Little is known about the cell division cycle of the compartmentalized planctomycetes, especially how membrane-bounded compartments such as those enclosing the nucleoid are transferred into the daughter cells. Results of phase contrast and fluorescence light microscopy combined with electron microscopy of thin-sectioned cells prepared by high pressure freezing/cryosubstitution can be used to derive a model for the cell cycle of Gemmata obscuriglobus. The G. obscuriglobus cell cycle is summarized in Fig. 6. In this model a mother cell forms a small bud with a narrow neck relative to mother cell diameter, and this bud gradually enlarges until it is similar in size to the mother cell, a stage which then lasts for a time considerably longer than other stages of cell division. Both the mother cell and the finally released bud are capable of further cell division – the bud can only start this after a lag period which is much longer than the lag observed for the second mother cell budding. There appears to be a distinct reproductive pole, since division seems to occur repeatedly at the same pole. A new bud is formed at the same pole position of the mother cell where the previous bud was formed, matured and separated from the mother cell.
During cell division, the earliest visible bud stage does not possess DNA via DAPI-staining and a fibrillar nucleoid is not observed in such sectioned buds. Electron microscopy of cells prepared by high pressure freezing/cryosubstitution shows that a nucleoid is initially visible in the bud before a complete nucleoid envelope is formed, suggesting that a nucleoid-containing nuclear body is not transferred intact as a membrane-enclosed structure to the bud. The origin of the new nucleoid envelope found in the bud from both ICM of mother cell and ICM of the bud implies that the G. obscuriglobus nucleoid envelope does not form from mother cell nucleoid envelope, but that it is formed by de novo membrane synthesis as an extension of ICM membrane – this may happen at each cell division. The nucleoid thus appears to be transferred to the bud at first in a naked form, and nucleoid envelope is synthesized around it as extensions of ICM membranes from both mother cell and bud. In Fig. 6 describing one possible model for the cell cycle of G. obscuriglobus, critical stages are those between Fig. 6B and Fig. 6C, where the nucleoid appears in the bud (which initially shows no nucleoid) and then becomes surrounded by two closely apposed membranes to form a new nucleoid envelope for the new bud nucleoid. We have no evidence relating to exactly how the naked nucleoid enters the bud, or whether this involves (as it might) the opening of the mother cell nucleoid envelope membranes to allow passage of a new nucleoid to the bud, so we have not shown such a possible stage in Fig. 6. It should be noted also that Fig. 6D illustrates a model consistent with the appearance of the cell in Fig. 5C, in that the nucleoid (blue) of the bud is now surrounded by two membranes an inner membrane (in light purple) continuous with mother cell ICM (dark purple), and an outer membrane (also in light purple) which shows continuity with the ICM of the bud (in dark purple) at two regions on either side of the bud neck. A possible intermediate stage between Fig. 6C and 6D is suggested by the appearance of the bud in Fig. 4C and 4D and Fig. 5B, where the nucleoid is only partially surrounded by membranes derived from mother ICM and bud ICM. In Fig. 5B, there appear to be multiple regions at which ICM of the bud is continuous with the outer membrane of the nucleoid envelope, so the outer nucleoid envelope membrane at least may not necessarily be formed at a single or only two points of continuity with the bud cell's ICM, as might be implied by the late stage in Fig. 5C and Fig. 6D where the nucleoid is almost completely surrounded by the mature nucleoid envelope consisting of two apposed membranes, one derived from the mother cell's ICM and one derived from the bud cell's ICM.
There thus appears to be an intimate relation between the ICM and nucleoid envelope in G. obscuriglobus. This is consistent with the continuity noted previously between the outer nucleoid envelope membrane and the ICM in cryosubstituted G. obscuriglobus cells [5]. The distribution of the nucleus and nucleoid in G. obscuriglobus is thus not analogous to closed mitosis in some eukaryotes such as yeasts where the nucleus and its envelope is distributed to the daughter cell intact [19], nor to open mitosis of other eukaryotes such as animal cells where the nucleoid envelope breaks down during mitosis and reassembles afterwards [12, 20, 21]. However, there is now evidence that even in open mitosis, newly assembles nuclear membrane actually derives from existing endoplasmic reticulum membrane [22]. In the metazoan Xenopus, the new nucleoid envelope (NE) formation is initiated by endoplasmic reticulum (ER) tubule-end binding and subsequent tethering of the ER network on chromatin. Chromatin was shown to play an active role in reshaping of the ER during NE formation; therefore the process of nuclear membrane formation in G. obscuriglobus, which coincides with the presence of the nucleoid in the buds of G. obscuriglobus, may be analogous in some ways. The occurrence of an early bud without a nucleoid followed by migration of nucleus into the new bud is analogous to the order of nucleoid appearance occurring in budding yeast [23], but presumably without an M phase mitotic segregation of chromosomes into the bud. Overall, the cell division of G. obscuriglobus displays some unique features not known in cells of either prokaryotes or eukaryotes. For example, a nucleoid envelope forms in the daughter cell around the nucleoid (not occurring in other prokaryotes) and there does not seem to be a process of eukaryote-like mitosis in these bacteria. Such a process would be expected to involve in open mitosis a disassembly of mother cell nucleoid envelope and in closed mitosis no stage at which a naked nucleoid (without nucleoid envelope) occurs and in both forms of mitosis, mitotic spindles composed of microtubules. Unlike any form of mitosis we can identify in eukaryotes, although the mother cell in dividing G. obscuriglobus retains enveloped nucleoids, the nucleoid in the bud is initially naked or only enveloped by ICM of the bud, and new nucleoid envelope is then apparently derived from existing intracellular membranes of both mother and daughter cells. The new bud outer nucleoid envelope membrane can easily be conceived to form by a formation of vesicle blebbing of the daughter bud cell ICM membrane since we have evidence for multiple vesicles forming the outer nucleoid envelope of the bud and still connected to the ICM. Continuity of the new inner nucleoid membrane with the mother cell ICM is seen and more difficult to explain mechanistically; the mother cell nucleoid envelope does not appear to be directly involved in formation of the bud nucleoid envelope but it may be that it must open at some point, perhaps the bud neck, in order for the nucleoid to pass to the bud – this stage is not clearly captured in our micrographs however so has not been assumed in our model for this process. In an alternative model, it may also be that the new bud nucleoid envelope membranes form via de novo membrane synthesis but that the ICM membranes act as seed points. The existing data do not discriminate between such models involving extension of existing membrane or formation of new membrane on a framework of existing membranes.
