rapid readout detector captures protein time-resolved waxs
TRANSCRIPT
nature methods | VOL.7 NO.10 | OCTOBER 2010 | 775
correspondence
Rapid readout detector captures protein time-resolved WAXS
To the Editor: Time-resolved wide-angle X-ray scattering (WAXS) is emerging as a powerful approach for visualizing glob-al conformational changes in proteins in real time. WAXS-based studies to date include following the light-triggered reactions of hemoglobin1 and myoglobin2,3 in complex with carbon monoxide, and structural characterization of the photocycles of bacteriorho-dopsin and proteorhodopsin4. Compared to time-resolved Laue diffraction and trapped intermediate studies5, time-resolved WAXS is a solution-based method and therefore permits visualization
Color blindness
To the Editor: As Bang Wong notes in his recent column on proper color coding in figures, just picking suitable colors is not always sufficient1. In particular, one important, but often neglect-ed, issue is color blindness because it affects a substantial portion of the human population (addi-tional information is available at MedlinePlus: http://www.nlm.nih.gov/medlineplus/colorblind ness.html). One well-known exam-ple of frequent color vision defects is red-green deficiency, which occurs, in case of Northern European ancestry, in about 8 percent of males and 0.5 percent of females2. Therefore, it is advisable to select figure colors accordingly, and journals might introduce color guide-lines to increase the authors’ awareness of this accessibility issue. For instance, software such as Adobe Photoshop offers color blindness proofing filters that simulate what an image will look like to individu-als with color blindness. Authors can then adjust the image to make it universally accessible (see Adobe Photoshop CS5 accessibility over-view at http://www.adobe.com/accessibility/products/photoshop/overview.html).
competInG FInancIaL InterestsThe author declares no competing financial interests.
mario albrecht
Max Planck Institute for Informatics, Campus E1.4, Saarbrücken, Germany. e-mail: [email protected]
1. Wong, B. Nat. Methods 7, 573 (2010).2. Deeb, S.S. Clin. Genet. 67, 369–377 (2005).
Nature Methods replies: We agree with Albrecht that color blind-ness in readers is an important but often overlooked consideration when authors choose figure colors1. In 2007 Nature Publishing Group began recommending that authors choose color combi-nations other than green and red when they prepare their manu-scripts for submission. Before accepting any manuscript, we again ask authors to recolor red-and-green graphs, schematics and mod-els in which the colors are arbitrary. Unfortunately, as can be seen in the pages of this issue and issues of other Nature Publishing Group journals, there are still instances of red and green being used in a manner inconsistent with these guidelines. We hope to highlight this issue in a future Points of View column that provides more detailed guidelines on choosing colors that are accessible to
Figure 1 | Time-resolved WAXS difference data recorded from detergent-solubilized samples of proteorhodopsin. (a) Experimental difference data for the indicated time windows after phototriggering. ΔS is the change in X-ray scattering intensity, and q = 4π sin(θ)/λ, where θ is half the angle through which X-rays are deflected and λ is the X-ray wavelength. (b) Singular value decomposition of the experimental difference data: the principal component (difference WAXS basis spectrum; experiment) and the results of structural refinement against this basis spectrum (model; see ref. 4). (c) Time evolution of the principal component shown in b as recovered by singular value decomposition (experiment) and the fit to a monoexponential decay (fit). (d) The difference WAXS basis spectrum recorded from proteorhodopsin4 using short polychromatic X-ray pulses at the European Synchrotron Radiation Facility (ESRF) overlaid upon that recorded using monochromatic radiation at the Swiss Light Source (SLS) after convolution with the polychromatic X-ray spectrum.
0.2 1.00.80.60.4 1.2
0.2 1.00.80.60.4 1.2
−50 500 100
Experiment
Model
Experiment
Fit
SLS
ESRF
ΔS (
a.u.
)A
mpl
itude
(a.
u.)
ΔS (
a.u.
)
q (Å−1)
Time (ms)
q (Å−1)
a b
d
c
q (Å−1)0.2 1.00.80.60.4 1.2
ΔS (
a.u.
)
1
0
–1
1
0
–1
1
0
–1
1
0
–1
60–80 ms
40–60 ms
20–40 ms
0–20 ms
our readers with color blindness. In the meantime, we encourage authors to visit http://www.vischeck.com/ for more information and to download their free Photoshop and ImageJ plug-ins for recoloring images.
1. Albrecht, M. Nat. Methods 7, 775 (2010).
© 2
010
Nat
ure
Am
eric
a, In
c. A
ll ri
gh
ts r
eser
ved
.
776 | VOL.7 NO.10 | OCTOBER 2010 | nature methods
correspondence
experiment4, the difference WAXS basis spectrum recovered was almost indistinguishable from that previously reported (Fig. 1d).
From these results we established that the structural evolu-tion of the photocycle of proteorhodopsin can be followed with a temporal resolution of 10 ms by reading a Pilatus X-ray detec-tor with a repetition rate of 100 Hz. As this detector does not add readout noise to the data8, averaging over many repeats can compensate for the lower X-ray flux of a monochromatic X-ray beam. Moreover, the detector’s rapid readout enables the entire time history of the protein’s photocycle to be recorded for every reaction-triggering event. These properties will also be advanta-geous in time-resolved WAXS studies of protein reactions that are not reversible. Current developments in pixel-detector technology are expected to push the temporal resolution achievable using this approach into the microsecond realm9, facilitating the study of more rapid protein conformational changes. As this method can be implemented at almost any synchrotron radiation–based beamline suitable for protein diffraction or X-ray scattering, our approach paves the way toward a broader application of time-resolved WAXS for characterizing protein reaction dynamics.
