植物叶绿体分裂的分子机制

叶绿体是植物细胞内一种重要的细胞器．它不仅是光合作用的场所,还是其它多种中间代谢的场所．叶绿体起源于蓝细菌,与其原核祖先类似,通过二分裂方式进行增殖．最近的研究表明,叶绿体的分裂装置包含原核起源和真核起源的蛋白质,它们在叶绿体的内膜内侧和外膜外侧协同作用以完成叶绿体的分裂．在过去十几年里,包括丝状温度敏感蛋白Z（FtsZ）、Min系统蛋白、质体分裂蛋白（PDV）和ARC蛋白等在内的多个叶绿体分裂相关组分被分离鉴定．本文简要介绍了叶绿体分裂装置各成员的发现、叶绿体被膜的收缩和叶绿体分裂位点的选择机制．另外,植物发育过程中叶绿体分裂可能受到细胞的控制,但目前对细胞如何调控叶绿体分裂知之甚少．本文对该领域的最新研究进展也进行了综述．     

Expression of the nucleus-encoded chloroplast division genes and proteins regulated by the algal cell cycle

Chloroplasts have evolved from a cyanobacterial endosymbiont and their continuity has been maintained by chloroplast division, which is performed by the constriction of a ring-like division complex at the division site. It is believed that the synchronization of the endosymbiotic and host cell division events was a critical step in establishing a permanent endosymbiotic relationship, such as is commonly seen in existing algae. In the majority of algal species, chloroplasts divide once per specific period of the host cell division cycle. In order to understand both the regulation of the timing of chloroplast division in algal cells and how the system evolved, we examined the expression of chloroplast division genes and proteins in the cell cycle of algae containing chloroplasts of cyanobacterial primary endosymbiotic origin (glaucophyte, red, green, and streptophyte algae). The results show that the nucleus-encoded chloroplast division genes and proteins of both cyanobacterial and eukaryotic host origin are expressed specifically during the S phase, except for FtsZ in one graucophyte alga. In this glaucophyte alga, FtsZ is persistently expressed throughout the cell cycle, whereas the expression of the nucleus-encoded MinD and MinE as well as FtsZ ring formation are regulated by the phases of the cell cycle. In contrast to the nucleus-encoded division genes, it has been shown that the expression of chloroplast-encoded division genes is not regulated by the host cell cycle. The endosymbiotic gene transfer of minE and minD from the chloroplast to the nuclear genome occurred independently on multiple occasions in distinct lineages, whereas the expression of nucleus-encoded MIND and MINE is regulated by the cell cycle in all lineages examined in this study. These results suggest that the timing of chloroplast division in algal cell cycle is restricted by the cell cycle-regulated expression of some but not all of the chloroplast division genes. In addition, it is suggested that the regulation of each division-related gene was established shortly after the endosymbiotic gene transfer, and this event occurred multiple times independently in distinct genes and in distinct lineages.     

Plastid division in an evolutionary context

Plastids are derived from free-living cyanobacteria that were engulfed by eukaryotic host cells through the process of endosymbiosis and, like their cyanobacterial ancestors, divide by binary fission. Over the last decade the continued identification and functional analysis of plastid division components, coupled with ever-increasing genomic resources, have yielded insights into the origins and evolution of the plastid division mechanism in higher plants. Here we review the current understanding of the evolution of the chloroplast division proteins and present a model of how the machinery has developed to execute plastid division in Arabidopsis.     

The evolution of the regulatory mechanism of chloroplast division

Chloroplasts arose from a cyanobacterial endosymbiont and multiply by division, reminiscent of their free-living ancestor. However, chloroplasts can not divide by themselves, and the division is performed and controlled by proteins that are encoded by the host nucleus. The continuity of chloroplasts was originally established by synchronization of endosymbiotic cell division with host cell division, as seen in existent algae. In contrast, land plant cells contain multiple chloroplasts, the division of which is not synchronized, even in the same cell. Land plants have evolved cell and chloroplast differentiation systems in which the size and number of chloroplasts (or other types of plastids) change along with their respective cellular function by changes in the division rate. We recently reported that PLASTID DIVISION (PDV) proteins, land-plant specific components of the chloroplast division apparatus, determined the rate of chloroplast division. The level of PDV protein is regulated by the cell differentiation program based on cytokinin, and the increase or decrease of the PDV level gives rise to an increase or decrease in the chloroplast division rate. Thus, the integration of PDV proteins into the chloroplast division machinery enabled land plant cells to change chloroplast size and number in accord with the fate of cell differentiation.Key words: chloroplast division, cell cycle, cell differentiation, cytokinin, endosymbiosis, evolution     

