ARTICLE
pubs.acs.org/JPCA
Primary Reactions of Bacteriophytochrome Observed with Ultrafast
Mid-Infrared Spectroscopy
)
K. C. Toh,†, Emina A. Stojkovic,‡,^ Alisa B. Rupenyan,†,# Ivo H. M. van Stokkum,† Marian Salumbides,†
Marie-Louise Groot,† Keith Moffat,‡,§ and John T. M. Kennis*,†
†
Biophysics Group, Department of Physics and Astronomy, Faculty of Sciences, VU University, De Boelelaan 1081,
1081HV Amsterdam, The Netherlands
‡
Department of Biochemistry and Molecular Biology, The University of Chicago, Chicago, Illinois 60637, United States
§
Institute for Biophysical Dynamics, The University of Chicago, Chicago, Illinois 60637, United States
bS Supporting Information
ABSTRACT: Phytochromes are red-light photoreceptor proteins that regulate
a variety of responses and cellular processes in plants, bacteria, and fungi. The
phytochrome light activation mechanism involves isomerization around the
C15dC16 double bond of an open-chain tetrapyrrole chromophore, resulting in
a flip of its D-ring. In an important recent development, bacteriophytochrome
(Bph) has been engineered for use as a fluorescent marker in mammalian
tissues. Bphs covalently bind a biliverdin (BV) chromophore, naturally
abundant in mammalian cells. Here, we report an ultrafast time-resolved mid-infrared spectroscopic study on the Pr state of two
highly related Bphs from Rps. palustris, RpBphP2 (P2) and RpBphP3 (P3) with distinct photoconversion and fluorescence
properties. We observed that the BV excited state of P2 decays in 58 ps, while the BV excited state of P3 decays in 362 ps. By
combining ultrafast mid-IR spectroscopy with FTIR spectroscopy on P2 and P3 wild type and mutant proteins, we demonstrate that
the hydrogen bond strength at the ring D carbonyl of the BV chromophore is significantly stronger in P3 as compared to P2. This
result is consistent with the X-ray structures of Bph, which indicate one hydrogen bond from a conserved histidine to the BV ring D
carbonyl for classical bacteriophytochromes such as P2, and one or two additional hydrogen bonds from a serine and a lysine side
chain to the BV ring D carbonyl for P3. We conclude that the hydrogen-bond strength at BV ring D is a key determinant of excitedstate lifetime and fluorescence quantum yield. Excited-state decay is followed by the formation of a primary intermediate that does
not decay on the nanosecond time scale of the experiment, which shows a narrow absorption band at ∼1540 cm-1. Possible origins
of this product band are discussed. This work may aid in rational structure- and mechanism-based conversion of BPh into an efficient
near-IR fluorescent marker.
’ INTRODUCTION
Phytochromes are red-light photoreceptors that play a critical
role in regulating various cellular functions in plant, fungal, and
bacterial kingdoms.1-6 Bacteriophytochromes (Bphs) RpBphP2
(P2) and RpBphP3 (P3) from Rhodoseudomonas palustris in
tandem regulate the synthesis of light harvesting LH4
complexes.7 The photosensory core of these proteins contains
the PAS, GAF, and PHY domains that are covalently attached to
an output/effector domain histidine kinase (HK). Bphs undergo
reversible photoconversion between two metastable photoexcited states denoted as Pr, absorbing near 700 nm, and Pfr,
absorbing near 750 nm. These states are interphotoconvertible
through an isomerization mechanism of the C15dC16 double
bond of a linear tetrapyrrole cofactor, biliverdin (BV). Both P2
and P3 bind BV autocatalytically at ring A through a covalent
linkage with a conserved cysteine in the PAS domain. P2 and P3
have a distinct light response manifested in their absorption
spectra. P2 undergoes classical Pr-Pfr photoconversion,
whereas P3 is the only Bph to date that forms a Pnr state with
a blue-shifted absorption spectrum, peaking at 645 nm. Like their
r 2010 American Chemical Society
counterparts in plant, the photochemistry of Bphs proceeds
through several intermediate stages before attaining a conformation of 15Ea of BV in light state (e.g., Pfr state) upon light
activation.8-11
In an important recent development, Bph has been engineered
for use as a fluorescent marker in mammalian tissues.12 Bph
fluoresces in the near-IR at ∼720 nm, a wavelength less prone to
scattering that can penetrate more deeply into tissue than light
emitted by GFP-derived fluorescent proteins. The BV cofactor is
a naturally occurring cofactor in mammalian tissue that covalently binds to a conserved cysteine in the bacteriophytochrome,
and hence Bph can readily be genetically encoded. It shares such
properties with flavin-binding photoreceptors such as LOV and
BLUF domains for use as photonic switch or sensor.13-17
Special Issue: Graham R. Fleming Festschrift
Received: July 23, 2010
Revised:
December 9, 2010
Published: December 30, 2010
3778
dx.doi.org/10.1021/jp106891x | J. Phys. Chem. A 2011, 115, 3778–3786
The Journal of Physical Chemistry A
Figure 1. (A) Biliverdin (BV) binding site in the X-ray structure of
Rhodopseudomonas palustris P3 (Protein Data Bank code 2OOL.20 (B)
BV binding site in the X-ray structure of Deinoccocus radiodurans BPh
(1ZTU).18 (C) BV chromophore in a ZZZssa configuration with ring
and atom numbering. (D) Partial protein sequence alignment of
bacteriophytochromes P3 and P2 from Rhodopseudomonas palustris,
DrBphP from Deinoccocus radiodurans, AtBphP1 from Agrobacterium
tumafeciens, PaBphP from Pseudomonas aeruginosa, and cyanobacterial
phytochrome Cph1 from Synechocystis sp. pcc 6803. Numerical values
indicate positions of amino acids in P3 primary sequence.
Phytochrome photochemistry is thus of considerable significance
for biomedical research and technology.
The recent determination of crystal structures of various BPhs
and the cyanobacterial phytochrome Cph1 has explored the
light-activated function of phytochromes.8,18-21 The linear tetrapyrrole chromophore is bound to the PAS (BPh) or GAF
(Cph1) domain at ring A through covalent linkage to a conserved
cysteine. BV is largely engulfed by the GAF domain, which
provides most of the hydrogen bonding networks and steric and
hydrophobic interactions to secure the chromophore in position.
The PHY domain forms an extension to the photosensory core of
phytochromes that works in tandem with the GAF domain for
tuning of spectral properties and implementing photochemical
effectiveness. In the Pr state the chromophore assumes a ZZZssa
configuration (Figure 1) and is positioned in its binding pocket
through steric interactions and hydrogen bonds from protein
residues to the pyrrole rings and propionate side chains. Recent
ARTICLE
studies have indicated that 15Za to 15Ea isomerization of the
chromophore at the C15dC16 double bond, which causes a flip of
pyrrole ring D, accompanies formation of the Pfr state.8 The
primary photoproduct, denoted Lumi-R, is formed on the
10-100 ps time scale and adopts the 15Ea configuration.22-27
The X-ray crystal structure of a classical Bph from Deinococcus
radiodurans DrBphP and cyanobacterial phytochrome Cph1
shows a single hydrogen bond between a conserved histidine
residue and the bilin chromophore ring D carbonyl in the
chromophore binding pocket (Figure 1B).18,28,29 In the P3
X-ray crystal structure (Figure 1A), besides the conserved histidine,
Lys-183 and Ser-297 are within hydrogen bonding distance to the
ring D carbonyl.20 Importantly, P3 is the only Bph with three
potential H-bonding partners interacting with D-ring carbonyl
group in the Pr state, as shown in the sequence alignment
(Figure 1D). These different aspects of classical Bph and P3 present
an opportunity to study the influence of the chromophore binding
pocket on the phytochrome photochemistry.
