurred 11 days later at a lower rate (25% vs 50% in FTS). Lastly, the hydrological parameters without fresh seawater addition derived slightly (pH fell to 7.8 and salinity increased to 38%) but not in RAS 25%. The ammonium and nitrites were less 1mg L-1 and nitrates reached 12 mg L-1 in RAS 0%. The next experiment aimed to determine the limit of the larval densities accepted by RAS with 10% fresh seawater addition per hour compared with FTS. Three densities were tested in both systems: 50, 150 and 300 larvae mL-1. The highest survival and growth rates were obtained in FTS at 50 larvae mL-1 while equivalent performances were recorded at 150 larvae mL-1 in both systems. At 300 larvae mL-1 mortalities increased in one replicate of RAS (60%) and in two replicates (100% and 90%) of FTS. Finally the RAS seems to play a role of “buffer” with less of a gap between extreme densities. For C. gigas larvae, RAS at 50 larvae mL-1 and 10% fresh seawater addition can be used, although it is less efficient than FTS, mainly at metamorphosis.
iv) Mussel, Mytilus edulis (IMARES)
M. edulis larval growth and survival were tested in three identical recirculation systems (Filter Rack S), flow through (FT) and batch cultures. In each rearing unit, 10 larvae.mL-1 were grown for a period of three weeks (prior to metamorphosis). Mussel larvae showed highest survival percentage in the batch treatment and lowest larvae survival in RAS. Length at the end of the experiment was higher in the flow through, followed by batch and RAS. No differences between treatments were found in pH, oxygen, salinity and NH4 concentrations. Temperature in the RAS was 3ºC throughout the experiment, which is higher than in batch and flow through. The larval biomass in RAS was too low to maintain a good functioning biofilter, which resulted in high NO2 concentrations at the beginning of the experiment. This was also indicated by low ammonium values. The effect of different NH4, NO2 and NO3 concentrations on the development of D-larvae was evaluated. No D-larvae were found above concentrations of 3.3 mg NH4/L and no effect was found for the different NO2 and NO3 concentrations tested.
Task 3.2 Study of microbiota and test methods
1) Study of microbiota
Deliverable D3.2
i) Oyster, Crassostrea gigas (Ifremer)
The microbiota of bivalve larvae and their environment were analysed by pyrosequencing during the three experiments in RAS and several new insights can be drawn from these analyses. The first is the great diversity of bacterial species that occurred in the different compartments of RAS (seawater, biofilter, larvae). The bacterial assemblages were complex and generally composed of 50 to 200 OTUs (Operational Taxonomic Units at 97% similarity). However, in some cases, predominant bacteria were established at a high level in a generally transient manner. The second main finding was that microbiota was niche specific. The similarity did not exceed 30% between the microbiota of each niche at the same sampling date. It tended to be homogenous regardless of the rearing conditions The third main finding is that the bacterial assemblages varied as a function of age. Thus, different bacterial assemblages succeeded mainly in seawater and larvae while composition remained more stable in the biofilter. γ-Proteobacteria were poorly represented, with especially low levels of Vibrio spp in the biofilter and seawater (<0.1%) and still less in larvae. Some bacterial species were sometimes predominant, such as Ruegeria sp, Phaeobacter sp, Devosia sp Delftia sp, Bacteriovorax sp, Bdellovibrio sp. These were found systematically from one experiment to another. Some are durably established in algal culture. They could represent a risk for the safety of hatchery rearing if they exerted detrimental effect.
Members of Burkholderiales were established in the larvae at the end of larval culture or just after the metamorphosis and seem to constitute a group firmly linked to oyster postlarvae. Bacteriovorax spp and Bdellovibrio spp, predatory bacteria, occurred systematically in seawater and in algal culture but not in other niches. These may participate to the control of bacterial proliferation and prevent vibrio development.
The role of the biofilm could also be harmful if it develops excessively, as observed in the Flow Through system with 300 larvae mL-1. During its formation, the α-Proteobacteria, especially Rhodobacteraceae, colonized the substrate first, then filamentous bacteria including were established and tended to cover them. These bacteria can be suspected of having negative effects since their abundance coincided with mortality in FTS and in RAS while a few vibrios were present.
In the biofilter, the autotrophic nitrifying bacteria were at very low level or absent (ammonia oxidisers). Possibly, heterotrophic nitrifiers could replace them, as described in some publications, but this means of nitrification has rarely been reported in the natural environment and never in aquaculture. Finally the small quantity of ammonium produced by larvae may be more greatly eliminated by the fresh seawater addition than by nitrification.
The sources of bacteria that colonized the different niches are diverse. Bacteria came from broodstock, from algal culture and many others probably from aerosol and seawater.
ii) Scallop, Pecten maximus (USC/ UiB)
Experiments were performed comparing the microbiota associated with scallop larvae in RAS and FTS systems. Differences were observed between the two systems in the diversity of the bacterial populations, as well as in the taxonomic groups present in the different compartments. In addition, a clear evolution of the microbiota was observed in both systems through the time. In RAS, a decrease in the bacterial diversity was observed through the experiment. Some taxa were present in all compartments but others were exclusive to some compartments in the hatchery. These results were confirmed when DGGE and pyrosequencing (metagenomic approaches) were employed for the analysis of the broodstock microbiota.
iii) Clam, Ruditapes decussatus (USC)
Studies on the microbiota of carpet-shell clams (R. decussatus) in FTS were carried out in collaboration with the subcontractor CIMA and the private hatchery Ostreira S.L.. The effects on the bacterial populations of different conditioning protocols (different temperatures and food regimes) of the broodstock and the use of antibiotic were evaluated.
A number of Vibrio isolates from broodstock and the larvae were studied, most of which were identified as species belonging to the Splendidus clade..Some of them are pathogenic for larvae and not previously recognized,. Interestingly, similar strains in both groups had been detected in the larvae as well as in the broodstock. This fact supports the hypothesis of vertical transmission of bacteria from adults to larvae, reinforcing the importance of the broodstock conditioning to control the entry of opportunistic pathogens into the distinct hatchery compartments.
The conditioning of broodstock at 20ºC without food for at least 10 days, demonstrate to be a good way to reduce the microbiota in them and improves the survival of larvae. That reinforces the results obtained in Task 2.1 (IPMA)
However, the use of antibiotics, enhanced the fast appearance and dominance of resistant strains, despite the fact all of the isolates from broodstock were sensitive to chloramfenicol.

2) Probiotic methods for controlling the microbiota
(Deliverable D 3.3)
1) Oyster, Crassostrea gigas (Ifremer)
In order to enhance larval culture, three probiotic bacteria (Phaeobacter gallaeciensis, Pseudoalteromonas sp, Vibrio sp) were added at 105 cells mL-1 final concentration to two RAS with 10% fresh seawater addition (in one RAS two times during the first week and in the other RAS every 48h througthout the rearing ). The effect of these probiotics was negative, since they decreased the growth and survival rates by around 20%. That proved that the microbiota can significantly influence the larval physiology. The bacterial composition shifted after the distribution of probiotics. A convergence of microbiota occurred between all RAS with and without probiotics in the second week, accompanied by reestablishment of the growth rate in RAS receiving probiotics.
2) Scallop, Pecten maximus (USC, UiB)
Probiotic mixtures were added 3 times a week. Larvae were settled in sieves in FTS tanks containing 30 sieves each. Salinity, temperature, pH, ORP, fluorescence, nitrate, nitrite and ammonium were estimated. Results showed that larval growth was good in all larval tanks. Best growth was observed in the standard FTS tank, but all tanks with probiotics also showed good growth. Best larval survival was observed in the tank receiving treatment with probiotics B and C. Survival after the larval stage was between 50 and 60 % and the fraction ready to settle was between 30.6 and 56.5 %. Yield of larvae treated with probiotic B was significant higher than all other groups. The larvae from standard flow through (FTS) had significant higher settlement than larvae given probiotic B, but similar settlement to larvae given probiotics A and C. These results showed that the high potential for use of probiotic added to larval tanks is a promising new development for rearing sensitive larvae..
A patent is being prepared as soon as possible
Task 3.3 Study of immunology and physiological status of bivalve larvae (CSIC, UNIPD)
Deliverables D3.4 D3.5 D3.6