Plastid division coordination across a double-membraned structure

Chloroplasts still retain components of the bacterial cell division machinery and research over the past decade has led to an understanding of how these stromal division proteins assemble and function as a complex chloroplast division machinery. However, during evolution plant chloroplasts have acquired a number of cytosolic division proteins, indicating that unlike the cyanobacterial ancestors of plastids, chloroplast division in higher plants require a second division machinery located on the chloroplast outer envelope membrane. Here we review the current understanding of the stromal and cytosolic plastid division machineries and speculate how two protein machineries coordinate their activities across a double-membraned structure.     

Plastid division: evolution, mechanism and complexity

BACKGROUND: The continuity of chloroplasts is maintained by division of pre-existing chloroplasts. Chloroplasts originated as bacterial endosymbionts; however, the majority of bacterial division factors are absent from chloroplasts and the eukaryotic host has added several new components. For example, the ftsZ gene has been duplicated and modified, and the Min system has retained MinE and MinD but lost MinC, acquiring at least one new component ARC3. Further, the mechanism has evolved to include two members of the dynamin protein family, ARC5 and FZL, and plastid-dividing (PD) rings were most probably added by the eukaryotic host. SCOPE: Deciphering how the division of plastids is coordinated and controlled by nuclear-encoded factors is key to our understanding of this important biological process. Through a number of molecular-genetic and biochemical approaches, it is evident that FtsZ initiates plastid division where the coordinated action of MinD and MinE ensures correct FtsZ (Z)-ring placement. Although the classical FtsZ antagonist MinC does not exist in plants, ARC3 may fulfil this role. Together with other prokaryotic-derived proteins such as ARC6 and GC1 and key eukaryotic-derived proteins such as ARC5 and FZL, these proteins make up a sophisticated division machinery. The regulation of plastid division in a cellular context is largely unknown; however, recent microarray data shed light on this. Here the current understanding of the mechanism of chloroplast division in higher plants is reviewed with an emphasis on how recent findings are beginning to shape our understanding of the function and evolution of the components. CONCLUSIONS: Extrapolation from the mechanism of bacterial cell division provides valuable clues as to how the chloroplast division process is achieved in plant cells. However, it is becoming increasingly clear that the highly regulated mechanism of plastid division within the host cell has led to the evolution of features unique to the plastid division process.     

An Arabidopsis homolog of the bacterial cell division inhibitor SulA is involved in plastid division

Plastids have evolved from an endosymbiosis between a cyanobacterial symbiont and a eukaryotic host cell. Their division is mediated both by proteins of the host cell and conserved bacterial division proteins. Here, we identified a new component of the plastid division machinery, Arabidopsis thaliana SulA. Disruption of its cyanobacterial homolog (SSulA) in Synechocystis and overexpression of an AtSulA-green fluorescent protein fusion in Arabidopsis demonstrate that these genes are involved in cell and plastid division, respectively. Overexpression of AtSulA inhibits plastid division in planta but rescues plastid division defects caused by overexpression of AtFtsZ1-1 and AtFtsZ2-1, demonstrating that its role in plastid division may involve an interaction with AtFtsZ1-1 and AtFtsZ2-1.     

Plastid division: across time and space

Plastid division is executed by the coordinated action of at least two molecular machineries--an internal machinery situated on the stromal side of the inner envelope membrane that was contributed by the cyanobacterial endosymbiont from which plastids evolved, and an external machinery situated on the cytosolic side of the outer envelope membrane that was contributed by the host. Here we review progress in defining the components of the plastid division complex and understanding the mechanisms of envelope constriction and division-site placement in plants. We also highlight recent work identifying the first molecular linkage between the internal and external division machineries, shedding light on how their mid-plastid positioning is coordinated across the envelope membranes. Little is known about the mechanisms that regulate plastid division in plant cells, but recent studies have begun to hint at potential mechanisms.     

