4 INTERNATIONAL S ON CRYO-EM 3D IMAGE ANALYSIS 2020Sciences, UCSF, San Francisco, CA, USA. 8:40 –...

4TH INTERNATIONAL SYMPOSIUM

ON CRYO-EM 3D

IMAGE ANALYSIS 2020

March 18 –21, 2020 Granlibakken Conference Center & Lodge Lake Tahoe, CA USA

Program Schedule Abstracts

Registered Participants

Cryo-EM 3D Image Analysis Symposium 2020

http://cryoem.bcm.edu/cryoem/events/view_workshop/1

2

The Organizing Committee gratefully acknowledges financial support from the following sponsors:

SPONSORS


https://www.igneous.io/

https://www.igneous.io/

https://www.panasas.com/

https://www.thermofisher.com/us/en/home.html



3

4th International Symposium on Cryo-3D Image Analysis 2020

March 18–21, 2020

Welcome to the 4th biennial CryoEM 3D Image Analysis Symposium to be held at Granlibakken in March 2020. As in the past, the symposium provides a GRC-like atmosphere with a strong focus on technical aspects of algorithms for image processing in CryoEM/CryoET. A significant amount of time in the schedule has been reserved for discussions, and there are several social events we encourage people to use for deeper discussions. To help encourage open discussion and to avoid distractions during talks, please note that this meeting follows a strict no photography and no audio/video recording policy. The organizers will take a few non-science photos to document the event, and these will be shared on social media. If your vision prevents you from easily reading the screen, we encourage you to sit closer rather than pulling out your camera. The poster session, however, will permit opt-in photography. Each poster presenter will have the option of putting a tag on their poster indicating that photography of that specific poster is permitted. This should eliminate the need for participants to bring printed copies of posters they wish to disseminate more widely. If you see this tag (will include picture in schedule) on a poster, you may take pictures of it. Please respect the request of authors who request no photography. Non-photographic social media posts are also permitted. Please use the hashtag #3DEMImgAn.

ORGANIZERS

Dorit Hanein

Sanford Burnham Prebys Medical Discovery Institute

La Jolla, CA, USA

Steve Ludtke Baylor College of Medicine

Houston, TX USA

ORGANIZING COMMITTEE

José-María Carazo Garcia Centro Nacional de Biotecnología, Madrid, SPAIN

Edward Egelman UVA Medical School Charlottesville, VA, USA

Masahide Kikkawa University of Tokyo Tokyo, JAPAN

Pawel A. Penczek UTHealth Houston, TX, USA

Scott Stagg Florida State University Tallahassee FL, USA

Niels Volkmann Sanford Burnham Prebys Medical Discovery Institute La Jolla, CA, USA



i 4

Table of Contents

Program 5 –13 Speaker Abstracts 14 – 37 Poster Abstracts 38 – 56 Registered Participants 59 – 65 Meeting Room Diagram 66

5

Program

WEDNESDAY, MARCH 18TH

3:00 – 6:00 PM

Check in (Hotel Lobby)

Meeting registration materials will be provided at check-in

Posters should be up for the entire meeting (Bay Room).

5:00 – 6:00 PM Welcome Wine and Cheese (Granhall)

6:00 – 7:30 PM Dinner (Granhall)

7:30 – 7:45 PM

Welcome Introduction (Mountain Lake)

Dorit Hanein, Program Chair Professor, Immunity and Pathogenesis, Sanford Burnham Prebys Medical Discovery Institute, La Jolla, CA, USA

David DeRosier, Session Chair Professor Emeritus of Life Sciences, National Academy of Sciences, Department of Biology Brandeis University, Waltham, MA, USA

7:45 – 8:45 PM

8:45 – 11:00 PM

Helical Polymers: A Personal Perspective

Edward H. Egelman, Keynote Speaker Professor, Department of Biochemistry and Molecular Genetics, University of Virginia, Charlottesville, VA, USA

Social/Cash Bar (Cedar House)


6

THURSDAY, MARCH 19TH

7:00 – 8:00 AM Breakfast Buffet (Granhall)

Session 1: Tomography (Mountain Lake)

8:00 – 8:15 AM Introduction, Scott Stagg, Session Chair Associate Professor, Department of Chemistry, Florida State University, Tallahassee FL, USA

8:15 – 8:40 AM Pushing Water Uphill with a Rake: Resolving High-Resolution Structure of the Nuclear Pore Complex with Subtomogram Averaging Beata Turoňová Staff Scientist, Structural and Computational Biology European Molecular Biology Laboratory, Heidelberg, GERMANY

8:40 – 9:05 AM Incorporating Tilted Image Processing into cisTEM Benjamin Himes Instructor, Howard Hughes Medical Institute Janelia Research Campus Ashburn, VA, USA

9:05 – 9:30 AM A Complete Data Processing Workflow for CryoET and Subtomogram Averaging Muyuan Chen Instructor, Biochemistry Baylor Collage of Medicine, Houston, TX, USA

9:30 – 9:55 AM Alignment of Tilt Series Daniel Castaño Senior Scientist, Center for Cellular Imaging and Nanoanalytics at Biozentrum, University of Basel, Basel, SWITZERLAND

9:55 –10:20 AM Coffee/Tea Break

Session 2: Conformational Variability (Mountain Lake)

10:20 –10:35 AM Introduction, Pawel Penczek, Session Chair Professor & Co-Director, Structural Biology Imaging Center, Department of Biochemistry and Molecular Biology, University of Texas-Houston Medical School, Houston, TX, USA


7

10:35 –11:00AM “Structure” in Structural Biology Abbas Ourmazd Distinguished Professor of Physics, University of Wisconsin, Milwaukee, WI, USA

11:00 –11:25 AM Non-Classification-Based Methods for Deciphering Continuous Conformational Variability: Methods Combining Image Analysis, Molecular Mechanics Simulation, and Machine Learning Slavica Jonic Research Director, CNRS Impmc-cnrs umr 7590

Sorbonne University, Paris, FRANCE

11:25 –11:50 AM What Is The Resolution Limit In Single Particle Cryo-EM? Holger Stark Professor, Department of Structural Dynamics Max-Planck Institute for Biophysical Chemistry University of Göttingen, Göttingen GERMANY

11:50 –12:15 PM Continuous Heterogeneity Using Spectral Volumes Joakim Andén Research Scientist, Numerical Analysis Center for Computational Mathematics Flatiron Institute, New York, NY, USA

12:15 – 1:15 PM Lunch Buffet (Granhall)

Afternoon free until Poster Session

Poster Session / Vendor Talks

4:30 – 6:00 PM Poster Session / Wine & Cheese Reception (Bay Room)

4:30 – 5:15 PM Vendor short talks (concurrent with poster session)

4:30 – 4:45 PM

Advances in Single Particle Data Acquisition

Bart van Knippenberg, PhD. Thermo Fisher Scientific Eindhoven, The Netherlands


8

4:45 – 5:00 PM

Igneous: A software solution for backup and archive of CryoEM data Adam Marko, PhD Scientific Solutions Lead Igneous, Inc., Seattle, WA, USA

5:00 – 5:15 PM

Development of an ideal electron counting detector for cryoEM at 100 kV Sacha De Carlo, Ph.D. Business Development Manager EM Dectris Ltd., Baden, SWITZERLAND

6:00 – 7:30 PM Dinner (Granhall)

Session 3: Deep Learning (Mountain Lake)

7:30 – 7:45 PM Introduction, José-María Carazo Garcia, Session Chair Professor, Biocomputing Unit Centro Nacional de Biotecnología, Madrid, SPAIN

7:45 – 8:10 PM CryoDRGN: A Tool for Reconstructing Highly Heterogeneous Structural Ensembles from Cryo-Electron Micrographs Joseph Davis Assistant Professor of Biology, Whitehead Career Development, Massachusetts Institute of Technology, Cambridge, MA, USA

8:10 – 8:35 PM Predicting Backbone Atomic Structure from High Resolution Cryo-EM Density Maps of Protein Complexes

Dong Si Assistant Professor, Data Analysis & Intelligent Systems University of Washington, Bothell, WA, USA


9

8:35 – 9:00 PM Methods For Increasing The Confidence And Completeness Of CryoEM Particle Picks Alex Noble Research Scientist, New York Structural Biology Center, New York, NY, USA

9:00 – 9:25 PM

9:30 – 11:00 PM

Protein 3D Structure Modeling From Medium-Resolution Cryo-EM Density Maps

Daisuke Kihara Professor, Biological Sciences

Purdue University, Lafayette, IN

Social Hour/Cash bar (Cedar House)

FRIDAY, MARCH 20TH

7:00 – 8:00 AM Breakfast Buffet (Granhall)

Session 4: Map and Model Improvement and Validation (Mountain Lake)

8:00 – 8:15 AM Introduction, Niels Volkmann, Session Chair Professor, Immunity and Pathogenesis Sanford Burnham Prebys Medical Discovery Institute, La Jolla, CA USA

8:15 – 8:40 AM

Validating Models and Assessing Map Quality by Density Sampling

James Fraser Associate Professor, Department of Bioengineering and Therapeutic Sciences, UCSF, San Francisco, CA, USA

8:40 – 9:05 AM Validation at 3 to 4Å -- What Works and What Doesn't Jane Richardson James B. Duke Professor of Medicine Professor of Biochemistry Duke University School of Medicine, Durham, NC, USA


10

9:05 – 9:30 AM Domain Contact Prediction as a Tool for CryoEm Map Interpretation Frank DiMaio Assistant Professor, Department of Biochemistry University of Washington, Seattle, WA, USA

9:30 – 9:55 AM Improvement and Validation of Atomic Models of Cryo-EM Reconstructions using Coot and Friends Paul Emsley

Investigator, Molecular Biology

MRC Laboratory of Molecular Biology

Cambridge, England, UK

9:55 – 10:15 AM Group Photo & Coffee/Tea Break

Session 5: Automation in CryoEM/ET (Mountain Lake)

10:15 –10:30 AM Introduction, Masahide Kikkawa, Session Chair Professor of Cell Biology and Anatomy, Graduate School of Medicine, The University of Tokyo, JAPAN

10:30 –10:55 AM SerialEM and IMOD for Automated Acquisition and Reconstruction of Cryo-tomograms David Mastronarde Professor, Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, USA

10:55 –11:20 AM Fast Tomography Grant Jensen Professor of Biophysics and Biology; Investigator, Division of Biology and Biological Engineering, Caltech, Howard Hughes Medical Institute, Pasadena, CA, USA

11:20 –11:45 AM What is Next for Automation? Anchi Cheng Senior Scientist, New York Structural Biology Center, New York, NY USA


11

11:45 –12:10 PM

12:15 – 1:15 PM

Automation and Integration of Image Processing Workflows for Electron Tomography Carlos Oscar Sanchez Sorzano Professor, Biocomputing Unit, Centro Nacional Biotecnología -CSIC, Madrid, SPAIN

Lunch Break (Granhall) Afternoon free until Poster Session

Poster Session

3:30 – 4:30 PM Poster Session/ Wine & Cheese Reception (Bay Room)

Round Table Discussion 1: Data Compression and Archival in CryoEM

Round table participants may optionally give a brief (5 minutes, 2-3 slides) opinion statement on

some aspect of this problem.