The knowledge of the larval and immune transcriptome of the four studied species was increased by Next Generation Sequencing. Roche 454 and RNA sequencing were employed to obtain the sequence of a large number of transcripts expressed in adult tissues, haemocytes, oocytes and larvae. Once the transcriptome scaffolds had been built and annotated by sequence similarity searches against model species, a microarray platform (8x60K) was constructed for Mytilus galloprovincialis, Ruditapes decussatus and Pecten maximus.
A detailed analysis of genes expressed in haemocytes was carried out in order to improve the knowledge of bivalve immunology. Identification and annotation of the most relevant immune-related genes allowed us to evaluate their expression during larval development of the four bivalve species, providing a full picture of immune system ontogeny.

i) Challenge of scallop larvae by mechanical stress (bubbling) and a pathogen Vibrio pectenicida (UNIPD, Ifremer)
The genes expressed during these challenges were examined by pyrosequencing. Despite the regulation of important immune transcripts such as those encoding a Baculoviral IAP repeat-containing protein 7, Big defensin and Macrophage mannose receptor 1, results indicated the involvement of a mRNA encoding Collagen alpha-4. Collagen is an abundant protein that is part of the extracellular matrix and deposited or broken down as part of the process of tissue growth and repair, thus suggesting that larval development and growth can be affected by the presence of V. pectenicida or bubbling.

ii) Ontology of immunity (CSIC)
The genes obtained after massive sequencing of the larval transcriptome of the 4 bivalve species in D3.4 were searched using the immune-related genes identified and annotated in D 3.5 to obtain a full picture of ontogeny of immune system.
After identifying the immune genes in the haemocyte transcriptome (D3.5) a list of keywords was constructed with the terms observed in the bivalve haemocyte genes and also using the genes from relevant immune pathways present in other organisms.
For each species, a oneway ANOVA analysis was carried out by considering the expression profiles of all the collected stages: D-larvae, ready to settle larvae, settled larvae and seed. A probe was considered differentially expressed when the adjusted p-value was lower than 0.01.
Finally, among the set of differentially expressed genes (DEGs), the list of “immune” key words was searched, allowing us to identify the expression profiles of important immune genes during development. The expression values of DEGs in each developmental stage are reported in the tables in the deliverable (D. 3.6).
Experimental infections carried out by the USC team, using clam larvae and different bacterial strains isolated from them, both pathogens and non-pathogenic ones, allowed to detect differential expresion of some immune genes in clam.

Work package 4: Improvement of metamorphosis and settlement

Task 4.1 Metamorphosis inducers (BU, USC, IPIMAR [now IPMA])
Deliverable D 4.1
There are difficulties in the culture of bivalve molluscs associated with the transition from larval to juvenile stages, with protracted metamorphosis, low development synchronicity and variable survival. The use of exogenously applied chemical agents, as demonstrated in other bivalve species, has been viewed as a means of resolving these issues. In order to improve control over the timing of metamorphosis in bivalve cultures, a variety of biotic and chemical inducers have been used to induce settlement and metamorphosis in larvae of different bivalve species.
REPROSEED examined a number of potential chemical inducers on competent larvae of M. edulis, P. maximus and R. decussatus as well as experimental manipulation of fatty acid profiles in the case of M. edulis and a biotic inducer for R. decussatus. Following completion of the experiments on M. edulis and P. maximus, chemicals tested were considered for their suitability for application in commercial trials.
A common experimental protocol was adopted across the partners, which was modified as required during the trials. A range of concentrations for each chemical were tested over exposure periods of 24, 48 and 72 hours and rates of metamorphosis were assessed after these periods and after a further 4 days (i.e. 7 days after the commencement of the trial).
In the case of M. edulis, larval susceptibility to these chemicals proved to be limited as none of the assessed chemicals (L-DOPA, epinephrine, KCl, GABA (γ-aminobutyric acid) and NH4Cl) induced a significant lasting improvement in metamorphosis compared with the untreated control. High concentrations generally proved to be toxic and led to lower levels of development. Excess K+ did however induce a slight improvement in larval growth at concentrations of 1.3 mM and 13.4 mM (1.3x10-3 and 1.3x10-2 respectively) after exposure for 24 and/or 48 hours. Manipulation of dietary fatty acid profiles did not result in significant improvements in metamorphosis rates. The results of the present study provide additional information on the role of chemicals in this species and the genus as a whole; however, a continued search for suitable agents is required.
The effects of the chemicals KCl, NH4Cl, acetylcholine chloride, GABA, L-DOPA and epinephrine on the induction of larval metamorphosis and larval toxicity were examined for P. maximus. A range of concentrations for each chemical, applied over a 48-hour period, were assessed. Larval metamorphic response was low, and concentration-dependent, whilst all chemicals proved toxic within the range of concentrations tested. Among the tested chemicals, KCl at 20mM and L-DOPA at 10-6M induced significantly (P<0.05) higher rates of larval development, improving development by 208% and 128% respectively compared to the controls after 1 week. Whilst the KCl treatment was toxic, reducing survival by 33% compared with the control (P<0.05) the L-DOPA treatment was found to significantly increase survival by 49% compared with the control (P<0.05). Furthermore, the influence of these two chemicals on larval development varied, with KCl only promoting dissochonch growth and L-DOPA only promoting gill development, suggesting that the pathways influenced by these two chemicals may be distinct.
The effect of the chemical inducers epinephrine and GABA at different concentrations and exposure times on R. decussatus metamorphosis were evaluated at concentrations of 10-2, 10-4 and 10-6 M. Based on the results obtained with different concentrations of GABA, another trial with GABA at a concentration of 10-2 M was conducted to evaluate the effect of exposure time (1, 2 and 3 days). Epinephrine and GABA were both identified as active metamorphosis inducers for this species; however, GABA also caused larval mortality. All concentrations of epinephrine induced higher metamorphosis when compared to a control after 3 days, with maximum metamorphic rate achieved with exposure to 10-4 M and 10-6 M epinephrine; double that of the control larvae. At the end of the experiment (day 7), all larvae had undergone metamorphosis. Nevertheless, no significance difference were observed (P>0.05) probably due to the high variability observed between replicates. In GABA, concentrations of 10-4 M and 10-6 M with exposure periods of 3 days, did not present a significant increase of metamorphosis rate compared with the control. However, a higher concentration of GABA (10-2 M) had a positive effect on metamorphosis, with an optimum exposure period of 2 days. GABA at this concentration also resulted in an increase in larval mortality at day 7 compared to the control, with mortality increasing with the length of exposure period. Thus epinephrine proved to be the most effective chemical inducer for R. decussatus with maximum metamorphic rate observed at 10-6M and no increase in mortality rate, this could be considered an active and fast metamorphosis inducer for this species.
For all species it is important to note that larval mortality arising from chemical treatments did not always happen within the exposure period, and therefore a minimum trial length of 7 days from initial exposure is recommended.
The potential application of chemical inducers at the commercial scale was considered for M. edulis and P. maximus. In the case of M. edulis, commercial scale production ceased after the start of REPROSEED and this, combined with the lack of significant results, meant that no commercial trials were feasible. In the case of P. maximus, the use of KCl was carefully considered at a meeting between BU and SCALPRO, but closer analysis revealed that improved metamorphosis was accompanied by an increase in mortality that was considered an unacceptable risk; therefore, commercial trials were not taken past the development of a protocol.

