Access to this full-text is provided by IUCr.
Content available from Acta Crystallographica Section D: Structural Biology
This content is subject to copyright.
research papers
224 https://doi.org/10.1107/S2059798323000931 Acta Cryst. (2023). D79, 224–233
Received 1 December 2022
Accepted 1 February 2023
Edited by R. J. Read, University of Cambridge,
United Kingdom
‡ These authors contributed equally.
Keywords: serial crystallography; lattice-
translocation defects; membrane proteins;
lipidic cubic phase; G protein-coupled
receptors.
PDB references: bovine rhodopsin in lipidic
cubic phase, dark state, SACLA, 7zbc; dark
state, SwissFEL, 7zbe; 1 ps light-activated, 8a6c
Supporting information:this article has
supporting information at journals.iucr.org/d
Correction of rhodopsin serial crystallography
diffraction intensities for a lattice-translocation
defect
Matthew J. Rodrigues,
a
*‡ Cecilia M. Casadei,
a,b
‡ Tobias Weinert,
a
Valerie
Panneels
a
and Gebhard F. X. Schertler
a,b
*
a
Laboratory of Biomolecular Research, Paul Scherrer Institute, 5232 Villigen PSI, Switzerland, and
b
Department of
Biology, ETH-Zurich, Zurich, Switzerland. *Correspondence e-mail: matthew.rodrigues@psi.ch,
gebhard.schertler@psi.ch
Rhodopsin is a G-protein-coupled receptor that detects light and initiates the
intracellular signalling cascades that underpin vertebrate vision. Light sensitivity
is achieved by covalent linkage to 11-cis retinal, which isomerizes upon photo-
absorption. Serial femtosecond crystallography data collected from rhodopsin
microcrystals grown in the lipidic cubic phase were used to solve the room-
temperature structure of the receptor. Although the diffraction data showed
high completeness and good consistency to 1.8 A
˚resolution, prominent
electron-density features remained unaccounted for throughout the unit cell
after model building and refinement. A deeper analysis of the diffraction
intensities uncovered the presence of a lattice-translocation defect (LTD) within
the crystals. The procedure followed to correct the diffraction intensities for this
pathology enabled the building of an improved resting-state model. The
correction was essential to both confidently model the structure of the
unilluminated state and interpret the light-activated data collected after photo-
excitation of the crystals. It is expected that similar cases of LTD will be
observed in other serial crystallography experiments and that correction will be
required in a variety of systems.
1. Introduction
The primary event in mammalian vision is the absorption of a
photon by 11-cis retinal, which is covalently linked to a lysine
side chain in rhodopsin via a protonated Schiff base (Naka-
nishi, 1991). Upon light absorption, retinal isomerizes to the
all-trans form and the receptor transitions through a number
of spectroscopically distinct intermediate states (Lewis &
Kliger, 1992; Mathies & Lugtenburg, 2000). Once in the active
state, the G protein-coupled receptor (GPCR) catalyses the
exchange of GDP for GTP in the transducin G protein to
initiate intracellular signalling cascades that result in neuronal
signalling (Bennett et al., 1982; Emeis et al., 1982).
Structural and spectroscopic studies have exploited the
abundance of rhodopsin in bovine retina to investigate how it
performs its critical role in vision. The initial resting-state
crystal structure of rhodopsin was followed by further crystal
and cryo-EM structures of the receptor in the active state, in
complex with signalling partners and in cryo-trapped inter-
mediate states following light activation (Palczewski et al.,
2000; Standfuss et al., 2011; Tsai et al., 2018, 2019; Nakamichi &
Okada, 2006a,b). However, until recently it has not been
possible to determine how the structure of the protein changes
as a function of time at physiological temperatures.
In the past decade, time-resolved serial femtosecond crys-
tallography (TR-SFX) experiments have shone light on the
photochemistry driving the activity of light-sensitive proteins
ISSN 2059-7983
Published under a CC BY 4.0 licence
(Poddar et al., 2022). In particular, the molecular mechanisms
of several retinal-dependent microbial opsins, including the
proton pump bacteriorhodopsin (Nango et al., 2016; Nogly et
al., 2018; Nass Kovacs et al., 2019; Weinert et al., 2019), the
sodium pump KR2 (Skopintsev et al., 2020), the chloride pump
NmHR (Yun et al., 2021; Mous et al., 2022) and the C1C2
channelrhodopsin cation channel (Oda et al., 2021), have been
deciphered. TR-SFX experiments with rhodopsins typically
rely on the delivery of microcrystals embedded in a viscous
medium to the interaction region (Nogly et al., 2016; James
et al., 2019), where they are illuminated by an optical pulse.
After a controlled time delay, the crystals are probed by an
X-ray pulse from a free-electron laser (FEL). As only one, still
diffraction image can be measured from an individual crystal
before it is destroyed by the FEL X-ray pulse (Chapman et al.,
2014), a single data set is typically composed of images
collected from tens of thousands of randomly oriented crys-
tals.
While TR-SFX experiments are uniquely capable of
visualizing protein conformational changes with atomic spatial
and subpicosecond temporal resolution, the sample require-
ments are often challenging to fulfil. Crystallization in the
lipidic cubic phase (LCP) has been the method of choice for
most membrane proteins prepared for TR-SFX experiments
(Landau & Rosenbusch, 1996), and LCP can also be used as a
medium to deliver crystals to the X-ray beam via a high-
viscosity sample injector (Weinert & Panneels, 2020).
Optimization of the bulk purification and LCP crystal-
lization protocols for bovine rhodopsin yielded large quan-
tities of well diffracting crystals that were suitable for
time-resolved experiments at FELs (Wu et al., 2015; Weinert &
Panneels, 2020; Gruhl et al., 2022).
