Are the multicellular volvocine algae monophyletic?

One of the strengths of the volvocine algae as a model system is that they span a range of sizes and degrees of complexity. Sizes range from tens of microns to a couple of millimeters, cell numbers range from one to 50,000 or so, some species do and some don’t have cellular differentiation, and some do and some don’t undergo inversion during development. This variation makes the volvocine algae ripe for comparative analyses, which I and many others have done. It also allows many of the intermediate steps between unicellular and complex multicellular life to be identified, as David Kirk did in his “twelve-step” paper.

The volvocine algae have clearly taken some of those steps more than once. Cellular differentiation, for example, has evolved at least three times, in the genus Astrephomene, in the so-called Volvox section Volvox (a.k.a. Euvolvox), and in the lineage that includes Pleodorina and the other Volvox species. One thing they seem to have only done once, though, is to evolve multicellularity itself.

There have been dozens of studies addressing the evolutionary relationships among various species of volvocine algae. Most have been from Hisayoshi Nozaki’s lab, though I and many others have weighed in as well. Nearly all of them, at least those that address the topic, agree that the three families that make up the multicellular volvocine algae–the Tetrabaenaceae, Goniaceae, and Volvocaceae–uniquely descend from a common ancestor. In other words, the multicellular volvocine algae are monophyletic.

Three important cladistic terms are used to summarize the evolutionary relationships among a group of species. If all of the members of the group descend from a common ancestor, and nothing else descends from that ancestor, the group is called monophyletic. Mammals, for example, are monophyletic. A monophyletic group is also called a clade. If all group members are descended from a common ancestor, but so are some non-group members, the group is called paraphyletic. Reptiles, for example, are paraphyletic, because there is no clade that includes all reptiles that doesn’t also include birds. The word ‘paraphyletic’ should nearly always be followed by ‘with respect to’: reptiles are paraphyletic with respect to birds.

The bottom of the barrel, in terms of evolutionary relationships, is polyphyly. A group is considered polyphyletic if its members don’t share a recent common ancestor at all, in other words, if they have multiple evolutionary origins. Flying animals are polyphyletic. Algae are polyphyletic. The genus Volvox is polyphyletic. Polyphyletic taxa are the scum of the phylogenetic Earth. Telling a taxonomist that a group she has named is polyphyletic is a deadly insult.

The prevailing view of volvocine evolutionary relationships is that the family Volvocaceae is sister to the Goniaceae (that is, each is the other’s closest relative), and the Tetrabaenaceae are sister to the Volvocaceae + Goniaceae. Two new papers infer relationships among volvocine algae and their unicellular relatives, and one of them challenges the view of multicellular monophyly.

If the multicellular volvocine algae are truly para- or polyphyletic, that’s a big deal, because it means that one of two things has to be true: either they evolved multicellularity two or more times, or some modern unicellular species descend from multicellular ancestors. There are few known examples of reversions from multicellular to unicellular life, and there are good theoretical reasons to expect that kind of change to be rare. Two independent origins of multicellularity would be less surprising, given the large number of times this has happened in various groups of green algae, but it would allow for some interesting comparisons.

The ‘canonical’ volvocine phylogeny, supported by lots of papers from several groups of researchers (myself included), has the Goniaceae (Gonium and Astrephomene) sister to the Volvocaceae (Colemanosphaera, Eudorina, Pandorina, Platydorina, Pleodorina, Volvox, Volvulina, and Yamagishiella), with the Tetrabaenaceae (Basichlamys and Tetrabaena) sister to Goniaceae+Volvocaceae:

Volvocine phylogeny

‘Canonical’ volvocine phylogeny, modified from Hanschen et al. 2017. T: Tetrabaenaceae, G: Goniaceae, V: Volvocaceae.

Small variations have been inferred, but the basic structure (Unicells, (T, (G, V))) usually holds up. A new paper by Thomas Pröschold and colleagues, however, says otherwise:

Strains of [unicellular] C[hlamydomonas] debaryana are almost identical in SSU and ITS rDNA sequences and formed the sister group of the Tetrabaenaceae (containing the genera Basichlamys and Tetrabaena)…the strains of C. debaryana were more closely related by DNA sequence to the Tetrabaenaceae than to Chlamydomonas. This was confirmed by approximately unbiased test using the program CONSEL. The tree with the artificial setting of C. debaryana as sister of the three Chlamydomonas species was significantly worse (SSU/ITS: AU 0.0348) than the best tree topology presented in Fig 3.