In our previous femtosecond time-resolved absorption studies
on P2 and P3 PAS-GAF-PHY constructs, we have shown that the
excited-state lifetimes and the spectra of P3 are very different
from P2 and other classical Bphs. Strikingly, the BV excited state
of P3 decayed with a monoexponential time constant of 330 ps,
significantly longer than observed in P2 and other phytochromes,
which we related to the hydrogen bond strength at ring D of the
BV chromophore.27 We determined that the two additional polar
residues, lysine and serine located at the immediate vicinity of BV
ring D, are responsible for a lowering the Lumi-R quantum yield
and increasing the BV excited-state lifetime. In addition, we
identified excited-state proton transfer (ESPT) from the BV
pyrrole rings to the protein backbone or a bound water as the
process that deactivates the BV excited state to the Pr state.27
Taken together, the fluorescence quantum yield of P3 is significantly higher than that of classical Bph and with detailed
knowledge about its excited-state dynamics, P3 forms an attractive starting material to generate a highly fluorescent deep-tissue
fluorescent probe by means of rational structure- and mechanismbased engineering.27
The vibrational spectrum of a protein or a protein-bound
chromophore contains a wealth of information about its structure,
the interaction with the environment, and electronic properties.
Time-resolved IR spectroscopy is a powerful tool that can reveal
many of the dynamic structural and physical-chemical properties
of chromophores involved in (photo)biological reactions.30,31 In
addition, it can reveal the involvement of those parts of the protein
that partake in the ongoing reactions. As the primary reactions in
biological photoreceptors proceed on the ultrafast time scale,
femtosecond mid-IR spectroscopy is a method of choice to identify
reaction mechanisms of biological photoreceptors.32-42 The femtosecond time-resolved infrared absorption study on Cph1 had
shown that methine bridges, ring A/D of the phycocyanobilin
(PCB) chromophore, were involved in structural changes in its
primary photochemistry.24 The application of femtosecond IR on
another bacteriophytochrome, Agp1, had given a similar
conclusion.42 Structural changes related to the methine bridges
was also inferred by steady-state resonance Raman studies on the
cryo-trapped intermediate states of various phytochromes.43 However, information about the structural evolution of BV in the early
photochemistry of P2 and P3 is lacking.
In this work, we extend our previous studies by comparing the
early photochemistry of P2 and P3 in the Pr state using ultrafast
mid-IR spectroscopy. We have studied the full photosensory
3779
dx.doi.org/10.1021/jp106891x |J. Phys. Chem. A 2011, 115, 3778–3786
The Journal of Physical Chemistry A
ARTICLE
PAS-GAF-PHY core of P2 and P3, as well as their short PASGAF constructs. Our results support our earlier observation that
excited-state decay is significantly slower in P3 as compared to
P2. Comparison of time-resolved IR spectra and FTIR spectra of
P2 and P3 and site-directed mutants where hydrogen bonding to
ring D was modified provides direct spectroscopic evidence that
hydrogen bonding to the ring D carbonyl is indeed significantly
stronger in P3 than in P2.
’ MATERIAL AND METHODS
Sample Preparation. The detailed preparation of wild type
P2 (PAS-GAF), P2 (PAS-GAF-PHY), and the P2 (PAS-GAFPHY) M169K/A382S (P2KS) mutant and of wild type P3 (PASGAF), P3 (PAS-GAF-PHY), and the P3 (PAS-GAF-PHY)
K183M/S297A (P3MA) mutant bacteriophytochrome proteins
was described previously.20 For the ultrafast mid-IR experiments,
the proteins were dissolved in D2O buffer (20 mM Tris 3 HCl, pD
8 at room temperature). For the FTIR experiments, the proteins
were dissolved in H2O buffer (20 mM Tris 3 HCl, pH 8 at room
temperature).
Femtosecond Mid-IR Spectroscopy. The experimental setup is a home-built spectrometer based on a 1 kHz amplified Ti:
sapphire laser system operating at 1 kHz (Spectra Physics
Hurricane) that allows visible pump/mid infrared probe in a time
window from 180 fs to 3 ns, as previously described.30,32 The red
excitation pulse was generated by means of a noncollinear optical
parametric amplifier and centered around 680 nm, at an excitation
energy of 150-250 nJ. The infrared probe had a spectral width of
200 cm-1, was spectrally dispersed after the sample, and was
detected with a 32-element array detector, leading to a spectral
resolution of 6 cm-1. Vibrational spectra between 1780 and
1450 cm-1 were taken in two intervals and simultaneously
analyzed. Spectra were recorded at 100 time delay points between -20 ps and þ2.8 ns. During the experiments, the sample cell
was continuously translated with a Lissajous scanner, which
ensured sample refreshment after each laser shot and a time interval
of 1 min between successive exposures to the laser beams. Background illumination to photorevert the Bph sample to Pr was
provided with a LED with a center wavelength at 750 nm (P2 PASGAF-PHY, P2 PAS-GAF and P3 PAS-GAF) or 650 nm (P3 PASGAF-PHY).
Data Analysis. The time-resolved data can be described in
terms of a parametric model in which some parameters, such as
those descriptive of the instrument response function (IRF), are
wavenumber-dependent, whereas others, such as the lifetime of a
certain spectrally distinct component, underlay the data at all
wavenumbers. The presence of parameters that underlay the data
at all wavenumbers allow the application of global analysis
techniques, which model wavenumber-invariant parameters as
a function of all data.44 The femtosecond transient absorption
data were globally analyzed using a kinetic model consisting of
sequentially interconverting evolution-associated difference
spectra (EADS), i.e., 1 f 2 f 3 f ... (Figures 2, 3 and 5A) in
which the arrows indicate successive monoexponential decays of
increasing time constant, which can be regarded as the lifetime of
each EADS. The first EADS corresponds to the time-zero
difference spectrum. This procedure enables us to clearly visualize the evolution of the (excited and intermediate) states of the
system. It is important to note that a sequential analysis is
mathematically equivalent to a parallel (sum-of-exponentials)
analysis.44 The analysis program calculates both EADS and
Figure 2. Time-resolved spectroscopy of the Rps. palustris P2 PASGAF-PHY construct. (A) Evolution-associated difference spectra
(EADS) and their corresponding lifetimes resulting from global analysis
of ultrafast mid-IR experiments upon excitation at 680 nm. (B) kinetic
traces at 1702, 1591, and 1540 cm-1.
decay-associated difference spectra (DADS), and the time constants that follow from the analysis apply to both. In general, the
EADS may well reflect mixtures of molecular states such as may
arise, for instance, from heterogeneous ground states or branching
at any point in the molecular evolution or inverted kinetics.45-52
Throughout the manuscript, the EADS are shown in the main
text and the corresponding DADS are shown in the Supporting
Information. The advantage of showing EADS over DADS is that
the former are intuitively more easily interpreted. A detailed
account of the global analysis methodology is given in the
Supporting Information.