4:30 – 5:15 PM Steven Ludtke, Moderator Professor, Department of Biochemistry and Molecular Biology, Director, CryoEM/ET Core Baylor College of Medicine, Houston, TX, USA

Roundtable Participants:

Clint Potter EM Co-Directorr National Resource for Automated Molecular Microscopy National Center for CryoEM Access and Training Simons Electron Microscopy Center New York Structural Biology Center, NY, USA

José-María Carazo Garcia Professor, Biocomputing Unit Centro Nacional de Biotecnología, Madrid, SPAIN

David Mastronarde Professor, Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, USA

Graig Yoshioka Research Assistant Professor, Biomedical Engineering Co-director PNCC Oregon Health & Science University, Portland, OR, USA


12

Wen Jiang Professor, Biological Sciences, Director of Purdue Cryo-EM Facility,

Markey Center for Structural Biology, Department of Biological Sciences, Purdue University, West Lafayette, IN, USA Gerard Kleywegt Professor, Structural Molecular Biology EMBL-European Bioinformatics Institute, Cambridgeshire, UK

Round Table Discussion 2: Data Acquisition Workflows and Best Practices

5:15 – 6:00 PM Dorit Hanein, Moderator Professor, Immunity and Pathogenesis, Sanford Burnham Prebys Medical Discovery Institute, La Jolla, CA, USA

Roundtable Participants:

Wah Chiu Photon Science Directorate, Professor, Bioengineering, Microbiology & Immunology, Stanford University, Stanford, CA, USA

Alex Noble Research Scientist, New York Structural Biology Center, New York, NY, USA

Michael A. Cianfrocco Research Assistant Professor, Department of Biological Chemistry, Life Sciences Institute, University of Michigan, Ann Arbor, MI, USA

Grant Jensen Professor of Biophysics and Biology; Investigator, Division of Biology and Biological Engineering, Caltech, Howard Hughes Medical Institute, Pasadena, CA, USA

6:00 – 7:30 PM Gala Dinner (Granhall)


13

Session 6: Short Talks / Poster Lectures (Mountain Lake)

Selected poster talks, 10 min + 5 min for Q&A

7:30 – 9:30 PM Dorit Hanein, Program Chair Professor, Immunity and Pathogenesis, Sanford Burnham Prebys Medical Discovery Institute, La Jolla, CA, USA

Poster Selection Committee:

Kenneth Taylor Professor, Department of Biological Sciences, Florida State University, Tallahassee, FL, USA

Tamir Gonen Professor of Biological Chemistry and Physiology; Investigator, Howard Hughes Medical Institute, David Geffen Medical School, UCLA, CA, USA

David DeRosier Professor Emeritus of Life Sciences, Department of Biology, National Academy of Sciences, Brandeis University, Waltham, MA, USA

End of Scientific Program

9:30 –11:00 PM Social Hour / Cash Bar (Cedar House)

SATURDAY, MARCH 21TH

7:30 – 9:00 AM Breakfast Buffet (Granhall) Check out / Departures


14

Speaker Abstracts Continuous Heterogeneity Using Spectral Volumes Joakim Andén, Research Scientist, Numerical Analysis Center for Computational Mathematics Flatiron Institute, New York, NY, USA Alignment of Tilt Series Daniel Castaño, Senior Scientist, Center for Cellular Imaging and Nanoanalytics at Biozentrum, University of Basel, Basel, SWITZERLAND A Complete Data Processing Workflow for CryoET and Subtomogram Averaging Muyuan Chen, Instructor, Biochemistry, Baylor Collage of Medicine, Houston, TX, USA What is Next for Automation? Anchi Cheng, Senior Scientist, New York Structural Biology Center, New York, NY USA CryoDRGN: A Tool for Reconstructing Highly Heterogeneous Structural Ensembles from Cryo-Electron Micrographs

Joseph Davis, Assistant Professor of Biology, Whitehead Career Development, Massachusetts Institute of Technology, Cambridge, MA, USA Domain Contact Prediction as a Tool for CryoEm Map Interpretation Frank DiMaio, Assistant Professor, Department of Biochemistry; University of Washington, Seattle, WA, USA Helical Polymers: A Personal Perspective Edward H. Egelman, Professor, Department of Biochemistry and Molecular Genetics, University of Virginia, Charlottesville, VA, USA Improvement and Validation of Atomic Models of Cryo-EM Reconstructions using Coot and Friends Paul Emsley, Investigator, Molecular Biology ,MRC Laboratory of Molecular Biology Cambridge, England, UK

Validating Models and Assessing Map Quality by Density Sampling James Fraser, Associate Professor, Department of Bioengineering and Therapeutic Sciences; University of California, San Francisco, San Francisco, CA, USA

Incorporating Tilted Image Processing Into cisTEM Benjamin Himes, Postdoctoral Associate, Howard Hughes Medical Institute, Janelia Research Campus, Ashburn, VA, USA Fast Tomography Grant Jensen, Professor of Biophysics and Biology; Investigator, Division of Biology and Biological Engineering, Caltech, Howard Hughes Medical Institute, Pasadena, CA, USA Non-Classification-Based Methods for Deciphering Continuous Conformational Variability: Methods Combining Image Analysis, Molecular Mechanics Simulation, and Machine Learning


15

Slavica Jonic, Research Director, CNRS Impmc-cnrs umr 7590, Sorbonne University, Paris, FRANCE Protein 3D Structure Modeling From Medium-Resolution Cryo-EM Density Maps Daisuke Kihara, Professor, Biological Sciences, Purdue University, Lafayette, IN SerialEM and IMOD for Automated Acquisition and Reconstruction of Cryo-tomograms David Mastronarde, Professor, Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, USA Methods For Increasing The Confidence And Completeness Of CryoEM Particle Picks Alex Noble, Research Scientist, New York Structural Biology Center, New York, NY, USA “Structure” in Structural Biology Abbas Ourmazd, Distinguished Professor of Physics, University of Wisconsin, Milwaukee, WI, USA Validation at 3 to 4Å -- What Works and What Doesn't Jane Richardson, James B. Duke Professor of Medicine, Professor of Biochemistry Duke University School of Medicine, Durham, NC, USA

Predicting Backbone Atomic Structure from High Resolution Cryo-EM Density Maps of Protein Complexes

Dong Si, Assistant Professor, Data Analysis & Intelligent Systems; University of Washington, Bothell, WA, USA

Automation and Integration of Image Processing Workflows for Electron Tomography Carlos Oscar Sanchez Sorzano, Professor, Biocomputing Unit, Centro Nacional Biotecnología -CSIC, Madrid, SPAIN Reconstruction of Average Subtracted Tubular Regions (RASTR) Enables Structure Determination of Tubular Filaments by Cryo-EM Scott Stagg. Associate Professor, Department of Chemistry, Florida State University, Tallahassee FL, USA

What is the resolution limit in single particle cryo-EM? Holger Stark, Professor, Department of Structural Dynamics; Max-Planck Institute for Biophysical Chemistry, University of Göttingen, Göttingen GERMANY

Pushing Water Uphill with a Rake: Resolving High-Resolution Structure of the Nuclear Pore Complex with Subtomogram Averaging Beata Turoňová, Staff Scientist, Structural and Computational Biology, European Molecular Biology Laboratory, Heidelberg, GERMANY


16

Speakers Abstracts

Continuous Heterogeneity Using Spectral Volumes

Joakim Andén Research Scientist, Numerical Analysis

Center for Computational Mathematics, Flatiron Institute, New York, NY, USA

The set of 3D configurations exhibited by a biological macromolecule constitutes a low-dimensional manifold in the space of density maps. We propose a new method for estimating this manifold using a three-step process. We first estimate the linear subspace which captures most of the 3D variability in the dataset. By restricting maps to this subspace, we may then create low-resolution 3D reconstructions from each image. These are in turn used to construct a graph Laplacian over the set of images, whose eigenvectors are used to characterize the underlying low-dimensional manifold. These eigenvectors, known as “spectral volumes,” may then be used to study the topology of the manifold, but also study the principal modes of variability and create higher-resolution 3D reconstructions. To demonstrate this method, we provide results on both synthetic and experimental data.


17

Alignment Of Tilt Series

Daniel Castaño University of Basel, Biozentrum

Basel Switzerland

Modern cryo-electron tomography projects may involve the alignment of several hundreds of tilt series. Robust automated methods for aligning them are thus becoming an indispensable tool for tomography practitioners. The tomography software package Dynamo incorporates now a new tool for fiducial-based alignment of tilt series. It bases on exhaustively testing projection models against the set of located gold beads, and it includes specific mechanisms to cope with characteristic pitfalls that arise in the processing of real data sets (sparseness, clustering or irregular spatial distribution of markers, heterogeneity of markers and marker quality across the tilt series). The performance of the package has been tested in fully automated modus against all public data sets deposed in the EMPIAR database, as well as on several hundreds in-house data sets, systematically providing good balances between automation and alignment quality. In addition to the capacities for unsupervised processing of large data test, the software includes an extensive repertory of tools for visual, interactive operation on individual tilt series.


18

Complete Data Processing Workflow for CryoET and Subtomogram Averaging

Muyuan Chen1, James M. Bell, Xiaodong Shi, Stella Y. Sun, Zhao Wang,

Steven J. Ludtke 1Instructor, Biochemistry, Baylor Collage of Medicine, Houston, TX, USA

Electron cryotomography (CryoET) is currently the only method capable of visualizing cells in 3D at nanometer resolutions. While modern instruments produce massive amounts of tomography data containing extremely rich structural information, it takes tremendous effort to process those data and reach biological findings. Here we present an integrated workflow that covers the entire tomography data processing pipeline, from automated tilt series alignment and particle selection to subnanometer resolution subtomogram averaging. Resolution enhancement is made possible through the use of per-particle-per-tilt CTF correction and orientation refinement. This workflow greatly reduces human effort and increases throughput of CryoET data processing, and is capable of determining protein structures at state-of-the-art resolutions for both purified macromolecules and cells.


19

What is Next for Automation?

Anchi Cheng New York Structural Biology Center, New York, NY USA

A high-degree of data-collection and pre-processing automation has become a reality in recent years for both single-particle analysis (SPA) and tomography cryo-EM. As demand for human effort is relieved, demand on instrumentation increases.

But what stops us from routinely collect 10000 SPA reconstruction-worthy micrographs in a 24 hour session that includes grid screening? I will discuss the hardware and software limitations, as well as physical and psychological factors that affect the realization of this goal.


20

CryoDRGN: a tool for reconstructing highly heterogeneous structural ensembles from cryo-electron micrographs

Ellen D. Zhong, Tristan Bepler, Bonnie Berger, Joseph H. Davis

Massachusetts Institute of Technology, Cambridge, MA, USA

In performing their essential functions, multi-subunit macromolecular complexes often undergo large conformational changes. Some such complexes exhibit seemingly continuous conformational changes, whereas others transition between disparate states in a highly cooperative manner such that conformational intermediates are sparely populated. To understand the extent of complex heterogeneity, the degree of cooperativity in their conformational changes, and to estimate ensembles of such structures, we have developed a neural-network based single particle analysis framework. This approach, named cryoDRGN, maps individual particles to a low-dimensional latent space, which we interpret as a conformational energy landscape. Direct inspection of this latent space provides insights into the degree of structural heterogeneity in the dataset, provides estimates of particle abundance in each state, and relates the observed states to one another. Additionally, we demonstrate that three-dimensional structures can be sampled from this latent space, allowing users to visualize conformational trajectories sampled along the data manifold. In this talk, we detail our method and apply it to exemplar simulated and experimental datasets to illustrate its utility in analyzing both continuously and discretely heterogeneous complexes. We find that these neural networks accurately estimate the conformational landscape of simulated datasets, reveal significant discrete and continuous conformational heterogeneity in publicly available datasets, and accurately produce ensembles of medium-resolution density maps, which can be used to understand how the complex’s structural changes drive its function.


21

Domain Contact Prediction as a Tool for CryoEm Map Interpretation

Frank DiMaio

Assistant Professor, Department of Biochemistry University of Washington, Seattle, WA, USA

Recent advances in protein contact prediction have enabled robust prediction of residue-residue contacts in proteins, given only a modest number of homologous sequences. Using a deep convolutional neural network to generated contact predictions, followed by minimization in Rosetta, such methods are frequently able to accurately predict domain structures at the "topology" level. We show how this per-domain structure prediction (as implemented in trRosetta) may be used in combination with a novel algorithm for EM-guided model selection and domain assembly. This combined approach allows structure determination from medium-resolution cryoEM reconstructions when homologous structures are not available.


22

Helical Polymers: A Personal Perspective

Edward H. Egelman Professor, Department of Biochemistry and Molecular Genetics,

University of Virginia, Charlottesville, VA, USA

It was over fifty years ago that the first 3D reconstruction from electron microscopy was published (DeRosier and Klug, 1968), and this happened to be a helical phage tail. It was no accident that this was a helical specimen, and not just because helical polymers are so abundant in biology. From one point of view helical objects are the simplest to reconstruct: a single image of a helical filament may provide all of the information needed for a 3D reconstruction. Fourier-Bessel methods of reconstruction dominated the field for the next 30 years, but now real-space approaches have become the standard. With the advent of direct-electron detectors, reaching a near-atomic resolution for helical polymers has become the standard, rather than the exception. However, problems in determining the correct helical symmetry can persist even when one reaches a resolution of ~ 5 Å. In addition, the highly anisotropic environment present in thin films prior to vitrification, the compressional forces associated with these thin films, and fluid flow with associated shear, can all impact the structure of the filaments being examined. I will discuss these problems and potential solutions while giving an overview of some of our own efforts involving everything from viruses infecting hosts living in nearly boiling acid to microbial nanowires, with stops along the way at bacterial and archaeal flagellar filaments and bacterial and archaeal pili.


