Speaker
Description
Analysis of diffraction data from time-resolved experiments suggests
incoherent addition of populations of states
Current refinement of crystal structures with excited-state
populations commonly relies on extrapolated data sets (Schmidt et al.,
2003; De Zitter et al., 2022). These are derived from differences
between diffraction data collected from crystals containing mixed
ground- and excited-state populations and data from crystals in the
ground state, each measured independently.
Having atomic models for both ground and excited states available, we
sought to refine both structures simultaneously against diffraction
data from crystals with mixed populations. Instead of extrapolation,
we partitioned the observed data into contributions from each
structural state using coherent addition of structure
factors. Difference electron-density maps were calculated and
structures refined using an updated version of the MAIN software
(Turk, 2013).
Initial tests were performed using the PYP data set collected 3 ps
after laser excitation (Pande et al., 2016; PDB 5HDS). During these
analyses we observed that coherent addition of ground- and
excited-state contributions did not perform satisfactorily. Compared
with refinement against extrapolated data, refinement against
coherently split data yielded higher partial R-factors and visually
distorted electron-density maps, particularly in omit regions of
interest. We therefore explored an alternative formulation based on
incoherent addition, analogous to diffraction from merohedral twins.
Unexpectedly, this approach produced clearer difference
electron-density maps and lower R-factors. Moreover, refinement
converged toward larger excited-state populations (approximately 30%)
compared with the initially assumed 10%. Difference maps calculated
using observed intensities (Iobs) rather than amplitudes (Fobs), while
using phases from the dark structure, further improved map clarity,
largely independent of weighting schemes (I, I/σ, SVD, Q- or
K-weighting). In the next step, the analysis was extended to
ligand-bound crystal structures of SARS-CoV-2 Mpro with partial ligand
occupancy. In these cases as well, refinement based on incoherent
data splitting consistently outperformed the coherent formulation.
Because I-based metrics apply equally to coherent and incoherent
models, we propose that refinement and goodness-of-fit assessment in
such systems should preferentially be formulated in terms of
intensities rather than structure-factor amplitudes.
These results motivate a re-examination of how structural populations
combine in diffraction from mixed-state crystals from the time resolved, ligand binding to heavy atom isomorphues replacement data interpretation.
Acknowledgements: Henry Chapman, Oleksandr Yefanov, Chufeng Li are gratefully acknowledged for constructive discussions. Marius Schmidt is gratefully acknowledged for providing data from the PYP project.
References
De Zitter, E., Coquelle, N., Oeser, P., Barends, T.R.M., Colletier,
J.-P., 2022. Xtrapol8 enables automatic elucidation of low-occupancy
intermediate-states in crystallographic studies. Commun Biol 5,
640. https://doi.org/10.1038/s42003-022-03575-7
Pande, K., Hutchison, C.D.M., Groenhof, G., Aquila, A., Robinson,
J.S., Tenboer, J., et al. 2016. Femtosecond structural
dynamics drives the trans/cis isomerization in photoactive yellow
protein. Science 352, 725–729. https://doi.org/10.1126/science.aad5081
Turk, D., 2013. MAIN software for density averaging, model building,
structure refinement and validation. Acta Crystallogr D Biol
Crystallogr 69, 1342–1357. https://doi.org/10.1107/S0907444913008408
Schmidt, M., Rajagopal, S., Ren, Z., Moffat, K., 2003. Application of
Singular Value Decomposition to the Analysis of Time-Resolved
Macromolecular X-Ray Data. Biophysical Journal 84,
2112–2129. https://doi.org/10.1016/S0006-3495(03)75018-8
| Scientific Topics | Biology |
|---|