Mitochondrial genome of a tertiary endosymbiont retains genes for electron transport proteins

Mitochondria and plastids originated through endosymbiosis, and subsequently became reduced and integrated with the host in similar ways. Plastids spread between lineages through further secondary or even tertiary endosymbioses, but mitochondria appear to have originated once and have not spread between lineages. Mitochondria are also generally lost in secondary and tertiary endosymbionts, with the single exception of the diatom tertiary endosymbiont of dinoflagellates like Kryptoperidinium foliaceum, where both host and endosymbiont are reported to contain mitochondria. Here we describe the first mitochondrial genes from this system: cytochrome c oxidase 1 (cox1), cytochrome oxidase 3 (cox3), and cytochrome b (cob). Phylogenetic analyses demonstrated that all characterized genes were derived from the pennate diatom endosymbiont, and not the host. We also demonstrated that all three genes are expressed, that cox1 contains spliced group II introns, and that cob and cox3 form an operon, all like their diatom relatives. The endosymbiont mitochondria not only retain a genome, but also express their genes, and are therefore likely involved in electron transport. Ultrastructural examination confirmed the endosymbiont mitochondria retain normal tubular cristae. Overall, these data suggest the endosymbiont mitochondria have not reduced at the genomic or functional level.     

Ribosomal RNA Analysis Indicates a Benthic Pennate Diatom Ancestry for the Endosymbionts of the Dinoflagellates Peridinium foliaceum and Peridinium balticum (Pyrrhophyta)

ABSTRACT. The establishment of chloroplasts as cellular organelles in the dinoflagellate, heterokont (stramenopile), haptophyte, and cryptophyte algae is widely accepted to have been the result of secondary endosymbiotic events, that is, the uptake of a photosynthetic eukaryote by a phagotrophic eukaryote. However, the circumstances that promote such associations between two phylogenetically distinct organisms and result in the integration of their genomes to form a single functional photosynthetic cell is unclear. The dinoflagellates Peridinium foliaceum and Peridinium balticum are unusual in that each contains a membrance-bound eukaryotic heterokont endosymbiont. These symbioses have been interpreted, through data derived from ultrastructural and biochemical investigations, to represent an intermediate stage of secondary endosymbiotic chloroplast acquistion. In this study we have examined the phylogenetic origin of the P. foliaceum and P. Balticum heterokont endosymbionts through analaysis of their nuclear small subunit ribosomal RNA genes. Our analyses clearly demonstrate both endosymbionts are pennate diatoms belonging to the family Bacillariaceae. Since members of the Bacillariaceae are usually benthic, living on shallow marine sediments, the manner in which establishment of a symbiosis between a planktonic flagellated dinoflagellate and a botton-dwelling diatom is discussed. In particular, specific environmentally associated life strategy stages of the host and symbiont, coupled with diatom food preferences by the dinoflagellate, may have been vital to the formation of this association.     

Serial replacement of a diatom endosymbiont in the marine dinoflagellate Peridinium quinquecorne (Peridiniales, Dinophyceae)

To infer the phylogeny of both the host and the endosymbiont of Peridinium quinquecorne Abé, the small subunit (SSU) ribosomal DNA (rDNA) from the host and two genes of endosymbiont origin (plastid‐encoded rbcL and nuclear‐encoded SSU rDNA) were determined. The phylogenetic analysis of the host revealed that the marine dinoflagellate P. quinquecorne formed a clade with other diatom‐harbouring dinoflagellates, including Kryptoperidinium foliaceum (Stein) Lindeman, Durinskia baltica (Levander) Carty et Cox and Galeidinium rugatum Tamura et Horiguchi, indicating a single endosymbiotic event for this lineage. Phylogenetic analyses of the endosymbiont in these organisms revealed that the endosymbiont of P. quinquecorne formed a clade with a centric diatom (SSU data indicated it to be closely related to Chaetoceros), whereas the endosymbionts of other three dinoflagellates formed a clade with a pennate diatom. The discrepancy between the host and the endosymbiont phylogenies suggests a secondary replacement of the endosymbiont from a pennate to a centric diatom in P. quinquecorne.     

