dynamical structural science

Acta Crystallographica Section A

Foundations of
Five-dimensional crystallography
Crystallography
ISSN 0108-7673 Marius Schmidt,a* Tim Graber,b Robert Henningb and Vukica Srajerb

a
Received 31 August 2009 Physics Department, University of Wisconsin-Milwaukee, Milwaukee, WI, USA, and bBioCARS,
Accepted 15 December 2009 Center for Advanced Radiation Sources, The University of Chicago, Chicago, IL, USA.
Correspondence e-mail: m-schmidt@uwm.edu

A method for determining a comprehensive chemical kinetic mechanism

in macromolecular reactions is presented. The method is based on five-
dimensional crystallography, where, in addition to space and time, temperature
is also taken into consideration and an analysis based on singular value
decomposition is applied. First results of such a time-resolved crystallographic
study are presented. Temperature-dependent time-resolved X-ray diffraction
measurements were conducted on the newly upgraded BioCARS 14-ID-B
beamline at the Advanced Photon Source and aimed at elucidating a
comprehensive kinetic mechanism of the photoactive yellow protein photocycle.
Extensive time series of crystallographic data were collected at two
temperatures, 293 K and 303 K. Relaxation times of the reaction extracted
from these time series exhibit measurable differences for the two temperatures,
hence demonstrating that five-dimensional crystallography is feasible.

1. Introduction crystallographic data (Schmidt et al., 2003). Application of the

SVD-based global analysis made possible for the first time a
1.1. Time-resolved crystallography theoretically sound analysis of the time-resolved X-ray data
Time-resolved macromolecular crystallography is a unique (Schmidt et al., 2004; Rajagopal et al., 2004, 2005; Ihee et al.,
method to unify chemical kinetics with X-ray structure 2005). Although the structures of the intermediates can be
determination (Moffat, 1998, 2001). Experiments are of the determined objectively, a candidate kinetic mechanism is not
pump–probe type using ultra-short laser pulses as the pump to unique. Multiple candidate mechanisms selected from a
initiate a reaction in a protein crystal followed by a short and general mechanism (see below) can explain the data equally
brilliant X-ray pulse of  100 ps duration to probe the struc- well. For a general understanding of protein function it is
ture at various time delays t. This way, nanosecond and sub- imperative to determine whether intermediates are on or off
nanosecond time resolution can be achieved to determine the the catalytic pathway. Hence, the number of degenerate
kinetics of reactions in biological macromolecules as well as candidates must be reduced to only one unique mechanism.
the structures of short-lived intermediates from the beginning As a corollary, the general kinetic mechanism has to be
of the reaction to the very end. Time-resolved crystallography covered comprehensively.
has been used to investigate reactions in heme-containing
proteins such as myoglobin (Srajer et al., 1996, 2001; Schmidt
et al., 2005a; Bourgeois et al., 2003, 2006; Schotte et al., 2003, 1.2. Chemical kinetics in time-resolved crystallography
2004), FixL (Key et al., 2007) and hemoglobin (Knapp et al., Chemical kinetics (Steinfeld et al., 1989) describes a reac-
2006). In addition, the photocycle of the photoactive yellow tion in terms of schemes similar to the ones shown in Fig. 1,
protein was investigated extensively (Perman et al., 1998; Ren where a mechanism is depicted that employs only first-order
et al., 2001; Schmidt et al., 2004; Raja-
gopal et al., 2005; Ihee et al., 2005) to
determine the structures of the inter-
mediates of its photocycle that form and
decay conjointly with the chemical
kinetic mechanism that connects them.
A major advance in extracting struc-
tures of intermediates and the kinetic
mechanism from the time-dependent
crystallographic data involved the Figure 1
Kinetic mechanisms. General: general mechanism with three intermediates I1, I2 and I3 , plus the
application of singular value decom- reference (dark) state. The reaction is started by a laser pulse. DE: dead-end candidate, SP: semi-
position (SVD) to the time-resolved parallel candidate.

198 doi:10.1107/S0108767309054166 Acta Cryst. (2010). A66, 198–206

 dynamical structural science
reactions. Each step is characterized by its respective rate that the time-dependent concentrations of the intermediates
coefficient. Rate coefficients cannot be measured directly. match (almost) exactly at a certain temperature, here 300 K.
Instead, macroscopic rates (or their negative inverse, the At this temperature these two mechanisms are degenerate,
relaxation times) are directly observable in kinetic measure- because they give the same relaxation times. Now assume that
ments. They are functions of all rate coefficients (Matsen & in the DE mechanism intermediate I3 is very stable. That is,
Franklin, 1950; Fleck, 1971). It is important to make a clear-cut the barrier of activation to revert to intermediate I2 is quite
distinction between experimentally observable (macroscopic) high. Then, the rate coefficient k3 becomes particularly small
rates and the hidden (microscopic) rate coefficients of the at lower temperatures. On the other hand, in the semi-parallel
mechanism (Rajcu et al., 2009). It is the purpose and the goal mechanism SP, intermediate 3 may branch directly to the final
of any kinetic experiment to ultimately determine the under- (dark) state, crossing a smaller energy barrier. If the
lying kinetic mechanism with all of its rate coefficients. Almost temperature is lowered to, say, 273 K, these two mechanisms
any kinetic model is mathematically underdetermined if become separable (Fig. 2b), although they were indis-
relaxation times from a time series at only one temperature tinguishable at 300 K. Consequently, the temperature adds
are available, because a large number of rate coefficients must observables that can be used to determine the unknowns (the
be determined from a smaller number of measured relaxation rate coefficients) of the mechanisms.
times. Since a time-resolved crystallographic experiment is a The rate coefficients at two temperatures are related. In the
true kinetic experiment, it depends on the kinetic mechanism simplest case this relationship is given by the Arrhenius
in fundamentally the same fashion. equation (see, for example, Cornish-Bowden, 1999), where kB
As an example we simulated relaxation times similar to is the Boltzmann constant and T is the temperature,
those of a recent time-resolved crystallographic experiment