Task 4.2 Seed behaviour (BU)
Deliverable D4.3
In this task, mobile seed behaviour of M. edulis and the impact of environmental conditions on the attachment and detachment of juvenile P. maximus were investigated with the aim of increasing the success of settlement (in the case of M. edulis) and ease of management and culture.
M. edulis possesses highly mobile early settlement stages able to relocate to new habitats; however, the factors affecting seed retention and dispersal are yet to be fully understood. This study assessed the impact of substrate, attachment period, food availability, water agitation, and density on attachment and retention of seed mussel between 0.97 and 3.85mm in shell length. Attachment was affected by both substrate type and attachment period. Seed showed a clear preference for smooth glass and natural fibre wool felt, with attachment rates of up to 89.2% and 95.2% after 6 hours, respectively, with seed requiring at least 3 hours to attach, after which there was no significant difference (P>0.05). Attachment to other natural fibre materials and slate was lower, though not necessarily significantly (P>0.05). Feeding, water agitation, density and substrate all significantly influenced seed retention, though not mortality, over the course of the study. Feeding seed significantly increased detachment and dispersal compared to seed receiving limited or no food (P<0.05). Fed seed were more likely to undertake active dispersal, with ≥75% of detachment associated with dispersal to alternative attachment sites. However, seed on restricted feeding regimes, although less likely to detach, were more likely to remain in situ, since <40% of detached seed relocated. Water agitation significantly increased the subsequent retention of individuals (P<0.05) and negated the influence of feeding seed, decreasing dispersal behaviour to alternative attachment sites by between 62 to 100% and total detachment by 55 to 100%. The stimulation of increased and stronger attachment by water movement dramatically increased the likelihood of retention. Results showed that seed detachment and dispersal was proportional to seed density, with detachment rates of between 0.16 to 0.07 seed cm-2 and dispersal rates of 0.06 to 0.02 seed cm-2. However, the rate of detachment and active dispersal was significantly influenced by substrate type, with the rate of loss higher from the smooth glass than from the fibrous wool material (P<0.05). From an aquaculture perspective, substrate type has a significant impact upon stocking density, with some substrates offering higher retention of more seed. The present study also demonstrated that although pedal crawling could not be used as a means of predicting either attachment or subsequent retention, a link between seed activity and seed mortality was identified; seed displaying increased mobility exhibited higher survival, with 80 to 100% lower mortality over our study than sedentary seed. This indicates that seed activity could be used as a simple predictor for seed performance, with activity related to level of fitness.
Environmental factors play a key role in the control of byssus attachment and detachment of many bivalve species. For P. maximus, the impact of substrate type, substrate pre-conditioning, attachment period, feeding ration, agitation and illumination on attachment of juvenile scallops between 1.9 and 7.2mm in shell length was assessed. Testing a variety of substrates including slate, nylon mesh, glass, and the natural fibre materials, wool felt, hemp, cotton and soya, showed that scallop juveniles have a preference for a hard textured (slate) surface, with mean attachment rates up to 75.6±14.4% after 24 hours. Periods longer than 1 hour are essential to allow juveniles sufficient time to attach, as significantly higher attachment to all substrate material occurred after 24 hours (ANOVA P<0.05). Attachment was boosted by pre-conditioning substrates in flow-through tanks of unfiltered seawater for 1 to 2 weeks, with significant improvements seen on substrates conditioned for 2 weeks over clean substrates (Tukey P<0.05). However, parameters including agitation (static vs agitated conditions), illumination (24 hours illumination vs 24 hours darkness) and feeding regime (starvation, 0.025 g, 0.05 g and 0.15 g organic weight of microalgae g-1 spat (live weight) week-1) had no impact on juvenile attachment. The impact of substrate type, substrate pre-conditioning and feeding ration on detachment of juvenile seed was also examined in a benthic water flume over increasing velocities (0, 7.2±0.1 10.1±0.2 12.2±0.1 and 12.6±0.2 cm second-1). In each experiment, detachment increased with increasing water velocity. However, of the substrates examined (wool, nylon and slate), although retention was variable over the course of the study, the hard textured surface (slate) ensured the greatest chance of retention, with attachment up to 100% at 12.6±0.2 cm second-1. Comparison of substrate pre-conditioning showed that under increasing water velocities conditioning for 2 weeks significantly compromised juvenile retention (Tukey P<0.05). Feeding regime also impacted retention, with detachment highest by starved animals and retention significantly higher in juveniles fed 0.025g microalgae. Feeding higher rations showed no improvement in retention over starved juveniles or juveniles fed just 0.025 g. Based upon the current findings, recommendable parameters for maximising juvenile P. maximus attachment and retention are; water velocities up to 12.6±0.2cm second-1, utilising a slate type substrate pre-conditioned for 1 week, having fed juveniles a diet of at least 0.025g microalgae.g-1 juveniles.week-1 and allowing at least 24 hours for attachment.

Task 4.3 Genes involved in changes through metamorphosis and settlement (BU, CSIC, UNIPD)
Deliverables D4.4 D4.5
In order to make a reliable microarray tool for each species, it was essential to obtain high quality gene sequences with confident annotation. For this reason, at the beginning of the project, all partners agreed to place additional effort on sequencing to obtain a good representation of the transcriptome of each species, especially of the larval stages. The goal of this additional sequencing was to increase the transcriptome characterization of specific developmental stages and cell types.
In the studied species, adult tissues included muscle, foot, mantle, gill, male and female gonads and digestive gland; larval developmental stages included sperm and oocytes in addition to trochophore, veliger and pediveliger larvae, settled larvae and one-month-old spat. Immune stimulation of haemocytes was carried out following an in vitro approach with Pathogen Associated Molecular Patterns (PAMPs) -Poly I:C, LPS, Peptidoglycan, Lipoteichoic acid, CpG and Zymosan, and an in vivo approach injecting mussels with heat-inactivated Vibrio anguillarum, and subsequently sampling stimulated haemocytes.
The samples were preserved and RNA was isolated and purified. Concentration and purity of RNA were measured and 1 μg of RNA was pooled and used as the source of starting material for cDNA synthesis. The sequencing was conducted by means of two Next Generation Sequencing technologies: Roche 454 sequencing and Illumina RNA sequencing.
All the sequenced reads were used to construct 4 species-specific scaffold transcriptomes and the putative function of each contig was obtained by BlastX similarity searches against UniProtKB/SwissProt database and several protein databases available on Ensembl Genome Browse. Once the transcriptome scaffolds had been annotated, a microarray platform (8x60K) was constructed for three of the four target species (M. galloprovincialis, R. decussatus and P. maximus).
High mortality of bivalve larvae occurs in mollusc aquaculture at the settlement and metamorphosis stage. For this reason, we tried to identify candidate genes in P. maximum, M. edulis and R. decussates and also in M. galloprovincialis (used as a model and because the availability of samples) that can be involved in changes through metamorphosis and settlement. Several different families were sampled, with 3 biological replicates of stimulated larvae or controls used at each larval stage. Larvae were concentrated, RNA isolated and purified, before 100 ng of each sample was used to obtain cDNA. Specific PCR primers were designed from the selected sequences according to qPCR restrictions. Real-time quantitative PCR was performed in the 7300 Real Time PCR System (Applied Biosystems). All reactions were performed as technical triplicates.
Before the molecular tools developed in this project were available, we selected a set of several immune genes previously identified and characterized in mussel haemocytes (Balseiro et al., 2011; Romero et al., 2011) to follow their expression in the different M. galloprovincialis larval stages, especially at metamorphosis stage, both at constitutive level and after an immune treatment. The studied genes were Myticin C (Myt C), Mytimycin (MMG1), Macrophage migration inhibition factor (MIF), Myticin B (Myt B), Complement C1q (C1q), Lysozyme; Mytilin B (Mytl B), Membrane Attack Complex/perforin (MACP); Fibrinogen-related protein (FREP). When the constitutive expression of selected immune-genes were examined, with results related to oocyte gene expression, antimicrobial peptides, Myticin B and Mytilin B, appeared up-regulated from the metamorphosis stage. MacP and fibrinogen-related protein also were regulated but not in settled larvae. The precursor of mytimycin 1 and lysozyme were down-regulated. Following infection of larvae in metamorphosis with live or heat-inactivated V. anguillarum, it was found that V. anguillarum did not affect the gene expression of larvae in metamorphosis, while live V. anguillarum induced a strong down-regulation of the antimicrobial peptides Myticin C, Myticin B and Mytilin B.
Having designed the microarrays for each species, the gene expression modulated at the metamorphosis was studied. Genes involved in cell cycle (up modulated in oocytes) are now down modulated (Mitotic apparatus protein p62, Nucleoplasmin-like protein ANO39). Comparing the Gene Ontology terms associated with the most expressed genes before and during metamorphosis, revealed that the “immune system process” seems to start to be expressed at this stage, in agreement with what we observed previously. Other processes such as “cell proliferation” were up modulated already in veliger larvae and this was found to continue in the pediveliger stage although we can see it also as a down-modulated process. This could suggest that some genes related to proliferation have started to be less expressed compared with the previous stages. We focused our attention at the immune genes because we can see that at pediveliger stage, they seem to be up and down modulated. Genes such as macrophage expressed 1, Beta-1,3-glucan-binding protein and Serine protease inhibitor Cvsi-2 were up modulated. This indicates that immune genes start to be expressed at this stage acquiring a more mature immunocompetence. However, when we analyse the normalized data, a gene family that is specifically down regulated at this stage was detected: Myticin B. After the pediveliger stage, myticins start to be highly expressed, reaching maximum rates in adults. Looking at the genes with an increment in this stage, several are related to tubuline (tubulin beta-5 chain, Tubulin beta chain, Os11g0247300, Os03g0718100)..