Protein import into cyanelles and complex chloroplasts

Higher-plant, green and red algal chloroplasts are surrounded by a double membrane envelope. The glaucocystophyte plastid (cyanelle) has retained a prokaryotic cell wall between the two envelope membranes. The complex chloroplasts of Euglena and dinoflagellates are surrounded by three membranes while the complex chloroplasts of chlorarachniophytes, cryptomonads, brown algae, diatoms and other chromophytes, are surrounded by 4 membranes. The peptidoglycan layer of the cyanelle envelope and the additional membranes of complex chloroplasts provide barriers to chloroplast protein import not present in the simpler double membrane chloroplast envelope. Analysis of presequence structure and in vitro import experiments indicate that proteins are imported directly from the cytoplasm across the two envelope membranes and peptidoglycan layer into cyanelles. Protein import into complex chloroplasts is however fundamentally different. Analysis of presequence structure and in vitro import into microsomal membranes has shown that translocation into the ER is the first step for protein import into complex chloroplasts enclosed by three or four membranes. In vivo pulse chase experiments and immunoelectronmicroscopy have shown that in Euglena, proteins are transported from the ER to the Golgi apparatus prior to import across the three chloroplast membranes. Ultrastructural studies and the presence of ribosomes on the outermost of the four envelope membranes suggests protein import into 4 membrane-bounded complex chloroplasts is directly from the ER like outermost membrane into the chloroplast. The fundamental difference in import mechanisms, post-translational direct chloroplast import or co-translational translocation into the ER prior to chloroplast import, appears to reflect the evolutionary origin of the different chloroplast types. Chloroplasts with a two-membrane envelope are thought to have evolved through the primary endosymbiotic association between a eukaryotic host and a photosynthetic prokaryote while complex chloroplasts are believed to have evolved through a secondary endosymbiotic association between a heterotrophic or possibly phototrophic eukaryotic host and a photosynthetic eukaryote.     

Comparative and evolutionary aspects of cyanobacteria and plant plastid division study

The review summarizes the current understanding of cyanobacterial cell division mechanisms in comparison with those of eukaryotic plastids. It also sheds light on the present knowledge of the nature of evolutionary transformations of the cyanobacterial cell division apparatus that could have occurred during the establishment of modern plastid division complex. Peculiar properties of cyanobacterial cell division process are discussed as well as the features of primary and secondary plastid replication.     

Origin and evolution of the chloroplast division machinery

Chloroplasts were originally established in eukaryotes by the endosymbiosis of a cyanobacterium; they then spread through diversification of the eukaryotic hosts and subsequent engulfment of eukaryotic algae by previously nonphotosynthetic eukaryotes. The continuity of chloroplasts is maintained by division of preexisting chloroplasts. Like their ancestors, chloroplasts use a bacterial division system based on the FtsZ ring and some associated factors, all of which are now encoded in the host nuclear genome. The majority of bacterial division factors are absent from chloroplasts and several new factors have been added by the eukaryotic host. For example, the ftsZ gene has been duplicated and modified, plastid-dividing (PD) rings were most likely added by the eukaryotic host, and a member of the dynamin family of proteins evolved to regulate chloroplast division. The identification of several additional proteins involved in the division process, along with data from diverse lineages of organisms, our current knowledge of mitochondrial division, and the mining of genomic sequence data have enabled us to begin to understand the universality and evolution of the division system. The principal features of the chloroplast division system thus far identified are conserved across several lineages, including those with secondary chloroplasts, and may reflect primeval features of mitochondrial division. Shin-ya Miyagishima is the recipient of the Botanical Society Award for Young Scientists, 2004.     

Crystal structure of a conserved domain in the intermembrane space region of the plastid division protein ARC6

The chloroplast division machinery is composed of numerous proteins that assemble as a large complex to divide double‐membraned chloroplasts through binary fission. A key mediator of division‐complex formation is ARC6, a chloroplast inner envelope protein and evolutionary descendant of the cyanobacterial cell division protein Ftn2. ARC6 connects stromal and cytosolic contractile rings across the two membranes through interaction with an outer envelope protein within the intermembrane space (IMS). The ARC6 IMS region bears a structurally uncharacterized domain of unknown function, DUF4101, that is highly conserved among ARC6 and Ftn2 proteins. Here we report the crystal structure of this domain from Arabidopsis thaliana ARC6. The domain forms an α/β barrel open towards the outer envelope membrane but closed towards the inner envelope membrane. These findings provide new clues into how ARC6 and its homologs contribute to chloroplast and cyanobacterial cell division.     

Complex origins of chloroplast membranes with photosynthetic machineries: multiple transfers of genes from divergent organisms at different times or a single endosymbiotic event?

