I saw this paper, and it was just asking be blogged about here. I figured I’d give it a shot.
Disclaimer – Not my work, never met any of the authors (although I’m sure they’re all within six degrees of me scientifically). The paper is open access, which I think is a good policy for me to adhere to in any future efforts along these lines.
Citation: L.A. Clifton, et al. “Low Resolution Structure and Dynamics of a Colicin-Receptor Complex Determined by Neutron Scattering.” The Journal of Biological Chemistry. Vol. 287, No. 1, pp. 337-346; January 2, 2012.
Among the many things that bacteria can do, one of them is knocking off other bacteria. There are a number of ways to go about this critical task, not surprisingly, and one of them involves proteins known as bacteriocins. These are proteins that the bacterium uses to kill off potential competitors, as they typically go after closely related bacteria. In this paper, the authors are focusing on Colicin N (ColN), a bacteriocin produced by E. coli. ColN depolarizes the inner membrane of Gram-negative bacteria by forming pores in the inner membrane, resulting in cell death.
The question the authors address is a fundamental one – how does ColN get past the lipopolysaccarhide-decorated outer membrane of a bacterium? It is ~ 40 kDa in size - so, clearly, not going to be able to easily masquerade as an ion or small molecule and pass unhindered through a pore in the outer membrane. The authors note that past research on ColN demonstrated that it is dependent on the presence of an outer membrane protein, OmpF (or related porins), to be effective. Cells that are OmpF-deficient will not be killed off by ColN. I should note that OmpF is a trimeric porin that permits the passage of ions and small molecules through the outer membrane. It was suggested that ColN could pass through the OmpF pore, but would need to be completely unfolded to do so. So there is clearly something going on here that is interesting.
The paper describes a multipronged approach to this question – the authors integrate microscopy, neutron reflectivity, and small angle neutron scattering (SANS). The authors step through their case – they first present the thin film imaging (Brewster’s angle microscopy) and neutron reflection data for their model of the OmpF/phospholipid monolayer. The microscopy suggests similar stability for the OmpF/phospholipid monolayer, although different topography and compression behavior (the formation of domains appears less evenly distributed in the OmpF/phospholipid monolayer, and there are “kinks” in the isotherm for the phospholipid-only monolayer compared to the OmpF-containing one). The neutron reflection data also seems to support the existence of an OmpF/phospholipid bilayer, despite Fig. 3B being mislabeled by my eye. Normally the neutron “refractive index” - neutron scattering length density, aka nSLD – is plotted as a distance away from some reference (e.g., an easily determined interface or a metal layer on which your sample is ultimately deposited). It seems that is what they intended to write (the x-axis seems to be labeled as such) but is mislabeled with the “Q/A-1” tag.
In any case, much of biologically-oriented neutron scattering is dependent on the existence of contrast variation in the nSLD. You can purchase deuterated compounds (such as lipids), prepare buffers in deuterium oxide, and even express & purify deuterated proteins. You then mix and match your deuterated and protonated components to see what each component looks like when in complex with everything else. It is a low-resolution means of doing so, but the benefits can outweigh the disadvantages.
The authors move onto the ColN portion of their work, showing the microscopy and neutron reflection data for ColN interacting with the OmpF/lipid monolayer. The time-lapse microscopy of ColN with the pure lipid monolayer and the OmpF/lipid monolayer shows increased image intensity, but appears to “smear” homogeneously with the pure lipid monolayer while forming larger, brighter spots with the OmpF/lipid monolayer. Their analysis of the neutron reflectivity data indicates the presence of the ColN in the same layer with the OmpF, and not just interacting with its surface, as they see in the ColN + pure lipid monolayer sample. Given the contrast variation matching, they state that they are able to see ColN extend as it inserts into the lipid region, suggesting that it is unfolding to some extent. The increase in surface pressure would suggest that it is not going through the OmpF pore but is, instead, inserting into the lipid region next to the OmpF. If it was inserting through the pore channel, the surface pressure might be expected to level off and not keep increasing.
The SANS data round out the story – they’re looking at the ColN/OmpF complex in detergent. (I know, I know.) Anyway, their data-derived model has one of the ColN domains slithering down between the cleft between OmpF monomers, while the remainder of ColN remains protruding outward. If you look at Fig. 6C, the blue distance distribution (where you are only looking at scattering from ColN) has two peaks, one that overlaps with the red trace (where one is only looking at OmpF) and a separate peak. So this at least makes sense. They do discuss the potential for translocation via the pore, and some recent literature on that possibility.
Mostly, I thought that this was a really interesting bit of research – while there is the obligatory mention of potential application to antibiotic development, it’s pretty obvious that the fundamental scientific question of “how does a largish protein get across a cell membrane where the cell has no interest in letting it inside?” I think that the experiments were reasonable, were carefully done, and did not set off too many massive alarms in my brain while reading. I would like to think that you could use something like nanodiscs or bicelles for the SANS studies so you could at least approximate a native membrane environment – clearly, sample homogeneity is a concern, as scattering methods can be notoriously sensitive. (Did I ever tell you about the time I spent a good afternoon into evening washing banjo cells for SANS experiments since said cells were just disgusting?) I haven’t worked with nanodiscs – although I’ve heard and read more than I can shake a stick at - and my experience with bicelles hasn’t been quite so detail-oriented, so maybe it would require sublime experimental mastery beyond the typical.*
Anyway. That was kind of fun. Also, how many of you saw Ohm’s Law Survives to the Atomic Scale? I imagine people will want to confirm this, as it is definitely seems really cool. Clearly, it was custom-made by “hand” (well, scanning tunneling microscope), so no immediate applications to large-scale mass production any week soon, but that isn’t why we do science.
Now, off to think about thermodynamics for a while. I need to come up with a reasonable explanation of some data today…..
3 hours ago