seeing the small with electron crystallography · tim grüne electron crystallography in biology...
TRANSCRIPT
WIR SCHAFFEN WISSEN — HEUTE FÜR MORGEN
Dr. Tim Grüne :: Paul Scherrer Institut :: [email protected]
Seeing the Small with Electron Crystallography
Center for Chemistry and Biomedicine Innsbruck31st May 2017
Tim Grüne Electron Crystallography in Biology
1 - Seeing the small
CCB Innsbruck 31st May 2017 1/38
Tim Grüne Electron Crystallography in Biology
Microscopy
Light Object Lense Image Plane (Detector)
Enlarged image from object
CCB Innsbruck 31st May 2017 2/38
Tim Grüne Electron Crystallography in Biology
Microscopes
800nm 400nm
Radio Micro Infrared X−raysVisible UV −raysγ
30cm10km 1mm 1nm 10pm
wavelength
123keV1.23keV3.09eV1.54eV0.00123eV4.12µV energy
• Light has a wavelength: 800nm = red . . . 400nm = blue
• Resolution (possibility to distinguish between two neighbouring points) ≈ λ
• Size of interest for molecules: ≈ 1Å = 0.1nm
• No lenses = no microscopes available for such short wavelengths!
CCB Innsbruck 31st May 2017 3/38
Tim Grüne Electron Crystallography in Biology
Workaround I: Electron Microscopy
• Light / Electromagnetic waves: no microscopes
• Use electrons as wave and use magnetic lenses: even at very short wavelength
• Since quantum mechanics: electrons are waves, de Broglie wavelength
mev√1− (v/c)2
= hλ
• Commonly used: Electron energies 200–300keV, wavelength λ = 0.025−0.017Å.
CCB Innsbruck 31st May 2017 4/38
Tim Grüne Electron Crystallography in Biology
Electron Microscope: Imaging Mode
Electron Wave Object Lense Image Plane (Detector)
Detector noise and radi-ation senstivity requirelow contrast images
Martinez-Rucobo et al. Mo-lecular Cell (2015) 58, 1079–1089
CCB Innsbruck 31st May 2017 5/38
Tim Grüne Electron Crystallography in Biology
Workaround II: Crystallography
Wave Crystal No lens!• Crystal requires no lens.• Crystal concentrates its signal on indi-
vidual spots aka reflections
CCB Innsbruck 31st May 2017 6/38
Tim Grüne Electron Crystallography in Biology
Crystallography in Brief
• Crystals amplify the signals from atoms so that the sig-nal can be detected.
• Data are spot positions and spot intensities• Data are not atoms, some calculations are required• Crystal structures provide high level of detail insight
Photoactive Yellow Protein (2ZOI): Turns lightinto molecular movement. PDB Molecule of theMonth March 2017
CCB Innsbruck 31st May 2017 7/38
Tim Grüne Electron Crystallography in Biology
Electron Crystallography
CCB Innsbruck 31st May 2017 8/38
Tim Grüne Electron Crystallography in Biology
Combining Electron Radiation (and Microscopes) with Crystallography
typical protein crystal size for X-rays(0.2mm = 200µm)
typical protein crystal size for elec-trons, 0.1×0.14×1.7µm3
���������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������
���������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������
1m
1m
1m
volumes compare like 6 bath tubs ofwater vs. a 10µl drop from a pipette
The combination of electron radiation (as inEM) with crystallography permits crystallo-graphic data collection from tiny crystals.
