New probes, singularities, and a reason for RNAi mismatches in vivo (Part 2)

Me in the rain.jpg

Me, Michael Shannon, in rainy Taipei

Takeharu Nagai (part 2/2): Singularities in cells: leaders, followers and citizens

Takeharu describes a familiar situation. Our results, whether its molecular behaviour or cellular behaviour are generally an averaged description of phenomena. We get an idea of what’s going on, but ignore the minority phenomena: the singularities that lead to group behaviours.

After a brief rundown of things he considers singularities: the big bang, the benign to malignant switch, formation of iPS stem cells with only 4 genes altered, populist fascism under Trump etc, he sets out to investigate this idea armed with a plethora of new fluorescent probes, and some cool model systems.

Here, he focuses on cAMP signalling in social amoeba, which transition from single celled entities to an intercommunicating mass, a multicellular being. To initiate this switch, all you have to do is starve them of nutrients.

By using two markers, Flamindo2 (Odaka) and R-FlincA (Horikawa, in press) he derives a ratiometric measurement for cAMP activity, known to be associated with this switch in amoeba behaviour. Combining this with a high speed microscope tiling technique, he is able to look at the behaviour of both single cells and the whole population of cells.

What he finds is amazing – single amoeba cells become leaders, displaying a burst of cAMP, before setting off their neighbours. “Early followers” then signal to “late followers”, and within a matter of hours, to “citizens”, setting off a continuous spiral of cAMP signalling which causes the amoeba to group together and become multicellular. The fluorescence videos of this are quite amazing – and while the paper isn’t out yet, you can view some similar behaviours online at Take’s website.

Interestingly, several leaders seem to be selected, but only one gains ultimate dominance as the seed of the spiral of signalling. One of the goals of the Nagai lab now is to find out how this leader is selected.

Okay, next up, Sua Myong

Sua Myong – How does RNAi actually work?

Sua employs FISH, a super resolution technique, to view RNA interference in single cells. What she finds is the first insight into the function of RNAi with reference to biologically relevant miRNAs since Fire and Mello won the nobel prize for the work and revolutionised the field.

The investigative technique works by targeting particular mRNAs with search RNAi strands loaded with 30 to 40 fluorophores each. By watching the transient binding of these RNAs in TIRF, the PSFs can be localised and the relative number of RNAs between conditions can be quantified.

The group tried to figure out which parts of shRNA were important in terms of its structure and its interactions with DICER and RISC, the proteins that bind it to genes of interest and cut the genes, respectively. To do this they altered the shRNA supplied to the cell by first changing the size of the hairpin loop, before measuring silencing using the FISH technique described above.

Longer loop size improved silencing, and it was found that this was dependent on better association with the DICER protein.

Introducing mismatches in the nucleotides was also trialled, as this is common in biological settings – many endogenous microRNAs have these mismatches which prevent them from binding the target gene perfectly.

What they found was that the altered shRNA had not problem binding DICER, but was inhibited in its handover to the RISC complex.

This is important, because it may be a way for cells to control the power of microRNAs, in cases where protein translation must be fine tuned.

Thanks for reading – that’s me over and out for this meeting.

Bicycle repair shack.jpg

A local bicycle repair shack in Taipei

Michael Shannon (Dylan Owen lab, KCL)


Sensor and probe development, day 4 of the BPS conference in Taipei

Michael Lawson Amy Palmer Julie Biteen.jpg

Michael Lawson, Amy Palmer, and Julie Biteen


Here are the talks on Sensor and Probe development, all on the last day of the conference (Part 1)

Amy Palmer probes the cellular ionosome


First up today, Amy Palmer develops new probes to investigate the ‘ionosome’* – the complex but under-recognised flow of metal ions within living cells. This is a relatively untapped field – 30 % of proteins in cells require a metal cofactor to function, yet only really Calcium has been addressed with fluorescent probes, or even recognised as a regulator of cell function.

Zinc ions are particularly interesting, and that’s the focus for today’s talk. It’s level is sensed by cells, which adjust their metabolism in response. 10% of proteins require Zinc, so it does a lot of different jobs.

OK so first of all the probe made by Amy and her team. This is a FRET probe, composed of two Zinc binding domains between the donor and acceptor. When Zinc binds, conformational change occurs and the probes come within the magic distance, facilitating energy transfer which can be detected. In addition, a signal sequence can be added to direct the construct to somewhere specific eg – membrane, nucleus, cytoplasm.

To prove the probe works, the group carried out in situ calibrations – within the same cells they will be testing. To soak up all the Zinc for a low signal readout, they used a chelator (in this case TPEN) and to get a high Zinc signal they used a Zinc carrier (Zinc Pyrithione). This works pretty nicely, and their yellow FRET signal is ratiometric with reference to the green donor signal alone. There also doesn’t seem to be any perturbation of endogenous Zinc concentration due to the probe itself, which is nice.

Next they looked at the kD values for binding of Zinc to some of the proteins it regulates, and it turns out that they are not fully occupied under physiological conditions. This is important as it points towards Zinc indeed being a regulator, functioning by binding on and off to the protein of interest.

