A Tertiary Plastid Gains RNA Editing in Its New Host

Christopher J. Jackson,z Sebastian G. Gornik, and Ross F. Waller*

School of Botany, University of Melbourne, Victoria, Australia
z
Present address: Biology Department, University of New Brunswick, New Brunswick, Canada
*Corresponding author: E-mail: r.waller@unimelb.edu.au.
Associate editor: Hervé Philippe
New sequences are available in GenBank (accession nos. JX292160–JX292162, JX292164–JX292175, and JX945981–JX945986) or in
the supplementary data, Supplementary Material online.

Abstract
Dinoflagellates are known for their development of highly aberrant organelle genetic systems. Both their plastid and
mitochondrial genomes are extremely reduced in gene number and rearranged into numerous unconventional genomic
elements. Transcription processes are also elaborately modified including extensive RNA editing and trans-splicing. Some

Downloaded from http://mbe.oxfordjournals.org/ at Universidade Federal de Viçosa on July 17, 2013

dinoflagellates have replaced their original plastid through serial endosymbiotic events. Karlodinium veneficum is such an
example that now contains a haptophyte plastid. This tertiary plastid provides a case of a more conventional genetic
system introduced into a cellular environment with a known penchant for genetic oddities. Here, we show that
K. veneficum plastid transcripts undergo extensive substitutional editing. The substitution types are more diverse than
those seen in most other plastids but are similar to those of dinoflagellate organelles. There is no evidence for RNA editing
of plastid-encoded transcripts from extant haptophytes, suggesting that K. veneficum plastid editing developed after the
uptake of the tertiary endosymbiont.
Key words: RNA editing, plastid, endosymbiosis, dinoflagellate.

The development of endosymbiotic organelles requires two
cells to learn to live together. Plastids and mitochondria owe
their existence to bacteria that became resident inside a host
eukaryote. The development of these organelles required
metabolic integration of the two cells, as well as molecular
harmonization of their genetic functions. Typically, the endo-
symbiont surrenders much of its genetic content to the host
cell nucleus but still maintains a reduced genome expressed
semiautonomously using retained bacterial expression ma-
Letter
replaced their peridinin plastid by a subsequent, or serial,
endosymbiosis (Archibald 2009). The sources of these new
plastids have been diverse, including from haptophytes, dia-
toms, cryptomonads (all resulting in tertiary plastids), and
green algae (serial secondary plastids) (Archibald 2009).
Dinoflagellates, therefore, present multiple opportunities for
examining the process of host–endosymbiont cell
integration.
The best studied of these recent endosymbioses in

chinery (Gould et al. 2008; Barbrook et al. 2010). All mito-
chondria are derived from a single ancient endosymbiotic
event, whereas plastids have migrated horizontally into new
hosts through several subsequent endosymbioses. These plas-
tid transfers occur when a eukaryote, already bearing a pri-
mary plastid, becomes an endosymbiont itself in a new host.
Plastids derived this way are known as secondary plastids or
tertiary plastids if the endosymbiont already bore a secondary
plastid (Archibald 2009). With each new endosymbiosis, a
new round of molecular harmonization must commence be-
tween the host cell and symbiont.
Dinoflagellate algae are unusual among eukaryotes in that
this phylum has gained plastid endosymbionts multiple dif-
ferent times. The ancestor of dinoflagellates and their sister
phylum, Apicomplexa, most likely inherited a common plas-
tid originally derived from a red algal endosymbiont
(Archibald 2009; Janouškovec et al. 2010). In photosynthetic
dinoflagellates (approximately half have now lost photosyn-
thesis), this ancestral plastid contains the distinctive pigment
peridinin. Four dinoflagellate groups have more recently
dinoflagellates is the gain of a haptophyte to create the
tertiary plastid found in three dinoflagellate genera,
Karlodinium, Karenia, and Takayama. The origin of this
new endosymbiont is clear as it contains the distinctive
haptophyte accessory pigments (190 -hexanoyloxy-
0
fucoxanthin and 19 -butanoyloxy-fucoxanthin) in place of
peridinin and phylogenies using plastid genes strongly
group it with haptophytes (Tengs et al. 2000; Ishida and
Green 2002; Patron et al. 2006). This endosymbiont is now
well integrated into its dinoflagellate host, with the endosym-
biont nucleus eliminated and only the photosynthetic plastid
retained, surrounded by 3–4 membranes (Dodge 1989). Most
endosymbiont genes are relocated to the host nucleus and,
therefore, are under the direct control of the new host (Ishida
and Green 2002; Takishita et al. 2004; Nosenko et al. 2006;
Patron et al. 2006; Shalchian-Tabrizi et al. 2006; Gabrielsen
et al. 2011). In Karlodinium veneficum, only 70 protein
genes are retained in the plastid, encoded on the plastid
genome inherited from the haptophyte (Gabrielsen et al.
2011).