CCB Innsbruck 31st May 2017 9/38
Tim Grüne Electron Crystallography in Biology
Typical Protein Crystallisation Trials
Luft, Wolfley, Snell, Crystal Growth& Design (2011), 11, 651–663
CCB Innsbruck 31st May 2017 10/38
Tim Grüne Electron Crystallography in Biology
Drops viewed through TEM
Stevenson,. . . , Calero, PNAS (2014) 111, 8470–8475 / Calero, . . . , Snell, Acta Cryst (2014) F70, 993–1008
CCB Innsbruck 31st May 2017 11/38
Tim Grüne Electron Crystallography in Biology
Drug Research: The Novartis Library
• 2,000,000 compounds of potential drug targets
• 30-40% suitable for X–ray powder analysis
• 10% suitable for single crystal X–ray analysis
Dr. Trixie Wagner (2012)
CCB Innsbruck 31st May 2017 12/38
Tim Grüne Electron Crystallography in Biology
Powders contains Single Crystals
Novartis IRELOH:Ø= 1,700nm = 1.7µm
Novartis EPICZA:Ø= 500nm = 0.5µm
Zeolite (Prof. Bokhoven):Ø= 300nm = 0.3µm
CCB Innsbruck 31st May 2017 13/38
Tim Grüne Electron Crystallography in Biology
2 - Structures
CCB Innsbruck 31st May 2017 14/38
Tim Grüne Electron Crystallography in Biology
Macromolecular Structures
• Protein Data Bank (PDB, www.pdb.org) contains about 13 protein structures from 3D electron diffraction
• Started 2013 (PDB ID 2013)
• (Mostly) commonly known protein — not a new structure to date.
CCB Innsbruck 31st May 2017 15/38
Tim Grüne Electron Crystallography in Biology
Structures from Test Proteins
Cruz et al., “Atomic-resolution structures from fragmented protein crystals with the cryoEM method MicroED”, NatureMethods(2017), 14, 399–405
CCB Innsbruck 31st May 2017 16/38
Tim Grüne Electron Crystallography in Biology
Thermolysin (sample courtesy Ilme Schlichting)
• Spacegroup P6122• Unit Cell 94.3 94.3 130.4 90◦ 90◦ 120◦
• dmin = 3.5Å• 72.4% completeness• MR with 3DNZ poly Alanine: TFZ=26.4, LLG=433• Buccaneer: side chain extension 315/316• Refmac5: R1/“Rfree” = 28.0% / 29.9% (4N5P w/o
water)
CCB Innsbruck 31st May 2017 17/38
Tim Grüne Electron Crystallography in Biology
Lysozyme
1. MR (Phaser) from poly Ala monomer determines spacegroup P21212 (TFZ=19.8, LLG=335.3)
2. Side chain completion with Buccaneer all except 27 atoms
3. Refinement with refmac5
After MR: difference density for bulky sidechains
Refined map guides model completion
CCB Innsbruck 31st May 2017 18/38
Tim Grüne Electron Crystallography in Biology
Organic Structures
• Pioneers: ZG Pinsker, BK Vainshtein (1940s +; 1990s)
• D Dorset (1995: Textbook Electron Crystallography)
• U Kolb (recording of 3D diffraction patterns, ADT, 1997+)
• X Zou, S Hovmóller (recording of 3D diffraction patterns with beam precession, RED, 2008+)
CCB Innsbruck 31st May 2017 19/38
Tim Grüne Electron Crystallography in Biology
Pharmaceutical I: Visualisation of Hydrogen Atoms
H–atom positions can be refined against electron diffraction dataCCDC: IRELOH, Dai et al., Eur. J. Org. Chem (2010), 6928-6937
Sample courtesy Novartis
• Field of view: 3µm• Crystal: 1.6µm×400nm
• dmin < 0.8Å• P212121: 85% completeness
with 3 crystals• a=8.06Å b=10.00Å c=17.73Å
• Hydrogen atoms in difference map evenwith poor model
• 1334 reflections, 195 parameters, 156 re-straints (RIGU)
• R1 = 15.5%,Rcomplete = 18.5%
CCB Innsbruck 31st May 2017 20/38
Tim Grüne Electron Crystallography in Biology
Pharmaceutical II: Differentiation of Atom Types
Data quality: recognition of atom types, C vs. O vs. N etc. (CCDC: EPICZA)
0
20
40
60
80
100
0 0.1 0.2 0.3 0.4 0.5 0.6 2 3 4 5 6 7 8 9 10 11 12
5 2.5 1.25 1.0
Com
plet
enes
s [%
]
I/σI
d* [1/Å]
d [Å]
CompletenessI/σI
• Field of view: 3µm• Crystal: 400nm diameter
• dmin = 0.87Å• a=11.35Å, b=12.7Å, c=13.0Å• P212121: completeness with 4
crystals: 86%
• 2545 refl., 258 param., 267 restraints(RIGU)
• all data: R1 = 15.9%,Rcomplete = 19.1%• R1 = 14.7%,Rcomplete = 18.0%
CCB Innsbruck 31st May 2017 21/38
Tim Grüne Electron Crystallography in Biology
Pharmaceutical II (EPICZA): Structure Solution Process
Direct methods reveal H atoms HFIX: all except 1 H Final Structure=data quality =model quality
CCB Innsbruck 31st May 2017 22/38
Tim Grüne Electron Crystallography in Biology
Summary: Electron Diffraction of Organic Compounds
• Structures can be solved with X–ray knowledge and methods.