It was possible to collect diffraction data sets from ‘dark-
state’ crystals, which were not illuminated, and also from
crystals 1, 10 and 100 ps after light activation at the Swiss Free
Electron Laser (SwissFEL) and the SPring-8 Angstrom
Compact Free-Electron Laser (SACLA) (Gruhl et al., 2022;
Table 1). An initial model of the dark-state structure could
be obtained by molecular replacement, after which several
iterations of model building and refinement were carried out.
Inspection of the electron-density maps revealed significant
density features in the solvent channels that could not be
successfully modelled by placing solvent molecules.
After in-depth analysis of the electron-density maps and
diffraction intensities, we were able to identify the presence of
a lattice-translocation defect within the rhodopsin crystals.
Procedures to correct rotation data collected from single
crystals exhibiting such defects have been described
previously (Wang, Kamtekar et al., 2005; Wang, Rho et al.,
2005). Here, we describe the procedure performed to detect,
characterize and correct serial crystallography data sets for
this pathology. The correction was essential to confidently
model both the dark and light-activated structures of
rhodopsin, and therefore to elucidate the first molecular
rearrangements in vertebrate vision.
2. Data collection and initial processing
2.1. Data collection
Bovine rhodopsin was purified from the native source and
crystallized in LCP as has been described in detail (Gruhl et
al., 2022). The microcrystals embedded in LCP were delivered
to the X-ray beam by a high-viscosity sample injector
(Weierstall et al., 2014) and diffraction data were collected
from crystals either without illumination by the pump laser
(dark sample) or at a defined time delay after illumination.
Data were collected as still images at SwissFEL and SACLA,
with separate dark-state data sets measured at both FELs.
Experimental parameters are detailed in Gruhl et al. (2022).
2.2. Initial processing
As described in Gruhl et al. (2022), all data collected
at SwissFEL (SF dark, 1 ps, 10 ps) were indexed using
indexamajig (White et al., 2012) with the XGANDALF algo-
rithm (Gevorkov et al., 2019). The MOSFLM,DirAx and
XGANDALF methods (Powell, 1999; Duisenberg, 1992;
Gevorkov et al., 2019) were used for the data collected at
SACLA (SACLA dark, 100 ps). Diffraction images collected
at SwissFEL were indexed in space group P22
1
2
1
with unit-cell
parameters a= 61.51, b= 91.01, c= 151.11 A
˚and data
collected at SACLA were indexed in the same space group
with unit-cell parameters a= 61.29, b= 90.81, c=150.51A
˚.
All data were integrated using indexamajig in CrystFEL
(White et al., 2012). The integration radius was set to two
pixels for SwissFEL data and three pixels for SACLA data,
while the background annulus was set to between four and six
pixels for SwissFEL data and between four and seven pixels
for data collected at SACLA. The crystal-to-detector distance
was optimized by sampling detector distances between 91.5
and 97.5 mm at SwissFEL and 47.5 and 53.5 mm at SACLA
first in 200 mm increments and then in 20 mm increments to
determine the detector distance at which the standard devia-
tions of the unit-cell parameters were minimized.
Initial phasing by molecular replacement was performed
with Phaser (McCoy et al., 2007) using a deposited structure of
research papers
Acta Cryst. (2023). D79, 224–233 Matthew J. Rodrigues et al. Lattice-translocation defects in serial crystallography 225
Table 1
Merging statistics for dark-state data sets collected at SwissFEL and
SACLA (Gruhl et al., 2022).
Values in parentheses are for the highest resolution shell.
Data set SwissFEL SACLA
Resolution range (A
˚) 16.10–1.80 (1.86–1.80) 10.47–1.80 (1.86–1.80)
a,b,c(A
˚) 61.51, 91.01, 151.11 61.29, 90.81, 150.51
,,() 90.0, 90.0, 90.0 90.0, 90.0, 90.0
Space group P22
1
2
1
P22
1
2
1
Measured reflections 65870940 (4371475) 105018485 (7436654)
Unique reflections 79305 (7852) 78209 (7715)
Multiplicity 830.6 (556.7) 1342.8 (963.9)
Completeness (%) 100.0 (100.0) 100.0 (100.0)
hI/(I)i7.62 (0.95) 8.87 (1.32)
R
split
(%) 8.21 (109.06) 6.75 (77.70)
CC* 0.9982 (0.8947) 0.9994 (0.9219)
CC
1/2
0.9926 (0.6672) 0.9977 (0.7391)
Translation vector (t
d
) 0.245 0.243
Translated fraction
0
= 0.01,
1
= 0.19
0
= 0.00,
1
= 0.15
rhodopsin at cryogenic temperature (PDB entry 1u19; Okada
et al., 2004), with solvent and ligand molecules removed, as a
search model. The asymmetric unit contains two rhodopsin
molecules arranged as an antiparallel dimer. Rotation of this
antiparallel dimer around the crystallographic screw axes
generates translational noncrystallographic symmetry (tNCS),
which was detected as a peak with a magnitude 65% of the
origin peak height at position 0.5a+ 0.378b+0.5cin the native
Patterson map with phenix.xtriage (Liebschner et al., 2019).
Multiple iterations of model building in Coot (Casan
˜al et al.,
2020) and refinement with phenix.refine (Liebschner et al.,
2019) were carried out until convergence of the R
free
statistics.
Most metrics of model quality were reasonable at this stage
(see Tables 2 and 3). In particular, model refinement against
dark-state data collected at SwissFEL yielded R
work
and R
free
values of 26.65% and 28.27%, respectively, in the resolution
range 16.1–1.8 A
˚, showing that the modelled structure
explained the experimental observations well.
3. Indications of lattice-translocation disorder in real
space
Although the R-factor statistics, which quantify the agreement
between the observed and model structure factors, converged
to values which are commonly considered to be acceptable in
serial crystallography, inspection of the electron-density maps
indicated that our model did not appropriately describe the
underlying data. As is common practice, we calculated
A
-
weighted electron-density difference maps, where the Fourier
coefficients (mF
o
DF
c
)exp(i’
c
), with observed structure-
factor amplitudes F
o
and model amplitudes and phases F
c
,’
c
,
include weighting factors mand Dto account for the coordi-
nate errors within the model and the partiality of the modelled
structure (Read, 1986). These maps are less affected by model
bias compared with (F
o
F
c
)exp(i’
c
) maps and are particu-
larly useful to identify errors in the model and aid model
building. In particular, positive difference peaks correspond to
features of the electron density which are unaccounted for by
the model.
