One of my Slug Club students recently asked me whether it is really worth pursuing the study of obscure, squishy creatures in an era during which biological science seems obsessively focused on “translational” research to apply scientific discoveries to clinical problems.  My response was that, while translational research is unquestionably important, the study of the biologically weird is where we discover unexpected fundamental truths about the way organisms function.  This new insight often lays the groundwork for new areas of research, and can, in time, lead to improvements in people’s lives.

Which brings us to the paper at hand: Zan et al. (2019) “A microbial factory for defensive kahalalides in a tripartite marine symbiosis.”  Science 364: eaaw6732 (published online ahead of print).  This paper not only taught me something new about Elysia, but I discovered methods and biochemical pathways I had never heard of.

For those who do not want to bob and weave their way through the entire journey of discovery, here is the quick summary: The authors discovered a bacterium living within the species of Bryopsis algae upon which Elysia rufescens feeds. This bacterium is responsible for the production of a chemical compound called kahalalide F (hereafter “KF”).  The following observations are what make the story more interesting:

1. The bacterium makes KF by stitching together amino acids using a biochemical pathway very different from that used by animals and plants.
2. The bacterium has adapted to the comfortable life inside plant cells by getting rid of many of its genes. 
3. Although E. rufescens concentrates KF in its tissues to deter predators, it does not maintain the bacteria in its tissues.  This contrasts with the chloroplasts from Bryopsis, which the slug separates from the other cellular components and continue to photosynthesize.

If that sounds interesting to you, then read on.

It was very exciting to discover a paper about E. rufescens and its food plant Bryopsis, in the journal Science during my lunchtime reading this week.  Science is one of the two most-read scientific journals in the world, so any article must be splashy and important enough for a wide readership.  That a study about Elysia was deemed worthy of publishing there pleased me to no end.  To be honest, the paper is not really about E. rufescens, nor about Bryopsis, but about a bacterium that lives inside Bryopsis, but the alga and slug still play prominent roles.

A quick literature search for E. rufescens (or a peek at the E. rufescens page on this site), will show you that the species has been mostly studied because it contains a defensive molecule called Kahalalide F (KF).

Kahalalide F: A Complex Defensive Compound

At first glance, KF looks like an organic chemist’s nightmare (or dream, I suppose), but on closer inspection it is a peptide, i.e., a short string of amino acids. As described below, KF has a number of unusual features for a peptide. From an ecological perspective, Becerro and colleagues showed that KF is a strong deterrent to predation. KF has also been found to have anti-cancer activity, so there is a medical motivation to study it.

The source of KF was not completely clear. It was assumed to be produced by the Bryopsis species (not yet assigned a species name, so I will simply call it “Bryopsis“) on which E. rufescens feeds. The slugs not only avoid the toxic effects of KF, but they concentrate the compound in their tissues at a higher level than is found in Bryopsis. KF has a few weird chemical features that suggested to researchers that it might be made by microbes and not the alga. For example, there is a fatty acid stuck to one end of the molecule, and it contains a number of “non-proteogenic” amino acids, such as ornithine or dehydrobutyrine, which are features of compounds produced by bacteria.

To test their hypothesis that KF has a microbial origin, and to better understand how it is synthesized, the team used a wide variety of methods, including metagenomic, metatranscriptomic, evolutionary genomic, and biochemical analyses, along with some more traditional fluorescence microscopy. If all the “meta-” and “-omics” has your brain in a knot, don’t panic yet; I will do my best to explain the methods as we go.

First, they chemically extracted the Bryopsis from their collection location to be sure it contained KF. That being done, they extracted DNA from the Bryopsis and any microbes that happened to be in or on the algae when it was collected. Using high-throughput sequencing, which breaks up DNA into fragments and sequences the whole genome, they analyzed the ribosomal RNA from the sample. All organisms have ribosomes that enable them to synthesize proteins, each ribosome contains RNA that helps it function, and the ribosomal RNA from each species has a unique sequence. The Bryopsis sample contained RNA largely from three bacterial classes, one of which was the Flavobacteriia. Of the Flavobacterial sequences, one species comprised 74 to 92% of the sample, making it the most abundant species associated with Bryopsis. To jump ahead again, they ultimately found that this bacterium was a new species, that it was responsible for the synthesis of KF, and gave it the provisional name “candidatus Endobryopsis kahalalidefaciens” (candidate kahalalide-making from inside Bryopsis).