• Radiation damage present, but not (always) limiting
• Kinematic approximation sufficient for high quality structures
1.15
1.2
1.25
1.3
1.35
1.4
1.45
1.5
1.55
1.6
1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55 1.6
Bond
leng
ths
X m
odel
[Å]
Bond lengths e- model [Å]
IRELOH
xline fit f(x) = 0.92 * x + 1.06e-01
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2
2.1
2.2
1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2
Bond
leng
ths
X-ra
y m
odel
[Å]
Bond lengths e- model [Å]
EPICZA
xline fit f(x) = 0.96 * x + 3.52e-02
CCB Innsbruck 31st May 2017 23/38
Tim Grüne Electron Crystallography in Biology
3 - Technical Aspects
CCB Innsbruck 31st May 2017 24/38
Tim Grüne Electron Crystallography in Biology
Dynamic Scattering
Data from SAPO–34: I(−2,−1,1)> Idirect beam (Eiger chip, 256x256 px)
CCB Innsbruck 31st May 2017 25/38
Tim Grüne Electron Crystallography in Biology
Kinemtic (X–ray) and Dynamic (e−) Scattering
• Kinematic Theory of Diffraction: Every photon / electron / neutron scatters once in the crystal
• |Fideal(hkl)| ∝√
Iexp(hkl)
• Dynamic Scattering: Multiple Scattering events occur
• Electron Diffraction: Multiple Scattering occurs even with nanocrystals
• For data from proteins: Currently no satisfactory treatment
CCB Innsbruck 31st May 2017 26/38
Tim Grüne Electron Crystallography in Biology
Multiple (Dual) Scattering
~Si
~S1o
~S2o
• Outgoing ray ~S1o acts as incoming ray for re-
flection ~S2o.
• Probability of re–reflection thickness de-pendent
CCB Innsbruck 31st May 2017 27/38
Tim Grüne Electron Crystallography in Biology
Multiple (Dual) Scattering
~Si
~S1o
~S2o ~S2
o
Laue Conditions (accordingly~b and~c):
(~S1o−~Si) ·~a = h1
(~S2o−~S1
o) ·~a = h′
(~S2o−~Si) ·~a = h1+h′ = h2
Requirement for detrimental effect on I(h2,k2, l2)• I(h1k1l1) must be strong• I(h′,k′, l′) must be strong• I(h2k2l2) must be weak• I(h1k1l1) and I(h2k2l2) on same frame
CCB Innsbruck 31st May 2017 28/38
Tim Grüne Electron Crystallography in Biology
Multiple (Dual) Scattering
I(h 1k1l1) I(h 2k2l2)
I(hkl)
1
3
5
10
2
���������
���������
������
• Re-reflection more likely for thicker crystal(path)• Percentage similar for all reflections on frame (2θ ≈ 0)• 10% of strong reflection affects weak reflection⇒ Measured intensities “shifted” from strong to weak⇒ Low resolution reflection under–, high resolution reflec-
tions overestimated
CCB Innsbruck 31st May 2017 29/38
Tim Grüne Electron Crystallography in Biology
Electron Detectors for Diffraction
TVIPS CMOS 1024x1024 Timepix 256x256 Eiger (PSI)20≤ I ≤ 21 cts 0≤ I ≤ 1 ct Lysozyme (invd) 0≤ I ≤ 1 ct SAPO–34 crystal
≈ 1kHz,50µm×50µm ≤ 23kHz,75µm×75µmcut–off: 11809 cut–off: 16, 64, or 4096 (4, 8, 12 bit)dead time ≈ 0.01s dead time 3µs
CCB Innsbruck 31st May 2017 30/38
Tim Grüne Electron Crystallography in Biology
The Lens System
���������������������������������������������������������������������������������������������������������������������������������
���������������������������������������������������������������������������������������������������������������������������������
��������������������������������������������������������������������������������
C3
C2
C1
Gun / Source
Objective lens
Sample
Diffraction Plane
Imaging Plane
• Lenses C1–C3 shape beam• Crystallography: Parallel beam• Objective lens: sets effective detector distance to back-
focal plane = diffraction mode• C3 not present in all microscopes
Lenses cause distortions.