23

Improvement and Validation of Atomic Models of Cryo-EM Reconstructions using Coot and Friends

Paul Emsley

Investigator, Molecular Biology MRC Laboratory of Molecular Biology

Cambridge, England, UK

Recent developments, using multi-threading, in Coot have improved the speed and the user interface of the real space refinement. Whereas before, one would hestitate to refine more than 10 or so residues, then new refinement system can refine substantially larger fragments, for example, a domain. Secondly, the refinement now includes interactive analysis of rotamer and ramachandran probabilities using coloured platonic solids and atom overlap markup also. Thirdly and finally, previously Coot spend its time in the main thread between drawing the graphics and refining/optimizing the coordinates - this lead to slow refinement or jerky animation - and often both. Now, the real space refinement has its own thread, so there graphics can be drawn smoothly (at a high frame rate). These technical improvements in animation and validation feeback, have lead to a more pleasant and game-like user experience. The scripting API in Coot allows developers to identify problematic regions of the model, and create a number of alternative hypotheses for atom positions, that might, for example, vary by usage of restraints, restraints weight or (in the case of loop rebuilding) the rank of the fit to density of the candidates. Such a scoring system should reduce the amount of interactivity required to rebuild problematic regions.


24

Validating Models and Assessing Map Quality by Density Sampling

James Fraser Associate Professor, Department of Bioengineering and Therapeutic Sciences;

University of California, San Francisco, CA, USA

Advances in high-resolution cryo-electron microscopy (cryo-EM) require the development of validation metrics to independently assess map and model quality. We will present EMRinger, a tool that assesses the precise fitting of an atomic protein model into the map during refinement, and qPTxM (quantifying Post-Transcriptional Modifications), which validates the presence of assigned rRNA modifications. In addition to assessing the quality of the model, these metrics provide a sensitive readout of map quality by probing the local density around high resolution features. Code and Coot plugins are available.


25

Incorporating Tilted Image Processing Into cisTem

Benjamin Himes Postdoctoral Associate,

Howard Hughes Medical Institute, Janelia Research Campus Ashburn, VA, USA

Images of a tilted specimen are important not only in tomography, but also increasingly in single particle workflows plagued by preferential orientation as well as cutting edge in situ particle detection approaches. I will discuss how we have extended cisTEM to handle the challenges presented by images of tilted specimen: with respect to data processing, our improved aberration estimation and orientation refinement; with respect to image simulation, steps we have taken to accurately an efficiently simulate tilts, including mutlislice wave propagation at angles non-coincident with the z-axis taking advantage of GPU acceleration.


26

Fast Tomography

Grant Jensen Caltech, HHMI

Pasadena, CA, USA The main limitation for many electron cryo-tomography (cryo-ET) projects today is low

throughput. Recently, we characterized a "high precision" single-axis-tilt cryoholder (Thermo Fisher) for the Titan Krios microscope that uses static arms to improve

eucentricity. As a result, we can skip the sample tracking step and collect tilt series in just a couple minutes. The method has been named Fast Incremental Single Exposure (FISE). Using beam shifts to record images of multiple targets at each tilt angle should further reduce acquisition times. All this progress calls for new software development to automatically identify targets on grids, automatically align tilt-series, automatically pick

particles from tomograms, and automatically calculate sub-tomogram averages.


27

Non-Classification-Based Methods for Deciphering Continuous Conformational

Variability: Methods Combining Image Analysis, Molecular Mechanics Simulation, And Machine Learning

Slavica Jonic

Research Director, CNRS Impmc-cnrs umr 7590 Sorbonne University, Paris, FRANCE

IMPMC - UMR 7590, Sorbonne Université/CNRS/MNHN, Paris, France Single-particle cryo electron microscopy (cryo-EM) has become comparable to X-ray crystallography with regards to the obtainable resolution of structures of biomolecular complexes, which are now increasingly determined at near-atomic resolution. To achieve such high resolutions, the classical approach is to classify a large number of cryo-EM images of complexes (particles), collected at random (unknown) orientations within a thin layer of vitreous ice, into an initially set number of classes (2D or 3D) and remove the particles not contributing to the highest-resolution class averages (i.e., keep only the particles with the most consistent views and conformations). Such “selection” of particles may obscure the information on a possibly larger conformational variability. More precisely, some conformational states may be thrown away blindly instead of being clearly elucidated, whereas the characterization of different coexisting conformations could be essential for understanding how the complexes function. Classification-based image analysis methods are suited for studying conformational changes with countable intermediate states and are referred to as discrete-state methods. The development of non-classification-based image analysis methods, assuming uncountable intermediate states, is in progress. These methods aim at describing continuous conformational changes by visualizing the full distribution of states in a low-dimensional space (usually 2D or 3D space) and are referred to as continuous-state methods. My team works on the development of continuous-state methods based on combining image analysis, molecular mechanics simulation, and machine learning methods. We have been working on the development of such methods for single particle cryo-EM and, since recently, we also work on such methods development for in vitro and in situ cryo electron tomography (cryo-ET). In this talk, I will describe the current progress and future prospects of these works.


28

Protein 3D Structure Modeling From Medium-Resolution Cryo-EM Density Maps

Daisuke Kihara Professor, Biological Sciences

Purdue University, Lafayette, IN

The significant progress of the cryo-electron microscopy (cryo-EM) poses a pressing need for software for structural interpretation of EM maps. Particularly, protein structure modeling tools are needed for EM maps of around 4 A resolution or worse, where building a main-chain structure is challenging. Our group is developing methods for protein structure modeling by applying various algorithms, including deep learning. In this presentation, I focus two of such methods from our lab. We have developed a de novo modeling tool named MAINMAST (MAINchain Model trAcing from Spanning Tree) for EM maps of up to about 4 A resolution. MAINMAST builds main-chain traces of a protein in an EM map from a minimum spanning tree constructed by connecting high-density points. MAINMAST showed better modeling performance than existing methods. The method is further enhanced recently to be able to model and segment symmetric protein complexes and ligand (drug) molecules that bind to a protein in a map. Moreover, to provide structure information for maps determined at even lower resolution (5~10 A), we have recently developed a new tool, Emap2sec, which uses convolutional deep neural network (CNN) for detecting structures of proteins. Emap2sec scans an EM map with a 3D voxel and assigns a type of protein structure class, i.e. alpha helix, beta strand, or coil, from density patterns of the voxel and its neighbors.


29

SerialEM and IMOD for Automated Acquisition and Reconstruction of Cryo-

tomograms

David Mastronarde Professor, Molecular, Cellular and Developmental Biology,

University of Colorado, Boulder, CO, USA In this talk I will describe a pipeline for automated acquisition of cryo-tomograms using SerialEM for acquisition and IMOD for reconstruction. SerialEM can acquire tilt series in a batch fashion from multiple, previously mapped locations. The series can be acquired either by the built-in tilt series controller, which can now do dose-symmetric series, or with scripts for more specialized situations. Movie frames from direct detector cameras can be either aligned during acquisition or saved to frame stacks. In the latter case, the IMOD program Framewatcher can run the alignment for all the frame stacks when the series completes, producing a single tilt series file ready for reconstruction. The program Serieswatcher can detect a completed tilt series produced by either means and launch the batch reconstruction program Batchruntomo, using reconstruction parameters that were set in the Etomo user interface. This interface allows virtually complete control over the parameters, the type of alignment to use, and the particular steps to be run. This pipeline works best for automated processing of routine data sets where proper parameters are known in advance; it would also be useful to get coarse reconstructions quickly for monitoring and initial screening.


30

Methods For Increasing The Confidence And Completeness Of CryoEM Particle Picks

Alex Noble

Research Scientist New York Structural Biology Center, New York, NY, USA

Small, non-globular, and asymmetric cryoEM particles often prove difficult or sometimes impossible to sufficiently pick due to incomplete picking and an overabundance of false positives. We introduce two complementary convolutional neural network (CNN) methods for increasing the confidence and completeness of cryoEM particle picks. First is Topaz-Denoise, a Noise2Noise-based denoiser which includes pre-trained denoising models, allowing for high-confidence initial manual identification of small, low-density particle views. Second is Topaz picker, a novel positive-unlabeled CNN requiring only sparse, high-confidence training picks, allowing for picking of many more real particles of any shape and size while picking very few false positives. As a package, Topaz and Topaz-Denoise substantially decrease the difficulty in picking conventionally difficult particles. Topaz is free, open-source, comes with a GUI, and is integrated into CryoSparc and Appion.


31

“Structure” in Structural Biology

Abbas Ourmazd Distinguished Professor of Physics

University of Wisconsin Milwaukee, WI, USA It has been famously said that the limits of one’s language define the limits of one’s world. If so, a key question facing structural biology is this: What do we mean by “structure”? The primary reason for studying biomolecular structure is that it sheds light on biological function. But function involves changes in structure. Until recently, we were content to believe structural changes involve “jumps” between a small number of discrete structures separated by high energy barriers. There is mounting evidence that the barriers are not high - at most a few times the thermal energy available under physiological conditions. This means many “structures” coexist, each with a sizable probability. So many, in fact, that the concept of one, or even a few discrete structures is inadequate, if not outright misleading. I will discuss means for determining structural continua from random cryo-EM snapshots.


32

Validation at 3 to 4Å -- What Works and What Doesn't

Jane Richardson James B. Duke Professor of Medicine, Professor of Biochemistry

Duke University School of Medicine, Durham, NC, USA Much significant biology can be discovered about organization, functional relationships, and conformational states from the exciting new cryoEM structures at 3-4A resolution. But tasks such as understanding chemical mechanism and designing drugs or other interaction partners require a finer level of detail than routinely achievable. At ≤2.5A, model validation has proven effective both at identifying local regions where detail is unreliable and also at providing guidance for corrections that improve local accuracy. At present, however, it has turned out that those model criteria are invalidated at 3-4Å because well-behaved, convergent refinement requires using most of them as explicit refinement targets. We and others are pushing to develop new validation criteria that can meet this challenge, by using independent information and covering wider than a single-residue span. CaBLAM, which identifies incorrectly oriented peptides, is the first example with proven effectiveness and community acceptance, and others are in the pipeline.


33

Predicting Backbone Atomic Structure from High Resolution Cryo-EM Density Maps of Protein Complexes

Dong Si Data Analysis & Intelligent Systems

University of Washington, Bothell, WA, USA

Accurately determining the atomic structure of proteins represents a fundamental problem

in the field of structural bioinformatics. A solution would be significant as protein structure

information could be utilized in the medical field, e.g. in the development of vaccines for

new viruses. This talk focuses on predicting the protein structure based on 3D images of

the proteins captured through cryo-electron microscopes (cryo-EM). A fully automated

computationally efficient protein structure prediction method would be particularly

beneficial in the field of cryo-EM as the technology allows researchers to photograph

multiple large protein complexes in a single study, which means that a fast prediction

method could allow for a high throughput of derived protein structures.

We present a deep learning approach, DeepTracer, for predicting locations of the backbone atoms, secondary structure elements, and the amino acid information. We noted that the prediction runtime of DeepTracer is significantly improved compared to other methods. It predicts a large protein complex structure of more than 30,000 amino acids in only 2 hours.


34

Automation and integration of image processing workflows for Electron Tomography

Carlos Oscar Sanchez Sorzano

Centro Nacional Biotecnología -CSIC, Madrid, SPAIN

Scipion is a workflow engine specifically developed for Image Processing in CryoEM. It is a well established for Single Particle Analysis and it is now being extended for image processing in Electron Tomography. In this talk we will present its main features and the tools currently available for the analysis of this kind of data.