One example of this is CDK2 in the cell cycle. Zn fluctuations accompany high CDK2 cytoplasmic recruitment/activity, when the cell has just exited mitosis. The group found that these zinc fluctuations are required for the decision to translocate CDK2 from nucleus (inactive) to cytoplasm (active).

One of the next steps for the group might be to target the probes to particular organelles, to investigate Zinc’s role there. Very interesting stuff.

*me and Amy agreed that this was a cooler name for this network than metallosome – what do you think?

Takeharu Nagai (part 1/2) – Acid resistant fluorescent protein for super resolution

Takeharu’s group looks in strange places for new fluorophores.

First up is the work of Hajime Shinoda, a “very handsome and cool student” in Take’s words. He has developed a new fluorescent protein, called Gamillus isolated from a flower hat jellyfish, which survives at low pH, where EGFP and others lose their signal.

That’s because it has a trans-isomeric conformation, instead of a cis one. It can’t gain H+ ions in a way that would disrupt the aromatic rings in the chromophores of the rest of the green cis proteins, so it is compatible for imaging low pH environments, like the inside of lysosomes. A nice control is shown: use GFP, you can’t see lysosomes, use Gamillus, and suddenly little polkadots appear in strategic cellular locations. It works!

Next, handsome Hajime altered the protein, so that it might be compatible for super resolution. He used the very problem that Gamillus solves, transition to cis isomerism, to achieve photoswitchability. Reversibly switchable Gamillus blinks on exposure to UV light, by switching between the inactive cis and active trans form. (Shinoda, unpublished data).

Michael Shannon


Taekjip Ha kicks off the mechanobiology section with tension

He’s going to talk about his now pretty famous (if you are into that kind of thing) tension sensors. This is important because all cells sense their outside environment and respond to it through mechanosensitive membrane proteins, which transmit this important information to the cytoplasm of the cell.

He uses short lengths of DNA with a fluorophore on each end. FRET occurs at low tension, and at high tension FRET is abated as the molecule is stretched.

The tension sensors can be programmed to lose their FRET at pN scales, on a range from 4 pN to 60 pN. So it’s a pretty direct way to measure how much force cells need.

Amazingly to me (I study integrin nanoclustering) Takejip shows that only one or two strongly binding integrins are required to induce successful binding of receptors to much less strong integrin adhesions and induce cell spreading.

[continuing our fairy tale theme, Takejip compares this phenomenon to the princess and the pea – tenuous but I’ll take it]

Next, he shows that Notch, another membrane protein, is depended on between 4 pN and 12 pN – a really small amount of force, in order to drop Gal 4 and allow it to go to the nucleus. At the end, Jagged1 is mentioned, which has a profile similar to that of notch – it needs 4 to 12 pN of force to activate, and seems to follow the ‘catch bond’ schema.

Pakorn Kanchanawong unfolds vinculin with 3D super resolution


Pakorn is very well known by now, but I’ve never seen him talk. He uses iPALM to get  sub 15 nm isotropic resolution in x y and z, and brought out some pretty memorable papers on the ultrastructure of focal adhesions.

Here he has done the same in cadherin based adhesions, in adherens junctions between cells. These are  super complex adhesions, and iPALM only currently works in the TIRF zone…so Pakorn and his team created a fake cell like substrate on the coverslip, so that the cell makes a junction that can be imaged.

Next he showcases a whole host of information about the cadherin adhesions, which appear fairly clearly segregated into three layers. The middle one, the interface zone, contains vinculin and its on this that he focuses.

By tagging both the N and C terminals of vinculin, Pakorn find that it stretches from 5 nm to 30 nm in length! To stretch it must be phosphorylated by Abl kinase, as well as having mechanical force applied across it. In addition, it turns out that Zyxin and VASP are taken with the vinculin, reaching nearly the height of the cortical actin.

A complex and rather elegant use of drugs and precise measurements (the best combination) from Pakorn once again, showing that mechanics and phosphorylation couple to produce vinculin stretch and subsequent molecular clutch engagement.

Ashley Nord describes the mechanosensitivity of flagellar machinery in bacteria

Ashley Nord for the final talk in this section. She describes this minimalist machinery that operates the bacterial flagellar – the alien like tendrils that drive bacterial swimming.

She describes mechanosensitivity in the ‘stators’ and does some pretty clever measurements on them. These are ion channels, which lend energy to the engine that is the flagellar machinery. The first thing she does is prove that more stators translate to a faster motor.

She also finds that the stators turn over dynamically – after a little while, they leave and are replaced by new stators. Crucially, in viscous solutions, the number of stators goes up to the maximum amount available. In less viscous solutions, the number of these stators goes down to about 4.

Okay and that’s me out for now. I will let the others take over!

Michael Shannon


Structured light, deep imaging and automated super resolution

Morning everyone. Everyone is looking pretty fresh and ready for the day, and appear to have entirely dodged jetlag. I myself am dangerously full of caffeine and ready to go.

