The Poison in the Leaf: Macro Evolutionary Patterns in Plant-Herbivore Co-Evolution

Plants and their herbivores have an interesting and complex relationship. It has been true for quite some time (many tens of millions of years) that terrestrial plants do not move around while animal herbivores do (though I've got friends from Texas who claim that there is a Texan tree that will move from one side of your yard to the other if it is pleased to do so). Generally speaking, a plant can not avoid being consumed by the herbivores by running away. So, it must have a defensive strategy or two that work in situ, and most likely these strategies evolved in relation to the also-evolving strategies of the mobile herbivores. The complex interconnected dynamics of herbivore-plant co-evolution provides much of the fabric for the ecology of any given terrestrial biome.

There are probably three broad categories of plant strategies to fight herbivory (I'm oversimplifying here). One is to escape the herbivory. A plant cannot run away, but it is possible to affect when and where the edible bits will be most available in a way that thwarts herbivory. This might be done by growing fast, producing long-distance dispersing seeds, and being weed-like in strategy, or it might involve confining certain growth and all reproduction to randomly selected years or prime-number years. The former is escaping across space; the latter is escaping across time. Both presume that there will be herbivores, but the plant avoids a situation in which the herbivores will have a chance to settle in on this particular plant as a medium or long term resource.

Another strategy is to recover quickly. So, you get eaten by a herbviore? What's to worry? Just grow back! This may involve positioning the growth bits in a way that herbivory does not matter as much as it otherwise might. Grass grows from the base of the plant while spruce trees grow from the tip of the plant. So if a large herbivore snacks on the top of the baby spruce tree in early summer, that's it for growth for the rest of the year. If a large herbivore snacks on the top of the grass plant, it just keeps growing. Within phylogenetically related plants, though, the relevant variation here is probably in how fast regrowth occurs. Once the plant gets munched by herbivores, who then lose interest and move on, the plant tries to grow back fast enough to replace energy-producing and storing tissues and reproductive parts. One way to do this might be to put extra energy into growing roots. Extra root tissue provides the stored energy (and possibly water) to regrow damaged above ground tissues more quickly.

The third major strategy is chemical or physical defense. With a bit of a cost (in energy and materials) a plant can make hard parts, spiny sharp parts, or poison, which may deter herbivores. These deterrents are never absolute ... there is always something out there that can get past the defense mechanism. The trick here may be similar to the old African adage: How fast do you need to be to out run a lion? Answer: A little bit faster than the guy you are with. Plants that put up a particular defense may simply be less preferred ... at that particular time and place ... than their neighboring plants, at least to the generalist herbivore, who, tasting the poison in the leaf or stem, moves along and eats the other guy.

A paper just out in the Proceedings of the National Academy of Sciences (PNAS) explores the whole plant defense thing. In particular, the paper looks at several species of milkweed to test a set of different hypotheses about the evolution of plant defensive strategies. This particular approach to analysis of this problem is fairly novel. Experimental manipulation of individuals across 38 species of milkweed (for which there is a DNA based phylogeny, or evolutionary family tree) resulted in various outcomes indicating the degree to which each species uses growth versus any one of three chemical factors as a defense.

You would have to be into some pretty esoteric plant biology to care about (or be able to fully grasp without some painful extra work) the nuances of each of the hypotheses being tested, so I'm going to summarize them in a somewhat simplified form without messing up the main point of this research too much.

It has been proposed that there is a tradeoff between being able to grow quickly and being able to produce the complex molecules that are used as poison in plant defense. This study showed that among milkweeds alive today, this is not apparently happening. This is not terribly surprising, because one would expect this tradeoff to be larger scale .... whereby a plant is either a slow growing form that invests a lot in chemical defense, such as an oak tree, vs. fast growing and highly edible, like a birch tree. The difference among milkweeds may be just too small in terms of fundamental strategic differences.