35

Reconstruction of Average Subtracted Tubular Regions (RASTR) Enables Structure Determination of Tubular Filaments by Cryo-EM

Dr. Peter S. Randolph1 and Dr. Scott M. Stagg1,2

1 Institute of Molecular Biophysics, Florida State University, Tallahassee, FL 32306,USA 2 Department of Chemistry and Biochemistry, Florida State University, Tallahassee, FL

32306, USA Correspondence to Scott M. Stagg: Institute of Molecular Biophysics, Florida State

University , 91 Chieftan Way, Tallahassee, FL 32306, USA, [email protected] As the field of electron microscopy advances, the increasing complexity of samples being produced demand more involved processing methods. In this study, we have developed a new processing method for generating 3D reconstructions of tubular structures. Tubular biomolecules are common throughout many cellular processes and are appealing targets for biophysical research. Processing of tubules with helical symmetry is relatively straightforward for electron microscopy if the helical parameters are known, but tubular structures that deviate from helical symmetry (asymmetrical components, local but no global order, etc.) present myriad issues. Here we present a new processing technique called Reconstruction of Average Subtracted Tubular Regions (RASTR), which was developed to reconstruct tubular structures without applying symmetry. We explain the RASTR approach and quantify its performance using three examples: a simulated symmetrical tubular filament, a symmetrical tubular filament from cryo-EM data, and a membrane tubule coated with locally ordered but not globally ordered proteins.


36

What is the resolution limit in single particle cryo-EM?

Holger Stark Max-Planck Institute

University of Göttingen, Göttingen GERMANY

The resolution in single particle cryo-EM has been constantly improved over the last couple of years and has now reached a level of well below 2 Angstrom. This is due to the massive improvement in electron microscopic instrumentation including direct detectors but also due to image processing software that allows movie frame alignments, powerful 2D/3D classification, accurate CTF estimation and beam tilt and Ewald sphere corrections. It therefore becomes interesting to discuss the resolution limits within this framework of available hardware and software. We have recently installed a new Titan Krios electron microscope which is equipped with a monochromator and a next generation spherical aberration corrector (BCOR). The contribution and usefulness of this equipment within the context of high-resolution structure determination will be discussed in the presentation.


37

Pushing Water Uphill with a Rake: Resolving High-Resolution Structure of the Nuclear Pore Complex with Subtomogram Averaging

Beata Turoňová

Staff Scientist, Structural and Computational Biology European Molecular Biology Laboratory, Heidelberg, GERMANY

Cryo Electron Tomography (cryoET) allows us to study the structure of large macromolecular complexes in their close to native environment. In combination with subtomogram averaging (SA) we can now resolve structures in great detail. While few structures have already reached sub-nanometer resolution, most of the complexes studied with cryoET and SA remain at medium to low resolution (10-30A). This is also true for the human nuclear pore complex (NPC) with the best reported resolution of ~20A. The NPC is a very challenging complex for both tomography and SA. First, the inherent flexibility of larger subcomplexes, membrane embedded nature of NPC, and its complex topology have prevented its entire reconstitution in vitro causing the image acquisition to be performed on samples with isolated nuclear envelopes. These samples are thus very thick and result in extremely low signal to noise ratio which complicates all subsequent image processing. Second, the NPC is a very large complex with ~120 nm diameter and has relatively low abundance within the tomograms, which severely complicates reliable subtomogram alignment. In order to improve the resolution on the NPC, we try to tackle both of these problems. We are currently working on improving and optimizing our tomographic workflow to obtain high-quality tomograms in reasonable time and using the geometric properties of NPCs we develop new routines for high-resolution structure determination.


38

Poster Abstracts

LAFTER and SIDESPLITTER - Approaches for Handling Conformational Flexibility and Overfitting

Christopher Aylett

Imperial College London London, United Kingdom

Single particle analysis entails the reconstruction of high-resolution volumes from tens of thousands of particle images with low individual signal-to-noise. Imperfections in this process result in substantial variations in the local signal-to-noise ratio within the resulting reconstruction, complicating the interpretation of molecular structure. Furthermore, during iterative refinement it is possible to stably incorporate noise into the reconstructed density. Such “overfitting” can lead to misinterpretation of the structure, and thereby flawed biological results. An effective local de-noising filter can therefore improve interpretability and maximize the amount of useful information obtained from cryo-EM maps, while during refinement it can prevent overfitting and improve maps and the final resolution. We present LAFTER and SIDESPLITTER: LAFTER is a local de-noising algorithm based on a pair of serial realspace filters. It compares independent half-set reconstructions to identify and retain shared features that have power greater than the noise. It is capable of recovering features across a wide range of signal-to-noise ratios, and we demonstrate recovery of the strongest features at Fourier shell correlation (FSC) values as low as 0.144 over a 256^3-voxel cube. A fast and computationally efficient implementation of LAFTER is freely available. SIDESPLITTER involves the application of a similar filtering process during refinement through the application of a LAFTER-derived local signal-to-noise filter, which we show to be capable of reducing overfitting in both idealized and experimental settings, while maintaining independence between the two sides of a split refinement. SIDESPLITTER can also improve the final resolution in refinements of structures prone to severe over-fitting, such as membrane proteins in detergent micelles. Finally, we propose a new way to evaluate the effectiveness of realspace filters for noise suppression, based on the correspondence between two FSC curves: 1) the FSC between the filtered and unfiltered volumes, and 2) Cref, the FSC between the unfiltered volume and a hypothetical noiseless volume, which can readily be estimated from the FSC between two half-set reconstructions.


39

Non-Uniformity of Projection Distributions Attenuates Resolution in Cryo-EM

Philip Baldwin Salk Institute La Jolla, CA

Virtually every single-particle cryo-EM experiment currently suffers from specimen adherence to the air-water interface, leading to a non-uniform distribution in the set of projection views. Non-uniform (anisotropic) distributions can negatively affect map quality, elongate structural features, and in some cases, prohibit interpretation altogether. Although some consequences of non-uniform sampling have been described qualitatively, we know little about how sampling quantitatively affects resolution in cryo-EM. Here, we show how inhomogeneity in any projection distribution scheme attenuates the global Fourier Shell Correlation (FSC) in relation to the number of particles and a single geometrical parameter, which we term the sampling compensation factor (SCF). The reciprocal of the SCF is defined as the average over Fourier shells of the reciprocal of the per-particle sampling and normalized to unity for uniform distributions. The SCF ranges from one to zero, with values close to the latter implying large regions of poorly sampled or completely missing data in Fourier space. Using two synthetic test cases, influenza hemagglutinin and human apoferritin, we demonstrate how any amount of sampling inhomogeneity always attenuates the FSC compared to a uniform distribution.

Structural Organization of the C1b supercomplex within the Ciliary Central Apparatus

Kai Cai Ut Southwestern Medical Center

Dallas, Tx ‘9+2’ motile cilia contain 9 doublet microtubules and a central apparatus (CA) composed of two singlet microtubules with associated projections. CA plays crucial roles in regulating ciliary motility. Defects in CA assembly usually result in motility-impaired or paralyzed cilia, which in humans causes disease. Despite their importance, the protein composition and functions of the CA projections are largely unknown. Here, we combined biochemical and genetic approaches with cryo-electron tomography and subtomogram averaging to compare the CA of wild-type Chlamydomonas with CA mutant. Our results show that two conserved proteins FAP42 and FAP246 are localized to the c1b projection of CA. Loss of the above proteins leads to destabilization of the structure of C1b projection and impaired ciliary motility. These data provide insight into the subunit organization and 3D structure of the CA, which is critical for understanding the molecular mechanisms by which the CA regulates ciliary beating.


40

2SDR-Pre: a Fast Pre-Processor for Single-Particle Cryo-EM through Enhancing 2D Clustering

Szu-Chi Chunga, Hsin-Hung Linb, Po-Yao Niua, Shih-Hsin Huangb, I-Ping Tua* and Wei-Hau

Changb*

aInstitute of Statistical Science, Academia Sinica, Taiwan bInstitute of Chemistry, Academia Sinica, Taiwan

128 Academia Road, Section 2, Nankang, Taipei 11529, Taiwan

Abstract 2D clustering plays a pivotal role in cryo-electron microscopy (cryo-EM) image analysis. It aims to group images of similar view and sift a homogeneous subset in silico to suit high-resolution 3D structure determination. Due largely to the presence of heavy noise, the clustering results are often non-ideal while the converging process is further burdened by rapid increase in the number and size of images. Here, we introduce a fast and loss-less pre-processing strategy, based on a novel dimension-reduction method (2SDR), to initialize 2D clustering. By implementing this 2SDR pre-processor prior to representative clustering algorithms, RELION and ISAC, we compare the performances with and without the pre-processor. Tests using multiple cryo-EM experimental datasets reveal the pre-processor gives benefits of saving the time on clustering, increasing the yield of particles, and improving the quality of classes. Remarkably, test with a huge dataset of 80S ribosome demonstrates the cost of the preprocessing is low while re-sifting a TRPV1 ion-channel dataset with the pre-processor has lifted the overall resolution of 3D structure by up to 0.2 Angstrom with concomitant enhancement in the interpretability of map. Our findings suggest the 2SDR pre-processor, with light computation, is applicable for boosting the performance of 2D clustering algorithms.


41

Interactions of AAV and its Cell Receptor; Holistic Multi-resolution

Cryo-EM/ET for Flexible Complexes

Mark Silveria1, Edward Large1, Nancy Meyer1,2, Guiqing Hu3, Grant Zane1, Qing Xie2,

Omar Davulcu2, Scott Stagg3 & Michael S. Chapman1,2

1Dept. Biochem., Univ., Missouri, Columbia, MO 65211;

2Dept. Biochem. & Mol. Biol., Oregon Health & Sci. Univ., Portland, OR 97239;

2Inst. Mol. Biophys., Florida State Univ., Tallahasee, FL 32306.

Adeno-associated virus (AAV) has become the vector of choice in early gene therapies

approved by the FDA. A molecular understanding of its host interactions is needed for more

specific targeting and immune evasion. Earlier attempts at structure cast doubt on the reported

identity of cell receptors, leading us to full genome screening and identification of a hitherto

uncharacterized membrane protein that is essential for cell entry and trafficking to the nucleus.

Interaction domains were narrowed down by x-linking/mass-spectroscopy, competitive

inhibition and mutation. Cryo-Electron Tomography (ET), at ~3nm resolution, revealed several

configurations of a flexible chain of 5 polycystic kidney disease (PKD) receptor domains

anchored to AAV2 at one end. A complex with the same fusion protein revealed, by single

particle analysis and sub-volume reconstruction, a single domain at ~1nm resolution. A 2-

domain AAVR construct eliminated flexibility and revealed (now at 2.5 Å resolution) the tightly

bound domain to be PKD2.

Our earlier studies had revealed differences between AAVs, with measurable binding to AAV5

for only PKD1. Its structure, at ~2.5 Å resolution, shows PKD1 bound at an AAV5 site different

from the PKD2-AAV2 site. Our EM structures provide a molecular rationale for differences in

receptor binding among AAVs, and lead to re-appraisal of the mechanisms of antibody

neutralization.

In characterizing interactions between flexible partners, our work highlights the benefits of

holistic analyses, in terms of biological sources, expression constructs and different EM

techniques, together providing atomic detail for molecular interactions with biologically relevant

overview. We also note that our atomic refinement program (RSRef) fits 3D atoms (rather than

atom centers) to the EM reconstruction. This allows refinement of additional parameters - for

local molecular flexibility, ligand occupancy, EM magnification, and a low-pass filter describing

the effective resolution of an atomic model. For well-refined EM structures, “model resolution”

is consistent with FSC, and it indicates the detail at which a model can reasonably be

interpreted.


42

Revealing The Polarity Of Actin Filaments In Intact Cells

Bruno Martins1, Simona Sorrentino1, Wen-Lu Chung1, Meltem Tatli1, Ohad Medalia1, and Matthias Eibauer1

1 Department of Biochemistry, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland

The actin cytoskeleton plays a fundamental role in numerous cellular processes, such as cell motility, cytokinesis, and adhesion to the extracellular matrix. Revealing the polarity of individual actin filaments at their native sites in the cell, would foster an unprecedented understanding of these processes and the associated mechanical forces. Cryo-electron tomography provides the means for high-resolution structural imaging of cells. However, the low signal-to-noise ratio of cryo-tomograms obscures the high frequencies and therefore the polarity of actin filaments cannot be directly measured. Here, we developed the ActinPolarityToolbox that enables to determine the polarity of each actin filament in cellular cryo-tomograms. We applied this method to reveal the actin polarity distribution in focal adhesions, and show a linear relation between actin polarity and distance from the apical boundary of the adhesion site. Our method facilitates the reconstruction of the actin cytoskeleton at sub-nanometric precision.

