Note: I wrote this "news and views" style piece last (academic) year for a graduate Cell Biology course at UC Berkeley. I just reread it and was a little disappointed with the closing, so if you have any suggestions I'm open to edits. Thanks to Abby Dernburg for assigning super interesting papers.
A vestige of my suburban upbringing, springtime always calls to mind spring cleaning. The thought conjures the lemon smell of Lysol, images of whirling dust and near-bursting trash bags. Also, neatly organized files and the bittersweet feeling of boxing unexamined books and magazines to be hid indefinitely in the attic. The notion of organization implicit in spring cleaning is one of concerted work – expelling accumulated entropy from the house – to achieve boundedness – as all things go in their appropriate containers.
Crowded, dense, filled with broken machines in the process of being disassembled, fixed and remade, cells hum with activity and mess much like a modern city; cells are hardly “clean” or “organized” in the spring cleaning sense. Amidst all the mess, Eukaryotic cells are nonetheless elegantly organized. DNA is mostly contained in the nucleus. Only certain proteins can enter through the nuclear pore and only at certain times. Many such proteins bind DNA and alter the expression of genes, but may bind only to specific sites. Oxidative phosphorylation takes place across mitochondrial membranes and hundreds of proteins encoded by the nuclear genome are directed specifically to that organelle. Exquisite vesicle transport machinery delivers cargo destined for degradation to lysosomes and proteins destined for secretion through the endoplasmic reticulum to the Golgi apparatus and to the cell surface (1).
Focusing on the organelles and the motions of molecules and vesicles between them is akin to thinking about cells cleaning in spring – all the things in the right boxes. Of course Eukaryotic cells use membranes to delineate their interior, to enable spatial and temporal separation of their components, but are there other modes of cellular organization? This question is particularly pointed with respect to bacteria, archaea and to organelles themselves, which lack internal membrane-bounded structures. Are bacteria simply miniscule bags of molecules diffusing and interacting at the will of the thermal bath? Or do they benefit from other sorts of organization? Two recent papers (2, 3) shed light on modes of organization that are different from membrane-bounding in that they are relative (this stuff doesn't mix with that stuff) but not topological (this stuff goes in this box). These papers raise the possibility that cells of all stripes are even more exquisitely organized than we'd thought.
If bacteria were “bags of molecules,” you'd expect all of their components – nucleic acids, proteins, metabolites, etc. – to diffuse freely. Indeed, free diffusion is thought to be important for bacteria in evenly distributing cytoplasmic components between their offspring. The diffusion equations puts this more precisely – the mean squared displacement () of a freely-diffusing molecule should be inversely proportional to its size () until it meets a boundary like the cell membrane that stops it from wandering further (4). In a recent paper in Cell, Bradley Parry and Ivan Surovtsev were struck by an observation at odds with free diffusion: a large bacterial filament called crescentin (~900 nm in length) appears to diffuse freely in growing bacteria, but slows to a crawl when those same bacteria are starved or subjected to a respiratory uncoupler called DNP (2). Diffusion is a passive process due to thermal fluctuations (4), but the motions of crescentin in Caulobacter crescentus are somehow driven by metabolism. Parry and Surovtsev demonstrate that metabolism-dependent motion of large biological molecules is both general and size-dependent. Small molecules like GFP diffuse freely independent of hosts' metabolic state while molecules like plasmids, crescentin and a synthetic fluorescent probe (called μNS) with effective radii larger than ~30 nm (three times the size of a ribosome) display metabolism-dependent motions in both E. coli and C. crescentus, bacteria separated by at least a billion years of evolution (2).
What's most fascinating about Parry and Surovtsev's work is their demonstration of “glassy dynamics” in the bacterial cytoplasm. When the drunk man of myth walks, he doesn't prefer any particular direction – he's so drunk that he forgot his last step, not to mention the way home. As such, the displacement of freely-diffusing molecules should follow a Gaussian distribution (2, 4). Since the drunkard is unbiased by his past, watching one drunk for a long time is much the same as observing many different drunks each for a short time – they all wander randomly and without bias, exploring a mean squared distance according to the diffusion equations. But this is not the case for larger molecules in the bacterial cytoplasm. Irrespective of the presence of DNP, the movements of μNS probes is not well-fit by a Gaussian (2). What's more, in the presence of DNP, the motions of μNS probes are “non-ergodic” in that probes of similar sizes may wander at very different speeds even in the same cell – the past movements of a probe have an effect on its future movements, suggesting that some molecules are confined or “caged” by their neighbors (2). Altogether, Parry and Surovtsev's beautiful work bears witness to a crowded and viscous bacterial cytoplasm at a near standstill during starvation and actively mixed by some unknown process during growth. These findings may be crucial in understanding how bacteria survive the long periods of starvation which are common in the wild. Moreover, the heterogeneity and non-ergodicity observed suggest that whatever process mixes the bacterial cytoplasm may be selective and perhaps even directional, opening a door on a much more nuanced view of how bacteria organize their cytoplasm.