(Schmidt et al., 2004). Three relaxation times are observed. ki ðTÞ ¼ Ai ðTÞ exp Gi# =kB T



The number of relaxation times is equal to (or smaller than) ¼ Ai ðTÞ exp Si# =kB exp Hi# =kB T : ð1Þ
the number of intermediates. A general chemical kinetic
mechanism of a cyclic reaction with three intermediates plus For each rate coefficient ki we would need to determine three
the final (dark) state is shown in Fig. 1. The general parameters: the enthalpy (H #) and entropy (S #) differ-
mechanism contains 12 rate coefficients (that from I1 directly ences to the transition state, as well as a pre-factor A(T). In
to D is not shown). It is immediately clear that the three the simplest case, A(T) is proportional to the temperature (see
relaxation times are not sufficient to uniquely determine all also Cornish-Bowden, 1999). Hence, the fit would include a
rate coefficients of the general mechanism. Therefore, two linear term, a constant and an exponential term. If we were
likely candidate mechanisms, each involving four rate coeffi- able to determine experimentally only the relaxation times, we
cients, were picked in a rather subjective way (DE: dead-end would need 12 different temperatures to account for the 36
candidate, SP: semi-parallel candidate). Fig. 2(a) shows that free parameters in the general mechanism of this example. If
the rate coefficients for the two likely mechanisms can be such we were only interested in the G # values in equation (1), we
would need eight temperatures (24 free parameters). The
critical question is whether measurement over a limited
physiological temperature range would allow us to determine
all the unknowns from measured relaxation times.
However, we are in fact considerably better off, because we
can exploit the so-called absolute scale present in crystal-
lography. Measured structure factor amplitudes can always be
placed on an absolute scale by scaling them to the calculated
structure factor amplitudes from a precise known structural
model. Electron density, when on an absolute scale, directly
relates to occupancy; that is, to fractional concentration. For
example, if a side chain has alternate conformations, the
electron density is directly related to the occupancy of that
conformation. The same is true for difference electron
densities on the absolute scale. In contrast, in optical
absorption spectroscopy, for example, absorption usually does
not directly relate to concentration as there is a linear factor,
the absorption coefficient, which is unknown a priori for
Figure 2 various intermediates. In crystallography this linear factor is
Concentration profiles and approximate relaxation times for the three simply absent. Consider an experiment in which a CO mole-
intermediates in the semi-parallel and dead-end mechanisms. (a) 300 K, cule is photo-dissociated from the iron of a heme protein. By
semi-parallel mechanism: intermediate I1 (dashed line), intermediate I2
integrating the negative difference electron density at the CO
(dashed dotted line) and intermediate I3 (dotted line). The thin lines
represent the corresponding concentrations for the same intermediates in binding site, seven electrons are obtained. Since CO has 14
the dead-end mechanism. (b) 273 K, lines the same as in (a). electrons, 50% of the CO has been photo-dissociated. If the

Acta Cryst. (2010). A66, 198–206 Marius Schmidt et al.  Five-dimensional crystallography 199
dynamical structural science
concentration of protein is 50 mM in the crystal, the remaining enhances greatly the ability to determine a comprehensive
bound CO concentration is 25 mM (50% of the 50 mM). Note chemical kinetic mechanism from time-resolved crystal-
that this calculation is only valid on the absolute scale. lographic data.
Now, we need to connect the structural result, the electron With the relaxation times  k and the amplitudes [Pjk] we
density, to the kinetics. Equation (2) is the result of integrating gain at least six observables per temperature to determine the
the coupled differential equations that describe the 36 free parameters (plus one initial condition, the extent of
mechanism involving three intermediates assuming exponen- reaction initiation) in the general mechanism. That would
tial kinetics. Equation (2) constitutes the mathematical base of reduce the number of required temperatures to only six or
simple chemical kinetics described in textbooks such as seven. It is extremely exciting to see that we actually have
Steinfeld et al. (1989). The time-dependent concentrations [Ij] a realistic chance to determine a comprehensive general
can be calculated from the rate coefficients, because both the mechanism with measurements at a relatively small number
relaxation times  k, which are the eigenvalues of the so-called of temperatures. However, the collection of an extensive
coefficient matrix, and the pre-factors [Pjk], which are the spatially and temporally complete time series of Laue data is
elements of the eigenvectors of the coefficient matrix, are a tedious time-consuming process let alone the collection of
functions of the rate coefficients; the [Pjk] contain further time series at multiple temperatures. It is a major goal of this
specific initial conditions, for example [I1] = 1.0, [I2] = [I3] = 0 paper to show that this can now be achieved.
at t = 0;
      1.3. Data acquisition
I1 ðtÞ ¼ P11 expðt=1 Þ þ P12 expðt=2 Þ
  BioCARS 14-ID-B is the only beamline in the USA dedi-
þ P13 expðt=3 Þ; cated to time-resolved macromolecular crystallography. It has
     
