Multiple active site residues are important for photochemical efficiency in the light-activated enzyme protochlorophyllide oxidoreductase (POR)

Abstract

Protochlorophyllide oxidoreductase (POR) catalyzes the light-driven reduction of protochlorophyllide (Pchlide), an essential, regulatory step in chlorophyll biosynthesis. The unique requirement of the enzyme for light has provided the opportunity to investigate how light energy can be harnessed to power biological catalysis and enzyme dynamics. Excited state interactions between the Pchlide molecule and the protein are known to drive the subsequent reaction chemistry. However, the structural features of POR and active site residues that are important for photochemistry and catalysis are currently unknown, because there is no crystal structure for POR. Here, we have used static and time-resolved spectroscopic measurements of a number of active site variants to study the role of a number of residues, which are located in the proposed NADPH/Pchlide binding site based on previous homology models, in the reaction mechanism of POR. Our findings, which are interpreted in the context of a new improved structural model, have identified several residues that are predicted to interact with the coenzyme or substrate. Several of the POR variants have a profound effect on the photochemistry, suggesting that multiple residues are important in stabilizing the excited state required for catalysis. Our work offers insight into how the POR active site geometry is finely tuned by multiple active site residues to support enzyme-mediated photochemistry and reduction of Pchlide, both of which are crucial to the existence of life on Earth.

Full text

Multiple active site residues are important for photochemical
efficiency in the light-activated enzyme protochlorophyllide
oxidoreductase (POR)☆
Binuraj R.K. Menon, Samantha J.O. Hardman, Nigel S. Scrutton ⁎,DerrenJ.Heyes⁎
a
Centre for Synthetic Biology of Fine and Speciality Chemicals, Manchester Institute of Biotechnology, School of Chemistry, The University of Manchester, Manchester, M1 7DN, UK
abstractarticle info
Article history:
Received 6 April 2016
Received in revised form 17 May 2016
Accepted 30 May 2016
Available online 01 June2016
Protochlorophyllide oxidoreductase (POR) catalyzes the light-driven reduction of protochlorophyllide (Pchlide),
an essential, regulatory step in chlorophyll biosynthesis.The unique requirementof the enzyme for light has pro-
vided the opportunityto investigate how light energy can be harnessed to power biologicalcatalysis and enzyme
dynamics. Excited state interactions between the Pchlide molecule and the protein are known to drive the sub-
sequent reaction chemistry. However, the structural features of POR and active site residues that are important
for photochemistry and catalysis are currently unknown, because there is no crystal structure for POR. Here,
we have used static and time-resolved spectroscopic measurements of a number of active site variants to
study the role of a number of residues, which are located in the proposed NADPH/Pchlide binding site based
on previous homology models, in thereaction mechanism of POR. Our findings, which are interpreted in the con-
text of a new improved structural model, have identified several residues that are predicted to interact with the
coenzymeor substrate. Severalof the POR variants have a profound effecton the photochemistry, suggesting that
multiple residues are important in stabilizing the excited state required for catalysis. Our work offers insight into
how the POR active site geometry is finely tuned by multiple active site residues to support enzyme-mediated
photochemistry and reduction of Pchlide, both of which are crucial to the existence of life on Earth.
© 2016 The Authors. Published by Elsevier B.V. Thisis an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
Keywords:
Protochlorophyllide oxidoreductase (POR)
Light activation
Photochemistry
Site-directed mutagenesis
Enzyme catalysis
1. Introduction
The reduction of protochlorophyllide (Pchlide) to chlorophyllide,
catalyzed by the enzyme protochlorophyllide oxidoreductase (POR), is
the key light-driven reaction that leads to a profound transformation
in plant development [1–5]. In non-flowering plants, algae and
cyanobacteria there is also a light-independent Pchlide reductase,
consisting of 3 separate subunits, whichcan catalyzethis same penulti-
mate step in the chlorophyll biosynthetic pathway [6,7]. In light-depen-
dent POR the reaction proceeds via a highly endergonic light-driven
hydride transfer from the NADPH coenzyme to the C17 position of the
Pchlide molecule followed by an exergonic thermally activated proton
transfer from a conserved Tyr residue to the C18 position of Pchlide
(Fig. 1A) [8–12]. Recent advances in our understanding of the POR
reaction mechanism that illustrate POR is an important model for
studying light-activated enzyme dynamics and how light energy can
be harnessed to power biological catalysis [13–18].