Here is the Figure 3 referred to in the quote:

Pröschold et al. 2018 Fig. 3

Figure 3 from Pröschold et al. 2018. Maximum likelihood molecular phylogeny of Chlamydomonas and representatives belonging to the Tetrabaenaceae, Goniaceae, and Volvocaceae based on SSU and ITS rDNA sequence comparisons. The branches in bold are highly supported in all analyses (posterior probabilities > 0.95; bootstrap values > 70%).

Pröschold and colleagues don’t make a big deal of this; paraphyly of the multicellular volvocines is not the main point of the paper. Rather, it is the description of a new species of Chlamydomonas (Chlamydomonas schloesseri) and a partial revision of Chlamydomonad phylogeny, including the establishment of a new genus, Edaphochlamys. These are both important results, since C. schloesseri appears to be a close relative of the model alga C. reinhardtii, and since the taxonomy of this group of algae is badly in need of revision. Nevertheless, I’m going to focus on the suggestion that the multicellular species are not monophyletic, since this would substantially change our view of the evolution of multicellularity in this group.

The tree in Figure 3 is based on the DNA sequence that encodes the small subunit of the nuclear ribosome (SSU) and the internal transcribed spacer region of the ribosomal operon (ITS). Pröschold and colleagues also inferred trees based on ITS alone and on the chloroplast gene encoding the large subunit of the rubisco enzyme (rbcL):

Pröschold et al. 2018 Fig. 4

Figure 4 from Pröschold et al. 2018. Maximum likelihood molecular phylogeny of Chlamydomonas and representatives belonging to the Tetrabaenaceae, Volvocaceae, and Goniaceae based on rbcL and ITS rDNA sequence comparisons. The branches in bold are highly supported in all analyses (posterior probabilities > 0.95; bootstrap values > 70%).

The three trees tell different stories, but none supports monophyly of the multicellular species. I have added a big red asterisk to each tree to indicate the most recent common ancestor of the multicellular species; in every case that ancestor also has unicellular (ChlamydomonasEdaphochlamys) descendants.

I’ll get back to evaluating this evidence, but first let’s look at another new paper, this one by Takashi Nakada and colleagues. There is one indication in their results suggesting non-monophyly of the multicellular species:

Nakada et al. 2018 Fig. 1

Figure 1 from Nakada et al. 2018. Bayesian phylogenetic tree of core-Reinhardtinia based on 18S rRNA gene sequences. Corresponding posterior probabilities (≥0.90; top left) and bootstrap proportions (≥50%) from maximum likelihood (top right) and neighbor-joining (bottom) analyses are shown next to the branches.

Pleodorina illinoisensis, which I’ve labeled in red, appears as sister to one strain of Chlamydomonas reinhardtiiPleodorina is a member of the family Volvocaceae, so if this were true, it would be very surprising. In spite of the high support values (posterior probabilities and bootstrap proportions are measures of statistical confidence, and both are essentially maxed out for this relationship), there are good reasons to be skeptical. First, this is a single-gene tree, based on sequences from the 18S rRNA gene, and single-gene trees are often unreliable (for reasons I’m not going to get into here). Second, Dr. Nakada was kind enough to send me the alignments from this paper, and although the 18S alignment has a large number of informative characters, none of them uniquely support this relationship (that is, there is no position at which P. illinoisensis and C. reinhardtii JinCheon1, and only those two taxa, have the same base). Finally, it is possible that C. reinhardtii JinCheon1 is simply misidentified. This sequence was not generated by the authors but downloaded from GenBank, and GenBank is known to have a non-trivial proportion of sequences that are not from the species listed (for example, see Bridge et al. 2003, Nilsson et al. 2006). The fact that this strain does not cluster with the other C. reinhardtii strains makes this possibility seem quite credible.