Significant positive difference electron-density features
were visible both in the solvent channels between rhodopsin
molecules and overlapping with the rhodopsin dimer (Fig. 1).
This difference density could not be sensibly modelled by the
placement of solvent molecules. Furthermore, some of the
observed features resembled amino-acid side chains (Fig. 2a).
However, there was insufficient space between the protein
molecules to accommodate an additional rhodopsin chain.
We were able to model a single tryptophan residue into a
particularly prominent positive difference electron-density
feature in the solvent channel (Fig. 2a) and then to least-
squares align tryptophan-centred tripeptides from the
rhodopsin structure to the placed tryptophan residue in Coot
(Fig. 2b). In this way, we could assign this unexplained density
feature to Trp265. Least-squares alignment of the entire
rhodopsin dimer to this tryptophan residue resulted in the
model overlapping well with the unexplained density (Fig. 2c).
This second copy of the dimer was related to the original
structure by a translation of 22.5 A
˚along the unit-cell baxis,
corresponding to approximately one-quarter of the length of
the baxis. As the original dimer and the translated copy
spatially overlap, it is not possible for both copies to be
present in the same unit cell.
This observation was reminiscent of several previous
observations of lattice-translocation disorder (LTD), namely
the presence of translation-related domains in the crystals, in a
number of systems (Bragg & Howells, 1954; Pickersgill, 1987;
Wang, Rho et al., 2005; Hare et al., 2009; Tsai et al., 2009;
Ponnusamy et al., 2014; Li et al., 2020). We therefore sought to
determine whether this was another such case and whether the
data could be corrected as previously described in Wang,
Kamtekar et al. (2005).
As a first step, a Fourier transform of the merged intensities
was performed to produce a Patterson map, which represents
the autocorrelation function of the electron density in real
space and exhibits prominent peaks at positions corre-
sponding to highly frequent interatomic and intermolecular
vectors in the structure. By inspecting this map, we observed a
prominent peak at the position 0.5a+ 0.378b+0.5c, where
research papers
226 Matthew J. Rodrigues et al. Lattice-translocation defects in serial crystallography Acta Cryst. (2023). D79, 224–233
Table 2
Refinement statistics for the SwissFEL dark-state model refined against
the original and the corrected diffraction intensities.
Original intensities Corrected intensities
R
work
/R
free
(%) 26.65/28.27 21.43/24.61
No. of atoms
Protein 4971 4971
Ligand 526 526
Water 174 174
Bfactors (A
˚
2
)
Protein 31.60 32.20
Ligand 46.77 49.08
Water 38.20 40.23
Ramachandran statistics
Favoured (%) 96.48 96.81
Allowed (%) 3.52 3.19
Outliers (%) 0.00 0.00
R.m.s. deviations
Bond angles () 0.009 0.007
Bond lengths (A
˚) 0.928 0.847
Table 3
Refinement statistics for the SACLA dark-state model refined against the
original and the corrected diffraction intensities.
Original intensities Corrected intensities
R
work
/R
free
(%) 22.38/23.68 19.50/22.11
No. of atoms
Protein 4970 4970
Ligand 539 539
Water 168 168
Bfactors (A
˚
2
)
Protein 30.48 31.16
Ligand 47.53 49.74
Water 38.13 39.88
Ramachandran statistics
Favoured (%) 96.64 97.14
Allowed (%) 3.36 2.86
Outliers (%) 0.00 0.00
R.m.s. deviations
Bond angles () 0.005 0.009
Bond lengths (A
˚) 0.738 0.977
a,b,care the unit-cell axes, which was accounted for by the
noncrystallographic symmetry operation relating the two
copies of the molecule in the asymmetric unit. We also
observed a second prominent peak, with a magnitude of 18%
of the origin peak for the SACLA data and 25% of the origin
peak for the SwissFEL data, at 0.245b(Fig. 3). We attributed
this peak to the vector t
d
relating the two translation-related
domains.
4. Indications of lattice-translocation disorder in
reciprocal space
We assume that terms from translation-related domains
corresponding to integer multiples of the fundamental trans-
lation t
d
contribute to the structure factors of an LTD-affected
crystal. Following and generalizing the treatment of LTD
presented in Wang, Kamtekar et al. (2005), we model such
structure factors with the weighted sum of an infinite number
of translation-related terms,
FoðhÞ¼ P
1
n¼0
nexp½i2ðntdÞh
FdðhÞ;ð1Þ
where hdenotes the reciprocal-lattice point ha*+kb*+lc*
with integer h,k,land reciprocal-lattice basis vectors a*, b*, c*.
Real-space domain translations are given by the set {nt
d
}of
integer multiples of the fundamental translation t
d
, and the
weights
n
represent the unit-cell fractions pertaining to each
domain. The single-domain structure factor F
d
is given by the
usual sum over contributions from all atoms within one unit
cell,
FdðhÞ¼P
j
fjexpði2xjhÞ;ð2Þ
where f
j
is the atomic form factor and x
j
is the position of atom
jin the unit cell. The structure factor of a translated domain is
Ftr
d¼P
j
fjexp½i2ðxjþntdÞh¼expði2ntdhÞFdðhÞ:ð3Þ
Inspection of the Patterson map provides the estimate
t
d
=t
d
b= 0.245b. Because high-order terms are progressively
less relevant, we truncate the summation in equation (1) to the
term with n= 2, obtaining
research papers
Acta Cryst. (2023). D79, 224–233 Matthew J. Rodrigues et al. Lattice-translocation defects in serial crystallography 227
Figure 2
Rhodopsin dimer in cartoon format (pink) overlaid with
A
-weighted F
o
F
c
difference density contoured at 2.0(green mesh). (a) Trp265 (cyan) fitted
into
A
-weighted F
o
F
c
difference density. (b) Cys264, Trp265 and Leu266 (cyan) aligned to the fitted Trp265. Transmembrane helices 5 (left) and 6
(right) are visible behind the translated peptide. (c) A translated rhodopsin dimer (cyan stick format) aligned with Trp265 overlaid with the main
rhodopsin dimer.