I2 ðtÞ ¼ P21 expðt=1 Þ þ P22 expðt=2 Þ been recently upgraded and features now the capability for
  ð2Þ
þ P23 expðt=3 Þ; 100 ps time resolution and an entirely new design to focus and
      deliver the X-rays to the sample. A single exposure with one
I3 ðtÞ ¼ P31 expðt=1 Þ þ P32 expðt=2 Þ
  153 ps X-ray pulse in the hybrid mode of the APS storage ring
þ P33 expðt=3 Þ: can produce an acceptable Laue diffraction pattern. This
dramatically speeds up data collection making the rapid
Now assume that we are able to observe the [Pjk] directly. That
collection of the entire time-series of Laue data feasible. In
would give us, in the best scenario, another nine independent
addition, it opens up new opportunities to investigate non-
observables to determine the underlying rate coefficients.
cyclic reactions in biomolecules. For the data presented here
However, some of the [Pjk] could be very small, close to zero,
we used a nanosecond laser for reaction initiation; however, a
because each intermediate might have only one transient. In
picosecond laser system implemented at BioCARS by Philip
our example in Fig. 2, [P12] and [P13] are in fact zero (and some
Anfinrud and colleagues (NIDDK/NIH) is also available. It
others are also negligible in magnitude and not independent).
delivers 30 ps pulses to the sample and makes sub-ns time-
This reduces the number of useful [Pjk]. In any case, we would
resolution possible. New timing hardware and user-friendly
obtain at least three additional observables if the [Pjk] could
software control modules enable fast automated data collec-
be observed. And indeed, the [Pjk] can be observed directly in
tion. To minimize radiation damage, larger crystals can be
crystallography. How can that be done? This approach has
translated during data collection so that some fresh crystal
been the foundation of the ‘posterior analysis’ developed by
volume is exposed each time to the intense X-ray pulses.
Schmidt et al. (2003, 2004) for post-SVD analysis of time-
Here we show an optimized way to rapidly collect
resolved data. Details are given also by Schmidt et al. (2005b)
comprehensive spatially and temporally complete data sets
and Schmidt (2008). In short, once the preliminary structures
from which the relaxation times of the kinetics can be
of the intermediates are determined, we can calculate time-
extracted. We demonstrate how the relaxation times vary by
independent difference electron densities for each inter-
changing the temperature by only 10 K, from 293 K to 303 K.
mediate by subtracting the structure factors of the known
These results will pave the way to a more comprehensive
initial (reference) D state from those of the intermediates. We
coverage of the available temperature range. Crystallography
assume a chemical kinetic mechanism with initial values for
becomes five-dimensional, involving space, time and
the rate constants and generate with these calculated time-
temperature, and with this the determination of a unique
dependent difference electron densities calc t ðtÞ that can be mechanism becomes feasible.
compared directly with the observed difference densities
obs
t ðtÞ. In a large fitting routine the rate coefficients and the
initial condition, namely the concentration of activated 2. Material and methods
molecules at the beginning of the reaction, are refined by
minimizing globally the difference between the observed and 2.1. Data collection
calculated difference maps. Both the amplitudes [Pjk] and the Two comprehensive time series consisting of 28 Laue data
measured relaxation times  j are used as observables this way sets (27 time-dependent data sets plus the reference dark data
to refine the numerical values of the rate coefficients ki . The set) were collected on crystals of the photoactive yellow
fact that concentration is directly related to electron density protein (PYP) at the BioCARS 14-ID beamline at the

200 Marius Schmidt et al.  Five-dimensional crystallography Acta Cryst. (2010). A66, 198–206
 dynamical structural science
Advanced Photon Source, Argonne National Laboratory, volume of the reciprocal space with an angular spacing of 3 .
Argonne IL, USA. One time series was collected at 293 K, the At 303 K we used only 12 or 11 different settings with an
other at 303 K. For each time series only one long needle- angular spacing of about 6 . All diffraction images were
shaped PYP crystal was used, each of dimensions 80  80  collected on a Mar165 CCD detector.
1000 mm. Crystals were mounted in glass capillaries of 1 mm
diameter. X-ray beam size was 100 (h)  75 (v) mm (Fig. 3). A 2.2. Data reduction and difference maps
reaction in the crystal was initiated by an intense laser pulse of Laue data were indexed and integrated using Precognition
6 ns full width half-maximum (FWHM) at 485 nm (Vibrant and scaled using Epinorm (RenzResearch, http://www.renz
Nd:YAG pumped OPO laser, Opotek), focused to a spot with research.com/). Table 1 shows the statistics of the compre-
a diameter of 110 mm. The total pulse energy was about 50 mJ hensive time series. The reference (dark) data were brought to
which corresponds to an energy density of  5.3 mJ mm2 at the absolute scale by scaling them to structure factor ampli-
the crystal. Temperature was controlled by a cryojet sample tudes calculated from the PYP model 2phy from the Protein
cooler (Oxford Instruments). Data were collected with time as Data Bank (Berman et al., 2000). The time-dependent Laue
the fast variable (Rajagopal et al., 2004). That means that at amplitudes were then scaled to their respective reference data.
each crystal orientation data were collected at a series of time This way, both the reference data and the time-dependent data
delays, then the crystal orientation was changed to a new were on the absolute scale. Weighted difference structure
value. The time delays were equidistant in logarithmic time, factor amplitudes were calculated for each time point using
from 4 ns to 256 ms. For each time point, a Laue diffraction w = 1/[1 þ ð 2 =h 2 iÞ + ðF 2 =hF 2 iÞ] as weighting factors (see
pattern was obtained by using seven (at 303 K) or eight (at Ren et al., 2001). Here,  2 corresponds to an individual
293 K) repeated 153 ps X-ray exposures, each preceded by a squared difference amplitude uncertainty relative to the mean
laser pulse. Based on previous experience (Schmidt et al., square uncertainty found in the entire data set, h 2i. F 2 is
2004; Ihee et al., 2005), the waiting time between the exposures the squared difference amplitude and is compared with the
was 2 s at 303 K and 4 s at 293 K. After all time points were mean square difference amplitude found in the entire data set
completed at given crystal orientation, the crystal was rotated hF 2i. This ensures that observations with large uncertainties
to collect data from another part of reciprocal space. The new and those with excessively high difference amplitudes are
orientation was selected in such a way that, if data collection down-weighted for map calculation. Absolute scale is main-
would terminate prematurely, reciprocal space would have tained by dividing (normalizing) the data by the average
been covered in an approximately uniform manner. In order weighting factor. Difference maps were calculated with the
to reduce radiation damage, the crystal was also translated weighted difference structure factor amplitudes using phases
approximately 30 mm along its axis for each new crystal calculated from the reference model mentioned above.
orientation. With a new setting all time points were again
collected holding the crystal stationary. This was repeated 2.3. Kinetic analysis
until reciprocal space was covered by crystal settings a few
degrees apart. The space group of PYP crystals is P63 . At The time series of electron density maps were analyzed by
293 K, 24 different orientations were used to cover the unique singular value decomposition using the program SVD4TX
(Schmidt et al., 2003; Zhao & Schmidt, 2009). The SVD
program masks out a given region in real space and performs
an SVD on the region within the mask. The volumes occupied
by the following residues were masked out: two different 4-
hydroxycinnamoyl (HC4) chromophore structures found in
the reference model and that of the pB2 photocycle inter-
mediate (Schmidt et al., 2004), Cys 69, Arg 52, Tyr42 and
Glu46. These residues basically cover the chromophore
binding pocket that contains the strongest signal ideal for this
feasibility study. For a more comprehensive structural analysis,
other parts of the PYP can also be masked out (Ihee et al.,
2005). The initial mask was evolved by only allowing grid
points in the mask that contained difference electron density
features larger than +3.0 or smaller than 3.0 that occur at
least in one time point.

