Biological controls to manage Acropora-eating flatworms in coral aquaculture

Coral aquaculture is expanding to supply the marine ornamental trade and active coral reef restoration. A common pest of Acropora corals is the Acropora-eating flatworm Prosthio stomum acroporae, which can cause colonial mortality at high infestation densities on Acropora spp. We investigated the potential of 2 biological control organisms in marine aquaria for the control of P. acroporae infestations. A. millepora fragments infested with adult polyclad flatworms (5 flatworms fragment−1) or single egg clusters laid on Acropora skeleton were cohabited with either sixline wrasse Pseudocheilinus hexataenia or the peppermint shrimp Lysmata vittata and compared to a control (i.e. no predator) to assess their ability to consume P. acroporae at different life stages over 24 h. P. hexataenia consumed 100% of adult flatworms from A. millepora fragments (n = 9; 5 flatworms fragment−1), while L. vittata consumed 82.0 ± 26.76% of adult flatworms (mean ± SD; n = 20). Pseudocheilinus hexataenia did not consume any Prosthiostomum acroporae egg capsules, while L. vittata consumed 63.67 ± 43.48% (n = 20) of egg capsules on the Acropora skeletons. Mean handling losses in controls were 5.83% (shrimp system) and 7.50% (fish system) of flatworms and 2.39% (fish system) and 7.50% (shrimp system) of egg capsules. Encounters be tween L. vittata and P. hexataenia result in predation of P. acroporae on an Acro pora coral host and represent viable biological controls for reducing infestations of P. acroporae in aquaculture systems.

Control of pests of Acropora spp. coral is highly desired, given that it is the most represented genus imported into many countries globally (Rhyne et al. 2014), and Acropora spp. are commonly used for reef restoration efforts (Barton et al. 2017). A problematic coral pest, Prosthiostomum acroporae (Rawlinson, Gillis, Billings, & Borneman, 2011), commonly known as the Acropora-eating flatworm, has plagued hobbyist aquaria for many years (Delbeek & Sprung 2005). P. acroporae is an obligate associate of Acropora spp. and actively consumes coral tissue, which results in characteristic ~1 mm circular pale feeding scars, often resulting in coral tissue necrosis. Infestations are associated with colonial mortality at high densities in captivity (Nosratpour 2008). P. acroporae infestations are challenging to detect because of their highly cryptic nature, which facilitates their spread into new systems undetected. Infestations impact coral health through reduction of host coral fluorescence over time and hinder the coral's ability to photoacclimate to changes in lighting conditions (Hume et al. 2014). Infestations are often not de tected until compromised host health is observed through visual signs, at which point flatworm population density is high and colonial mortality of the coral may occur. There is no current empirical evidence to support effective treatment or prevention measures for P. acroporae infestations, although Barton et al. (2019) examined the life cycle under a range of temperature conditions and suggested timed intervention to disrupt the life cycle.
The aim of the present study was to evaluate the potential of 2 biological controls to reduce infestation by the Acropora-eating flatworm P. acroporae on coral. Biocontrol candidates included the peppermint shrimp L. vittata (Stimpson, 1860), which has been previously reported to remove parasites on fish and in the environment (Vaughan et al. 2017(Vaughan et al. , 2018a, and the wrasse Pseudocheilinus hexataenia (Bleeker, 1857), based on anecdotal evidence that it may reduce P. acroporae populations in aquaria through active foraging (Delbeek & Sprung 2005). This study examined the efficacy of potential biocontrols on adults and eggs of Prosthiostomum acroporae in captive systems over a 24 h period in vivo.