PORs originating froma variety of photosynthetic organisms, includ-
ing cyanobacteria and higher plants, have been studied in detail using
ultrafast and cryogenic spectroscopic techniques [8,16,18–21]. Excited
state interactionsbetween the Pchlide molecule and active site residues
in the enzyme are proposed to result in a reactive charge-separated
state that facilitates the sequential hydride and proton transfer reac-
tions on a microsecond timescale [7,15,18,22]. The light-driven hydride
transfer step is coupled to enzyme motions within the lifetime of the
Pchlide excited state and is followed by the proton transfer step,
which is reliant on solvent dynamics and an extended network of mo-
lecular motions [22]. The catalytic cycle of POR also comprises a series
of conformational changes that are associated with ordered substrate
binding and product release steps to form a reactive active site confor-
mation [7,23,24]. However, despite a detailed molecular understanding
of the catalytic cycle from femtoseconds to seconds the structural fea-
tures of POR, and active site residues that contribute to light activation
and the reaction dynamics, is currently unknown due to the lack of a
crystal structure.
Journal of Photochemistry & Photobiology, B: Biology 161 (2016) 236–243
☆Funding source statement: NSS is an Engineering and Physical Sciences Research
Council (EPSRC) Established Career Fellow (EP/J020192/1). Ultrafast spectroscopy was
performed at the Ultrafast Biophysics Facility, Manchester Institute of Biotechnology, as
funded by BBSRC Alert14 Award BB/M011658/1. The work is a contribution from the
EPSRC/BBSRC Synthetic Biology Research Centre SYNBIOCHEM (BB/M017702/1), which
provided infrastructure funding.
⁎Corresponding authors at: Manchester Institute of Biotechnology, The University of
Manchester, Manchester, M1 7DN, UK.
E-mail addresses: nigel.scrutton@manchester.ac.uk (N.S. Scrutton),
Derren.Heyes@manchester.ac.uk (D.J. Heyes).
http://dx.doi.org/10.1016/j.jphotobiol.2016.05.029
1011-1344/© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Contents lists available at ScienceDirect
Journal of Photochemistry & Photobiology, B: Biology
journal homepage: www.elsevier.com/locate/jphotobiol

Sequence similarities with other short-chain dehydrogenase/reduc-
tase (SDR) enzymes identifies POR as a member of the classical SDR
family of enzymes. The conserved glycine rich-Rossmann dinucleo-
tide-binding domain (GxxxGxG) and the catalytic YxxxK motifs of
POR have suggested a similar catalytic mechanism in POR and other
SDR enzymes [25–27]. Consequently, although the overall sequence ho-
mology and pair wise identity between PORand any individual SDR en-
zyme is below the requirement to select an empirical structural
template, different homology models of POR have been proposed by
using closely related SDR enzymes as the structural template [5,27–
30]. These homology models were composed of 7 β-sheets surrounded
by 9 α-helices (Fig. 1B). The extensions to some of the central helices
provide the Pchlide binding cleft, where the YxxxK catalytic motif was
included as part of an α-helix that results in residues Tyr193 and
Lys197 interacting with the Pchlide molecule. A unique 33 amino acid
residue insertion in POR was included as a loop region, as no corre-
sponding structural regions were present in the template structures,
and was predicted to beinvolved inestablishing an oligomerisation do-
main for POR inbarley [27,31].Basedonthesepreviousmodelswehave
now selected a number of active site variants to study the role of resi-
dues located in the proposed NADPH and Pchlide bindingsites in the re-
action mechanism of POR (Fig. 2). Detailed binding, multiple turnover
and single turnover laser studies have shown that although many of
the variants have reduced catalytic activity they all retain the ability to
bind substrates, albeit with reduced affinity in some cases. Several of
the active site variants show impaired photochemical behaviour, sug-
gesting that multiple residues are likely to be involved in the formation
of the excited state ‘reactive’charge transfer state that is required prior
to hydride transfer. In the ongoing absence of a crystal structure we
have rationalized these findings by using multiple SDR enzymes as a
structural template to produce an improved homology model of POR
from Thermosynechocystis elongatus.