3. Results
Figure 3
Typical laser and X-ray beam illumination geometry at the crystal. 3.1. Data collection efficiency and crystallographic quality
Needle-like PYP crystals of length > 1 mm and diameter 80 mm were
Owing to the high polychromatic X-ray flux at the sample at
used. Only the region that was not compromised by the small satellite
crystals was used for measurements. Arrows shown in cyan indicate the 14-ID beamline, accumulation of only eight 153 ps X-ray
translation and reorientation of the crystal during data collection. pulses in the hybrid mode of the APS storage ring was suffi-

Acta Cryst. (2010). A66, 198–206 Marius Schmidt et al.  Five-dimensional crystallography 201
dynamical structural science
Table 1
Data collection statistics for data at (a) 293 K and (b) 303 K.
(a) Every second time point is listed. 24 crystal orientations were used to cover the unique volume of reciprocal space. Rmerge is given for the entire resolution range
to 1.6 Å. Rscale is the result from scaling time-dependent and dark (reference) data sets to a resolution of 1.76 Å. Difference maps were calculated to 1.8 Å.
t (light)
Dark 4 ns 16 ns 64 ns 256 ns 1 ms 4 ms 16 ms 64 ms 256 ms 1 ms 4 ms 16 ms 64 ms 256 ms 4 s†

Completeness, 1–1.76 Å (%) 89.2 89.2 88.9 88.3 88.3 87.9 87.9 87.6 87.3 87.6 87.3 87.4 87.1 86.6 86.4
Completeness, 1–3.5 Å (%) 91.2 91.4 91.3 91.3 92.0 91.4 91.4 91.5 91.2 91.2 91.6 91.1 91.2 91.6 91.6
Last shell 78.4 76.9 76.6 73.8 74.6 74.0 74.6 73.4 73.9 71.3 74.9 72.3 72.2 69.7 69.6
I/I 18.0 17.2 17.0 11.5 15.0 15.0 14.2 14.4 16.0 17.1 14.4 16.4 13.2 13.0 14.1
Rmerge on | F | (%) 3.6 4.2 4.4 6.4 4.5 4.7 4.8 4.9 4.8 4.4 4.7 4.8 4.8 4.9 5.0
Rmerge on | F |2 (%) 6.4 7.4 7.7 11.1 7.8 8.8 8.4 8.5 8.5 7.6 8.3 8.5 8.4 8.4 8.8
Rscale (%)‡ 8.8 9.1 11.4 10.6 10.6 11.0 11.6 10.9 11.3 12.2 11.9 11.4 11.8 13.4 12.1
hF i (e) 12.1 13.7 20.4 15.1 16.8 17.0 17.6 17.7 16.9 17.6 17.7 18.5 11.8 20.1 25.5
( F) 3.7 4.1 4.5 4.1 4.3 4.4 4.5 4.5 4.5 4.5 4.4 4.7 4.8 4.8 4.1
max/min / § 7.6 7.0 5.0 7.0 6.4 6.7 6.0 6.0 4.7 4.5 5.1 5.0 5.1 4.9 4.6
+6.0 +5.0 +5.0 +6.2 +6.1 +5.6 +5.4 +4.8 +4.7 +4.5 +5.1 +4.5 +4.8 +4.5 +4.5

(b) Every second time point is listed. Up to 32 ms 12 crystal orientations and for larger time delays 11 crystal settings were used to cover the unique volume of the
reciprocal space. Rscale is the result from scaling time-dependent and dark (reference) data sets to a resolution of 1.76 Å. Difference maps were calculated to 1.8 Å.
t (light)
Dark 4 ns 16 ns 64 ns 256 ns 1 ms 4 ms 16 ms 64 ms 256 ms 1 ms 4 ms 16 ms 64 ms 256 ms 2 s}†

Completeness, 1–1.76 Å (%) 70.7 70.4 69.8 69.9 69.5 69.6 68.8 68.9 65.7 65.7 65.7 65.3 65.1 64.9 65.0
Completeness, 1–3.5 Å (%) 74.3 74.0 74.2 74.2 74.3 73.9 72.2 73.7 73.5 71.9 71.9 71.8 71.4 71.6 72.0
Last shell 52.7 51.2 51.2 50.8 49.9 50.6 48.4 48.3 48.5 45.3 45.6 43.9 44.0 44.3 43.9
Rmerge on | F | (%) 3.2 3.5 3.4 3.6 3.5 3.9 3.8 3.8 3.8 3.8 3.8 3.9 3.8 3.8 3.9
Rmerge on | F |2 (%) 5.7 6.1 6.0 6.4 6.2 6.9 6.7 6.6 6.6 6.8 6.8 6.9 6.6 6.7 6.9
Rscale (%)††‡ 9.3 9.6 11.0 10.4 11.3 11.3 10.9 11.9 11.3 12.9 12.0 11.4 12.4 13.3 12.2
hFi (e) 11.4 12.9 15.0 14.7 15.9 16.7 15.7 16.8 17.1 18.0 18.7 18.0 18.4 18.9 25.2
h Fi 4.5 4.6 4.8 4.8 4.9 4.9 5.1 5.1 5.2 5.3 5.4 5.5 5.5 5.7 4.1
max/min / ‡‡§ 5.5 5.9 6.4 6.1 5.1 6.5 6.2 4.5 4.3 4.5 4.5 4.7 4.5 4.8 4.3
+4.4 +6.2 +5.8 +8.1 +6.1 +5.1 +5.7 +4.5 +4.3 +4.5 +4.7 +4.4 +4.5 +4.6 +4.5