CCB Innsbruck 31st May 2017 31/38
Tim Grüne Electron Crystallography in Biology
Garnet Andradite
• The garnet Andradite, Ca3Fe23+(SiO4)3, radiation hard
• 2 grids courtesy Xiaodong Zou (Stockholm)• Space group Ia3̄d, a = 12.06314(1)Å (ICSD No. 187908)
(Wikipedia)
• Summed images from Garnet (200keV)• 66.8◦ rotation• good coverage of detector surface
CCB Innsbruck 31st May 2017 32/38
Tim Grüne Electron Crystallography in Biology
Spatial Correction for the Detector Surface
• Spot positions calculated from Laue Conditions
~S.~a = h~S.~b = k~S.~c = l
• Data Processing: Deviations between calculated and observed positions
• e.g. XDS: per–pixel look-up tables for X– and Y–coordinates
• Independent of Source of Error
CCB Innsbruck 31st May 2017 33/38
Tim Grüne Electron Crystallography in Biology
Spatial Correction for the Detector Surface
XDS Correction Table X–coordinate and Y–coordinate
CCB Innsbruck 31st May 2017 34/38
Tim Grüne Electron Crystallography in Biology
Directly Visible Improvements
Garnet Data set processed before spatial correction:
BEAM DIVERGENCE: 0.16◦
REFLECTING RANGE: 0.47◦
Garnet Data set processed after spatial correction:
BEAM DIVERGENCE: 0.15◦
REFLECTING RANGE: 0.28◦
CCB Innsbruck 31st May 2017 35/38
Tim Grüne Electron Crystallography in Biology
Improved Cell Accuracy with Look–up Tables
1. Collect data from garnet2. Change as little as possible3. Collect data from target sample4. Process using garnet correction tables
Sample Courtesy Roche C31H29Cl2F2N3O4, SG P21
Data Collection and Processing: Max Clabbers
a b c α β γ
XRPD 6.405 18.206 25.829 90.000 92.180 90.000XDS uncorrected 6.556 18.728 26.276 90.500 92.243 90.540XDS corrected 6.564 18.721 26.254 90.064 92.171 90.137
CCB Innsbruck 31st May 2017 36/38
Tim Grüne Electron Crystallography in Biology
4 - Conclusions
• Electron Diffraction = Structures from very small crystals < 1µm
• Applications: Single Crystal Structures, where X–rays only see powder
• High quality data + structures for organic compounds
• Proteins: Radiaton damage currently limits competing data resolution: new ways of data collection required
CCB Innsbruck 31st May 2017 37/38
Tim Grüne Electron Crystallography in Biology
5 - Acknowledgements
• Prof. J. P. Abrahams, Dr. E. van Genderen, M. Clabbers, Dr. T. Blum, C. Borsa, J. Heidler, Dr. R. Pantelic
• Novartis (Compounds)
• Roche (Compounds)
• Dr. I. Nederlof, ASI (Medipix / Timepix)
• Dr. B. Schmitt, PSI Detector group
• Prof. K. Diederichs (XDS)
• Dr. W. Kabsch (XDS)
• Dr. D. Waterman (DIALS)
• Prof. J. van Bokhoven, ETH Zürich
• Prof. X. Zou and S. Hovmöller, University Stockholm (garnet grids)
CCB Innsbruck 31st May 2017 38/38