COMMENTARY

The tiny enslaved genome of a rhizarian alga

Thomas Cavalier-Smith* Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS, United

t least twice in the history of life a predatory nonphotosyn- thetic protozoan cell ate a eu- karyotic algal cell and enslaved Ait internally instead of digesting it, thereby becoming a chimeric photoph- agotrophic cell with two distinct nuclei and remarkably complex membrane topology—far surpassing that of or . In their descendants, the host Fig. 1. Apart from plants formed by permanent enslavement of a cyanobacterium to create the first (20), there are three groups of photosynthetic . All of these groups belong to the nucleus became dominant, whereas that and were formed when enslaved from the kingdom to make – of the algal slave shrank by transfer of eukaryote cellular chimeras. A red alga was enslaved to form chromalveolates, and were most of its genes into the main nucleus enslaved to form euglenoid algae within and chlorarachnean algae within , two and retargeting the proteins they encode otherwise entirely heterotrophic protozoan infrakingdoms. The enslaved algal nucleus persists in minia- back into the enslaved . Some turized form as a tiny ‘‘’’ only in within the chromalveolates and in the descendants managed to transfer all es- chlorarachnean cercozoan algae in the Rhizaria. Gilson et al. (5) have sequenced the tiny nucleomorph sential genes and lose the enslaved nu- genome of the flagellate chlorarachnean . clei altogether while retaining the algal chloroplast and plasma membrane (e.g., ans. Chlorarachnea is Greek for green Organisms are not optimized, let alone brown seaweeds), but in two groups of spider, referring to the web-like body of designed, because selection is powerless algae, cryptomonads (1) and chlorarach- the archetypal Chlorarachnion,in without the requisite chance mutations. neans (2–4), the enslaved nucleus re- Thus, the nucleomorph of Bigelowiella mains in the chimeric cell many millions which green-chloroplast-containing cells of years afterward, raising fascinating form a multicellular network by tempo- was retained solely because 17 genes are questions for cell and evolutionary biol- rary fusions of thread-like pseudopods essential for the chloroplast, just as were ogy. How are functions shared between that trap diverse prey for phagocytic the because two evolutionarily unrelated nuclei, and engulfment and digestion. of the 30 chloroplast proteins they en- how do their proteins integrate into one Compared with other algae, code (1). If, by chance, copies of all of harmonious cell? The first inklings of chlorarachneans are little known partly these few genes had moved into the nu- answers came with the first nucleo- because they live largely in subtropical cleus and their products successfully re- morph genome sequence from a crypto- waters. Japanese researchers found the targeted into the chloroplasts through monad (1). In this issue of PNAS, biodiversity of chlorarachneans to be the four surrounding membranes, the Gilson et al. (5) report the first genome greater than previously recognized (11) nucleomorph genome would have disap- sequence of a chlorarachnean nucleo- and that they include classical amoeboid peared. Such loss occurred in chromal- morph, the tiniest nucleus in nature. forms, others with cell walls, and flagel- veolates, both in the chromobiote sisters Fig. 1 shows how these two unusual lates like Bigelowiella (12) with potential of cryptomonads and in (e.g., algal groups fit on the eukaryote evolu- as a laboratory model, being easier to and sporozoa like the tionary tree (6, 7). Cryptomonads grow; many display all three growth malaria parasite) (10), but never in belong to a vast branch, the chromalveo- forms at different life-cycle stages. chlorarachneans. lates, formed by a single enslavement Thus, chlorarachneans and crypto- However, euglenoid algae (members of a red alga (8, 9) and containing eight monads represent entirely independent of Excavata) also have green phyla: two mainly photosynthetic natural experiments in nuclear genome acquired from a green alga long ago. [ (including , miniaturization that are most interesting Their ancestor must once have had a , and eight other algal classes) to compare (13). What do such compar- nucleomorph and transferred all essen- and Haptophyta]; two with one algal isons tell us? As expected, Bigelowiella tial chloroplast protein genes into its class each (dinoflagellates and crypto- nucleomorphs have far fewer genes than own nucleus, retargeted their proteins to monads) but many ; and those of cryptomonads; only 331 pro- the chloroplast, and then lost the four entirely nonphotosynthetic (e.g., tein-coding genes compared with 464 in nucleomorph entirely. It is possible that protozoa, and Pseudofungi) (10). the cryptomonad Guillardia (1). As in retargeting was easier for euglenoids, in By contrast, chlorarachneans belong to Guillardia, most genes are for house- which proteins have to cross only three Rhizaria, a typically amoeboid group keeping, merely to maintain the nucleo- membranes and the targeting machinery characterized by long, thin-branching morph itself and the that via Golgi vesicles that fuse with the out- pseudopods that often anastomose as a make its proteins. The only end-product ermost membrane copes with proteins net. Rhizaria have two phyla: , functions of obvious direct value to the temporarily stuck in vesicle membranes wherein chlorarachneans are the only host are provided by 17 genes that en- by hydrophobic sequences (14). It is algae, and , including the largely code proteins imported into the former widely assumed that chlorarachneans marine