What does it take to become multicellular?

One of the key innovations in the history of life—at least, looking back at it from the viewpoint of a complex species—was the origin of multicellularity.  Having many cells that are differentiated, so that different cells can do different things, allows the specialization of tissues and organs that opens up many new ways of life.  Kidneys, wings, legs, leaves, brains—all of these things require differentiation of cells.  A single cell on its own has to do it all:  reproduction, movement, excretion, and nutrition. Multicellularity and cell differentiation means some cells can change their function, becoming mouths or anuses or livers without disrupting, say, the other cells that are responsible for reproduction.  It’s a way of overcoming pleiotropy by farming out tasks to specialized tissues. In their important book The Major Transitions in Evolution, John Maynard Smith and Eörs Szathmáry considered multicellularity one of the nine most critical evolutionary innovations.

So how did multicellularity come about? How much genetic change was needed to get those single cells on their own to form colonies, with some of them specializing in reproduction and the others in locomotion, nutrition, and the like?  Did it take a wholesale restructuring of the genome?

That was the question that Simon Prochnik and his colleagues asked—and partially answered—in a new paper in Science.  They had a clever approach:  look at two species that were fairly similar, but one of which was multicellular, showing some differentiation among cells, and the other was not.  Here are the species they used.  The “simple” one was the unicellular alga Chlamydomonas reinhardtii, which has flagella, threadlike organelles that can be whipped about to move the cell though the water.  This species has been used extensively in studies of movement and organelle differentiation.  Its genome was sequenced in 2007.  Here’s what it looks like (note the flagella):

Chlamydomonas reinhardtii

The “multicellular” species compared to C. reinhardtii was a multicellular green alga, the famous Volvox carterii. Here it is:

Volvox carteri

What you see is a spherical multicellular organism, but one that’s not very highly differentiated.  Nevertheless, it has crucial features that differentiated uni- from multicellular species.  There are two types of cells.  First, there are about 2000 “somatic” cells with flagella that lie on the surface of a ball of jelly (or, as the authors call it, “glycoprotein-rich extracellular matric”).  These somatic cells allow coordinated swimming so that the whole ball can swim in one direction (But there are also 16 large reproductive cells (“gonidia”) that can form sperm and eggs, which can cross-fertilize another Volvox and produce a zygote.  These zygotes undergo meiosis (becoming haploid again, like their parents), and grow inside the “female”, eventually bursting free as the mother goes to her just reward. They can also reproduce asexually, produces little Volvoxes that are genetically identical to the parent (you can see some forming in the photo above).

The Joint Genome Institute website outlines why this relatively simple species is a good model for studying multicellularity:

The 48-hour life cycle allows easy laboratory culture and includes an embryogenesis program that features many of the hallmarks of animal and plant development. These features include embryonic axis formation, asymmetric cell division, a gastrulation-like inversion, and differentiation of germ and somatic cells.

In their new paper Prochnik et al. asked a simple question: if we sequence the whole genome of Volvox carteri, how much difference does it show from its fairly close unicellular relative C. reinhardtii?  Does the initial evolution of multicellularity require many new types of genes, or will a few simple changes suffice?

The answer seems to be the latter.  Comparing the V. carteri genome with the already-published C. reinhardtii genome, Prochnik et al. showed this:

  • The two species have almost exactly the same number of genes: 14,520 protein-coding genes in Volvox and 14,516 protein-coding genes in Chlamydomonas.
  • There seem to be only 32 expressed genes in Volvox that don’t have homologs in Chlamydomonas.
  • The number of proteins per gene family (i.e., those genes that probably arose from one ancestor by duplication) is about the same for both species. However, two groups of genes have a lot more copies in Volvox. One family comprises those genes that produce the glcyoproteins composing the extracellular jelly in the middle of the ball. The other produces “cyclins,” which are involved in cell division  The expansion of the cyclin family in Volvox may have something to do with the differentiation of its cells.
  • The genome size of Volvox is slightly larger than that of Chlamydomonas (138 million bases as opposed to 118 million), but this is due largely to an increase in the amount of noncoding repeated DNA and in the length of “introns” (noncoding bits of DNA that interrupt genes) in Volvox.
  • Finally, the authors looked at those genes that they thought would be key players in the evolution of multicellularity: genes involved in intercellular communication (membrane traffic, secretion, and the like), formation of cell structure, and regulation of cell division. The types of genes involved in these structures were also quite similar between the two species.

The authors’ conclusion?  Not many new genes have to change to turn a single cell into a multicellular, proto-differentiated species.  In the Science news piece on this article, plant biologist Arthur Grossman comments: “The findings suggest that it doesn’t take very large changes in gene content to transition from a single-cell to a multicellular lifestyle.”