Fates of Evolutionarily Distinct,Plastid‐type Glyceraldehyde 3‐phosphate Dehydrogenase Genes in Kareniacean Dinoflagellates

The ancestral kareniacean dinoflagellate has undergone tertiary endosymbiosis, in which the original plastid is replaced by a haptophyte endosymbiont. During this plastid replacement, the endosymbiont genes were most likely flowed into the host dinoflagellate genome (endosymbiotic gene transfer or EGT). Such EGT may have generated the redundancy of functionally homologous genes in the host genome—one has resided in the host genome prior to the haptophyte endosymbiosis, while the other transferred from the endosymbiont genome. However, it remains to be well understood how evolutionarily distinct but functionally homologous genes were dealt in the dinoflagellate genomes bearing haptophyte‐derived plastids. To model the gene evolution after EGT in plastid replacement, we here compared the characteristics of the two evolutionally distinct genes encoding plastid‐type glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) in Karenia brevis and K. mikimotoi bearing haptophyte‐derived tertiary plastids: “gapC1h” acquired from the haptophyte endosymbiont and “gapC1p” inherited from the ancestral dinoflagellate. Our experiments consistently and clearly demonstrated that, in the two species examined, the principal plastid‐type GAPDH is encoded by gapC1h rather than gapC1p. We here propose an evolutionary scheme resolving the EGT‐derived redundancy of genes involved in plastid function and maintenance in the nuclear genomes of dinoflagellates that have undergone plastid replacements. Although K. brevis and K. mikimotoi are closely related to each other, the statuses of the two evolutionarily distinct gapC1 genes in the two Karenia species correspond to different steps in the proposed scheme.     

The Phylogenetic Position of Alpha- and Beta-Tubulins from the Chlorarachnion Host and Cercomonas (Cercozoa)

Alpha and beta-tubulin genes from Chlorarachnion and an alpha-tubulin gene from Cercomonas have been characterised. We found the Cercomonas and Chlorarachnion alpha-tubulins to be closely related to one another, confirming the proposed relationship of these genera. In addition, the Chlorarachnion host and Cercomonas also appear to be more distantly related to Heterolobosea. Euglenozoa, chlorophytes. Heterokonts, and alveolates. Chlorarachnion was also found to have two distinctly different types of both alpha- and beta-tubulin. one type being highly-divergent. Chlorarachnion contains a secondary endosymbiont of green algal origin, raising the possibility that one type of Chlorarachnion tubulins comes from the host and the other from the endosymbiont. Probing pulsed field-separated chromosomes showed that the highly-divergent genes are encoded by the host genome, and neither alpha- nor beta-tubulin cDNAs were found to include 5' extensions that might serve as targeting peptides. It appears that Chlorarachnion has distinct and divergent tubulin paralogues that are all derived from the host lineage. One Chlorarachnion beta-tubulin was also found to be a pseudogene, which is still expressed but aberrantly processed. Numerous unspliced introns and deletions resulting from mis-splicing are contained in the mRNAs from this gene.     

Symbiotic origin of a novel actin gene in the cryptophyte Pyrenomonas helgolandii

Cryptophytes are photosynthetic protists that have acquired their plastids through the secondary symbiotic uptake of a red alga. A remarkable feature of cryptophytes is that they maintain a reduced form of the red algal nucleus, the nucleomorph, between the second and third plastid membranes (periplastidial compartment; PC). The nucleomorph is thought to be a transition state in the evolution of secondary plastids, with this genome ultimately being lost when photosynthesis comes under full control of the "host" nucleus (e.g., as in heterokonts, haptophytes, and euglenophytes). Genes presently found in the nucleomorph seem to be restricted to those involved in its own maintenance and to that of the plastid; other genes were lost as the endosymbiont was progressively reduced to its present state. Surprisingly, we found that the cryptophyte Pyrenomonas helgolandii possesses a novel type of actin gene that originated from the nucleomorph genome of the symbiont. Our results demonstrate for the first time that secondary symbionts can contribute genes to the host lineage which are unrelated to plastid function. These genes are akin to the products of gene duplication or lateral transfer and provide a source of evolutionary novelty that can significantly increase the genetic diversity of the host lineage. We postulate that this may be a common phenomenon in algae containing secondary plastids that has yet to be fully appreciated due to a dearth of evolutionary studies of nuclear genes in these taxa.     

