Culture of Ciona intestinalis in closed systems

Authors


Abstract

Improvements in closed-system culturing methods for marine invertebrates are important prerequisites for the generalized use of transgenic lines. We discuss here the effects of several closed-system conditions on the growth and survival of the solitary ascidian, Ciona intestinalis. In Shimoda, close to the sea, a small-tank system was used to ensure that tanks and systems were reasonably equipped, water exchange was rapid, and animals separated to minimize the risk of infection. In Gif-sur-Yvette, an inland site, we tried to determine the optimal conditions to limit handling operations, and to save artificial seawater by avoiding water pollution. A mixture of at least two types of live algae was better than any single-organism diet. With these maintenance protocols, we were able to obtain several generations of Ciona intestinalis, including several transgenic lines. Because these systems make it easier to rear Ciona intestinalis in laboratories, they increase the potentialities of this model organism for research. Developmental Dynamics 236:1832–1840, 2007. © 2007 Wiley-Liss, Inc.

INTRODUCTION

Ciona intestinalis is an ideal model system for research in several domains of biology (see other reports in this special issue). Several methods have been applied to this simple chordate for functional studies of genes and regulatory elements. Egg electroporation and the microinjection of DNA, RNA, or morpholinos are routinely performed on animals collected in the wild, requiring only short-term maintenance (Satoh et al.,2003). Ciona intestinalis have been maintained in laboratories since the middle of the last century. Berrill (1937,1947) observed late stages of development and juveniles, and Costello and Henley (1971) used Berrill's method, and cultured Ciona in the laboratory, on a diet of diatoms. About 20 years ago, several Japanese researchers set up a long-term inland culture system to breed successive generations of Ciona (joint project directed by Motonori Hoshi, grant MEXT no. 60880009, 1987). This project resulted in 19 generations of successive inbreeding (Satoh,1994), demonstrating the feasibility of long-term culture for marine invertebrates. Inland culture systems are useful as they provide easy access for checking the status of the animals and for experimentation throughout the year. Chemical mutagenesis (Moody et al.,1999; Sordino et al.,2000,2001) and stable transgenesis methods (Deschet et al.,2003; Sasakura et al.,2003) have recently been developed. These methods clearly require improvements in culture techniques for this species. Two laboratories have constructed ascidian culture facilities close to the sea, facilitating large-scale mutant analyses (Cirino et al.,2002; Hendrickson et al.,2004; see also Fig. 1). At the facility developed by Smith et al. at Santa Barbara (CA, USA), natural seawater, containing plankton for the filter feeders, is pumped into the tanks. Cirino et al. (2002) developed a different system, based on the use of artificial foods. Both facilities are open-sea systems and, therefore, benefit from an unlimited supply of seawater. The techniques used by these facilities are, however, difficult to apply to inland laboratories, where seawater is in short supply, or to marine stations handling genetically modified organisms (GMOs), requiring strict measures to prevent escape. Thus, there is a need for closed seawater systems for inland laboratories and marine research stations.

Figure 1.

Recommended strategies based on seawater availability and purpose.

One of the chief priorities for inland laboratories is limiting the use of expensive artificial seawater. The geographical location of these laboratories acts as a natural barrier to the dissemination of genetically modified marine animals. The recent development of transgenic ascidians has obliged marine stations to limit their use of seawater, not due to shortage, but due to the need to treat water and effluent to prevent GMO dissemination in the wild. Closed systems also prevent contamination with wild-type animals.

Over the last few years, we have devoted considerable efforts to improving culture systems. The parameters tested include tank size, and type and amount of food and seawater. A limited amount of routine maintenance is required. The system may be scaled down in terms of both the size and number of tanks, according to research needs, as previously described (Sasakura et al.,2003; Hendrickson et al.,2004). However, many apparently minor points are of critical importance in closed systems for establishing populations of transgenic animals with various insertions and genetic backgrounds, particularly as these organisms must often be maintained reliably over long time periods. We, therefore, provide a detailed description of our closed systems for culturing Cionidae in laboratories and discuss the setting of the various parameters for laboratories wishing to set up a new ascidian facility.

RESULTS AND DISCUSSION

Strategies for Closed Systems

Different designs and strategies may be used, depending on the location, available space, funding, available manpower, and purpose of the facilities (Fig. 1). A closed system was set up at Gif (Gif-sur-Yvette, Essonnne, France) due to limited seawater availability. The Shimoda facility has unlimited access to seawater but needs to prevent GMO escape. A closed system was, therefore, also chosen for this site.