Assessing the Impact of Bit-Reduction Compression on Single Particle Refinement

Adam Fluty

Baylor College Of Medicine Houston, TX, USA

Due to increasing detector sizes and data collection rates, the size of CryoEM datasets have been rapidly growing. As a result, the archival of CryoEM data has become a prominent issue in the field. Counting-mode detectors have enabled the use of lossless TIFF compression on movie-mode data, yet the composite image of each movie is still often stored as a 32 bit floating point file. This practice retains numerical precision far beyond the level of uncertainty for the pixel values, implying that bit reduction may be a viable method to reduce data size without any information loss. In order to test this approach, we picked a data-limited beta-galactosidase dataset because of its particular sensitivity to information loss. We reduced the bits of each image by adjusting the individual pixel values to nearby discrete values that were calculated based on a tunable number of standard deviations from the mean; the images were then losslessly compressed to reduce the file size further. We find that 5 bits is a safe level for virtually all single particle reconstruction projects and that, in many cases, as few as 2 or 3 bits may be enough to preserve reconstruction quality. We also find that reduction with lossless compression

outperforms lossy JPEG compression at similar compression levels.


43

CryoEM Structure Determination of Alpha-synuclein interacting with Tau

Alimohammad Hojjatian1, Anvesh KR Dasari2, Dianne W. Taylor1, Nadia Daneshparvar1, Fatemeh Abbasi Yeganeh1, Kwang Hun Lim2, Kenneth A. Taylor1

1Institute of Molecular Biophysics, Florida State University, Tallahassee, FL, USA

2Department of Chemistry, East Carolina University, Grenville, NC 27858, USA

With the resolution revolution in cryo-EM structure determination, near atomic resolution of 3Dreconstruction became almost routinely possible opening the way for a “future which is not crystallized”. Due to the low Signal to Noise Ratio of ~3% in electron micrographs of frozen hydrated biological samples, a vast number of particles must be aligned iteratively, starting from low resolution frequencies, in order to produce a near atomic resolution 2D or 3D reconstruction while avoiding overfitting or incorrect fitting. The presence of symmetry can be of enormous significance, as it can basically limit the alignment to a portion of the molecule with the rest replicated using symmetry. Helical symmetry, for example, a Helix can normally be described with three parameters: helical rise, helical twist and handedness of the helix. Helical symmetry was of enormous help in solving the first high resolution structure of a biological sample by cryoEM using Fourier-Bessel inversion. However, every high-level symmetry comes with great risk of amplification of noise in search of signal that can result in overfitting or wrong structure determination. a-synuclein fibrils, as amyloids, are of great pathological significance in so-called synucleinopathies but their structural features are very difficult to reconstruct in 3D. Amyloid fibril structure is generally described as long pitch helix of densely stacked (4.75. spacing) flat subunits (subunit sheets are oriented perpendicular to the helical axis) in an assumed left handed helix. Subunit packing within the filament is also known as cross-beta-sheet quaternary structure which leaves one direction of structure without any structural features except for the 4.75. axial spacing between adjacent subunits. Implementation of iterative helical real-space reconstruction algorithm in RELION enabled us to reconstruct and correct for imperfect helices instead of using Fourier-Bessel inversion which could not be applied to imperfect helices, as many studies show helical cross-over distance variations from one filament to the next. To avoid possible local optima, we checked different values for helical parameters and after all of the refinements our structure resolution went down to 4.0. with almost 200,000 segments, starting from roughly 1,000,000 segments. Using MonoRes and Localdeblur gave us better sharpening of the map. Due to the relatively low resolution, we were not able to build the atomic model from de novo, as well the handedness determination of filaments was not possible for resolutions lower than 2.3A. However, our ssNMR data showed that the resonances for certain residues in our filaments are as those of a previously reported structure (PDB: 6rt0) for alpha synuclein filaments. Using this as an initial model, we used Coot and added alanine residues, mutated them with the correct amino acids and refined the atomic model. Continuing efforts involve removing or lowering the clashes between a couple of the residues and improving the Ramachandran scores which are not satisfactory. Side chains are not generally fully resolved which makes the process more complicated. A tomography data set was collected from which the filaments were shown to be left handed. Our study is differentiated from other studies by the presence of Tau monomers with the α-synuclein. Tau is a pathologically related amyloid has implications in synucleinopathies. Previous studies shown that tau interacts with α-synuclein and accelerate α-synuclein aggregation. In this study, we investigated the effect of tau on α-synuclein aggregates and found that our α-synuclein filaments have distinct protofilament interactions than those of previously reported α-synuclein filaments. Despite 2D results that show extra densities on the outer surface of the 𝛼-synuclein filaments, we were not able to resolve them in 3D reconstructions. However,

the tomography data set combined with subvolume averaging may resolve the extra densities (projected to be Tau proteins) in 3D without the need to impose helical symmetry. The refinement


44

of the atomic model, validation and subvolume averaging is still in progress. The Study is funded by NIH.

Watching Protein Degradation in Action with Time-resolved Electron Cryo- Microscopy

Demian Keihsler1, Martin Columbini1, Tatyana Bodrug2, Nicholas Gene Brown2,

David Haselbach1 1 IMP - Research Institute of Molecular Pathology, Vienna, Austria

2 Department of Pharmacology and Lineberger Comprehensive Cancer Center, University of North, Carolina School of Medicine, Chapel Hill, NC, USA

Processes in the cell such as intracellular proteolysis rely on macromolecular machines which perform their tasks by assuming specific conformational states, guided by barriers in their potential energy landscape. The technological advance of electron cryo-microscopy in combination with modern computational and biophysical methods has allowed for a better understanding of the structures of these macromolecular machines and their behavior. However, to gain insights into the work cycles of these molecules complete active reaction mixtures should be measured, best in a time-resolved manner. In the past years, methods for time-resolved electron cryo-microscopy have emerged that are able to achieve time steps in the millisecond range. The aim of our ongoing research project is to develop a procedure with minimal time between sample mixing and grid freezing, allowing us to control reaction time with millisecond-precision and by thus visualising the states of the Ubiquitin-Proteasome system step-by-step. To achieve this, we are developing a custom-built modular device that uses microfluidic chips for sample mixing and a gas-assisted or piezoelectric sprayer to generate active reaction mixtures frozen in thin ice on the grid in a minimal time frame. Several technical questions remain open, such as what the absolute minimum achievable time between first contact of sample and substrate and complete freezing is, since it is ultimately limited by how fast a grid can be frozen. Further it remains to be shown how it can be justified on a theoretical level that the 3D-classes of grids frozen after different reaction times really are timesteps of the same process.


45

Streamlining the Cryo-FIB Lift-Out Procedure for Cryo-Electron Tomography Studies of Multicellular Organisms

Sven Klumpe

Max-Planck-Insitute for Biochemistry Munich, GERMANY

For more than a century, Drosophila melanogaster has been one of the most prevalent model organisms in developmental biology. Due to the loss of ultrastructural information in classical plastic-embedded samples [1], the structural details of most processes in this fascinating organism remain elusive. Structural analysis of multicellular organisms has recently been made possible by cryo-FIB milling and lift-out at cryo-temperatures [2]. However, the procedure is tedious and time-consuming [3, 4]. Here, we present software and hardware solutions to not only improve the throughput, but furthermore to streamline it for a reliable and robust workflow. With the help of this development we are now able to study the early development of D. melanogaster at molecular resolution. The optimizations are designed to be useful to the largest community: the software is adaptive to different applications and the hardware relies on simple and easy-to-realize modifications of the FIB microscope. 1. Winey, M., et al., Conventional transmission electron microscopy. Mol Biol Cell, 2014. 25(3): p. 319-23. 2. Schaffer, M., et al., A cryo-FIB lift-out technique enables molecular-resolution cryo-ET within native Caenorhabditis elegans tissue. Nat Methods, 2019. 16(8): p. 757-762. 3. Schaffer, M., et al., Cryo-focused Ion Beam Sample Preparation for Imaging Vitreous Cells by Cryo-electron Tomography. Bio Protoc, 2015. 5(17). 4. Hayles, M.F., et al., The making of frozen-hydrated, vitreous lamellas from cells for cryo-electron microscopy. J Struct Biol, 2010. 172(2): p. 180-90.

High-Throughput Cryo-EM Enabled by User-Free Preprocessing Routines

Yilai Li, Jennifer N. Cash, John. J.G. Tesmer, Michael A. Cianfrocco

University of Michigan Ann Arbor, MI

The growth of single-particle cryo-EM into a mainstream structural biology tool has allowed for many important biological discoveries. Continued developments in data collection strategies alongside new sample preparation devices heralds a future where users will collect multiple datasets per microscope session. To make cryo-EM data processing more automatic and user-friendly, we have developed an automatic pipeline for cryo-EM data preprocessing and assessment using a combination of deep learning and image analysis tools. We have verified the performance of this pipeline on a number of datasets and extended its scope to include sample screening by the user-free assessment of the qualities of a series of datasets under different conditions. We propose that our workflow provides a decision-free solution for cryo-EM, making data preprocessing more generalized and robust in the high-throughput era as well as more convenient for users from a range of backgrounds.


46

Tools for Investigating Fourier Space Sampling and Anisotropy in Cryo-EM

Philip Baldwin1 and Dmitry Lyumkis1

1. The Salk Institute for Biological Studies, La Jolla, CA 92037, USA.

Recent cryo-EM advances have enabled routine structural biology applications. However, virtually

all specimens currently adapt one or several “preferential orientations” on cryo-EM grids1. This

leads to a non-uniform sampling distribution associated with the reconstructed object. We recently

showed that any amount of non-uniformity in sampling always attenuates cryo-EM resolution,

while severe variations in sampling additionally lead to elongated, anisotropically resolved maps

and/or prevent structural studies altogether2. Non-uniformity in sampling, and the resulting

resolution anisotropy, remains a fundamental limitation in cryo-EM that affects virtually all

specimens prepared using current vitrification technologies. To address the issue, we proposed

a generalizable experimental strategy that ameliorates the effects of non-uniform sampling

caused by preferred orientation by tilting the stage during data acquisition3. Here, I will provide an

update on our efforts to understand the implications of non-uniform sampling, including: (1) A

discussion of the Sampling Compensation Factor (SCF) to explain the relationship between

Fourier space coverage and resolution attenuation; (2) A new set of tools embedded within a

graphical user interface to directly visualize Fourier space coverage and to determine how

sampling and resolution attenuation are affected when the specimen is tilted in the electron

microscope. The ideas provide a general framework for understanding sampling, with specific

experimental recommendations for data acquisition and analysis. References 1. Noble, A. J., Dandey, V. P., Wei, H., Brasch, J., Chase, J., Acharya, P., Tan, Y. Z., Zhang,

Z., Kim, L. Y., Scapin, G., Rapp, M., Eng, E. T., Rice, W. J., Cheng, A., Negro, C. J., Shapiro, L., Kwong, P. D., Jeruzalmi, D., Georges, des, A., Potter, C. S. & Carragher, B. Routine single particle CryoEM sample and grid characterization by tomography. elife 7, 32 (2018).

2. Baldwin, P. R. & Lyumkis, D. Non-Uniformity of Projection Distributions Attenuates Resolution in Cryo-EM. Prog. Biophys. Mol. Biol. (2019). doi:10.1016/j.pbiomolbio.2019.09.002

3. Tan, Y. Z., Baldwin, P. R., Davis, J. H., Williamson, J. R., Potter, C. S., Carragher, B. & Lyumkis, D. Addressing preferred specimen orientation in single-particle cryo-EM through tilting. Nature Methods 14, 793–796 (2017).


47

Visibility, Backup, and Archive for your CryoEM Data

Adam Marko, Scientific Solutions Lead, [email protected] Christian Smith, VP of Product Kiran Bageshpur, CEO Igneous Ingenious Systems, Seattle, WA

Abstract: As laboratories add new microscopes and improved detectors, data sizes in CryoEM often overwhelm scientific staff and storage systems. Here we describe a software-based solution that offers data visibility, backup, and archive of CryoEM and other scientific data to cloud or local storage. With our solution, storage can be visualized, searched, managed, and compressed through a single interface. A customer use case on image analysis will also be presented.