2. Materials and Methods
2.1. Homology Modelling
The homology model of POR from Synechocystis sp. [27] was used as
the initial template for the development of a structural model of
Thermosynechococcus elongatus (BP-1) POR, using SWISS-MODEL ho-
mology server [32]. The NADPH coenzyme and Pchlide substrate were
docked in to the model with AutoDock and the ligand-docked structures
were relaxed by means of MD simulations by using the Amber ff94 force
field implemented inGromacs 4.6.5. Prior to this, the base structure was
corrected and topology files for Gromacs were generated by using the
online tools from MDweb [33]. The quality of developed homology
model was analyzed using QMEAN server [34]. After optimisation the
final model showed a QMEAN score of 0.50 and Z score of 2.95. In addi-
tion, the SWISS-MODEL homology server was used to determine the 15
SDR enzymes of known structures in the PDB that share the highest se-
quence identity with T. elongatus POR (Tables S1 and S2, Fig. S1). Mus-
tang-MR structural sieving server was used for the multiple structural
alignments of these SDR enzymes with POR and to remove residues
from the alignment that are below a threshold root mean squaredevia-
tion (RMSD) of 1.8 Å (Fig. S2) [35]. A putative uncharacterized SDR pro-
tein (PDB id: 3RD5) that has the highest sequence identity with POR
was used as the reference structure and thesieving procedure removed
68% of the 3RD5 residues at 1.8 Å sieving level using Mustang-MR siev-
ing server. The unsieved structural regions of 3RD5 were mapped on to
the homology model of T. elongatus POR by structural alignment to de-
termine the regions that are conserved between SDR enzymes and
POR and will therefore be modeled most accurately.
2.2. Sample Preparation and Site Directed Mutagenesis
All chemicals were obtained from Sigma-Aldrich. Recombinant POR
from the thermophilic cyanobacteria T. elongatus BP-1 was
overexpressed in Escherichia coli and purified as described previously
[36]. Site-directed mutagenesis of the por gene was performed using
the QuikChange kit (Stratagene) to mutate residues. The primers that
were used (MWG Eurofins, Germany) are shown in the supplementary
information (Table S3). The correct mutations were confirmed by DNA
sequencing (MWG Eurofins, Germany) and variant protein waspurified
as described previously [36]. The Pchlide pigment was produced and
purified as described previously [37].
2.3. Steady-state Activity and Substrate Binding Measurements
Steady-state activity measurements were carried out as described
previously using a Cary 50 spectrophotometer (Agilent Technologies)
[38]. The binding of NADPH coenzyme was monitored using fluores-
cence energy resonance transfer in a Cary Eclipse fluorimeter (Agilent
Technologies) [38] and the binding of Pchlide was measured by follow-
ing the red-shift in absorbance at 642 nm, essentially as described [24].
2.4. Laser Flash Photolysis
Absorption transients at 696 nm were measured at 298 K using an
LKS-60 flash photolysis instrument (Applied Photophysics Ltd.) with
the detection system (comprising probe light, first monochromator,
sample, second monochromator and photomultiplier) at right angles
to the incident laser beam. Dark assembled enzyme–NADPH–Pchlide
ternary complex samples were excited in a cuvette of 1-cm pathlength
as described previously [10] with a 6 ns laser pulse at450 nm (~30 mJ)
Fig. 1. The light-activated reduction of Pchlide catalyzed by POR. A, trans addition of
hydrogen across the C
17
_C
18
carbon double bond of Pchlide to form Chlide in the
chlorophyll biosynthesis pathway is catalyzed by protochlorophyllide oxidoreductase
(POR). B, the three-dimensional homology model of POR from Synechocystis, based on
the crystal structure of 7α-hydroxysteriod dehydrogenase as a structural template [27].
237B.R.K. Menon et al. / Journal of Photochemistry & Photobiology, B: Biology 161 (2016) 236–243

using an OPO of a Q-switched Nd-YAG laser (Brilliant B, Quantel). The
reaction samples were prepared by taking 15 μM Pchlide and 250 μM
NADPH in the presence of 60.0 μM enzyme for wild-type POR. Based
on the dissociation constant for Pchlide an equal concentration of terna-
ry complex was used for the variant enzymes. Hence, the reaction sam-
ples (1 ml) were prepared by using 60 μM enzyme concentrations for
R38V, K42A, T147S and S189A T230S and 55 μM for S16C, G19A and
T230S. Higher concentrations of enzymes were used for N39V
(128 μM), N90A (106 μM), Y94F (78 μM), T145 A (160 μM), T230 A
(78 μM), T230F (175 μM), T147F (74 μM), N149 V (100 μM) and
H236A (71 μM) variant enzymes. Rate constants were measured from
the average of at least five time dependent absorption measurements
by fitting to a single exponential function.