The paradigm “cyanobacterial origin of chloroplasts” is currently viewed as an established fact. However, we may have to re-consider the origin of chloroplast membranes, because membranes are not replicated by their own. It is the genes for lipid biosynthetic enzymes that are inherited. In the current understandings, these enzymes became encoded by the nuclear genome as a result of endosymbiotic gene transfer from the endosymbiont. However, we previously showed that many enzymes involved in the synthesis of chloroplast peptidoglycan and glycolipids did not originate from cyanobacteria. Here I present results of comprehensive phylogenetic analysis of chloroplast enzymes involved in fatty acid and lipid biosynthesis, as well as additional chloroplast components related to photosynthesis and gene expression. Four types of phylogenetic relationship between chloroplast enzymes (encoded by the chloroplast and nuclear genomes) and cyanobacterial counterparts were found: type 1, chloroplast enzymes diverged from inside of cyanobacterial clade; type 2, chloroplast and cyanobacterial enzymes are sister groups; type 3, chloroplast enzymes originated from homologs of bacteria other than cyanobacteria; type 4, chloroplast enzymes diverged from eukaryotic homologs. Estimation of evolutionary distances suggested that the acquisition times of chloroplast enzymes were diverse, indicating that multiple gene transfers accounted for the chloroplast enzymes analyzed. Based on the results, I try to relax the tight logic of the endosymbiotic origin of chloroplasts involving a single endosymbiotic event by proposing alternative hypotheses. The hypothesis of host-directed chloroplast formation proposes that glycolipid synthesis ability had been acquired by the eukaryotic host before the acquisition of chloroplast ribosomes. Chloroplast membrane system could have been provided by the host, whereas cyanobacteria contributed to the genes for the genetic and photosynthesis systems, at various times, either before or after the formation of chloroplast membranes. The origin(s) of chloroplasts seems to be more complicated than the single event of primary endosymbiosis.

     

The function and diversity of plastid protein import pathways: a multilane GTPase highway into plastids

The photosynthetic chloroplast is the hallmark organelle of green plants. During the endosymbiotic evolution of chloroplasts, the vast majority of genes from the original cyanobacterial endosymbiont were transferred to the host cell nucleus. Chloroplast biogenesis therefore requires the import of nucleus-encoded proteins from their site of synthesis in the cytosol. The majority of proteins are imported by the activity of Toc and Tic complexes located within the chloroplast envelope. In addition to chloroplasts, plants have evolved additional, non-photosynthetic plastid types that are essential components of all cells. Recent studies indicate that the biogenesis of various plastid types relies on distinct but homologous Toc-Tic import pathways that have specialized in the import of specific classes of substrates. These different import pathways appear to be necessary to balance the essential physiological role of plastids in cellular metabolism with the demands of cellular differentiation and plant development.     

Dynamics of the Synechococcus elongatus cytoskeletal GTPase FtsZ yields mechanistic and evolutionary insight into cyanobacterial and chloroplast FtsZs

The division of cyanobacteria and their chloroplast descendants is orchestrated by filamenting temperature-sensitive Z (FtsZ), a cytoskeletal GTPase that polymerizes into protofilaments that form a “Z ring” at the division site. The Z ring has both a scaffolding function for division-complex assembly and a GTPase-dependent contractile function that drives cell or organelle constriction. A single FtsZ performs these functions in bacteria, whereas in chloroplasts, they are performed by two copolymerizing FtsZs, called AtFtsZ2 and AtFtsZ1 in Arabidopsis thaliana, which promote protofilament stability and dynamics, respectively. To probe the differences between cyanobacterial and chloroplast FtsZs, we used light scattering to characterize the in vitro protofilament dynamics of FtsZ from the cyanobacterium Synechococcus elongatus PCC 7942 (SeFtsZ) and investigate how coassembly of AtFtsZ2 or AtFtsZ1 with SeFtsZ influences overall dynamics. SeFtsZ protofilaments assembled rapidly and began disassembling before GTP depletion, whereas AtFtsZ2 protofilaments were far more stable, persisting beyond GTP depletion. Coassembled SeFtsZ–AtFtsZ2 protofilaments began disassembling before GTP depletion, similar to SeFtsZ. In contrast, AtFtsZ1 did not alter disassembly onset when coassembled with SeFtsZ, but fluorescence recovery after photobleaching analysis showed it increased the turnover of SeFtsZ subunits from SeFtsZ–AtFtsZ1 protofilaments, mirroring its effect upon coassembly with AtFtsZ2. Comparisons of our findings with previous work revealed consistent differences between cyanobacterial and chloroplast FtsZ dynamics and suggest that the scaffolding and dynamics-promoting functions were partially separated during evolution of two chloroplast FtsZs from their cyanobacterial predecessor. They also suggest that chloroplasts may have evolved a mechanism distinct from that in cyanobacteria for promoting FtsZ protofilament dynamics.     

