Rosette is a rosette is a... (1|3)
by Christoph
I dare say that every biologist was blown away when, about ten years ago, they first saw pictures of spherical rosettes that choanoflagellates – single-celled protists, eukaryotes – can and happily do form when nudged by certain bacteria. Is this emergent multicellularity? We here at STC felt the same way, and Roberto's recent piece Inside a Choanoflagellate attests to that (see here for an animated 3D reconstruction). Skilled microscopists had long been aware of such "rosettes" among (some) bacteria, but they never made it into the canon of bacterial shapes as they only form under certain growth conditions, and most likely as an intermediate between single‑celled, planktonic growth, and multicellular growth in a biofilm.
Figure 1. Roseobacter strain 27–4 (now Phaeobacter inhibens strain 27–4) grown in MB at 25°C under static conditions. Pictures are from phase contrast microscope at ⨯1,000 magnification. One bacterial cell is approximately 1 by 3 µm. Frontispiece: detail of Fig. 3b from Source
Roseobacter rosettes
Among the various images of bacterial rosettes that you find by an image search, those shown in Figure 1 from a roseobacter, Phaeobacter inhibens, are among the most meaningful. For although they are less sharp and detailed than the few rosettes in Figures 2+3, you see so many of them, each formed from about a dozen cells, that there is no doubt that virtually all the inoculated cells in this static culture formed rosettes almost in sync, and not just a few that had gone astray (or were just preparation artifacts).
Figure 2. Scanning electron micrographs of Roseobacter strain 27–4 grown in MB under stagnant conditions. (a) Cells were only exposed to osmium vapor and subsequently coated with gold‑palladium. (b) Standard fixation and dehydration procedure was used. Source
Bruhn et al. (2005) observed that rosettes – they fondly talk of stars – formed primarily at the liquid/air interface in a static Phaeobacter culture (Figure 1) and that these subsequently compacted into a biofilm in which individual rosettes could no longer be visualized. The formation of a biofilm at the liquid/air interface is well known from Bacillus subtilis and is referred to as a "pellicle" (you may also know it as the thin slimy layer that forms on fermenting apple juice when you leave the bottle standing for a few days without shaking).
Phaeobacter cells divide by binary fission, and in agitated liquid medium the daughter cells do not stick together for long (see here, full Figure). Other reserchers have found efficient rosette formation also in agitated liquid medium but it was tricky to get the conditions right (ref. here and here). The currently favored idea, in short, is that moving planktonic cells attach to a preferred solid substrate when they have sensed it, that is, touched it with the tips of their flagella or of their pili. The attachment is accomplished by shedding the flagella and secretion of EPS that coats the first colonizing cell(s), and subsequently their daughter cells, and affixes them to the substrate. Voilá, you now have a young biolfilm.
Figure 3. Scanning electron micrographs of the axenic dinoflagellate Prorocentrum minimum CCMP 1329 co-cultivated with Phaeobacter inhibens DSM 17395 for two days under static conditions. Scale bar for part C, D and E is 2 µm and for part B is 1 µm. See parts B, D, and E here. Source
This is exactly what Frank et al. (2015) observed (in the electron microscope) when they examined the colonization of the dinoflagellate Prorocentrum minimum by Phaeobacter inhibens. A mutant Phaeobacter lacking the ability to secrete EPS was unable to colonize the dinoflagellate, but the wild-type could, and occasionally, but not consistently, formed clearly visible rosettes (Figure 3). It is therefore evident that Phaeobacter can form rosettes both during biofilm formation at the liquid/air interface and during colonization of submerged solid surfaces. It is equally evident from the other panels of Figure 3 (see here) that rosette formation is not a prerequisite for biofilm formation by Phaeobacter, and thus certainly not a defined developmental stage.