† For the 4 s time point the reference (dark) data at 303 K were subtracted from the reference (dark) data collected at 293 K. ‡ Rscale given here is for the last shell: 1.87–
1.76 Å. § Highest positive and lowest negative difference electron density feature (), in units of  value of the difference map. † } For the 2 s time point the reference (dark) data
at 293 K were subtracted from the reference (dark) data collected at 303 K. ‡ †† Rscale given here is for the last shell: 1.87–1.76 Å § ‡‡ Highest positive and lowest negative
difference electron density feature (), in units of  value of the difference map.

cient to record very well exposed diffraction images. The total collected using only one crystal. The scale factors Rscale
elapsed time for collecting the comprehensive time series at between the reference data and the time-dependent data for
293 K was about 6.5 h even with 4 s waiting time between the the last resolution shell increase slightly with each time point,
X-ray exposures. With only 2 s between the exposures an indicative of some radiation damage (Table 1). However, the
entire comprehensive time series can be collected in about
3.5 h. Using two fast Linux computers, data reduction of 56
PYP Laue data sets (28 data sets each at two different
temperatures) can be completed in a few days using the
Precognition/Epinorm software package. For crystals of other
space groups that require covering a larger part of reciprocal-
space data, data reduction time increases accordingly. Using
24 crystal orientations, PYP Laue data completeness is quite
good to 1.8 Å and Rmerge is around 7% throughout the entire
resolution range, demonstrating the exquisite quality of crys-
tallographic data obtained at 14-ID-B on relatively small
diffracting volumes (80  75  100 mm) (Table 1a). With 12
different crystal orientations, completeness is somewhat
reduced (Table 1b). However, the signal in the difference Figure 4
Difference electron density maps at a time delay t of 256 ns at (a) 293 K
maps resulting from data collected with 12 orientations is and (b) 303 K. Contour levels: 3/4 (red/white); +3/+4/+5 (blue/
comparable with that in the maps resulting from data with 24 dark-blue/cyan). The atomic model displayed is that of the reference
crystal orientations as judged by the largest peak intensities of (dark) state (PDB 2phy structure) with the HC4-chromophore in its trans
configuration. The electron density feature  shows the flip of the
the difference features (Table 1 and Fig. 4). The data for the
carbonyl oxygen of the HC4 tail and  shows the trans to cis isomerization
respective time series are remarkably homogeneous in terms about the double bond, while  indicates the displacement of the Cys69
of completeness and Rmerge . This is because each series was sulfur to which the chromophore is bound.

202 Marius Schmidt et al.  Five-dimensional crystallography Acta Cryst. (2010). A66, 198–206
 dynamical structural science
scale factors even at the highest dose do not substantially
exceed those from scaling two reference data sets collected on
different crystals [listed in the last columns of Tables 1(a) and
1(b)]. This homogeneity of the data throughout the time series
greatly contributes to the quality of the SVD analysis.

3.2. Radiation damage

The absorbed dose was calculated by estimating that
roughly 3  1010 photons with an average energy of 12 keV
are in a 153 ps X-ray pulse. Eight X-ray pulses were used to
acquire a single CCD frame recording of a Laue diffraction
pattern, containing a total of 2.4  1011 photons. With an
X-ray beam size of 75 mm (v)  100 mm (h), and crystal
diameter of 80 mm, a crystal volume of 6  107 cm3 (beam
size times crystal diameter) is exposed to these photons for
each diffraction pattern (Fig. 3). As the translation of the
crystal along the long crystal dimension (Fig. 3) before each
new crystal orientation is 30 mm and the beam size is 100 mm,
only one-third of fresh crystal volume is exposed to the X-rays
each time the crystal is translated after the collection of 28
time points. This means that each diffracting crystal volume is
actually exposed 3.3 times per time point. With a typical
protein density of 1.35 g cm3 (White et al., 2007) the exposed
volume contains about 810 ng of material. 2.4  1011 photons
at 12 keV have a total energy of 0.46 mJ. Owen et al. (2006)
give an approximate absorption coefficient of 1 mm1 for a
protein crystal. For the 80 mm PYP crystal, this means that
7.7% of the X-ray photons are absorbed and contribute to the
absorbed dose given in Gray (Gy = J kg1). Each volume Figure 5
along the needle-shaped crystal absorbs a dose of about 0.14  (a) Estimation of absorbed dose (full circles) and completeness (full
106 Gy per time point (3.3  0.077  0.46 mJ/810 ng). In squares) as a function of time. Since the time points were collected in
addition, the absorbed dose was also determined using the consecutive order, dose becomes a function of time. Black dashed lines:
guides to the eye. Red dashed lines: lower and higher dose rate limits. (b)
program Raddose (Murray et al., 2004; Paithankar et al., 2009). Absorbed dose as calculated by program Raddose (full circles). I/ I (full
The absorption coefficient determined by Raddose was squares) as a function of time is also shown. Red dashed lines: lower and
0.22 mm1. The absorbed photons generated a temperature higher dose rate limits.
jump of < 0.1 K for each X-ray pulse. The accumulated dose
determined from our crude estimation and that from Raddose The pre-factors Ai,1, Ai,2 and Ai,3 are (linear) fit parameters for
are compared for all time points in Fig. 5. each individual rsv, and the relaxation times  1 ,  2 and  3 are
common fit parameters for both rsvs. The result of the fit is
3.3. Results from the SVD analysis shown in Fig. 7 by the solid lines through the data points of the
first two rsvs. Table 2 shows the resulting values for relaxation
The singular values and the right singular vectors (rsv)
times. All relaxation times are shorter at 303 K, indicating a
obtained from the SVD analysis of the time series of differ-
faster reaction as expected.
ence maps are shown in Figs. 6 and 7, respectively. Two
obviously significant singular values are present (Fig. 6) at
both temperatures. The rsv contain the kinetic information
about the reaction. Those rsvs that belong to the two largest 4. Discussion
singular values are shown by circles and squares. Other, not so 4.1. Radiation damage
significant, rsvs are also shown in Fig. 7 by triangles and
The results presented here address the feasibility of five-
diamonds. These other rsvs might also contain some signal.
dimensional crystallographic experiments and allow us to
However, since all relevant relaxation times are common to all
assess the suitability of such data for SVD-based global
significant rsvs, only the first two rsvi (i = 1, 2) were analyzed
analysis of a kinetic mechanism. One of the major concerns is
here. These rsvs were globally fit by a sum of three exponential
the radiation damage of the crystals owing to the total X-ray
functions with relaxation times  1,  2 and  3 ,
dose that is necessary for collecting a comprehensive time
rsvfit
i ¼ Ai;1 expðt=1 Þ þ Ai;2 expðt=2 Þ þ Ai;3 expðt=3 Þ:
series of data (Figs. 5a and 5b). Fig. 5 compares our estimation
with the more accurate calculation by the program Raddose
ð3Þ
(Murray et al., 2004; Paithankar et al., 2009). Figs. 5(a) and 5(b)