Figure 1
Rhodopsin dimer (pink) and symmetry-related rhodopsin molecules (pale orange) shown in cartoon format with
A
-weighted F
o
F
c
difference density
contoured at 4.0overlaid (green). The unit cell is shown as a green box.
FoðhÞ’½0þ1expði2tdhÞþ2expði4tdhÞFdðhÞ
¼½0þ1expði2tdkÞþ2expði4tdkÞFdðhÞ:ð4Þ
As a result of squaring the complex structure factors to obtain
diffraction intensities Io¼FoF
o, interference terms (
0
1
+
1
2
)[exp(i2t
d
k) + exp(i2t
d
k)] and
0
2
[exp(i4t
d
k)+
exp(i4t
d
k)] appear. This gives rise to a characteristic
oscillatory behaviour of the observed intensities as a function
of reciprocal space index k,
IoðhÞ’½2
0þ2
1þ2
2þ2ð01þ12Þcosð2tdkÞ
þ202cosð4tdkÞIdðhÞ;ð5Þ
with Id¼FdF
d. This oscillatory behaviour can be recognized
in Fig. 4, where measured intensities, averaged over reciprocal-
space indices (h,l), are shown as a function of index k.
5. Correction of diffraction intensities
The inspection and analysis of the electron-density maps, the
Patterson map and the merged intensities supported the
diagnosis of LTD within the rhodopsin crystals. Tentative
domain fractions
0
,
1
and
2
were sampled with steps of 0.01
in the range [0, 1]. For each triplet (
0
,
1
,
2
) satisfying
2¼101ð6Þ
with
00121;ð7Þ
we estimated individual domain intensities by inversion of
equation (5) as follows:
Idðh;0;
1;
2Þ’½2
0þ2
1þ2
2þ2ð01þ12Þcosð2tdkÞ
þ202cosð4tdkÞ1IoðhÞ:ð8Þ
For each set of corrected intensities I
d
(h;
0
,
1
,
2
), we
calculated the Patterson map value at positions t
d
band 2t
d
b.
The values of j~
Id
Idð0;tdb;0Þj and j~
Id
Idð0;2tdb;0Þj, where ~
Id
Idis the
Fourier transform of I
d
, are reported in Figs. 5 and 6 for the
SwissFEL and SACLA data, respectively, as a function of
(
0
,
1
). We optimized the domain fractions by selecting the
research papers
228 Matthew J. Rodrigues et al. Lattice-translocation defects in serial crystallography Acta Cryst. (2023). D79, 224–233
Figure 4
Diffraction intensities from SwissFEL (a) and SACLA (b), averaged over
reciprocal-space indices hand l, as a function of index kbefore and after
correction with
0
= 0.01,
1
= 0.19,
2
= 0.80 (SwissFEL) and
0
= 0.00,
1
= 0.15,
2
= 0.85 (SACLA).
Figure 3
Patterson map values for SwissFEL data (a) and SACLA data (b) along the direction 0^
aa þy^
bb þ0^
cc, where ^
aa;^
bb;^
cc are unit vectors aligned with the
corresponding unit-cell basis vectors before and after correction of the intensities with
0
= 0.01,
1
= 0.19,
2
= 0.80 (SwissFEL) and
0
= 0.00,
1
= 0.15,
2
= 0.85 (SACLA). The peak height at 0.245bis equivalent to 25% and 18% of the origin peak heights for uncorrected SwissFEL and SACLA data,
respectively.
triplet resulting in small Patterson peaks at both examined
positions. We obtained
0
= 0.01,
1
= 0.19 and
2
= 0.80 for the
SwissFEL data set and
0
= 0.00,
1
= 0.15 and
2
= 0.85 for the
SACLA data set. Fig. 3 shows a comparison of the Patterson
map values before and after correction along the direction of
the unit-cell axis b. Upon correction of the intensities using the
optimized values, a dampening of the oscillations of (h,l)-
averaged values as a function of index kwas observed (Fig. 4).
With the definitions
0¼2
0þ2
1þ2
2;ð9Þ
1¼2ð01þ12Þ;ð10Þ
and
2¼202;ð11Þ
equation (5) can be rewritten
IoðhÞ’½0þ1cosð2tdkÞþ2cosð4tdkÞIdðhÞ;ð12Þ
with weights
0
= 0.6762,
1
= 0.3078,
2
= 0.016 for the
SwissFEL data and
0
= 0.745,
1
= 0.255,
2
= 0 for the
SACLA data. The ratios
1
/
0
(0.46 and 0.34 for the SwissFEL
and SACLA data, respectively) and
2
/
0
(0.02 and 0 for the
SwissFEL and SACLA data, respectively) express the relative
importance of interference terms for increasing relative
displacement and show that it is well justified to neglect terms
with n> 2 in equation (1).
Translation correction factors were determined separately
for each of the data sets. Refinement of the rhodopsin dark-
state atomic model against the LTD-corrected intensities from
SwissFEL yielded an immediate improvement in both the
R
work
and R
free
statistics (21.43% and 24.61%, respectively)
and the model geometry. More importantly, the amount of
unexplained mF
o
DF
c
difference density was much reduced
research papers
Acta Cryst. (2023). D79, 224–233 Matthew J. Rodrigues et al. Lattice-translocation defects in serial crystallography 229
Figure 6
Absolute values of the Patterson function of corrected SACLA intensities with varying
0
,
1
and
2
=1
0
1
at real-space positions t
d
b(a) and 2t
d
b
(b). Correction of intensities with
0
= 0.00,
1
= 0.15 (black circle) dampens the Patterson peaks at both positions simultaneously.