Nevertheless, it is worth taking a closer look at Figure 2. Bruhn et al. (2005) removed by careful washing most of the EPS sheaths (and salt crystals) from rosettes seen in Figure 2a, but their "cleansed" rosettes remained connected (Figure 2b). They say: "The individual cells were anchored to one another with fibrils." "Fibrils" are not precisely defined, neither in terms of their dimensions (length, diameter) nor in terms of their molecular composition as polymers (protein, polysaccharides, nucleic acid). The authors leave it at these "fibrils", and these can indeed be seen quite clearly at the "base" of the two larger rosettes in Figure 2b.
Figure 4. (A) Polar polysaccharide at the center of rosettes. Fluorescent overlay image of membrane stained cells (red) and Alexa 488-conjugated WGA lectin (green). Scale bar corresponds to 1 μm. (B) Kinetics of rosette formation. Rosettes form by cell encounters. Two independent aliquots from the same synchronized culture were stained with either Alexa 488-conjugated WGA (green) or Alexa 594-conjugated WGA (red) and were mixed for 30 minutes at room temperature in buffer prior to imaging. Rosettes with dual-labeled centers are shown. Images are overlay of phase contrast (gray) with green and red fluorescence channels. © Rosette complexity increases over time. The number of cells per rosette was quantified over time. Error bars indicate standard deviation of two biological replicates; n > 300 cells. Source
But perhaps not only there? In many cells in Figure 2b, I see nipple‑shaped protrusions on the old pole of cells that are about to divide. They vaguely resemble – emphasis on vaguely – an early stage of stalk formation in the Stalking Caulobacter (another, but distantly related, Alphaproteobacterium). It is easy to imagine – but completely speculative, I must add – that in two daughter cells directly after division, the proto-"fibrils" extend and attach to each other. The "folding over" of one of the two cells required for this would be physically facilitated by the fact that they are in cahoots, that is enveloped by a shared polysaccharide sheath. This could be the starting point for the formation of a new rosette, which could then be enlarged by repeating the process until there is no more space for the attachment of newly "incoming" cells (which would result in rosettes of approximately the same size as you see in Figure 1). If I were not writing a post for STC here, I might run to the lab to look for a Phaeobacter mutant that no longer makes these protrusions – and then see if it can still form rosettes.
I could have saved myself the last two sentences – I apologize for wasting your time – if I had done more careful research into the literature. It is actually quite normal that if you look more closely, you will find that others have already had the same or a very similar idea... and have already done the experiments. Just like here. To avoid unnecessary bias, I don't always look at papers in which Roberto was involved as an author first. I should have done it here. Segev et al. (2015) studied Morphological Heterogeneity and Attachment of Phaeobacter inhibens and found that 1. an GlcNAc-containing polysaccharide detected by fluorescently labeled lectin on the poles of some cells, the "nipples," and at the center of all rosettes, the "fibrils" (Figure 4A), and 2. that rosettes are formed by the attachment of individual cells at the polysaccharide-containing pole rather than by cell division (Figure 4B,C). Covering of rosettes by EPS does apparently not play a role in their formation, and my Gedanken Experiment was off the mark in this respect.
A side note. The genus Roseobacter in the Roseobacteraceae family of the Alphaproteobacteria, like Phaeobacter inhibens strain 27–4, got it's ICNP-approved name for the rose-colored colonies that the type strains R. litoralis and R. denitrificans form on solid media or the staining of liquid cultures due to their expression of bacteriochlorophyll a. It's impossible, of course, to see whether the individual ovoid or rod-shaped cells have rosy cheeks, but the name does not reflect their talent to form rosettes. Elio wrote A Whiff of Taxonomy – The Roseobacters earlier, along with a colored version of the frontispiece used here.
Stay tuned for part 2, in which I will turn to rosette formation in Planctomycetes and Nevskiales, before looking at another Alphaproteobacterium, Caulobacter and, inevitably, Escherichia coli in part 3.
Do you want to comment on this post? We would be happy about it! Please send us an email, or comment on Mastodon, or on Bluesky.