and , green algal chloroplast. The 20-fold which often cultivate algae temporarily more numerous housekeeping genes are in their cells, much as corals do, but kept merely to allow expression of these Conflict of interest statement: No conflicts declared. never permanently enslave them as a 17 and are an evolutionary load, reflect- See companion article on page 9566. true chloroplast with its own protein- ing the sometimes bizarre imperfections *E-mail: [email protected]. import machinery like in chlorarachne- of evolution by mutation and selection. © 2006 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0603505103 PNAS ͉ June 20, 2006 ͉ vol. 103 ͉ no. 25 ͉ 9379–9380 Downloaded by guest on October 2, 2021 enslaved a green alga independently of by Guillardia. Given originally Ϸ1,000 living green plants; some occur in the euglenoids (4), but the possibility that chloroplast protein genes in the enslaved same position (four are even shared this occurred once only in a hypothetical algal nuclei, shared retention of these with Guillardia and must predate the common ancestor of Rhizaria and Exca- two is probably just a coincidence. red͞green algal divergence), implying vata and that heterotrophic members of There appears no rhyme or reason why that 135–570 My of selection for small both groups all lost chloroplasts (6, 8) is certain genes were successfully trans- size failed to streamline the genome as not yet decisively ruled out. ferred and others were not (5). It is much as a designer could have. Yet in- Deciding which history is true is probably through historical accidents tron size was dramatically reduced to important for estimating the age of that chlorarachneans and cryptomonads 18–21 nt (mostly 19 nt), suggesting a chlorarachnean nucleomorphs. Judging failed to transfer all of their genes to novel, possibly length-dependent mecha- from sequence trees (15), the last com- nism of splicing. mon ancestor of chlorarachneans may Unlike in cryptomonads, the nucleo- be only a quarter of the age of Cerco- Degree of compaction morph lost genes for tubulins and pro- zoa, but, because their lineage branches teasomes, raising the question whether early, the first chlorarachnean could is not simply proteasome proteins are imported into have been 70% of that age. Thus, if a the periplastid space or are dispensed single green alga were enslaved by an related to elapsed with. Previously, divergent nuclear- excavate͞rhizarian ancestor, chlorarach- coded tubulins were found with leaders, nean nucleomorphs must be marginally evolutionary time. suggestive of targeting signals (19); thus, older than Cercozoa, which probably go it is likely that they are imported and back 540 million years (My) (16). If, in- that the nucleomorph divides by a relict mitotic spindle as in Guillardia (1). stead, chlorarachneans acquired plastids the nucleus, whereas chromobiotes, al- What remains for the future? With independently of euglenoids, their veolates, and euglenoids independently unpublished sequences of Bigelowiella nucleomorphs are probably 135–380 My succeeded. Retention of Toc75͞Tic20 chloroplast and mitochondrial genomes old. Cryptomonad nucleomorphs are but not Tic22 genes for chloroplast pro- completed in Canada and Japan, we as old as chromalveolates, probably tein import, unlike Guillardia, also is Ϸ greatly need nuclear genome sequences 570 My (16). Thus, chlorarachnean likely accidental. for Bigelowiella and Guillardia to under- nucleomorphs are no older than those Determinism is evident in the way stand how four genomes cooperate to of cryptomonads and are possibly much intergenic spacers have been pared make such complex chimeric cells. Rhi- younger, yet they did shrink more. Thus, down to as little as 1 nt or less: some zaria are the only major group in the degree of compaction is not simply re- genes overlap by up to 110 nt. These living world with no complete nuclear lated to elapsed evolutionary time, as very short spacers and overlapping genome sequences. For fuller under- also shown by chromalveolates that lost genes must reflect systematically greater standing of the cell biology and lifestyles nucleomorphs and were no older than reproductive success of deletions com- of these unusual organisms, which may cryptomonads that did not. pared with insertions. Elimination of be the most abundant predators on Nucleomorph genome reduction in- intergenic noncoding DNA, pseudo- Earth as Cercozoa dominate the soil volves three processes: transfer of gene genes, and gene duplicates attest to and Retaria dominate marine sediments copies to the nucleus, their acquisition strong selection for small genomes (18), and the oligotrophic ocean vastnesses, of bipartite targeting signals for protein but chlorarachnean nucleomorph genes we need genome sequences from diver- import across four membranes (17), and riddled with [852 in all, 160 in- gent heterotrophic Rhizaria, especially deletion of the original nucleomorph dependently confirmed as spliced by Cercozoa (far easier to cultivate). A copy. This gene-transfer process has parallel sequencing of DNA comple- final answer to the conundrum whether both chance and deterministic aspects. mentary to mRNAs (5)] show that such chlorarachnean and euglenoid plastids The role of historical accident is shown selection is mutationally limited: precise were enslaved in a common ancestor or by only 2 of the 17 chloroplast protein deletion is probably difficult. not may depend on elucidating the genes of Bigelowiella nucleomorphs be- Gilson et al. (5) show that den- chlorarachnean protein-targeting ing the same as any of the 30 retained sity is essentially the same as in free- systems.

