The goal is to gain insight into a protein’s structure and conformational ensemble by combining smFRET experiments as input parameters with in-silico smFRET built on a powerful Monte-Carlo conformation sampling tool (see Fig 2A). Progress include the creation of the software ‘COMPLEXES’ for conformational structure sampling. Fig 2B shows a test on CHMP1B.
I simulated CHMP1B in COMPLEXES, and then picked FRET pairs that would be optimal (Fig2 C). Mutations were performed via 2-step PCR (Fig 2D) and the proteins expressed, purified, labeled and imaged on a custom built TIRF microscope in the Hurley lab. For data analysis, I wrote ‘py2FRET’ (Fig 2F, left, a donor/acceptor puncta pair; right, FRET intensities over time). For structure analysis of ESCRT proteins in their respective assembly structures, Ist1-CHMP1B tubules were formed. Fig 2G depicts the Frost lab’s structure of the assembly. Fig 2H depicts my successful efforts to recreate these tubules.
For ~15 years, researchers tried to reconstitute the ESCRT pathway with limited success. I proposed to do an optical tweezer experiment on ESCRT filled membrane vesicles that could fully reveal the pathway: The optical tweezer, applied from the outside would induce an outward bud and tubule in the vesicle. The ESCRTs on the inside would recognize this negative curvature and the tubule as their native template and act on the membrane. Once a working reconstitution of the ESCRT assembly was found, a subsequent step could monitor ESCRT action in a functionally relevant state via FRET and other structure-probing methods.
Through Dan Goldman, I had built a connection to the Bustamante lab. We chose to upgrade the ‘Mini-Tweezer’ (Fig 3A) to allow for the first membrane tweezing experiments in the lab. I built a custom-made pressure controller (Fig 3B), and later exchanged it to a Fuigent model (Fig 3C). Fig 3D depicts the experimental setup: A giant unilamellar vesicle (GUV) is held by a micropipette. A tweezing bead, functionalized to stick to the membrane, is held in an optical trap above the GUV. A membrane nanotube is pulled from the vesicle by slightly touching and pulling away with the bead. The tube pulling will result in a characteristic force pulling curve. ‘sigma sampling’ can be used (Fig 3E) to measure bending rigidity. Visualizing the nanotubes was done together with Maurizio Righini on the ‘Fleezers’ instrument, which we modified by adding imaging capability (Fig 3F). The resulting images of our membrane nanotube (Fig 3G) convinced us that we were on the right track.
Tweezing experiments are performed in microfluidic chambers (Fig 4A). To develop a UV-light based ATP uncaging setup, outside illumination (Fig 4B) was replaced with an optical fiber that could deliver the UV light precisely to the tip of the aspiration pipette (Fig 4D). Multiple iterations led me to develop a protocol to encapsulate all yeast ESCRT-III components (snf7, vps2, vps24, vps4) and caged ATP into a vesicle (Fig 4E). Tweezing and ATP uncaging on these vesicles led to a surprising and exciting result: very large forces were generated.
I used fluorescence based protein localization using confocal microscopy and two micromanipulator controlled pipettes (Fig 4G). Nanotubes pulling and protein decoration of the tubes from the inside could be confirmed (Fig 4H). To fully combine force and fluorescence experimentation, we decided to build the West Coast’s first confocal plus optical tweezer instrument (only a handful worldwide). I led the effort of building the microscope and writing the operation software together with Shannon Yan and Il-Hyung Lee from the Bustamante and Hurley labs. We named the instrument Confleezer 1.0. (Fig 4I, J). With this new instrument, we could show that the ESCRT proteins go specifically into the nanotube tube and become enriched in the tube after ATP release while the force rises. Most importantly, we could reproducibly demonstrate ESCRT specific and ATP and Vps4 dependent membrane nanotube scission – a first in this field!
The work is published: Schöneberg et al. NRM 2016 (doi:10.1038/nrm.2016.121) and under review (bioRxiv:
https://doi.org/10.1101/262170(opens in new window)). Confleezers 1.0 operating software is disclosed to UC Berkeley’s intellectual property management