Acta Cryst. (2010). A66, 198–206 Marius Schmidt et al.  Five-dimensional crystallography 203
dynamical structural science
Table 2
Relaxation times.
Numbers in parentheses are for data at 293 K where only 12 frames were
merged to match the number of frames in the 303 K data set.
T 1 2 3

293 K 17.7 ns (12.0 ns) 254 ms (266 ms) 2.5 s (2.5 s)

303 K 10.3 ns 147 ms 1.4 s

Figure 6
The first six (out of 27) singular values, SV, for the data at 293 K (blue)
and 303 K (red). Two singular values are clearly significant.

also plot the completeness as well as I/ I in the last resolution

shell as a function of time. Since the time points are collected
in consecutive order from one crystal setting, the crystals are
exposed to a dose increasing by 0.14  106 Gy per time point
(see above) as estimated by our crude approximation. The
dose calculated by Raddose was substantially smaller, only
0.037  106 Gy per time point due to an about fourfold smaller Figure 7
absorption coefficient of 0.22 mm1. At room temperature, The right singular vectors (rsv) plotted as a function of time. Full circles:
crystal damage has been observed at 0.38  106 Gy (South- first significant singular vector; full squares: second significant singular
vector. The third and fourth rsvs are shown by triangles and diamonds,
worth-Davies et al., 2007) suggesting that, between two and ten respectively. Solid black lines through the first two significant rsvs are the
time points (dependent on the dose estimation), the crystals result of a global fit using equation (3). (a) 293 K; insert: data set includes
are damaged (lower dose rate limit). Interestingly, at room only 12 diffraction frames to match the number of frames in the 303 K
data set. (b) 303 K.
temperature the damage threshold D1/2 is largely dependent
on, and increases with, the rate the dose is applied to the
crystals (Southworth-Davies et al., 2007). Even more, this Figs. 5(a) and 5(b)], indicative of only a minimal effect of
effect is roughly independent on the elapsed time between the radiation damage. At cryogenic temperatures the radiation
exposures. In our experiments 5300 Gy is absorbed in one damage threshold is given by the Henderson limit which is
single 153 ps X-ray pulse, which gives a dose rate of 3.5  around 3  107 Gy (Henderson, 1990; Owen et al., 2006), and
1013 Gy s1. Systematic investigations on radiation damage at of the order of 200 to 800 time points could be collected there.
these high dose rates do not exist so far. D1/2 values for dose We believe that, owing to the extremely high dose rate, the
rates of only 10 Gy s1 are around 1.8  106 Gy (higher dose dose threshold that limits the number of collectable Laue data
rate limit in Fig. 5), suggesting a potential crystal damage after sets is significantly higher than the high dose rate limit
12 time points (Fig. 5a). The dose calculated by Raddose reported above and lower than the Henderson limit. At least
stayed well below the higher dose limit, so the damage seems with PYP, radiation damage seems to be small even after 28
to remain small (Fig. 5b). The completeness in the last reso- time points. In addition, as we compare results for two crystals
lution shell of our PYP crystals as well as I/ I in the last used in an identical fashion at two temperatures, any radiation
resolution shell changes only slightly [Tables 1(a) and 1(b); damage should affect both crystals equally.