WP5 Low cost production of high quality seed

Task 5.1 High density seed production systems (Ifremer, IMARES, UiB, IPMA)
Deliverables 5.1 5.2 5.3 5.12

For a successful implementation of high density seed production systems, the relation between algal densities, seed size, seed density on the one hand and algal uptake rate on the other was determined and the waste matrix calculated. Use of recirculating aquaculture systems (RAS) was tested at the commercial scale for scallop seed by Scalpro and oyster seed by Satmar. For the other two species, there was no commercial hatchery available. Thus, the tests were performed at the experimental scale by IMARES for mussels at (32 L rearing tanks) and IPMA for clams (0.5 L rearing tanks).

i) Oyster, Crassostrea gigas (IFREMER, Satmar)

A mixed diet of two algal species (Isochrysis galbana (T) and Chaetoceros gracilis (Cg) was shown to be a good ration for C. gigas larval development, although diet information for oyster spat was lacking. The effects of five phytoplankton assemblages were tested for C. gigas spat: 100% Cg, 25%T/75%Cg, 50%T/50%Cg, 75%T/25%Cg, 100%T. The best diet for C. gigas spat was a balanced mixed diet 50%T/50% or a diet with more diatoms 25%T/75%Cg.

Mass balance calculations for oyster spat were performed including tank size and biomass, TAN (total ammoniacal nitrogen) mass balance calculations, biofilter sizing calculations, solids mass balance and oxygen mass balance calculations. Based on this, an estimated required flow rate of 0.3 l min-1 was found for oyster spat in a 0.5-L system.

Experiment in 0.5-L rearing tanks
Fourteen-day-old post set spat was added in March 2013 to each unit at 6 g total biomass (≈ 6500-8500 individuals) and reared in RAS (recirculation system). Spat was reared in RAS at three fresh seawater addition rates, 20%, 10% and 5%, and compared to a FTS (flow through system) at 100% fresh seawater addition.
Satisfactory weight growth development was recorded in RAS 20% fresh seawater addition, close to FTS performances, but spat size was lower. In RAS 10%, the loss of growth attained 20% but, as the consumption of algae was correlated with the growth, this is acceptable. With 5% fresh seawater the loss was ≈ 50% of spat weight growth despite a similar length of development.
The RAS 10% should be a good compromise but further economic studies need to be performed to evaluate the benefit for heat economy.

Pilot-scale Experiment
At Satmar the cost of the water treatment is much lower at the larval stage than at the micronursery stage. This is the reason why a RAS experiment was preferred. In March 2014, an experiment started at Satmar with oyster spat (500 ml oyster spat at 1-mm size) in a RAS (total volume of the filter: 840L and flow rate: 1.5m3h-1 and 10% fresh water addition per hour) showed that the spat was very healthy and its growth (the grow rate of the first week was 31%-and the second week 9.3%/day) similar to that in classical raceways. A trial at production scale is planned.

ii) Mussel, Mytilus edulis (IMARES)

For an efficient use of a recirculation system (RAS) we made a feeding method operational for M. edulis spat in which the actual algal cell density in the rearing container determined the addition of algae to the seed. The method makes use of a feed-back mechanism with a fluorescence sensor. This algae aqua-feed regulator was able to generally keep the cell concentration within 10% of the set level range. Thus, the system offers a good method for constant optimal feeding. System tests were carried out with a mixture of the algal species Pavlova lutheri and Chaetoceros muelleri. This diet proved to be the best for M. edulis spat in the BLUE SEED project. Mussel seed up to 5 mm filtered 40,000 cells/ml without producing pseudo-faces, the threshold was 50,000 cells/ml for 5-10 mm seed, 70,000 cells/ml for 10-15 mm seed and 80,000 cells/ml for 15-20 mm seed.

Mass balance calculations for mussel spat were performed including tank size and biomass, TAN mass balance calculations, biofilter sizing calculations, solids mass balance and oxygen mass balance calculations. Based on this an estimated required flow rate of 3.72 l min-1 was found for mussel spat in a 127-L system with 32-L rearing tanks.

Four trials in 32-L rearing tanks
Comparison of spat performances was made between RAS 10% freshwater addition per day and FTS 520% renewal per day with feed from outdoor batch cultures of Skeletonema costatum and Tetraselmis suecica (20 L per tank per day) or Reed mariculture diet (10 mL per tank per day). Water quality parameters remained within an acceptable range. Temperature was the most difficult parameter to keep at a desired level for RAS. Due to low algal density in outdoor batch cultures, the fresh seawater addition exceeded 10% daily in RAS. Similar growth was measured in FTS and RAS. However, the growth rate was much lower than in previous experiment. This may be related to the varying amount of food (outdoor batch) and/or low quality of diet (Reed Mariculture) that colonized the batch cultures. The need for a supply of sufficient amount of food, combined with the wish to reduce flow rate calls for use of concentrated, high quality feed.

iii) Scallop, Pecten maximus (UiB, Scalpro)

Spat were fed with a mixture of the single celled algae Isochrysis sp. (T-iso strain CCAP 927/14), Pavlova lutheri CCAP 931/1, Chaetoceros muelleri CCAP 1010/3 and Skeletonema costatum CCAP 1077/5 at a ratio of 3:2:3:1. Algal cell density was monitored 1-2 times a week and adjusted to maintain a sufficient algal concentration level in the tank (10-15 cell.µL-1). In the FTS system, the algal mixture was added to new inflowing seawater, and in the RAS additional algal feed was added directly to the tank.

