[Publication] A fibril with encoded water channel

Our cryo-EM work on the hamster prion (sequence 108-144) fibril is recently accepted in the Journal of the America Chemical Society. The hard works done by Wendy and Jess definitely made the project move smoothly. The work is a collaboration with Dr. Rita Chen’s group (Institute of Biological Chemistry, Academia Sinica).

A “fake” cover art I made to celebrate our fibril paper published in JACS.

Rita’s group prepared the fibrils and took the cryo-EM images with the great assistances from the two EM managers Dr. Sean Wu and Dr. Hsiung Wang from the Academia Sinica CryoEM facility.

We took over the data process, analysis and model building. We had learned from our synuclein work (not published yet, in revision) that RELION 3 didn’t do a good job picking up correct fibrils. Therefore, Wendy and I have decided to manually pick up all fibrils. Yes, by human eye balls! Wendy had checked 6000+ images and manually picked >180,000 fibrils in 3 weeks. The eye-checked fibrils were processed by RELION through multiple steps including 2D and 3D classification, optimization of helical parameters and refinements. The resolved cryo-EM map of this fibril finally achieved to 2.23Å — a very high quality map to present sidechains and molecular interactions.

During our fibril picking process, we also evaluated the accuracy performed by crYOLO (see my another Blog post). crYOLO did a quite good job in few hours for 6000+ images. We absolutely will use crYOLO as a sharp tool to accelerate fibril data analysis for the future cases.

In contrast to conventional amyloid fibrils, this hamster prion 108-144 fragment formed a fibrils with “4 protofilaments”. The 2.23 Å map indeed greatly helped us to identify the individual protofilaments. Jess and I built the molecular structures of this prion fibril and found some unidentified EM maps inside the central pore of this 4-protofilament fibril. With careful inspection and modeling, we found out these signals are water molecules as they form a helical water wire followed by the 4.7 Å helical rise and 108˚ helical twist rules. The water molecules interact with Tyrosine (residue 128) through stable hydrogen bonds further confirming the formation of helical water wire. We were superb exciting of the findings as it is the first (to our knowledge) report showing waters in the amyloid fibrils. The tight assembled 4-protofilament fibril stabilizes water wire inside the central pore.

We believe this structure provides novel information of fibril formation. It is way beyond our current understanding. The water wire could be used an tool for future artificial design of water channel, too.

The work is accepted on July 8 in J. Am. Chem. Soc. and published online on July 20th. I also have a tweet for it immediately online posted.