204 Marius Schmidt et al.  Five-dimensional crystallography Acta Cryst. (2010). A66, 198–206
 dynamical structural science
4.2. Limits 4.3. Effects of the laser that initiates the reaction
Another issue related to the success of five-dimensional Empirical evidence shows that the excessive laser energy
crystallography is the ability to measure reliably relatively density at the sample induces increased transient and even-
small changes in relaxation rates as a function of temperature. tually permanent increase in crystal mosaicity as observed by
In the measurements presented here, the PYP HC4 chromo- elongation of diffraction spots, as well as gradual loss of
phore clearly photo-isomerized from trans to cis as evident diffraction at higher resolution. The pulse energy density we
from the difference map shown in Fig. 4. The positive differ- used is well below those density levels. In addition we used
ence electron density features  and  are characteristic of 4 ns laser pulses with a substantially lower peak power
such isomerization and represent one of the early inter- compared with the picosecond laser pulses that have also been
mediates in the PYP photocycle (Ihee et al., 2005) that used in the past for time-resolved experiments (Schotte et al.,
subsequently relaxes towards later intermediates (Schmidt et 2003, 2004). Therefore, we believe that we avoided crystal
al., 2004; Ihee et al., 2005). The relaxation times of a reaction in damage by the laser pulses as much as possible.
general shift to shorter times if the temperature is increased. A laser-induced temperature jump is also a potential chal-
It is empirically observed that the reaction velocity doubles lenge for five-dimensional crystallography. For the data
when the temperature is increased by 10 K (Q10 rule, see presented here, a laser beam with a 110 mm diameter delivered
Cornish-Bowden, 1999). Here, we find that the second about 90% of the total pulse energy or  45 mJ into the crystal
relaxation time  2 changes by a factor of about 1.7, being volume of 110  80  80 mm. Using the typical crystal density
253 ms at 293 K and 147 ms at 303 K in good agreement with of 1.35 g cm3 given above, this volume contains 9.5 
the Q10 rule. The change in the first relaxation time  1 is more 1010 kg. The heat capacity of protein crystals is of the order
difficult to determine since the data are noisy in this time of 3 to 8 kJ kg1 K1 (Miyazaki et al., 1993, 2000). Assuming
regime. However, the general trend that the relaxation times that all of the laser pulse energy delivered to the crystal is
become shorter with increasing temperature can be observed absorbed within the crystal volume above (which is consistent
also at the fast times ( 1 = 10 ns at 303 K and 17 ns at 293 K). with the optical density of the crystal at 485 nm), the
The same trend is also observed if only 12 diffraction frames temperature jump in the crystal generated by the laser pulse
were merged at 293 K, to match the completeness of data ranges from 6 to 16 K. Thermal diffusion times in the crystal
collected at 303 K (Table 2, data in parentheses and Fig. 7, are of the order of 50 ms (Moffat et al., 1992). We can expect
insert). This shows that the reduced reciprocal-space that on the fast time scale, up to a few tens of milliseconds, the
completeness at 303 K had no severe influence on the same temperature offset owing to laser-induced heating will be
relaxation times. This is the first time that such a shift of present at all temperatures. However, special attention has
relaxation times has been systematically observed for time- to be given to the slower time scales. The time-dependent
resolved crystallographic data. This was achieved by changing temperature profile in the crystal should be initially modelled
the temperature by only 10 K. The temperature range in which based on crystal and laser beam geometry as well as taking
these experiments will be feasible spans from the freezing laser power, heat capacity and heat conductance in protein
point of the liquor in the crystals which should be substantially crystals into account (Noelting, 1998). This profile can be
below 273 K to the temperature where crystal stability is likely refined using the five-dimensional crystallographic data
compromised by protein denaturation. PYP melts in solution at later stages of the analysis.
at around 358 K (Meyer et al., 2003). In crystals the melting
temperature could be even larger. Hence, it should be possible
to vary the temperature between 268 K and around 343 K.
Seven to nine temperature settings separated by 10 K intervals 4.4. Mechanisms
will be achievable this way. In recent years, a number of papers have reported effects on
A comprehensive mechanism containing N states (N  1 kinetic variables owing to molecular crowding that occurs
intermediates plus the ground state) can have N(N  1) rate under physiological conditions in the cell (see Zhou et al., 2008
coefficients. Schmidt et al. (2003) determined the limits of the for a recent review). Binding constants and reaction rates are
SVD-based analysis using three intermediates plus the ground altered compared with dilute solution. Protein crystals are
state and five orders of magnitude in time covered by about ultimately crowded. It is not clear whether crystals or dilute
three time points per logarithmic decade. Under those solutions best resemble the situation in vivo.
conditions, good results were obtained by the SVD analysis. Finally, it has been observed that the simple exponential
With nine orders of magnitude in time covered, as presented approach to reaction kinetics fails at cryogenic temperatures
here, at least five intermediates (N = 6) with 30 rate coeffi- where the time scales of conformational fluctuation are dras-
cients in the general mechanism could be accommodated. tically reduced or frozen, leading to an ensemble of molecules
Incidentally, this is also the number of intermediates observed with a distribution of rate coefficients (Austin et al., 1975).
by Ihee et al. (2005). In the spirit of the above discussion, Since time-resolved crystallographic experiments are
90 free variables would face about ten observables per performed at ambient temperatures, the protein molecules are
temperature point. For a comprehensive mechanism, nine flexible (Parak, 2003; Schmidt et al., 2009) and the ensemble
different temperature settings would be necessary. likely obeys classical chemical kinetics (Tetreau et al., 2004) as