Figure 5
Absolute values of the Patterson function of corrected SwissFEL intensities with varying
0
,
1
and
2
=1
0
1
at real-space positions t
d
b(a) and
2t
d
b(b). Correction of intensities with
0
= 0.01,
1
= 0.19 (black circle) dampens the Patterson peaks at both positions simultaneously.
in the resulting electron-density maps. This allowed further
model building to remove inappropriately placed solvent
molecules, fix distorted amino-acid side-chain rotamers and
build more of the flexible loops in the protein. The dark-state
final models therefore accounted better for the experimental
data. Improved model phases were available for calculation
of the light-to-dark electron-density difference maps, with
Fourier coefficients ðFlight
oFdark
oÞexpði’dark
cÞ, derived from
the observed amplitudes Fdark
oand Flight
oof the dark and light-
activated state, respectively, and dark model phases ’dark
c.
These maps are essential to interpret the conformational
changes following light activation.
There is a stark contrast between the mF
o
DF
c
maps
produced after refinement of the final dark-state model
against the diffraction intensities with and without correction,
demonstrating the importance of the correction in obtaining
an accurate dark-state model (Figs. 7 and 8). As the dark-state
model serves as the foundation for interpreting light-induced
structural changes, the correction described was required to
confidently derive biological insights into the very first
processes in receptor activation (Gruhl et al., 2022).
6. Discussion
Several examples of LTD in crystals of soluble proteins have
been published since the first observation in 1954 (Bragg &
Howells, 1954). Furthermore, occurrences of LTD may be
underreported, as this pathology is difficult to detect and may
often be ignored when the translated fraction is small
(Lovelace & Borgstahl, 2020). LTD has also been identified in
LCP-grown crystals of the rice silicon transporter Lsi1 (Saitoh
et al., 2021). Similar to our observation of LTD in rhodopsin
crystals, the Lsi1 lattice translation appears to occur parallel
to the plane of the membrane. Indeed, lattice translations
perpendicular to the plane of the LCP bilayer seem improb-
able as the hydrophobic transmembrane region of the protein
would be displaced from the hydrophobic lipid environment
to the aqueous environment between the lipid bilayers.
Instead, type I LCP crystals, which assemble as stacks of 2D
crystals, appear to be susceptible to translations in the relative
positions of these 2D layers, which result in translations
parallel to the plane of the membrane.
In the previously observed cases of LTD the data were
collected using the rotation method (Dauter, 1999) and it was
possible to diagnose the disorder by direct inspection of the
diffraction images. In Wang, Kamtekar et al. (2005), for
example, a translation of 0.5 in the caxis in Bacillus phage ’29
DNA polymerase crystals resulted in the appearance of weak
and streaky spots for reflections with odd values of land
strong spots for even values of l, caused by the main lattice and
the translated lattice scattering out of phase and in phase with
odd and even values of l, respectively.
Such an analysis is not possible with serial crystallography
data, as the measured intensities from individual frames are
partial and the intensity of a given reflection cannot be
determined by inspection of a single diffraction image.
Instead, it is necessary to search for non-origin peaks in the
Patterson map that are too close to the origin to be explained
by tNCS.
Despite working with a single protein purification and
crystallization protocol, we observed some variability in the
determined values of the translated domain fractions
research papers
230 Matthew J. Rodrigues et al. Lattice-translocation defects in serial crystallography Acta Cryst. (2023). D79, 224–233
Figure 7
Map showing mF
o
DF
c
difference electron density after refinement of the SwissFEL dark-state model against (a) the original diffraction intensities
(contoured at 0.514 e
A
˚
3
, equivalent to 4.00) and (b) the corrected intensities (contoured at 0.514 e
A
˚
3
, equivalent to 4.96).
depending on the examined data set. While data sets from the
SwissFEL beamtime showed
0
and
1
values close to 0.01 and
0.19, respectively, we found that data sets collected during a
previous beam time at SACLA had lower translated domain
fractions, with values of
0
= 0.00 and
1
= 0.15. The unex-
plained difference density was therefore less prominent after
model refinement using uncorrected intensities from the
SACLA data set. Although the crystalline sample was
prepared using the same protocol for both beam times, subtle
differences in the sample preparation such as the crystal-
lization temperature or precipitant composition may affect the
probability of a lattice translocation occurring during crystal
growth. Such a variation in the translated fraction has
previously been observed in crystals of lentiviral integrase in
complex with LEDGF and was ascribed to differences in
crystal-growth conditions (Hare et al., 2009). It is also possible
that the smaller beam size used at SACLA, 1 1mm
compared with 5 5mm at SwissFEL, reduced the chance of
collecting data from a region of the crystal containing a lattice-
translocation defect.
The translated domain fractions were found to be similar
for all data sets collected during the same beam time. This is to
be expected as both dark and light data sets were collected
from the same sample batches and the translated domain
fractions are the result of an average over all diffracting
crystals contributing to each data set.
We attempted the refinement of a composite atomic model
containing the main dimer at occupancy (1 ) and a second
model of the dimer with occupancy , translated from the main
model by t
d
, against the uncorrected data, as was previously
successful for the 1918 H1N1 neuraminidase (Zhu et al., 2008).
The introduction of an additional copy of the dimer during
refinement with our serial crystallography rhodopsin data
necessitates the use of strict NCS constraints to avoid signifi-
cantly worsening the data-to-parameter ratio. However, the
refinement of a composite model against uncorrected data was
deemed impractical with bovine rhodopsin, as interpretation
of the 2mF
o
DF
c
electron-density maps was challenging in
the region where the two models overlapped. As previously
observed by Zhu et al. (2008), we found that correction of the
data for LTD was required to model solvent molecules in the
overlapped region accurately. The positions of ordered water
molecules are of key interest in studies of receptor dynamics
as they often stabilize hydrogen-bond networks in the transi-
tions between conformational states.