Species selection, husbandry, and culture
Twenty Lysmata vittata and 10 Pseudocheilinus hexataenia were purchased from Cairns Marine, Cairns, Australia, and maintained for 1 mo before any experimentation. Because of space limitations, shrimps were housed together in one 50 l flowthrough aquarium system (10 turnovers d −1 ) with approximately 5 kg of 'live' rock for hiding and protection between molts. P. hexataenia were housed individually in 50 l flow-through aquarium systems (10 turnovers d −1 ) with a 60 mm PVC tee (3-way junction) each for shelter. Filtered seawater (0.04 µm nominal pore size) at 27°C was used to supply the system. Shrimps and fish were fed twice daily to satiation with a mixture of thawed Tasmanian mysid shrimp, Ocean Nutrition ® Marine Fish Eggs, Ocean Nutrition ® Cyclopods, and Vitalis ® Platinum formulated feed. Animals were fed the morning prior to the commencement of each experimental trial but not during their trial period.
Adult Prosthiostomum acroporae were collected from a culture of infested captive Acropora spp. colonies. Flatworms were maintained in culture using established methods (see Barton et al. 2019).

Coral fragment preparation, infestation, and egg collection
To provide A. millepora for biological control trials, 96 A. millepora fragments (approximately 50 mm height; 30 mm width) were generated from donor colo nies harvested from 2 colonies sourced from Davies Reef, Australia (harvested September 2017; GBRMPA Permit: G12/35236.1), and 5 captive colonies originating from Orpheus Island, Australia (harvested May 2016; G14/36802.1). A combination of bone cutters and a band saw (Gyrphon ® Aquasaw XL) was used to prune A. millepora fragments, which were then fixed onto aragonite coral plugs (32 mm diameter) with cyanoacrylate glue.
To infest A. millepora fragments with P. acroporae, fragments were housed temporarily in individual 5 l containers. Before the start of each experimental trial, 5 P. acroporae individuals, approximately 3 mm in size, were directly pipetted onto each A. millepora fragment. After 60 s, each fragment was gently shaken to ensure P. acroporae had laterally appressed themselves to the host coral's tissue and were not stuck in the coral mucus (flatworms can dislodge if stuck in mucus). Any worms that detached were attempted to be reattached once and then discarded for another specimen if unsuccessful.
Egg capsules were naturally laid on Acropora skeleton in the P. acroporae culture and then harvested using bone cutters to remove the section of skeleton with these eggs. The underside of each subsequent skeletal fragment was glued onto clean aragonite disks or 'frag plugs' with cyanoacrylate glue. The number of eggs per cluster was determined by counting them under a dissecting microscope (Leica EZ4, 10−40× magnification) while immersed in seawater to prevent desiccation. Only fragments of coral skeleton bearing unhatched and undamaged egg capsules were selected for experimentation.

L. vittata experiments
Experiments with L. vittata were conducted on 4 separate trial days (i.e. 6 control and 6 treatment replicates per trial; n = 24 control; 24 treatment). On the day before each L. vittata trial, a random number generator was used to designate treatments and controls to aquaria. PVC blocks (80 × 80 × 25 mm; 32 mm diameter depression with central 10 × 15 mm hole to hold 32 mm diameter aragonite plugs in all replicates) were placed in each aquarium (3.5 l) before each trial. After their morning feeding, 6 L. vittata were haphazardly caught from their holding system using a 500 ml wide−mouth container and placed into their respective experimental tanks. L. vittata were given a minimum of 2 h to acclimate to their surroundings in the replicate experimental flowthrough aquaria (5 l h −1 ) maintained at 27 ± 0.1°C. L. vittata were considered acclimated once they settled on the bottom of each aquarium.
A. millepora fragments (1 per aquarium) infested with 5 P. acroporae each were introduced to each of the 3.5 l aquaria (treatment and control) for 24 h to determine if the presence of L. vittata (treatment) influenced the number of remaining flatworms on each coral fragment. The number of flatworms remaining was determined using a seawater screen-ing method (Barton et al. 2019). In addition, the PVC blocks and clear tanks were inspected for flatworms with the naked eye after each trial, with any flatworms found added to the remaining total of flatworms. Experiments examining the influence of L. vittata on P. acroporae egg capsules were conducted using the same approach, with the exception of egg capsules being counted before and after the trial under a stereo microscope (Leica EZ4, 10−40× magnification). Skeletal fragments (n = 48) were divided equally across treatments and controls (i.e. n = 24 control, 24 treatment) in L. vittata trials with 47.27 ± 19.09 (mean ± SD) egg capsules per fragment. L. vittata do not forage immediately before or after molting (D. Vaughan pers. comm.), therefore any shrimps that molted during the 24 h trial were excluded (i.e. 4 replicates were removed due to molting; n = 20).