Note: Supplementary information is available on the Nature Methods website.
acKnoWLedGmentsWe thank B. Henrich, E.F. Eikenberry and H. Billich for technical support. This work was supported by grants from the Human Frontiers Science Program and the Swedish Science Research Council.
competInG FInancIaL InterestsThe authors declare no competing financial interests.
sebastian Westenhoff1,6, erik malmerberg1,6, david arnlund1, Linda Johansson1, elena nazarenko1, marco cammarata2, Jan davidsson3, Vincent chaptal4, Jeff abramson4, Gergely Katona1, andreas menzel5 & richard neutze1
1Department of Chemistry, Biochemistry and Biophysics, University of Gothenburg, Gothenburg, Sweden. 2Stanford Linear Accelerator Center National Accelerator Laboratory, Menlo Park, California, USA. 3Department of Physical Chemistry, Uppsala University, Uppsala, Sweden. 4Department of Physiology, David Geffen School of Medicine, University of California, Los Angeles, California, USA. 5Paul Scherrer Institut, Villigen, Switzerland. 6These authors contributed equally to this work. e-mail: [email protected]
1. Cammarata, M. et al. Nat. Methods 5, 881–886 (2008).2. Ahn, S., Kim, K.H., Kim, Y., Kim, J. & Ihee, H. J. Phys. Chem. B 113, 13131–
13133 (2009).3. Cho, H.S. et al. Proc. Natl. Acad. Sci. USA 107, 7281–7286 (2010).4. Andersson, M. et al. Structure 17, 1265–1275 (2009).5. Westenhoff, S. et al. Acta Crystallogr. A 66, 207–219 (2010).6. Ihee, H., Wulff, M., Kim, J. & Adachi, S. Int. Rev. Phys. Chem. 29, 453–520
(2010).7. Cammarata, M. et al. Rev. Sci. Instrum. 80, 015101 (2009).8. Henrich, B. et al. Nucl. Instrum. Methods Phys. Res. A 607, 247–249 (2009).9. Dinapoli, R. et al. Nucl. Instrum. Methods Phys. Res. A 617, 384–386 (2010).
of large conformational changes that might otherwise be restrict-ed by protein-protein contacts in a crystal lattice. Conversely, because an ensemble of randomly oriented protein molecules is probed, structural information is averaged over all orientations and is therefore less detailed than that recovered using diffrac-tion methods6. Nevertheless, structural refinement of α-helical movements against time-resolved WAXS difference data has been reported2,4.
Previous time-resolved WAXS studies of proteins1–4 have used a laser pulse to initiate a photoreaction, and the samples had been probed using a short, highly intense polychromatic X-ray pulse, the duration of which was controlled by a mechanical X-ray chopper7. The time delay between the arrival of the laser and X-ray pulses had been controlled electronically, and the temporal resolution had been limited by the X-ray pulse dura-tion, which could be as short as 100 ps at synchrotron radiation sources. However, to our knowledge, only two highly specialized polychromatic beamlines in the world have the performance required for time-resolved WAXS studies of proteins: 1D09B of the European Synchrotron Radiation Facility and ID14B of the Advanced Photon Source.
We developed an alternative data acquisition strategy that uses a conventional monochromatic X-ray beam but records time-dependent WAXS data by using a rapid-readout, two-dimension-al photon-counting pixel detector8. We carried out experiments at the coherent small-angle X-ray scattering (cSAXS) beamline of the Swiss Light Source using samples of detergent-solubi-lized and purified proteorhodopsin concentrated to 15 mg ml–1 (Supplementary Methods). We delivered liquid samples into the monochromatic X-ray beam (11.2 keV, ~5 × 1011 photons s–1 in a 200 µm × 300 µm spot) using a 1-mm-diameter quartz capillary and photoexcited them using 5-ms green-light pulses at a repeti-tion rate of 5 Hz (Supplementary Methods and Supplementary Fig. 1). We recorded a sequence of WAXS images on a single-module Pilatus detector8 (83.8 mm × 33.5 mm; pixel size, 172 µm × 172 µm) using an integration time of 7 ms and a readout delay of 3 ms, equating to a readout frequency of 100 Hz. We recorded ~4,500 acquisition cycles (each 200-ms window cor-responding to one laser pulse) and averaged them during 3 h of experimental time.
We plotted WAXS difference data (photoactivated state minus resting state) recorded from solubilized proteorhodopsin as a func-tion of the time delay after photoactivation (Fig. 1a), with the effects of laser- and X-ray–induced heating removed4 (Supplementary Methods and Supplementary Fig. 2). A singular value decomposi-tion of these data yielded the principal component (Fig. 1b), and we approximated the decay of this component by a monoexponen-tial decay with a time constant of 34 ms (Fig. 1c). After convolu-tion with the polychromatic X-ray spectrum used in our earlier
© 2
010
Nat
ure
Am
eric
a, In
c. A
ll ri
gh
ts r
eser
ved
.