Feet and Glue

October 7, 2011

Tags: silk, evolution, Theraphosidae, Rind, Gorb, Aphonopelma, Perez-Miles, Grammostola, Brachypelma, Peattie, Sahni, Latrodectus

[This post contains later amendments that may be instructive to other science writers. See the later posts on purported tarantula foot silk to see why.]

Three papers published this summer might at first seem unrelated. But read together, they pull the entire arc of spider silk evolution into sharper focus. Two papers indirectly address the evolutionary origins of spider silk production. The other demonstrates that the evolution of silk proteins has been central to spider evolution even after the extraordinary proliferation of silks that made the vertical orb web possible.

First, F. Claire Rind and colleagues reported that at least some tarantulas (which belong to the family Theraphosidae) do indeed secrete silk from their feet. This report appears to settle a controversy that first broke out in 2006, when a team of researchers led by Stanislav Gorb announced that they had persuaded a Costa Rica zebra tarantula, Aphonopelma seemanni, to walk on a nearly vertical surface covered with glass microscope slides. The researchers claimed that as the tarantula started to slip, it left behind “footprints” made up of miniscule silk fibers. If this observation held true, it could have important implications concerning the origin of spider silk production. Like all spiders, tarantulas secrete silk through abdominal spinnerets, small appendages ending in multiple spigots that are the outlets for the abdominal silk glands. Genetic studies have shown that spider spinnerets are the evolutionary descendants of the gill branches of ancient arthropod limbs. If spiders secreted silk from their limbs as well as through their spinnerets, this fact might not only cement the limb-spinneret connection but also suggest new hypotheses for the earliest origins and survival value of spider silk.

The Gorb report left some questions open, however. The authors made no mention of finding silk glands in the tarantula’s feet. Also, because they had studied only one species of tarantula, it was impossible to know whether their finding applied only to A. seemanni, only to a few species of tarantula, to all tarantulas, or to all mygalomorphs, the 3000-species-strong branch of the spider evolutionary tree that encompasses the tarantulas. If the finding was true for many tarantulas or other mygalomorphs, it could—for the reasons outlined above—have implications concerning the oldest uses of spider silk. But, if it turned out that just one or only a few closely related species of tarantula secrete silk from their feet, it could mean that this ability didn’t evolve until about 50 million years into the course of spider evolution. That would still be interesting, but it would simply confirm what we already know about silk protein evolution in arthropods—that silk production has evolved and disappeared and evolved again in numerous arthropod lineages. It’s common stuff. What makes spiders special is not that they make silk but the fact that so many uses and types of their abdominal spinneret silk have evolved.

And what if A. seemanni actually didn’t secrete silk from its feet? After all, no one else had ever reported such an observation. A different team, led by Fernando Perez-Miles, set out to test the Gorb findings. These researchers conducted a trial that attempted to eliminate another possibility: that the Gorb team was mistaking spinneret silk for foot silk. So they used paraffin wax to seal the spinnerets of four A. seemanni tarantulas, let them climb the vertical walls of a container lined with glass slides, and then examined the slides. In 2009, they reported that they could find no silk “footprints.” They also couldn’t find any silk glands or silk conduits when they sliced and microscopically examined the tarantulas’ feet.


Until the Rind paper appeared this summer. Rind and colleagues come down strongly on the side of foot silk, and they also take the related experiments to a new level. First, they tested three different tarantula species for evidence of foot silk secretion: the Chilean rose (Grammostola rosea), the Indian ornamental (Poecilotheria regalis), and the Mexican flame knee (Brachypelma auratum), which come from widely separated branches of the Theraphosidae family tree. They filmed all three species climbing the glass-slide-coated vertical surface of a holding tank, taking note of how their eight feet came into contact with the surface and making sure that if the abdominal spinnerets came into contact with a slide that slide was excluded from study. They then challenged the two different Chilean rose tarantulas by gently shaking the tank while the spiders were on the vertical surface, causing them to slip. In each of eight trials, they found silk on the slides that [4/27/12 amendment by Leslie: it seemed (note to self: skepticism does not equal cynicism; stay alert!)] could only have been secreted by the spiders’ feet.