We did not observe dramatic changes in the time-resolved
Flight
oFdark
odifference maps as a result of the correction
(Fig. 9). This is because the light and dark data sets were
collected from the same sample using an interleaved sequence
of visible light laser pulses to pump the sample and X-rays to
probe the structure (Nango et al., 2019). The systematic error
due to LTD is therefore very similar in both data sets, usually
with less than 1% difference in the translated populations.
Instead, the most significant effect of LTD was to increase the
number of poorly modelled regions in the dark-state models.
As the dark-state model was taken as a starting point for the
refinement of light-state structures into the extrapolated maps
generated from the light-activated data sets, correction for
LTD improved our models of both the dark and light-
activated rhodopsin structures.
research papers
Acta Cryst. (2023). D79, 224–233 Matthew J. Rodrigues et al. Lattice-translocation defects in serial crystallography 231
Figure 8
Map showing mF
o
DF
c
difference electron density after refinement of the SACLA dark-state model against (a) the original diffraction intensities
(contoured at 1.071 e
A
˚
3
, equivalent to 4.00) and (b) the corrected intensities (contoured at 1.071 e
A
˚
3
, equivalent to 11.04).
7. Conclusions
TR-SFX experiments provide unprecedented opportunities to
understand ultrafast biological processes, including the photo-
isomerization of retinal in rhodopsin, which is the first event in
mammalian vision. Having overcome many of the challenging
barriers to performing a TR-SFX experiment with rhodopsin,
we identified the LTD in the crystals at the data-processing
stage. The described correction facilitated the interpretation
of the data collected during these time- and resource-intensive
experiments (Gruhl et al., 2022). Given the demanding pres-
sures on sample preparation and beam time, it is often not
possible to repeat an experiment after optimization of the
crystals to avoid the LTD.
Here, we show the identification and characterization of
LTD in a serial crystallography experiment. As these experi-
ments, including time-resolved studies, become more wide-
spread, it is likely that the present method of LTD correction
will be of choice.
Acknowledgements
Open access funding provided by ETH-Bereich Forschung-
sanstalten.
Funding information
M. J. Rodrigues received funding from the European Union’s
Horizon 2020 research and innovation programme under
Marie Skłodowska-Curie grant agreement No. 701647. This
research project was funded by Swiss National Science
Foundation Grant No. 192760 to G. F. X. Schertler.
References
Bennett, N., Michel-Villaz, M. & Ku
¨hn, H. (1982). Eur. J. Biochem.
127, 97–103.
Bragg, W. L. & Howells, E. R. (1954). Acta Cryst. 7, 409–411.
Casan
˜al, A., Lohkamp, B. & Emsley, P. (2020). Protein Sci. 29, 1069–
1078.
Chapman, H. N., Caleman, C. & Timneanu, N. (2014). Phil. Trans. R.
Soc. B,369, 20130313.
Dauter, Z. (1999). Acta Cryst. D55, 1703–1717.
Duisenberg, A. J. M. (1992). J. Appl. Cryst. 25, 92–96.
Emeis, D., Ku
¨hn, H., Reichert, J. & Hofmann, K. (1982). FEBS Lett.
143, 29–34.
Gevorkov, Y., Yefanov, O., Barty, A., White, T. A., Mariani, V.,
Brehm, W., Tolstikova, A., Grigat, R.-R. & Chapman, H. N. (2019).
Acta Cryst. A75, 694–704.
Gruhl, T., Weinert, T., Rodrigues, M., Milne, C. J., Ortolani, G., Nass,
K., Nango, E., Sen, S., Johnson, P. J. M., Cirelli, C., Furrer, A., Mous,
S., Skopintsev, P., James, D., Dworkowski, F., Ba
˚th, P., Kekilli, D.,
Ozerov, D., Tanaka, R., Glover, H., Bacellar, C., Bru
¨nle, S., Casadei,
C. M., Diethelm, A. D., Gashi, D., Gotthard, G., Guixa
`-Gonza
´lez,
R., Joti, Y., Kabanova, V., Knopp, G., Lesca, E., Ma, P., Martiel, I.,
Mu
¨hle, J., Owada, S., Pamula, F., Sarabi, D., Tejero, O., Tsai, C.-J.,
Varma, N., Wach, A., Boutet, S., Tono, K., Nogly, P., Deupi, X.,
Iwata, S., Neutze, R., Standfuss, J., Schertler, G. F. & Panneels, V.
(2022). bioRxiv, 2022.10.14.511948.
Hare, S., Cherepanov, P. & Wang, J. (2009). Acta Cryst. D65, 966–973.
James, D., Weinert, T., Skopintsev, P., Furrer, A., Gashi, D., Tanaka,
T., Nango, E., Nogly, P. & Standfuss, J. (2019). J. Vis. Exp., e59087.
Landau, E. & Rosenbusch, J. (1996). Proc. Natl Acad. Sci. USA,93,
14532–14535.
Lewis, J. W. & Kliger, D. S. (1992). J. Bioenerg. Biomembr. 24, 201–
210.
Li, L., Dai, S., Gao, G. F. & Wang, J. (2020). Acta Cryst. D76, 1057–
1064.
Liebschner, D., Afonine, P. V., Baker, M. L., Bunko
´czi, G., Chen,
V. B., Croll, T. I., Hintze, B., Hung, L.-W., Jain, S., McCoy, A. J.,
Moriarty, N. W., Oeffner, R. D., Poon, B. K., Prisant, M. G., Read,
R. J., Richardson, J. S., Richardson, D. C., Sammito, M. D., Sobolev,
O. V., Stockwell, D. H., Terwilliger, T. C., Urzhumtsev, A. G.,
Videau, L. L., Williams, C. J. & Adams, P. D. (2019). Acta Cryst.
D75, 861–877.
Lovelace, J. J. & Borgstahl, G. E. O. (2020). Crystallogr. Rev. 26, 3–50.
