Within the Fungi, simple linear multicellularity of hyphae occurs in all major clades (see below), but only Ascomycota and Basidomycota display more complex two- and three- dimensional multicellularity in the form of sexual spore- producing fruiting bodies. In both of these groups, reversals to unicellular lifeforms have occurred, for example, Saccharomyces and many other related yeasts in the Saccharomycotina (Ascomycota) or Cryptococcus albidus and related species in the hymenomycete clade of Basidiomycota (de Hoog et al. 2000, p. 130).
Medina, M., A. G. Collins, J. W. Taylor, J. W. Valentine, J. H. Lipps, L. A. Amaral Zettler and M. L. Sogin (2003). "Phylogeny of Opisthokonta and the evolution of multicellularity and complexity in Fungi and Metazoa." International Journal of Astrobiology 2(3): 203-211. doi:10.1017/S1473550403001551
Update: The authors responded to criticism like this on The Loom, here's an excerpt:
Our yeast are not utilizing ‘latent’ multicellular genes and reverting back to their wild state. The initial evolution of snowflake yeast is the result of mutations that break the normal mitotic reproductive process, preventing daughter cells from being released as they normally would when division is complete. Again, we know from knockout libraries that this phenotype can be a consequence of many different mutations. This is a loss of function, not a gain of function. You could probably evolve a similar phenotype in nearly any microbe (other than bacteria, binary fission is a fundamentally different process). We find that it is actually much harder to go back to unicellularity once snowflake yeast have evolved, because there are many more ways to break something via mutation than fix it.