One of the main purposes of our closed culture system is to make possible the isolation, maintenance, and experimental use of stable transgenic animals. Sperm maturation, which comes earlier and easier than egg maturation, is sufficient for this purpose, because reporter genes are expressed as dominant markers in transgenic lines. Conversely, culture for egg maturation is required only for limited purposes, such as mutant screening with recessive phenotypes, and the observation of maternal reporter gene expression in eggs. Our systems are, therefore, designed for the culture of large numbers of sperm-producing animals and a proportion of well-grown adults with mature eggs.

We describe here two major strategies for laboratories with limited access to seawater and cost concerns, such as that at Gif, and for laboratories with ready access to seawater, such as that at Shimoda.

Design of Tanks

At Gif, an inland site, the cost of seawater is a fundamental concern. Frequent seawater exchange is not possible. We estimate that 20,000 liters of artificial seawater are required each year for the 80 tanks at Gif, at an overall annual cost of about 3,300 Euros. Electric pumps are used to ensure water flow, essential for maintenance of the oxygenation and quality of seawater. We initially decided to raise juveniles in separate rat cages (15 l each) for experiments, to obtain adequate amounts of food and to prevent contamination. About a week before the start of the experiment, these tanks were filled with artificial seawater (Reef Crystals, Instant Ocean; see Supplemental Material #1, which can be viewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat) or seawater collected from the English Channel and obtained from Hippocampe SA (Antony, France). Several optional pieces of equipment were tested: lights (two standard neon tubes, Grolux, Sylvania), one 320 l/h pump (Micro-jet ML320, Aquarium Systems), and one air-pulsed internal filter filled with “pozzolana” volcanic lava (fitted in the tanks 2 weeks before the experiment). Air is expelled from the filter via a bent tube at the surface, resulting in moderate, but probably beneficial mixing of the water surface and no projection, which would otherwise result in higher levels of evaporation and salt deposits on tanks. All these pieces of equipment were found to increase growth rate and are now routinely used (see Supplemental Materials #2 and #3).

The marine station at Shimoda is supplied with large quantities of seawater: about 180,000 litters of natural sea water are used each year for the 60 tanks when the facility runs at full power. A different strategy was used here. Tanks are duplicated to ensure that fresh seawater is supplied at the correct temperature during seawater exchange (see Supplemental Materials #4 and # 5). The circulation and filtration system have been kept as simple as possible, thanks to the availability of large amounts of seawater.

There is no need to circulate seawater with a pump and long pipes, avoiding laborious maintenance of the blockage of pipes with sticky living diatoms. Frequent seawater exchange also overcomes the need for filtration on sands or corals. The Shimoda system involves a flow of seawater controlled by a synchronous motor (30 rpm) connected to one shaft and two propellers (see Supplemental Material #2B–E). All the seawater in this system is exchanged completely twice per week. We have now 60 paired tanks, making it possible to culture about 6,000 animals simultaneously. The washing of these tanks—drainage, washing, and addition of fresh seawater (see Supplemental Material #5)—takes two people about 2 hr. The daily care of a few transgenic lines requires only about 10 min. The simplicity of the tank equipment used minimizes the time required for water exchange. One of the advantages of this small-tank system for large-scale cultures is that all the Ciona juveniles at the same age can be placed in tanks at the same time. Juveniles in different ages or size are not added. The growth of Ciona in the tank is almost synchronous.

Basic Flows for Reproduction, Growth, and Ageing

The standard schedule for Ciona growth in our systems is shown in Figure 2 (refer to Supplemental Material #6 for more detail). The maturation of sperm and eggs takes about 2 and 3 months, respectively, at Shimoda. Maturation is slightly slower in the Gif system, presumably due to the limiting quantities of food. Ciona lives for 6 months or more. This long lifespan makes it possible to stock adults and delay production of the next generation. However, old animals sometimes reproduce improper eggs in which germinal vesicle breakdown does not occur, while functional sperm can be obtained.

Figure 2.

Breeding schedules in the two breeding systems.

