CORE Metadata, citation and similar papers at core.ac.

uk
Provided by Elsevier - Publisher Connector

Current Biology, Vol. 12, R62–R64, January 22, 2002, ©2002 Elsevier Science Ltd. All rights reserved. PII S0960-9822(01)00675-3

Chloroplast Evolution: Secondary Dispatch

Symbiogenesis and Multiple Losses

T. Cavalier-Smith and acquired plastids by numerous independent sym-

bioses. But because each symbiogenetic origin of an
organelle is evolutionarily complex, requiring novel
Chloroplasts originated from cyanobacteria only organelle-specific protein-targeting machinery and the
once, but have been laterally transferred to other acquisition by over a thousand genes of appropriate
lineages by symbiogenetic cell mergers. Such targeting signals, I have long maintained that symbio-
secondary symbiogenesis is rarer and chloroplast genesis is very rare and chloroplast loss distinctly
losses commoner than often assumed. commoner [2,3,7]. Although photosynthesis has been
repeatedly lost, complete plastid loss has never been
demonstrated within the plant kingdom — glauco-
During or following the global warming that thawed phytes, red algae and green plants — which arose by
the last ‘snowball earth’ glaciation about 580 million the primary origin of chloroplasts from cyanobacteria.
years ago [1], chloroplasts originated from a However, molecular phylogeny has established that
cyanobacterial symbiont in a biciliate protozoan [2,3]. chloroplast loss is indeed relatively frequent in het-
The resulting cellular chimaera was so successful that erokonts [8] and dinoflagellates [9], and also occurred
it rapidly diversified into two primary lineages of in euglenoids [10].
eukaryotic algae: the now rare glaucophytes like Secondary symbiogenesis has left remarkable
Cyanophora, which retained the cyanobacterial traces of its evolutionary role in the more complex
peptidoglycan wall within their chloroplast envelope, topology of the membranes surrounding all non-plant
and the green plant/red algal lineage, which lost the chloroplasts. The kingdom Chromista was established
peptidoglycan. The latter split into red algae, which in 1981 to embrace all algae — cryptomonads, het-
retained the cyanobacterial phycobilisome pigments, erokonts and haptophytes — with chloroplasts located
and green algae, which replaced them by chlorophyll within the lumen of the rough endoplasmic reticulum
b to adapt to different light frequencies. Soon after- (ER), plus their heterotrophic relatives like the het-
wards, a red alga and two different green algal cells erokont oomycetes which are deemed to have
were implanted into yet other biciliate hosts to form evolved from them by chloroplast loss [2]. Chloro-
three further groups of eukaryotic algae: a process plasts of all chromists are separated from the sur-
called secondary symbiogenesis (see Figure 1). New rounding rough ER by a smooth membrane, the
light on these early events in eukaryotic evolution periplastid membrane, the relic of the plasma mem-
comes from host genes encoding proteins that were brane of the red alga phagocytosed by the ancestor of
secondarily imported into the acquired plastids [4] all chromists. This unique topology probably arose
and, as reported by Andersson and Roger [5] in this when the former food vacuole membrane fused with
issue, from symbiont genes that were apparently the nuclear envelope of the engulfing host. In cryp-
retained in the host nucleus after the symbiont was tomonads, the nucleus of the enslaved red alga per-
lost. Secondary symbiogenesis has also been greatly sists because it has 30 genes encoding proteins
clarified by the complete sequence of a cryptomonad essential for the secondarily acquired plastid [6]. But
nucleomorph [6], an evolutionarily miniaturised relic of in other chromists — heterokonts and haptophytes,
the red algal nucleus that was enslaved over collectively called chromobiotes as they share the
500 million years ago. brown carotenoid fucoxanthin that colours brown sea-
Several well-established eukaryote groups comprise weeds and diatoms — these thirty genes were trans-
a mixture of photosynthetic algae and non-photosyn- ferred to the host nucleus, allowing the nucleomorph
thetic heterotrophs, notably dinoflagellates, het- to be lost with its evolutionary burden of hundreds of
erokonts, cryptophytes and Euglenozoa. We now know housekeeping genes (needed only for expressing
that all these acquired chloroplasts secondarily from those 30 genes).
other eukaryotes (red or green algae). Early in the In all chromistan algae, thousands of plastid
twentieth century, however, it was thought that such proteins encoded by nuclear genes must be imported
groups with mixed nutritional properties were ances- across four topologically and chemically distinct
trally photosynthetic and their non-photosynthetic membranes: the rough ER, the periplastid membrane,
members evolved by losing chloroplasts. After the and the outer and inner chloroplast envelope
1960s revival of Mereschkowsky’s symbiogenetic membranes. Their bipartite amino-terminal targeting
theory of chloroplast origins, the alternative dogma sequences direct this: distally is a signal sequence
arose that such groups were ancestrally heterotrophic mediating import into the rough ER lumen [11], and
proximally a chloroplast transit peptide that mediates
Department of Zoology, University of Oxford, South Parks import across the chloroplast envelope and possibly
Road, Oxford, OX1 3PS, UK. E-mail: tom.cavalier- also the periplastid membrane [12]. Traditionally,
smith@zoo.ox.ac.uk chromistan algae were grouped as ‘chromophyte
Current Biology
R63