3780
dx.doi.org/10.1021/jp106891x |J. Phys. Chem. A 2011, 115, 3778–3786
The Journal of Physical Chemistry A
ARTICLE
with a nitrogen cooled photovoltaic MCT detector (20 MHz, KV
100, Kolmar Technologies, Inc.). Two LEDs, emitting at 680 and
750 nm, were used to convert P2 to its light or dark states,
respectively. For P3, LEDs emitting at 680 and 650 nm were used
instead. The light minus dark FTIR data were obtained, by
subtracting an initially recorded protein dark-state spectrum as
the background spectrum, from the light activated (using the 680
nm LED) protein spectrum. Background and sample interferogram data were averaged from 500-2000 interferogram scans, at
4 cm-1 spectra resolution. Measurements were repeated by
illuminating the sample with a 750 nm (on P2) or a 650 nm
(on P3) light to deactivate the light state of the protein and by
taking a background and a light activated spectrum. The FTIR
sample was prepared using a drop of 2 mL of sample at OD700 nm
of ∼100 (in 20 mM Tris/HCl pH8 buffer) and spread between
two tightly fixed CaF2 windows.
’ RESULTS AND DISCUSSION
Figure 3. Time-resolved spectroscopy of the Rps. palustris P2 PAS-GAF
construct. (A) Evolution-associated difference spectra (EADS) and their
corresponding lifetimes resulting from global analysis of ultrafast mid-IR
experiments upon excitation at 680 nm. (B) Kinetic traces at 1708, 1588,
and 1541 cm-1.
To account for unresolved fast relaxation processes within the
instrument response that become apparent as a sharp peak
around zero time delay, a pulse follower was included in the
global analysis procedure. To avoid any effect of prezero signals
arising from perturbed free induction decay (FID)53 and unresolved relaxation dynamics around zero delay on the outcome
of global analysis procedure, the prezero to 0.5 ps spectra were
given a low weight.
Differential Fourier-Transform Infrared (FTIR) Spectroscopy. The differential FTIR data were recorded at room temperature using a FTIR spectrometer (IFS 66s Bruker) equipped
Ultrafast Mid-IR Spectroscopy of P2. The reaction dynamics of P2 (PAS-GAF-PHY) in the Pr state was investigated
from a subpicosecond time scale up to 3 ns by means of ultrafast
mid-IR spectroscopy. The sample was excited at 680 nm, and a
spectral range of 1470-1780 cm-1 that covers the CdO and
CdC vibration regions was monitored. The data were globally
analyzed in terms of a kinetic scheme with sequentially interconverting species, where each species is characterized by an
EADS that has a specific lifetime. One decay lifetime of 58 ps and
a nondecaying component were required for an adequate fit of
the data. The EADS are shown in Figure 2A, the corresponding
DADS are shown in Figure S1 of Supporting Information
Figure 2B shows kinetic traces at selected wavenumbers. Note
that prezero signals in the kinetics, most apparent in the 1702
cm-1 trace in Figure 2B, arise from perturbed FID53 and do not
relate the Bph photophysics. In some of the kinetics, an
unresolved fast relaxation within the instrument response becomes apparent as a peak around zero delay. It is taken into
account by including a pulse follower in the global analysis
procedure and not further interpreted.
The 58 ps component is assigned to the BV excited state. In our
visible pump-probe experiments, we observed a major excitedstate decay component of 50 ps, consistent with the present data. In
addition, 0.4, 4, and 250 ps components were observed with visible
transient absorption spectroscopy.27 The absence of these components in our femtosecond mid-IR data is likely due to limited signalto-noise, or to a relative insensitivity of the IR spectra to the
dynamics associated with these time constants. The 58 ps EADS
shows major bleach bands in the range 1570-1640 cm-1,
attributed to the BV chromophore CdC methine bridges vibrational
bands.24,42 These bands are located at 1591, 1613, and 1635 cm-1.
Their frequencies closely resemble those in the Pr state of plant
phytochrome A (PhyA), DrBphP, and Agrobacterium bacteriophytochrome 1 (Agp1), as observed with resonant Raman
spectroscopy.54,55 The largest bleach band at 1591 cm-1 in our
femtosecond mid-IR data is assignable to the ring C-D methine
bridge stretch band56 (see Figure 1C for BV ring and atom
numbering). The bleach band at ∼1630 cm-1 is assigned to the
ring A-B methine bridge (C4dC5) vibration band.56 Due to the
reduction of bond orders in the excited states, these bleach bands
are expected to be downshifted in the S1 state.24
The 58 ps EADS shows major bleach bands at 1728 and
1702 cm-1 (Figure 2A). In PhyA, the high-frequency carbonyl
3781
dx.doi.org/10.1021/jp106891x |J. Phys. Chem. A 2011, 115, 3778–3786
The Journal of Physical Chemistry A
Figure 4. Time-resolved spectroscopy of the Rps. palustris P3 PASGAF-PHY construct. (A) Evolution-associated difference spectra
(EADS) and their corresponding lifetimes resulting from global analysis
of ultrafast mid-IR experiments upon excitation at 680 nm. (B) Kinetic
traces at 1678, 1588, and 1541 cm-1.
band at 1730 cm-1 was assigned to ring A C1dO stretching
through 18O isotope labeling of PCB at this site.57,58 Thus, the
1728 cm-1 band can confidently be assigned to the BV ring A
C1dO stretch vibration. It follows that the 1702 cm-1 band is
associated with the BV ring D C19dO vibration, which is
consistent with assignments made for Agp1 by Diller and coworkers26 and Bartl and co-workers.59 The CdO stretching
bands observed here have a lower frequency than those found in
Cph1 (located at 1738 and ∼1720 cm-1 respectively), probably
due to a different environment of the chromophore. Also, the
different conjugation at ring A between BV and PCB may play a
ARTICLE
role. The 58 ps EADS shows a bleach band at 1540 cm-1 that was
observed previously in femtosecond mid-IR spectroscopy on
phytochromes but not interpreted.24,42
The final nondecaying component is formed from the BV
excited state in 58 ps and persists through our experimental time
scale of 3 ns. It is assigned to the primary photoproduct Lumi-R.
The nondecaying component has a low amplitude and most of its
features do not rise above the noise level, in keeping with the low
quantum yield of Lumi-R formation of 0.13.27 Surprisingly,
however, is the occurrence of a very sharp absorption band at
1541 cm-1, at the same frequency as the bleach in the 58 ps
EADS. Figure 2B shows the kinetic trace that demonstrates the
rise of this positive-amplitude feature. Its possible origin will be
discussed further on in the paper.
We also performed ultrafast IR experiments on the P2 PASGAF construct that undergoes limited photoconversion20 under
identical experimental conditions. Two time constants of 4 and
175 ps and a long-lived component were required for an
adequate fit of the data. Figure 3A shows the resulting EADS,
while kinetic traces at selected wavelengths are shown in
Figure 3B. Figure S2 (Supporting Information) shows the
DADS. The BV excited-state lifetime is significantly longer at
175 ps than in the P2 PAS-GAF-PHY construct, which agrees
with our findings from ultrafast visible spectroscopy.60 The
infrared signature of excited-state BV is essentially the same
as that in P2 PAS-GAF-PHY, with CdO bleaches at 1734 and
1708 cm-1, methine bridge stretches at 1588, 1611, and 1635 cm-1,
and a bleach at 1536 cm-1. As in the PAS-GAF-PHY construct, the
long-lived photoproduct shows a pronounced induced absorption
band at 1541 cm-1.
In the P2 PAS-GAF construct, a 4 ps component was resolved.
