by Christoph Weigel
When it comes to cellular differentiation in bacteria, two model systems stand out in my mind: sporulation in Bacillus subtilis and the two different cell morphologies in Caulobacter crescentus. Both systems have in common that dramatic changes in cell morphology are easily observed in the microscope. In the case of B. subtilis spore formation a number of stress signals can derail the cell from its default cell cycle program: cell growth > chromosome replication > cell division > cell growth > … . A mother cell that switches to sporulation 'sacrifices' itself in the process of producing the spore (see here in STC ). In C. crescentus the cellular differentiation is part and parcel of its normal cell division cycle. A stalked surface-attached and mother cell grows, replicates its chromosome, and, upon cell division, releases a flagellate 'swarmer' daughter cell before resuming growth, chromosome replication, and so on, the 'default' cell cycle (Fig. 1, lower arrow). Chromosome replication is suppressed in the swarmer cell until it finds a surface to which it can attach (Fig. 1, upper arrow). Upon attachment via a 'holdfast' patch of secreted polysaccharide the swarmer morphs into a stalked cell, which then resumes growth, chromosome replication, and so on. Both model systems have in common that the execution of the differentiation program is conducted by a dozen or so 'key regulators', while the complete orchestra comprises well over hundred proteins that perform specific localized actions in a coordinated timely order (see here for a visual summary of genes involved in Caulobacter differentiation ). Finally, both model systems were initially chosen for studying differentiation because the organisms were considered less complex than eukaryotic cells. What a beautiful delusion!
Detach Caulobacter from a surface – if you can
As a consequence of studying Caulobacter differentiation for many years, Yves Brun and his coworkers became interested in finding out how strongly the stalked cells attach to microscope slides (Fig. 2). They already knew well that the caulobacters, once they have attached to a glass surface, cannot simply be removed by rinsing the slides with water. The attached cells resist even flushing with strong jets of water (working in the 'wet lab' can be extremely fun! ). To more precisely measure the adhesion force, they "seeded" cells onto a thin flexible pipette whose force constant could be calibrated by AFM (see below ). A suction pipette mounted to a micromanipulator was then used to grab the body of an attached cell and pull it away from the flexible pipette (Fig. 3). Because of the large force required, the cell body had to be sucked into the pipette, which was then bent away from the pulling direction. In this setup, the adhesion force could be calculated from the amount of bending required to break the cell–pipette contact (see here for a video of one experiment ). Calculating the force needed for detachment required a careful calibration, and to this end the researchers employed Atomic Force Microscopy (AFM), a technique that allows ultra-precise measurements of sub-nanometer distances by controlling the forces exerted on the AFM cantilever that 'scans' a surface, or, in this case, bends a pipette tip. They could thus calibrate the support pipette and obtained correlation curves for the forces required for its deflection over a certain range of distances at different cell attachment points. For the final calculation, they measured the holdfast geometry since the detachment force acts over the entire area with which the holdfast sticks to the flexible pipette. The result was stunning, as they wrote:"At 68 N/mm2, the adhesion strength of the holdfast is stronger than gecko's toes setae and far stronger than all other known cellular attachments. C. crescentus holdfast covering 1 cm2 would have the potential to hold a weight of 680 kg on a wet surface, making it an excellent candidate as a biodegradable or even surgical adhesive. The lower limit of the attachment strength measured for the holdfast in this study is stronger than the commercial dentin adhesive, which can provide a bond strength up to 30 N/mm2 ". The minor remaining problem with holdfast™ as a commercial product is precisely its stunning super‑adhesive property: to handle the substance, one would need a complete production pipeline with 'lotus effect' surfaces, which are still in the process of development.
How do caulobacters profit from such strong attachment to surfaces in their natural habitat ? When they thrive in small creeks with rapidly moving water, small waves, and eddies, they will frequently get into contact with air-liquid interfaces (=bubbles ) that easily exert shear forces sufficient to remove most other microorganisms attached to pebbles. The adhesion force of C. crescentus measured by Tsang et al. appears to be beyond the estimated threshold value but within an order of magnitude.
What is this super "super glue" made of ? The holdfast is a tiny blob of polysaccharide – crosslinked ß-1,4 N‑acetylglucosamine polymers with a diameter of ~100 nm, to be more precise – much like what many bacteria produce and secrete as part of the extracellular matrix when they make biofilms. Members of the Caulobacterales, a taxonomic order (=subfamily ) of the Alphaproteobacteria, are special in that they produce small localized patches of this 'glue' for attachment instead of covering themselves entirely with 'slime'. Polysaccharide secretion is triggered when caulobacters sense a surface via mechanical distortion of their pili and/or impediment of their flagellar movement. C. crescentus secretes the holdfast polysaccharide close to the pole where the force-sensing pili and the flagellum are located, and start growing the stalk pretty exactly at that point in the membrane, by protruding it, where the cell has already attached to a surface. Thus, the C. crescentus holdfast is located at the end of a stalk. Such 'stalks,' also called prosthecae, differ from pili and flagella in that they are not composed of protein multimers anchored to/in the cell membrane but actually are membrane protrusions containing cytoplasm (what drives the stalks to assume their cylindrical form and narrow diameter is not well understood ). The Caulobacterales are particularly playful as their stalks are not always found at one cell pole as in C. crescentus . Some members of the order have more than one stalk, as for example, Asticcacaulis biprosthecum, and others, including Hirschia baltica and Hyphomonas neptunium, produce their progeny cells at the distal tip of their stalk while passing the replicating chromosome through the stalk's internal 'channel' (this is extravagant if anything. Both have ftsZ genes and are thus likely to employ the conventional bacterial-type of cell division, though ). Since most Caulobacterales thrive in nutrient-poor environments, it is reasonable to assume that their stalks serve primarily to increase the surface‑to-volume ratio for gathering scarce resources more efficiently. It has been shown that stalk membranes are particularly rich in nutrient receptors. Indeed, C. crescentus grows its stalk to ~30× the normal length(!) when starved for phosphate (here is a picture of cells with an intermediate 'long stalk' phenotype ).