Template-Free Detection and Classification of Membrane-Bound Complexes

Antonio Martinez-Sanchez1,2, Zdravko Kochovski, Ulrike Laugks, Johannes Meyer zum Alten Borlogh, Saikat Chackraborty, Stefan Pfeffer, Wolfgang Baumeister, Vladan Lucic

1 Max Planck Institute of Biochemistry, Martinsried, Germany

2 University of Goettingen Medical Center, Goettingen, Germany

With faithful sample preservation and direct imaging of fully hydrated biological material, cryoelectron tomography provides an accurate representation of molecular architecture of cells. However, detection and precise localization of macromolecular complexes within cellular environments is aggravated by the presence of many molecular species and molecular crowding. We developed a template-free image processing procedure for accurate tracing of complex networks of densities in cryo-electron tomograms, a comprehensive and automated detection of heterogeneous membrane-bound complexes and an unsupervised classification (PySeg). Applications to intact cells and isolated endoplasmic reticulum (ER) allowed us to detect and classify small protein complexes. This classification provided sufficiently homogeneous particle sets and initial references to allow subsequent de novo subtomogram averaging. Spatial distribution analysis showed that ER complexes have different localization patterns forming nanodomains. Therefore, this procedure allows a comprehensive detection and tructural analysis of complexes in situ. PySeg is a sophisticate software implemented in Python which interacts with other software packages to complete the subtomogram averaging process as well as to prepare the inputs and display the outputs. To introduce the users in the utilization of this software here we also present a guided tutorial freely available on GitHub. This tutorial also includes a procedure for generating synthetic membranes with protein complexes associated. As a consequence, each workflow step is reproducible to ensure users a correct progress. The completion of this tutorial before processing in-house experimental data is recommended to learn the full potential of PySeg as well as its limitations.


48

Optimal Box/Particle Size based on Defocus Distribution for 200keV Single Particle CryoEM of <200kD proteins.

Toshio Moriya

High Energy Accelerator Research Organization (Kek) Ibaraki, Japan

Recently, it has been demonstrated that single particle analysis (SPA) using 200keV CryoEM paired with direct electron detector (DED) is capable to reconstruct <200kDa protein structures at resolution higher than 3.0A. However, the majority of near-atomic resolution cryoEM structures has been determined using 300keV CryoEMs equipped with DEDs. As a consequence, many of typical parameter settings for CryoEM session and image processing steps are based on the accumulated experience of 300keV CryoEMs, such as defocus range for EM sessions and amplitude contrast for CTF estimation. Therefore, we revised the parameters for 200keV acceration voltage, and found out merely optimizing box size and particle size based on defocus distribution of dataset can improve the resolution. As a result, using our Thermo Fisher Talos Arctica (200 keV) equipped with a Falcon III DED, we determined 2.85A resolution structure of the 110kDa C3 protein.

CCP-EM: Software for Cryo-EM

Colin M. Palmer, Tom Burnley, Agnel Praveen Joseph, Jola Mirecka and Martyn Winn Scientific Computing Department, STFC Rutherford Appleton Laboratory, Didcot, UK

The CCP-EM software suite [1] includes an ever-growing range of tools for working with cryo-EM data and provides a complete workflow for single-particle analysis. This begins with the well-known RELION package for single-particle reconstruction. To assess and optimise the 3D reconstruction, CCP-EM provides several map optimisation tools including MRC to MTZ, LocScale, LAFTER and Confidence Maps. Atomic models can be fitted into the map using programs such as MOLREP, Dock-EM and Flex-EM, or built de novo with Coot, Buccaneer and Nautilus. Atomic coordinates and B-factors can be refined with REFMAC5, and finally CCP-EM provides a combined model validation tool to assess both model geometry and model-to-map fit by calculating a number of metrics (e.g. FSC, SMOC score, Molprobity & CaBLAM) simultaneously.

References: [1] Recent developments in the CCP-EM software suite. Acta Cryst. D73, 469-477, 2017


49

Structural Basis for Strand Transfer Inhibitor Binding to HIV Intasomes

Dario Oliveira Passos

Salk Institute La Jolla, Ca

The HIV intasome is a large nucleoprotein assembly that mediates the integration of a DNA copy of the viral genome into host DNA. Intasomes are targeted by the latest generation of antiretrovirals, integrase (IN) strand transfer inhibitors (INSTIs). Challenges associated with lentiviral intasome biochemistry have hindered high resolution structural studies of how INSTIs bind to their native drug target. Here, we present high-resolution cryo-electron microscopy (cryo-EM) structures of HIV intasomes bound to the latest generation INSTIs. These structures highlight how small changes in the IN active site can have significant implications for drug binding and design and provide mechanistic insights into why a leading INSTI retains efficacy against a broad spectrum of drug resistant variants. The data have implications for expanding effective treatments available for HIV-1 infected individuals.

Cryo-Electron Tomography of Protein Nanocrystals for Structure Determination

Ariana Peck1, Qing Yao1, Aaron Brewster2, Nicholas Sauter2, John Heumann3, Grant Jensen1,4

1 Division of Biology and Biological Engineering, California Institute of Technology 2 Molecular Biophysics and Integrated Bioimaging Division, Lawrence Berkeley National

Laboratory 3 Department of Molecular, Cellular, and Developmental Biology, University of Colorado Boulder,

4 Howard Hughes Medical Institute

Structure determination is critical for revealing biochemical mechanisms and accelerating drug

discovery. X-ray crystallography has provided most of our knowledge of protein structure.

However, this method is constrained by the need for large, well-ordered crystals and the loss of

phase information. The rapidly developing methods of serial femtosecond crystallography, micro-

electron diffraction and single-particle reconstruction circumvent the first of these limitations by

enabling data collection from nanocrystals or purified proteins. However, the first two methods

also suffer from the phase problem, while many proteins fall below the molecular weight threshold

required by single-particle reconstruction. Cryo-electron tomography of protein nanocrystals has

the potential to overcome these obstacles of mainstream structure determination methods. Here

we describe a data processing workflow that leverages crystallographic algorithms to solve

protein structures from tomograms of nanocrystals. This pipeline is validated using simulated

crystals, and the effects of tilt errors and radiation damage on the accuracy of the recovered phase

information are explored.


50

Helical Parameters and Their Uncertainties: A Potential Filtering Tool

Hamidreza Rahmani, Nadia Daneshparvar, Dianne Taylor, Kenneth Taylor Institute of Molecular Biophysics

Florida State University Helical reconstruction has been a part of 3D electron microscopy for decades and new developments in Relion (after relion 2) seem to be promising. However, when working with elical ystems with larger helical rise like myosin filaments (with the rise of 145Å) helical symmetry offers very little improvement over rotational symmetry during reconstruction. This is a challenge and an opportunity: it makes is difficult to achieve high resolution, but it provides an opportunity to diagnose and study the helical parameters without enforcing or assuming them. Muscle activation and its structural process can be explained by several models but signaling mechanism for thick filaments is currently unknown. Related processes are observed among different types of muscle: vertebrate filaments raise their average axial repeat, isolated smooth muscle filaments dissolve when relaxed by adding ATP, and IFM filaments change their helical twist. These changes could be different manifestations of the same underlying process in the backbone. The importance of this underlying process is highlighted by the fact that IFM and cardiac muscle both respond to stretch activation. In this project, we can investigate this process at high resolution in IFM. Accurate measurement of the helical twist in ordered (relaxed), disordered (activated), and proteolyzed (headless) myosin filaments can confirm the relationship between the helical twist and the order of the myosin heads. Helical twists of our structures can be measured from reconstructions with only rotational symmetry imposed. The most common way to verify the helical twist is to reconstruct multiple subsets of segments but the error we get from this method is less than ~0.01°. We suspect this error value is under-estimated and ill-defined since the iterative process will end up with the same reference and bias the reconstruction. The accuracy of helical twist is better defined if evaluated using the raw data and the alignment parameters. New developments in relion2 have allowed us to track individual segments back to the raw data and their corresponding filaments. I have written a program to evaluate the helical twist of each individual filament using the prior information from relion2 and the alignment information from cisTEM. The population of filaments provides us with a statistical error for the helical twist. Currently, we have the value and uncertainty only for the relaxed filaments. Verifying our hypothesis is not the only application of reviewing the alignment of each filament. We can also separate the best filaments and the best particles using this method. We have had some early successful experience where the resolution improved when the particles were filtered based on their helical twist and alignment with respect to the filament, but we are at the early stages.


51

The Rocking Phase Plate – Another Step Towards Improved Stability

Barragán Sanz K1*, Irsen S1

Research Center caesar; Ludwig-Erhard-Allee 2, 53175 Bonn, Germany *Author for correspondence: [email protected]

Phase plates are promising tools for enhancing contrast, especially in cryo-transmission electron microscopy (cryo-TEM). As alternative to defocusing, contrast of weak phase objects -like most cryo samples- can be improved in close to focus images by inserting a phase plate into the back focal plane of the TEM1. The phase plate adds an additional phase shift of ideally π/2 to the scattered electrons and thus enhances image contrast. Their advantages for electron tomography and single particle acquisition have been demonstrated recently2. Nevertheless, phase plates are still no tool for routine application. This is mostly due to difficult handling and limited durability of actual phase plate designs. For more than 15 years, carbon based Zernike phase plates have been the most used type 3. Recently, a new type of hole free phase plates -Volta phase plate- has been made available. The hole free design has reduced ringing artifacts due to the missing hole edge. This is on cost of a variation of the phase shift over time. Here, we present our experimental results from a rocking phase plate. We use a classical Zernike type phase plate which is based on a thin iridium film instead of carbon. This improves the long term stability of the phase plate. To overcome the ringing artifact problem, we move the phase plate on a circular path during acquisition. This rocking mode virtually smooths the edge of the central hole. Additionally, the diameter of the central hole can be larger compared to classical Zernike phase plates which facilitates the positioning of the phase plate inside the TEM. The rocking mode is possible due to a special, piezo based positioning system, which can position the phase plate with nanometer precision4. We were able to show that the rocking mode settings can be used without interference of the. phase plate hole edge during image recording. Furthermore, we could not find any resolution loss caused by the moving phase plate. No cutoff can be detected in the power spectra of the rocking phase plate in comparison to the power spectrum calculated from a micrograph recorded with a not moving phase plate. In conclusion, the rocking mode could be an alternative method to acquire phase plate data. Currently, we are recording high-resolution single particle datasets to demonstrate up to which resolution the rocking phase plate can be used. References: 1.- Danev, R., Nagayama, K. (2001). Journal of the Physical Society of Japan 70, 696–702. 2.- Fan, X., Zhao, L., Liu, C., Zhang, J.-C., Fan, K., Yan, X., et al. (2017). Structure 25, 1623–1630. 3.- Nagayama, K., Danev, R. (2008). Philos Trans R Soc Lond B Biol Sci, 363, 2153–2162. 4.- Kurth, P., Pattai, S., Rudolph, D., Overbuschmann, J., Wamser, J., Irsen, S. (2014). Microscopy and Microanalysis 20, 220–221.

mailto:[email protected]


52

Mechanistic Insights Into Actin-Based Force Generation During Clathrinmediated Endocytosis Via in Situ Cryo-Electron Tomography

Daniel Serwas(1), Matthew Akamatsu(1), Karen M. Davies(2), David G. Drubin(1)

(1) Department of Molecular and Cell Biology, University of California Berkeley,

Berkeley, California 94720, USA (2) Molecular Biophysics and Integrated Bioimaging Division, Lawrence Berkeley National

Laboratory, Berkeley, California 94720, USA

Networks of interconnected actin filaments act in essential cellular processes, for example by providing force generated through polymerization. One of these processes is clathrin-mediated endocytosis (CME), a major pathway for down-regulation of signaling receptors and uptake of extracellular material, which involves a series of plasma membrane remodeling steps resulting in the formation of clathrin-coated vesicles (CCVs) containing cargo. How actin precisely functions in CME remains elusive, largely due to a lack of structural information on actin organization. Here, we made use of recent technological advances in cryo-electron tomography (cryo-ET) to discover the 3D actin cytoskeleton organization in mammalian CME at near-native state conditions. Combined with mathematical modeling, this analysis promises to reveal how actin functions in CME. We found the conditions that allow us to visualize individual actin filaments at CME sites and CCVs of fully hydrated cells using cryo-ET. Actin networks associated with CME sites and CCVs are complex arrangements consisting of branched as well as unbranched filaments. We determined filament polarity based on branch point geometry and found an accumulation of potentially growing plus-ends at positions that allow for polymerization driven plasma membrane deformation and CCV transport. Moreover, branched filaments do not arise from a common single filament, contradicting the theory that a single mother filament gives rise to the entire CMEassociated network. Surprisingly, we also found varying amounts of very long, heavily bent actin filaments at CME sites. Based on mathematical modeling, these filaments are predicted to store elastic energy, which augments force generation by actin polymerization to support membrane deformation and thus represents a previously unrecognized mode of actin-based force production in CME. Taken together, our results reveal the complex actin filament organization at endocytic sites in unprecedented detail. We are in the process of in-depth quantitative analysis of our data, which will, in combination with mathematical modeling, ultimately result in a detailed understanding of actin function in CME.


