Thread by @wc_ratcliff

Well, this has been a long-awaited day- the first paper on our multicellularity Long Term Evolution Experiment (LTEE) is on the BioRxiv. Ever wonder how simple multicellular organisms evolve to become larger and more complex over thousands of gens? 1/35

biorxiv.org/content/10.110…

TLDR: Over 3,000 generations, snowflake yeast evolved macroscopic size, increasing from ~100 to ~450,000 cells / group. This required sustained biophysical adaptation: individual snowflake yeast went from being weaker than gelatin to as strong and tough as wood. 2/35

This work represents ~4 years of sustained effort from @ozan_g_b, who pioneered and shepherded this experiment, and @dahaj1897 & @yunkerlab, who led the biophysics. Other key contributors were @pennyckahn, grad students @thomas_day_, @Iishiiyaa, and Kai Tong, and @evadyer. 3/35

Before jumping into the science, I want to contextualize this work a bit. Inspired by the foundational work of @RELenski, we have always wanted to set up a multicellularity LTEE. It just took us half a decade to figure out how to make it work (see tweet 30 for the solution!) 4/35

We hope that this is just the first chapter in a long story of multicellular discovery using experimental evolution. 5/35

The early evolution of multicellularity has historically been difficult to study. While we know that simple groups of cells readily evolve from single-celled ancestors, we don’t really know how cells become integrated into increasingly complex multicellular organisms. 6/35

We started out our experiment with snowflake yeast- dumb clumps of cells that initially don’t have multicellular adaptations. But they are not randomly assembled- they have a tree-like growth form, resulting from daughter cells remaining attached to their mothers after repro 7/35

Snowflake yeast have a simple life cycle- as the group grows larger, strain arising from cellular packing results in cell-cell separation, breaking off a branch of cells, and a new baby snowflake cluster is born. This is necessary for groups to be evolutionary units. 8/35

We grow our yeast in shaking incubators, and once a day select for larger cluster size by sedimentation (only the bottom of the bottom of the pellet gets to survive, so it’s an arms race). 9/35

Why select on size? The benefits of multicellularity stem in part from advantages of being in a group, and most multicellular lineages have undergone selection to form larger, more mechanically-robust multicellular bodies at some point in their evolutionary history. 10/35

Large size is also thought to play a role in the evolution of increased complexity, both because growing larger requires biophysical innovation and because larger size may select indirectly for cellular differentiation. But these hypotheses have never been directly tested. 11/35

OK, enough background. Over 600 rounds of selection (~3,000 generations), snowflake yeast evolved to become macroscopic. We’re talking mm scale- bigger than a fruit fly. They go from ~100 cells per cluster (left tube) to nearly half a million cells per cluster (right tube)! 12/35

Adaptation in the five replicate populations proceeded nearly identically for the first 100 d. But then they hit a period of evolutionary stasis. Two pops broke through rather quickly, while three others took about a thousand generations to resume evolving larger size. 13/35

Somewhat surprisingly, macroscopic snowflake yeast continue to grow like their ancestors. They remain clonal, and continue to grow through mother-daughter cell adhesion. If you squish them, they break into little modules that resemble their snowflake yeast ancestor. 14/35

How did snowflake yeast evolve to become ~20,000x larger? This is a real biophysical challenge. Initially, if you break a single cell-cell connection, you break the group. This is great for endowing groups with a life cycle, but it makes for *terrible* material properties. 15/35

So how do they actually do it? In our experiment, snowflake yeast evolve far more elongate cells. This reduces the density that cells pack in the group, slowing down strain accumulation and fracture. These evolved cells look more like E. coli than yeast! 16/35

But wait- looking above, we see something interesting. At first (aspect ratio between 1.2-2.2. AR is just the ratio of cell length to width.), more elongate cells increase cluster size slowly, linearly. Then all hell breaks loose. What’s going on? 17/35

Well, we can look at the relationship between cellular aspect ratio and cell packing density. At first, experimental measurements match up really well with expectations from a 3D biophysical simulation…but then it diverges! 18/35

Paradoxically, in our experiments, evolving longer cells ends up making the clusters MORE densely packed, not less. What is going on? 19/35

We needed to look inside the clusters, to see if changes in the geometry and topology of the clusters held the answer. Unfortunately, light scatters too quickly for normal microscopy. Instead, we imaged 1000's of ultrathin slices on a SEM and we had our internal structure. 20/35

We were shocked to find that disconnected branches. In the ancestor, this would have caused a propagule to separate. Then we realized what was happening: our branches were becoming entangled. The cells in the cluster had evolved vine-like behavior, wrapping about each other 21/35

Entanglement can make things very strong. To fracture the material, one needs to sever many cellular bonds, not just 1. We percolated cellular entanglement by using a convex hull approach, and found that the vast majority of the cells in random sub-volumes were entangled 22/35

We also tested the materials properties of our ancestor and evolved macroscopic snowflake yeast, and found that the evolved yeast exhibited ‘strain stiffening’, a hallmark of entangled systems. 23/35

To put some perspective on the evolution of their materials properties: the ancestor was a terrible material: about 100x weaker than gelatin.

Macroscopic snowflake yeast were about 10,000x more biophysically tough. These guys have the strength and toughness of WOOD! 24/35

We sequenced yeast from all 5 replicate populations and the results make sense: we see lots of mutations in genes that cause cell elongation (cell cycle and filamentous growth) and budding (most appear to increase bud scar size, strengthening cell-cell connections). 25/35

One of the beautiful things about exp. evolution is that it can both test hypotheses and explore the unknown. This paper does a bit of both, and we learned some neat things:

1) We got front row seats to a major transition, examining MC adaptation in real evolutionary time. 26/35

2) We showed how selection for larger organismal size, a fundamental multicellular trait, can drive the evolution of novel multicellular adaptations making larger size possible. 27/35

3) We show that genetically regulated development is not required, a priori, for sustained multicellular adaptation. Selection readily acts on emergent multicellular traits arising from mutations that only directly affect cell-level phenotypes. 28/35

4) It underscores the importance of interdisciplinary research for studying the evolution of multicellularity- I don’t think we could understand this transition without considering its biophysics. Multicellular bodies are Darwinian materials, after all! 29/35

Earlier I mentioned that it took us ~5 years to figure out how to do this experiment. I left out an important point. @ozan_g_b actually did 3x the amount of experimental evolution I described above, continuing the experiment described here: nature.com/articles/s4146…. 30/35

@ozan_g_b evolved snowflake yeast with three different types of oxygen metabolism- anaerobes, mixotrophs (can ferment and respire), and obligate aerobes. Only the anaerobes evolved macroscopic size. 31/35

All of our previous attempts at exp. evol. used O2-loving mixotrophs, and they all stalled out, never evolving to be very large. O2, it turns out, creates a powerful constraint on the evolution of multicellularity- larger organisms are less able to efficiently consume it. 32/35

Where to now? Well, we’re continuing the experiment! It’s an exploration and we don’t yet know what we will find, but we’re especially interested in using these yeast to study the origin of multicellular development & cellular differentiation. 33/35

If you are interested in this kind of stuff, our lab is recruiting PhD students for 2022. We’re interested in both pure biologists and, in collaboration with @Yunkerlab, biophysicists- drop me / us an email and we can chat. 34/35

Finally, if you want to hear the detailed version of this story, invite @ozan_g_b for a seminar! He gives a killer talk and has deep thoughts on multicellularity, earth science, ecology and evolution. And huge thanks to the @PackardFdn and @NIH for funding this work! 35/35