Regarding chemical defense, it was hypothesized that the three chemicals milkweed use for this purpose would be trading off ... if a plant makes a lot of one chemical, it would make less of another. However, according to the study, "[c]ontrary to [this] prediction ... we found no tradeoffs between the three most potent resistance traits."

In general, the widespread thinking on tradeoffs between defense strategies was not supported by this study, though it is almost certainly true that tradeoffs do happen. The give and take between multiple alternative strategies likely happens at different scales than exposed by this study for milkweed.

There is, however, a very interesting result from this work. The researchers applied Pagel's methods to the data to explore growth vs. chemical defense strategies in a phylogenetic context. I don't need to explain Pagel's method to get across the key finding of this study, but I'm going to anyway because it is important.

Pagels' method (which grows from Harvey and Pagel's work on phylogenetic methods) solves Galton's problem. Galton's problem comes from Anthropology, and ultimately applies to evolutionary studies and comparative studies in general. Here's the problem.

Suppose I told you that in a sampling of world cultures, it turns out that there are 1000 matrilineal societies and 500 patrilineal societies. That might make you think that matrilineal societies have some kind of increased likelihood of emerging or surviving over time compared to patrilineal societies. Nice hypothesis. But further investigation shows that 899 of the matrilineal societies are all cultures of the same language subfamily that recently diversified. We might believe that all 899 of these societies emerged recently from one society. Meanwhile, the patrilineal societies in this sample seem to have popped up all over the place, including among the societies in our big matrilineal language subfamily.

In other words, if we go back in time just a little bit, a very large number of the matrilineal societies, it turns out, 'emerged' as a very small number of events, while the patrilineal societies emerged many many times independently. Patrilineality would then be considered more likely to arise or survive over time.

The same sort of thing applies in evolutionary terms. Say we collect 1000 small mammal species, and observe stuff about them, and discover that 500 of them can fly. Does this mean that flight evolved 500 times, and is thus very likely to evolve (like, there is a 50-50 chance of flight evolving)? Well, if 499 of those mammals are bats then no, it means that flight evolved twice, once in an ancestral bat and once in some other mammal species. Flight, it would turn out, hardly ever emerges among mammals. Without considering the phylogenetic context, one would come to a very misleading conclusion.

Pagel developed methods of analyzing phylogenetic data to explore evolutionary questions. These methods essentially allow the comparison of observed patterns to a null model of random-walk evolution across a series of phylogenetic events (represented by the best available phylogenetic tree).

In the present study, a phylogenetic look at the data reveals a very interesting pattern.

It turns out that early in the evolution of milkweed, chemical defense was important, evolved a bunch of times, and diversified, but later on, during the diversification of the milkweeds and subsequently, it became less and less important. Meanwhile, growth-related adaptations became more important.

The authors suggest that this is because the milkweeds are fed on primarily by specialist herbivores. Specialist herbivores can evolve mechanisms for coping with the chemical defenses used by plants. In fact, they may even end up with mechanisms to use these chemical defenses as resources, in some cases. As I suggested above, we would expect that plants fed on primarily by generalist herbivores to benefit nicely from chemical defenses.

So, this study identifies an interesting macro-co-evolutionary trend relating closely related groups of plants and their herbivores to an overall defense syndrome. Early diversification of the milkweeds may have occurred because of the rise of chemical defenses, which in turn diversified. Generalist herbivores would have avoided the plant, leaving room open for specialists to emerge and/or evolve in relation to the milkweeds, as these specialists (or would be specialists) would have had less competition from the generalists put off by the chemical defenses. Then, the evolutionary trajectory of the milkweeds would shift as the specialists adapted to the plants' chemical defenses, rendering these defenses less potent.

Across the great diversity of plants and animals, there must be thousands of stories just like this one, but each with its own twist. We are only beginning to scratch the surface.

Agrawal, A., Fishbein, M. (2008). Phylogenetic escalation and decline of plant defense strategies. Proceedings of the National Academy of Sciences, 105(29), 10057-10060.