Chloroplast DNA Replication Is Regulated by the Redox State Independently of Chloroplast Division in Chlamydomonas reinhardtii

Chloroplasts arose from a cyanobacterial endosymbiont and multiply by division. In algal cells, chloroplast division is regulated by the cell cycle so as to occur only once, in the S phase. Chloroplasts possess multiple copies of their own genome that must be replicated during chloroplast proliferation. In order to examine how chloroplast DNA replication is regulated in the green alga Chlamydomonas reinhardtii, we first asked whether it is regulated by the cell cycle, as is the case for chloroplast division. Chloroplast DNA is replicated in the light and not the dark phase, independent of the cell cycle or the timing of chloroplast division in photoautotrophic culture. Inhibition of photosynthetic electron transfer blocked chloroplast DNA replication. However, chloroplast DNA was replicated when the cells were grown heterotrophically in the dark, raising the possibility that chloroplast DNA replication is coupled with the reducing power supplied by photosynthesis or the uptake of acetate. When dimethylthiourea, a reactive oxygen species scavenger, was added to the photoautotrophic culture, chloroplast DNA was replicated even in the dark. In contrast, when methylviologen, a reactive oxygen species inducer, was added, chloroplast DNA was not replicated in the light. Moreover, the chloroplast DNA replication activity in both the isolated chloroplasts and nucleoids was increased by dithiothreitol, while it was repressed by diamide, a specific thiol-oxidizing reagent. These results suggest that chloroplast DNA replication is regulated by the redox state that is sensed by the nucleoids and that the disulfide bonds in nucleoid-associated proteins are involved in this regulatory activity.Chloroplasts are semiautonomous organelles that possess their own genome, which is complexed with proteins to form nucleoids and also certain machinery needed for protein synthesis, as is the case in prokaryotes. It is generally accepted that chloroplasts arose from a bacterial endosymbiont closely related to the currently extant cyanobacteria (Archibald, 2009; Keeling, 2010). In a manner reminiscent of their free-living ancestor, chloroplasts proliferate by the division of preexisting organelles that are coupled to the duplication and segregation of the nucleoids (Kuroiwa, 1991) and have retained the bulk of their bacterial biochemistry. However, chloroplasts have subsequently been substantially remodeled by the host cell so as to function as complementary organelles within the eukaryotic host cell (Rodríguez-Ezpeleta and Philippe, 2006; Archibald, 2009; Keeling, 2010). For example, most of the genes that were once in the original endosymbiont genome have been either lost or transferred into the host nuclear genome. As a result, the size of the chloroplast genome has been reduced to less than one-tenth that of the free-living cyanobacterial genome. Thus, the bulk of the chloroplast proteome consists of nucleus-encoded proteins that are translated on cytoplasmic ribosomes and translocated into chloroplasts. In addition, chloroplast division ultimately came to be a process tightly regulated by the host cell, which ensured permanent inheritance of the chloroplasts during the course of cell division and from generation to generation (Rodríguez-Ezpeleta and Philippe, 2006; Archibald, 2009; Keeling, 2010).Chloroplast division is performed by constriction of the ring structures at the division site, encompassing both the inside and the outside of the two envelopes (Yang et al., 2008; Maple and Møller, 2010; Miyagishima, 2011; Pyke, 2013). One part of the division machinery is derived from the cyanobacterial cytokinetic machinery that is based on the FtsZ protein. In contrast, other parts of the division machinery involve proteins specific to eukaryotes, including one member of the dynamin family. The majority of algae (both unicellular and multicellular), which diverged early within the Plantae, have just one or at most only a few chloroplasts per cell. In algae, the chloroplast divides once per cell cycle before the host cell completes cytokinesis (Suzuki et al., 1994; Miyagishima et al., 2012). In contrast, land plants and certain algal species contain dozens of chloroplasts per cell that divide nonsynchronously, even within the same cell (Boffey and Lloyd, 1988). Because land plants evolved from algae, there is likely to have been a linkage between the cell cycle and chloroplast division in their algal ancestor that was subsequently lost during land plant evolution. Our recent study showed that the timing of chloroplast division in algae is restricted to the S phase by S phase-specific formation of the chloroplast division machinery, which is based on the cell cycle-regulated expression of the