ß The Author 2012. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For permissions, please
e-mail: journals.permissions@oup.com

Mol. Biol. Evol. 30(4):788–792 doi:10.1093/molbev/mss270 Advance Access publication November 28, 2012 788
A Tertiary Plastid Gains RNA Editing in Its New Host . doi:10.1093/molbev/mss270
In this study, we present evidence that in K. veneficum,
following endosymbiosis within the dinoflagellate host, plas-
tid genetic processes have been re-engineered with the ac-
quisition of an RNA editing system. From the recent 454
assembly of the K. veneficum plastid genome, many gene
sequences were noted as being unusual, including unex-
pected sequence divergence, the presence of canonical stop
codons interrupting genes, and five genes with apparent
frameshifts (Gabrielsen et al. 2011). To examine transcrip-
tional processes in this endosymbiont, we generated gene
and transcript sequences corresponding to 14 plastid-
encoded genes (psaA, psbC, rbcL, atpG, atpI, rpl5, rps13,
rps18, rpl33, rpl36, petD, rpoB, rpoC2, and secY—totaling
7,373 nucleotides) including all five purported instances of
frameshifts, as well as cases of internal stop codons. We used a
proof reading polymerase (Platinum Taq DNA Polymerase
High Fidelity) and generated multiple independent amplicons
of each to verify sequencing fidelity (primer binding sites were
excluded from sequence data). These data revealed extensive
substitutional RNA editing occurring in the K. veneficum plas-
tid consisting exclusively of the four transitions, predomin-
antly A to G and U to C (table 1 and fig. 1). Editing affects on
average 3% of nucleotides (up to 8.6% for some genes), is
heavily biased for codon positions 1 and 2 (93%), and
almost always results in nonsynonymous changes (95%). All
cases of internal stop codons examined (in psaA, rps13, rpoB,
and secY) were edited to restore open reading frames (e.g.,
fig. 1). We found no evidence of insertional/deletional editing
based on comparison of gene to transcript sequences. In four
of the five genes reported by Gabrielsen et al. (2011) to con-
tain frameshifts (rpl5, rpoB, rpoC2, and secY), our sequences
lacked the frameshifts, indicating 454 sequencing errors (that
often occurred in regions of homopolymers). In the fifth case
(petD), transcript data suggest that this gene is simply shor-
tened by 18 nucleotides by a new but genuine stop codon.
Our data, therefore, suggest that no frameshifts are present in
K. veneficum plastid genes. In free-living haptophytes, no re-
ports of RNA editing have been made, and we independently
generated genomic DNA (gDNA) and cDNA sequences for
three plastid genes (psaA, psbC, and rbcL) from the hapto-
phyte Emiliania huxleyi and found no evidence of editing
(70 edits occur in the corresponding K. veneficum genes).
This suggests that RNA editing in the K. veneficum plastid
was acquired after endosymbiosis in the dinoflagellate host.
One effect of RNA editing in the K. veneficum plastid is that
transcript-encoded proteins share greater identity to homo-
logs in related organisms than do the direct translations of the
gene sequences. We compared conceptual translations of
K. veneficum gDNA and cDNA sequences to proteins from
haptophyte E. huxleyi (no other complete gene data set for a
haptophyte or dinoflagellate peridinin plastid exists to enable
broader comparison) and found an average increase of 2.0%
identity after editing and up to 6.3% for one gene (rpl36)
(table 1). Random nucleotide edits would most often
reduce protein identity, so the predominant increases seen
indicate that editing events are under selection to maintain
protein similarity and function (it is not possible to know
whether small decreases in identity seen in rps13, rps18,
MBE
rpoB, and rpoC2 compared with E. huxleyi proteins affect
function in these cases). A second consequence of editing is