53

The Heptameric Assembly of Tnsc Bound to DNA Reveals the Activation Mechanism of Tn7 Transposition

Yao Shen1,5, Josue Gomez-Blanco2,5, Ruben Sanchez-Garcia3, Kaustuv Basu2, José María

Carazo3, Joseph Peters4, Joaquin Ortega2,5, Javier Vargas2,5, and Alba Guarné1,5

1Department of Biochemistry and 2Department of Anatomy and Cell Biology, McGill University, Montreal, QC, CA.3Biocomputing Unit, National Center of Biotechnology (CSIC), Instruct Image Processing Center, Madrid, Spain.4Department of Microbiology, Cornell University, Ithaca, NY,

USA.5McGill Centre de Recherche en Biologie Structurale.

The ability of transposons to move across genomes and modify gene expression makes them exceptional tools for genetic manipulation. Tn7 is emerging as a robust system due to its multiple target selection pathways. Beyond a heterodimeric transposase (TnsA/TnsB) and its activating subunit (TnsC), Tn7 elements encode a variety of proteins to direct transposition to specific sites. TnsD directs Tn7 transposition to a conserved site in bacterial chromosomes at high frequency. TnsE promotes Tn7 transposition to conjugal plasmids, ensuring horizontal spread. Many Tn7-like elements lack TnsE, but they have co-opted CRISPR/Cas elements that mediate RNA-guided transposition, underscoring the potential of Tn7 for genetic engineering applications [1-4]. All targeting mechanisms rely on TnsC to activate the TnsA/TnsB transposase, but how TnsC reads signals from diverse target selection proteins and activates transposition remains poorly understood. TnsC is an ATP-dependent DNA-binding protein from the AAA+ ATPase superfamily and nucleotide binding/hydrolysis is required for its different functions. We have determined the crystal structure of TnsC bound to ADP and a 3.7Å cryo-electron microscopy structure of TnsC bound to AMPPnP and DNA. Comparison of these structures reveals that TnsC undergoes extensive conformational rearrangements upon AMPPnP binding, resulting in the formation of a heptameric ring that encircles DNA. This reorganization changes the orientation of the C-terminal domain of TnsC and defines distinct N- and C-terminal faces of the ring. Combining the known information about the TnsC-mediated interactions with other Tn7-encoded proteins [5, 6] and using machine learning-based methods, we predict a model where the N-terminal face of the TnsC ring interacts with the target selection proteins while the C-terminal face activates the transposase. Our ongoing structure-guided functional analysis will provide mechanistic understanding of how TnsC responds to different targeting signals to activate Tn7 transposition, thereby facilitating the development of Tn7 system for genetic engineering applications.

Reference 1. Strecker, J., et al., RNA-guided DNA insertion with CRISPR-associated transposases.

Science, 2019. 365(6448): p. 48-53. 2. Klompe, S.E., et al., Transposon-encoded CRISPR-Cas systems direct RNA-guided DNA

integration. Nature, 2019. 571(7764): p. 219-225. 3. Halpin-Healy, T.S., et al., Structural basis of DNA targeting by a transposon-encoded

CRISPR-Cas system. Nature, 2019. 4. Peters, J.E., et al., Recruitment of CRISPR-Cas systems by Tn7-like transposons. Proc

Natl Acad Sci U S A, 2017. 114(35): p. E7358-E7366. 5. Choi, K.Y., J.M. Spencer, and N.L. Craig, The Tn7 transposition regulator TnsC interacts

with the transposase subunit TnsB and target selector TnsD. Proc Natl Acad Sci U S A, 2014. 111(28): p. E2858-65.

6. Ronning, D.R., et al., The carboxy-terminal portion of TnsC activates the Tn7 transposase through a specific interaction with TnsA. EMBO J, 2004. 23(15): p. 2972-81.


54

Dissecting Pathogenesis of Central Nervous System (CNS)-parasites by Electron Cryotomography (Cryo-ET)

Stella Y Sun1, Muyuan Chen2, Jason T Kaelber 3, Xiaoduo Dong4, Jian Shi4, Yasaman Nematbakhsh5, Chwee Teck Lim5, Michael F Schmid6, Li-av S Zarko 7, John C Boothroyd 7,

Cynthia Y He4, Wah Chiu1,6,7

1Stanford University, Bioengineering, Palo Alto, CA, 2Baylor College of Medicine, Biochemistry and Molecular Biology, Houston, TX,

3Rutgers University, Institute for Quantitative Biomedicine, Piscataway, NJ, 4 National University of Singapore, Biological Sciences, Singapore, Singapore

5National University of Singapore, Mechanobiology Institute, Singapore, Singapore, 6SLAC National Accelerator Laboratory, CryoEM center, Menlo Park, CA,

7Stanford University, Microbiology and Immunology, Palo Alto, CA

Electron Cryo-tomography (Cryo-ET) is a bioimaging technique that combines electron microscopy with cryopreservation, without chemical fixation or dehydration. Cryo-ET reveals the 3D ultrastructure and function of interest objects of a biological system in their native environment. The size and complexity of the systems can range from nanometers to micrometers and can be combined with other specimen preparation and imaging modalities. Protozoan parasites have evolved diverse mechanism to cross the host barriers and reach deeper tissues where they proliferate and lead to severe infectious diseases. Trypanosoma brucei (T. brucei), a flagellated protozoan parasite as the causative agent of human African sleeping sickness, is a highly invasive pathogen. They are able to penetrate deeply into the host central nervous system (CNS). To understand how flagellar motility facilitates cell penetration, we used cryo-ET to visualize two genetically anucleate mutants with different flagellar motility behaviors. We found that the T. brucei cell body is highly deformable in response to flagellar beating and environmental conditions. Based on the cryo-ET models, we proposed a mechanism of how flagellum motility is coupled to cell shape changes, which may facilitate penetration through size-limiting barriers. To date and give the size and complexity, studying parasite invasion by Cryo-ET has been extremely challenging, especially for the intracellular parasite Toxoplasma gondii (a congenital pathogen) involving entering a host cell. With the development of Cryo-electron microscopy, we are able to visualize a 3D structure of the remarkable invasion machinery inToxoplasma tachyzoites. The resulting 3D structures provide a testable model for its role in the invasion process.


55

A Dimension Reduction Method for Cryo-EmImage Processing

I-Ping Tu Institute of Statistical Science, Academia Sinica, Taipei, Taiwan

Principal component analysis (PCA) is arguably the most popular dimension reduction method for vector type data. When applied on image data, PCA demands the images to be portrayed as vectors. The resulting computation is heavy because it would solve an eigenvalue problem of a covariance matrix whose size equals the square of the pixel number. To mitigate the computation burden, multi-linear PCA that uses column and row basis with a Kronecker product to compose the matrix structure was proposed, for which the success was demonstrated on face image sets. However, when we apply MPCA on the particle images of the single particle cryo-electron microscopy (cryo-EM) experiments, the results are not satisfying. Here, we propose a dimension reduction method called Two Stage Dimensional Reduction (2SDR) where we first apply MPCA to extract its projection scores, and then apply PCA on these scores to further reduce the dimension. Tests using single particle cryo-EM benchmark experimental data sets demonstrate that 2SDR reduce huge computation costs compared to PCA, and show 2SDR can reconstruct better quality images than MPCA. Further application of 2SDR on a cryo-EM micrograph data set significantly reduces the noise to clearly reveal the individual particles. Remarkably, the de-noised particles boxed out from the micrograph allow subsequent structural analysis to reach a high-quality 3D density map. This is a joint work with Szu-Chi Chung, Po-Yao Niu, Su-Yun Huang and Wei-Hau Chang.

Improved SPA Throughput in EPU By Using an Enhanced Acquisition Strategy

Bart van Knippenberg, Andreas Voigt, Fanis Grollios, Dmitri Klenov

Thermo Fisher Scientific; Eindhoven, The Netherlands

Democratizing cryo-em by lowering the cost of operation and the ease of use has become an increasingly important topic in recent years. Reducing the required instrument time by increasing the number of acquired images per hour has therefore become a high priority for our data acquisition software (EPU). The recently introduced aberration-free image shift (AFIS) is used in combination with a novel target location clustering strategy. This drastically reduces the number of stage moves and centering steps during automated acquisition in EPU. However, using fewer centering steps poses the risk of inaccurate navigation. This could compromise the usability of the acquired data for reconstruction or even lead to targeting incorrect locations. Moreover, incorrect navigation might reduce the number of usable particles per image, thus actually lowering the effective throughput. To prevent this, we improved the target location detection and introduced a novel calibration procedure to account for any additional inaccuracies. No manual setup, monitoring or intermediate input is needed to use these improvements. All steps and calibrations are fully automatic, which ensures the high level of usability that our users expect from EPU. Measurements show that the navigation error remains below 100 nm, which is sufficient for the SPA use case. The combination of target location clustering and AFIS contributes up to a factor 2 in throughput without compromising image quality. Combined with fringe-free imaging and recent detectors, a total throughput increase up to 4x can be achieved. Experiments were performed on cryo samples, showing that reconstructions of 2Å are possible with the proposed enhancements.


56

Structure of Microbial Nanowires Reveals Stacked Hemes that Transport Electrons over Micrometers

Fengbin Wang

University of Virginia Charlottesville, VA

Long-range (> 10 μm) transport of electrons along networks of Geobacter ulfurreducens protein filaments, known as microbial nanowires, has been invoked to explain a wide range of globally- important redox phenomena. These nanowires were previously thought to be type IV pili composed of PilA protein. Here we report a 3.7 Å resolution cryo-electron microscopy structure which surprisingly reveals that, rather than PilA, G. sulfurreducens nanowires are assembled by micrometer-long polymerization of the hexaheme cytochrome OmcS, with hemes packed within ~3.5-6 Å of each other. The inter-subunit interfaces show unique structural elements such as inter-subunit parallel-stacked hemes and axial coordination of heme by histidines from neighbouring subunits. Wild-type OmcS filaments show 100-fold greater conductivity than other filaments from an ΔomcS strain, highlighting the importance of OmcS to conductivity in such structures. This structure explains the remarkable capacity of soil bacteria to transport electrons to remote electron acceptors for respiration and energy sharing.