Mathies, R. & Lugtenburg, J. (2000). Molecular Mechanisms in Visual
Transduction, edited by D. Stavenga, W. DeGrip & E. Pugh, pp. 55–
90. Amsterdam: North-Holland.
McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D.,
Storoni, L. C. & Read, R. J. (2007). J. Appl. Cryst. 40, 658–674.
Mous, S., Gotthard, G., Ehrenberg, D., Sen, S., Weinert, T., Johnson,
P. J. M., James, D., Nass, K., Furrer, A., Kekilli, D., Ma, P., Bru
¨nle, S.,
Casadei, C. M., Martiel, I., Dworkowski, F., Gashi, D., Skopintsev,
P., Wranik, M., Knopp, G., Panepucci, E., Panneels, V., Cirelli, C.,
Ozerov, D., Schertler, G. F. X., Wang, M., Milne, C., Standfuss, J.,
Schapiro, I., Heberle, J. & Nogly, P. (2022). Science,375, 845–851.
Nakamichi, H. & Okada, T. (2006a). Angew. Chem. Int. Ed. 45, 4270–
4273.
Nakamichi, H. & Okada, T. (2006b). Proc. Natl Acad. Sci. USA,103,
12729–12734.
Nakanishi, K. (1991). Am. Zool. 31, 479–489.
research papers
232 Matthew J. Rodrigues et al. Lattice-translocation defects in serial crystallography Acta Cryst. (2023). D79, 224–233
Figure 9
Isomorphous Flight
oFdark
odifference electron density within 3.2 A
˚of the
retinal chromophore 1 ps after light activation. Dark-state (SwissFEL) C
atoms of Lys296 are shown in red, N atoms in purple and retinal atoms in
pink. 1 ps light-activated (SwissFEL) C atoms of Lys296 are shown in
dark green, N atoms in cyan and retinal atoms in light green. (a) Original
diffraction intensities (contoured at 4.0, equivalent to 0.075 e
A
˚
3
) and
(b) corrected intensities (contoured at 4.0, equivalent to 0.075 e
A
˚
3
).
Nango, E., Kubo, M., Tono, K. & Iwata, S. (2019). Appl. Sci. 9, 5505.
Nango, E., Royant, A., Kubo, M., Nakane, T., Wickstrand, C., Kimura,
T., Tanaka, T., Tono, K., Song, C., Tanaka, R., Arima, T., Yamashita,
A., Kobayashi, J., Hosaka, T., Mizohata, E., Nogly, P., Sugahara, M.,
Nam, D., Nomura, T., Shimamura, T., Im, D., Fujiwara, T.,
Yamanaka, Y., Jeon, B., Nishizawa, T., Oda, K., Fukuda, M.,
Andersson, R., Ba
˚th, P., Dods, R., Davidsson, J., Matsuoka, S.,
Kawatake, S., Murata, M., Nureki, O., Owada, S., Kameshima, T.,
Hatsui, T., Joti, Y., Schertler, G., Yabashi, M., Bondar, A. N.,
Standfuss, J., Neutze, R. & Iwata, S. (2016). Science,354, 1552–1557.
Nass Kovacs, G., Colletier, J.-P., Gru
¨nbein, M. L., Yang, Y., Stensitzki,
T., Batyuk, A., Carbajo, S., Doak, R. B., Ehrenberg, D., Foucar, L.,
Gasper, R., Gorel, A., Hilpert, M., Kloos, M., Koglin, J. E.,
Reinstein, J., Roome, C. M., Schlesinger, R., Seaberg, M., Shoeman,
R. L., Stricker, M., Boutet, S., Haacke, S., Heberle, J., Heyne, K.,
Domratcheva, T., Barends, T. R. M. & Schlichting, I. (2019). Nat.
Commun. 10, 3177.
Nogly, P., Panneels, V., Nelson, G., Gati, C., Kimura, T., Milne, C.,
Milathianaki, D., Kubo, M., Wu, W., Conrad, C., Coe, J., Bean, R.,
Zhao, Y., Ba
˚th, P., Dods, R., Harimoorthy, R., Beyerlein, K. R.,
Rheinberger, J., James, D., DePonte, D., Li, C., Sala, L., Williams,
G. J., Hunter, M. S., Koglin, J. E., Berntsen, P., Nango, E., Iwata, S.,
Chapman, H. N., Fromme, P., Frank, M., Abela, R., Boutet, S.,
Barty, A., White, T. A., Weierstall, U., Spence, J., Neutze, R.,
Schertler, G. & Standfuss, J. (2016). Nat. Commun. 7, 12314.
Nogly, P., Weinert, T., James, D., Carbajo, S., Ozerov, D., Furrer, A.,
Gashi, D., Borin, V., Skopintsev, P., Jaeger, K., Nass, K., Bath, P.,
Bosman, R., Koglin, J., Seaberg, M., Lane, T., Kekilli, D., Brunle, S.,
Tanaka, T., Wu, W., Milne, C., White, T., Barty, A., Weierstall, U.,
Panneels, V., Nango, E., Iwata, S., Hunter, M., Schapiro, I.,
Schertler, G., Neutze, R. & Standfuss, J. (2018). Science,361,
eaat0094.
Oda, K., Nomura, T., Nakane, T., Yamashita, K., Inoue, K., Ito, S.,
Vierock, J., Hirata, K., Maturana, A. D., Katayama, K., Ikuta, T.,
Ishigami, I., Izume, T., Umeda, R., Eguma, R., Oishi, S., Kasuya, G.,
Kato, T., Kusakizako, T., Shihoya, W., Shimada, H., Takatsuji, T.,
Takemoto, M., Taniguchi, R., Tomita, A., Nakamura, R., Fukuda,
M., Miyauchi, H., Lee, Y., Nango, E., Tanaka, R., Tanaka, T.,
Sugahara, M., Kimura, T., Shimamura, T., Fujiwara, T., Yamanaka,
Y., Owada, S., Joti, Y., Tono, K., Ishitani, R., Hayashi, S., Kandori,
H., Hegemann, P., Iwata, S., Kubo, M., Nishizawa, T. & Nureki, O.