Inspection of the corresponding DADS (Figure S1, Supporting
Information) reveals a pattern of alternating negative and
positive bands at similar amplitudes, with a negative band at
1690 cm-1, a positive band at 1655 cm-1, and a broad negative
band near 1590 cm-1. The pattern of the 4 ps DADS does not
resemble BV excited-state decay, as the negative feature at
1690 cm-1 does not correspond to a BV CdO vibration in the Pr
state, and the broad negative feature near 1590 cm-1 does not
resemble the BV methine bridge stretches in Pr as observed for the
175 ps DADS. Hence, we conclude that the 4 ps component
represents a relaxation process in the excited state.27
Ultrafast Mid-IR Spectroscopy of P3. We investigated the
reaction dynamics of the P3 PAS-GAF-PHY construct in the Pr
state by means of ultrafast mid-IR spectroscopy to uncover
structural aspects of its photoreaction. The sample was excited
at 680 nm and a spectral range of 1470-1780 cm-1 that covers
the CdO and CdC vibration regions was monitored. Global
analysis indicated that three lifetime components of 37 and 362
ps and a nondecaying component were required for an adequate
fit of the data. The EADS are shown in Figure 4A, whereas kinetic
traces at selected wavenumbers are shown in Figure 4B. Figure S3
(Supporting Information) shows the DADS.
The first EADS (Figure 4A, circles) has a lifetime of 37 ps,
similar to the 53 ps component observed with ultrafast visible
spectroscopy. Its origin will be discussed below. It evolves into
the EADS that has a lifetime of 362 ps (Figure 4A, solid
diamonds). The 362 EADS corresponds to a relaxed form of
the BV excited state similar to, although somewhat shorter than,
that observed with ultrafast visible spectroscopy, which had a
lifetime of 450 ps.27 It has an overall shape similar to that of the
BV excited state of P2 (Figure 2A) but differs in some important
3782
dx.doi.org/10.1021/jp106891x |J. Phys. Chem. A 2011, 115, 3778–3786
The Journal of Physical Chemistry A
aspects. It shows strong bleach bands at 1588, 1603, and
1630 cm-1 that represent the BV chromophore CdC methine
bridges vibrational bands. The strongest bleach band at
1588 cm-1 is assigned to the BV ring C-D methine bridge
(C15dC16) stretching. The 1630 cm-1 bleach band is assigned
to the ring A-B methine bridge (C4dC5) stretching. These
bands are similar to those observed in P2 (Figure 2A), Cph1 and
Agp1.24,42 Inspection of the carbonyl region (1680-1740 cm-1)
reveals a pattern that is quite different from that of P2: The ring A
CdO stretching band is located at 1736 cm-1 (Figure 4A, solid
diamonds), a frequency comparable to those found in P2, Cph1
and Agp1.24,42 Strikingly, a pronounced bleach is observed at
1678 cm-1, a frequency where P2 and other phytochromes show
no such signal. Given its frequency and its prompt rise within the
instrument response, it must correspond to a CdO mode of the
BV chromophore. As the ring A C1dO was already firmly
assigned to the 1736 cm-1 band,58 the 1678 cm-1 band most
likely corresponds to the BV ring D C19dO stretch mode. We
will substantiate this assignment below by means of FTIR
spectroscopy on P2, P3, and site-directed mutants. This observation implies that in P3, the ring D CdO stretch mode has a
significantly lower frequency than in other phytochromes. We
will demonstrate below that the downshift of the ring D CdO
frequency results from the increased hydrogen bond strength to
ring D in P3. As in P2, a strong and broad bleach band at 1541
cm-1 is observed in the BV excited state of P3.
The BV excited state evolves in 362 ps to the nondecaying
EADS (open diamonds), which is assigned to the primary
photoproduct Lumi-R. This EADS has a very low amplitude, in
keeping with the low Lumi-R quantum yield of P3 of 0.06.27 As in
P2, a sharp induced absorption band at 1541 cm-1 is observed.
Figure 4B shows the rise of this product band. Other bands in this
EADS are mostly buried in the noise and will not be further
considered.
The spectral evolution shows a 37 ps component in addition to
the 362 ps and nondecaying components. Inspection of the
DADS (Figure S3, Supporting Information) shows negative
bands at 1730, 1690, and 1655 cm-1, a broad negative feature
between 1580 and 1630 cm-1, and positive bands at 1675 cm-1
and at frequencies below 1530 cm-1. The mostly negative
pattern suggests that the 37 ps component mainly represents a
BV excited-state decay process. However, the negative signals at
1690 at 1660 cm-1 do not match the carbonyl frequencies in Pr,
the bands at 1580-1630 are shifted by ∼6 cm-1 with respect to
those of Pr, and we conclude that the 37 ps does not represent
decay of the Pr excited state. We conclude that this component
remains difficult to interpret in specific molecular terms.
We also performed ultrafast IR experiments on the P3 PASGAF construct under identical experimental conditions. P3 PASGAF does not form the Pnr state and undergoes photoconversion to a Meta-R like state.20 Figure 5A shows the resulting
EADS, kinetic traces at selected wavelengths are shown in
Figure 5B. Figure S4 (Supporting Information) shows the
DADS. The BV excited-state lifetime is significantly longer at
435 ps than in the P3 PAS-GAF-PHY construct, which agrees
with our findings from ultrafast visible spectroscopy.60 In addition, the 9.5 ps component observed in the P3 PAS-GAF-PHY
construct does not appear in these data. The infrared signature of
excited-state BV is essentially the same as that in P3 PAS-GAFPHY, with the ring A CdO bleaches at 1741 cm-1, the ring D
CdO bleach at 1677 cm-1, methine bridge stretches at
1588, 1612, and 1635 cm-1, and a bleach at 1546 cm-1. As in
ARTICLE
Figure 5. Time-resolved spectroscopy of the Rps. palustris P3 PAS-GAF
construct. (A) Evolution-associated difference spectra (EADS) and their
corresponding lifetimes resulting from global analysis of ultrafast mid-IR
experiments upon excitation at 680 nm. (B) Kinetic traces at 1677, 1588,
and 1541 cm-1.
the PAS-GAF-PHY construct, the long-lived photoproduct
shows a pronounced induced absorption band at 1541 cm-1.
FTIR Spectroscopy. In the ultrafast IR experiments on P3, we
observed a BV CdO band at a particularly low frequency of
1678 cm-1 (Figures 4A and 5A) as compared to P2 (Figures 2A
and 3A) and other classical (bacterio)phytochromes24,42 and
assigned it to the ring D carbonyl stretch mode. The lowering of
the ring D carbonyl frequency is most likely due to the stronger
hydrogen bonding at this site: the P3 X-ray structure shows 2-3
amino acids hydrogen bonding to the carbonyl of ring D in dark
state, i.e., the conserved His-299, Lys-183, and Ser-297.20 In
classical phytochromes, only a conserved His hydrogen bonds to
ring D18,28,61 (Figure 1). To investigate this idea, we performed
3783
dx.doi.org/10.1021/jp106891x |J. Phys. Chem. A 2011, 115, 3778–3786
The Journal of Physical Chemistry A
light-minus-dark FTIR spectroscopy on wild type P3 and the
P3MA mutant,27 where the two polar amino acids Lys-183 and
Ser-297 are mutated to methionine and alanine, respectively,
eliminating the two hydrogen bonds. Also, FTIR spectra were
taken on the classical Bph P2 and its P2KS mutant, where Met169 and Ala-382 were replaced by polar residues Lys and Ser. All
FTIR experiments were performed on PAS-GAF-PHY constructs. We note that the P2KS and P3MA mutants retain their
respective wild-type photoconversion properties, i.e., P2KS converts to Pfr and P3MA converts to Pnr.20
Figure 6A shows the light-minus-dark FTIR spectra of P3 wild
type (solid line) and the P3MA mutant (crossed symbol line). In
P3 wild type, carbonyl bleaches are observed at 1734 and 1685
cm-1 (Figure 6A, solid line), assigned to ring A C1dO and ring
D C19dO, respectively. This result is consistent with those of
ultrafast IR spectroscopy, which indicated ring A and D carbonyl
frequencies at 1736 and 1678 cm-1 (Figure 4A) (note that the
FTIR spectra were taken in H2O and ultrafast IR spectra in D2O,
giving rise to slightly different frequencies). In the P3MA mutant,
the bleach at 1685 cm-1 has disappeared and a new bleach at
1711 cm-1 has appeared (Figure 6A, crossed symbol line),
indicating that the BV ring D carbonyl shifts up by 26 cm-1
upon replacement of the hydrogen bonding amino acids Ser and
Lys by nonpolar amino acids Met and Ala.
