In the continuing quest to understand the benefits of kleptoplasty, the Slug Club students and I set up some experiments to determine whether Elysia mucus caused predators to avoid them, and, if so, which components.  The logic behind the experiment was that 1) both the literature and my own observations indicated that slugs taste bad to predators; and 2) kleptoplasts produce organic compounds that appear in the mucus (Trench et al., 1972, Biol. Bull. 142:335); so maybe kleptoplasty plays a role in making defensive chemicals that keep fish and predators from eating them.

For these experiments, Zvi Kellman at IBBR was kind enough to provide some space for a predator setup.  We got to set the system up, and not have to worry about vacating the space for the entire semester.

The first small hurdle was to find some predators.  Because my preliminary experiments were done with fish, I had originally planned on testing fish from the same range as the E. clarki, maybe bluehead wrasse or small puffers. Then I was reminded of the paperwork for doing experiments with vertebrates, which is required even if we just watch what they eat and don’t do anything mean to them.  Dreading the possibility that time would evaporate while I assembled forms and waited for replies, I became much more enthusiastic about invertebrate predators.

But which invertebrate predators?  The primary requirements were that they were plausible natural predators, from the same geographic locations (i.e., sympatric), that we could obtain multiple individuals of the same species, and that they would not be too expensive or hard to keep.  I was hoping to find some small, nasty crabs that local aquarists or shops had extracted from their tanks, but nobody seemed to have any available.  Mike Middlebrooks had tried blue crabs, but they seemed rather large for our facilities.  I finally settled on peppermint shrimp (Lysmata wurdemanni).  They are sympatric, and are certainly willing to tear apart and eat other small, squishy invertebrates.  KP Aquatics sells them in lots of 5, so I could set up multiple groups of them without going broke.  Even if they don’t naturally prey on Elysia (nobody knows), they require no encouragement to eat, so we could at least test whether they are discouraged by slug slime.

We obtained a used 40 gallon breeder tank from a local aquarist (Thanks Rik), and moved in.  We had a sink and a plentiful supply of purified water located a few feet away, so it was pretty easy to get set up.

Bench space at IBBR, with 40 breeder tank. 2/2/18

I siliconed in some acrylic dividers to have five independent sections.  Each section had a drain leading to the sump, which had a  skimmer and a small filter.  The output of the return pump fed each section independently, and flow to each was adjusted with a ball valve.  Each section also had a plastic plant and cave assembled from 1.5″ PVC fittings.  At the beginning, each section had 4-5 shrimp.

Tank with dividers in place, just after adding water and salt. Each section drains to the 10 gallon sump on the right, with a protein skimmer and filter. Water is pumped back to shrimp enclosures, with flow controlled by ball valves that can be seen at the top of each section. 2/17/18

We messed around for a few weeks, then settled on an approach inspired by Becerro et al. (2001; J. Chem. Ecol. 27:2287).  They made cubes based on powdered fish food and carrageenan, added potentially aversive compounds, and then placed them in the ocean to examine the effects on feeding by fish.

Slug Club Spring 2018, making food cubes. From left: Kathleen, Claudia and Elyson. 4/18/18.

The students took the lead in working out how best to make carrageenan based food.  They figured out how to get the carrageenan dissolved, when to add the powdered food (Zoplan freeze dried zooplankton by Two Little Fishies), and the best way to add mucus or slug tissue to experimental cubes.  The hot liquid food was poured into silicone ice cube trays to make as many 1 cm cubes as needed.  Silicone O-rings, used later to secure the food cubes in the shrimp tanks, were held in place with plastic rods during pouring.

Ready to pour food. Silicone O-rings, threaded onto plastic rods, are placed into compartments of ice cube tray.

The food cubes were then placed in the shrimp enclosures.  Each was secured by its O-ring to a plastic paperclip tied to a monofilament line suspended vertically in water column.  Each line had multiple possible attachment sites.  The shrimp were not shy, and rapidly moved in to explore the potential food source.

Testing food assay system. Food cubes are secured by O-rings to plastic paper clips that were tied to monofilament line. Line is weighted by PVC elbow at bottom, hooked to input hose above.  Shrimp rapidly approach and sample the cubes. 3/28/18.

Experiment 1 tested to effect of mucus on feeding.  Two cubes, one control and one experimental, were placed in each section.  The control cubes were made with powdered food, carrageenan and artificial seawater (ASW); the ASW was replaced by mucus in the experimental cubes.

Mucus dripping from annoyed slugs in dip net. 4/9/18.

Collecting the mucus was not trivial.  Even though the slugs readily secrete mucus when annoyed (such as when crowded into a dip net), the sticky mucus is not easy to separate from the slugs.  Ultimately, we just let it drip into a beaker for 30 minutes or so, then returned the irritated by unharmed slugs to their tank. The mucus was then mixed into the food, and the unused portion frozen for later use.

Detail of food cube, showing embedded red silicone O-ring connected to plastic paper clip (yellow). 3/28/18.

To control for effects of position in the tank, food cubes from both groups were placed at different levels.  We recorded the weight before adding to the tanks, and the position in the tank, then left the cubes with the shrimp for 24 hours.  They were collected, and each was weighed again. The amount consumed was then calculated.