Figure 1.
Kingdom Plantae Chromalveolates Phylogeny of the five eukaryote king-
Rhodophyta Kingdom Chromista doms, highlighting the primary symbio-
Viridaeplantae Cryptophyta Chromobiotes genetic origin of chloroplasts from
(green plants) Heterokonta Haptophyta cyanobacteria which created the plant
Glaucophyta kingdom, and the three secondary sym-
R biogenetic events which created eukary-
ote–eukaryote chimaeras by laterally
Chloroplast Miozoa Alveolates implanting pre-existing chloroplasts into
Ciliophora other protozoan lineages (dashed lines). A
Excavates single implantation of a red algal chloro-
Euglenozoa Cortical alveoli
Percolozoa plast (R) created the typically brown
G Kingdom Kingdom chromalveolates, in which several lin-
eages, such as the ciliates (Ciliophora),
Archezoa Parabasalia Animalia Fungi
later lost photosynthesis. Cryptomonad
Metamonada
Loukozoa Choanozoa
nucleomorphs are relics of the endosym-
G Cercozoa biotic red alga’s nucleus that have per-
Retaria sisted in these cellular chimaeras for half
Apusozoa a billion years. Two separate implanta-
Opisthokonts tions of green algal cells (G) took place.
Biciliate eukaryotes
Amoebozoa One gave rise to euglenoid chloroplasts
(probably in the common ancestor of
Ancestral heterotrophic uniciliate eukaryote Euglenozoa and Percolozoa or an even
Current biology
earlier excavate). The other was into a
cercozoan flagellate giving the chloro-
plasts and nucleomorphs of chlorarachnean algae. At least two independent tertiary replacements of typical dinoflagellate chloro-
plasts by phylogenetically different ones also occurred [9,13], but are not shown (dinoflagellates belong in Miozoa with Sporozoa, which
are non-photosynthetic but mostly retained plastids for lipid synthesis). The 12 phyla in black belong in the kingdom Protozoa (see
[17], where the rationale behind this phylogeny is detailed). Though Percolozoa are shown as sisters to Euglenozoa, some rRNA trees
[17] suggest that they may be sisters to Archezoa instead; if this were true, Archezoa also must have had photosynthetic ancestors.