P. hexataenia experiments
P. hexataenia (n = 9) were acclimated for approximately 2 wk to their randomly allocated flow-through aquaria at 27 ± 0.1°C with PVC blocks in place. The 50 l aquaria (n = 9 with wrasse, 9 without) were separated by black plastic because of the acute eyesight and territorial behavior of P. hexataenia. After acclimation, each fish regularly accepted food and did not exhibit signs of physical or behavioral stress.
Following morning feeding of P. hexataenia, infested A. millepora fragments (5 flatworms each) were introduced to each 50 l aquarium and left for a duration of 24 h to assess if the presence of the wrasse influenced the number of flatworms remaining on each coral fragment. Flatworms were recovered using an established screening method (Barton et al. 2019). The surfaces of the aquaria and the PVC blocks holding the fragment plugs were inspected visually for any remaining worms, which were added to the total remaining flatworms if present. Experiments examining the influence of P. hexataenia on P. acroporae egg capsules were conducted similarly, but egg capsules were counted before and after in spection with a stereo microscope (Leica EZ4, 10−40× magnification). The 18 skeletal fragments used in P. hexataenia trials (n = 9 treatment, 9 controls) had 42.33 ± 16.95 (mean ± SD) egg capsules per skeletal fragment.

Statistical analysis
Binomial generalized linear mixed models (GLMMs) and generalized linear models (GLMs) were gener-ated in RStudio (Version 1.0.143; R packages 'car, ' Fox &Weisberg 2019, and'lme4,' Bates et al. 2015) to assess the effect of L. vittata treatments on P. acroporae egg capsules and individual flatworms. Treatment was considered a random effect and trial identity a fixed effect in the model to ensure that there were no effects that changed the results significantly (p < 0.05) between L. vittata trials. Lacking any significant effects from trial identity in both experiments testing L. vittata egg and individual consumption, the GLM with pooled data denoted any significant effects (p < 0.05) of treatment on consumption for each experiment. Four replicates were removed from statistical analysis of the L. vittata vs. egg capsule experiment because these replicates molted during the experimental trial. Kruskal-Wallis tests were used to assess the results of P. hexataenia experiments with a significance threshold of α = 0.05.
Lysmata shrimps use their setaecovered antennules to detect chemical cues (via cuti cular sensilla) from their environment and locate suitable prey items (Zhu et al. 2011, Caves et al. 2016. Because they do not use visual mechanisms to locate and capture prey, L. vittata predation on P. acroporae is not hindered by the camouflage of these flatworms. However, L. vittata must physically encounter P. acroporae eggs or individuals while foraging to consume them, thus potentially limiting their ability to control P. acroporae populations in larger aquaria (aquaria > 3.5 l were not tested in this study), where the probability of a direct encounter would be limited by proximity and the availability of alternate food sources (L. vittata were not fed during the trials). Despite this possible limitation, L. vittata remain useful as a potential treatment of P. acroporae infestations because intimate cohabitation with Acropora enables shrimp to scavenge among coral branches and consume P. acroporae individuals and egg capsules. L. vittata are also an aggregating species and can be kept in high numbers when provided with sufficient food and shelter (Vaughan et al. 2018b). Future research could examine diet preferences of L. vittata, which may contribute to their efficacy in removing flatworms from Acropora colonies (e.g. Grutter & Bshary 2004).
Experimental trials with Pseudocheilinus hexataenia demonstrated that these fish are effective at reducing the P. acroporae population, with their presence having a significant effect on flatworm abundance remaining on A. millepora fragments (Kruskal-Wallis; p < 0.001). All P. acroporae exposed to P. hexataenia were removed over 24 h (100%; n = 9), compared to a loss of 7.5 ± 13.92% of flatworms (mean ± SD; n = 9) in controls. In contrast, all egg capsules were recovered intact in the experimental treatments (100%; n = 9) when cohabited with P. hexataenia. In the control, 2.39 ± 3.84% egg capsules (mean ± SD; n = 9) were not recovered, resulting in significant differences between treatment and control (Kruskal-Wallis; p < 0.05), likely from incidental mechanical damage to egg capsules through handling.