(2021). eLife,10, e62389.
Okada, T., Sugihara, M., Bondar, A. N., Elstner, M., Entel, P. & Buss,
V. (2004). J. Mol. Biol. 342, 571–583.
Palczewski, K., Kumasaka, T., Hori, T., Behnke, C. A., Motoshima,
H., Fox, B. A., Le Trong, I., Teller, D. C., Okada, T., Stenkamp, R.
E., Yamamoto, M. & Miyano, M. (2000). Science,289, 739–745.
Pickersgill, R. W. (1987). Acta Cryst. A43, 502–506.
Poddar, H., Heyes, D. J., Schiro
`, G., Weik, M., Leys, D. & Scrutton,
N. S. (2022). FEBS J. 289, 576–595.
Ponnusamy, R., Lebedev, A. A., Pahlow, S. & Lohkamp, B. (2014).
Acta Cryst. D70, 1680–1694.
Powell, H. R. (1999). Acta Cryst. D55, 1690–1695.
Read, R. J. (1986). Acta Cryst. A42, 140–149.
Saitoh, Y., Mitani-Ueno, N., Saito, K., Matsuki, K., Huang, S., Yang,
L., Yamaji, N., Ishikita, H., Shen, J., Ma, J. F. & Suga, M. (2021).
Nat. Commun. 12, 6236.
Skopintsev, P., Ehrenberg, D., Weinert, T., James, D., Kar, R. K.,
Johnson, P. J. M., Ozerov, D., Furrer, A., Martiel, I., Dworkowski,
F., Nass, K., Knopp, G., Cirelli, C., Arrell, C., Gashi, D., Mous, S.,
Wranik, M., Gruhl, T., Kekilli, D., Bru
¨nle, S., Deupi, X., Schertler,
G. F. X., Benoit, R. M., Panneels, V., Nogly, P., Schapiro, I., Milne,
C., Heberle, J. & Standfuss, J. (2020). Nature,583, 314–318.
Standfuss, J., Edwards, P. C., D’Antona, A., Fransen, M., Xie, G.,
Oprian, D. D. & Schertler, G. F. X. (2011). Nature,471, 656–660.
Tsai, C.-J., Marino, J., Adaixo, R., Pamula, F., Muehle, J., Maeda, S.,
Flock, T., Taylor, N. M., Mohammed, I., Matile, H., Dawson, R. J. P.,
Deupi, X., Stahlberg, H. & Schertler, G. (2019). eLife,8, e46041.
Tsai, C.-J., Pamula, F., Nehme
´,R.,Mu
¨hle, J., Weinert, T., Flock, T.,
Nogly, P., Edwards, P. C., Carpenter, B., Gruhl, T., Ma, P., Deupi, X.,
Standfuss, J., Tate, C. G. & Schertler, G. F. X. (2018). Sci. Adv. 4,
eaat7052.
Tsai, Y., Sawaya, M. R. & Yeates, T. O. (2009). Acta Cryst. D65, 980–
988.
Wang, J., Kamtekar, S., Berman, A. J. & Steitz, T. A. (2005). Acta
Cryst. D61, 67–74.
Wang, J., Rho, S.-H., Park, H. H. & Eom, S. H. (2005). Acta Cryst.
D61, 932–941.
Weierstall, U., James, D., Wang, C., White, T. A., Wang, D., Liu, W.,
Spence, J. C. H., Doak, R. B., Nelson, G., Fromme, P., Fromme, R.,
Grotjohann, I., Kupitz, C., Zatsepin, N. A., Liu, H., Basu, S.,
Wacker, D., Han, G. W., Katritch, V., Boutet, S., Messerschmidt, M.,
Williams, G. J., Koglin, J. E., Seibert, M. M., Klinker, M., Gati, C.,
Shoeman, R. L., Barty, A., Chapman, H. N., Kirian, R. A.,
Beyerlein, K. R., Stevens, R. C., Li, D., Shah, S. T. A., Howe, N.,
Caffrey, M. & Cherezov, V. (2014). Nat. Commun. 5, 3309.
Weinert, T. & Panneels, V. (2020). Methods Mol. Biol. 2127, 321–338.
Weinert, T., Skopintsev, P., James, D., Dworkowski, F., Panepucci, E.,
Kekilli, D., Furrer, A., Bru
¨nle, S., Mous, S., Ozerov, D., Nogly, P.,
Wang, M. & Standfuss, J. (2019). Science,365, 61–65.
White, T. A., Kirian, R. A., Martin, A. V., Aquila, A., Nass, K., Barty,
A. & Chapman, H. N. (2012). J. Appl. Cryst. 45, 335–341.
Wu, W., Nogly, P., Rheinberger, J., Kick, L. M., Gati, C., Nelson, G.,
Deupi, X., Standfuss, J., Schertler, G. & Panneels, V. (2015). Acta
Cryst. F71, 856–860.
Yun, J.-H., Li, X., Yue, J., Park, J.-H., Jin, Z., Li, C., Hu, H., Shi, Y.,
Pandey, S., Carbajo, S., Boutet, S., Hunter, M. S., Liang, M., Sierra,
R. G., Lane, T. J., Zhou, L., Weierstall, U., Zatsepin, N. A., Ohki,
M., Tame, J. R. H., Park, S.-Y., Spence, J. C. H., Zhang, W., Schmidt,
M., Lee, W. & Liu, H. (2021). Proc. Natl Acad. Sci. USA,118,
e2020486118.
Zhu, X., Xu, X. & Wilson, I. A. (2008). Acta Cryst. D64, 843–850.
research papers
Acta Cryst. (2023). D79, 224–233 Matthew J. Rodrigues et al. Lattice-translocation defects in serial crystallography 233
Available via license: CC BY 4.0
Content may be subject to copyright.