Water recirculation, which may lead to contamination in or between tanks, is a potential problem in closed systems. At Gif, we used filters to minimize this risk but we also found that excluding mature egg-bearing animals from the main facility system was effective. This finding influences current strategies for the handling of transgenic ascidians at Gif (Fig. 1, see Supplemental Material #3): the next generation is produced as soon as the adults start to produce sperm. The remaining genitors are placed in large tanks under constant lighting, with artificial feed supplementation (800 l), as previously described (Hendrickson et al.,2004). The smallest animals are eliminated from the trays. The trays are periodically examined for the presence of potentially contaminating juveniles, which are eliminated immediately.

Density of Juveniles and Adults in Tanks

Animal density is a critical parameter in closed systems, in which access to seawater is limited and costly. High animal density results in slower growth and lower rates of egg and sperm maturation. Lower densities yield higher proportions of animals with mature eggs but the maintenance of smaller numbers of animals in each tank requires the use of a larger number of tanks for a given number of transgenic lines. Animal density should, therefore, be determined by egg and sperm requirements, the size of the facility, and holding supports (Supplementary Material #7). Density is optimized according to the intended purpose: (1) a few animals producing sperm only are used to obtain sperm surgically to produce the next generation; (2) animals with mature eggs are used for self-fertilization for the screening of mutants; and (3) large stocks of animals are required to ensure supply when wild animals are unavailable and to replace animals accidentally stripped from the Petri dishes. Many of the transgenic animals generated by electroporation die during metamorphosis and a significant proportion of these animals are malformed or grow slowly. These animals should be eliminated as soon as possible because they do not develop gonads.

At Gif, animals are cultured at a density of about 1 adult individual per liter. The density typically used in our systems is thus three adults per 3-l tank, 3 trays with a total of 20 to 30 adults in a 20-l tank, and a maximum of around 2,500 adults in our 2,500-l facility, consisting of 80 tanks. Higher densities would require an increase in food supply, leading to toxicity and poor water quality. The number of Ciona is adjusted according to the purpose of the culture and the size of the facility. Juveniles require less food than adults, so larger numbers of animals can be used per tray (typically 50). The utilization of bench space is optimized by using rectangular cell culture trays (Omni tray cell culture treated, 165218, Nalge Nunc International). Water is changed daily and antibiotic treatment (e.g., gentamycin, Sigma, G1272, St. Louis, MO) is implemented occasionally. Larvae are plated at a density of about 50 per dish (200 per dish for weaker transgenic animals resulting from electroporation), and dishes with about 30 healthy animals are selected (transparent tunic matrix, filtering activity and blood cell movements) and transferred, after one week, to 20-l tanks for culture. The key to success is to control the number of juveniles, by regular removal of the smallest and weakest animals, with wooden toothpicks and cell scrapers.

The conditions at Shimoda are similar but the strategy used is slightly different. Animals are kept in 9-cm Petri dishes for 3 weeks. For transgenic lines, which require both sperm and eggs, about 10 healthy juveniles per Petri dish are retained at 3 to 5 days after fertilization. More juveniles are retained when culturing transformed animals to generate founders, as egg maturation is not necessitated. Large numbers of animals should be required for the screening of large populations. In this case, the culture is initiated with about 20 juveniles. It is important to retain enough healthy juveniles for the transfer of 5–6 (for maintenance of transgenic lines) or 10 (for creating founders) animals per Petri dish to tanks after 3 weeks. Sixteen Petri dishes are placed in each tank, giving a total of 80 to 160 Ciona per tank according to the purposes. Animal density may thus reach 2 to 5 animals per liter.

Selection of Food

Food is the most important parameter in Ciona culture. It should fulfill several criteria: it should be easy to obtain and to prepare, should provide sufficient nutrients for growth and reproduction, and should pollute seawater as little as possible. A live natural diet (e.g., live algae) is the most appropriate, particularly for juveniles, due to its small particle size. Natural food is also less polluting, a major advantage over artificial foods, particularly in situations in which seawater is limited. Large quantities of algae must, therefore, be cultivated, and this may be highly time-consuming. Detailed protocols are provided in Supplemental Material #8.

We have carried out five sets of experiments investigating several artificial diets and species of algae (for 4 to 8 weeks in each case). We used a progressive approach, comparing several types of tank equipment and foods. Each set of experiments was carried out on a different batch of animals, with all batches originating from the same location (see below) but not necessarily collected in the same season. The possibilities for inter-experiment comparisons were, therefore, limited. However, conditions that proved efficient in successive experiments were regarded as a genuine improvement of laboratory culture and used as a reference for subsequent experiments.