How did they get there? First, they had to figure out the general way the peptide was made. If you are like me, and focus narrowly on multicellular creatures such as animals and plants, then you could be forgiven for thinking there is only one way to make a peptide, which is essentially a short protein. Turns out there are at least two.

RiPP or NRPS?

In the more canonical approach to making a protein, the cell uses a DNA sequence as a set of instructions to assemble a chain of amino acids. The DNA is read by an enzyme that makes (“transcribes”) messenger RNA (mRNA), and that mRNA is read (“translated”) by a ribosome, which sticks amino acids together in the correct order. Because some of the amino acids in KF are weird, they would have had to be modified after being assembled, so the authors termed this possibility “ribosomal peptide synthesized and posttranslationally modified peptide” pathway, or RiPP.

There is, however, another way that bacteria make peptides, called non-ribosomal peptide synthesis, or NRPS. Unlike a ribosome, which can make any protein or peptide that can be encoded by an mRNA sequence, NRPS requires a collections of enzymes customized for each peptide. NPRS enzymes are composed of modules that assemble the amino acids, modify them as needed, add fatty acids, and so on. This was completely new to me, but there is apparently quite a lot known about it.

In the good old days, it might have taken a lifetime to identify the biochemical pathway that the bacterium uses to make KF. Now that researchers are capable of sequencing, assembling and analyzing whole genomes, it just takes a lot of really hard work. The researchers sequenced all of the genes from organisms associated with Bryopsis, looking for either:

1. A gene that encodes the sequence of a predicted peptide, which could then be transcribed to mRNA and translated into a peptide. This would indicate that KF is made via RiPP.
2. A cluster of genes that encode the number and type of modules that would be expected to synthesize KF, based on the sequences of known NRPS enzymes. Finding this would support NRPS as the source of KF.

They did not find a gene with the sequence predicted to encode the KF peptide, suggesting to them that KF was unlikely to be produced via RiPP.

KF Is Made Through NRPS

They combined the data they had from the high-throughput sequencing with a second method of sequencing that produced longer fragments of DNA sequence (“longer reads”), and assembled the entire genome of a bacterium. The genome contained multiple clusters encoding NRPS enzymes, one of which, which they called NRPS-8, had features expected to be required for KF synthesis:

1. A predicted condensation domain, which would be needed to chemically bind the fatty acid chain (red in the diagram above) to the first amino acid.
2. A total of 13 predicted NRPS modules, as would be expected for an enzymatic pathway that makes a peptide consisting of 13 amino acids.
3. Eight of the NRPS modules contain domains that are expected to convert normal L-amino acids to D-amino acids (shown in blue in diagram).

Based on these features, the authors were confident that they had discovered the enzyme that produces KF, and therefore that KF is synthesized by the bacterium through the NRPS pathway.

Including NRPS-8, which the genome sequence of the bacterium contained genes encoding a total of 20 NRPS pathways. Intriguingly, previous work had extracted at least 15 kahalalides from Bryopsis and E.rufescens. When the authors compared the predicted NRPS pathways to the structures of the kahalalides, they could match eight of the “chemotypes” (chemical compounds with the same amino acid sequence, but differing in modifications) to bacterial NRPSs. They conclude that “Candidatus Endobryopsis kahalalidefaciens” (hereafter “cEK”) is responsible for at least nine, and probably all, of the kahalalides produced by Bryopsis.

The Bacterium Is Losing Genes

Based on the complete genome sequence, it was clear that cEK has lost some genes. When they compared its genome to a closely related, but free-living species, they found hundreds fewer genes. The missing genes encoded proteins that performed such functions as DNA repair, detoxification, chemotaxis (moving toward or away from chemical stimuli), adaptation to unusual conditions, and, notably, synthesis of amino acids. All of these would be needed for living in a cruel, variable world, but may not be necessary for a comfortable life associated with an algal cell that provides stability and nourishment.