High-throughput pyrosequencing of the chloroplast genome of a highly neutral-lipid-producing marine pennate diatom, Fistulifera sp. strain JPCC DA0580

The chloroplast genome of the highly neutral-lipid-producing marine pennate diatom Fistulifera sp. strain JPCC DA0580 was fully sequenced using high-throughput pyrosequencing. The general features and gene content were compared with three other complete diatom chloroplast genomes. The chloroplast genome is 134,918 bp with an inverted repeat of 13,330 bp and is slightly larger than the other diatom chloroplast genomes due to several low gene-density regions lacking similarity to the other diatom chloroplast genomes. Protein-coding genes were nearly identical to those from Phaeodactylum tricornutum. On the other hand, we found unique sequence variations in genes of photosystem II which differ from the consensus in other diatom chloroplasts. Furthermore, five functional unknown ORFs and a putative serine recombinase gene, serC2, are located in the low gene-density regions. SerC2 was also identified in the plasmids of another pennate diatom, Cylindrotheca fusiformis, and in the plastid genome of the diatom endosymbiont of Kryptoperidinium foliaceum. Exogenous plasmids might have been incorporated into the chloroplast genome of Fistulifera sp. by lateral gene transfer. Chloroplast genome sequencing analysis of this novel diatom provides many important insights into diatom evolution.     

A gene in the process of endosymbiotic transfer

Background

The endosymbiotic birth of organelles is accompanied by massive transfer of endosymbiont genes to the eukaryotic host nucleus. In the centric diatom Thalassiosira pseudonana the Psb28 protein is encoded in the plastid genome while a second version is nuclear-encoded and possesses a bipartite N-terminal presequence necessary to target the protein into the diatom complex plastid. Thus it can represent a gene captured during endosymbiotic gene transfer.

Methodology/Principal Findings

To specify the origin of nuclear- and plastid-encoded Psb28 in T. pseudonana we have performed extensive phylogenetic analyses of both mentioned genes. We have also experimentally tested the intracellular location of the nuclear-encoded Psb28 protein (nuPsb28) through transformation of the diatom Phaeodactylum tricornutum with the gene in question fused to EYFP.

Conclusions/Significance

We show here that both versions of the psb28 gene in T. pseudonana are transcribed. We also provide experimental evidence for successful targeting of the nuPsb28 fused with EYFP to the diatom complex plastid. Extensive phylogenetic analyses demonstrate that nucleotide composition of the analyzed genes deeply influences the tree topology and that appropriate methods designed to deal with a compositional bias of the sequences and the long branch attraction artefact (LBA) need to be used to overcome this obstacle. We propose that nuclear psb28 in T. pseudonana is a duplicate of a plastid localized version, and that it has been transferred from its endosymbiont.     

An assessment of vertical inheritance versus endosymbiont transfer of nucleus-encoded genes for mitochondrial proteins following tertiary endosymbiosis in Karlodinium micrum

Most photosynthetic dinoflagellates harbour a red alga-derived secondary plastid. In the dinoflagellate Karlodinium micrum, this plastid was replaced by a subsequent endosymbiosis, resulting in a tertiary plastid derived from a haptophyte. Evolution of endosymbionts entails substantial relocation of endosymbiont genes to the host nucleus: a process called endosymbiotic gene transfer (EGT). In K. micrum, numerous plastid genes from the haptophyte nucleus are found in the host nucleus, providing evidence for EGT in this system. In other cases of endosymbiosis, notably ancient primary endosymbiotic events, EGT has been inferred to contribute to remodeling of other cell functions by expression of proteins in compartments other than the endosymbiont from which they derived. K. micrum provides a more recently derived endosymbiotic system to test for evidence of EGT and gain of function in non-plastid compartments. In this study, we test for gain of haptophyte-derived proteins for mitochondrial function in K. micrum. Using molecular phylogenies we have analysed whether nucleus-encoded mitochondrial proteins were inherited by EGT from the haptophyte endosymbiont, or vertically inherited from the dinoflagellate host lineage. From this dataset we found no evidence of haptophyte-derived mitochondrial genes, and the only cases of non-vertical inheritance were genes derived from lateral gene transfer events.     