2.5. Ultrafast Pump-Probe Transient Absorption Spectroscopy
A Ti:sapphire amplifier system (Spectra Physics Solstice Ace) pro-
duced 6 mJ of800 nm pulsesat 1 kHz with 100 fs pulse duration.A por-
tion of the output of the amplifier was used to pump a Topas Prime OPA
with associated NirUVis unit which was used to generate the pump
Fig. 2. The sequence alignment and conservation of amino acid among homologous POR enzymes. The protein sequence of T. elongatus POR is aligned with different homologous POR
enzymes. The amino acid residues that were targeted in this study are indicated by arrows. The conserved regions are indicated by black lines. PRALINE; protein multiple sequence
alignment web server tool (by Centre for Integrative Bioinformatics, Vrije Universiteit Amsterdam, Netherlands) was used for the alignment of protein sequences.
238 B.R.K. Menon et al. / Journal of Photochemistry & Photobiology, B: Biology 161 (2016) 236–243

beam centred at 450 nm, with FWHM of ca. 10 nm. A broad band ultra-
fast pump-probe transient absorbance spectrometer ‘Helios’(Ultrafast
systems LLC) was used to collect data (at random time points) from
~1 ps to 2.6 ns with a time resolution of around 0.2 ps. The probe
beam consisted of a white light continuum generated in a sapphire crys-
tal and absorbance changes were monitoredbetween 475 and 725 nm.
A broad band sub-nanosecond pump-probe transient absorbance spec-
trometer ‘Eos’(Ultrafast systems LLC) was used to collect data (at ran-
dom time points) up to 2 μs with a time resolution of around 0.5 ns. A
2 kHz white-light continuum fibre laser was used to generate the
probe pulses and the delay between pump and probe was controlled
electronically. For both sets of measurements samples were excited at
450 nm with 0.5 μJ power and a beam diameter of ~200 μm. Samples
(1.5 ml) were flowed at a rate of approximately 30 ml/min through a
0.2 mm pathlength quartz cell (at room temperature) to ensure that a
different area of the sample is excited with each pump laser pulse. Sam-
ples were measured upon until theproportion of Chlide in the reaction
mixture became higher than 10%, which resulted in data collection
times of b5 min for the wild-type and in the range of 8–20 min for the
variants. Samples were prepared in the dark containing 500 μM POR,
200 μM Pchlide and 4 mM NADPH in activity buffer with 10% glycerol,
0.1% 2-mercaptoethanol, and 0.5% Triton X-100.
2.6. Global Analysis
The datasets from the Helios and Eos measurements were merged
by selecting data with the same (small) proportion of product present
and scaling the Eos dataset by a fixed factor to match the intensity of
the ground state bleach feature to that in the Helios data. The data
were then analyzed globally using the open-source software Glotaran
[39]. This procedure reduces the matrix of change in absorbance as a
function of time and wavelength, to a model of one or more exponen-
tially decaying time components, as described in the main manuscript
[39], each with a corresponding difference spectrum (speciesassociated
difference spectra (SADS)). Errors quoted with the lifetime values are
the standard errors calculated during the global analysis. The lifetimes
quoted for the conversion between states also include contributions
from the rates of ground state recovery through both radiative and
non-radiative processes. For the analysis, the pre-excitationbackground
was subtracted, and Helios datasets corrected for spectral chirp, and the
datasets were fitted to a simple sequential model where one species
converts to another, which then persists for the lifetime of the
experiment.
3. Results
3.1. A New Structural Model for POR
As there is currently no crystal structure available for the POR en-
zyme further insights into therole of putative active site residues in ca-
talysis has been gained by producing a moreaccurate homology model
for the enzyme. This has been achieved by selecting multiple templates
of known PDB structures with a high sequence similarity with POR,
rather than a single template, and by identifying the core regions con-
served across all of the SDR enzymes (Figs. S1 and S2). A total of 15
SDR enzymes were selected as the targets of sequence and structure
alignment with POR, where the identity of individual templates with
POR varied between 19 and 30% (Table S2, Fig. S1)[32]. By using a struc-
tural sieving server, the highly conserved regions were mapped to the
homology model to optimise the POR structure (Fig. 3). The overall
structural model of T. elongatus POR is similar to previous models (Fig.