1. Douglas, S., Zauner, S., Fraunholz, M., Beaton, 7. Richards, T. A. & Cavalier-Smith, T. (2005) Na- 14. Slavikova, S., Vacula, R., Fang, Z., Ehara, T., M., Penny, S., Deng, L. T., Wu, X., Reith, M., ture 436, 1113–1118. Osafune, T. & Schwartzbach, S. D. (2005) J. Cell Cavalier-Smith, T. & Maier, U. G. (2001) Nature 8. Cavalier-Smith, T. (1999) J. Eukaryotic Microbiol. Sci. 118, 1651–1661. 410, 1091–1096. 46, 347–366. 15. Cavalier-Smith, T. & Chao, E. E. (2003) 2. Hibberd, D. J. & Norris, R. (1984) J. Phycol. 20, 9. Patron, N. C., Rogers, M. B. & Keeling, P. J. 154, 341–358. 310–330. (2004) Eukaryotic Cell 3, 1169–1175. 16. Cavalier-Smith, T. (2006) Phil. Trans. R. Soc. 3. Gilson, P. & McFadden, G. I. (1995) Chromosoma 10. Cavalier-Smith, T. (2004) in Organelles, Genomes London B 361, 969–1006. 103, 635–641. and Eukaryote Phylogeny, eds. Hirt, R. P. & Hor- 17. Rogers, M. B., Archibald, J. M., Field, M. A., Li, 4. Ishida, K., Cao, Y., Hasegawa, M., Okada, N. & ner, D. S. (CRC, London), pp. 75–108. C., Striepen, B. & Keeling, P. J. (2004) J. Eukary- Hara, Y. (1997) J. Mol. Evol. 45, 682–687. 11. Ishida, K., Nakayama T. & Hara, N. (1996) Phycol. otic Microbiol. 51, 529–535. 5. Gilson, P. R., Su, V., Slamovits, C. H., Reith, Res. 44, 37–45. 18. Cavalier-Smith, T. (2005) Ann. Bot. 95, 147–175. M. E., Keeling, P. J. & McFadden, G. I. (2006) 12. Moestrup, Ø. & Sengco, M. (2001) J. Phycol. 37, 19. Keeling, P. J., Deane, J. A. & McFadden, G. I. Proc. Natl. Acad. Sci. USA 103, 9566–9571. 624–646. (1998) J. Eukaryotic Microbiol. 45, 561–570. 6. Cavalier-Smith, T. (2003) Eur. J. Protistol. 39, 13. Cavalier-Smith, T. (2003) Phil. Trans. R. Soc. 20. Cavalier-Smith, T. (2000) Trends Plant Sci. 5, 338–348. London B 358, 109–134. 174–182.

9380 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0603505103 Cavalier-Smith Downloaded by guest on October 2, 2021