algae’ with dinoflagellates, which have a different replacement in chromists and alveolates [4], collec-
brown carotenoid, peridinin. But dinoflagellate plas- tively called chromalveolates [12].
tids are apparently not within the rough ER lumen, and Fast et al. [4] discovered that, in two chromist
they have an envelope of only three smooth mem- groups (cryptomonads and heterokonts) and two alve-
branes. It has therefore long been assumed that their olate groups (dinoflagellates and Sporozoa), the glyc-
plastids had a separate secondary symbiogenetic eraldehyde phosphate dehydrogenase (GAPDH) gene
origin; many also discounted the arguments for a that originally encoded the chloroplast GAPDH protein
single symbiogenetic origin for all chromistan plastids has been replaced by a duplicate of the radically dif-
[13]. Discovery of the remarkable single-gene circles ferent host gene encoding cytosolic GAPDH. Further-
of dinoflagellate chloroplasts [14] and their phyloge- more, they found that the chromalveolate algal
netic analysis [15] led to a radical reappraisal of replacement genes branch on trees with the ciliate
chromist and dinoflagellate relationships [12]. Not only gene for the cytosolic protein, not with those of non-
chromists and dinoflagellates, but also Sporozoa, alveolates. If these replacements happened indepen-
such as malaria parasites and Toxoplasma, apparently dently, one would have to suppose that the same
had plastids of red algal origin [15]. gene underwent four identical duplications and
A substantial simplification of eukaryote phylogeny retargetings from cytosol to plastid by acquiring
was effected by proposing that all these groups are bipartite targeting signals for import across four (or in
related, and that their red algal chloroplasts had been dinoflagellates three) membranes. Furthermore, the
acquired not independently in five separate sym- retargeted cytosolic enzymes would have had to
bioses, as many then supposed [13], but in a single change their cofactor specificity from NAD to NADPH
secondary symbiogenetic event to form a photosyn- independently.
thetic common ancestor of all chromophytes [12]. It Such multiple convergence of duplication, acquisi-
was already known that dinoflagellates and Sporozoa tion of targeting sequences, and enzymatic remould-
are mutually related by descent from an obscure ing is simply incredible, especially when we bear in
group of flagellates, the protalveolates, grouped with mind that the ancestral chimaeric cells also had host
them in the phylum Miozoa [12]. Miozoa are related to and symbiont genes for distinct mitochondrial GAPDH
the apparently non-photosynthetic ciliate protozoa, enzymes and symbiont cytosolic genes, any of which
the two groups being collectively called alveolates in principle could have undergone duplication and
because of related ultrastructural features of their cell retargeting, and that gene replacement is unlikely
cortex. If alveolates and chromists really do have a even to have been physiologically necessary — it was
photosynthetic common ancestor, ciliates and pro- probably just an evolutionary accident. It is far simpler
talveolates must also be derived from it and have lost to invoke a single gene replacement in the photosyn-
photosynthesis. That this is almost certainly so is now thetic common ancestor of all chromalveolates [4,12]
dramatically shown by a remarkable case of gene and accept that all heterotrophic chromalveolates
Dispatch
R64