a net increase in the A + T nucleotide content of plastid genes
compared with transcripts, and this occurs predominantly at
codon position 1 then position 2 (table 2). The strong overall
A + T bias of this genome (72.9%, Gabrielsen et al. 2011) is
greatest at codon position 3 as this position least often con-
strains the amino acid specified. Hence, the predominance of
A to G and U to C editing at positions 1 and 2 has allowed
further A + T drift to occur at these more limiting positions.
Together these data indicate that RNA editing has allowed a
more extreme A + T rich genome to develop, while amelio-
rating potential negative impacts to protein sequence and
function.
RNA editing in the K. veneficum plastid presents a dramatic
Downloaded from http://mbe.oxfordjournals.org/ at Universidade Federal de Viçosa on July 17, 2013
change from a nonediting to editing transcriptional system.
So how and why has this system of RNA editing been
acquired? To address how an editing system might have
evolved requires knowledge of the molecular machinery
used for editing. Throughout eukaryotes, RNA editing has
evolved many times, is most common in organelles, and is
most often limited to one or two substitution types (Knoop
2011; Gray 2012). However, the biochemistry of few substitu-
tional editing systems is well understood. The best known is
from plant organelles, where a family of sequence-specific
RNA-binding proteins is thought to direct cyti-
dine deaminases that catalyze the base conversion C to U,
the major editing event in plants (Salone et al. 2007;
Chateigner-Boutin and Small 2010). Karlodinium veneficum
plastid RNA editing consists of four substitution types, imply-
ing a more complex machinery than this. In other dinofla-
gellate organelles, even more extensive and diverse
substitutional RNA editing types occur, including transver-
sions. Dinoflagellate mitochondrial editing is best docu-
mented; it occurs widely in this phylum, 11 of a possible 12
substitution types are known to occur and editing densities of
up to 6% are seen (cox3, K. veneficum) (Lin et al. 2008; Waller
and Jackson 2009; Jackson et al. 2012). Fewer reports of editing
in the peridinin plastid have been made, but from
Heterocapsa triquetra and Lingulodinium polyedrum, up to
eight different substitution types have been reported
(Wang and Morse 2006; Dang and Green 2009). Although
the mechanism of editing in dinoflagellate organelles is en-
tirely uncharacterized, enzymatic base conversion is unable to
account for these more diverse types of transitions and trans-
versions, and a nucleotide excision and replacement mech-
anism is envisaged for such editing systems.
It might be that the K. veneficum plastid has independently
developed editing machinery de novo. However, the relative
complexity of this editing system (four substitution types),
and the presence of complex substitutional editing in both
mitochondria and peridinin plastids of dinoflagellates, pre-
sents an alternative hypothesis of adoption of pre-existing
editing machinery. In plants, common components of the
"editosome" are used in both plastids and mitochondria,
showing that a common machinery can drive RNA processing
in both organelles (Takenaka et al. 2012). Further, in K. vene-
ficum, several plastid genes from the pre-existing peridinin
789
Jackson et al. . doi:10.1093/molbev/mss270 MBE
Table 1. RNA Editing Events in 14 Tertiary Plastid-Encoded Transcripts.
Gene Sequence RNA Edits Number gDNA Coding cDNA Coding Increase to
Identity to Identity to E. hux Protein
E. hux (%) E. hux (%) Identity After Editing (%)
A – G 26
a G – A 2
psaA (1,227 nt) C – U — 72.8 76.7 3.9
U – C 6
A – G 15
G – A 2
psbC (904 nt) C – U 1 79.0 82.0 3.0
U – C 7
A – G 4
G – A 1
rbcL (1,062 nt) C – U 2 83.3 83.6 0.3
U – C 4
A – G 7

Downloaded from http://mbe.oxfordjournals.org/ at Universidade Federal de Viçosa on July 17, 2013