Sectioned G. obscuriglobus budding cells prepared via high-pressure freezing/cryosubstitution display nucleoids which always appear in condensed form, whether in the mother cell or the bud. This contrasts with the normal case in prokaryotic cells such as E. coli where DNA does not condense during cell division. It suggests that either the nucleoid remains condensed during division or is recondensed after a strand unfolding if that occurs during division. It may be most likely that a condensed nucleoid is transferred to the bud since there is a very early stage of budding without DAPI-stainable DNA, and the next stage distinguishable has a condensed nucleoid as seen via TEM of thin-sectioned high pressure frozen/cryosubstituted cells.
In buds with nucleoids and nucleoid envelopes, ribosomes aggregate and are arranged linearly at both sides of the nucleoid envelope, that is, along both the inner and the outer envelope membranes. Such ribosome arrangement along the nucleoid envelope membranes could also be seen in the mother cell and might be a phenomenon of the nucleoid envelope of G. obscuriglobus preserved when prepared by high-pressure freezing. It may form a useful marker of nucleoid envelope in such cells. Such arrangement implies that co-translational protein secretion might occur across nuclear body membranes at some stages e.g. in a newly formed bud, making them analogous to eukaryote ER in some ways.
The only other planctomycete for which the cell division cycle has been described is a freshwater strain once classified as Morphotype IV of the 'Blastocaulis-Planctomyces' group, ICPB 4232, closely related to ATCC35122 and therefore to Pirellula staleyi [24], and in that strain a motile swarmer daughter develops at one pole of a non-motile daughter cell [25]. We have examined negatively stained Gemmata obscuriglobus cells and single cells of mother cell size possess a tuft of flagella, and this view that mother cells before or between budding are motile is also confirmed in phase contrast microscopy of wet mount, where most or all cells are motile. In ICPB 4232, as much as 30 hours were required for the swarmer to initiate a new budding cycle after becoming sessile, while G. obscuriglobus buds also exhibited a lag in bud formation but this lag was only 3–5.5 hours. In ICPB 4232, budding formation and maturation to separation occupied 3 hours under the conditions used, while in G. obscuriglobus budding cell division occupied approximately 12 hours from initial bud formation until separation from the mother cell. In ICPB 4232, there is a 'resting phase' lag of 7–9 hours before a new bud forms, while in G. obscuriglobus there is also a lag in new mother cell budding, of 2–4 hours. Of course such times may vary in any case with different culture media and strain growth rate. The structure of the cell division cycle may be similar in these two planctomycetes, but G. obscuriglobus may not differentiate into a distinct swarmer stage.
The closest analogs to the G. obscuriglobus mode of cell division within the Bacteria are the budding prosthecate bacteria Hyphomicrobium and Hyphomonas. In Hyphomicrobium sp. strain B522, nucleoids are absent in very young buds [26] and this is similar to the situation observed in G. obscuriglobus. In the prosthecate bacterium Hyphomonas, the nucleoid DNA is partitioned to the swarmer cell and transferred to the swarmer cell through a prosthecate hypha via 'pseudovesicles' surrounded by cytoplasmic membrane and containing ribosomes as well as DNA [16]. This does not appear to occur in G. obscuriglobus, but the stage where actual nucleoid transfer occurs has not been captured, perhaps because this is relatively rapid and thus cells displaying it occur in very low numbers in thin sections.
The phylum Planctomycetes to which Gemmata obscuriglobus belongs has been proposed to be a member of the PVC superphylum comprising at least the phyla Verrucomicrobia and Chlamydiae as well as the Planctomycetes [4, 27, 28]. Members of the phylum Chlamydiae display a distinctive life cycle during the infection of eukaryote cells by these pathogens, including a reticulate body stage capable of division by binary fission and an infective elementary body cell stage [29]. The latter displays condensed nucleoids analogous to those of G. obscuriglobus, but this condensation is released in the reticulate dividing stage, unlike the situation in dividing G. obscuriglobus. There is also as yet no evidence of cell compartmentalization in chlamydiae.