The Chemosynthesis Hypothesis: Do Clams and Bacteria Work Together?

Close-up of a white clam with a long arm protruding from its shell, on a black background

Dorsal view of Xylophaga dorsalis shell and body, with g pointing to white spots. 

This is the fifth in a series of posts by Associate Curator of Invertebrate Zoology Janet R. Voight as she heads out on an expedition to Norway. There, she and colleagues will look for a wood-boring clam, Xylophaga dorsalis, to study its, well, poo. Read the first post and stay tuned for upcoming posts to find out what they discover.

Here's one hypothesis that we're exploring: for this particular species of wood-boring clams, the clam's feces in the borehole produce chemosynthetic bacteria—and that these bacteria actually help the clams. But how do we test that? It’s tougher than I hoped. We expect only low concentrations of sulfide in wood; at hydrothermal vents, sulfide concentrations are so high that they really stink. But sunken wood doesn’t stink, even with a lot of feces. In addition, low levels of sulfide react really fast with oxygen. If we expose the feces to oxygen, any sulfide might well disappear, maybe without a trace. We can’t do that and get good data.

The hole the clam bores into the wood has one small opening to the outside. The clam reaches the wood as a tiny (and very lucky) swimming larva that picks out where to settle. It changes its morphology and starts to bore in. It makes a tiny hole and the siphon sticks out of the opening. Theoretically, the best method to document the presence and concentration of sulfide would be to use a chemical microprobe, inserting it into a soft uniform substrate and moving it bit by bit to detect chemicals. But microprobes are very delicate and expensive instruments. And we can’t tell without opening the borehole (and losing the sulfide) where the clam shell is, or where the feces stop and the wood begins. If we poked the shell or the wood once (just once!) with the fragile microprobe, then the $1200 probe would be toast—meaning that its meter, positioning system, and the software that make it work would be useless.   

So we have to put together different methods that will each give us a clue. We hope to combine different clues and get a definitive answer (just like Perry Mason used to do). First, we will sample the clam tissue to look for evidence of microbial input. Microbes are different from you and me because they use a different form, or isotope, of elements, like carbon and nitrogen and maybe sulfur. If the clams benefit from microbial nutrition, we should see a microbial signature in their flesh.   

Second, we are going to use molecular methods to determine whether the microbes that we know change sulfide to sulfate are physically close to the clam. Let’s face it: we expect microbes to be in feces, so just finding them there doesn’t mean much. We will have to identify those microbes, using molecular methods (they do not have a lot of morphology). How do we know they do this chemical reaction? Previous work identifies microbial groups that have different ways of living. If sulfide-oxidizing microbes help feed the clam, they will be close to it. If they aren’t helping the clam, they will likely be closer to the wood, which is the source of the sulfide.

Third, when microbes change sulfide to sulfate, they often leave tiny calcium deposits. These clams have white spots at the base of their siphons. We are going to see if the white spots are made of calcium.  

Fourth, we are going to take the feces away from some clams in boreholes and see what happens. If they need the feces to live, they will get sick.  

Fifth, we are going to measure specimens of the clam to test whether their excurrent siphon (the one that carries poo) grows faster than the incurrent siphon. This is rare in animals. The siphons share a wall, so why would they grow at different rates? I suspect it is to deposit poo on top of the pile as the pile gets bigger. I showed that this happens in a pretty closely related species, but we need to see if it’s also true in this one. These methods are a bit indirect; they won’t give us sulfide concentrations in areas separated by the width of a human hair like a microprobe would have, but combined, we will be able to really test this hypothesis and the other, competing hypotheses I’ll write about next time.  

Read on in Part 6: Journey to Norway.

Funding for this project was provided by the Robert A. Pritzker Center for Meteoritics and Polar Studies established by a grant from the Tawani Foundation.