At any rate, Nakada and colleagues don’t base any of their conclusions on this putative relationship, only mentioning the result in passing as part of their description of this “barely resolved” tree. Their multi-gene tree, based on 18S plus five chloroplast genes, is consistent with monophyly of the multicellular group, though with low support at one critical node.


Nakada et al. 2018 Fig. 2

Figure 2 from Nakada et al. 2018. Bayesian phylogenetic tree of core-Reinhardtinia based on combined 18S-atpB-psaA-psaB-psbC-rbcL gene sequences. Corresponding posterior probabilities (≥0.90; left) and bootstrap proportions (≥50%) from maximum likelihood (middle) and neighbor-joining (right) analyses are shown next to the branches. Branch lengths and scale bars represent the expected number of nucleotide substitutions per site. Metaclades (MC; 1.00 posterior probabilities), unicellular lineages shown in Figs. 1 and S1 (in parentheses), and clades 1–4 are indicated.

C. reinhardtii JinCheon1 is not included in the multigene analysis, probably because the chloroplast genes were not available. This strain was also not included in Dr. Pröschold’s analyses, so an apples-to-apples comparison is not possible.

So between these two papers, we have one tree consistent with monophyly of the multicellular species and four that contradict it, though for different reasons. I’ve discussed the 18S tree from Nakada et al., but what about those of Pröschold and colleagues? In their discussion, they note that

…in contrast to the phylogenies using chloroplast genes, where Chlamydomonas and Vitreochlamys species were often at the base of the Volvocales sensu stricto, the phylogenetic analyses of SSU and ITS rDNA sequences always demonstrated that the unicellular taxa are distributed among the colonial lineages (Nakada et al., 2016; this study).

However, the SSU/ITS tree (their Figure 3 above) does not actually demonstrate this with strong support. I was unable to obtain the support values for all of the nodes, but only those subtended by branches in bold received “strong support”, i.e. Bayesian posterior probabilities > 0.95 and bootstrap values > 70%. None of the nodes that would contradict monophyly of Tetrabaenaceae+Goniaceae+Volvocaceae are strongly supported. The same is true of the ITS and rbcL trees in Figure 4, and, with the exception of the same C. reinhardtii JinCheon1 / Pleodorina illinoisensis relationship reported in Nakada et al. 2018, the same is true of the 18S tree in Nakada et al. 2016. However, the trees are not the only evidence presented by Pröschold and colleagues:

The genera Chlamydomonas and Edaphochlamys were topologically sisters of the families Goniaceae/Volvocaceae and Tetrabaenaceae, respectively. These results were partially confirmed by the activity tests of gamete lytic enzymes (GLE) derived from Chlamydomonas reinhardtii. In these tests the GLE dissolved not only the cell walls of several Chlamydomonas species such as C. reinhardtii, C. incerta, and partly C. debaryana (only SAG 26.72, but no reaction by SAG 4.72 and SAG 14.72), it also degraded those of the colonial genera Gonium, Astrephomene, Basichlamys (= Gonium sacculiferum), and Tetrabaena (= Gonium sociale; Matsuda et al., 1987, Matsuda, 1988). All these data indicated the close relationship of these taxa. In contrast, the gamete autolysin had no influence on the other colonial Volvocaceae.

This is indeed suggestive, but the relationships it suggests are not consistent with those inferred from the genetic data. Neither these trees nor any published trees I’m aware of suggest a particularly close relationship among ChlamydomonasBasichlamysTetrabaenaand the Goniaceae (Gonium and Astrephomene) to the exclusion of the Volvocaceae. Setting aside support values for a minute, the topologies of the ITS and SSU/ITS trees are consistent with the ‘canonical’ view that the Goniaceae are sister to the Volvocaceae, which means (by definition) that both are equally closely related to Chlamydomonas. So the single character of lytic enzyme compatibility does suggest a relationship at odds with multicellular monophyly, but it is a different relationship from that suggested by the genetic data. A single loss of compatibility near the base of the family Volvocaceae seems more likely.

