Microbiologists study microbes with microscopes because they are not visible to the naked eye. Unless they are. This was exactly the case for "Candidatus Thiomargarita magnifica", described in a recent BioRχiv preprint by Vollard et al. (2022): "A centimeter-long bacterium with DNA compart..." (more about the second part of the title below.)
A centimeter-long bacterium with DNA... You don't necessarily think "mon dieu, des microbes!" when you see tiny whitish threads the length and diameter of eyelashes on Rhizophora mangle leaves while diving among the mangrove shallows by the coast of Guadeloupe island, Lesser Antilles (which belongs to France.) Or when you see them "portrayed" next to a dime, as on the frontispiece. Gatherings of nematodes come to mind, or colonies of ciliates, or of marine fungi. As a marine biologist you may eventually think of filamentous cyanobacteria like Anabaena or Oscillatoria, but these do not appear whitish when you see them floating in the water. There are in fact visible filamentous bacteria that appear whitish ─ they contain sulfur granules aplenty, thus whitish ─ and eventually form sessile colonies like Thiolava veneris ("Venus' hair"), a member of the larger Thioploca clade of filamentous sulfur‑oxidizing bacteria (see here and here in STC.) And finally, there are these beautiful chains of "sulfur pearls", strung giant globular cells of Thiomargarita species that reach up to 0.75 mm in diameter (Merry wrote about them here.) So, who are these tiny whitish threads you see in Figure 1, some of which reach 1 cm or more in length?
16S sRNA phylotyping revealed that "Ca. Thiomargarita magnifica" is a member of the genus Thiomargarita, with "Ca. T. nelsonii" being its closest relative (both with the "Candidatus" prefix, formally, as they are yet uncultivated.) The researchers were in for no small surprise when they looked at these filaments ─ now under the microscope ─ stained with the common fluorescent membrane dye FM 1-43x (Figure 2). Unlike the bacterial filaments used as controls, the T. magnifica filaments showed no septation at all (note the different scales in panels A─D in Figure 2.) A single Thiomargarita magnifica cell as in panel D with a length of ~2 mm and a width of ~50 µm is, well, simply magnificent. The authors saw even longer filaments as those shown in Figure 1, the longest one was about 2 cm long. Let that sink in.
You see in Figure 1 single cells attached by one end to their substrate, the Rhizophora mangle leaves, while the apical end shows rather regular constrictions (seen particular well in Figure 5D). Volland et al. observed that the terminal buds are eventually dropped. They conclude from length measurements of such terminal buds and the smallest substrate-attached filaments, which were in good agreement, that T. magnifica has a dimorphic life cycle: a sessile mother cell and dispersive daughter cells. They explicitly refer to the dimorphic life cycle of Caulobacter ─ the sessile stalk cell produces swarmer progeny cells ─ but clear morphological distinctions are not evident in T. magnifica mothers and daughters and, for example, it is not known if the daughters have a preferred cell pole for attachment to the substrate.
Vollard et al. probed cells with a fluorescent Thiomargarita-specific probe and DAPI to localize DNA. The signals they obtained were evenly distributed as "dots" along the entire cell length, and exclusively localized in a narrow strip ─ actually more a layer since cells are three-dimensional ─ just below the cell wall (Figure 3). More than 70% of the cell volume is occupied by an enormous vacuole that extends through the entire length of the cell, including the apical segment. This appears to be a "family trait" because the globular Thiomargaritas also have this huge vacuole, which restricts the space for the bacterial cytoplasm to a narrow layer just below the cell wall.
They assumed that the strongly fluorescent "dots" in Figure 3 are nucleoids that contain one or more genomes, depending on the signal intensity. By counting the dots and normalizing their size, they were able to calculate the approximate genome copy number in a cell: on average 37,000 (±8,000) genome copies per millimeter of filament. That's what I call polyploidy! A Gedankenexperiment: if you stuff such a millimeter of T. magnifica, just the cytoplasmic cytoplasmic layer of course, with E. coli cells with one chromosome each you would end up with approximately 25,000 genome copies. The "action radius" of a genome for its transcription and translation products is thus largely comparable in both cases.