Figure 6B shows the light-minus-dark FTIR spectra of P2 wild
type (solid line) and the P2KS mutant (crossed symbol line). In
P2 wild type, which forms only a single hydrogen bond from a
conserved His to ring D, bleaches are observed at 1732 and 1703
cm-1 (Figure 6B, solid line), assigned to the ring A C1dO and
ring D C19dO, respectively. With ultrafast IR, similar frequencies are observed at 1728 and 1702 cm-1 (Figure 2A). In the
P2KS mutant, where M169 was replaced by Lys and A283 by Ser
(the equivalent amino acids in P3), it is expected that two
additional hydrogen bonds are formed to the ring D carbonyl.
Indeed, the FTIR spectrum of the P2KS mutant (Figure 6B,
crossed symbol line) shows a downshifting of ring D C19dO
stretching frequency from 1703 (in WT) to 1676 cm-1 (in
P2KS) in the dark state.
We conclude that in P3 wild type in the dark, hydrogen
bonding to the ring D carbonyl is significantly stronger than in
P2, consistent with the X-ray structures of P3 and classical (B)ph
(Figure 1).20,28,61 This result further corroborates our earlier
work,27 where we identified the factors that determine the BV
excited-state lifetimes and isomerization quantum yields of wildtype and mutants of P2 and P3. We concluded that the hydrogenbond strength to ring D is rate-limiting for isomerization and the
excited-state lifetime and that the quantum yields of fluorescence
and isomerization are determined by excited-state deprotonation
of biliverdin at the pyrrole rings, in competition with hydrogenbond rupture between the D-ring and the apoprotein.27
Origin of 1540 cm-1 Band. An interesting observation of the
present work is a band near 1540 cm-1 that appears as a broad
bleach in the BV excited state (thus corresponding to Pr) and a
narrow induced absorption at essentially the same frequency in
the primary photoproduct in all Bph variants studied here, i.e., P2
PAS-GAF-PHY, P2 PAS-GAF, P3 PAS-GAF-PHY, and P3 PASGAF. Because the 1540 cm-1 band in the primary photoproduct
has an amplitude that exceeds all other product bands, it might
correspond to a molecular state that plays an integral role in the
primary photochemistry of bacteriophytochrome. The question
arises what specific vibrational mode(s) of BV these bands
belong to. We first note that the negative (Pr) and positive
ARTICLE
Figure 6. (A) Light-minus-dark FTIR spectroscopy of wild type P3
PAS-GAF-PHY (solid line) and the P3MA PAS-GAF-PHY mutant
(crossed line). (B) Light-minus-dark FTIR spectroscopy of wild type
P2 PAS-GAF-PHY (solid line) and the P2KS PAS-GAF-PHY mutant
(crossed line).
(photoproduct) bands do not necessarily relate to the exact same
vibrational mode. In previous ultrafast mid-IR experiments on
Cph1 and Agp1, a similar but smaller bleach band at ∼1540
cm-1 was observed.24,42 With resonant Raman spectroscopy on
plant PhyA in D2O, Pr shows a band at 1547 cm-1, and cryotrapped Lumi-R exhibited a sharp band at 1541 cm-1.43 In
DrBphP, a band at 1546 cm-1 was observed in D2O with
resonant Raman spectroscopy.55 In neither of these studies were
the bands near 1540 cm-1 interpreted. It does not belong to the
N-H in-plane bending vibrational band (at ∼1570 cm-1 in
PhyA, Agp1 and DrBphP in H2O10,54,55) because this band
downshifts to below 1100 cm-1 upon deuteration.43,55
As a first possible origin of the 1540 cm-1 band, calculations
and IR absorption experiments on the model compound biliverdin dimetyl ester have indicated a mode at 1543 cm-1 in D2O
that corresponded to the CdC ring stretch and C-vinyl stretch
band of ring D.57 As a second possibility for the origin of
the ∼1540 cm-1 band, recent DFT calculations have indicated
that in a pyrrole-N deuterated PCB chromophore in a ZZZssa
configuration, a band near 1540 cm-1 arises that includes mainly
stretching coordinates from ring B and, to a minor content, from
the B-C and A-B methine bridges. In the ZZEssa configuration, this band slightly shifts to 1542 cm-1.62 It is difficult to
3784
dx.doi.org/10.1021/jp106891x |J. Phys. Chem. A 2011, 115, 3778–3786
The Journal of Physical Chemistry A
understand how such a band assignment would relate to Lumi-R
formation because no significant changes are thought to occur at
ring B upon isomerization about the C15dC16 double bond. We
note, however, that the long-lived 1540 cm-1 absorption does
not necessarily relate to Lumi-R and may correspond to a
ground-state intermediate on the pathway of Pr reformation.
Such ground-state intermediates were recently observed for
various phytochromes.26,63 At this stage the origin of the
1540 cm-1 bands remain unclear and will be the subject of
further studies.
’ CONCLUSIONS
Here, we have reported an ultrafast mid-IR study of two
related bacteriophytochromes: P2, which shows classical Pr-Pfr
photochemistry and P3, which shows an unusual Pr - Pnr
photochemistry. In P2, BV excited-state decay occurs with a time
constant of 58 ps, largely consistent with our results from visible
transient absorption spectroscopy which indicated a biexponential decay with a main decay component of 60 ps. Excited-state
decay in P3 is significantly slower with a time constant of 362 ps,
which is also consistent with visible transient absorption
results.27 In our previous work, we proposed that the slower
excited-state decay of P3 is related to an increased hydrogen
bond strength at ring D, with three amino acid side chains (His,
Lys, and Ser) competing for hydrogen bonding to the ring D
carbonyl in P3.20,27 In P2, only one such hydrogen bond can form
from conserved His to ring D.18,61 Here, we obtained direct
spectroscopic evidence for increased hydrogen bond strength at
ring D in P3: ultrafast IR spectroscopy on P2 and P3, and FTIR
spectroscopy on the P2 and P3 wild types and P3MA and P2KS
mutants indicated that in P3, the ring D C19dO stretch mode has
an unusually low vibrational frequency at 1685-1678 cm-1. In
contrast, in P2 the ring D C19dO stretch mode is located at
1703 cm-1, which demonstrates that P3 has one or two additional
hydrogen bonds to ring D.
’ ASSOCIATED CONTENT
Supporting Information. DADS spectra. Discussion of
model based data analysis. This material is available free of charge
via the Internet at http://pubs.acs.org.
bS
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: john@nat.vu.nl. Phone þ31205987212.
)
Present Addresses
Imperial College, London, United Kingdom.
Department of Biology, Northeastern Illinois University, Chicago.
#
Chemistry Department, University of Amsterdam, Amsterdam,
The Netherlands.
