My Thesis
My PhD thesis entitled “Functional and Structural Studies of the Human Voltage-Gated Proton Channel” has recently been published online by the Rockefeller University website here. The Thesis covers six years of extensive research that I carried out in the Laboratory of Molecular Neurobiology and Biophysics at the Rockefeller University under the tutelage of Dr Roderick MacKinnon. I worked closely with many postdocs in the lab during my tenure in the lab and would like to especially thank Dr Seok-Yong Lee (now at Duke University, check out his website here) and Dr Joel Butterwick along with all other members of the MacKinnon lab.
—
The thesis constitutes the first ever biochemical characterisation of the human voltage-gated H+ channel. It includes one-and-a-half chapters of published data then an additional two-and-a-half chapters of unpublished data, including a never before seen crystal structure (you will have to read the thesis to understand why we never published this structure before now). Additional unpublished data includes extensive mutagenesis studies and NMR characterisation of the detergent solubilised channel. I think that anyone interested in voltage-gated ion channel biochemistry and biophysics will appreciate this work. For anyone who works on voltage-gated H+ channels I think you will really get a kick out of all the unpublished work and encourage you to contact me it you have any questions or comments.
Click to see find and download my thesis from the Rockefeller University library archive.
—
Below is also a link to a recording of the public presentation portion of my Thesis defence. This presentation does not by any means cover all the work presented in my Thesis but I do discuss some of the major finding. This video has also been posted under the ‘My work’ page of this blog and on my linkedin profile so you may have come across it before. Hope you enjoy.
—
Please feel free to use the comments section below to ask me any questions or comment about any of this work.
LettsScience 
Congrats James!
Thanks!
Great job! I am very glad that my dissertation defense was NOT recorded! It was excruciating! Seriously, though, you were able to determine a structure for a chimera of hHv1 with KvAP, but you did not publish it because you felt it was in a non-native configuration. The Osaka group published a structure of a chimaera of mHv1 with CiVSP and S. cerevisae GCN4 (does that make it a trimaera?). The obvious questions is, do you think this might be a native structure?
Hi Tom! Thanks for the comment and question. Let me start off by saying that Takeshita et al. got much closer to the native structure than I ever did. However, I have some doubts about wether their structure represents a native structure of the Hv channel. As you mention in order to get their channel to crystallise they had to make the “trimaera” as you put it and I can certainly sympathise Hv is a very difficult protein to work with in detergent as my thesis attests.
This trimaera-ization, however, does seem to introduce some non-native elements to the structure, most pronounced is the fact that the structure they solved is that of a trimer and to the best of my knowledge native Hv has not been shown adopt a trimeric configuration in the membrane but is a dimer. They also show that their construct is dimeric in the membrane under the same conditions for which they show function. Therefore they don’t specifically provide evidence that the trimeric form of the channel is functional.
Some aspects of the structure seem reasonable, such as the positions of the S4 arginines relative to the Phe gap, however, even though they are able to register the helix using Se anomalous signal, higher resolution is required to really see the details of the side chain interactions (there is basically no density for the side chains at all).
I think the Zinc binding site looks reasonable, although I know you have your own theories about that 🙂
My major criticism of this paper is their model of the dimer that they generate. This model does not agree at all with the published cross-linking data and they don’t really address this fact to my satisfaction in the paper. They build this dimer model from the trimer structure since they know that the functional channel is a dimer in the membranes. I would like to see a study of site-specific cysteine cross-linking of their construct in its dimeric membrane embedded form. I would predict that the dimer in the membrane would display a cross-linking pattern similar to wild-type Hv and that the dimer model based on their trimer structure is wrong.
One final thing, I am not entirely convinced by the fact that the coiled-coil and S4 are a single continuous helix. There is certainly some functional evidence that seems to point to this being the case, however, in my thesis I present limited proteolysis data that indicates that the linker between S4 and the coiled-coil is very sensitive to proteolysis indicating that there isn’t any significant secondary structure. Nonetheless, it would not surprise me if there is a transition between helical and non-helical structure in this region depending on the gating state of the channel (or reduction state of the coiled-coil Cys).
Thanks again for the question and comment. Takeshita’s structure is clearly the culmination of a lot of really intricate molecular biology, electrophysiology, biochemistry and just plain hard work. It is a testament to how difficult Hv is to work with that after all their work (and mine for that matter) this is the closest we have come. Like I said they certainly got closer to the native structure than I did but I still think more structural work needs to be done. Now I am interested to know what you think of the structure?
One thing in the structure especially surprised me. This regards the “hydrophobic plug.” We all know about the Tao et al (2010) paper describing the highly-conserved Phe (F150 in hHv1, F146 in mHv1, F278 in Shaker, F233 in paddle chimera) residue that is the external boundary of the “charge transfer center.” Bezanilla and colleagues (e.g., Campos et al, 2007) including Perozo, include this Phe as part of the “hydrophobic plug” that separates In from Out, and seems to act in much the same way Tao et al had described for Phe alone.
In the Li/Perozo structures of CiVSP, the hydrophobic gasket comprises three residues all at the same level, forming a collar through which S4 passes during gating. The three hydrophobic residues are all at the same level (with each other) both in Up and Down states. In our Open model for hHv1, the three corresponding residues also align horizontally, with Asp112 above and Arg211 below them. When I map the corresponding residues onto the closed mHv1 structure, F146 and V105 seem to be in their expected locations, but one residue (V174 in S3) seems to be much lower and pointing away from the pore.
I asked Yasushi about this (in view of this residue being near the splice with CiVSP) and he had noticed it. They believe their structure reflects the native one, and they speculate that most likely the role of V174 in the gasket (assuming an analogous thing exists in Hv1) is played by F178 in mHv1. I suppose there is no reason that everything should be identical in every VSD-containing molecule. So I have no opinion, but will be interested to see what new information may appear.