The fact that the authors were unable to culture the bacterium separately from the alga supports the idea that it is dependent on Bryopsis for survival. Further, by labeling DNA specific to cEK, they showed that the bacterium is found inside the plant cells (see panel from Figure 4, below).

Figure 4D from Zan et al., showing the location of bacteria inside a cell of Bryopsis sp. Yellow dots indicate bacteria labeled with a probe specific for “Candidatus Endobryopsis kahalalidefaciens,” and are indicated by arrowheads labeled “cEK.” Other bacteria that do not contain the cEK marker, are labeled red, and are indicated by the arrowheads labeled “OB.” The cell wall of the alga is labeled blue.

Therefore, cEK appears to be an “endosymbiont,” an organism that lives symbiotically inside the cells of another, and is probably completely dependent on its host alga for survival. In exchange, the bacterium provides a collection of toxic compounds that reduce grazing on the alga. The importance of kahalalide production to the bacterium is underscored by the observation that a full 20% of the genome is devoted to genes devoted to NRPS pathways.

Slugs Do Not Host cEK

As a devotee of Elysia, you are familiar with kleptoplasty, the phenomenon by which the slugs take the chloroplasts from their algal food and maintain them in their bodies as an energy source. If the slugs have a propensity to keep useful cellular components in their bodies, one might ask whether they also store the cEK as a constant source of KF to deter predators from eating them.

The authors confirmed previous work that E. rufescens‘ tissue contains KF, and went on to look for evidence of cEK bacteria in the slugs’ bodies. Using highly sensitive methods to hunt for cEK DNA in the tissues, they found no evidence that the slugs contained significant populations of the bacteria. Attempts at labeling tissues with markers specific for cEK were also unsuccessful, despite positive labeling of other bacterial species. They conclude that, in contrast to their theft and use of algal chloroplasts, E. rufescens digest the bacteria fully when they suck the sap out of the Bryopsis cells.

Conclusions

Once again, biology comes up with a story that is better than fiction. A bacterium infects an alga, takes up residence, and starts using an unusual biochemical pathway to produce nasty compounds that protect the alga from being eaten. Natural selection favors those algae that contain the toxins (and can tolerate them), so the bacteria are now found throughout the species, and possibly even more broadly. The sheltered existence inside the alga allows the bacterium to lose many of the genes it would normally rely on to survive in the outside world, eventually rendering it incapable of independent life. Elysia rufescens can tolerate the toxins produced by the bacteria, and develops mechanisms to concentrate the kahalalides in its tissues.

Do Byopsis pennata or B. plumosa, the favorite foods of E. clarki, contain cEK or a relative? I expect we will know soon enough.

Because the goal of this post was to bring attention to the original work, I have glossed over a huge amount of information regarding the methods used, the biochemistry of kahalalides, and the genomic structure of cEK. I would encourage anyone who is interested to get access to the original article at Science Magazine.

Ah, the Good Old Days, when “curiosity-driven” science wasn’t naughty.  Of course, one might not want to get too nostalgic when the woman whose work is the subject of this post is referred to as “Miss Lillian Russell” in the running head of the paper.  Apparently, gender and marital status were considered relevant in a scientific publication 90 years ago.

The primary goal of the Solar Sea Slug project is to study the neurobiological specializations of photosynthetic molluscs.  In order to chart the input and output pathways, we need a good roadmap of the central nervous system (CNS) and the nerves that connect it with the rest of the animal.  I had hoped that someone had generated at least a crude diagram, and had found some drawings of the nervous systems of Elysia viridis (Huber, 1993, J. Moll. Stud. 59:381) and E. crispata (Gascoigne, 1972, Trans. R. Soc. Edinburgh 69:137), but had not found much more.  I fully expected to map the peripheral nerves by myself, and we had even done some preliminary experiments with fluorescent labels for axons for Slug Club during spring semester.