Mass balance calculations for scallop spat were performed including tank size and biomass, TAN mass balances calculations, biofilter sizing calculations, solids mass balance and oxygen mass balance calculations. Based on this, an estimated required flow rate of 3.72 L min-1 was found for scallop spat in a 4898-L system.

Trials (4) in RAS
In this experiment, larvae reared in RAS (20 % fresh water addition /day) and FTS (100% renewal/day) were exposed to probiotics (Prob A, Prob B and Prob C) and compared to controls RAS and FTS (see WP3). After metamorphosis, the postlarvae were distributed into sieves. Twenty sieves were placed in the larval tank (3.500 L) keeping the same larval batches. Several benefits of RAS were observed at the postlarval stage:
• Similar or higher settlement
• More rapid settlement
• Similar or significantly larger sizes of postlarvae
• No pH reduction, but increased nitrite concentrations at times
Benefits of probiotics:
• Higher larval yield (2 of 3 tanks)
• Larger postlarval sizes (3 of 3 groups)
• Higher settlement (3 of 3 groups)
• Prob B gave the highest settlement
• To mix RAS and probiotics induce highr growth

The RAS system functions well for scallop postlarvae with regard to settlement, survival and shell growth. The high potential for RAS to reduce water and energy use in scallop spat nurseries was therefore clearly demonstrated. The economic evaluation showed a reduction of approximately 10% of the total costs compared to the flow-through system. An additional point is the stabilization of water quality meaning reduced labour for the same production or increased production for same labour. There is a need to go further to improve the biofilter system setup and find limits regarding maximum biomass capacities for Pecten maximus.

iv) Clam, Ruditapes decussatus (IPMA)

A mixed diet composed by two microalgae species Isochrysis aff galbana and Chaetoceros calcitrans (1:1 in cells/ml) has shown to be a good ration for R. decussatus juvenile development. However, for a successful implementation of high density R. decussatus seed production systems, the relation between algal density, seed size, seed density and algal uptake rate needs to be determined. The clearance rate of R. decussatus spat also provides substantial information on the microalgae consumption. To supply the algal cells more precisely as a function of clearance rate of seed, a dosing machine should be implemented to supply the algae more precisely.
The feasibility of rearing R. decussatus spat in a recirculating aquaculture system (RAS) was tested in three production tanks in shaped 500-mL funnels for spat rearing. The water pumped from the sump unity entered inside each of the three rearing tanks by “upwelling”. Two different flows were tested: 132 mL and 66 mL per minute. The results showed that, the dimensions of the RAS system were probably undersized for the rearing of R. decussatus juveniles with the size used in the experiment. Although no imbalance (physical, chemical or microbiological) of the system was found, seed did not grow significantly. We supposed that the food consumption by the system could be the cause of this weakness. The low growth that occurred with the highest fresh seawater addition (50%) could have been caused by a malfunctioning biofilter, leading to toxic compounds that disturbed the postlarval growth.

Microbiota in RAS for seed rearing
Deliverable D5.4

i) Oyster, Crassostrea gigas (Ifremer)
In the experiment of rearing of C. gigas postlarvae in RAS with different fresh seawater additions per hour (RAS1:20%, RAS2:10%, RAS3:20%) the microbiota was characterized by pyrosequencing in postlarvae at day 0 (or day 30 post fertilisation : dpf) and at day15 (dpf 45). For the last sampling, each replicate of RAS and the seawater per RAS were analysed. The postlarvae came from the experiment undertaken to test the probiotic effect during the larval culture in which the microbiota was also analysed. After this larval stage, the 3 original batches were kept separated during the metamorphosis stage and each was then distributed to constitute one replicate per RAS. The pyrosequencing of region V1-V3 of 16S rRNA gene was performed according to the same protocol u described previously for larvae. After treatment by Qiime software, the number of sequences per sample ranged from 2637 to 10710. The microbiota of postlarvae appeared as complex as in larvae and specific in comparison to the microbiota of seawater. It was constituted by a large number of OTUs (47 to 110), of which certain were sometimes predominant. The similarity before and after the metamorphosis was weak and ranged from 21 to 30%. The microbiota shifted between the beginning (15 dpf) and the end of this phase (30 dpf). Many OTUs disappeared, especially those affiliated to Rhodobacteraceae (α-Proteobacteria), Flavobacteriaceae and Planctomycetes. Members (6 OTUs) of Caulobacteraceae (α-Proteobacteria) and of Burkholderiales (β-Proteobacteria), including mainly Delftia sp (13 OTUs), replaced them. This last group took a notable place with 27±11%. Then, the microbiota became more stable and several OTUs established at the beginning of postlarval phase maintained similar abundance until the end, such as Delftia spp (except in RAS1), members of Caulobacteraceae (α-Proteobacteria), and Xanthomonadaceae (γ –Proteobacteria). The ANOSIM tests showed that RAS1 (20%) was significantly different to both the other RAS, while the bacterial assemblages of RAS2 and 3 were similar. Nevertheless, an unclassified bacterium accounting for 43% in RAS2 did not occur in RAS3. The differences in microbiota could be the reason for the different growth rates. Otherwise, distinct bacterial assemblages succeeded during the development stages and tended to be more diverse between replicates and batches at the end of postlarval rearing. Finally the fingerprint of hatchery in terms of bacteria does not seem so important for the ulterior development stages. Indeed, the microbiota did not appear definitely fixed but it evolved continually. It should change outside of hatchery on ongrowing intertidal sites even if some bacterial groups as Burkholderiales firmly linked to oyster pediveliger and postlarvae could persist. In conclusion, the hatchery–reared seed can be considered as a carrier of favourable or neutral microbiota without pathogenic bacteria especially vibrio.

ii) Scallop, Pecten maximus (UiB/USC)
Probiotic mixtures were added 3 times a week. Larvae were settled in sieves in FTS tanks containing 30 sieves each. Salinity, temperature, pH, ORP, fluorescence, nitrate, nitrite and ammonium were estimated. Results showed that larval growth was good in all larval tanks. Best growth was observed in standard FTS tank, but all tanks with probiotics also showed good growth. Best larval survival was observed in the tank receiving treatment with probiotics B and C. Survival after larval stage was between 50 and 60% and fraction ready to set was between 30.6 and 56.5%. The yield of larvae treated with probiotic B was significantly higher than with all other groups. The larvae from standard flow through (FTS) had significantly higher settlement than larvae given probiotic B, but similar to settlement of larvae given probiotics A and C. These results showed that the high potential for use of probiotics added to larval tanks is a new promising development for rearing sensitive larvae.

Task 5.2 Optimisation of phytoplankton production
Deliverables D5.5 D5.6 D5.7

1) Optimization of growth rate for Thalassiosira weissflogii, Chaetoceros gracilis and Skeletonema marinoi: a factorial approach (IFREMER) (Deliverable 5.5)

Optimal growth conditions for the three diatoms Chaetoceros gracilis, Thalassiosira weissflogii and Skeletonema marinoi were investigated using a factorial design approach. C. gracilis showed higher growth rates than T. weissflogii and S. marinoi under the conditions tested. Optimal growth was predicted for C. gracilis at temperature 33°C, irradiance 930 µmole.m-2.s-1 and pH 9.0. T. weissflogii optimal conditions were: temperature higher than 34°C, irradiance higher than 430 µmole.m-2.s-1 and pH 9.0. For S. marinoi, optimal growth was predicted for temperature 25 °C, irradiance 500 µmole.m-2.s-1 and pH 9.0. Interestingly, interactions between factors did not significantly affect growth for C. gracilis, while some did for T. weissflogii and S. marinoi. The three species demonstrated similar optimal conditions for irradiance and pH, but S. marinoi required a lower temperature for optimal growth. The three species appear to be better adapted for spring or summer outdoor cultivation.