Do Caulobacter cells age ?
The fact that C. crescentus mother cells have clearly visible assymetry and are very sticky allowed Martin Ackermann, Stephen Stearns, and Urs Jenal to address a long-standing question in bacterial physiology: Are bacterial cells immortal or do they senesce? Bacterial cells can die, of course. Ask bacteriophages, predatory bacteria like Bdellovibrio bacteriovorans, or protists – including our own macrophages, which much behave like the latter – and all will confirm that bacteria can be killed with ease. Or take Bacillus subtilis , where the mother cell lyses, or in other words, dies after producing a spore. Also, bacteria eventually die when deprived of nutrients. Not immediately and not always, though, as for example bacteria in marine sediments have been found to somehow survive on a minimal diet (including cannibalism ) for millennia with 1 – 2 doublings per century(!) on average. But do bacterial cells age? That is, do they become 'senescent' as Peter Medawar defined it back in 1951 as:"ageing accompanied by that decline of bodily faculties and sensibilities and energies which ageing colloquially entails" . For bacteria, such an 'ageing process' (note Medawar's British spelling ) could possibly be observed by a decline of their 'reproductive output', meaning a decrease in the production of progeny cells over time under constant growth conditions. Seen naïvely, one would assume that bacterial cells don't age because growing E. coli cells, for example, divide symmetrically at midcell, which leaves the observer with hardly a chance to tell the mother from the daughter. But what Ackermann and colleagues did was to grab C. crescentus by the stalk due to the stickiness of the holdfast. Thus, they were able to follow the reproductive output of individual mother cells for extended periods of time (see Fig. 1).
Their experimental setup was outright minimalistic: they let Caulobacter cells attach to the surface of a microscopic flow chamber and then applied a constant flow of fresh growth medium across the surface to which the cells kept firmly attached (by the adhesive holdfast at the distal tip of their polar stalk ). The medium flow did not only serve to keep the growth conditions constant over time but also to flush away most of the progeny swarmer cells emerging from division (Fig. 4). They followed groups of cells of the same age for about 300 hours under the microscope and recorded their reproductive events (this sounds more tedious than it actually is because modern camera equipment can do the recording automatically and software can assess the images ). They calculated the number of progeny produced per individual as a function of its age. This quantity represents the age-specific reproductive output, combining both survival and division. Although some cells produced up to 130 progeny cells in 300 hours, many stopped dividing or divided more slowly with increasing age, resulting in a marked decrease in reproductive output with age (see Fig. 5; the example is taken from one of their other experiments). They found in control experiments that young swarmer cells born in the chamber after 250 hours were indistinguishable (±10%) in terms of cell division time from young cells measured at the beginning of an experiment.
The results obtained by Ackermann et al. were clear cut: yes, Caulobacter cells do age and cease dividing after ~60 generations, on average. But as for any clear experimental answer to a properly asked question a whole basket of new questions pops up immediately. Is senescence in Caulobacter a consequence of its dimorphic life cycle or do bacteria that divide symmetrically also become senescent over time ? How can aging stalked cells produce juvenile daughters, swarmer cells ? Is aging genetically programmed or a consequence of accumulating junk ? How could a hypothetical genetic program for aging (now in American English ) be executed exclusively in the mother cell while sparing the offspring ? How could the accumulation of junk protein ('aggregates') be confined to the continuously growing mother cell (cell constriction at midcell starts late in the Caulobacter cell cycle, not before >50% of all components of the daughter cell are already synthesized ) ? And finally, what could be a selective advantage for Caulobacter of having evolved senescence ? As it turns out, some of these questions have been addressed and we now know a lot more about bacterial senescence. So much so that the subject will be treated in a follow-up post. Diving any deeper into that subject now would loosen my holdfast to the topic of this post: the Caulobacter stalk.
Frontpage: C. crescentus cells in pre-divisional stage. Future swarmer cells with flagella (left) and stalked cells (right) with clearly visible stalks and holdfasts (blob at the distal end of stalk). No scale given. Source.
Figures linked in the text: A The Caulobacter cell-cycle control circuit. A Schematic of the cell cycle. The θ-like structure in the cell represents the replicating chromosome. Colors indicate which master regulator(s) is present in each cell type or cell compartment. B The genetic network controlling the cell cycle. Source; B Dimorphic cell cycles, polar adhesives and stalks are conserved across diverse Alphaproteobacteria. Maximum likelihood tree is inferred from GyrA sequences of various Alphaproteobacteria, the Gammaproteobacteria E. coli and Pseudomonas aeruginosa serving as an outgroup. Source; C Growth of wild-type and mutant C. crescentus und phosphate-limiting conditions. Source
Christoph Weigel is lecturer at Technical University Berlin, and at HTW, Berlin's University for Applied Sciences. He's an Associate Blogger for STC.