Nucleus-to-nucleus gene transfer and protein retargeting into a remnant cytoplasm of cryptophytes and diatoms

The complex plastid of the cryptophyte Guillardia theta and of the diatom Phaeodactylum tricornutum can both be traced back to an engulfed eukaryotic red alga. The eukaryotic origin of these plastids is most obvious in cryptophytes, where the organelle still possesses a remnant nucleus, the nucleomorph. The nucleomorph itself is embedded in the periplastid compartment (PPC), the remnant of the former red algal cytosol. In the cryptophyte and diatom, the complex plastid is surrounded by 4 membranes, the outer one being continuous with the host rough endoplasmatic reticulum. In a recent report, we have shown that a nuclear encoded PPC protein of G. theta expressed in P. tricornutum leads to a localization, recently described as being a "bloblike structure," which can be obtained by mutation of plastid protein-targeting sequences of the diatom itself. Here we present further nucleus-encoded PPC proteins from G. theta, such as the eukaryotic translation elongation factor-1alpha, evidence for their nucleus-to-nucleus gene transfer, and retargeting of the proteins. We also investigated the first nuclear encoded PPC-targeted protein of P. tricornutum (Hsp70) and analyzed it for in vivo localization together with the identified G. theta PPC proteins. This revealed that all localize to the bloblike structures, which we suggest is the highly reduced PPC of P. tricornutum. Furthermore, the described cryptophyte PPC proteins possibly allow the elucidation of the processes by which proteins are involved in different levels of host control over its eukaryotic organelle.     

Chloroplast genomes of the diatoms <Emphasis Type="Italic">Phaeodactylum tricornutum</Emphasis> and <Emphasis Type="Italic">Thalassiosira pseudonana</Emphasis>: comparison with other plastid genomes of the red lineage

The chloroplast genomes of the pennate diatom Phaeodactylum tricornutum and the centric diatom Thalassiosira pseudonana have been completely sequenced and are compared with those of other secondary plastids of the red lineage: the centric diatom Odontella sinensis, the haptophyte Emiliania huxleyi, and the cryptophyte Guillardia theta. All five chromist genomes are compact, with small intergenic regions and no introns. The three diatom genomes are similar in gene content with 127-130 protein-coding genes, and genes for 27 tRNAs, three ribosomal RNAs and two small RNAs (tmRNA and signal recognition particle RNA). All three genomes have open-reading frames corresponding to ORFs148, 355 and 380 of O. sinensis, which have been assigned the names ycf88, ycf89 and ycf90. Gene order is not strictly conserved, but there are a number of conserved gene clusters showing remnants of red algal origin. The acpP, tsf and psb28 genes appear to be on the way from the plastid to the host nucleus, indicating that endosymbiotic gene transfer is a continuing process.     

Nuclear and plastid DNAs from the binucleate dinoflagellates Glenodinium (Peridinium) foliaceum and Peridinium balticum

The binucleate dinoflagellates Glenodinium (Peridinium) foliaceum Stein and Peridinium balticum (Levander) Lemmermann were found to contain two major buoyant density classes of DNA. The heavier peak (1.730 g/cm3) was derived from the "dinokaryotic" nucleus and the lighter peak (1.706 g/cm3) from the "endosymbiont" nucleus and this allowed for the fractionation of G. foliaceum DNA in CsCl/EtBr density gradients. An initial CsCl/Hoechst Dye gradient removed a minor A-T rich satellite species which was identified as plastid DNA with a size of about 100-106 kb. Analysis of the nuclear DNA by agarose gel electrophoresis and renaturation studies showed that the endosymbiont nucleus lacked amplified gene-sized DNA molecules, however, this nucleus did have a comparatively high level of DNA. The total amount of DNA per cell and the relative contributions of the two nuclei appeared to vary between two strains of G. foliaceum (75 pg/cell in CCAP strain and 58 pg in UTEX strain). The only strain of P. balticum examined contained 73 pg cell. These results are discussed in relation to the status and possible functioning of the endosymbiont nucleus and the idea that these dinoflagellates provide model systems with which to study the evolution of plastids.     

Plastid proteome prediction for diatoms and other algae with secondary plastids of the red lineage

The plastids of ecologically and economically important algae from phyla such as stramenopiles, dinoflagellates and cryptophytes were acquired via a secondary endosymbiosis and are surrounded by three or four membranes. Nuclear‐encoded plastid‐localized proteins contain N‐terminal bipartite targeting peptides with the conserved amino acid sequence motif ‘ASAFAP’. Here we identify the plastid proteomes of two diatoms, Thalassiosira pseudonana and Phaeodactylum tricornutum, using a customized prediction tool (ASAFind) that identifies nuclear‐encoded plastid proteins in algae with secondary plastids of the red lineage based on the output of SignalP and the identification of conserved ‘ASAFAP’ motifs and transit peptides. We tested ASAFind against a large reference dataset of diatom proteins with experimentally confirmed subcellular localization and found that the tool accurately identified plastid‐localized proteins with both high sensitivity and high specificity. To identify nucleus‐encoded plastid proteins of T. pseudonana and P. tricornutum we generated optimized sets of gene models for both whole genomes, to increase the percentage of full‐length proteins compared with previous assembly model sets. ASAFind applied to these optimized sets revealed that about 8% of the proteins encoded in their nuclear genomes were predicted to be plastid localized and therefore represent the putative plastid proteomes of these algae.     