3A–C) [5,27–30] with the highly conserved Rossmann-fold region, sur-
rounding β-sheets and α-helices and the catalytic tetrad residues, to-
gether with two insertions (residues 232 to 253 and residues 151 to
186) that are not present in other SDR enzymes and are therefore struc-
turally less well-defined. Consequently, this refined model of POR
provides important insights about the potential role of individual resi-
dues in catalysis. Many of the residues selected for mutagenesis are in
close enough proximity to interact with the NADPH coenzyme. Ser16
and Gly19 are part of the highly conserved GxxxGxG motif, whereas
Arg38, Asn39 and Lys42 are located close to the 2′-phosphate of the
adenosine ribose group and Asn90 is located near to the adenine moiety
of NADPH (Figs. 3D, 4A). Other residues, such as Tyr94, Thr145, Thr147,
Asn149, Ser189, Thr230 and His236 are also positioned in theactive site,
close to both Pchlide and NADPH (Figs. 3D, 4A, B). Tyr94, Thr145,
Thr147 and Asn149 are all located in close proximity to the propionate
side-chain of the Pchlide molecule at the C17 position, whereas Thr230
is located at a key position near to the diphosphate chain joining the
adenosine and the nicotinamide ribose groups of NADPH. However, al-
though theSer189 and His236 residues arepositioned relatively nearto
the Pchlide molecule and have previously been implicated in Pchlide
binding from previous homology models [27], the new structural
model shows that this is unlikely as their distance from Pchlide is too
great to allow any direct interactions with the pigment (Fig. 4B).
3.2. Steady-state Characterization of Active Site POR Variants
The new structural model of POR, in addition to previous structural
homology models [5,27–30], have highlighted a number of residues
that may be important for catalytic activity. These residues, which can
potentially interact with the NADPH coenzyme [40,41], the Pchlide sub-
strate or are in close proximity to the binding site of both substrates [5,
27–30], were altered by site-directed mutagenesis and their catalytic
activity initiallydetermined under steady-state conditions (Table 1). Al-
though many of the variants retained close to wild-type levels of activ-
ity, several showed a significant reduction in catalytic activity. In
particular, mutations to Asn39, Asn90, Thr145, Thr147 and Asn149
had a major impact on steady-state activity, whereas replacement of
the Thr230 residues with a bulkier Phe residue showed a marked de-
crease in activity, in contrast to the limited effects caused by replace-
ment with Ala or Ser.
In order to provide a more in-depthrationale for this variation in cat-
alytic activities, the ability of all variant enzymes to bind the coenzyme
NADPH and Pchlide was measured (Table 1). The dissociation constant
for NADPH was determined by measuring the FRET signal from Trp res-
idue(s) in the protein tothe bound NADPH coenzyme [38]. Several var-
iants, including S16C, G19A, R38V, K42A, T147F, T230A andT230F, had a
significantly reduced affinity for NADPH compared to the wild-type en-
zyme, indicating a role for the targeted residues in coenzyme binding.
The binding of Pchlide to the enzyme–NADPH complex has been mon-
itored by measuring the red-shift in the absorbance maximum of the
Pchlide molecule upon forming a ternary POR–Pchlide–NADPH ternary
complex (Table 1)[24].Most of the variant POR enzymes exhibited only
minor changes in the Pchlide binding properties. However, exceptions
to this were the N39V, N90A, T145A, N149V and T230F variants,
which showed ~ 5–10 fold reduction in the affinity of the Pchlide
substrate.
3.3. Laser Excitation Measurements of the Site-directed Mutants
Laser photoexcitation measurements have been used to obtain in-
formation on the excited state and single-turnover kinetics of interme-
diate formation for wild-type POR and the variant enzymes. The
formation of a broad absorbance band at 696 nm represents the hydride
transfer chemistry from NADPH and the disappearance of this 696 nm
band represents the proton transfer step to the C18 position to form
the final Chlide product [10,11]. Hence, the kinetics and relative yield
of formation/decay of the absorbance band at 696 was measured in
dark-assembled enzyme-substrate ternary complex samples after exci-
tation with a 6 ns laser pulse. For those variants whose reaction could be
accurately measured there were only minor differences in the rates of
hydride and proton transfer (Table 1). More significantly, many of the
239B.R.K. Menon et al. / Journal of Photochemistry & Photobiology, B: Biology 161 (2016) 236–243

mutations resulted in a major loss of the amplitude of the absorbance
signal at 696 nm upon photoexcitation (Table 1), suggesting that the
quantum yield or photochemical efficiency of the hydride transfer
step is impaired in those variants (example transients are shown in
Fig. 3). In some cases, notably the N149V, T145A, T147S and T230F var-
iants, the photochemical efficiency was so low (Table 1) that it was not
possible to measure accurate rates for the hydride and proton transfer
steps, indicating a potentially important role for these residues in pho-
tochemistry. Indeed, replacement of the Thr230 residue with a bulkier
Phe residue in the T230F variant reduced the photochemical efficiency
significantly compared to the T230A and T230S variants (Fig. 5).
