Spongeworthy genes


What are the key ingredients for making a multicellular animal, or metazoan? A couple of the fundamental elements are:

  • A mechanism to allow informative interactions between cells. You don’t want all the cells to be the same, you want them to communicate with one another and set up different fates. This is a process called cell signaling and the underlying process of turning a signal into a different pattern of gene or metabolic activity is called signal transduction.

  • Patterns of differing cell adhesion. But of course! The cells of your multicellular animal better stick together, or the whole creature will fall apart. This can also be an important component of morphogenesis: switching on a particular adhesion molecule (by way of cell signaling, naturally) can cause one subset of cells to stick to one another more strongly than to their neighbors, and mechanical forces will then sort them out into different tissues.

These are extremely basic functions, sort of a minimal set of cellular activities that we need to have in place in order to even begin to consider evolving a metazoan. Fortunately for our evolutionary history, these are also useful functions for a single celled organism, and while the metazoa may have elaborated upon them to a high degree, there’s nothing novel about the general processes in our make-up. The principles of signaling and transduction were first worked out in bacteria, and anyone who has a passing acquaintance with immunology will know about the adhesive properties of bacteria, and their propensity for modulating that adhesion to build complexes called biofilms.

So let’s take a look at the distribution of signaling and adhesion molecules in single-celled organisms, multicellular animals, and most interestingly, a group that is close to the division between the two (although more on the side of multicellularity), the sponges.

Phylogenetic position of O. carmela (asterisk) relative to model bilaterians. Dd, D. discoideum; Sc, Saccharomyces cerevisiae; Sp, Schizosaccharomyces pombe; Oc, O. carmela; Dm, Drosophila melanogaster; Ce, Caenorhabditis elegans; Gg, Gallus gallus; Mm, Mus musculus;
Hs, Homo sapiens.

This cladogram illustrates the relationships of the organisms studied, and also, unfortunately, requires you to note a few non-standard acronyms.

First, a few non-metazoans on the left. “Dd” is the slime mold, Dictyostelium, which has amoeboid cells, and “Sc” and “Sp” are both yeasts, single-celled eukaryotes.

On the right are the true multicellular animals, or eumetazoa: “Dm” is the fly, Drosophila, “Ce” is the worm C. elegans, “Gg” is the chicken, Gallus, “Mm” is the mouse, Mus, and finally, of course, “Hs” is us, Homo sapiens. All of these organisms were selected because of the availability of comprehensive genomic data, and because they span a range of single-celled to multicellular organisms.

In the middle of the cladogram with an asterisk above its name is Oscarella carmela, the subject of this study. It’s a sponge. Sponges are interesting because they’re just barely multicellular: they only have a few cell types, no organs, only one kind of tissue, and their organization is extremely simple (still, it seems unfair to kick them out of the Eumetazoa, and as we’ll see, their superficial simplicity hides an underlying complexity).

There’s the question: what is the status of these essential components of the recipe for the metazoa in these organisms, the sponges, that are just above the threshold for multicellularity? To answer that question, I’m going to show you some humungous tables of data that might be a little bit intimidating, but don’t panic— the message is going to be clear.

Here’s how it works: the authors started with one species, the sponge O. carmela, that is not well characterized. They also had a list of known important molecules and their sequences, so they went fishing in a library of expressed gene sequences from the sponge to find which of these genes are present…and, basically, almost all of them are. The table lists on the left hand side the names of these genes, and then in addition, the authors searched genomic databases for 9 other well characterized species to see if they also had these genes. If they did, the entry is marked with a bullet (•); if not, an open symbol (◊). I think you’ll see a pattern emerge fairly easily.

This first table lists the O. carmela signaling pathway genes. They fall into 6 classes: the Wnt pathway, TGFβ, Hedgehog, RTK, Jak/Stat, and Notch/Delta (there is a seventh common metazoan signaling family, the nuclear hormone receptor, which hasn’t been found in the sponge.) They’ve also tossed two “Sp”s into the table: the bilaterian one is Strongylocentrotus purpuratus, the sea urchin.