THE EVOLUTION OF RNA POLYMERASE II IN RED ALGAE

Cryptophytes are photosynthetic protists that have acquired their plastids through the secondary symbiotic uptake of a red alga. A remarkable feature of cryptophytes is that they maintain a reduced form of the red algal nucleus, the nucleomorph, between the second and third plastid membranes (periplastidial compartment, PC). The nucleomorph is thought to be a transition state in the evolution of secondary plastids with this genome ultimately being lost (e.g., as in heterokonts, haptophytes, euglenophytes) when photosynthesis comes under full control of the "host" nucleus. For this to happen, all genes for plastid function must be transferred from the nucleomorph to the nucleus. In this regard, it is generally assumed that nucleomorph genes with functions unrelated to plastid or PC maintenance are lost. Surprisingly, we show here the existence of a novel type of actin gene in the host nucleus of the cryptophyte, Pyrenomonas helgolandii , that has originated from the nucleomorph genome of the symbiont. Our results demonstrate for the first time that secondary symbionts can contribute genes to the host lineage that are unrelated to plastid function. These genes are akin to the products of gene duplication and provide a source of evolutionary novelty that could significantly increase the genetic diversity of the host lineage. We postulate that this may be a common phenomenon in algae containing secondary plastids that has yet to be fully appreciated due to a dearth of evolutionary studies of nuclear genes in these taxa.     

SYMBIOTIC ORIGIN OF A NOVEL ACTIN GENE IN THE CRYPTOPHYTE,PYRENOMONAS HELGOLANDII

Cryptophytes are photosynthetic protists that have acquired their plastids through the secondary symbiotic uptake of a red alga. A remarkable feature of cryptophytes is that they maintain a reduced form of the red algal nucleus, the nucleomorph, between the second and third plastid membranes (periplastidial compartment, PC). The nucleomorph is thought to be a transition state in the evolution of secondary plastids with this genome ultimately being lost (e.g., as in heterokonts, haptophytes, euglenophytes) when photosynthesis comes under full control of the “host” nucleus. For this to happen, all genes for plastid function must be transferred from the nucleomorph to the nucleus. In this regard, it is generally assumed that nucleomorph genes with functions unrelated to plastid or PC maintenance are lost. Surprisingly, we show here the existence of a novel type of actin gene in the host nucleus of the cryptophyte, Pyrenomonas helgolandii, that has originated from the nucleomorph genome of the symbiont. Our results demonstrate for the first time that secondary symbionts can contribute genes to the host lineage that are unrelated to plastid function. These genes are akin to the products of gene duplication and provide a source of evolutionary novelty that could significantly increase the genetic diversity of the host lineage. We postulate that this may be a common phenomenon in algae containing secondary plastids that has yet to be fully appreciated due to a dearth of evolutionary studies of nuclear genes in these taxa.     

Centrohelida is still searching for a phylogenetic home: analyses of seven Raphidiophrys contractilis genes