G – A —
atpG (178 nt) C – U 1 47.5 49.2 1.7
U – C 7
A – G 16
G – A 1
atpI (307 nt) C – U — 61.4 67.3 5.9
U – C 4
A – G 1
G – A —
rpl5 (139 nt) C – U — 77.8 80.0 2.2
U – C —
A – G 9
a G – A 1
rps13 (351 nt) C – U 1 50.0 49.2 0.8
U – C 11
A – G 6
G – A —
rps18 (335 nt) C – U — 26.3 25.0 1.3
U – C 3
A – G 10
G – A 1
rpl33 (181 nt) C – U — 53.8 59.6 5.8
U – C 4
A – G 3
G – A 1
rpl36 (153 nt) C – U 1 62.5 68.8 6.3
U – C —
A – G 4
G – A 1
petD (174 nt) C – U — 36.0 36.0 0.0
U – C 1
A – G 10
a G – A 4
rpoB (580 nt) C – U — 31.3 30.8 0.5
U – C 6
A – G 2
G – A —
rpoC2 (1,268 nt) C – U — 16.2 15.8 0.4
U – C 2
A – G 18
G – A 1
secYa (514 nt) C – U 2 25.1 25.7 0.6
U – C 12
A – G 131 (59.3%)
G – A 15 (6.8%)
Total (7,373 nt) C – U 8 (3.6%) Average 51.6 Average 53.6 Average 2.0
U – C 67 (30.3%)
NOTE.—Protein identity to haptophyte Emiliania huxleyi (E. hux) proteins was determined by three-way alignments with K. veneficum gDNA and cDNA conceptual translations.
New sequences are available in GenBank (accessions JX292160–JX292162, JX292164–JX292175, and JX945981–JX945986) or in the supplementary data, Supplementary Material
online.
a
Genes where internal stop codons are removed by editing.
790
A Tertiary Plastid Gains RNA Editing in Its New Host . doi:10.1093/molbev/mss270 MBE
1 99
rps13 RNA

Protein M
I F R I I Y
A L G V VH
100 198
rps13 RNA

Protein K N V Y R I I
E D A H K T T
199 297

rps13 RNA

Protein N K E Y Y*
D R GH HQ
298 351

rps13 RNA =A
=G
=U
Protein * =C

Downloaded from http://mbe.oxfordjournals.org/ at Universidade Federal de Viçosa on July 17, 2013

FIG. 1. Representative K. veneficum plastid gene, rps13, showing position and types of transitional mRNA editing events. Note multiple edits can act on
a single codon, and an internal stop codon (asterisk) is removed. Pre- and post-RNA-edited nucleotides are indicated above and below, respectively.
Encoded amino acids changes are similarly indicated in protein schematic.