To further explore this potential role in photochemistry the excited
state dynamics were investigated by pump-probe transient absorption
spectroscopy for 5 POR variants that showed an impaired photochemi-
cal efficiency and compared to those obtained for Pchlide only and a
wild-type POR–Pchlide–NADPH ternary complex. The time-resolved
absorption difference spectra from these measurements were then
modeled using global analysis to yield species associated difference
spectra (SADS). For clarity only the SADS are shown in the main manu-
script, whereas the raw time-resolved data (Figs. S3–S9), and kinetic
traces showing fits at selected probe wavelengths (Fig. S10) can be
found in the supporting information. In the wild-type POR ternary
Fig. 3. Structural models of PORfrom Synechocystis (A) [27],Arabidopis thaliana (B) [30] and from T. elongatus (C) described in the present work are shown for comparison. D, a close-up
view of the active site illustrating the relative positions of the active site residues characterized in this study, which are shown in cyan, and the active site Tyrand Lys residues, which are
shown in red as sticks. The proteinbackbone is representedas a ribbon, NADPH is shown in yellowand Pchlide is shown in green. In Fig.A, B and C, protein backboneis coloured in gray to
indicate regions in the model which are not conserved.
Fig. 4. A close-up view ofthe NADPH (A) and Pchlide (B) binding site,showing potentialactive site aminoacid residues thatcould interactwith the NADH or Pchlidemolecule basedon the
derived structural model of T. elongatus POR.
240 B.R.K. Menon et al. / Journal of Photochemistry & Photobiology, B: Biology 161 (2016) 236–243

enzyme-substrate complex the observed transient spectral changes can
be fitted to a branched model along ‘catalytic’and ‘non-catalytic’path-
ways after the formation of an intramolecular charge transfer state
(S
ICT
) as described previously [18]. In this model the formation of the
hydride transfer intermediate, with the characteristic absorbance band
at 696 nm, is observed with a lifetime of approximately 500 ns (Fig.
Table 1
The steady-state kinetic parameters, coenzyme and substrate binding constants and rates of hydrideand proton transfer for wild-type POR and variant enzymes. All rates and binding
constantswere measured at 25 °C. The ratesof hydride and proton transfer were measured from the average of at least five timedependent absorption measurements by laser excitation
of similar levels of ternary enzyme-substrate complex and by following absorbance changes at 696 nm. The amplitude of the absorbance change at 696 nm isalsoshown.
Enzyme
k
cat
(s
−1
)
K
d
NADPH
(nM)
K
d
Pchlide
(μM)
k
hydride
(×10
6
s
−1
)
k
proton
(×10
4
s
−1
)ΔmAbs 696 nm
Wild-type 0.17 ± 0.002 21 ± 1 5.6 ± 0.6 2.21 ± 0.06 2.72 ± 0.04 102 ± 1
S16C 0.16 ± 0.001 325 ± 90 3.5 ± 0.5 2.11 ± 0.03 2.68 ± 0.03 61 ± 3
G19A 0.16 ± 0.001 182 ± 31 3.5 ± 0.4 2.08 ± 0.07 2.67 ± 0.02 92 ± 2
R38V 0.16 ± 0.001 411 ± 76 5.4 ± 0.6 2.12 ± 0.03 2.69 ± 0.01 54 ± 4
N39V 0.02 ± 0.001 23 ± 6 28.9 ± 1.4 1.98 ± 0.02 2.71 ± 0.03 10 ± 0.4
K42A 0.16 ± 0.001 230 ± 81 4.5 ± 0.4 2.14 ± 0.09 2.67 ± 0.04 85 ± 3
N90A 0.07 ± 0.002 94 ± 7 21.4 ± 0.6 2.28 ± 0.04 3.17 ± 0.04 31 ± 2
Y94F 0.17 ± 0.003 63 ± 4 11.8 ± 0.7 2.23 ± 0.02 3.11 ± 0.01 80 ± 1
T145A 0.01 ± 0.001 33 ± 2 39.7 ± 2.1 n. d. n. d. 2 ± 0.2
T147S 0.02 ± 0.003 60 ± 3 6.5 ± 0.7 n. d. n. d. 2 ± 0.2
T147F 0.01 ± 0.003 117 ± 3 10.3 ± 0.6 n. d. n. d. 1 ± 0.1
N149V 0.01 ± 0.003 55 ± 2 19.2 ± 1.2 n. d. n. d. 2 ± 0.2
S189A 0.16 ± 0.001 38 ± 16 4.9 ± 0.4 2.15 ± 0.07 2.69 ± 0.01 69 ± 4
