Learning from a mustard

arabidopsis 3

Ask a molecular biologist about their favorite plant and they will likely answer with Arabidopsis thaliana. This mustard-family species is an excellent resource for evolutionary biology because of the wide range of genetic and genomic resources available for it. As a result it has become a model organism for plants in the same way that zebrafish are for human diseases.

For an ecologist who prefers to be in the field rather than in the lab, Arabidopsis is not a very inspiring organism. It is a small weedy-looking thing that completes its entire life cycle in about a month. Compared to the grandeur of a redwood or the beauty of an alpine primrose or the mysterious clonal lifestyle of an aspen, it has little to offer. Nevertheless, a few years back, I found myself embarking on a study with this species in collaboration with several Arabidopsis experts. Why?

The answer is that it is a marvelous system for experimenting on plant phenotypes and for then interpreting these experiments in an evolutionary context. Without too much trouble one can generate a range of individuals with known mutations causing known phenotypes, or with natural phenotypes representative of growth in very different environments worldwide. Arabidopsis is a tool.

I was interested in a question about leaf carbon economics. Leaves generally fall along a spectrum from ‘fast’ to ‘slow’ (Wright et al. Nature 2004), with some species having fast-photosynthesizing but cheap and low-lifespan leaves and others with slow-photosynthesizing but expensive an high-lifespan leaves. Why should carbon economics in plant leaves be restricted along this one-dimensional strategy axis? A number of theories have been proposed to explain this pattern, including some that I have helped to develop. But these theories had never been compared to each other using the same dataset because each required one to measure very different predictor variables. Knowing which (if any) theory was consistent with reality would help us understand the basis of this fundamental economic fact about plants.

arabidopsis 1

And so – Arabidopsis. It grows quickly, which makes it suitable for lab work. It can be forced into a range of phenotypes based on multiple approaches (recombinant inbred lines, near isogenic lines, natural ecotypes, knockout mutants). These phenotypes can then be traced back to known genetic changes. We decided to measure a range of economic and predictor variables in a wide set of Arabidopsis genotypes.

arabidopsis 2

I had little expertise or interest in growing the plants, but was lucky to work with an excellent French group of scientists who were already growing and phenotyping the species. Their previous work already showed that these phenotypes are associated with leaf economic trait variation (Vasseur et al. Ecology Letters 2012). They had developed an automated platform to go from seed to spreadsheet with a minimum of human intervention. I visited this facility last year and was highly impressed by its capability. Nearly everything is computer-controlled.

arabidopsis 4

One of the things we ourselves measured was the minor vein density of different genotypes – this variables characterizes the leaf’s vascular system (which transports water and sugar). A few theories (e.g. Blonder et al. Ecology Letters 2011) we developed suggest that this variable is a strong predictor of leaf economic traits. We also measured several other predictors including leaf dry matter content, a variable important to another theory (Shipley et al. Ecology 2006). And then we put it all in a statistical model to see which theories would survive the data.


The answer, surprisingly, was that very few theories were consistent with data. We found that minor vein density might play an important role (via its influence on an unmeasured but real variable), and that this vein density did seem to vary across genotypes in a way consistent with it playing a causal role. But ultimately, we were left with something of a mystery. We need more and better theory – reality is still too complex. I personally think that leaf venation networks do play an important role in leaf economics (but see a recent study in tropical forests – Li et al. Ecology Letters 2015) but in a way that is more complex than we first proposed.

You can read our study in Annals of Botany Plants (link, pdf). It took more than a few years to go from idea to publication, but I’m proud to see this finally out.

One of the wonderful things about science is changing one’s mind because of new data. Even if those data comes from an unremarkable small mustard plant.

arabidopsis 5

One Comment Add yours

  1. Interesting blog Ben. Thanks for the detail and links provided.

Leave a Reply