Is this surprising? Well, not really.  I’m not sure most biologists would have suggested that multicellularity requires a wholesale restructuring of the types of proteins present in one-celled species.  If that were true, it would be very difficult to go from one-cell to multi-cells in an adaptive, step-by-step fashion..  I would have thought that changes in protein sequence (not type) were important and, perhaps, changes in how genes are used—how they are turned on and off, and when. (This is a suggestion for which biologist Sean Carroll is famous.  I once questioned it, but now am coming around to his point of view.)

Perhaps the transition from one cell to a Volvox-type species involves changes in gene expression or timing.  Prochnik et al. didn’t look at “microRNAs” (miRNA), those bits of the genome that are involved in selectively silencing genes after they’re transcribed into messenger RNA but before they produce protein.  Nor did they investigate expression patterns of genes (something that’s obviously going to happen soon, as it’s not too hard to do), or look at the regulatory regions of protein-coding genes.  Finally, it’s still possible that although multicellularity here didn’t require new types of proteins, it did rest in a crucial way on changes in the sequences of those proteins.  That would be a hard thing to study.

Prochnik et al. recognize this: as they say in the paper, “Further studies of gene regulation and the role of noncoding RNAs will be enabled by the Volvox genome sequence, allowing a more complete understanding of the transformation from a cellularly complex Chlamydomonas-like ancestor to a morphologically and developmentally complex ‘fierce roller.'”

It seems, then, that at least this critical step in the original of multicellular species may require not wholesale changes in the types of genes in the genome, but a few critical tweaks in how those genes are expressed.

And let us not forget that the building blocks for all of this rested on things that had already evolved in one-celled organisms, which in themselves are fantastically complex, with elaborate networks of genes for metabolism, excretion, protection, movement, DNA replication, DNA translation into protein, and cell division. Single cells may look simple, but Lord, they’re not!

As one my my colleagues commented after reading this paper, “Maybe all the hard work was done by bacteria.”


Prochnik, S. E. et al. (28 authors) 2010.  Genomic analysis of organismal complexity in the multicellular green alga Volvox carteri. Science 329:223-226.

27 thoughts on “What does it take to become multicellular?

  1. Very cool. And quite expected.

    I’ll bet a box of donuts that:

    * Answers in Genesis will reject the findings because it’s “impossible to add information” to an organism.

    * The Discovery Institute will agree with the creationists and then declare these findings confirm their assertion that there is no such thing as junk DNA.

    * BioLogos will say that god tweaked how the relevant genes were expressed and regulated; therefore Jesus totally is real.

    Hypothesis generation is FUN!

    1. Ooo, another one…

      AIG will declare that because the scientists did not witness C. reinhardtii “turn into” V. carteri, that it is impossible for it to happen.

      I LOVE this game!

  2. I once questioned it, but now am coming around to his point of view.

    What! So maybe deep down you are a cissy after all?

    1. Umm. . . the only thing I took issue with was that when the gene-regulation hypothesis was first tested, there weren’t much data to support it. I called for caution. Data now seems to be supporting it, though. Stay tuned.

      1. Tchoh – that’s the problem with scientists – ALWAYS changing their minds when the facts don’t fit. Hey Kevin, you’re right! This IS a fun game!

      2. If epigenetic (i.e. gene regulatory) mechanisms are sufficient for vast differences in cell phenotypes and organ morphology programs in complex eukaryotes, why would if be a leap to suggest that epigenetic mechanism could underlie transitions in evolutionary linaeages (I’m asking sincerely, not rhetorically)?

        1. changes in DNA sequences that regulate expression of nearby genes are genetic, not epigenetic. “Gene regulatory” and “epigenetic” are not synonyms.

          Epigenetic regulation (e.g. histone methylation) is transient in evolutionary terms: its effects last a couple of generations at best, and thus cannot serve as the basis for evolutionary change.

          Mutations in gene regulatory sequences, on the other hand, are permanent.

          1. I’m not referring to changes in DNA sequence-that’s the point-simply alteration in the regulation (say, by changes in the transcription factor/cofactor milieu). And yes, all gene regulatory mechanism are epigenetic, using the traditional definition. Referring to only chromatin & DNA modifications exclusively as “epigenetic” is a recent phenomenon. Transcription factors have direct effects in transcription (in addition to targeting chromatin modifications, which require DNA-binding transcription factors for targeting to specific loci)

            That said, your point makes complete sense. The stable heritability of histone modifications is indeed unclear. Thanks.

            1. I should add that the questionable mitotic heritability of histone modifications is why there is still resistance to refering to it as epigenetic. DNA methylation obviously does satisfy this criteria, though.