By recent advance in evolutionary biology, the majority of eukaryotes are classified into six eukaryotic assemblages called as "supergroups". However, several eukaryotic groups show no clear evolutionary affinity to any of the six supergroups. Centrohelida, one of major heliozoan groups, are such an unresolved lineage. In this study, we newly determined the genes encoding translation elongation factor 2 (EF2), cytosolic heat shock protein 70 (HSP70), and cytosolic heat shock protein 90 (HSP90) from the centroheliozoan Raphidiophrys contractilis. The three Raphidiophrys genes were then combined with previously determined actin, alpha-tubulin, beta-tubulin, and SSU rRNA sequences to phylogenetically analyze the position of Centrohelida in global eukaryotic phylogeny. Although the multi-gene data sets examined in this study are the largest ones including the centroheliozoan sequences, the relationships between Centrohelida and the eukaryotic groups considered were unresolved. Our careful investigation revealed that the phylogenetic estimates were highly sensitive to genes included in the multi-gene alignment. The signal of SSU rRNA and that of alpha-tubulin appeared to conflict with one another: the former strongly prefers a monophyly of Diplomonadida (e.g., Giardia), Parabasalia (e.g., Trichomonas), Heterolobosea (e.g., Naegleria), and Euglenozoa (e.g., Trypanosoma), while the latter unites Diplomonadida, Parabasalia, Metazoa, and Fungi. In addition, EF2 robustly unites Rhodophyta and Viridiplantae, while the remaining genes considered in this study do not positively support the particular relationship. Thus, it is difficult to identify the phylogenetic relatives of Centrohelida in the present study, since strong (and some are conflicting) gene-specific "signals" are predominant in the current multi-gene data. We concluded that larger scale multi-gene phylogenies are necessary to elucidate the origin and evolution of Centrohelida.     

Host origin of plastid solute transporters in the first photosynthetic eukaryotes

Background  

It is generally accepted that a single primary endosymbiosis in the Plantae (red, green (including land plants), and glaucophyte algae) common ancestor gave rise to the ancestral photosynthetic organelle (plastid). Plastid establishment necessitated many steps, including the transfer and activation of endosymbiont genes that were relocated to the nuclear genome of the 'host' followed by import of the encoded proteins into the organelle. These innovations are, however, highly complex and could not have driven the initial formation of the endosymbiosis. We postulate that the re-targeting of existing host solute transporters to the plastid fore-runner was critical for the early success of the primary endosymbiosis, allowing the host to harvest endosymbiont primary production.     

After the primary endosymbiosis: an update on the chromalveolate hypothesis and the origins of algae with Chl <Emphasis Type="Italic">c</Emphasis>

The chromalveolate hypothesis proposed by Cavalier-Smith (J Euk Microbiol 46:347–366, 1999) suggested that all the algae with chlorophyll c (heterokonts, haptophytes, cryptophytes, and dinoflagellates), as well as the ciliates, apicomplexans, oomycetes, and other non-photosynthetic relatives, shared a common ancestor that acquired a chloroplast by secondary endosymbiosis of a red alga. Much of the evidence from plastid and nuclear genomes supports a red algal origin for plastids of the photosynthetic lineages, but the number of secondary endosymbioses and the number of plastid losses have not been resolved. The issue is complicated by the fact that nuclear genomes are mosaics of genes acquired over a very long time period, not only by vertical descent but also by endosymbiotic and horizontal gene transfer. Phylogenomic analysis of the available whole-genome data has suggested major alterations to our view of eukaryotic evolution, and given rise to alternative models. The next few years may see even more changes once a more representative collection of sequenced genomes becomes available.     

The ultrastructural features and division of secondary plastids

Plastids in heterokonts, cryptophytes, haptophytes, dinoflagellates, chlorarachniophytes, euglenoids, and apicomplexan parasites derive from secondary symbiogenesis. These plastids are surrounded by one or two additional membranes covering the plastid-envelope double membranes. Consequently, nuclear-encoded plastid division proteins have to be targeted into the division site through the additional surrounding membranes. Electron microscopic observations suggest that the additional surrounding membranes are severed by mechanisms distinct from those for the division of the plastid envelope. In heterokonts, cryptophytes and haptophytes, the outermost surrounding membrane (epiplastid rough endoplasmic reticulum, EPrER) is studded with cytoplasmic ribosomes and connected to the rER and the outer nuclear envelope. In monoplastidic species belonging to these three groups, the EPrER and the outer nuclear envelope are directly connected to form a sac enclosing the plastid and the nucleus. This nuclear-plastid connection, referred to as the nucleus-plastid consortium (NPC), may be significant to ensure the transmission of the plastids during cell division. The plastid dividing-ring (PD-ring) is a conserved component of the division machinery for both primary and secondary plastids. Also, homologues of the bacterial cell division protein, FtsZ, may be involved in the division of secondary plastids as well as primary plastids, though in secondary plastids they have not yet been localized to the division site. It remains to be examined whether or not dynamin-like proteins and other protein components known to function in the division of primary plastids are used also in secondary plastids. The nearly completed sequencing of the nuclear genome of the diatom Thalassiosira pseudonana will give impetus to molecular and cell biological studies on the division of secondary plastids.