^
’ ACKNOWLEDGMENT
We are grateful to Peter Hildebrandt of Technical University
Berlin for sharing unpublished results. We thank Jos Thieme for
technical support. K.C.T. and J.T.M.K. were supported by the
Earth and Life Sciences Council of The Netherlands Foundation
for Scientific Research (NWO-ALW) through a VIDI grant to J.
T.M.K. E.A.S and K.M. were supported by an NIH grant
GM036452 to K. M.
ARTICLE
Bph, bacteriophytochrome; BV, biliverdin; PCB, phycocyanobilin; EADS, evolution-associated difference spectrum; DADS,
decay-associated difference spectrum; ESPT, excited-state proton
transfer
’ REFERENCES
(1) Borthwick, H. A.; Hendricks, S. B.; Parker, M. W.; Toole, E. H.;
Toole, V. K. Proc. Natl. Acad. Sci. U.S.A. 1952, 38, 662.
(2) Butler, W. L.; Norris, K. H.; Siegelman, H. W.; Hendricks, S. B.
Proc. Natl. Acad. Sci. U.S.A. 1959, 45, 1703.
(3) Davis, S. J.; Vener, A. V.; Vierstra, R. D. Science 1999, 286, 2517.
(4) Hughes, J.; Lamparter, T.; Mittmann, F.; Hartmann, E.; Gartner,
W.; Wilde, A.; Borner, T. Nature 1997, 386, 663.
(5) Lamparter, T.; Michael, N.; Mittmann, F.; Esteban, B. Proc. Natl.
Acad. Sci. U.S.A. 2002, 99, 11628.
(6) Blumenstein, A.; Vienken, K.; Tasler, R.; Purschwitz, J.; Veith,
D.; Frankenberg-Dinkel, N.; Fischer, R. Curr. Biol. 2005, 15, 1833.
(7) Giraud, E.; Zappa, S.; Vuillet, L.; Adriano, J. M.; Hannibal, L.;
Fardoux, J.; Berthomieu, C.; Bouyer, P.; Pignol, D.; Vermeglio, A. J. Biol.
Chem. 2005, 280, 32389.
(8) Yang, X.; Kuk, J.; Moffat, K. Proc. Natl. Acad. Sci. U.S.A. 2008,
105, 14715.
(9) Inomata, K.; Hammam, M. A.; Kinoshita, H.; Murata, Y.; Khawn,
H.; Noack, S.; Michael, N.; Lamparter, T. J. Biol. Chem. 2005, 280, 24491.
(10) Borucki, B.; von Stetten, D.; Seibeck, S.; Lamparter, T.;
Michael, N.; Mroginski, M. A.; Otto, H.; Murgida, D. H.; Heyn,
M. P.; Hildebrandt, P. J. Biol. Chem. 2005, 280, 34358.
(11) von Stetten, D.; Seibeck, S.; Michael, N.; Scheerer, P.;
Mroginski, M. A.; Murgida, D. H.; Krauss, N.; Heyn, M. P.; Hildebrandt,
P.; Borucki, B.; Lamparter, T. J. Biol. Chem. 2007, 282, 2116.
(12) Shu, X.; Royant, A.; Lin, M. Z.; Aguilera, T. A.; Lev-Ram, V.;
Steinbach, P. A.; Tsien, R. Y. Science 2009, 324, 804.
(13) Drepper, T.; Eggert, T.; Circolone, F.; Heck, A.; Krauss, U.;
Guterl, J. K.; Wendorff, M.; Losi, A.; Gartner, W.; Jaeger, K. E. Nat.
Biotechnol. 2007, 25, 443.
(14) Schroder-Lang, S.; Schwarzel, M.; Seifert, R.; Strunker, T.;
Kateriya, S.; Looser, J.; Watanabe, M.; Kaupp, U. B.; Hegemann, P.;
Nagel, G. Nat. Methods 2007, 4, 39.
(15) Strickland, D.; Moffat, K.; Sosnick, T. R. Proc. Natl. Acad. Sci. U.
S.A. 2008, 105, 10709.
(16) Wu, Y. I.; Frey, D.; Lungu, O. I.; Jaehrig, A.; Schlichting, I.;
Kuhlman, B.; Hahn, K. M. Nature 2009, 461, 104.
(17) Chapman, S.; Faulkner, C.; Kaiserli, E.; Garcia-Mata, C.;
Savenkov, E. I.; Roberts, A. G.; Oparka, K. J.; Christie, J. M. Proc. Natl.
Acad. Sci. U.S.A. 2008, 105, 20038.
(18) Wagner, J. R.; Brunzelle, J. S.; Forest, K. T.; Vierstra, R. D.
Nature 2005, 438, 325.
(19) Wagner, J. R.; Zhang, J.; Brunzelle, J. S.; Vierstra, R. D.; Forest,
K. T. J. Biol. Chem. 2007, 282, 12298.
(20) Yang, X.; Stojkovic, E. A.; Kuk, J.; Moffat, K. Proc. Natl. Acad.
Sci. U.S.A. 2007, 104, 12571.
(21) Rohmer, T.; Lang, C.; Hughes, J.; Essen, L. O.; Gartner, W.;
Matysik, J. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 15229.
(22) Andel, F., III; Hansson, K. C.; Gai, F.; Anfinrud, P. A.; Mathies,
R. A. Biospectroscopy 1997, 3, 421.
(23) Heyne, K.; Herbst, J.; Stehlik, D.; Esteban, B.; Lamparter, T.;
Hughes, J.; Diller, R. Biophys. J. 2002, 82, 1004.
(24) van Thor, J. J.; Ronayne, K. L.; Towrie, M. J. Am. Chem. Soc.
2007, 129, 126.
(25) Dasgupta, J.; Frontiera, R. R.; Taylor, K. C.; Lagarias, J. C.;
Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 1784.
(26) Schumann, C.; Gross, R.; Michael, N.; Lamparter, T.; Diller, R.
ChemPhysChem 2007, 8, 1657.
(27) Toh, K. C.; Stojkovic, E. A.; van Stokkum, I. H. M.; Moffat, K.;
Kennis, J. T. M. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 9170.
(28) Essen, L. O.; Mailliet, J.; Hughes, J. Proc. Natl. Acad. Sci. U.S.A.
2008, 105, 14709.
3785
dx.doi.org/10.1021/jp106891x |J. Phys. Chem. A 2011, 115, 3778–3786
The Journal of Physical Chemistry A
(29) Cornilescu, G.; Ulijasz, A. T.; Cornilescu, C. C.; Markley, J. L.;
Vierstra, R. D. J. Mol. Biol. 2008, 383, 403-413.
(30) Groot, M. L.; van Wilderen, L. J. G. W.; Di Donato, M.
Photochem. Photobiol. Sci. 2007, 6, 501.
(31) Kotting, C.; Gerwert, K. ChemPhysChem 2005, 6, 881.
(32) Groot, M. L.; van Wilderen, L. J. G. W.; Larsen, D. S.; van der
Horst, M. A.; van Stokkum, I. H. M.; Hellingwerf, K. J.; van Grondelle, R.
Biochemistry 2003, 42, 10054.
(33) Herbst, J.; Heyne, K.; Diller, R. Science 2002, 297, 822.
(34) Heyne, K.; Mohammed, O. F.; Usman, A.; Dreyer, J.; Nibbering,
E. T. J.; Cusanovich, M. A. J. Am. Chem. Soc. 2005, 127, 18100.
(35) Schumann, C.; Gross, R.; Wolf, M. M. N.; Diller, R.; Michael,
N.; Lamparter, T. Biophys. J. 2008, 94, 3189.