In a paper in Nature from 2012, Pilong Li, Sudeep Banjade and Hui-Chun Cheng (3) address a very different question: why is it that so many Eukaryotic proteins contain multiple binding domains? This question is especially important for understanding cellular organization because multivalent proteins, particularly those binding RNA, are enriched in “cellular bodies” like Cajal bodies and C. elegans P granules. Cellular bodies are not membrane bounded, but they resemble organelles in that they are spatially separate and compositionally distinct from their surroundings and carry out particular biochemical tasks like RNA splicing (5). To interrogate the properties of multivalent protein bodies, Li, Banjade and Cheng used the SH3 domain and its ligand the “proline-rich motif” (PRM) to generate synthetic multivalent proteins of different valencies. When mixed, these synthetic proteins phase separate from buffer and form droplets that coalesce with each other but do not mix with the buffer, like water droplets on glass. Droplets can also be made to form in-vivo by co-expressing SH3 and PRM oligomers in HeLa cells, wherein multivalent proteins associate reversibly with the droplet phase and exchange with the cytoplasm in ~10 seconds (3).
Li et al. posit that regulated condensation of multivalent proteins into liquid droplets could coordinate local switch-like transitions in signaling or assembly processes by quickly concentrating an enzymatic activity in one place. Nephrin, for example, is a transmembrane protein found in the “foot projections” of kidney epithelial cells. Nephrin contains three phospho-tyrosine binding sites which bind SH2 domains of the protein NCK when phosphorylated. NCK, in turn, contains three SH3 domains which bind to proline rich motifs of N-WASP, which is known to stimulate the assembly of actin networks through it's interaction with the Arp2/3 complex (Figure 4A of 2, 5). These proteins phase separate in-vitro and the critical concentrations for phase separation are lowered by ~10 fold by supplying phosphorylated Nephrin. Li, Banjade and Cheng reconstitute a sharp phase transition in the rate of actin assembly by titrating an N-WASP mutant that can bind to NCK SH3 domains but not Arp2/3, demonstrating clearly that regulated phase-transitions can drive the condensation of wild-type N-WASP and, through it, the local assembly of actin filaments (3).
Physics at the scale of cells is very different from what we experience daily: surface tension is much stronger than gravity and inertia is negligible compared to viscous drag, such that a moving bacterium stops almost instantaneously when its flagellum stops spinning (4). It stands to reason that some modes of cellular organization would be similarly unintuitive. Imagine if our books were magnetically attracted to a corner of the living room. Ignore for a moment that you'd need an astronomically powerful magnet in the wall to overcome the books' inertia and imagine organizing books in that fashion – piled up in some random order against the wall. They'd certainly be out of the way, but it would be awful hard to find your favorite book. And yet, separating proteins and nucleic acids from each other by mutual affinity appears to be a common and powerful means use by cells to separate and concentrate activities (3, 7). Likewise, our homes are not nearly as crowded as the interiors of cells and nothing in our macroscopic world moves by diffusion. From this perspective, it is difficult to imagine why you would need activated diffusion to separate or mix cells interiors. And yet it seems that such systems exist and are retained by almost unrelated bacteria (2). These two papers call for us all to put on “physicist hats” to build intuition about the challenges and opportunities that are presented to cells in organizing their microscopic world.
1. Alberts B et al. (2010) Essential cell biology: an introduction to the molecular biology of the cell (Garland Pub., New York, New York, USA). 3rd Ed. Available at: http://books.google.com/books?id=inZWDr4h1aIC.
2. Parry BR et al. (2014) The bacterial cytoplasm has glass-like properties and is fluidized by metabolic activity. Cell 156:183–94. Available at: http://www.ncbi.nlm.nih.gov/pubmed/24361104 [Accessed March 19, 2014].
3. Li P et al. (2012) Phase transitions in the assembly of multivalent signalling proteins. Nature 483:336–40. Available at: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3343696&tool=pmcentrez&rendertype=abstract [Accessed March 21, 2014].
4. Phillips RB, Kondev J, Theriot J, Orme N, Garcia H (2009) Physical biology of the cell (Garland Science New York).
5. Matera a G, Shpargel KB (2006) Pumping RNA: nuclear bodybuilding along the RNP pipeline. Curr Opin Cell Biol 18:317–24. Available at: http://www.ncbi.nlm.nih.gov/pubmed/16632338 [Accessed March 24, 2014].
6. Jones N et al. (2006) Nck adaptor proteins link nephrin to the actin cytoskeleton of kidney podocytes. Nature 440:818–23. Available at: http://www.ncbi.nlm.nih.gov/pubmed/16525419 [Accessed April 28, 2014].
7. Brangwynne C, Eckmann C (2009) Germline P granules are liquid droplets that localize by controlled dissolution/condensation. Science (80- ) 5:1729–1732. Available at: http://www.sciencemag.org/content/324/5935/1729.short [Accessed April 2, 2014].