Acta Cryst. (2010). A66, 198–206 Marius Schmidt et al.  Five-dimensional crystallography 205
dynamical structural science
long as the rate of conformational averaging is fast compared Noelting, B. (1998). J. Phys. Chem. 102, 7506–7509.
with the reaction rates. Owen, R. L., Rudino-Pinera, E. & Garman, E. F. (2006). Proc. Natl.
Acad. Sci USA, 103, 4912–4917.
We thank Keith Moffat for providing PYP crystals and Jane Paithankar, K. S., Owen, R. L. & Garman, E. F. (2009). J. Synchrotron
Kuk from his lab for preparing the crystals used for this study. Rad. 16, 152–162.
Use of the Advanced Photon Source was supported by the Parak, F. G. (2003). Curr. Opin. Struct. Biol. 13, 552–557.
Perman, B., Srajer, V., Ren, Z., Teng, T. Y., Pradervand, C., Usrby, T.,
US Department of Energy, Basic Energy Sciences, Office of Bourgeois, D., Schotte, F., Wulff, M., Kort, R., Hellingwerf, K. J. &
Science, under Contract No. DE-AC02-06CH11357. Use of Moffat, K. (1998). Science, 279, 1946–1950.
the BioCARS Sector 14 was supported by the National Rajagopal, S., Anderson, S., Srajer, V., Schmidt, M., Pahl, R. &
Institutes of Health, National Center for Research Resources, Moffat, K. (2005). Structure, 13, 55–63.
under grant number RR007707. MS is supported in part by the Rajagopal, S., Schmidt, M., Anderson, S., Ihee, H. & Moffat, K.
(2004). Acta Cryst. D60, 860–871.
college of L&S, UW-Milwaukee, and by the National Science Rajcu, V., Schmidt, M. & Stoneman, M. (2009). Biochemical
Foundation, award ID 0952643. The time-resolved set-up at Applications of Nonlinear Optical Spectroscopy, Series: Optical
Sector 14 was funded in part through collaboration with Philip Science and Engineering, Vol. 138, edited by V. Yakovlev, pp. 1–32.
Anfinrud (NIH/NIDDK). USA: CRC press.
Ren, Z., Perman, B., Srajer, V., Teng, T. Y., Pradervand, C., Bourgeois,
D., Schotte, F., Ursby, T., Kort, R., Wulff, M. & Moffat, K. (2001).
References Biochemistry, 40, 13788–13801.
Austin, R. H., Beeson, K. W., Eisenstein, L., Frauenfelder, H. & Schmidt, M. (2008). Ultrashort Laser Pulses in Medicine and Biology,
Gunsalus, I. C. (1975). Biochemistry, 14, 5355–5373. edited by W. Zinth, M. Braun and P. Gilch. pp. 201–241. New York:
Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Springer Verlag.
Weissig, H., Shindyalov, I. N. & Bourne, P. E. (2000). Nucleic Acids, Schmidt, M., Achterhold, K., Prusakov, V. & Parak, F. (2009). Eur.
28, 235–242. Biophys. J. 38, 687–700.
Bourgeois, D., Vallone, B., Arcovito, A., Sciara, G., Schotte, F., Schmidt, M., Nienhaus, K., Pahl, R., Krasselt, A., Nienhaus, U., Parak,
Anfinrud, P. A. & Brunori, M. (2006). Proc. Natl. Acad. Sci. USA, F. & Srajer, V. (2005a). Proc. Natl. Acad. Sci. USA, 102, 11704–
103, 4924–4929. 11709.
Bourgeois, D., Vallone, B., Schotte, F., Arcovito, A., Miele, A. E., Schmidt, M., Pahl, R., Anderson, S., Srajer, V., Ren, Z. & Moffat, K.
Sciara, G., Wulff, M., Anfinrud, P. & Brunori, M. (2003). Proc. Natl. (2004). Proc. Natl. Acad. Sci. USA, 101, 4799–4804.
Acad. Sci. USA, 100, 8704–8709. Schmidt, M., Pahl, R., Ihee, H. & Srajer, V. (2005b). Methods in
Cornish-Bowden, A. (1999). Fundamentals of Enzyme Kinetics. Molecular Biology, Vol. 305, Protein-Ligand interactions: Methods
London: Portland Press. and Applications, edited by G. U. Nienhaus, pp. 115–154. Totowa:
Fleck, G. M. (1971). Chemical Reaction Mechanism. New York: Holt, Humana Press.
Rinehart and Winston. Schmidt, M., Rajagopal, S., Ren, Z. & Moffat, K. (2003). Biophys. J.
Henderson, R. (1990). Proc. R. Soc. London Ser. B, 241, 6–8. 84, 2112–2129.
Ihee, H., Rajagopal, S., Srajer, V., Pahl, R., Schmidt, M., Schotte, F., Schotte, F., Lim, M., Jackson, T. A., Smirnov, A. V., Soman, J., Olson,
Anfinrud, P. A., Wulff, M. & Moffat, K. (2005). Proc. Natl. Acad. J. S., Phillips, G. N. Jr, Wulff, M. & Anfinrud, P. A. (2003). Science,
Sci. USA, 102, 7145–7150. 300, 1944–1947.
Key, J., Srajer, V., Pahl, R. & Moffat, K. (2007). Biochemistry, 46, Schotte, F., Soman, J., Olson, J. S., Wulff, M. & Anfinrud, P. A. (2004).
4706–4715. J. Struct. Biol. 147, 235–246.
Knapp, J. E., Pahl, R., Srajer, V. & Royer, W. E. Jr (2006). Proc. Natl. Southworth-Davies, R. J., Medina, M. A., Carmichael, I. & Garman,
Acad. Sci. USA, 103, 7649–7654. E. (2007). Structure, 15, 1531–1541.
Matsen, F. A. & Franklin, J. L. (1950). J. Am. Chem. Soc. 72, 3337– Srajer, V., Ren, Z., Teng, T. Y., Schmidt, M., Ursby, T., Bourgeois, D.,
3341. Pradervand, C., Schildkamp, W., Wulff, M. & Moffat, K. (2001).
Meyer, T. E., Devanathan, S., Woo, T., Getzoff, E. D., Tollin, G. & Biochemistry 40, 13802–13815.
Cusanovich, A. (2003). Biochemistry, 42, 3319–3325. Srajer, V., Teng, T. Y., Ursby, T., Pradervand, C., Ren, Z., Adachi, S.,
Miyazaki, Y., Matsuo, T. & Suga, H. (1993). Chem. Phys. Lett. 213, Schildkamp, W., Bourgeois, D., Wulff, M. & Moffat, K. (1996).
303–308. Science, 274, 1726–1729.
Miyazaki, Y., Matsuo, T. & Suga, H. (2000). J. Phys. Chem. B, 104, Steinfeld, J. I., Francisco, J. S. & Hase, W. L. (1989). Chemical Kinetics
8044–8052. and Dynamics. Englewood Cliffs: Prentice Hall.
Moffat, K. (1998). Acta Cryst. A54, 833–841. Tetreau, C., Blouquit, Y., Novikov, E., Quiniou, E. & Lavalette, L.
Moffat, K. (2001). Chem. Rev. 101, 1569–1581. (2004). Biophys. J. 86, 435–447.
Moffat, K., Chen, Y., Ng, K., McRee, D., Getzoff, E., Rossmann, M., White, E. T., Tan, W. H., Ang, J. M., Tait, S. & Litster, J. D. (2007).
Hajdu, J. & Petsko, G. A. (1992). Philos. Trans. R. Soc. London Ser. Powder Technol. 179, 55–58.
A, 340, 175–190. Zhao, Y. & Schmidt, M. (2009). J. Appl. Cryst. 42, 734–740.
Murray, J. W., Garman, E. F. & Ravelli, R. B. G. (2004). J. Appl. Cryst. Zhou, H.-X., Rivas, G. & Minton, A. P. (2008). Annu. Rev. Biophys.
37, 513–522. 37, 375–397.

206 Marius Schmidt et al.  Five-dimensional crystallography Acta Cryst. (2010). A66, 198–206