The Gorb team had provided micrographs it said represented silk spigots found on the feet of A. seemanni. But the Perez-Miles team disputed this characterization of these structures, saying they looked nothing like spinneret spigots but instead resembled “thermosensory setae,” or well-known bristles found on tarantula feet used for sensing changes in temperature. The Rind team has now provided a fascinating set of micrographs showing [4/27/12 amendment by Leslie: what they say are] multiple spigots interspersed among the shorter bristles of the tarantulas’ feet. (You can see some of the micrographs as well as video of the experiments at Dave Mosher’s blog at Wired Science.) The pores through which silk is secreted are obvious on some of these spigots, and the micrographs even capture one spigot with a blob of silk forming at its tip. [4/27/12 comment by Leslie: I should have written: "These apparent spigots have obvious pores through which silk could emerge, and the micrographs even capture one spigot with a blob of liquid forming at its tip."]

Coincidentally, another research team, headed by Anne Peattie, also found tarantula foot silk at about the same time, although Peattie’s team was looking for something else. Peattie, like most researchers, accepted the consensus that spiders adhered to surfaces using a dry, fibrillar (or “hairy”) adhesive. Basically, the tiny “hairs” of the pads on their feet stick to surfaces due to the attraction between the molecules in the “hairs” and the molecules of the surface. She and her team wanted to compare spider feet and gecko feet adhesion. They used interference reflection microscopy to examine what happened as a G. rosea tarantula (as well as a jumping spider, a wandering spider, a solifugid, and two species of mite, which are all arachnids) hung upside down from a microscope slide. As the tarantula began to lose its grip and moved its feet to try to compensate, the researchers noticed streaks of fluid where its feet had been. They also found fluid secreted from the feet of all the other arachnids they studied, including the other spiders. In terms of understanding how different methods of adhesion have evolved, this paper would appear to represent a breakthrough—applying more advanced imaging methods, the researchers have overturned the previous assumption that most, if not all, arachnids rely solely on dry adhesion.

And, like Rind’s team, Peattie’s team spotted “silk-like secretions” left behind by their G. rosea. But even though the jumping spider and the wandering spider left behind foot fluid, they didn’t leave any foot silk, at least none that the researchers could see.

Putting these four papers together leads to some interesting questions. First, the Rind team thinks the Perez-Miles team didn’t find foot silk because they didn’t rattle their tarantula’s tank enough—they surmise that the tarantula didn’t feel itself in danger of falling. But no one except the Perez-Miles team tested an A. seemanni. It’s unlikely but still possible that A. seemanni doesn’t secrete silk from its feet. Still, the fact [4/27/12 amendment by Leslie: "the finding," not "the fact"] that three and probably four different tarantula species from widely separated branches of the Theraphosidae family tree have this ability means [4/27/12 amendment by Leslie: "would mean" not "means"] that it’s highly likely researchers will find more similarly endowed species.

But is foot-silk secretion just a tarantula thing? The jumping spider and wandering spider, which seem not to produce foot silk, are araneomorphs, spiders that possess more complicated silk systems than the tarantulas. Is foot-silk secretion an ability that didn’t evolve until the tarantulas evolved, or is it instead the ancestral condition for all spiders? Is it an ability that’s been lost by all the araneomorphs? The earliest spiders most likely lived almost entirely underground, like today’s mesotheles. The mesotheles are the oldest still-living branch of the spider evolutionary tree and seem little changed from the oldest mesothele fossil yet found, which is about 290 million years old. Unlike tarantulas, some of whom live in trees and many of whom roam in and over rocky terrains, the earliest spiders wouldn’t have had much need of a redundant means of holding on to vertical surfaces. Is it possible that foot silk first evolved not as a non-slip gripper but rather as a way of encasing eggs or of hiding oneself? Check out this video of web-spinners, insects belonging to the order Embiidina. Might spiders have used silk something like this before spinnerets evolved? On the other hand, given that no silk storage glands have been found in the tarantulas’ feet (whereas embiids do have silk storage glands), maybe ancestral spiders couldn’t produce enough silk from their feet for egg wrapping or burrow lining. It will be much harder to tell whether spiders use foot silk when lining burrows or encasing eggs—in fact, I can’t imagine a way of monitoring such activity that wouldn’t cause a spider to stop what it was doing, making monitoring impossible. But we may get more clues to this mystery if someone (1) runs similar experiments on various mesothele species and (2) examines the molecular sequence of the tarantula foot silk to see how similar it is to tarantula silk secreted through spinnerets and to mesothele silk.