At Gif, the first experiments clearly demonstrated that artificial foods are not suitable for juveniles, as the animals failed to grow and died quickly, probably due to water pollution. Seven different types of live algae were tested: two different green algae (Skeletonema costatum and Tetraselmis suecica), a tropical red alga (Isochrysis galbana “Tahiti” type), a diatom (Chaetoceros gracilis), and three other species (Chlorella sp., Dunaliella sp., and Platymonas sp.). The last three species were difficult to culture, so that their use was abandoned. Abnormal intestine filling was observed in juveniles supplied with a diet consisting of the two green algae. These algae (about 20 to 60 μm long) are probably too large for ascidian juveniles to eat. The juveniles failed to grow on this diet and the mortality rate was high. Growth was maximal with Isochrysis galbana and Chaetoceros gracilis. Rhodomonas has also been identified as a suitable food for Ciona (Berrill,1947), but the slightly larger size of these algae (> 10 μm) may make them less appropriate for juveniles, which seem to prefer small algae (Tyree,2001).

The harvesting of algae in laboratories is time-consuming and may require large items of specialist equipment. Moreover, contamination of algal cultures is frequent, requiring the re-establishment of cultures from sterile stocks (see Supplementary Material #8). In facilities with algal culture capacities large enough to scale up quantities of the number of animals, red algae (Isochrysis) or diatoms (Ch. gracilis) may replace artificial dried foods (see below). Concentrating the algae has no effect on cell viability. This may limit the nitrates or contaminating metals transferred to the culture with the algae. Centrifugation of algae could be carried out in laboratories with high-capacity centrifuges. In France, we are currently trying to identify private companies selling such products. This process would increase costs, but would reduce manual labor and would probably increase reliability. In Japan, Chaetoceros calcitrans can be purchased from a company (Nisshin Marinetech, CO., LTD, Yokohama, Japan), although this solution is costly for large-scale Ciona cultures. Sexual maturation can be achieved, even if this single species is given to young Ciona for the sake of simplicity (see below). However, a mixture of two species is more efficient (see Supplemental Material #9), as reported for the culture of echinoderm larvae (Wray et al.,2004). C. savignyi seems to prefer different types of food, and does not grow well on a single-diatom diet. Mixtures, therefore, seem preferable in all cases.

Rearing of Juveniles

Juveniles start to eat food about three days after fertilization. Just after metamorphosis, small juveniles are not very tolerant to poor seawater quality, resulting primarily from the presence of unconsumed food. Live diatoms are, therefore, the best food at this stage. At this stage, animals can be transferred to 2- to 3-l tanks in which the plates are kept vertical (Gif), although this system may not be suitable if the animals are to be observed continuously after the juvenile stages. This problem can be overcome by culturing them in smaller containers, such as tall Petri dishes (25-mm height and 92 mm in diameter, No. 906, BIO-BIK), facilitating complete changes of seawater (Shimoda). The greater height of these dishes results in diatoms floating for longer periods, increasing the chances of their capture by juveniles. Normal Petri dishes are not recommended in this strategy. Frequent seawater exchange maintains water quality without requiring large amounts of seawater. Juveniles kept in tall Petri dishes are supplied with Ch. gracilis or Ch. calcitrans. Live diatoms are available from several companies in Japan, but we prefer to use laboratory-harvested diatoms for small juveniles. Most commercially available diatoms are indeed supplied in very high-density formulations (10,000,000–100,000,000 cells per ml), containing some dead cells. These dead diatoms stick to the bottom of the dishes, and their accumulation can result in serious health problems in juveniles. Laboratory-cultured diatoms (<100,000 cells/ml) contain fewer dead cells and rarely cause such problems. The harvested diatoms are diluted to 3% (v/v) with seawater, and are added every other day, during seawater exchange. If the bottom of the dish becomes dirty, it should be cleaned with the fingers or cell scrapers. Once the juveniles have reached a length of 2 mm, they are transferred to 30-l tanks. Attention should be paid to the number of diatoms added to the tanks during the first week, because young Ciona are sensitive to poor water quality. Live diatoms (at this point, juveniles may be fed on laboratory-harvested or commercially produced diatoms) are supplied at a concentration of about 300 cells/ml, when the intestines of the juveniles are empty.

Rearing of Subadults and Adults

We have found that growth suddenly accelerates 3 weeks after fertilization. This time point corresponds roughly to the stage at which the gonads start to grow (Okada and Yamamoto, 1999). Around this stage, the amount of food should be increased, typically from 10 ml per species of algae to 100 ml in a 20-l tank. As our culture system yields a concentration of 500,000 algae per ml, the concentration is raised from 250 cells per ml to 2,500 cells per ml, as described in the previous section. At Gif, animals are fed twice daily, in the morning and at the beginning of the afternoon, and taps are kept closed during the day, isolating the tanks from the main circuit, to avoid the loss of algae in the circulating system.

At Shimoda, the Ciona reach a length of 5 to 7 mm within the first week (4 weeks after fertilization) of culturing in tanks. At this stage, their appetite increases and they become more tolerant to changes in water conditions. Diatoms are supplied every day at a concentration of 3,000 cells/ml. However, if diatoms are still present in the water the next day (as shown by the turbidity of the water) the number of diatoms added should be reduced.

At Shimoda, MSY (Microfeast, Spirulina, and Yeast) solution is introduced when the animals reach a length of 15 mm, giving a mixture of dried artificial diet and diatoms (3,000 cells/ml) (see Supplementary Materials #7 and #10). Dried diets are more likely to pollute seawater rapidly. Pollution can be minimized by keeping the amount of dried food the amount that Ciona can consume entirely within a few hours. MSY solution is provided twice daily, whereas diatoms are provided once daily. Where possible, it is better to increase the frequency of feeding without changing the total volume of MSY solution provided. As reported by Cirino et al. (2002), a perister pump can be used to supply MSY solution continually. However, this requires additional space so it is only really practical when high growth rates are desired.

Initially, 5 ml of the MSY solution is added to the tank twice per day. The amount of MSY added is gradually increased, 5 ml at a time, to 15 ml (i.e., 5, 10, 15 ml) according to the size and appetite of Ciona. A rough guide (Supplementary Material #7) is helpful but adjustments should be made at each feeding, according to the size of Ciona, the clarity of the seawater, and the presence of digested materials in the intestine. For example, if the water is clear and the intestines of the Ciona contain little or no digested material (white in color), then the amount of MSY should be increased. If sufficient MSY solution is given, the intestine contents should be dark green or brown. If the intestine is empty, the amount of the MSY solution should be increased. Conversely, if the water is turbid, the amount of MSY solution should be reduced. The amount of MSY solution added should not exceed a certain limit (more than 20 ml) because it may rapidly cause water pollution. Ciona more than 4 months old sometimes have less appetite. In such cases, the amount of MSY solution should also be reduced.

Adults are also given some dried food in addition to live algae, particularly in the Shimoda system. A mixture of foods is essential to ensure the supply of the nutrients required for reproduction. Artificial foods, such as “Spirulina” (Earthrise Nutritionals), “Microfeast” (Microfeast® PZ-20; Salt Creek, Inc., Salt Lake City, UT) or Marine Deluxe (Coralfood, H&S Aquaristik, Germany) are efficient in open systems (Cirino et al.,2002). In closed systems, large particles should be filtered out before dried diets are supplied to Ciona (see Supplementary Material #10). These diets seem to contain sufficient nutrients for Ciona growth and reproduction. In addition to dried diets, frozen yeasts are added at Shimoda (Kyowa Hakko Kogyo, Japan) to supply several unsaturated fatty acids, such as DHA (docosahexaenoic acid) and EPA (eicosapentaenoic acid).

CONCLUSION

Further studies of closed systems are required, but the methods described here can be used for Ciona culture in inland laboratories, making it possible to work with Ciona intestinalis and Ciona savignyi worldwide. Currently, 10 transgenic lines are maintained at Gif, and 181 at Shimoda. Most of these lines express reporter genes in a tissue-specific manner. They express enhanced green fluorescent protein (EGFP) or yellow fluorescent protein (YFP) named “Venus” in each of the anterior nervous system, epidermis, muscle, notochord, and neural tissues (Fig. 3). These transgenic lines are useful for investigations of the mechanisms of tissue formation during embryogenesis and morphogenesis, and can save us from the need for in situ hybridization or immunostaining. Transgenic lines and Minos transposon constructs that were generated by the Shimoda team are available from http://www.shimoda.tsukuba.ac.jp/˜ciona/index.html.

Figure 3.

Stable transgenic lines showing GFP or YFP expressions in a specific tissue of larvae (A–D) and of juveniles (E,F; H.A., unpublished data). The lines were created with the Minos transposon at Shimoda and Kyoto (A–D) or I-SceI meganuclease at Gif (E,F). A: A transgenic line of Ci-TnI-BW:GFP, showing GFP expression in the muscle. Scale bar = 100 μm. B: A transgenic line of Ci-Nut:GFP, showing GFP expression in the central nervous system. C: A transgenic line of Ci-Bra:GFP, showing GFP expression in the notochord. D: A transgenic line of Ci-Epi-I:GFP, showing GFP expression in the epidermis. E: First ascidian-stage transgenic juvenile for a Ci-ROR:Venus construct, showing expression at the top of the endostyle (arrow). F: Detail on the neural complex of an adult transgenic for the same construct as in E. Expression is observed in the ciliated funnel (arrowhead).

Sperm from these heterozygous transgenic animals is used to fertilize experimental eggs, and approximately 50% of animals show GFP expression in a non-mosaic fashion. Routine exchanges between laboratories and resource centers will be required in the future to accelerate research activity. We repeatedly tested the feasibility of collaboration between laboratories on different continents to collaborate by sending dry sperm from a transgenic line, simply chilled on ice, fertilized eggs, and adult animals from Gif to Santa Barbara, with no loss of fertility. This demonstrates that ascidians from different locations, with high levels of variability, are interfertile. This method should also make it possible to select the populations or lines most suitable for use in the laboratory.

Experiments are also underway to improve the storage and shipment of frozen sperm. Our facilities will soon establish custom-produced lines, which will be made available via the recently funded AMAGEN platform at Gif (http://amagen.inaf.cnrs-gif.fr/index.html/). It has been announced that an NIH-funded stock centre for C. intestinalis and C. savignyi will be set up at Santa Barbara (CA, USA) by M. Levine and W.C. Smith (http://www.ascidiancenter.ucsb.edu). The numbers and sizes of Ciona facilities increase in several laboratories. D. Jiang (Sars, Norway), R. Zeller (UCSD, USA), and B. Davidson (UA, USA) are now establishing new facilities (personal communications).

Our initial aim was to develop relatively small-scale closed culture systems. These systems can be built and used at an even smaller scale, and are sufficient for the maintenance of a small number of lines in inland laboratories. We hope that the methods described here will help many laboratories far from the sea to set up facilities, and that this, together with the increasing numbers of laboratories located close to the sea, will make the Ciona research network more efficient.

EXPERIMENTAL PROCEDURES

Seawater and Temperature

When available, natural seawater is preferred for cultures of Ciona. Periodic controls of chemical and organic pollution are required, but seawater remains the best and cheapest resource for coastal laboratories. Some laboratories located at some distance from the sea periodically transport and store natural seawater (e.g., Nori Satoh's laboratory, Kyoto, Japan). However, this transport and storage is costly and, furthermore, in some cases (we have tested various sources at Gif), transported natural seawater becomes unsuitable for Ciona culture, for unknown reasons. Artificial seawater may be more stable in quality, an important property for the maintenance of successive generations of important transgenic animals.

Despite its high cost, artificial seawater should, therefore, be considered, particularly for inland laboratories. We have found this medium to be appropriate for Ciona culture. We have tested several artificial seawaters, such as Red Sea Salts (Red Sea Fish Pharm, Eilat, Israel) and REEF CRYSTALS salts (Reef Crystals®, Aquarium Systems, USA). Both products are suitable for animal culture but the American product seems to be the most efficient. Water must be treated by reverse osmosis before use, because sea salts do not dissolve correctly in tap water, which already contains minerals. In addition, tap water may contain pollutants, such as nitrates or heavy metals. A typical salt concentration of 35 g/l was used, but concentrations can be adjusted according to salt conditions at the site of animal collection. Dybern (1967) investigated the effect of salinity on different local populations, and found that unusual conditions caused stress, particularly during embryonic development. It is, therefore, important to avoid salinity stress. Water salinity should be checked with a conductivity meter, to check that the artificial salts are correctly dissolved. If natural seawater is used, salinity checks are required only after long periods of rain.

Periodic monitoring of water characteristics is also required if seawater is not changed frequently. Nitrate concentration is a particular concern as nitrate is very toxic to ascidians at concentrations above 25 mg/l. Nutrient and mineral deficiencies can mostly be overcome by frequently exchanging seawater. However, iodine concentration tends to decrease over time, which lead us to add iodine to the water (potassium iodide concentrate, Reef Evolution®, Aquarium Systems, USA) and oligo-elements (trace element concentrate, Reef Evolution®, Aquarium Systems, USA), in accordance with the manufacturer's instructions.

Seawater temperature is maintained at 18° to 20°C at both Gif and Shimoda. Lower temperatures generally slow growth and inhibit egg maturation. Higher temperatures may lead to rapid water pollution. Empirically, 23°C seems to be the upper limit for rapid growth, but it is associated with rapid water pollution and increasing mortality rates (Cirino et al.,2002) and is, therefore, not generally recommended. Temperature-regulation devices can be used with small aquariums. We currently regulate water temperature by controlling air temperature throughout the facility. More information about the growth rates of ascidians originating from populations living at different sites would be useful (Dybern,1965).

Treatment of Seawater to Prevent the Escape of Genetically Modified Organisms (GMOs)

Sexually mature Ciona produces both eggs and sperm. If transgenic animals are cultured, the seawater in the tank is, therefore, likely to contain transgenic embryos and larvae. The water must be treated to kill animals later before its discharge. Hydrogen peroxide, detergents, or chlorine may be used, or the water may be heated or treated with UV light (Hendrickson et al.,2004). At Shimoda, used seawater is treated with 0.016% hydrogen peroxide overnight before discharge. Adults are not killed by this treatment, but are easily caught with a mesh. According to Carver et al. (2003), treatment with 5% of acetic acid, even for as little as 30 s, is effective, as is the exposure of the animals to freshwater conditions for a long period. This last method is particularly interesting given its low cost and the low risk of pollution.

Light and Shade

The light cycle is generally considered important for the maturation of the animals and for the generation of food algae in the tanks. A long photoperiod (14 to 16 h of daylight or 24 h daylight) may be applied. Round-the-clock lighting prevents daily gamete spawning, increasing the number of eggs available for experiments, but Ciona occasionally produces eggs in response to various stimuli such as water exchange, even under continuous lighting. Fertility does not decrease significantly in mild continuous lighting conditions, but ascidians seem to prefer shaded areas rather than direct “sunshine.” It may, therefore, be useful to cover holding supports with a roof (see Supplemental Material #7). This roof also helps to prevent the growth of macroalgae on the tunic of Ciona.

Specific light qualities should be used (Grolux, Sylvania). These conditions may lead to the growth of green or red macro-algae on the tank surfaces. For unknown reasons, animals grow less well in tanks covered with green algae. Red algal growth is a clear indicator of abnormal conditions (e.g., abnormal pH). In such cases, the seawater should be discarded and the tank replaced. Frequent seawater exchange, as in the Shimoda system, can prevent algal proliferation.

Animal Collection and Reproduction

The sexually mature adults used in Gif are collected from Roscoff Marine Station (North Brittany, France) and transferred to Paris within 24 hr, in oxygen-filled bags packed in polystyrene boxes, to which a bag of ice is added in summer. Animals are collected throughout the year, but some seasonal differences have been noted: for 3 or 4 months each year, typically from January to March, animals are less fertile and produce poor-quality eggs that are difficult to fertilize and cannot be electroporated. The animals are also sometimes adversely affected by heat during transfers to the laboratory in summer, with mass deaths observed on arrival. Sorted mature adults are allowed to rest for at least one or two days in 80-l tanks. They are then placed in tanks with continuous lighting for at least another two days, to stimulate gamete production. They can then be used in the following week.

Stable transgenic animals can be maintained in laboratory conditions for 5 to 6 months. During this period, they can be crossed with wild-type animals, with transgenic sperm generally used to fertilize wild-type eggs. Eggs from two wild-type animals are placed separately in Petri dishes (90 mm diameter, 25 mm high), and sperm collected from transgenic lines is added. The use of eggs from two animals ensures the production of healthy juveniles. The number of eggs is not controlled, but several hundred eggs are required to ensure continual metamorphosis. We use more eggs (about 1,000 or more) for founder screening than for isolating the next generation of transgenic lines, because the percentage of reporter gene transmission to the next generation is very low in some founders (reviewed by Sasakura in this issue, pages 1758–1767). Several hours after insemination, the seawater is removed with a pipette and fresh seawater is added. Fertilized eggs are cultured to the larval stage. Reporter gene-positive larvae are selected at this point if reporter gene expression occurs only before metamorphosis. The larvae are incubated for two days to allow metamorphosis to occur. In most cases, more juveniles are found around the edge of the dishes than at the bottom. For transgenic lines with reporter gene expression after metamorphosis, reporter gene-positive animals are selected at this point. Juveniles are staged as previously described (Chiba et al.,2004), and about 10 juveniles are selected for further culture.

In some cases, such as screening for potentially transgenic adults, it is sometimes necessary to obtain gametes from an individual without killing it. Progeny must be obtained by natural spawning in this case. The fertilization ratio is often low in 3-l glass beakers, due to the low sperm concentration. Transgenic animals cultured in facilities of this type are often small and it is harder to collect sperm surgically from these animals than from wild specimens. Pulled glass pipettes are, therefore, used to collect sperm from the top of the spermiduct, where most sperm is stored, according to Hendrickson et al.'s dissection method (2004). This technique makes it possible to collect enough sperm, even from small animals, for the fertilization of several batches of eggs.

Intact sperm can be kept for several days at 4°C and shipped overseas on ice. Frozen sperm can also be stored for long periods (see Corbo et al.,1997; Hendrickson et al.,2004). Genetic diversity is optimized and potential inbreeding depression avoided by collecting eggs from wild animals and using them for fertilization. However, we are currently trying to establish a healthy, fertile laboratory line, by crossing 20 pairs of animals independently and pooling the larvae (L.L., unpublished data). This should favor successive syngenic breeding when wild-type animals are not available due to seasonal conditions.

Contamination with Unfavorable Animals

Over time, other species may proliferate in and cover the surfaces of the tanks. These species include Mystiques (mollusks), serpulid annelids, sponges, and colonial ascidians. Copepods may be present in huge numbers. They seem to “visit” the pharynx of juvenile ascidians, hampering feeding by causing stress. These contaminating species should be eliminated if they become too numerous, for various reasons: they may act as reservoirs for pathogens, compete for food, and have long-term effects on the availability of oligoelements and iodine.

Contamination should be limited by keeping tanks containing wild-type animals as far as possible from tanks containing laboratory animals, preferably in another room. Seawater preparation tanks should be periodically checked for the absence of contamination. The clean dissection of gametes is recommended. For larvae and juveniles, seawater should be filtered and checked for contamination before placing plates in tanks. Tank walls should also be washed periodically with sponges. In Shimoda, tanks are cleaned by rinsing three times with tap water and once with seawater. The tap water could be evaporated or wiped off rather than rinsing with seawater. At Gif, water is initially filtered through 50-μ mesh and tanks are cleaned once a month by filling part of circuit with tap water.

Acknowledgements

We thank Bill Smith, Paolo Sordino, and Filomena Ristoratore for their helpful advice on ascidian rearing. We also thank Jean-Christophe Thiesen who helped to initiate algal cultures at Gif. Algal stocks were kindly provided from Roland Le Guedes and J.-P. Cadoret (IFREMER, Nantes, France). We thank Marc Vandeputte, who carried out the mathematical analysis of the results, Patrick Parra, Jean-Yves Tiercelin, and Pierre Noaro, who were involved in the design and construction of the rack equipment. We also thank Aurélien Drouard and facility members for the efficient and friendly environment. We thank Prof. Kazuo Inaba, Prof. Shigeki Fujiwara, Yuichi Oogai, Yoshikazu Okada, Yasuyo Kasuga, Yasutaka Tsuchiya, Toshihiko Sato, Yoshiko Harada, Katsutoshi Mizuno, Aru Konno and Hideo Shinagawa for helping us with culturing Ciona in Kyoto and Shimoda. We also thank all members of the Maizuru Fishery Research Station of Kyoto University, the International Coast Research Center of the Ocean Research Institute of the University of Tokyo, the Education and Research Center of Marine Bioresources, Tohoku University, Dr. Tatsuya Oshika, and members of Kobe Municipal Suma Aqualife Park, for collection of Ciona adults and seawater. The work carried out at Gif was supported by INRA, CNRS, the French GIS Institut de la Génomique Marine, and the European Commission (EaC key action QLK3-CT-2001-01890), Plurigenes (STREP project LSHG-CT-2005-018673), the Marine Genomics Network of Excellence (EU-FP6 contract no. GOCE-CT-2004-505403), and the ANR project CHORREGNET to J.S.J. T.M. was Predoctoral Fellows of JSPS. The research in Shimoda was also supported by Grant-in-Aids from the MEXT, Japan, to Y.S. (18770193), and N.S. and Y.S. (17207013).

Ancillary