Animal signaling-pathway components in the sponge O. carmela

Pathway components Functional roles Phylogenetic distribution of signaling gene families*




Hs Mm Gg Sp Dm Ce Sp Sc Dd

    Wnt Ligand
    Frizzled Receptor
    Dkk Antagonist
    Dixdc Positive regulator
    β-catenin Downstream effector
    Dvl Downstream effector
    Nemo Downstream effector
    Follistatin Antagonist
    Mad Downstream effector
    Hh Ligand
    Ptc Receptor
    Disp Ligand export
    Sufu Downstream effector
Receptor tyrosine kinase
    Egfr Receptor
    Igfr Receptor
    Fgfr Receptor
    Epha Receptor
    Ret Receptor
    Musk Receptor
    Ddr Receptor
    Jak Receptor cofactor
    Stat Transcription factor
    Stam Downstream effector
    Pias Antagonist
    Delta Ligand
    Notch Receptor

Dixdc, Dix domain containing; Dkk, Dickkopf; Dvl, Dishevelled; Nemo, Nemo-like kinase; Mad, mothers against decapentaplegic homolog; Hh, hedgehog homolog; Ptc, Patched; Sufu, suppressor of fused; Disp, Dispatched; Egfr, epidermal growth factor receptor; Igfr, insulin-like growth factor receptor; Fgfr, fibroblast growth factor receptor; Epha, Eph receptor A; Ret, Ret protooncogene; Musk, skeletal muscle tyrosine kinase receptor; Ddr, epithelial discoidin domain receptor; Jak, Janus kinase; Stam, signal transducing adaptor molecule; Stat, signal transducer and activator of transcription; Pias, protein inhibitor of activated STAT; Amb, Amoebozoa; Hs, Homo sapiens; Cf, Canis familiaris; Mm, Mus musculus; Gg, Gallus gallus; Sp, Strongylocentrotus purpuratus; Dm, Drosophila melanogaster; Ce, Caenorhabditis elegans; Sp, Schizosaccharomyces pombe; Sc, Saccharomyces cerevisiae; Dd, Dictyostelium discoideum.

*Filled circles (•) indicate the presence of gene homologs, not necessarily orthologs, in select taxa. Open lozenges (◊) indicate their absence.

See the pattern? The three rightmost columns, the fungi and the slime mold, generally lack homologs to all of these genes, with a few notable exceptions. The six leftmost columns and O. carmela, all possess these genes, with a few interesting exceptions. In this property of having a particular suite of signaling genes, the metazoa cluster together very nicely, including the sponges.

Let’s do it one more time, only this time looking at cell adhesion molecules rather than signaling genes.

Eumetazoan cell-adhesion machinery in the sponge O. carmela

Adhesion genes homologs Phylogenetic distribution of cell adhesion gene families*




Hs Mm Gg Sp Dm Ce Sp Sc Dd

Cell-contact and -adhesion proteins
ECM molecules
    40S ribosomal protein SA
Adhesion-related cytoskeletal linkers
    Rho1 GTPase

LASP-1, LIM and SH3 domain protein 1; ADAM, a disintegrin and metalloprotease; Fat, Fat protocadherin tumor suppressor; NCAM, neural cell adhesion molecule; ECM, extracellular matrix; Cthrc, collagen triple helix repeat containing; Col11a2, procollagen, type XI, α 2; Col4, Type IV collagen; Fras1, Fraser Syndrome 1; Amb, Amoebozoa; Hs, H. sapiens; Cf, C. familiaris; Mm, M. musculus; Gg, G. gallus; Sp, S. purpuratus; Dm, D. melanogaster; Ce, C. elegans; Sp, Sc. pombe; Sc, Sa. cerevisiae; Dd, D. discoideum.

*Filled circles (•) indicate the presence of gene homologs, not necessarily orthologs, in select taxa. Open lozenges (◊) indicate their absence.

Look at that: the same pattern! Fungi and Amoebozoa lack the repertoire of sticky cell surface genes that all the metazoa, including the sponges, possess. This is not to say that single-celled creatures lack their own special adhesive proteins or signaling molecules, but merely that the ones they use are often different than the ones we use, and that you can recognize our clade by the molecules we have elevated to the canon…and that sponges definitely belong in our clade.