2) Continuous outdoor cultures under different conditions (Deliverable 5.6)

The five selected algal species (four diatoms Skeletonema marinoi, Chaetoceros neogracile, Thallassiosira weiisflogii, Phaeodactylum tricornutum, and the flagellate Tetraselmis suecica) can be cultured in outdoor conditions using continuous or semi-continuous techniques. Different tests in raceways showed that the initial condition of the inoculum is of utmost importance. Furthermore, the continuous culture needs to be started at the moment that the culture is halfway through the log phase, outdoor cultures were accordingly inoculated with 3-day-old S. marinoi (indoor 300 L). The water needs to be filtered to 0.2 µm. A cheap medium containing ammonium, phosphorus (N:P ratio 9:1), silica, iron, manganese and vitamins seems an appropriate medium for the culture of diatoms. With a mean of 2 to 2.5 cells per colony, S. marinoi size generally remained small in our experimental conditions. With an average length of 8 to 12 µm, outdoor phytoplankton production fits well with the food particle size that most of mollusc spat require. The main bottleneck in large-scale outdoor continuous phytoplankton production is the longevity of the culture. The absorption ratio can indicate nutrient and light limitation during the growth phase. This can give the farmer a real time signal to add nutrients or dilute the culture, which may prolong the culture life.
The semi-continuous and continuous cultures yielded similar algal concentrations to batch cultures, but had a longer life time. This would reduce labour to 2 to 2.5 times less. In addition, continuous or semi-continuous conditions had lower levels of bacteria, which can also increase stability of the culture. Improvement of the system should focus on factors that can enhance the longevity of the culture. Transfer to end-users is established through the REPROSEED technology transfer workshop held in Brest.

3) Comparison of price of inert diet with costs of algal production NOFIMA (Deliverable 5.7)

In an attempt to provide cost-effective alternatives to fresh microalgae, substitution products have been tested over the last few years with variable results. To move ahead in the development of artificial diets, which theoretically could be used to replace live algae, feeding, completely or partly, knowledge mollusc feeding requirements is needed. The main objective of this task is the estimate the cost of a formulated diet for shellfish spat based on their essential needs.
The objective was to review the available literature on composition of various algae and other diets which were tested for shellfish larvae and spat and document the data sorted according the nutrient classes and the qualitative effects of the algae and/or diets in terms of growth and survival. The next goal was to use the data for producing a tentative diet formulation, which was sent to potential producers.
More than 100 publications, mainly articles and reports were reviewed. The relevant data were extracted to create a data base on MS Excel file. The data are sorted according to nutrient classes and also qualitatively sorted, between good diets the resulted in relatively good growth and survival, medium quality diets, and poor diets. The nutrient classes were total lipids, lipid classes, fatty acids and fatty acids groups (saturated un-saturated, n-3, etc.), sterol classes, total protein, amino acids composition, carbohydrates and ash content. In addition, non-quantitative data that seemed to be relevant were also registered.
The collected data were analysed using the Principle Components Analysis (PCA) statistical method and showed that the rations of n-3 and n-6 and triacylglycerols are positively correlated with good diets while DHA, 18:0, DHA/EPA and saturated fatty acids are to a lesser extent. Free fatty acids are strongly negative correlated with good diets. The analysis of another lipid class, sterols, suggests the importance of Isofucosterol and Campesterol in bivalve nutrition. There was a positive correlation between ash content and inferior diets. Essential amino acids, which are positively correlated with the beneficial diet and the non-essential amino acids, are in proximity of the inferior diets on the other side.
A particle size and chemical composition of the diet was derived from the analysis. The market size for spat diet is based on FAO data from 2008, 276 tons are produced per year in Europe for the feeding of molluscs. This amount of algae costs about $276 106 year-1 (Otero, 2010). To our regret, out of 13 potential producers that we contacted to give a cost estimation, only one responded. Aquafauna Bio-Marine produces a micro-diet (Algamac Protein Plus) which is a good live algae replacement for rotifer culture. The cost is €17/kg. Prices of life algae and algae paste were higher than €100/kg.
IMARES tested the Skretting rotifer diet Ori-culture (€85/kg) on mussel spat and did not find any growth, while the control treatment fed a 50/50 mixture of live Chaetoceros muelleri and Pavlova lutheri showed good growth.

Task 5.3 Quality of seed
Deliverables 5.8 5.9 5.11

1) Genetic diversity of seed (Ifremer)
Protocol of amplification with 2-3 microsatellite multiplexes in the three species C. gigas, P. maximus, R. decussatus (Deliverable 5.8.)
Genetic monitoring of cultivated species of economic interest is a major concern for the long term viability of aquaculture. Maintaining genetic diversity in cultured population is a key factor to limiting the impact of seeding and culture on native or non cultivated populations. Therefore, we successfully developed multiplex PCR microsatellites assays for the Pacific oyster Crassostrea gigas (Li et al., 2010) and great scallop Pecten maximus (Morvezen et al., 2013). We were able to estimate parental assignment power of these assays in both species. For Ruditapes decussatus, no sufficient spat production was obtained in the project and the development of the markers by other teams in parallel led to the decision of a non duplication of the work for this species.

Comparison of effective size and diversity for seed production of several species (Deliverable 5.9.)In C. gigas, parentage analysis of real offspring demonstrated that 97% of all offspring were unambiguously allocated to a pair of parents based on two multiplex PCRs with only a 4% error rate, and 100% of the offspring were correctly allocated to their parents when three multiplex PCRs were used. The sampling and genotyping was performed in French hatcheries in collaboration with SYSAAF and Labogena for C. gigas, but the low number of families finally sampled only allowed us to work on the assignment power of the panel of markers and could not conclude on effective population size and diversity.
In the context of restocking of P. maximus in Brittany, we were able to determine how many of the seed really come from crosses made in the hatchery. Finally we characterized the genetic diversity in seed production and showed, again for P. maximus, that there is no loss of genetic diversity between seed produced in hatchery in 2007 and the natural populations in Brittany. We did not either detect any temporal loss of diversity in three natural populations issued from reproduction in 2004-2005, 2007 and 2008-2009.

To conclude, these new tools and analyses will be very useful in the management of seed production and their genetic diversity in hatcheries, in contexts of both restocking and breeding programs.

2) Effects of first gonadic development on seed robustness (UCBN)

In adult oysters, gametogenesis is a period of negative energy budget (Soletchnik et al., 1997; Heude-Berthelin, 2000), detrimental for their defence mechanisms (Perdue et al., 1984). Moreover, these mechanisms are tightly regulated by temperature and food supply (Fabioux et al., 2005). At the beginning of this work, very little information was available on the first gonadic establishment, gametogenesis and sex determination in spat, especially regarding the potential influence of the environment. In order to address this question, mRNA expressions of actors of the molecular cascade (Cg-DMl, Cg-SoxE, Cg-β-catenin, Cg-Foxl2/Cg-Foxl2os) and of Oyvlg, a germ cell marker, were investigated by real-time PCR in spat grown at different temperatures (18, 22, 25 and 28°C). In parallel, gonadic differentiation, gametogenesis and sex ratios were assessed by histology at each of these temperatures. Whatever the temperature, Cg-DMl, Cg-SoxE, Cg-β-catenin and Oyvlg expressions peaked at the same developmental stage, always after Cg-Foxl2/Cg-Foxl2os (around 40–44 dpf for spat grown at 18°C). Temperature increased the kinetics of first gonadic differentiation and gametogenesis. At 25°C a significant switch occurred in sex ratio towards males and in the balance of expression between male and female genes, in favour of males. A slight gametogenesis disturbance was also observed.
From a practical point of view, our results will help (i) to better identify spat critical physiological time-windows and (ii) to provide some clues on how to use temperature and food to modulate physiological kinetics and sex-ratios.