evolved by losing photosynthesis. This was probably 4. Fast, N.M., Kissinger, J.C., Roos, D.S., Keeling, P.J. (2001).
easier in chromalveolates because, unlike plants, Nuclear encoded, plastid-targeted genes suggest a single
many retain a capacity for phagotrophy, often being common origin for apicomplexan and dinoflagellates plas-
tids. Mol. Biol. Evol. 18, 418–426.
mixotrophs able both to photosynthesise and phago-
5. Andersson, J.A., Roger, A.J. (2002). A cyanobacterial gene
cytose. Thus, in the protist world they fill a niche like in non-photosynthetic protists — an early chloroplast
corals among animals, which themselves rely on acquisition in eukaryotes. Curr. Biol., this issue.
temporarily cultured intracellular photosynthetic 6. Douglas, S., Zauner, S., Fraunholz, M., Beaton, M.J.,
dinoflagellates. Penny, S., Deng, L.-T., Wu, X., Reith, M., Cavalier-Smith,
So we must now accept that chromalveolates form T., Maier, U.-G. (2001). The highly reduced genome of an
enslaved algal nucleus. Nature 410, 1081–1086.
a major monophyletic branch of the eukaryotic tree.
7. Cavalier-Smith, T. (1993). The origin, losses and gains of
Andersson and Roger [5] have independently shown chloroplasts. In Origin of Plastids: Symbiogenesis,
that the heterotrophic oomycete heterokonts, once mis- Prochlorophytes and the Origins of Chloroplasts, R.A.
classified as fungi, did evolve from algal ancestors: they Lewin, ed. (New York : Chapman & Hall) pp. 291–348.
have a 6-phosphogluconate dehydrogenase (gnd) gene 8. Cavalier-Smith, T., Chao, E.E., Allsopp, M.T.E.P (1995).
closely related to that of diatoms and brown algae, Ribosomal RNA evidence for chloroplast loss within Het-
erokonta: pedinellid relationships and a revised classifica-
which must have come from the chromalveolate red tion of ochristan algae. Archiv. Protistenk. 145, 209–220.
algal symbiont. As glaucophytes, the most divergent 9. Saldarriaga, J., Taylor, F.J.R., Keeling, P.J., Cavalier-Smith,
plant group, have cortical alveoli like those of alveo- T. (2001). Dinoflagellate nuclear SSU rRNA phylogeny sug-
lates, plants and chromalveolates may actually be sister gests multiple plastid losses and replacements. J. Mol.
groups [12]. Though collectively called photokaryotes Evol. 53, 204–213.
[12], it is usually thought that their common ancestor 10. Preisfeld, A., Busse, I., Klingberg, M., Talke, S., Ruppel,
H.G. (2001). Phylogenetic position and inter-relationships
was non-photosynthetic. However, Andersson and
of the osmotrophic euglenids based on SSU rDNA data,
Roger [5] also found gnd genes clearly related to those with emphasis on the Rhabdomonadales (Euglenozoa). Int.
of plants and chromists in the purely heterotrophic J. Syst. Evol. Microbiol. 51, 751–758.
amoeboflagellate Heterolobosea (phylum Percolozoa); 11. Ishida, K., Cavalier-Smith, T., Green, B.R. (2000).
the simplest interpretation is that Percolozoa evolved Endomembrane structure and the chloroplast protein tar-
by plastid loss from photosynthetic ancestors. geting pathway in Heterosigma akashiwo (Raphido-
phyceae, Chromista). J. Phycol. 36, 1135–1144.
Either the primary origin of chloroplasts or a lateral
12. Cavalier-Smith, T. (1999). Principles of protein and lipid tar-
transfer by secondary symbiogenesis occurred earlier geting in secondary symbiogenesis: euglenoid, dinoflagel-
than usually thought. I favour the latter, as Percolozoa late, and sporozoan plastid origins and the eukaryote
are probably sisters to Euglenozoa [12,16,17], which family tree. J. Eukaryot. Microbiol. 46, 347–366.
acquired plastids from a green alga. If this took place 13. Delwiche, C.F. (1999). Tracing the thread of plastid diver-
in the common ancestor of Euglenozoa and Percolo- sity through the tapestry of life. Am. Nat. 154, S164–S177.
zoa, we need not postulate primary plastid loss (never 14. Zhang, Z., Green, B.R., Cavalier-Smith, T. (1999). Single
gene circles in dinoflagellate chloroplast genomes. Nature
clearly demonstrated) in any eukaryotic lineage, merely 400, 155–159.
additional losses of secondary plastids (already known) 15. Zhang, Z., Green, B.R., Cavalier-Smith, T. (2000). Phy-
in the ancestral percolozoan and within Euglenozoa. logeny of ultra-rapidly evolving dinoflagellate chloroplast
However, the suggestion [12] that euglenoid plastids genes: a possible common origin for sporozoan and
were acquired even earlier, in the same event as chlo- dinoflagellate plastids. J. Mol. Evol. 51, 26–40.
rarachnean chloroplasts and nucleomorphs [18,19], is 16. Baldauf, S.L., Roger, A.J., Wenk-Siefert, I., Doolittle, W.F.
(2000). Kingdom-level phylogeny of eukaryotes based on
contradicted by cytoskeletal phylogeny [17]. Chlo- combined data. Science 290, 972–977.
rarachnean chloroplast genome sequences should 17. Cavalier-Smith, T. (2002). The phagotrophic origin of
eventually confirm their independence. eukaryotes and phylogenetic classification of Protozoa.
Andersson and Roger [5] also noted highly divergent Int. J. Syst. Evol. Microbiol., in press.
gnd genes in trypanosomes (Euglenozoa). It is unclear 18. Gilson, P., McFadden, G.I. (1997) Good things in small
whether these came from the green algal symbiont or packages: the tiny genomes of chlorarachniophyte
endosymbionts. BioEssays 19, 167–173.
the host; the secondarily amitochondrial Archezoa
19. Ishida, K., Green, B.R., Cavalier-Smith, T. (1999). Diversifi-
also have comparably divergent gnd genes [5], con-
cation of a chimaeric algal group, the chlorarachniophytes:
ceivably related as Euglenozoa and Archezoa are both phylogeny of nuclear and nucleomorph small subunit rRNA
excavate protozoa [17]. Finding other analogous genes. Mol. Biol. Evol. 16, 321–331.
genetic relics of distant evolutionary events should
provide an even clearer picture of the broad lines of
early eukaryotic diversification.

References
1. Hoffman, P.F., Kaufman, A.J., Halverson, G.P., Schrag,
D.P. (1998). A neoproterozoic snowball earth. Science 281,
1342–1346.
2. Cavalier-Smith, T. (1982). The origins of plastids. Biol. J.
Linn. Soc. 17, 289-306.
3. Cavalier-Smith, T. (2000). Membrane heredity and early
chloroplast evolution. Trends Plant Sci. 5, 174–182.