UCSC BME 220 Homework 2: making pictures

Jeff Ferguson, May 4th 2009

H9 Haemagglutinin


Heamagglutinin is a protein found in the influenza A virus. There are 15 distinct HA variants known with H9 being solved after the respiratory virus outbreak in Hong Kong in 1999 in which it was involved. The HA protein is involved in receptor binding as well as membrane fusion making it the focus of study in flu strains and also one of a strain’s defining characteristics.

Figure 1

(a)
(b)

(c)
Figure 1 (a) H9 shown in its primary conformation with chain A colored red and chain b colored blue, disulfide bridge at cys4, cys137 shown in yellow spheres, antigenic loop shown in green sticks (b) close detail of cysteine inter-chain linkage (c) close detail of antigenic loop with residues labeled and atoms colored

The HA protein is a trimer takes two distinct physical conformations. The first conformation is formed when the protein first folds before the chain is cleaved into two chains. This conformation is the one the virus presents on its capsid and the one that binds to cell receptors inducing endocytosis. The second conformation is a dramatic change of fold resulting in helices extending into the cell membrane inducing fusion. The conformational change is induced by the lowering of pH in the enclosure around the virus and it’s been shown that this conformational change can be induced by several different destabilizing effects, not only changes of pH (Carr et al). The suggestion is that the protein is “spring loaded” with the second conformation being lower energy so that anything that starts to denature the protein can cause it to shift to the more favorable state. My thought when I read this was it would be interesting to run a secondary structure prediction on the protein to see if a helix is predicted for the segment (residues 60-74) that go from a coil region in the first conformation to an extended helical region in the second conformation.

To show the way the protein’s conformation changes I’ve taken a model of the extended conformation chain B fragment of H9 and aligned the chain B from the complete protein in its primary or folded conformation. To make this alignment I broke the protein in several points and aligned it’s fragments to the extended structure to show which parts correspond.

I was disappointed to find that there was nothing predicted for the coil though I’m unclear on how the alignment effects the secondary structure predictions. The protein’s only been solved (as a whole) in the first conformation so there’s certainly “conservation” showing the region as consistently being a coil region.

Figure 2
(a)
 (b)
(c)

 (d)
Figure 2 (a) The trimeric structure of the extended conformation of the HA protein, this is a fragment containing almost none of chain A and missing a large segment of chain B (b) a monomer of the restricted HA protein shown in extended conformation (c) the full chain B from H9 in its non-extended or folded conformation colored so that each color represents a segment that has been broken free of the rest of the chain for alignment (d) the alignment of the H9 chain B to the extended monomer showing which segments of H9 correspond to which segments of the extended conformation; the alignment is nearly perfect and is only misaligned in coil regions where I would have had to break and align each residue separately


From the alignment in figure 2 you can clearly see that the region colored in red is forced to extend when the inter-helical coil region stiffens into a helix. The region preceeding (toward the N-terminus) the red helix is what fuses with the cell membrane. It has no alignment because it is removed by the protease used to purify the extended conformation fragment.

References:

Chavela M. Carr, Charu Chaudhry, and Peter S. Kim, “Influenza hemagglutinin is spring-loaded by a metastable native conformation”, PNAS Vol. 94, pp 14306-14313, December 1997

Per A. Bullough, Frederick M. Hughson, John J. Skehel and Don C. Wiley, “Structure of influenza haemagglutinin at the pH of membrane fusion” NATURE, vol. 371, September 1 1994

Ya Ha, David J. Stevens, John J. Skehal and Don C. Wiley, “H5 avian and H9 influenza virus haemagglutinin structures: possible origin of influenza subtypes”, The EMBO Journal, vol 21 No.5, pp. 865-875, 2002

PyMOL

I chose PyMOL because it’s what I already had installed and because it’s what we were taught in BME110. It seemed that since it’s currently the most popular and it’s open source it would likely continue to be around and I wanted to learn a tool that would continue to be useful.

I took a lot longer getting some of the pictures I wanted than I expected. This was mostly because I wasn’t familiar with the functions I wanted to use such as the alignment tool and I wasted a lot of time trying to figure out the “sculpting” features which seem to exist only as a demo. Reading the PyMOL wiki page on the sculpting pretty well confirmed that I had set off in the wrong direction. Once I had learned about the align feature and started putting everything together the work went reasonably fast and PyMOL over all seems like it has plenty of power. The other reason it took me a long time getting my pictures is at first I didn’t have the trimer fragment structure so aligning the protein wasn’t an option. When I found that structure to use as a template breaking up my chain and aligning it was actually very straight forward. The alignment process seemed a little clumsy because there wasn’t an option (that I was aware of) to give the backbone some flex at certain points so the whole thing had to be broken into individual objects and aligned.

Because of the way PyMOL is built on Python it seems like to take full advantage of it I may want to pick up some Python. Even without that though there are plenty of scripting options. There’s also the issue of weak documentation though the wiki seems to have enough in it to at least point someone learning it in the right direction.