3) Gene expression associated with seed survival and growth under various experimental conditions (Deliverable 5.11) (CSIC, UNIPD)

Taking advantage of the genomic tools developed in this project we were able to follow the expression of immune genes until the seed stage.
Although, unfortunately, the gene expression profile could not be studied to compare the RAS and batch systems because there were not enough differences in the seed condition, in this WP we were able to describe the gene expression associated with bivalve ontogeny.
Several genes of mussels increase their expression through the different developmental stages until reaching a maximum level in seed, suggesting that at this moment they present a complete immunological response. We also compared the expression of some key genes in clam, R. decussatus seed exposed to a bacterial infection in order to evaluate differential values after significant stress. Immune genes showed a general increment after 48h, but genes involved in apoptosis showed an inhibition after the bacterial exposure. The differential expression of TLRs, main receptors of the innate immune system, after infection with pathogenic or non-pathogenic bacteria suggests a complex recognition of pathogens that should be studied further.

Potential Impact:
Potential impact, main dissemination activities and exploitation of results
1- Dissemination
The outcomes of REPROSEED were disseminated to (i) the scientific community, through publications in scientific journals and conference communications; (ii) the end-users, by the workshop organised at the end of the programme and the survey on this economic sector; and (iii) a wider audience, through two articles published in international journals dedicated to aquaculture and to general information on scientific programmes. The video-film and website, which will remain active over an extended period, will contribute to informing the public more broadly. At the local and national levels some participants were been interviewed about this programme by television and newspapers.
Twenty two articles from REPROSEED programme have been published or accepted in scientific journals. Further articles have been submitted and many others are in preparation. Six PhD theses have been defended including REPROSEED research. A patent for probiotics should soon be deposited by USC, UiB and Scalpro.
The other means of dissemination have been oral and poster communications at eight international and five national conferences. The international conferences were generally specialized (‘microbiology, genetics, marine genomics, immunology, physiology, etc.) and the national meetings were more focused on aquaculture of molluscs.
Two general presentations of the REPROSEED program were made: one at the EAS meeting in Rhodes (Greece) 18–21 October 2011, the other at the 1st Oyster World Congress, Arcachon (France), 28 November–02 December 2012.
A special session of REPROSEED was organised during the Physiomar 12 conference, on 4–8 September, 2012 in Santiago de Compostela to present the main results of REPROSSED after 30 months of the project. Twenty-three oral and poster communications were presented including both academic and applied research.
Communication to a broader audience was made through two articles on the REPROSEED programme published in International journals, one dedicated to the scientific programme in Europe. One article on REPROSEED was printed in a special issue untitled “Food for thought” (International Innovation in March 2013). The other was published in an issue of Global Aquaculture Alliance (April 2013).
Members of the CSIC-Vigo team were interviewed by regional television and described the REPROSEED programme. In addition, some articles about this research were published in regional journals.
An article on the workshop with end-users was published in “Cultures Marines” (January 2014).
An audiovisual document has been produced by the communication service of Ifremer, presenting an overview of bivalve production in Europe, the objectives and main outcomes of REPROSEED to secure and develop seed production.
The website is frequently visited and updated regularly. It will be active for at least one year after the end of REPROSEED programme.

2- Relationship with end-users
The participation of two commercial hatcheries (SATMAR and Scalpro) in REPROSEED was very useful, notably to draft the questionnaire for the enquiry into the industry and to test RAS at pilot or production scales.
The relationship with end-users was established through the survey on mollusc hatcheries in Europe. The objective of this enquiry was to make the first ever review of the general activities and practices of commercial shellfish hatcheries operating in Europe.
The survey provided a number of different types of information regarding management, staff, production and technological practices.
The enquiry confirmed the limited size of the hatchery sector, with only 38 companies in the whole of Europe at the time it was made. All companies were SMEs, with a high diversity of sizes. The number of employees varied between 1 to 50, with a median of four, consisting of one biologist, two technicians and one basic staff employee. This sector was characterized by the high skill level of the employees, but the woman/man ratio of 1/6 was far from balanced.
Standardization of processes is limited in this sector (water treatments, technologies, reproduction practices, nutrition, etc.). This is due to the relative newness and pioneer status of this activity, the great diversity of species but also to the different environmental requirements or characteristics.
The veterinary control and hygiene practices are very limited, even though the current sanitary rules are applied. Otherwise, this sector shows rapid adoption of innovative technologies (flow-through systems, water treatments, triploid seed production, genetic selection) and those still in development, such as RAS).
As a second means to establish a partnership with the end-users was the workshop, held in Ifremer centre Bretagne, on 28–29th October, 2013. Nineteen persons representing 14 European hatcheries participated in this meeting with 12 scientists involved in REPROSEED and a representative of EMPA. Syntheses of outcomes were presented by the leaders of workpackages and axes. All attendees came away pleased with this meeting. Different initiatives should now be taken to realize the collaborations between these end-users and scientists. Many hatcheries are keen to work in cooperation with research laboratories and other hatcheries. Different possibilities have been suggested to consolidate and extend this collaboration, such as a forum of discussion on the web the elaboration of the European implement some outcomes of REPROSEED, to participate in different European platforms (EATIP, FABRE-TP) so that the mollusc hatchery industry and its technical needs are better recognized.
The creation of an association of European mollusc hatcheries was not considered as viable because it the sector is too small and scattered. Finally, to pursue this collaboration it will be necessary to constitute rapidly a working party composed of scientists and end-users.
The relationship with EMPA (European Mollusc Producers Association) was cordial but contact was infrequent. Apart from the kick-off meeting and the workshop with the end-users, no representatives attended the REPROSEED meetings although they were formally invited. The minutes of these meetings were sent to EMPA.