An important lesson here is that molecular diversity preceded morphological complexity—the lineage leading to the metazoa accumulated diverse signaling and adhesion genes because they were useful in their niche, not ours, but that those genes also were useful in achieving multicellularity. The transition was seamless and required very little in the way of novel function.

Single-celled organisms mastered all the dance steps first; what multicellularity accomplished was to add choreography, coordinating all of the players into one beautiful, coherent, and elaborate dance. Even sponges—their dance may be fairly simple and repetitive, but it still requires all the basic properties of cell:cell interaction to be carried out successfully.

Nichols SA, Dirks W, Pearse JS, King N (2006) Early evolution of animal cell signaling and adhesion genes. Proc Nat Acad Sci USA 103(33):12451-12456.


  1. says

    Thanks so much for this post. I’m in the last stages of dissertation writing and hearing this basic stuff is like a wake-up call. I am in a very distantly related field, and yet you have no idea how hard I just smacked myself on the head when I read this. Cell-signaling! Adhesion! Of course!

  2. Vasha says

    Related to the uncertain position of sponges at the edge of multicellularity, see the two Deep Sea News posts on carnivorous sponges. The second one has gruesome video. Basically, tiny organisms get stuck on hooks on the outside of the sponge, then the sponge’s cells all migrate individually, surround the prey, and each digest a bit of it in an amoeba-like way.

  3. Scott Hatfield says

    Brilliant! And (since it’s PNAS) it will be available as PDF file for gratis at some point. I can’t wait to add this to my collection!….SH

  4. says

    What sort of bias is introduced by including yeasts as fungal representatives? I wonder what the genomes of multicellular fungi would reveal. Multicellularity has evolved multiple times in eukaryotes — what are the similarities and differences in the different ways of dancing the cell signaling/adhesion waltz?

  5. CanuckRob says

    PZ, thank you for another excellent post. I am not a scientist, just a fairly well read layman, and your science posts are some of the most interesting and comprehensible that I have found. No wonder Richard Dawkins said “Pharyngula blog can reliably be consulted for trenchant good sense”.

  6. JT says

    That’s so cool!! Looking through the tables, its interesting that Dicty has a frizzled receptor but no Wnt.

  7. says

    Hi Martin,

    “It will be cool to see this extended to choanoflagellates.”

    Nicole King, who was also involved in this study, has studied choanoflagellates and found the same thing; “choanoflagellates … express multiple members of gene families previously thought to be unique to animals.” (King N., 2004, “The Unicellular Ancestry of Animal Development”, Developmental Cell 7(3):313-325)

  8. Opisthokont says

    You claim that it is unfair to kick the sponges out of the Eumetazoa, but I would ask: what term you would use instead to indicate the group of animals with at least two layers of tissue differentiation and some sort of symmetry to their body-plans? Traditionally, the term “Metazoa” was used for exactly that group (sponges were in the subkingdom Parazoa), but for some reason the term “Metazoa” is used nowadays more for what we used to call the Kingdom Animalia. In order to alleviate confusion, what used to be called the “Metazoa” are now the “Eumetazoa”, “Eu” being the Greek prefix for “no, we mean really this time” (seen also in such groups as the Euarchonta and eudicots). Personally, I would rather keep the older terms, but the newer ones have the advantage of not having been redefined several times. So, PZ, please keep things from getting any more complicated than they need to be: sponges are not in the Eumetazoa, and never will be (barring some convincing and unexpected findings that they and Trichoplax are not the least-derived members of the animal kingdom)!

  9. says

    Keith Douglas: Do you have a link to more information about the killer mushrooms? A quick Google didn’t turn up anything, but my search engine fu is lamentably weak.

  10. says

    PZ – thanks for this. I was only able to follow it after printing it out and marking up the tables manually. But this is a great help in understanding how this all fits together.

    My sense of wonder at the workings of biology have been reawakened by your blog; if I could only turn back the clock I would have taken more biology in college.