The results from the first round of experiments were underwhelming, but informative.  The shrimp were unequivocally indifferent to the presence of mucus in the food.  The graph below shows that shrimp consumed about 0.5 grams per day of food, regardless of whether it contained mucus.

Experiment 1: Effect of mucus on feeding by shrimp (means and standard errors). Control food cubes contained 10 ml ASW, mucus cubes had 10 ml of mucus from Elysia clarki.  High food content (5g/90 ml).  Sample sizes are 25 cubes for each group.

So, shrimp do not seem to mind mucus in their food.

The next step was to determine whether Elysia even taste bad to shrimp.  Peppermint shrimp do, in fact, eat a number of nasty things, including the anemone Aiptasia.  We had predicated our experiments on the idea that Elysia contain compounds the shrimp would find distasteful, but we never actually tested this idea.  It was time for a “Friday Afternoon Experiment,” in which one tries a quick and dirty test to see what happens.  In this case, we threw some baby Elysia to the shrimp.

Shrimp devouring baby slug during first trial with live Elysia. 3/26/18

The results were horrifying and illuminating.  Shrimp rapidly grabbed, dismembered and ate the baby slugs.  However, they left fragments of uneaten slugs, which gave a glimmer of hope that they did not find them to be the most palatable meals.  We repeated the experiment twice more, with a week in between, and the shrimp consumed less at each session, as if they had learned not to eat them, but consistently attacked and tore at the slugs.

Although brutal, there are some lessons here.

First, shrimp should definitely be included in the list of possible predators.  Naive shrimp had no qualms about eating Elysia tissue, and they would be well placed to hunt down small mollusks during their nightly forays to rummage the reef.

Second, the damaged slugs often recovered.  The usual examples of toxic or unpalatable species (think monarch butterflies) are often killed when they are sampled by predators.  Kin might benefit from future behavior of the predator, but the sampled individual is out of luck.  In this case, the slug may be able to limp away after having been tasted and rejected, and benefit directly from containing aversive compounds.  This may also explain why Elysia lack the bright warning coloration often associated with unpalatable prey.  Maybe the first line of defense is to be inconspicuous, and, if that fails, expect to be spat out after being tasted.

Experiment 2 was based on a suggestion by one of the students: if shrimp don’t mind mucus, what about the slugs’ tissues?

We anesthetized slugs, and worked out how to produce tissue puree.  We quickly learned that a standard mortar and pestle has little effect on tissue is both slimy and rubbery.  Chopping with a razor blade was better, and produced adequate tissue for the first round of experiments.  Later, I had to cull a large number, and used a food processor to make a tissue homogenate.

Experiment 2: Effect of mucus and slug puree on feeding. Medium food content (2.5 g/90 ml).  20 cubes per group.

The results were consistent with our expectations.  Mucus alone had little or no effect on feeding.  Slug tissue appeared to reduce feeding by more than 35%.  However, the results were not statistically significant, possibly due to variability between groups of shrimp.

We accidentally performed an additional experiment on the effect of food concentration on the rate of feeding and the aversive effect of slug tissue.  For one batch of food, we ran out of powdered food, and used less than half the normal amount.

Experiment 3: Effect of mucus and slug puree on feeding. Low food concentration (1.16 g/90 ml).  12 cubes per data point.

The shrimp ate less overall, and the effect of including slug tissue appeared to be stronger, reducing feeding by more than half.  As above, the results were not statistically significant, presumably due to the small sample and variability between sections.  Nonetheless, balancing the attractiveness of food versus the distastefulness of slug tissue needs to be considered for future experiments.

Relationship between amount of food cubes eaten (means and standard errors) and concentration of food powder. All groups show a modest increase in consumption when cubes have more food content. Cubes containing slug tissue (green line) appear to be less palatable at both food concentrations tested.

For the moment, we can tentatively conclude that mucus from Elysia has no impact on feeding by peppermint shrimp, but that slug tissue inhibits feeding.

Because Mike Middlebrooks tried similar experiments on blue crabs with adult slugs, the last experiment of the semester was to put full grown Elysia into the shrimp enclosures.  Normally approach food rapidly, but were very cautious.  When slugs dropped right in front, picked with pincers, but did not tear at the slugs.

In the future, there are several improvements that could help to sharpen the results.

One issue that arose was the change in shrimp populations in the sections over time.  A few shrimp were lost (the absence of carcasses suggests cannibalism) over the three months of the experiment, yielding a final range number of 2-5 shrimp per section, which resulted in considerable inter-section variation in food consumed in a 24 hour period.  Doing a periodic census and adjusting populations will presumably reduce variability.

Another possible confounding factor as that some cubes appeared to be torn apart and left on the bottom, and were therefore scored as eaten when they may have only been consumed partially.  Using video analysis of time spent feeding on each type of cube would help distinguish between destruction and consumption of food.

One other issue was the growth of algae during the semester.  Despite relatively low light in the lab, protein skimming and periodic water changes, the addition of food supported algae growth.  Given the omnivorous nature of the shrimp, the algae may have served as an alternative food source. Giving the enclosures a quick scrub during weekly maintenance would force the shrimp to focus on the food cubes.