These results indicate that P. hexataenia is highly efficient at eating flatworms using well-developed eyesight (Gerlach et al. 2016) but does not interact with the hard shell of flatworm egg capsules. The implementation of P. hexataenia as biological controls must consider their ecology and husbandry requirements. In the wild, these fish actively forage in their established territory (Geange & Stier 2009, Geange 2010, generally only coming together for mating purposes (Kuwamura 1981). While their foraging behavior appears similar in captivity, the solitary and territorial nature of P. hexataenia renders keeping more than 1 individual in smaller aquaria (e.g. <1000 l) problematic. More than 1 individual could be kept in aquaculture systems large enough to avoid territorial confrontation, but the 'patrol' range of this territory may remain relatively constant. It is for this reason, combined with the fact that this fish does not interact with flatworm egg capsules, that they may not be as suitable for treating acute infestations of P. acroporae compared to L. vittata. However, their performance in our trials suggests that this colorful labrid is a useful tool for consuming adult flatworms, thus mitigating the chronic impacts of a given P. acroporae infestation by removing or reducing the P. acroporae density to non-lethal levels for the Acropora host.
P. hexataenia and L. vittata identify prey items in different ways while foraging, which has implications for how they are used in the captive environment and their ecological roles in native ecosystems. Little is understood about the dynamics of wild P. acroporae populations, although our results may provide further understanding of the trophic relationships between P. acroporae and natural predators in reef ecosystems. P. acroporae are cryptic and there are no documented infestations causing colonial mortality of Acropora colonies in the wild. It does remain likely that some proportion of wild mortality of Acropora colonies attributed to other causes (e.g. sedimentation and algal competition) is instead experiencing negative secondary effects on coral health from P. acroporae infestation. However, the presence of natural predators of P. acroporae (e.g. P. hexataenia and L. vittata) may reduce incidences of mortality in wild Acropora colonies.
In captive systems, pairing both of these biological control organisms with the manual removal of P. acroporae egg clusters is likely to be highly effective in reducing the overall infestation within a given aquarium system. However, consideration must be given to the sustainable supply of the organisms if used as biological controls. L. vittata are available through the ornamental trade and can be bred in captivity. Although peppermint shrimp species from other regions (e.g. L. wurdmenii, L. boggessi, Rhyne & Lin 2006) were not investigated in the present study, they could also be examined for their ability to interact analogously with P. acroporae and could be supplied sustainably for biocontrol of flatworm infestations. Although P. hexataenia is categorized as Least Concern (Bertoncini 2010;IUCN Red List 2010), overharvesting for use as biological controls in the ornamental trade could impact local populations. Lessons should be taken from the Scandinavian salmonid industry, where harvesting of wrasse broodstock used for biological control of sea lice parasites has exerted considerable pressures upon wild populations (Brooker et al. 2018, Powell et al. 2018.
In summary, this study provides the first empirical evidence of potential biological control organisms for P. acroporae in captivity. The ability of both L. vittata and P. hexataenia to consume P. acroporae renders them useful preventative measures of infestation in addition to potentially being used to treat colonies infested with adult flatworms and thereby drastically reducing the impact of this pest on captive colonies. While P. hexataenia had no apparent interest in P. acroporae egg capsules, L. vittata displayed the added benefit of consuming egg capsules through their foraging activities, with encounters with the egg clusters likely to further control the flatworm populations in captive systems. The addition of sustainable biological control organisms adds a valuable tool for flatworm control, which is suitable for both aquarium hobbyists and large-scale coral aquaculture facilities.