The lecture hall is filling up, with little biophysical interactions (ha) going on all around…we await the introduction by Jung-Chi Liao.

Jung-Chi asks out thanks for bringing all of the rain from across Asia – funny as it has been raining here for 4 days straight now [note – I ask someone later and this is not normal in Taipei!]. He reminds us of the upcoming Shilin night market and cultural tours plus the banquet at the end. A big thanks to the other organisers, particularly Wei-Chun.

Suliana Manley – Aesop’s fable by SIM


Achilles Kapanidis and Suliana Manley 

And up first is Suliana Manley : she uses Super Resolution live microscopy to reveal that cell wall manipulating proteins are also responsible for changing cell size after division.

Suliana opens with the popular Aesop’s story the ‘tortoise and the hare’. She uses the story to describe the ability of individual bacteria to decide how big they should be, a bit like the hare, choosing his/her moment to race through the cell cycle.

The model is Caulobacter crescentus – and she shows us that population of such bacterial cells plateaus to become all the same size – the larger ones get smaller and the smaller get larger. With this in mind, she uses Structured Illumination Microscopy to address the question of how this happens.

Its known, says Suliana, that the divisome, the group of proteins that manage the construction and destruction of the peptidoglycan cell wall are the ones that also form a contractile ring when cells are ready to divide in two. By using SIM, which is really gentle technique, the group can watch several cell cycles and collect ample data while tagging these two proteins.

What’s interesting is that by treating the bacteria with a pedtidoglycan inhibitor (fosfomycin), the cells get longer (by decreasing the rate of pedtidoglycan synthesis). It seems therefore that the availability of peptidoglycan intermediates might be a way to control the size of such cells.

Bi-Chang Chen – lattice light sheet extends to localisation

A swift change then, after some questions, to Bi-Chang chen – well known for his work on the lattice light sheet in Eric Betzig’s lab.

As usual, the lattice images never cease to amaze. Using multiple Bessel beams and a spatial light modulator, the team have been able to image whole cells, and even whole brains – like that of the fly – 185 um maps. That’s because beam energy is spread, photodamage is reduced and signal to noise is maintained. The sample is moved up and down to achieve magnificent images.

Of most interest to me is the single molecular side of things. In particular, swept z stacks of sox2 transcription factors in live cells are impressive. I quite fancied that I could see ‘hop’ diffusion, certainly there were regions where the Sox2 molecules were clustering together in a very interesting consistent looking way. Bi-Chang also describes some new self blinking dyes suitable for super resolution localisation microscopy. This type of dye is probably going to be invaluable for live cell localisation: especially for nanocluster tracking.

Shean-Jen Chen: Adaptive optics in a new guise

Next up Shean-Jen Chen who is going to talk to us about the availability of his multi-photon, which has recently been revamped and outfitter with some sweet new adaptive optics that he calls ‘temporal focusing’.

It’s a multi-photon ‘scope but not as we know it. He’s made it fully biotissue compatible, in all of its messy, undulating glory.

First, he has used a grating to create multiphoton at the sample plane where the now separate wavelength beams converge. Second, he uses a Digital Micromirror Device (DMD) basically a plate with 1000s of tiny controllable mirrors, with which he can structure the light going onto the sample, to get to that high spatial frequency information: only one lens as this is non -linear. The effective resolution laterally is 168 nm and axially is 1.3 um.

Finally, the juicy bit – he uses a deformable mirror to compensate for temporal distortion. He basically measures how the light should look and corrects for deformation in the image. Biotissue has diverse refractive indices, so correcting for them with adaptive optics like this is desirable.

The result of all of that is high speed (30 cell volumes per second) and much nicer resolution, which will be invaluable in years to come, as the biologists dream up new avenues, some of which might be 6 cells deep!


Takeharu Nagai and Yemima Riani


Masahiro Ueda: super resolution x 1000

OK so the final talk of this session comes from Masahiro Ueda. This one really stands out for me, something that has been in my mind for some time!

Masahiro has some very impressive statistics. That’s because he’s found a way to automate super resolution microscopy, as well as drug treatments of such cells. With this technique, we are looking at 1000s of cells per day – incredible: my record on the dSTORM at King’s is around 80 cells in one (long day). So this really is a big deal.

The set up involves a two 96 well plates, one full of chemicals or drugs, the other full of cells, which both go into a chamber with the microscope at 37 degrees, with 5 % CO2. The drugs are added to the cells at time points with a robot (and extensive washing – the video of this robot washing itself was strangely amusing to us, the jetlagged), while the cells are located using machine learning software, autofocusing (on the side of the aperture) and a quick fluorescence check to avoid unnecessary time wasting.

The example given here is the clear relationship between EGFR trimerization and an activation marker, the adaptor protein Grb2. The Grb2 proteins cluster and hover around the areas of no longer diffusing EGFR trimers. Very promising stuff.

Oh, and yes, they did also engineer an objective oil supply device for all of those stage shifts!

Phooph! That’s the end of that section *runs into the rain to reinvigorate*

Next up, mechanobiology!­