  3. As one my my colleagues commented after reading this paper, “Maybe all the hard work was done by bacteria.”

    I think it is likely that this idea will only gain more support the more we discover about this subject.

    I am not a biologist, and I am not trying to downplay the significance of this study as it is clearly necessary for a clear understanding of how multicellular organisms came about, but I think it will be much more interesting to compare the genomes of a “late model” single celled organism to successively “earlier model” single celled organisms. There are almost certainly much greater differences between early and late single cell organisms than between late single cell and early multicellular organisms.

    At least from this amateur’s admittedly not particularly well informed point of view.

  4. When I started reading this I expected they’d find some new membrane proteins in carteri. Very cool nevertheless!

  5. “Maybe all the hard work was done by bacteria.”

    As a layman noticing the usefulness of traits leading to biofilms and bacterial signaling, I was lead to believe that was the case.

    Though it doesn’t explain why only one type of bacteria went ahead and formed true multicellulars with body plans and whatnot, and quite late at that. I gather the cellular feeding principle is somewhat different (pores vs vacuoles in eukaryotes)? But surely one could get around that.

    an increase in the amount of noncoding repeated DNA and in the length of “introns” (noncoding bits of DNA that interrupt genes) in Volvox.

    Right, one data point. But it would be fun to know if this is a general observation.

    Surely the constraints on multicellular cells differ from unicellulars, so any dependence on cell size (say) that in turn depends somewhat on genome size would differ. Similar for genome protein productivity/latency time, in as much noncoding/intron length hampers that.

    Onion test and all that, but it may be that one can see a difference here. Wonder what TR Gregory says.

    “somatic” cells with flagella that lie on the surface of a ball of jelly

    What do you know, I’ve seen that picture before but never understood it. For some reason I got that as plastids scattered in transparent cells permeating the balls, which doesn’t make as much sense nor fit the picture as well. Thanks for a lucid description!

    Now I wonder how they manage that though. Directed carbohydrate gel (jelly) production (and orienting the flagella outwards could help that)? That could be a spin-off from biofilms, making their own “substrate” body.

    1. “why only one type of bacteria went ahead and formed true multicellulars with body plans and whatnot”

      I’ll have to amend that; AFAIU it happened independently several times (fungi vs animals plants). Still, it is the same ‘class’ (in the math sense) doing it.

      1. No, fungi, animals and plants all decended from a single ancestor. It only seems to have happened once, unless other efforts died out.

  6. Now the rest of the 138 million bases need to be studied and compared. That might take a few weeks or so. Someone has a lot of papers to write.

  7. How do the somatic cells coordinate their swimming? Do we know what sort of signalling is involved, and how “decisions” are made on what direction to move? Or is each cell simply responding to external stimuli?

    1. As Roseanne Roseannadanna used to say, “You ask a lot of questions.”

      Sadly, I don’t know the answers–I’m not a Volvox person or a physiologist–but I suspect you can find the answers in a library or online.

    2. Well, if they are anything like the cilia organizing our left-right symmetry they have a preferred rotation axis. The hairy ball theorem of topology (i.e. combing must leave a “cow-lick”) breaks the remaining symmetry, so quite naturally a Volvox would rotate as a consequence.

      The name implies that maybe that is all it does?

  8. It doesn’t seem that surprising that a bacterium that gets into the next generation through a fly’s gametes would work hard to undermine a parasite that causes fly sterility. It’s super cool though!

  9. Why do they assume that the most recent common ancestor of Chlamadomonas and Volvox was monocellular?

  10. “This is a suggestion for which biologist Sean Carroll is famous. I once questioned it, but now am coming around to his point of view.”

    I was recently reading your 2007 Evolution paper with Hoekstra along with several of Carroll’s papers, trying to sort out all them out. Is there anyway you could detail your “coming around to his point of view”? (Mainly because I am struggling with the concepts, but also because I think coming-around’s are educational in their own right.)

  11. Sean Carroll certainly deserves credit for promoting the idea that regulatory changes are perhaps more important for evolution than are structural differences in proteins, but it is not accurate to say that he “suggested” this idea. For example, Emil Zuckerkandl made this suggestion at least as early as 1963 and Alan Wilson developed it further in the the mid-1970s (e.g., Wilson, A.C. [1976] Gene Regulation in Evolution in “Molecular Evolution” edited by F.J. Ayala). Wilson’s papers, in turn, stimulated my own research in this area over the next decade and more. For reviews see Dickinson, W.J. (1989) Gene Regulation and Evolution. In “Speciation and the Founder Principle” edited by L.V. Giddings et al. and Dickinson, W.J. (1991) Evolution of Regulatory Genes and Patterns in Drosophila. In Evolutionary Biology, vol. 25 edited my M.K Hecht et al.

Comments are closed.