Rapid Brownian Motion Primes Ultrafast Reconstruction of Intrinsically Disordered Phe-Gly Repeats Inside the Nuclear Pore Complex in Scientific Reports, 2016
Just came across this recent work from the Mofrad lab (UCB and Lawrence Berkeley National Laboratory), and just blown away by it. The NPC contains nucleoporin proteins, which are ‘intrinsically disordered’ (i.e. highly dynamic rather than rigidly structured, see blog tag / Wiki) and as such come up ‘fuzzy’ to any imaging tool a lab cares to throw at it.
This dynamism is core [literally] to nuclear transport, letting hydrophobic FG [phenylalanine-glycine motif] repeats flail every which way in the confined space of the nuclear pore complex channel (often described as a ‘basket’ composed of ~30 different nucleoporins), creating an ‘oily spaghetti’ as it’s described here, i.e. a hydrogel which instills a phase separated boundary across which ~1000 molecules pass per second (described in notes I wrote here last year).
See also: 2013 post on Peter Tompa’s review of hydrogel formation by IDPs, ‘microtrabecular lattice’ reincarnate as he would have it.
Here, the hydrogel is presented as self-oscillating — compared to the fluctuations predicted by Alan Turing in 1952 (as written about a couple of years back, here). Belousov–Zhabotinsky reactions were observed in the 1960s, as Turing had foreseen from thinking about developmental processes, instantiated in a formal mathematical framework as reaction-diffusion systems.
In this piece the authors write:
The concentration fluctuations within the FG-meshwork are reminiscent of the cyclic motions of self-oscillating gels induced by an oscillatory chemical reaction called the Belousov-Zhabotinsky (BZ) reaction (Maeda 2008). In those systems, periodic chemical energy of the BZ reaction is converted to mechanical oscillations within the polymeric meshwork. Indeed, fluctuations in self-oscillating hydrogels has been harnessed for mass transport and cargo delivery purposes(Murase 2008, Shinohara 2008). However, there is a fundamental difference between those systems and the NPC in that no chemical reaction occurs inside the FG-meshwork, nor is there any external source of energy to wriggle the FG-meshwork.
Indeed, the strange beauty of IDPs is that their effects are purely entropic - of the possession of a jam-packed microcanonical ensemble, which can be viewed in thermodynamic terms or through information theory (mentioned regarding Boltzmann recently, and in the post last week on critical states, I touched on how these threads were pulled together by Shannon).
Back to Berkeley, the authors go on to present a totally new view of the situation:
Instead, we propose that the thermal noise, spreading through the geometrically confined NPC channel that hosts numerous transient, individually weak hydrophobic and electrostatic bonds, along with the delicate structures of the end-tethered FG-repeats, en masse produce incessant rapid Brownian motion, leading to continuous concentration fluctuations in the NPC channel.
I suppose thermal noise would have been taken as given in descriptions of motion at this scale by previous authors, but in this work it is explored quantitatively: the authors found the ‘wriggling’ induced by incessant fluctuations of Brownian motion produced by the reverbration of thermal noise upon these radially confined and channel-tethered molecules kept the NPC permanently sealed.
Butting in with a book recommendation to readers at this point: Howard C. Berg’s Random Walks in Biology.
This begs the question… if a molecule breaks through, how long does it take to ‘reconstruct’ the dense meshwork? Satisfyingly, they wrote that inter-FG motif hydrophobic bonds suction matters back together, ‘self-healing’ (but later this was disproven by simulation, hydrophobic crosslinks were suggested instead). The fine details of simulation take the wind out of this really nice result so I won’t go into their minutiae.
As well as the analogy to BZ reactions, the authors compare the system to the cell crawling mechanism (actin filaments driving lamellipodium protrusion, attachment, release/retraction and elastic-propelled procession along the surface).
We quantified the reconstruction pattern and time for the FG-meshwork in detail and proposed a time-dependent relation for this process which is biphasic with a rapid and a slow phase. The reconstruction occurs mainly during the first phase, which is mainly entropically driven; a cavity within a dense meshwork is entropically highly unfavorable, and thus, once the macromolecule passes through, the meshwork quickly rearranges itself. Once the configurational entropic cost is compensated, the density fluctuations within the cavity, ρ(t), continues more slowly with lots of ‘small-amplitude’ peaks and valleys during the saturating phase. Imaginably, favorable inter- and intra-FG-meshwork hydrophobic as well as electrostatic interactions play a more visible role in this phase.
…time of reconstruction, τ, lies anywhere between 0.44± 0.12μs and 7.91± 0.30μs, depending on the macromolecule size and shape. Given the transport time of a single macromolecule is anywhere between 3ms to several seconds, the reconstruction occurs three to seven orders of magnitude faster than the actual transport. This indicates that the reconstruction process is ultrafast and ‘instantaneous’ compared to the timescale of the entire transport process.
Note to readers who don’t come across biophysics much, in statistical mechanics angle brackets 〈around something〉 indicate ‘over an ensemble’, as in ‘averaged’ often (they can also be Dirac notation)
More precisely, τ is even shorter than the time of local diffusional motion of the macromolecule. Here, by local we mean the time it takes for a particle to diffuse a distance equal to its characteristic length, L. For example, for a globular cargo of 20 nm in diameter it takes about = /6D≅ 15.0μs to diffuse its diameter in cytoplasmic viscosity47. Since the cargo diffusion inside the NPC channel is slower than that within the cytoplasm45, 15.0μs is the lower bound on the local diffusional time of a 20-nm globular cargo inside the NPC. Yet, τ for the cavity created by the same cargo is 3.89±0.14μs (±SEM) (Table 1), meaning that the reconstruction process is at least fourfold faster than the local diffusional motion. This implies that the cargo does not leave any void behind itself, suggesting the NPC channel is perpetually sealed so that the traffic of macromolecules does not lead to breaking the permeability barrier or ‘leaking’
There are more conclusions which I won’t go into: within the radial zone of FG motifs’ wanderings there’s a ‘rod-like zone’ of particular rigidity, and various fascinating curiosities noted through experiment: such as how hydrophobic binding spots may scale with size of the macromolecular surface.
