Scansion

In case you were wondering, I eventually plumped for the retrograde neurotransmission project, meaning I am now at least notionally some kind of mongrel pharmacologist. I'm attempting to educate myself in a few of the things that that entails, and may sketch out some fragments on here as I go along.

To begin, let's address the concept of scanning ion conductance microscopy, which is almost certainly destined to play an important role in my PhD. SICM is one of several scanning probe imaging techniques developed in the last couple of decades on the back of some seminal piezoelectric technology.

A piezoelectric substance, usually some kind of crystal, transduces electrical energy into mechanical energy and vice versa. For our purposes, that means it changes size when a voltage is applied. The relationship between voltage and size is very precise, and this can be used to position a probe -- attached to separate piezos for each of three orthogonal directions -- with an accuracy measurable in nanometres -- that is, billionths of a metre.

That's pretty fucking accurate.

If the attached probe is designed to measure something at its location, that can be used to determine what's happening there -- and thereby map the space the probe is moving through in considerable detail.

The first major scanning probe technique was scanning tunnelling microscopy, in which the thing the probe measured was the ability of electrons to teleport across an insulating gap thanks to a particular kind of quantum magic.¹ Many related techniques followed, all of them dependent on nanometre-scale probe positioning and the measurement of some physical parameter at the probe location. In the case of SICM, the probe is a tiny glass tube or nanopipette and the measured value is the tube's ability to conduct electricity in the form of ions.

An ion, as anyone who's sat through a high school chemistry class will vaguely remember, is an atom or molecule with an electron or two out of place. A lot of common substances -- in particular, salts -- are composed of ions, which dissociate when dissolved in water. The world's most famous and popular salt, mealtime condiment sodium chloride (NaCl), typically encountered in crystalline form as a neat ion lattice, is a case in point: in brine it exists as a bunch of separate sodium (Na⁺) and chloride (Cl^-) ions.

In the sort of electrical conductors we're most familiar with -- metallic wires and electronic components -- current passes in the form of loose electrons. That's not the case in most of the biological systems we care about. In your body and mine -- and in pretty much every living thing you encounter -- electrical currents, when they pass at all, are flows of ions.

One of the most important ways in which such currents flow is through the walls of cells. That kind of flow is the basis of all sensation, perception, cognition and control in your body. It's a good chunk of the essence of life. Though our understanding is far from complete, there can be no doubt that it is fundamental to every aspect of our existence. A whole pile of Nobel Prizes have been awarded to investigations of such electrophysiology -- some of them right here at UCL. My own PhD, in its small way, will be part of that tradition.

SICM, however, is not, at least in the first instance, concerned with transmembrane conductance. It's purview is broader.

Let's say you place your pipette into a bath of electrolyte. Within the pipette is an electrode; outside, in the main volume, is another. Between these two electodes you apply a voltage -- a potential difference. The result is what used to be called an electromotive force, pushing the electrolyte ions from one electrode to the other: a current flows.

But.

The current flows because the ions are free to move from one electrode to the other through the pipette. These ions are very small, but they aren't infinitesimal. In order for current to flow the ions need to physically pass into (or out of) the pipette.

Now, let's say you position the pipette so that its opening is very close to the surface of some object -- it doesn't matter what the object is, but for our purposes it'll be a cell. As the pipette gets close, the gap through which the ions have to pass gets really small -- as a result the conductance goes down. How much it goes down depends on the pipette's distance from the surface. If you were to push the pipette all the way down against the cell membrane, the ion conductance would drop to nothing.² But you don't do that. Instead, you use a feedback system to control the distance so that the conductance is always some known value: really small, but not zero.

Which means it's always the same distance from the cell surface.

And, because you're controlling its position with the piezos and know what voltage you're applying to them, you always know exactly where the pipette tip is.

Therefore, you know exactly where the surface of the cell is. As you scan the tip across the cell, you can determine its topography with, pretty much, the resolution of your piezos. Which, I probably don't need to tell you, is a fuck of a lot finer than the resolution available to optical microscopy.

That, in a nutshell, is SICM. With it we can image a cell's surface topography in remarkable detail while keeping it alive in physiological conditions. This is a Good Thing. We can also use the same piezo-controlled pipette to patch clamp particular locations on the cell surface and make electrophysiological recordings of transmembrane currents -- determining the behaviour of embedded ion channels. This is another Good Thing. The combination of these two capacities will be the foundation of my PhD.

However.

Just so you know.

SICM is a splendid technique, but it's incredibly fragile. You're trying to measure a cell with a glass nanopipette. It has to be positioned exactly. It has a limited range of movement. If it encounters drastic surface gradients it may jam or break. For such reasons -- and many others -- a lot of scans will fail. The setup time is considerable. The control software (developed by one of my supervisors, so I just have to hope he never reads this) is dog awful. Scanning is agonisingly slow.

SICM is a splendid technique, but it also sucks.

Maybe, peripherally, I'll also work on that a bit. Maybe not. There's a lot to do. An unlimited amount, actually. At stake, potentially, is a significant chunk of how the mammalian nervous system works. Far more than one PhD.

But hey: eyes on the road. Hands on the wheel.

Just drive, she said.

---

1 The fact that STM works is one of those perfect confirmations of quantum theory: something that looks like a mathematical contrivance, in this case involving complex exponentials, turns out to exactly describe the way things really happen. Not that such confirmation was lacking by the time STM arrived -- the behaviour of the humble transistor, and thus every single electronic device we use every day, is a product of quantum effects. Anyone who says that quantum mechanics does not operate at the macro level is deluded -- or, at least, has never used teh interwebs.
2 Okay, that's a lie. It specifically ignores the transmembrane currents discussed earlier, which are incredibly important. The mediation of those currents by membrane protein ion channels will be absolutely central to my project. Nevertheless, for our present coarse discussion of SICM, we can gloss over them and consider the cell surface as an undifferentiated insulator.
Posted by matt at October 12, 2007 11:14 PM