(36) Stoner-Ma, D.; Jaye, A. A.; Matousek, P.; Towrie, M.; Meech,
S. R.; Tonge, P. J. J. Am. Chem. Soc. 2005, 127, 2864.
(37) van Thor, J. J.; Ronayne, K. L.; Towrie, M. J. Am. Chem. Soc.
2007, 129, 126.
(38) van Wilderen, L. J. G. W.; van der Horst, M. A.; van Stokkum,
I. H. M.; Hellingwerf, K. J.; van Grondelle, R.; Groot, M. L. Proc. Natl.
Acad. Sci. U.S.A. 2006, 103, 15050.
(39) Bonetti, C.; Mathes, T.; van Stokkum, I. H. M.; Mullen, K. M.;
Groot, M. L.; van Grondelle, R.; Hegemann, P.; Kennis, J. T. M. Biophys.
J. 2008, 95, 4790.
(40) Alexandre, M. T. A.; Domratcheva, T.; Bonetti, C.; van
Wilderen, L.; van Grondelle, R.; Groot, M. L.; Hellingwerf, K. J.; Kennis,
J. T. M. Biophys. J. 2009, 97, 227.
(41) Kennis, J. T. M.; Groot, M. L. Curr. Opin. Struct. Biol. 2007, 17,
623.
(42) Schumann, C.; Gross, R.; Michael, N.; Lamparter, T.; Diller, R.
ChemPhysChem 2007, 8, 1657.
(43) Kneip, C.; Hildebrandt, P.; Schlamann, W.; Braslavsky, S. E.;
Mark, F.; Schaffner, K. Biochemistry 1999, 38, 15185.
(44) van Stokkum, I. H. M.; Larsen, D. S.; van Grondelle, R.
Biochimica Et Biophysica Acta-Bioenergetics 2004, 1657, 82.
(45) Gradinaru, C. C.; Kennis, J. T. M.; Papagiannakis, E.; van
Stokkum, I. H. M.; Cogdell, R. J.; Fleming, G. R.; Niederman, R. A.; van
Grondelle, R. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 2364.
(46) Papagiannakis, E.; Kennis, J. T. M.; van Stokkum, I. H. M.;
Cogdell, R. J.; van Grondelle, R. Proc. Natl. Acad. Sci. U.S.A. 2002, 99,
6017.
(47) Gauden, M.; Yeremenko, S.; Laan, W.; van Stokkum, I. H. M.;
Ihalainen, J. A.; van Grondelle, R.; Hellingwerf, K. J.; Kennis, J. T. M.
Biochemistry 2005, 44, 3653.
(48) Gauden, M.; Grinstead, J. S.; Laan, W.; van Stokkum, H. M.;
Avila-Perez, M.; Toh, K. C.; Boelens, R.; Kaptein, R.; van Grondelle, R.;
Hellingwerf, K. J.; Kennis, J. T. M. Biochemistry 2007, 46, 7405.
(49) Berera, R.; Herrero, C.; van Stokkum, L. H. M.; Vengris, M.;
Kodis, G.; Palacios, R. E.; van Amerongen, H.; van Grondelle, R.; Gust,
D.; Moore, T. A.; Moore, A. L.; Kennis, J. T. M. Proc. Natl. Acad. Sci. U.S.
A. 2006, 103, 5343.
(50) Bonetti, C.; Alexandre, M. T. A.; van Stokkum, I. H. M.; Hiller,
R. G.; Groot, M. L.; van Grondelle, R.; Kennis, J. T. M. Phys. Chem.
Chem. Phys. 2010, 12, 9256.
(51) Berera, R.; van Stokkum, I. H. M.; Kodis, G.; Keirstead, A. E.;
Pillai, S.; Herrero, C.; Palacios, R. E.; Vengris, M.; van Grondelle, R.;
Gust, D.; Moore, T. A.; Moore, A. L.; Kennis, J. T. M. J. Phys. Chem. B
2007, 111, 6868.
(52) Berera, R.; van Grondelle, R.; Kennis, J. T. M. Photosynth. Res.
2009, 101, 105.
(53) Hamm, P. Chem. Phys. 1995, 200, 415.
(54) Andel, F., 3rd; Lagarias, J. C.; Mathies, R. A. Biochemistry 1996,
35, 15997.
(55) Wagner, J. R.; Zhang, J.; von Stetten, D.; Gunther, M.; Murgida,
D. H.; Mroginski, M. A.; Walker, J. M.; Forest, K. T.; Hildebrandt, P.;
Vierstra, R. D. J. Biol. Chem. 2008, 283, 12212.
(56) Margulies, L.; Toporowics, M. J. Am. Chem. Soc. 1984, 106, 7331.
(57) Smit, K.; Matysik, J.; Hildebrandt, P.; Mark, F. J. Phys. Chem.
1993, 97, 11887.
ARTICLE
(58) Foerstendorf, H.; Benda, C.; Gartner, W.; Storf, M.; Scheer, H.;
Siebert, F. Biochemistry 2001, 40, 14952.
(59) Piwowarski, P.; Ritter, E.; Hofmann, K. P.; Hildebrandt, P.; von
Stetten, D.; Scheerer, P.; Michael, N.; Lamparter, T.; Bartl, F. ChemPhysChem 2010, 11, 1207.
(60) Toh, K. C.; Stojkovic, E. A.; van Stokkum, I. H. M.; Moffat, K.;
Kennis, J. T. M. Manuscript submitted 2010.
(61) Wagner, J. R.; Zhang, J. R.; Brunzelle, J. S.; Vierstra, R. D.;
Forest, K. T. J. Biol. Chem. 2007, 282, 12298.
(62) Borucki, B.; von Stetten, D.; Seibeck, S.; Lamparter, T.;
Michael, N.; Mroginski, M. A.; Otto, H.; Murgida, D. H.; Heyn,
M. P.; Hildebrandt, P. J. Biol. Chem. 2005, 280, 34358.
(63) van Wilderen, L.; Clark, I. P.; Towrie, M.; van Thor, J. J. J. Phys.
Chem. B 2009, 113, 16354.
3786
dx.doi.org/10.1021/jp106891x |J. Phys. Chem. A 2011, 115, 3778–3786
DADS P2 PAS-GAF-PHY
2
1
-3
∆A (10 )
0
-1
-2
-3
-4
58 ps
non-decaying
-5
1750
1700
1650
1600
1550
1500
-1
Wavenumber (cm )
Fig. S1. Decay-associated difference spectra (DADS) that follow from global analysis of
time-resolved IR data of Rps. palustris P2 PAS-GAF-PHY bacteriophytochrome. The
excitation wavelength was 680 nm.
DADS P2 PAS-GAF
2
0
-3
∆A (10 )
1
-1
-2
-3
4 ps
175 ps
non-decaying
1750
1700
1650
1600
1550
1500
-1
Wavenumber (cm )
Fig. S2. Decay-associated difference spectra (DADS) that follow from global analysis of
time-resolved IR data of Rps. palustris P2 PAS-GAF bacteriophytochrome. The
excitation wavelength was 680 nm.
DADS P3 PAS-GAF-PHY
2
-3
∆A (10 )
1
0
-1
37 ps
362 ps
non-decaying
-2
1750
1700
1650
1600
1550
1500
-1
Wavenumber (cm )
Fig. S3. Decay-associated difference spectra (DADS) that follow from global analysis of
time-resolved IR data of Rps. palustris P3 PAS-GAF-PHY bacteriophytochrome. The
excitation wavelength was 680 nm.