The incipient constriction of the apical cell segment is clearly visible in Figure 3, but this constriction spans a distance of ~50 nm across, and it is therefore clear that this cannot be a conventional cell division of "normal" rod‑shaped bacteria like E. coli or B. subtilis (unfortunately the authors give no explanation for the cloudy staining of the septal region.) In the nearly completely sequenced genome (~12 Mb) of T. magnifica, Volland et al. found the dcw gene cluster, including ftsZ, conserved but the ftsA, zipA, and ftsE─ftsX genes encoding divisome proteins required for tethering the FtsZ ring to the septal membrane lacking. Also missing were the "late" divisome genes ftsI and ftsW, ftsQ, ftsL, ftsB, and ftsK. In contrast, the set of genes encoding cell elongation proteins was "complete", while three were even duplicated "in situ", that is, the duplication located directly adjacent to the original (a replication slippage?). Apparently, cell division by budding in T. magnifica follows a so far unknown route, and particularly the role of FtsZ in this division process would be well worth studying.*)
...with DNA compartmentalized in membrane-bound organelles A closer inspection of T. magnifica's cytoplasm by CLEM (correlative light‑electron microscopy) revealed that its genetic material is concentrated in compact nucleoids that align along the vacuolar membrane as single or small groups of "blobs" with more irregular spacing over the entire cell length (Figure 4, A─E). In Figure 5 D+G, it is clearly visible that these blobs also contain many small electron dense granules, confirmed to be ribosomes by FISH with probes targeting Thiomargarita rRNA sequences. The blobs virtually bulge the membrane of the vacuole, and it almost looks as if they were safely put in a kangaroo pouch. This is highly remarkable since bacterial nucleoids are usually confined to much smaller cell volumes where they appear to occupy a considerable percentage of the cytoplasm and are embedded in a "ribosome soup" (see here in STC.) Yet, as Xiang et al. (2021) have shown, the bacterial cytoplasm is a "poor solvent" for nucleoids and the notion of Volland et al. (2022) that they are "...spread throughout the cytoplasm, as is common in bacteria" is a misconception.
My skepticism about the authors' claim that the genome of T. magnifica is localized in membrane-bound compartments ─ which they call "pepins" ─ stems from the fact that their technically excellent EM images do not show a continuous membrane, at least not in one focal plane (Figure 4, F+G). I vividly recall that such a compartmentalized, that is, membrane-enclosed arrangement of the nucleoid was proposed for Gemmata obscuriglobus (Planctomycetes) (see here in STC.) It later turned out that the inner membrane of Gemmata is intricately invaginated but does not enclose the nucleoid (see here in STC.)
The Vollard et al. paper is currently a preprint, and it is difficult to gauge which of their claims regarding the "pepins" will stand up to reviewer scrutiny. To be clear, we at STC do not see ourselves in that role. Anyway, as microbe enthusiasts, we rejoice at every new find that shows us that the Small Things keep crawling out of every drawer we try to put them in. Today's drawer: the minimal and maximal size of single bacterial cells (see here for a recent review by Petra Levin and Esther Angert.) Thiomargarita magnifica busts the upper size limit carefully calculated by Kempes et al. (2016) and is – or can be, as there is no defined unit cell length for this bacterium – quite a bit longer than Elaine Newman's famous E. coli metK mutant (~750 mm).
*) Perhaps it is not impossible that FtsZ rings can enclose and constrict a cell diameter of ~50 nm after all. I came across an older paper by Szwedziak et al. (2014) in which the authors describe that overlapping FtsZ filaments can form a complete ring in liposomes, with a ring diameter of 90 nm. However, in this artificial system, the inside of the liposomes was "adorned" with FtsA oligomers, which promoted the attachment of the FtsZ filaments. T. magnifica lacks FtsA, and thus my question above remains unanswered for now.