components of the chloroplast division machinery (Miyagishima et al., 2012).Because chloroplasts possess their own genome, chloroplast DNA must be duplicated so that each daughter chloroplast inherits the required DNA after division. However, it is still unclear how the replication of chloroplast DNA is regulated and whether the replication is coupled with the timing of chloroplast division, even though certain studies have addressed this issue, as described below.Bacteria such as Escherichia coli and Bacillus subtilis possess a single circular chromosome. In these bacteria, the process of DNA replication is tightly coupled with cell division (Boye et al., 2000; Zakrzewska-Czerwińska et al., 2007), in which the initiation of replication is regulated such that it occurs only once per cell division cycle (Boye et al., 2000). In contrast, cyanobacteria contain multiple copies of their DNA (e.g. three to five copies in Synechococcus elongatus PCC 7942; Mann and Carr, 1974; Griese et al., 2011). In some obligate photoautotrophic cyanobacterial species, replication is initiated only when light is available (Binder and Chisholm, 1990; Mori et al., 1996; Watanabe et al., 2012). Replication is initiated asynchronously among the multiple copies of the DNA. Although the regulation of the initiation of DNA replication is less stringent than that in E. coli and B. subtilis, as described above, a recent study using S. elongatus PCC 7942 showed that this replication peaks prior to cell division, as in other bacteria.Chloroplasts also contain multiple copies of DNA (approximately 1,000 copies; Boffey and Leech, 1982; Miyamura et al., 1986; Baumgartner et al., 1989; Oldenburg and Bendich, 2004; Oldenburg et al., 2006; Shaver et al., 2008). In algae, chloroplast DNA is replicated in a manner that keeps pace with chloroplast and cell division in order to maintain the proper DNA content per chloroplast (i.e. per cell). In contrast, in land plants, the copy number of DNA in each chloroplast (plastid) changes during the course of development and differentiation, although contradictory results were reported about leaf development (Lamppa and Bendich, 1979; Boffey and Leech, 1982; Hashimoto and Possingham, 1989; Kuroiwa, 1991; Rowan and Bendich, 2009; Matsushima et al., 2011). Previous studies that synchronized the algal cell cycle by means of a 24-h light/dark cycle showed that chloroplast DNA is replicated only during the G1 phase, after which it is separated into daughter chloroplasts during the S phase by chloroplast division, implying that chloroplast DNA replication and division are temporally separated (Chiang and Sueoka, 1967; Grant et al., 1978; Suzuki et al., 1994). However, under these experimental conditions, G1 cells grow and the chloroplast DNA level increases during the light period. Cells enter into the S phase, chloroplast DNA replication ceases, and the chloroplasts divide at the beginning of the dark period. Thus, it is still unclear whether chloroplast DNA replication is directly controlled by the cell cycle, as is the case in chloroplast division, or chloroplast DNA replication occurs merely when light energy is available.We addressed this issue using a synchronous culture as well as a heterotrophic culture of the mixotrophic green alga Chlamydomonas reinhardtii. The results show that chloroplast DNA replication occurs independently of either the cell cycle or the timing of chloroplast division. Instead, it is shown that chloroplast DNA replication occurs when light is available in photoautotrophic culture and even under darkness in heterotrophic culture. Further experimental results suggest that chloroplast DNA replication is regulated by the redox state in the cell, which is sensed by the chloroplast nucleoids.     

ARC3 is a stromal Z-ring accessory protein essential for plastid division

In plants, chloroplast division is an integral part of development, and these vital organelles arise by binary fission from pre-existing cytosolic plastids. Chloroplasts arose by endosymbiosis and although they have retained elements of the bacterial cell division machinery to execute plastid division, they have evolved to require two functionally distinct forms of the FtsZ protein and have lost elements of the Min machinery required for Z-ring placement. Here, we analyse the plastid division component accumulation and replication of chloroplasts 3 (ARC3) and show that ARC3 forms part of the stromal plastid division machinery. ARC3 interacts specifically with AtFtsZ1, acting as a Z-ring accessory protein and defining a unique function for this family of FtsZ proteins. ARC3 is involved in division site placement, suggesting that it might functionally replace MinC, representing an important advance in our understanding of the mechanism of chloroplast division and the evolution of the chloroplast division machinery.