Table 2. Codon Positional A + T Content in 2,443 Plastid-Encoded
Codons.
Codon Position A + T% gDNA A + T% cDNA
1st 56.5 53.0
2nd 64.1 61.8
3rd 84.0 83.1
All 68.2 66.0
plastid have persisted in the host nucleus, and their proteins
are now targeted into the haptophyte tertiary plastid
(Nosenko et al. 2006; Patron et al. 2006; Waller et al. 2006).
Therefore, it is possible that either pre-existing plastid RNA
editing machinery, or that for the existing mitochondrial
system (or even both), could have been recruited to this
new plastid and contributed to its development of editing.
Confirmation of either hypothesis awaits knowledge of the
RNA editing machinery in dinoflagellates.
Why RNA editing has been recruited by this new dinofla-
gellate endosymbiont is a more challenging question and one
that is equally relevant to RNA editing in all systems. It is
tempting to consider that editing evolved to correct muta-
tions after they occurred, and, indeed, editing appears to re-
store K. veneficum plastid sequences to greater identity with
homologs (table 1), as is also the case in plant organelles
(Wakasugi et al. 2001). However, it is unlikely that evolution
or recruitment of the necessary editosome complexity could
occur quickly enough to rescue spontaneous deleterious mu-
tations. Moreover, across eukaryotes, the presence of organ-
elle RNA editing systems negatively correlates with genome
mutation rates, suggesting that mutational pressure does not
explain development of this form of genome complexity
(Lynch et al. 2006). The theory of “constructive neutral evo-
lution” presents an alternative hypothesis, positing that de-
velopment of the machinery for editing precedes its
requirement (Lukes et al. 2011; Gray 2012). In such a scenario,
the gradual development or recruitment of a nondeleterious
editing system would have no immediate function. However,
once established, it might allow subsequent mutations to
occur that are then reversed at the RNA level. In K. veneficum,
one evolutionary consequence of adopting editing is an in-
crease in the A + T nucleotide bias of plastid genes compared
with the transcripts (and also compared with E. huxleyi where
the average A + T content of homologous sequences equals
61%). It is possible that there is some positive advantage for
allowing this increase in A + T content; however, such an
advantage is not obvious. If the K. veneficum plastid did use
pre-existing editing machinery, then this would represent a
good case of neutral evolution, cultivated by the novel genetic
environment of the dinoflagellate host. It is interesting to
note that the K. veneficum plastid genome has apparently
undergone further accelerated change after entering its dino-
flagellate host, with substantial gene loss and genome re-
arrangement, although not yet approaching the level of
plDNA modification seen in peridinin plastids (Howe et al.
2008; Gabrielsen et al. 2011). Although the mechanism caus-
ing these genomic changes is unclear, they provide further
evidence of the peculiar influence that the dinoflagellate en-
vironment can have on its new endosymbiotic partners.
RNA editing in the related tertiary plastid of Karenia miki-
motoi has been simultaneously reported (Dorrell and Howe
2012). Interestingly, both transition and transversion edits are
reported for K. mikimotoi, suggesting possible ongoing devel-
opment of complexity of the plastid editing machinery in this
lineage.
Supplementary Material
Supplementary data are available at Molecular Biology and
Evolution online (http://www.mbe.oxfordjournals.org/).
Acknowledgments
This work was supported by an Australian Research Council
Discovery grant (DP0663590) and by the University of
Melbourne Science Faculty Scholarship to C.J.J.