One thing’s for sure: physicists still want to drill down to the finer details, as ever.
Nonetheless, from the viewpoint of polymer physics criteria, whether the FG-meshwork is truly a ‘gel’, or an entangled meshwork of polymers, or a confined polymer brush, or something else, awaits further in-depth mechanical investigation.
In his book on IDP structure and function (to my knowledge the only such text on the topic), Peter Tompa stuck to 'entropic brush' - I think there's acknowledgement that these are only convenient placeholders until later work follows up proceedings.
Ending with a nod to the systems biological analogies they left as they went, of oscillatory processes and self-regulation, the authors conclude:
More importantly, the current study also suggests that the aggregation of FG-repeats within the NPC channel can be viewed as a novel biopolymeric stimulus-responsive network that immediately changes its conformation in a stimulus-dependent manner. In a shape- and size-dependent way, FG-meshwork ‘rapidly opens up’ in response to hydrophobic affinity difference, while the intrinsic ultrafast Brownian motion of FG-repeat biopolymers ‘quickly closes’ the meshwork as soon as such a stimulus disappears. These call for new investigations on the NPC from novel biopolymer physics viewpoints, combined with shape effects
What’s that phrase, everything that connects computes…?
Some of my favourite IDP papers of late, starting with the father of the field:
Vladimir Uversky (2016) Dancing Protein Clouds: The Strange Biology and Chaotic Physics of Intrinsically Disordered Proteins. The Journal of Biological Chemistry
Vladimir Uversky (2016) Protein intrinsic disorder-based liquid–liquid phase transitions in biological systems: Complex coacervates and membrane-less organelles. Advances in Colloid and Interface Science
Bergeron-Sandoval (2016) Mechanisms and Consequences of Macromolecular Phase Separation. Cell
Wu and Fuxreiter (2016) The Structure and Dynamics of Higher-Order Assemblies: Amyloids, Signalosomes, and Granules. Cell
Schmidt and Görlich (2016) Transport Selectivity of Nuclear Pores, Phase Separation, and Membraneless Organelles. Trends in Biochemical Sciences
Nott et al. (2015) Phase Transition of a Disordered Nuage Protein Generates Environmentally Responsive Membraneless Organelles. Molecular Cell
Patel et al. (2015) A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease Mutation.. Cell
Csizmok et al. (2016) Dynamic Protein Interaction Networks and New Structural Paradigms in Signaling. Chemical Reviews
Pak et al. (2016) Sequence Determinants of Intracellular Phase Separation by Complex Coacervation of a Disordered Protein. Molecular Cell
...You get the picture - there really is a minor surge of papers on this topic, it's wonderful to watch a new aspect of biological regulation appear before our eyes. The word 'coacervate' just came back to me, it's oddly rarely used on my network but I've been seeing it in these papers regularly.
coacervate: a colloid-rich viscous liquid phase which may separate from a colloidal solution on addition of a third component.
Inevitable Wikipedia definition:
Coacervation is a unique type of electrostatically-driven liquid-liquid phase separation, resulting from association of oppositely charged macro-ions. The term "coacervate" is sometimes used to refer to spherical aggregates of colloidal droplets held together by hydrophobic forces.[1] Coacervate droplets can measure from 1 to 100 micrometres across, while their soluble precursors, are typically on the order of less than 200 nm. The name "coacervate" derives from the Latin coacervare, meaning "to assemble together or cluster".
The process of coacervation was famously proposed by Alexander Oparin and J. B. S. Haldane as crucial in his early theory of abiogenesis (origin of life/ prozikhozhdenic zhiney). This theory proposes that metabolism predated information replication, although the discussion as to whether metabolism or molecules capable of Template replication came first in the origins of life remains open and for decades the theory of Oparin and Haldane was the leading approach to the origin of life question.
My former research supervisor M. Madan Babu was off on a mountain at some unspecified conference the other week, a photo and audio recording of which popped onto my Twitter feed (Cell Conversations: Illuminating the Dark Proteome). In the conversation recorded, Madan asked “How many phase separated structures can actually coexist, and how do they interact with each other? What kind of biology do they drive?”
...so subsequently that's what I've been ticking over on too. The list above came from looking through the citations to the particularly wonderful Nott 2015 study ("disordered nuage" protein coacervates, tying together P-granules, Cajal bodies and other incunabula of some new textbook understanding... there really is a whole hidden level down there...) - as well as my Google Scholar alerts Twitter bot on the topic, which leaves a tweet trail of mine and others' readings and occasional comments (it can be hard to remember from a title alone if you've read a paper sometimes).
Google Scholar: citations to Nott et al. (2015) since 2016
...it's with regret I note that I did set up a nice blog to send these to but forgot to link it up to the feed :-( idp-papers.tumblr.com - watch this space perhaps...
My other notes from the Cell Conversations session (via Twitter)
New challenges: to decode combinatorial SLiM code & low complexity sequence patterns conferring ability to phase separate
“Barely scratched surface of signalling puncta”, eg: neural Shh Src sequestrat'n in cancer
Madan Babu: “How many phase separated structures can actually coexist, and how do they interact with each other? What kind of biology do they drive?”
Advice to next generation: “key would be to be bold, and apply a multidisciplinary approach… try to link IDP function directly to biology”