As is often the case, the breakthrough came as I was burrowing through the references from one of the older papers.  Huber’s 1993 paper reviews the structure of the CNS of a wide range of gastropod molluscs, and includes a nice diagram of the central nervous system of E. viridis, but provides no information about the periphery.  I had skimmed the section describing the image, but had mostly focused on the figure itself.  He made a passing reference to earlier work by Russell (1929, Proc. Zool. Soc. Lond. 14:197), so, for the sake of completeness, I ordered the paper through inter-library loan, making it clear that I would like them to include any photographic plates, which are often in a section separate from the main body of the paper.

When I downloaded the paper about a day later, I was pleasantly shocked.  As part of a study of the taxonomy of the ascoglossans (= sacoglossans), Lillian Russell had performed a thorough, detailed analysis of the nervous system, nerves, and internal anatomy of E. viridis and a nudibranch, Aeolidia papillosa. She provided an excellent description of the fusion of about seven ganglia to produce the nerve ring that comprises the CNS, along with multiple excellent drawings of the CNS, the peripheral nerves and internal organs of E. viridis.

It is hard to know which image is my favorite, but I am very fond of the drawing below, which shows the ganglia in the head, the nerves that originate from them, and the terminations of the nerves in sensory organs such as the eyes and rhinophores.

The drawing below, of the CNS and nerves removed from the animal, is also quite impressive, showing the individual ganglia that make up the CNS, and the origins of the many nerves that carry sensory information in, and motor commands out.

Aside from publishing a similar description of E. clarki (which probably only differs in slight details), and leaving a note in her will to let me know about it, I can’t imagine how she could have done any more to help me out.  Although I had intended to do the work myself, anatomical description of this quality requires considerable patience and artistic skill.  I am not very patient, nor am I a good artist.  Thanks to Russell’s hard work, I am free to do my behavioral experiments and electrical recordings, and be in my happy place.

So, why is this useful?  In a general sense, it gives a starting place for finding neurons and circuits that generate behavior.  If you want to know where sensory information goes in the nervous system, finding the nerve that carries the axons from the relevant sense organ or part of the body is a good start.  Same goes for finding the motor neurons that cause muscles in a certain part of the body to contract.  As a specific example, if we are trying to find how information about light gets to the brain, we first need to find the nerves that connect to the eyes or (hypothetically) other sense organs that detect light or photosynthetic activity. Then we can record monitor impulse activity in the nerves, or from the specific neurons with axons in those nerves.

This is a nice reminder about a few things.

1. Possibly most importantly, we truly are standing on the shoulders of others as we try to move science forward.  I hesitate to say “on the shoulders of giants,” because, for all I know, Lillian Russell was of average height.
2. You really never know when something you did will be useful to another scientist.  Ninety years ago (90!), nobody knew about kleptoplasty, and the study was carried out because nervous system structure was considered to be a good marker for evolutionary relationships.  It is not reasonable to argue that every experiment that can be imagined must be performed, but during this time in which research needs to be either “translational” or “transformative” in order to be funded, it would be nice to make a tiny bit of room for work that is simply of high quality.
3. It may be necessary to dig deep into the literature to avoid reinventing the wheel.  Many of us are guilty of looking over the figures of a paper and skimming the text, rather than reading every word.  Time is short and papers can be long.   Nonetheless, some of the most relevant information to people in the field is in the body of the text, rather than in figures or tables.  Who knows what else I will find as I work my way through the old literature.

The Solar Sea Slug project has progressed by leaps during the past half year or so.  In addition to finding the roadmap described above, and establishing a self-sustaining colony of E. clarki, some of the students in the Neurobiology Lab course helped me to work out procedures for recording from neurons in the CNS.  There are still a few improvements  to be made, such as more easily getting the electrode through the sheath that surrounds the ganglia, but we have a little real data.  Below is an action potential that was stimulated by injecting 1 nA of current into a large neuron in the abdominal ganglion (Pierce’s “Parker Cell?”).

How are we doing?

Self-sustaining culture: 

Neuroanatomical roadmap:

Dissection techniques: 

First intracellular recording:

This project is really starting to take off.