There is also the statistical test performed by Pröschold and colleagues:

..the strains of C. debaryana were more closely related by DNA sequence to the Tetrabaenaceae than to Chlamydomonas. This was confirmed by approximately unbiased test using the program CONSEL. The tree with the artificial setting of C. debaryana as sister of the three Chlamydomonas species was significantly worse (SSU/ITS: AU 0.0348) than the best tree topology presented in Fig 3.

This test doesn’t address multicellular monophyly, though. What it shows is that C. debaryana is not sister to VLE Group 1 (C. reinhardtii + C. incerta + C. schloesseri). The genus Chlamydomonas is known to be paraphyletic with respect to all sorts of things: the multicellular volvocine algae, VitreochlamysChlorogonium, Paulshulzia, etc. This test doesn’t support a sister group relationship between C. debaryana and the Tetrabaenacea; all it does is exclude one possible relationship among four Chlamydomonas species.

All in all, I find the evidence against the monophyly of the three multicellular families suggestive, but not convincing. Strong support for multiple origins of multicellularity, or one or more reversions to unicellularity, is an exciting possibility, and further studies are certainly warranted. If, for example, the Tetrabaenaceae evolved multicellularity separately from the Goniaceae+Volvocaceae, we could ask interesting questions about convergence versus parallelism: how similar were the genetic and cell biological changes leading to multicellularity in these two lineages? The two extant species of the Tetrabaenaceae diverged much more recently than the other multicellular lineages, meaning that it’s possible they evolved multicellularity much more recently than the Goniaceae+Volvocaceae; if so, it might be a partial explanation for why they seem to have undergone fewer changes. For the time being, I think we have to consider the question unresolved. My opinion, subject to my biases, is that the case for monophyly is currently stronger than that against, but that could certainly change with additional evidence. I look forward to seeing it.

I am sincerely grateful to Drs. Pröschold and Nakada for providing preprints, reprints, and additional information. I would, as always, welcome their comments, clarifications, or corrections.


Stable links:

Bridge, P. D., P. J. Roberts, B. M. Spooner, and G. Panchal. 2003. On the unreliability of published DNA sequences. New Phytol. 160:43–48. DOI: 10.1046/j.1469-8137.2003.00861.x

Hanschen, E. R., M. D. Herron, J. J. Wiens, H. Nozaki, and R. E. Michod. 2017. Repeated evolution and reversibility of self-fertilization in the volvocine green algae. Evolution 72:386–398. DOI: 10.1111/evo.13394

Kirk, D. L. 2005. A twelve-step program for evolving multicellularity and a division of labor. BioEssays 27:299–310. DOI: 10.1002/bies.20197

Libby, E., and W. C. Ratcliff. 2014. Ratcheting the evolution of multicellularity. Science 346:426–427. DOI: 10.1126/science.1262053

Nakada, T., T. Ito, and M. Tomita. 2016. 18S ribosomal RNA gene phylogeny of a colonial Volvocalean lineage (Tetrabaenaceae-Goniaceae-Volvocaceae, Volvocales, Chlorophyceae) and its close relatives. J. Japanese Bot. 91:345–354. DOI not available; Dr. Nakada kindly supplied a reprint.

Nakada, T., Y. Tsuchida, and M. Tomita. 2018. Improved taxon sampling and multigene phylogeny of unicellular chlamydomonads closely related to the colonial volvocalean lineage Tetrabaenaceae-Goniaceae-Volvocaceae (Volvocales, Chlorophyceae). Molecular Phylogenetics and Evolution 130:1–8. DOI: 10.1016/j.ympev.2018.09.013

Nilsson, R. H., M. Ryberg, E. Kristiansson, K. Abarenkov, K. H. Larsson, and U. Köljalg. 2006. Taxonomic reliability of DNA sequences in public sequences databases: A fungal perspective. PLoS One 1:e59. DOI: 10.1371/journal.pone.0000059

Pröschold, T., T. Darienko, L. Krienitz, and A. W. Coleman. 2018. Chlamydomonas schloesseri sp. nov. (Chlamydophyceae, Chlorophyta) revealed by morphology, autolysin cross experiments, and multiple gene analyses. Phytotaxa 362:21–38. DOI: 10.11646/phytotaxa.362.1.2

Leave a Reply