T230A 0.16 ± 0.001 181 ± 16 11.9 ± 1.4 1.63 ± 0.07 2.26 ± 0.03 51 ± 3
T230S 0.16 ± 0.001 48 ± 17 4.6 ± 0.5 1.71 ± 0.06 2.38 ± 0.06 60 ± 1
T230F 0.01 ± 0.001 436 ± 87 44.9 ± 3.5 n. d. n. d. 2 ± 0.2
H236A 0.10 ± 0.004 30 ± 11 9.4 ± 0.6 2.18 ± 0.03 2.72 ± 0.01 33 ± 1
Fig. 5. Hydride and proton transfer transients of wild-type and Thr230 variant POR
enzymes. Typical laser transients showing the absorbance change at 696 nm for the
wild-type POR and the Thr230 variants. The formation of the absorbance band at
696 nm represents hydride transfer and the disappearance of the 696 nm band
represents proton transfer [15]. The transients were measured using an equal ternary
complex concentration based on the Pchlide binding constant for each enzyme. The
experimental procedure is described in the materials and method section.
Fig. 6. Species associated difference spectra (SADS) resulting from a global analysis of the
time-resolved visible data for wild type and variant POR–Pchlide–NADPH ternary
complexes after excitation at 450 nm. The data for wild type (A) and N39V (B) were
fitted as described in the Supporting Information to a branched model (shown above
the panel), where 60% of the ICT state is converted to the solvated ICT state along the
‘non-catalytic’pathway and 40% is converted to a ‘reactive’ICT state along the ‘catalytic’
pathway [18]. The data for T145A (C), T147S (D), N149V (E) and T230F (F) were fitted
to the sequential model (shown above the panel) as described in Supporting
Information. Kinetic traces showing fits at selected energies are shown in Fig. S9.
241B.R.K. Menon et al. / Journal of Photochemistry & Photobiology, B: Biology 161 (2016) 236–243

6A), similar to that obtained in the earlier laser flash photolysis mea-
surements. The excited state dynamics observed for the N39V variant
could also be fitted to the same branched model with similar spectral
features and lifetimes (Fig. 6B), suggesting that photochemistry pro-
ceeds in an identical way to the wild type enzyme. In contrast, the hy-
dride transfer intermediate is completely absent in the N149V, T145A,
T147S and T230F variants (Fig. 6C-F), confirming that photochemistry
is impaired in these variants. In all of these cases the SADS required to
model the transient spectral data appear to be very similar to free
Pchlide (Fig. S10) and could be fitted to 3 sequentially evolving expo-
nential functions. This represents a simple, linear decay pathway for
the excited state dynamics in these variants, involving the solvation of
the ICT state, followed by decay of the S1/ICT excited state into a long-
lived triplet state on the ns timescale. The triplet state then relaxes
back to the ground state on the μstimescale[18].
4. Discussion
It is now known that light-induced interactions between the protein
and the pigment occur within the lifetime of the Pchlide excited state
and are essential for the subsequent chemical steps, involving sequen-
tial hydride and proton transfer reactions, on themicrosecond timescale
[8,15,18,21,22]. However, the lack of any crystal structure for POR has
made it difficult to understand the structural elements and active site
residues that are required for this unique light-driven catalysis. Conse-
quently, a number of residues located in the proposed substrate binding
sites of POR have now been altered and their role in the reaction mech-
anism of POR investigated by static and time-resolved spectroscopic
measurements. Moreover, the results have been rationalized in terms
of active site structure based on a new structural model of the enzyme.