Advances in Vibration Control Strategies for Cryo TEM, E-beam Metrology and other Nanotech tools

Wes Wigglesworth

TMC Vibration Control Peadbody, MA

Contact: Wes Wigglesworth, [email protected], 978-538-1135

Cryo Transmission Electron Microscopes, Focused Ion Beam and Electron Microscopes, and other ultra-precision instruments installed in nanotech, materials engineering, and life science research facilities are required to achieve extremely high resolution and precise measurements. Today’s instruments commonly can achieve sub-nanometer, and even sub-angstrom, resolution. Floor vibration often prevents such tools from meeting their design specifications. To achieve floor vibration specifications published by the tool manufacturers, architects work with consultants to design extremely quiet buildings at great cost. Is this efficient and effective? Is there an alternative? Architects can design buildings to meet moderate floor vibration levels without too much difficulty. Designing buildings to meet the extremely low vibration levels required for the nanoscale and other advanced science and technology facilities, however, requires significantly more cost with diminishing returns. Even to the extent that the building floors are quiet, once the building is populated with people and machinery, vibration sources are introduced negating much of the benefit of the quiet building design. Sources of vibration are discussed and different strategies for vibration mitigation are presented including passive vibration isolation, massive isolated plinths and point-of-use inertial active vibration control pedestals


57

Notes:


58

Notes:

59

Registered Participants

Ximena Barros Alvarez

Postdoc STANFORD UNIVERSITY

[email protected]

Joakim Andén Faculty

KTH ROYAL INSTITUTE OF TECHNOLOGY [email protected]

Christopher Aylett

Faculty IMPERIAL COLLEGE LONDON

[email protected]

Maia Azubel Staff

STANFORD MEDICAL SCHOOL [email protected]

Philip Baldwin

Staff SALK INSTITUTE

[email protected]

Alberto Bartesaghi Faculty

DUKE UNIVERSITY [email protected]

Jan Binovsky

Student CENTRAL EUROPEAN INSTITUTE OF

TECHNOLOGY [email protected]

Dominika Borek

Faculty UT SOUTHWESTERN MEDICAL CENTER

[email protected]

Jacob Brink Other

JEOL USA, INC. [email protected]

Raquel Bromberg Faculty

UT SOUTHWESTERN MEDICAL CENTER [email protected]

Alister Burt Student

INSTITUT DE BIOLOGIE STRUCTURALE [email protected]

Kai Cai PostDoc


Stephen Carter

Postdoc CALTECH

[email protected]

Daniel Castano Staff

UNIVERSITY OF BASEL [email protected]

Wei-Hau Chang

Faculty ACADEMIA SINICA

[email protected]

Michael Chapman Faculty

UNIVERSITY OF MISSOURI [email protected]

Muyuan Chen

Faculty BCM

[email protected]

Zhe (James) Chen Faculty

UNIVERSITY OF TEXAS SOUTHWESTERN MEDICAL CENTER

[email protected]

Anchi Cheng Staff

NEW YORK STRUCTURAL BIOLOGY CENTER [email protected]

Wah Chiu

Faculty STANFORD

[email protected]





















60

Hui-Ting Chou Scientist AMGEN

[email protected]

Nattaya Chutikamoltham Student

MAHIDOL UNIVERSITY [email protected]

Michael Cianfrocco

Faculty UNIVERSITY OF MICHIGAN

[email protected]

Joey Davis Faculty

MASSACHUSETTS INSTITUTE OF TECHNOLOGY [email protected]

Sacha De Carlo

Other DECTRIS LTD.

[email protected]

David Derosier Other

SBP INSTITUTE [email protected]

Frank Dimaio

Faculty U. WASHINGTON [email protected]

Edward Egelman

Faculty UNIVERSITY OF VIRGINIA

[email protected]

Matthias Eibauer Senior Scientist

UNIVERSITY OF ZURICH [email protected]

Paul Emsley

Staff MRC LABORATORY OF MOLECULAR BIOLOGY

[email protected]

Philipp Erdmann Postdoc

MAX PLANCK INSTITUTE OF BIOCHEMISTRY [email protected]

Mériem Er-Rafik Faculty

DTU, DENMARK TECHNICAL UNIVERSITY [email protected]

Adam Fluty

Student BAYLOR COLLEGE OF MEDICINE

[email protected]

James Fraser Faculty UCSF

[email protected]

José-María Carazo Garcia Other

NATIONAL CENTER FOR BIOTECHNOLOGY – CSIC

[email protected]

Cornelius Gati Staff

SLAC NATIONAL ACCELERATOR LABORATORY [email protected]

Tamir Gonen

Faculty UCLA

[email protected]

Iris Grossman-Haham Postdoc UCSF

[email protected]

Sahil Gulati Other

GATAN, INC. [email protected]

Yirui Guo

Other LIGO ANALYTICS, INC

[email protected]

Dorit Hanein Faculty

SANFORD-BURNHAM-PREBYS [email protected]

Eric Hanssen

Faculty THE UNIVERSITY OF MELBOURNE

[email protected]























61

Richard Held Postdoc

STANFORD UNIVERSITY SOM [email protected]

Benjamin Himes

Postdoc HHMI JRC

[email protected]

Alimohammad Hojjatian Student

FLORIDA STATE UNIVERSITY [email protected]

Yuri Iozzo

Other SANOFI

[email protected]

Rossi Irobalieva Staff ETHZ

[email protected]

Grant Jensen Faculty

CALIFORNIA INSTITUTE OF TECHNOLOGY [email protected]

Wen Jiang

Faculty PURDUE UNIVERSITY

[email protected]

Zachary Johnson

Scientist

PFIZER

[email protected]

Slavica Jonic Faculty

SORBONNE UNIVERSITY – CNRS [email protected]

Sebastian Kaspar

Program Manager Em DECTRIS

[email protected]

Demian Keihsler Research Technician

RESEARCH INSTITUTE OF MOLECULAR PATHOLOGY (IMP)

[email protected]

Heena Khatter

Other VERTEX PHARMACEUTICALS

[email protected]

Sagar Khavnekar Graduate Student

MAX PLANCK INSTITUTE OF BIOCHEMISTRY [email protected]

Daisuke Kihara

Faculty PURDUE UNIVERSITY

[email protected]

Masahide Kikkawa Faculty

THE UNIVERSITY OF TOKYO [email protected]

Dmitri Klenov

Other THERMO FISHER SCIENTIFIC

[email protected]

Gerard Kleywegt Staff

EMBL - EUROPEAN BIOINFORMATICS INSTITUTE

[email protected]

Sven Klumpe Student

MAX-PLANCK-INSITUTE FOR BIOCHEMISTRY [email protected]

Lubomir Kovacik

Postdoc UNIVERSITY OF BASEL

[email protected]

Julio Kovacs Staff

GATAN [email protected]

Thomas Lane

Staff SLAC

[email protected]

Roy Lederman Assistant Prof.

YALE UNIVERSITY [email protected]























62

Angela (Hsiang-Ting) Lei

Staff INCYTE RESEARCH INSTITUTE

[email protected]

Jeremy Leitz Postdoc

STANFORD UNIVERSITY [email protected]

Yilai Li

Postdoc UNIVERSITY OF MICHIGAN

[email protected]

Yuxi Liu Postdoc

CALIFORNIA INSTITUTE OF TECHNOLOGY [email protected]

Steven Ludtke

Faculty BAYLOR COLLEGE OF MEDICINE

[email protected]

Dmitry Lyumkis

SALK INSTITUTE FOR BIOLOGICAL STUDIES

[email protected]

Adam Marko Other

IGNEOUS SYSTEMS [email protected]

Antonio Martinez-Sanchez

Postdoc MAX PLANCK INSTITUTE OF BIOCHEMISTRY

[email protected]

Antonio Martinez-Sanchez Senior Research Scientist

UNIVERSTY OF GOETTINGEN MEDICAL CENTER

David Mastronarde

Faculty UNIVERSITY OF COLORADO

[email protected]

Jola Mirecka Postdoc STFC

[email protected]

Robert Monteverde Other

DIRECT ELECTRON [email protected]

Paul Mooney

Staff GATAN, INC.

[email protected]

Toshio Moriya Faculty

HIGH ENERGY ACCELERATOR RESEARCH ORGANIZATION (KEK)

[email protected]

Robert Murphy Other

PANASAS [email protected]

Alex Noble

Postdoc NEW YORK STRUCTURAL BIOLOGY CENTER

[email protected]

Zbyszek Otwinowski Faculty


Abbas Ourmazd

Faculty UNIVERSITY OF WISCONSIN-MILWAUKEE

[email protected]

Colin Palmer Staff

UKRI STFC [email protected]

Kanupriya Pande

Staff LAWRENCE BERKELEY NATIONAL LAB

[email protected]

Dario Oliveira Passos Staff

SALK INSTITUTE [email protected]

Ariana Peck

Postdoc CALTECH

[email protected]






















63

Pawel Penczek Faculty

THE UNIVERSITY OF TEXAS [email protected]

Clint Potter

Other NEW YORK STRUCTURAL BIOLOGY CENTER

[email protected]

Hamidreza Rahmani Student


Snekalatha Raveendran

Staff BAYLOR COLLEGE OF MEDICINE

[email protected]

Shaun Rawson Staff

HARVARD MEDICAL SCHOOL [email protected]

Bharat Reddy

Postdoc UNIVERSITY OF CHICAGO

[email protected]

Jane Richardson Faculty

DUE UNIVERSITY [email protected]

Smriti Sangwan

Postdoc UCSF

[email protected]

Kristina Barragan Sanz Student

RESEARCH CENTER CAESAR [email protected]

Andreas Schenk

Staff FRIEDRICH MIESCHER INSTITUTE FOR

BIOMEDICAL RESEARCH [email protected]

Michael Schmid

Other SLAC/STANFORD

[email protected]

Hande Boyaci Selcuk Postdoc

THE ROCKEFELLER UNIVERSITY [email protected]

Daniel Serwas

Postdoc UC BERKELEY

[email protected] Yao Shen Student

MCGILL UNIVERSITY [email protected]

Jian Shi

Staff NATIONAL UNIVERSITY OF SINGAPORE

[email protected]

Pitor Silz Faculty

HARVARD MEDICAL SCHOOL AND SBGRID [email protected]

Ji-Joon Song

Faculty KOREA ADVANCED INSTITUTE OF SCIENCE

AND TECHNOLOGY [email protected]

Carlos Oscar Sanchez Sorzano

Faculty CENTRO NACIONAL BIOTECNOLOGíA (CSIC)

[email protected]

Scott Stagg Faculty


Holger Stark

Faculty UNIVERSITY OF GÖTTINGEN

Marc Stroms

Other THERMO FISHER SCIENTIFIC [email protected]

Stella Sun

Postdoc STANFORD UNIVERSITY

[email protected]






















64

Mark Swift Staff

SANFORD BURNHAM PREBYS MEDICAL DISCOVERY INSTITUTE [email protected]

Chad Tabatt

Other NANOSCIENCE INSTRUMENTS

[email protected]

Florence Tama Faculty

NAGOYA UNIVERSITY [email protected]

Kenneth Taylor

Faculty FLORIDA STATE UNIVERSITY

[email protected]

Akihisa Tsutsumi Faculty

THE UNIVERSITY OF TOKYO [email protected]

I-Ping Tu Faculty

ACADEMIA SINICA [email protected]

Beata Turoňová Staff Scientist

EMBL [email protected]

Bart Van Knippenberg

Other THERMO FISHER SCIENTIFIC

[email protected]

Andreas Voigt Application Software Scientist

THERMO FISHER SCIENTIFIC [email protected]

Niels Volkmann

Faculty SBP

[email protected]

William Wan Faculty

VANDERBILT UNIVERSITY [email protected]

Jonathan Wagner Student

MAX PLANK INSTITUT

[email protected]

Fengbin Wang

Postdoc UNIVERSITY OF VIRGINIA

[email protected]

Kristopher White Postdoc

STANFORD UNIVERSITY [email protected]

Wes Wigglesworth

Other TMC

[email protected]

Rahel Woldeyes Postdoc

STANFORD [email protected]

Xiao-Ping Xu

Staff SBP MEDICAL DISCOVERY INSTITUTE

[email protected]

Qing Yao Postdoc

CALTECH [email protected]`

Craig Yoshioka

Faculty OREGON HEALTH AND SCIENCES UNIVERSITY

[email protected]

Xiaodi Yu Staff

JOHNSON & JOHNSON [email protected]

Weili Zheng

Postodc UNIVERSITY OF VIRGINIA

[email protected]

Ellen Zhong Student

MIT [email protected]























65

Huabin Zhou Postdoc

UNIVERSITY OF TEXAS SOUTHWESTERN MEDICAL CENTER

[email protected]

Ye Zhou Postdoc

DUKE UNIVERSITY [email protected]



66

Meeting Room Diagram

NOTE: All sessions will be held in the Mountain-Lake Room

Date post:	30-Jul-2020
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

4 INTERNATIONAL S ON CRYO-EM 3D IMAGE ANALYSIS 2020Sciences, UCSF, San Francisco, CA, USA. 8:40 –...

Documents