791
Jackson et al. . doi:10.1093/molbev/mss270
References
Archibald J. 2009. The puzzle of plastid evolution. Curr Biol. 19:
R81–R88.
Barbrook AC, Howe CJ, Kurniawan DP, Tarr SJ. 2010. Organization and
expression of organellar genomes. Phil Trans R Soc B. 365:785–797.
Chateigner-Boutin A-L, Small I. 2010. Plant RNA editing. RNA Biol. 7:
213–219.
Dang Y, Green BR. 2009. Substitutional editing of Heterocapsa triquetra
chloroplast transcripts and a folding model for its divergent chloro-
plast 16S rRNA. Gene 442:73–80.
Dodge JD. 1989. Phylogenetic relationships of dinoflagellates and their
plastids. In: Green JC, Leadbeater BSC, Diver WL, editors. The chro-
mophyte algae: problems and perspectives. Oxford: Clarenden Press.
p. 207–227.
Dorrell G, Howe CJ. 2012. Functional remodeling of RNA processing in
replacement chloroplasts by pathways retained from their prede-
cessors. Proc Natl Acad Sci U S A. 109:18879–18884.
Gabrielsen TM, Minge MA, Espelund M, et al. (11 co-authors). 2011.
Genome evolution of a tertiary dinoflagellate plastid. PLoS One 6:
e19132.
Gould S, Waller R, McFadden G. 2008. Plastid evolution. Annu Rev Plant
Biol. 59:491–517.
Gray MW. 2012. Evolutionary origin of RNA editing. Biochemistry 51:
5235–5242.
Howe CJ, Nisbet RER, Barbrook AC. 2008. The remarkable chloroplast
genome of dinoflagellates. J Exp Bot. 59:1035–1045.
Ishida K-I, Green BR. 2002. Second- and third-hand chloroplasts in dino-
flagellates: phylogeny of oxygen-evolving enhancer 1 (PsbO) protein
reveals replacement of a nuclear-encoded plastid gene by that of a
haptophyte tertiary endosymbiont. Proc Natl Acad Sci U S A. 99:
9294–9299.
Jackson CJ, Gornik SG, Waller RF. 2012. The mitochondrial genome and
transcriptome of the basal dinoflagellate Hematodinium sp.: charac-
ter evolution within the highly derived mitochondrial genomes of
dinoflagellates. Genome Biol Evol. 4:59–72.
Janouškovec J, Horák A, Obornik M, Lukes J, Keeling PJ. 2010. A common
red algal origin of the apicomplexan, dinoflagellate, and heterokont
plastids. Proc Natl Acad Sci U S A. 107:10949–10954.
Knoop V. 2011. When you can’t trust the DNA: RNA editing changes
transcript sequences. Cell Mol Life Sci. 68:567–586.
Lin S, Zhang H, Gray MW. 2008. RNA editing in dinoflagellates and its
implications for the evolutionary history of the editing machinery.
In: Smith HC, editor. RNA and DNA editing: molecular mechanisms
792
MBE
and their integration into biological systems. New York: Wiley Press.
p. 280–309.
Lukes J, Archibald JM, Keeling PJ, Doolittle WF, Gray MW. 2011. How a
neutral evolutionary ratchet can build cellular complexity. IUBMB
Life 63:528–537.
Lynch M, Koskella B, Schaack S. 2006. Mutation pressure and the evo-
lution of organelle genomic architecture. Science 311:1727–1730.
Nosenko T, Lidie KL, Van Dolah FM, Lindquist E, Cheng J-F, Bhattacharya
D. 2006. Chimeric plastid proteome in the Florida “red tide” dino-
flagellate Karenia brevis. Mol Biol Evol. 23:2026–2038.
Patron NJ, Waller RF, Keeling PJ. 2006. A tertiary plastid uses genes from
two endosymbionts. J Mol Biol. 357:1373–1382.
Salone V, Rüdinger M, Polsakiewicz M, Hoffmann B, Groth-Malonek M,
Szurek B, Small I, Knoop V, Lurin C. 2007. A hypothesis on the
identification of the editing enzyme in plant organelles. FEBS Lett.
581:4132–4138.
Shalchian-Tabrizi K, Minge MA, Cavalier-Smith T, Nedreklepp JM,
Klaveness D, Jakobsen KS. 2006. Combined heat shock protein 90
Downloaded from http://mbe.oxfordjournals.org/ at Universidade Federal de Viçosa on July 17, 2013
and ribosomal RNA sequence phylogeny supports multiple replace-
ments of dinoflagellate plastids. J Eukaryot Microbiol. 53:217–224.
Takenaka M, Zehrmann A, Verbitskiy D, Kugelmann M, Härtel B,
Brennicke A. 2012. Multiple organellar RNA editing factor
(MORF) family proteins are required for RNA editing in mitochon-
dria and plastids of plants. Proc Natl Acad Sci U S A. 109:5104–5109.
Takishita K, Ishida K-I, Maruyama T. 2004. Phylogeny of nuclear-encoded
plastid-targeted GAPDH gene supports separate origins for the
peridinin- and the fucoxanthin derivative-containing plastids of
dinoflagellates. Protist 155:447–458.
Tengs T, Dahlberg OJ, Shalchian-Tabrizi K, Klaveness D, Rudi K, Delwiche
CF, Jakobsen KS. 2000. Phylogenetic analyses indicate that the
190 hexanoyloxy-fucoxanthin-containing dinoflagellates have tertiary
plastids of haptophyte origin. Mol Biol Evol. 17:718–729.
Wakasugi T, Tsudzuki T, Sugiura M. 2001. The genomics of land plant
chloroplasts: gene content and alteration of genomic information by
RNA editing. Photosynth Res. 70:107–118.
Waller RF, Jackson CJ. 2009. Dinoflagellate mitochondrial genomes:
stretching the rules of molecular biology. Bioessays 31:237–245.
Waller RF, Slamovits CH, Keeling PJ. 2006. Lateral gene transfer of a
multigene region from cyanobacteria to dinoflagellates resulting in
a novel plastid-targeted fusion protein. Mol Biol Evol. 23:1437–1443.
Wang Y, Morse D. 2006. Rampant polyuridylylation of plastid gene
transcripts in the dinoflagellate Lingulodinium. Nucleic Acids Res.
34:613–619.