DADS P3 PAS-GAF
-3
∆A (10 )
1
0
-1
435 ps
non-decaying
1750
1700
1650
1600
1550
1500
-1
Wavenumber (cm )
Fig. S5. Decay-associated difference spectra (DADS) that follow from global analysis of
time-resolved IR data of Rps. palustris P3 PAS-GAF bacteriophytochrome. The
excitation wavelength was 680 nm.
Supporting Information to
The primary reactions of bacteriophytochrome observed with
ultrafast mid-infrared spectroscopy
K.C. Toh, Emina A. Stojković, Alisa B. Rupenyan, Ivo H.M. van Stokkum, Marian
Salumbides, Marie-Louise Groot, Keith Moffat, John T.M. Kennis
Model based data analysis
The aim of data analysis is to obtain a model-based description of the full data set in
terms of a model containing a small number of precisely estimated parameters, of which
the rate constants and spectra are the most relevant. The basic ingredient of kinetic
models, namely the exponential decays, will be described first, followed by use of these
ingredients for global and target analysis1-3 of the full data. Our main assumption is that
the time and wavelength properties of the system of interest are separable, which means
that spectra of species or states are constant. For details on parameter estimation
techniques the reader is referred to1-4. Software issues are discussed in5.
A. Modeling an exponential decay
Here an expression is derived for describing an exponentially decaying component. The
instrument response function (IRF) i(t) can usually adequately be modeled with a
Gaussian with parameters µ and ∆ for, respectively, location and full width at half
maximum (FWHM):
1
i (t ) = ~
exp(− log(2)(2(t − µ ) / ∆) 2 )
∆ 2π
~
where ∆ = ∆ /(2 2 log(2) ) . The convolution (indicated by an *) of this IRF with an
exponential decay (with rate k) yields an analytical expression which facilitates the
estimation of the IRF parameters µ and ∆:
S.1
1
k ∆% 2
t − ( µ + k ∆% 2 ))
c I (t , k , µ , ∆) = exp(− kt ) ∗ i (t ) = exp(−kt )exp(k ( µ +
)){1 + erf (
}
2
2
2∆%
The wavelength dependence of the IRF location µ can be modeled with a polynomial.
jmax
µ (λ ) = µλ + ∑ a j (λ − λc ) j
c
j =1
Typically, a parabola is adequate and the order of this polynomial ( jmax ) is two. The
reference wavelength λc is usually at the center of the spectrograph.
B. Global and target analysis
The basis of global analysis is the superposition principle, which states that the measured
data ψ (t , λ ) result from a superposition of the spectral properties ε l (λ ) of the
components present in the system of interest weighted by their concentration cl (t ) .
ncomp
ψ (t , λ ) =
∑ c (t )ε (λ )
l
l
l =1
The cl (t ) of all ncomp components are described by a compartmental model that consists
of first-order differential equations, with as solution sums of exponential decays. We
consider three types of compartmental models: (1) a model with components decaying
monoexponentially in parallel, which yields Decay Associated Difference Spectra
(DADS), (2) a sequential model with increasing lifetimes, also called an unbranched
unidirectional model6, which yields Evolution Associated Difference Spectra (EADS),
and (3) a full compartmental scheme which may include possible branchings and
equilibria, which yields Species Associated Difference Spectra (SADS). The last is most
often referred to as target analysis, where the target is the proposed kinetic scheme,
including possible spectral assumptions. In this paper we did not attempt a target
analysis. Instead, throughout the manuscript, the EADS are shown in the main text and
the corresponding DADS are shown in the Supporting Information.
(1) With parallel decaying components the model reads
ncomp
ψ (t , λ ) =
∑ c (k ) DADS (λ )
I
l
l
l =1
The DADS thus represent the estimated amplitudes of the above defined exponential
S.2
decays c I (kl ) . When the system consists of parallel decaying components the DADS are
true species difference spectra. In all other cases, they are interpreted as a weighted sum
(with both positive and negative contributions) of true species difference spectra.
(2)
A sequential model reads
ncomp
ψ (t , λ ) =
∑c
II
l
EADSl (λ )
l =1
where each concentration is a linear combination of the exponential decays,
l
clII = ∑ b jl c I (kl ) , and the amplitudes6 b jl are given by b11 = 1 and for j ≤ l :
j =1
l −1
b jl = ∏ km /
m =1
l
∏
( kn − k j )
n =1, n ≠ j
When the system consists of sequentially decaying components 1 → 2 → ... → ncomp the
EADS are true species difference spectra. In all other cases, they are interpreted as a
weighted sum (with only positive contributions) of true species difference spectra.
Equivalence of the parallel and the sequential model
It is important to note that the fit is identical when using a parallel or a sequential model.
Both the estimated lifetimes and the residuals from the fit are identical. This can be
demonstrated as follows. Since the concentrations of the sequential model are a linear
combination of the exponential decays we can write for the matrix of concentrations
(where column l corresponds to component l)
C II = C I B
where the upper triangular matrix B contains the elements b jl defined above.
Furthermore, in matrix notation the parallel model reads
Ψ = C I DADS T
where DADS T is the transpose of the matrix that contains the DADS of component l in
column l. Likewise, in matrix notation the sequential model reads
Ψ = C II EADS T
Combining these three equations we obtain the relations
DADS = EADS ⋅ BT
S.3
EADS = DADS ⋅ B −T
where the coefficients of the lower triangular matrix B −T are given by b1−l 1 = 1 = b1−l T and
for j ≤ l :
j −1
b −jl1 = ∏
n =1
( k n − kl )
= blj−T
kn
Thus the DADS are linear combinations of the EADS, and vice versa. Thus, the lth
EADS is a linear combination of the lth and following DADS. In particular, the first
EADS, which corresponds to the time zero difference spectrum, is the sum of all DADS;
and the final EADS is proportional to the final DADS.
In systems where photophysical and photochemical processes occur the sequential model
with increasing lifetimes provides a convenient way to visualize the evolution of the
(excited and intermediate) states of the system. Therefore, the EADS are shown in the
main text and the corresponding DADS are shown in the Supporting Information.
References
(1)
Holzwarth, A. R. Data Analysis of Time-Resolved Measurements. In
Biophysical Techniques in Photosynthesis; Amesz, J., Hoff, A. J., Eds.; Kluwer:
Dordrecht, The Netherlands, 1996; pp 75.
(2)
van Stokkum, I. H. M.; Larsen, D. S.; van Grondelle, R. Biochimica Et
Biophysica Acta-Bioenergetics 2004, 1657, 82.
(3)
van Stokkum, I. H. M.; van Oort, B.; van Mourik, F.; Gobets, B.; van
Amerongen, H. (Sub)-Picosecond Spectral Evolution of Fluorescence Studied with a
Synchroscan Streak-Camera System and Target Analysis. In Biophysical Techniques in
Photosynthesis Vol. II; Aartsma, T. J., Matysik, J., Eds.; Springer: Dordrecht, The
Netherlands, 2008; pp 223.
(4)
van Stokkum, I. H. M. “Global and Target Analysis of Time-resolved
Spectra,” Department of Physics and Astronomy, Faculty of Sciences, Vrije Universiteit,
Amsterdam, The Netherlands, 2005.
(5)
van Stokkum, I. H. M.; Bal, H. E. “A Problem Solving Environment for
interactive modelling of multiway data”, 2006.
(6)
Nagle, J. F.; Parodi, L. A.; Lozier, R. H. Biophysical Journal 1982, 38,
161.
S.4