Or maybe foot silk is more closely related to the foot fluid Peattie’s team found than to any current silk uses. Peattie’s team, whose long-term focus is adhesion, doesn’t think that the foot silk contributes much to attachment force. And their tarantula relied on dry adhesion until it began to slip—it’s not like the spider relies on wet adhesion most of the time. Also—let’s face it—clinging upside down to squeaky clean glass is not exactly analogous to any situation a tarantula was likely to find itself in hundreds of millions of years ago. So it seems possible that the foot fluid and the foot silk (perhaps in earlier evolved forms) first served a purpose different from adhesion and then were coopted. Perhaps they first evolved as a way of leaving a chemical trail on the ground for potential mates to follow. Perhaps the liquid and silk worked together at some other task. Did the foot silk evolve from the foot liquid or independently? There are a lot of unanswered questions, and it’s unlikely we’ll have clues to the answers until someone undertakes chemical and perhaps amino acid sequence analyses of the fluid and the silk, and also compares them to various silks secreted through the spinnerets.

Now jump 120 (assuming foot-silk evolution in the tarantulas) or more than 180 (assuming foot silk is an original spider trait) million years ahead to consider the research concerning aggregate silk protein gumfoot glue on cobwebs published by Vasav Sahni and team. Here’s a gumfoot web we found under a rocking chair in our house.

From an evolutionary standpoint, what’s interesting about such cobwebs is that they evolved after the orb web, probably in response to the evolution of spider-hunting wasps. (We talk about how cobwebs work as prey catchers and protection from predators as well as the evidence for their late evolutionary descent from orb webs in Chapter 11, “Beyond ‘Perfect’” in Spider Silk). This makes cobwebs some of the most recently evolved spider webs, although in human terms they’re still awfully old—about 120 million years old.

Orb weavers and cobweb weavers possess the same set of silk glands. Both coat the prey-capture threads in their webs with glue produced in their aggregate glands. But the two types of spiders use different silks, originating in different glands, to lay down these threads. In orb webs, the capture thread consists of flagelliform silk, which is superstretchy. In cobwebs, the capture threads consist of major ampullate silk, which evolved before flagelliform silk and is superstrong and stiffer than flagelliform silk. The Sahni team found that, although the two types of spiders deposit capture glue from the same aggregate glands, the glues are in fact chemically different and possess different mechanical properties. Each glue benefits its maker in a different way.

Both glues form droplets on their silk thread. But the droplets have different compositions. Orb-weaver aggregate glue droplets have a dense protein core surrounded by a translucent mixture of glycoproteins (protein molecules that contain carbohydrates) and a solution of water and salts. Gumfoot aggregate glue droplets lack this core. Although some constituents of the gumfoot drops have been identified, their exact content is not yet known. What is known now, thanks to this research, is that the different compositions of the two glues cause them to react to humidity levels differently. Spider webs hang in air, and depending on where a spider spins its web, the surrounding air can be quite humid, quite dry, or anywhere in between. Orb-weaver aggregate glue droplets are maximally sticky at intermediate humidity (40-60% relative humidity). If the air is very humid or very dry, the glue is less sticky, and prey would have a better chance of struggling free of the orb web. In contrast, humidity doesn’t affect gumfoot glue much. Orb weavers live in humid tropical or temperate areas, whereas widow spiders—whose gumfoot glue Sahni’s team examined—live in both humid and quite dry areas. The fact that gumfoot glue isn’t affected much by humidity may have allowed cobweb weavers to fan out into more habitats than they would have been able to had their aggregate glue not evolved in a different direction from orb-weaver aggregate glue.

Each of these papers opens up new paths of inquiry. And because they focus on the evolution of proteins, they are quite likely, long term, to have implications for the evolution of animals other than spiders. I’m really looking forward to seeing where this all leads.


"...a compelling introduction to evolution in action through the lens of spiders and their silks."

Quick Links

Find Authors