B- Technical and scientific advances
1-Broodstock management
In R. decussatus, both populations (Ria de Aveiro and Ria Formosa lagoons) could be considered as viable broodstock for intensive hatchery production of juveniles and the observed extended spawning periods have interesting implications for the implementation of profitable aquaculture. Moreover, this species presented a great capacity for gonadic regeneration, which, coupled with its high gonadal development rate, could enable larvae to be produced during most of the year without extensive and expensive broodstock conditioning. Temperature is a parameter that must be carefully managed to improve the reproductive conditioning of bivalves. The knowledge of the most profitable combination of diet and temperature to conditioning R. decussatus broodstock can contribute to improving global hatchery technological development of this emerging species, and consequently the production of high quality spat.
In Crassostrea gigas, a preconditioning treatment based on abundant alimentation under low temperature for 5 weeks before broodstock conditioning increased the gametogenesis effort, leading to production of large quantities of gametes. Moreover, characteristics of breeders were more homogeneous: males exhibited high and homogenous fertilization rate. Female exhibited high fertilization and D-larvae rates. Starved animals exhibited irregular performances.
2- Oocyte maturation and quality
In Ruditapes decussatus and Pecten maximus, differences in the mRNA expression of some important genes were detected, providing a more comprehensive picture of the key processes affecting R. decussatus and P. maximus oocyte maturation and competence acquisition. The transcripts which seemed to play a major role in the female gamete maturation and competence acquisition were those encoding proteins involved in the cell cycle progression, cytoskeleton organization, calcium regulation and WNT signalling. Although R. decussatus and P. maximus reproduction is still far from being completely controlled in hatchery, this study represents, as far as we know, the most effective attempt to improve the knowledge of oocyte maturation processes in these two species.
In Crassostrea gigas, female Pacific oysters do not all produce oocytes of equal quality. Using a 2-DE proteomic approach, several proteins that were differentially accumulated according to oocyte quality were identified. Characterization of these proteins provided important insight into the mechanisms of early development in this marine bivalve.
In Mytilus galloprovincialis, expression of genes encoding antimicrobial peptides (Myticins, C1q) could be used as markers of the quality of mussel oocytes.
3- Spermatozoa maturation and quality
In Pecten maximus, following an intragonadic serotonin injection, sperm could be collected in all mature males within 30 minutes. This sperm collection technique can be used during routine hatchery practices. In Crassotrea gigas and Pecten maximus, the knowledge of proteins involved in spermatozoa maturation processes was greatly increased. The comparison between proteomes has highlighted protein candidates that could be essential for maturation of spermatozoa and could be used as markers of sperm quality.
4- Control of Gamete release
The spawning of R.decussatus is unpredictable, hindering the establishment of a fixed protocol . However, some changes in the traditional stimuli of thermal shock led to higher percentages of spawners, and consequently to making the artificial production of this species in hatchery more attractive. To define an effective spawning induction protocol for R. decussatus, it will be necessary to evaluate other stimuli. It is the same for the P. maximus female for which no spawning protocol was identified. For these species, interesting perspectives may come from results of tests with C. gigas peptides, including peptides from egg-laying hormone and vitellogenin-like protein. If gamete release is induced, the perspective of this work is a large production of recombinant proteins in yeast system as possible mean to trigger spawning in respecting physiological conditions.
5- Improvement of rearing using RAS
Among the technical improvements made during REPROSEED, the most important is undoubtedly with RAS. Results fulfilled the objectives for the larval and postlarval rearing. They are particularly efficient for scallop when coupled with the addition of probiotics during the larval stages. For oyster larvae and postlarvae, RAS also appeared well adapted even if a slowdown in growth occurred lasting of one or two days of larval culture, but without a supplementary algal consumption that was correlated with growth. For the postlarval culture, an equivalent growth in terms of weight was achieved with 20% fresh seawater renewal and, at pilot scale; similar performances were obtained with 10%. Finally, RAS can replace the traditional flow-through spat rearing in different mollusc species with conditions to adapt it for each species. RAS systems reduce needs for water, energy and labour and create a more stable environment. Finally the cost reduction was estimated around 10% but is probably more when in taking account the improvement of the reliability of this system.
Otherwise this system could still be optimized by taking account the waste matrix. The biofilter in particular could be a more adapted size as the ammonia excretion by the bivalves is much weaker than that of fish. When the closure is important, the pH could be adjusted by the addition of calcium bicarbonate, which is also necessary to shell formation.
6- Microbiota in RAS
The cultivable bacteria in RAS did not exceed 5x104 mL-1 for oyster, clam or scallop larvae and the vibrios remained at a low level except when mortalities occurred. The UV treatment may be efficient but other factors as the presence of predatory bacteria may also contribute to limiting bacterial proliferation, especially vibrios.
The positive effect of probiotics on scallop larvae and postlarvae opens new perspectives for bivalve species that are difficult to produce. Thus, they could improve the larval rearing of bivalves such as R. decussatus, the flat oyster, Ostrea edulis, and other Pectinids. Although this research was started several years ago, application at production scale of bivalve has rarely been done. To best exploit this outcome, a patent should be deposited on these probiotic mixtures.
The microbiota may cause variability in larval and postlarval performances. At present the control of bacteria is carried out by a sanitary procedure including disinfection of broodstock, UV treatment and cleaning operations. This seems to be efficient to prevent infectious disease. According to the holobiont theory, the microbiota is an essential element for the health and development of its host. As a result of the sanitary control and consequent the relative impoverishment of bacterial diversity it might be expected that microbiota become deficient, and notably contain detrimental bacteria. However, the microbiota appeared more diverse at the seed stage and no pathogenic bacteria could be detected. Finally, it can be considered as safe and possibly favourable, without negative effects on the ulterior development of seed. In fact, the microbiota evolved continuously even if some bacterial groups seem more persistent. This conclusion could be extended to commercial hatcheries once similar studies have been conducted on their seed.
7- Immunology and genomics
Gentle bubbling in scallop larval rearing was sufficient to cause disturbance. This justifies the use of antibiotic in some commercial hatcheries. Such an effect is indicated by the overexpression of several genes, notably those involved in collagen metabolism. Two consequences can be drawn from this experiment. Firstly, some bivalve larvae are susceptible to any stress. Therefore, the rearing conditions must be adapted to each species. Secondly, the genomic approach made it possible to reveal precisely the effect of rearing stress on the immunity and physiology of the animals. Some proteins and gene markers would be useful tools to diagnose physiological disorders more rapidly and could first be applied to R. decussatus larvae and postlarvae.
The microarrays constructed on the basis of mass sequencing and annotations are now available and can be used for many scientific applications.
8- Metamorphosis
The different inducers of metamorphosis tested on Mytilus edulis did not significantly enhance the settlement despite the numerous trials. The change in dietary fatty acid profiles did not influence the metamorphosis rates either. For Pecten maximus, two chemicals (KCl and L-DOPA) significantly improved the metamorphosis rate. However, the gain was considered as insufficient to apply this method in commercial hatcheries with regards to the lack of reliability and the risk of mortalities. For R. decussatus, similar remarks can be made for epinephrine and GABA, which improved the metamorphosis rate. Finally, the conditions to reach a high metamorphosis rate must be sought among other factors, such as substrate with preconditioning, the food and the attachment period. These parameters were defined for scallop larvae and should be extended for the other bivalves
9- Algal production and inert diet
Continuous or semi-continuous algal cultures in raceways were perfected using a simple and cheap medium, increasing culture longevity and optimizing some other parameters, like pH and light. This method has some advantages over traditional batch culture including that it reduces labour. However some improvements should be introduced including the increase of longevity and the produced quantity.
The bibliographic review on the inert food leads us to conclude that the optimisation in composition of inert diet and production of microparticles appear possible. However, recent tests performed in this programme, like previous ones, on different types of inert food in comparison with live food were not fully conclusive. Taking account the size of market and the cost to develop an inert food, no concerned companies are interested in launching a new formulation of particles. However, the conditions for the production of inert food should evolve with scientific advances in this domain and the cost, which appears to be attractive.
10- Quality of seed
The quality of seed was investigated according to three criteria. The first concerned the genetic diversity of seed. The multiplex PCR microsatellites successfully developed for C. gigas and P. maximus will be very useful in the management of seed production and their genetic diversity in hatcheries, both in contexts of restocking and in breeding programs. When this method was applied to the scallop seed demonstrated that in seed produced in the hatchery there was no loss of the diversity compared with natural populations
It can be concluded that the method of production in the hatchery, including broodstock management, rearing parameters such as high temperature, ad libitum food and elimination of small larvae does not affect the diversity except if a breeding programme is performed.
The second criterion is the precocity of maturation in oyster that decreases the robustness of oyster seed. The experiment confirmed that high temperature and food supply increased gametogenesis. In addition, at 25°C, a switch occurred towards males in sex-ratio. However, the decrease in food supply and temperature are incompatible with hatchery production.
The third criterion was to establish a relationship between immunity status and seed robustness in emerging species. The efforts were concentrated on acquiring of new data on the transcriptome and to describe the genes associated differentially to each stage can be very useful as a baseline to future works to evaluate the quality of the seed using these.

List of Websites: