News & Events

CTSA Project Update: Polychaete Culture in Hawaii: A Potentially Valuable Feed for Shrimp Hatcheries

Jul 27, 2020

Polychaete Culture in Hawaii: A Potentially Valuable Feed for Local Shrimp Hatcheries
Dustin Moss, Oceanic Institute of Hawaii Pacific University

Captive reproduction of penaeid shrimp requires conditioning and maturation of broodstock to stimulate gonadal development and induce mating, spawning, and ultimately the hatching of eggs to produce viable larvae. Broodstock diet is significant in the maturation process, especially in stimulating ovarian development in females. Most hatchery managers feed broodstock a mixed diet of raw, wet feeds such as squid, marine polychaetes, Artemia biomass, and shellfish, as well as formulated feeds. Polychaetes from several genera are a common component of maturation diets (used by 66% of respondents in a survey by Global Aquaculture Alliance). Popularity of polychaetes is related to their high concentrations of specific fatty acids. Because penaeid shrimp have a limited ability to synthesize the n-6 and n-3 families of fatty acids de novo, including poly-unsaturated linoleic and linolenic acids, or to elongate and de-saturate these into highly unsaturated fatty acids (HUFAs) such as arachidonic, eicosapentanoic, and decosahexanoic acids, high concentrations of these important HUFAs found in the ovaries of female broodstock have been attributed to the dietary intake of HUFA-rich items, such as marine polychaete worms. It should be noted that pelleted shrimp feeds do not contain high levels of HUFAs, nor do female shrimp readily develop mature ovaries/oocytes when fed pelleted diets.

Hawaii is a leader in the genetic improvement of penaeid shrimp and a major source of broodstock to farms locally and around the world. Currently, ~400,000 shrimp broodstock are sold by Hawaii-based breeding companies each year with an export value of ~US$20 million. To support breeding activities, extensive captive reproduction is needed. It is estimated (based on usage of polychaetes at OI extrapolated to the entire industry) that >10,000 kg of frozen marine polychaetes are imported into Hawaii annually to support shrimp breeding/hatchery activities (cost >$400,000 per year). The primary sources are wild-caught Glycera dibranchiata from the Northeast coast of the US (~$50/kg including freight) and cultured polychaetes (Nereis virens) imported from Europe (cost ~$33/kg). Major shrimp farms in Asia and Central America typically use live, wild-caught and/or cultured, local polychaetes. These worms are much cheaper (less than $10/kg), but are not a viable alternative to imported, frozen worms for Hawaii shrimp hatcheries due to biosecurity risks (bacterial and viral pathogens).

Researchers at Oceanic Institute of Hawaii Pacific University (OI), with support from CTSA, have collected and evaluated several local polychaete species for their aquaculture potential and use as a shrimp maturation feed, including Marphysa sanguinea, Lumbrineris japonica, Sabellastarte spectabilis, Malacoceros indicus, and Chaetopterus variopedatus. M. sanguinea (Fig. 1) was selected as the primary culture candidate based on its large size (up to 25 cm), high survival in culture, high palatability to shrimp broodstock, and excellent biochemical composition. OI has established a large breeding population of M. sanguinea and the population has tested negative for all known shrimp pathogens since the original founders were collected in 2013. Below is a summary of our findings to date.

Fig 1

Figure 1. Adult M. sanguinea (3.4 g) cultured at OI.

Basic Biology
M. sanguinea is a cosmopolitan species with a global distribution at sub-tropical and temperate latitudes. M. sanguinea typically exhibit patchy distribution within the intertidal to subtidal zones, where they are found in soft-sediment habitats. They are omnivorous, and leave their soft-sediment burrows at night to scavenge for food. There is a paucity of information on the natural diet of M. sanguinea, but it is believed that this species feeds on benthic macroalgae, benthic microalgae, and detritus.

Two locations on the windward coast of Oahu has been identified as collection areas for M. sanguinea. M. sanguinea are in moderate to high abundance (1-3 worms per shovel load of sediment) at these locations. However, M. sanguinea are mixed with other worm species and can be difficult to identify in the field (especially small specimens). In addition, field collection of tissue samples and PCR screening for disease are time consuming and expensive. Thus, collection of M. sanguinea (or any other worm species) in commercial quantities is not possible or at least not commercially viable. Furthermore, it would likely be very difficult to obtain a State of Hawaii collection permit for large-scale harvesting of worms.

A trial comparing the attractability and palatability of cultured and wild-caught M. sanguinea and commercial, frozen bloodworms (Glycera dibranchiata) was conducted. Results from this trial, showed no statistical difference in response time (time from introduction to shrimp finding the worm), holding/consumption time, and total consumption among the three worm types for Pacific white shrimp, Penaeus (Litopenaeus) vannamei. Thus, palatability of cultured M. sanguinea is not a concern.

Biochemical Composition
Biochemical analysis was performed on wild and cultured M. sanguinea, and frozen, commercially available bloodworms, G. dibranchiata to determine nutrient composition and fatty acid profiles. Eicosatetraenoic acid (PUFA Omega-3) was only found in M. sanguinea samples (both wild and cultured). The specific impacts/value of this fatty acid on shrimp reproduction is unknown; however, other Omega-3, PUFAs are known to be beneficial to shrimp reproduction (e.g. docosapentanoic and docasahexaenoate acid). The total quantity of fatty acids known to be beneficial to shrimp reproduction was highest by dry wet in wild M. sanguinea and approximately double the amount found in cultured M. sanguinea and G. dibranchiata. These results suggests that there is room for improvement in our feeding regime of M. sanguinea to optimize the nutritional value of the worms. Crude protein by wet weight was similar for M. sanguinea (wild and cultured) and G. dibranchiata. Crude lipid and lipids known to be beneficial to shrimp reproduction by wet weight were higher in M. sanguinea (wild and cultured) than in G. dibranchiata. Thus, M. sanguinea is potentially a superior maturation feed compared to G. dibranchiata, although this needs to be verified with shrimp reproduction trials.

Captive Reproduction and Sediment Evaluation
M. sanguinea larvae and juveniles have been collected from OI culture systems. M. sanguinea lay eggs in an “egg mass” within the sediment. Once eggs hatch, larvae (stage unknown) migrate to the surface where they, presumably, disperse as plankton. At OI, initial larvae collections were made by diverting effluent from a tank stocked with adults through a small tank lined with a 75 micron mesh bag. Larvae and juveniles of multiple sizes class were collected, suggesting that multiple spawning events occurred (Fig. 2).

Fig 2

Figure 2. Larval/juvenile M. sanguinea estimated to be 4 days old (A), 2 weeks (B), or 3 weeks (C).

Up to 150 larvae were collected in a single day from a standing crop of ~100 adults. Most larvae were competent to settle or at post-settlement stages at time of collection. Based on the known larval biology of this species, it is estimated that the larvae were 4-10 days old at collection. Despite strong evidence that worms were spawning in captivity, collection of large quantities of larvae needed for commercial farming operations was not achieved with this method.
For this reason, we investigated the idea of “inoculating” a virgin culture area with adult M. sanguinea and allowing the worms to mature (if needed) and naturally populate the area. For this trial, we divided two tanks into four quadrants each, with each tank having either commercially available sand or commercially available, small coral rubble (7.5 cm sediment depth; Fig. 3). Two quadrants of each tank were randomly assigned the low density treatment (15 adult worms) and the other two quadrants were assigned to the high density treatment (30 adult worms). Worms in each quadrant were fed larval shrimp feed every other day. After 273 days, 0.2 m3 of sediment was randomly collected from each quadrant (~20% of each quadrant) and the number of harvestable worms (>~5 mm) were collected and weighed by quadrant.   

Fig 3

Figure 3. Tank “A” divided into four quadrants and filled with commercially available sand.

Data (by quadrant) is provided in Tables 1 and 2 below. Neither sediment type nor stocking density had a statistically significant effect on mean weight of harvested worms or number of worms produced (#/m2 or #/m3). However, worms harvested from coral rubble had a mean weight of 0.618 g compared to 0.093 g for worms harvested from sand. Several worms larger than 2 g each were observed for coral rubble with the largest worm being 3.4 g (Fig. 1). Conversely, sand quadrants produced 568 worms/m2 (7,562 worms/m3) compared to 264 worms/m2 (3525 worms/m3) for coral rubble. Simply, the sand sediment produced a large number of smaller worms compared to the coral rubble that produced fewer larger worms. Furthermore, large numbers (>500) of very small, un-harvestable worms were observed in sand sediment, but were absent from coral rubble. Sediment type did have a significant effect on biomass produced with coral rubble producing 0.145 kg of worms/m2 (1.93 kg/m3) compared to 0.045 kg/m2 (0.59 kg/m3) for sand.

Table 1

Table 2

These results show that M. sanguinea will naturally propagate in either sand or coral rubble. However, it appears that sand is better for reproduction/larval survival (due to higher numbers of smaller worms) and coral rubble is better for growout of large worms (due to larger worm size and biomass production per unit area). Furthermore, these results show that commercially relevant production levels of M. sanguinea can be achieved with coral rubble quadrants having an average production of 0.145 kg/m2 or 1.93 kg/m3. Importantly, we feel (based on additional experiences) that these densities could be achieved in ~180 days (compared to 273 days for this trial); however, this may be seasonal and impacted by water temperature and natural spawning periods.

Demonstration of Commercial Scale Growout
A 125-m2 raceway was prepped and filled with 20 m3 of sediment. Neither sand nor small coral rubble (previously tested sediments) were available locally in large quantities, so small pea gravel (of volcanic origin) was used instead. Culture of worms in pea gravel had not been previously attempted at OI; however, the pea gravel is roughly the same size as the previously used coral rubble, so no significant problems related sediment type were anticipated. The raceway was “seeded” with larvae via effluent (containing planktonic larvae) from two tanks containing established populations of adult M. sanguinea. The worms were later consolidated into one (1) tank to make management easier.

After 220 days of seeding, the raceway had a worm density of 2,198 worms/m3, worms had a mean weight of 0.554 g, and production was 1.2 kg of worms/m3. Worm density and production were a bit lower than previous cultures in coral rubble (Table 2). The reason(s) for this are currently unclear. However, it is possible that reproduction in the adult worm tank(s) were relatively low and this resulted in a lower number of larvae being “seeded” in the raceway. Possible causes of low reproduction are: (1) worm stress associated with moving them to tanks adjacent to the raceway, (2) worm stress caused by consolidating the two adult tanks, and (3) seasonal impact (a previous study showed that maximum larvae production is May to September). It is also possible that the pea gravel is not as suitable as coral rubble; however, the worm population in the raceway was approximately the same average size as previous cultures in coral rubble (Table 2).

Figure 4. Harvesting system for Marphysa sanguinea

Figure 4

A harvesting system using a small concrete mixer was developed. The mixer is used to slowly churn the sediment while a seawater hose (~100 lpm) is used to flush the worms that have been separated from the sediment into a harvesting net (Fig. 4). The system can process ~0.03 m3 of sediment in about 3 minutes and yield ~36 g of worms. So, ~4 harvesting cycles (~0.12 m3 of sediment) were needed to obtain the amount of worms required each day (~140 g) for a shrimp maturation trial (Fig. 5; see section below). While this system works very well for harvesting relatively small amounts of worms, a better system will be needed if cultures (and harvesting needs) are expanded. 

These results show that commercial scale culture of M. sanguinea is possible. However, work on optimizing sediment, understanding reproduction output/trends, and further exploring harvesting techniques are likely needed to fully commercialize the culture of this species.

Figure 5. A sample (~80 g) of freshly harvested worms.

Figure 5

Impact on Reproductive Output of Shrimp
Reproductive output of shrimp broodstock (most notably the production of viable nauplii) is vitally important to hatchery operators. So, any new/replacement broodstock feeds need to be thoroughly tested prior to their use in commercial operations. To determine the efficacy of M. sanguinea as a maturation feed, a trial was conducted to compare the reproductive output of shrimp broodstock fed either live M. sanguinea or frozen, imported N. virens (an industry standard). Two round, shrimp maturation tanks at OI (15 m2) were each stocked with 33 female P. vannamei broodstock (Table 3). A third tank was stocked with 66 male broodstock. This resulted in a 1:1 sex ratio which is common in commercial hatcheries using this species. All shrimp were about the same age (i.e. ~6 month old) and were from the same sub-population within OI’s breeding program.

Table 3

The male tank was fed only maturation pellets at a rate of 4.0-4.3% BW/day. Female tanks initially received a daily diet consisting of 14% BW chopped squid, 6% BW polychaete worms, and 4% commercially available maturation pellets. One female tank was fed live M. sanguinea cultured at OI (Fig. 5. “Live” tank), while the other was fed cultured N. virens (“Frozen” tank) imported from a European supplier. Feed rates were adjusted several times during the trial based on predicted tank biomass, but the daily feed amounts were the same for both female tanks. Due to differences in actual growth and survival, with the Live tank being higher for both traits, the overall daily feed rates were slightly different between female tanks: 13.3% BW chopped squid, 6.1% BW polychaete worms, and 3.6% BW pellets for the Live tank; 14.3% BW chopped squid, 6.5% BW polychaete worms, and 3.8% BW pellets for the Frozen tank. 

Figure 6

Figure 6. Mature P. vannamei female broodstock eating cultured M. sanguine.

Matings began 5 days after stocking due to the presence of a large number of females with ovarian development. Each afternoon, mature females were collected and transferred to the male tank for mating. About 4 hr later, females were recaptured. Mated females were transferred to individual, 300-L tanks for spawning. Un-mated females were returned to their respective maturation tanks (as determined by tag code). About 07:00 the following morning, females in spawning tanks were returned to their respective maturation tanks. At 10:00-10:30 (shortly after hatching was completed), spawning tanks were vigorously mixed and three 100-ml samples were collected. The number of unhatched eggs and nauplii were counted in each sample to estimate total eggs (sum of eggs and nauplii), total nauplii, and hatch rate [(total nauplii / total eggs) x 100] for each spawn. The total number of mature females, total number of matings, and total number of spawns were also recorded for each female tank.

The trial lasted at total 36 days, with the mating period being 31 days. Mating were carried out on 19 days with no matings occurring on weekends or holidays due to associated labor demands. A total of 122 mature females (19% of population/day) were sourced in the Live tank, compared to 80 females (13% of population/day) in the Frozen tank. There were also more females mated for the Live tank (59; 9.4% of population/day) compared to the Frozen tank (43; 7.2% of population/day).

Table 4

Mean eggs/spawn and total egg production were much higher in the live polychaete tanks (30% and 74%, respectively) (Table 4; Fig. 7 & 8). Mean eggs/spawn was similar for the two tanks during the first few days of mating (Fig. 7), but was higher for the Live tank from day-6 of sourcing onward. Likewise, total egg production between the tanks was approximately equal for the first 7 days of matings, but then higher for the Live tank afterwards, with the magnitude increasing each day of mating (Fig. 8). 

Figure 7

Figure 8

Figure 9

Figure 10

Figure 7. Trend in mean eggs per spawn. Presented as 3-day averages to reduce variation.
Figure 9. Trend in mean nauplii per spawn. Presented as 3-day averages to reduce variation.
Figure 8. Trend in cumulative egg production for females fed live or frozen polychaetes.
Figure 10. Trend in cumulative nauplii production for females fed live or frozen polychaete.

Mean nauplii/spawn and total nauplii production were much higher in the Live tank (87% and 151%, respectively) (Table 4; Fig. 9 & 10). The extremely large difference in nauplii production (8.8 million vs 3.5 million) was a result of the Live tank producing more mature females, more spawns, more eggs/spawn, and a higher hatch rate (57% vs 40%). Mean nauplii/spawn (when using 3-day averages) was higher in the Live tank for the entire trial with the difference between female tanks being fairly constant (Fig. 9). 

In total, these results clearly show that females fed live M. sanguinea far outperform females fed frozen, culture N. virens. The reproductive superiority of females fed live M. sanguinea is supported by the fact that the females also grew faster than females fed frozen N. virens. It is unclear whether the positive effects of live M. sanguinea are simply due to the worms being live and/or if there are inherent differences in nutritional quality between M. sanguinea and N. virens.

M. sanguinea is an excellent culture candidate (for use as a shrimp broodstock feed) based on its large size (up to 25 cm), high palatability to P. vannamei broodstock, high survival in culture, and its acceptable to excellent biochemical composition (with regards to shrimp nutrition/maturation). Basic culture techniques have been developed and commercial scale culture of M. sanguinea has been demonstrated. In addition, female shrimp broodstock fed these worms have superior reproductive output compared females fed imported, frozen polychaete worms. Importantly, these results support the continued research and development needed to make M. sanguinea farming in Hawaii a reality. The availability of live, local worms could reduce or (hopefully) eliminate the need to import polychaetes by offering a superior product at a potentially lower price (freight charges can be >30% of imported worm costs).