The majority of SDR enzymes contain a highlyconserved nucleotide-
binding domain, termed the Rossmann fold [40–44], which is character-
ized by a TGxxxGxG motif (TGASSGVG in T. elongatus POR). Mutations
to Ser16 and Gly19 in this region were found to result in a lower affinity
for NADPH, although the catalytic activity remained unaffected. Similar-
ly, changes to Arg38 and Lys42 also reduced NADPH affinity and the
close proximity of these residues to the 2′-phosphate and hydroxyl
groups of the adenosine ribose of the coenzyme in the structural
model support their important role in defining the coenzyme specificity
in the SDR family [45]. In addition, Thr230 may also be important for the
optimal positioning of the NADPH coenzyme in the active site as it ap-
pears to be located at a key position near to the diphosphate chain join-
ing the adenosine and the nicotinamide ribose groups. Although the
affinity for NADPH was reduced for both the T230 A and T230S variants
the effect was much greater when a bulkier Phe residue was placed at
this position.
In terms of Pchlide binding Asn39, Asn90, Thr145 and Asn149 ap-
pear to be important as mutations to these residues reduce both the cat-
alytic efficiency of the enzyme and the affinity of the substrate. In the
majorityof SDR enzymes there is a catalytic tetrad of active site residues
that are essential for activity, consisting of Asn-Ser-Tyr-Lys [44,46].It
has been suggested that the Ser stabilizes the substrate and the Tyr
acts as a proton donor, whereas the Lys lowers the pKa of the Tyr-OH
to promote proton transfer and the Asn residue provides important in-
teractions with theLys to maintain its position inthe active site and pro-
mote the proton relay mechanism [44,46].The importance of the active
site Tyr andLys residues in POR catalysis have already been reported in
previous studies [11]. However, in POR from T. elongatus the Ser residue
from the catalytic tetrad is replaced by a Thr (Thr145), which is located
in close proximity to the catalytic Tyr and Lys and also to the NADPH co-
enzyme in the structural model. Hence, changes to this residue may sig-
nificantly influence any interactions between the Tyr, Lys, NADPH and
Pchlide substrate. The Asn from the catalytic tetrad is Asn90 in POR,
which is likely to play a similar role to other SDR enzymes by stabilizing
the active site geometry through interactions with the ribose hydroxyl
group of the nicotinamide and the active site Lys [44,46]. However,
the present work has revealed that changes to two other residues,
Ser189 and His236, which had been suggested to be important for
Pchlide binding based on previous homology models [27],onlyhavea
minimal effect on substrate binding and catalytic activity. Although
the new structural model of POR described in the present work is simi-
lar to previous homology models [5,27–30], the Pchlide binding sitedif-
fers significantly and indicates that the proposed role of His236 in
chelating the central Mg ion of Pchlide is unlikely as there is no possibil-
ity for direct interaction with Mg
2+
. In addition,the previous suggestion
that Ser189 may be important in the correct positioning of the Pchlide
towards the pro-S face of NADPH [27] is also unlikely as the S189 A var-
iant retained wild-type levels of substrate binding and catalytic activity.
The crystal structure of the unrelated light-independent Pchlide reduc-
tase shows that thereis no requirement forthe Pchlide substrate to have
adirectMg
2+
coordination with amino acid residues in the active site
[6,7]. Hence, based on the new model, His236 together with other aro-
matic residues such as Phe233 and Phe237, may be involved in Pchlide
binding byproviding a pi stacking interaction with the substrate (Fig. 4).
However, His236 is replaced by serine in some homologous POR en-
zymes, which indicates that this residue is not an absolute necessity.
Moreover, Pro96 and Leu232 are close to the Mg
2+
in the new model
and it is possible that these residues are also involved either in a direct
or indirect coordination with the central metal ion (Fig. 4).
Importantly, the present work highlights how minor changes to the
architecture of the active site can have a profound effect on the efficien-
cy of photochemistry. Previous studies have shown that the photo-
chemistry is driven by excited state interactions between the active
site Tyr and the carboxyl group of the propionate side chain at the C17
position of Pchlide, which pulls electron density away from the C17–
C18 double bond to form a ‘reactive’charge transfer state [18].Thiscre-
ates an electron-deficient site across the double bond, which triggers
the subsequent transfer of the negatively charged hydride from
NADPH [18]. Several of the active site variants, including T145A,
T147S, T147F, N149V and T230F exhibit impaired photochemical be-
haviour,implying a role for theseresidues inthe formation of the excit-
ed state ‘reactive’charge transfer state. Based on the structural model
the Thr147 and Asn149 residues are in close enough proximity to inter-
act directly with the propionate side-chain at the C17 position of
Pchlide. As this region of the Pchlide molecule is essential for photo-
chemistry [18] it is likely that both of these residues are essential for
the excited state interactions that create the ‘reactive’charge transfer
state. As discussed above, the close proximity of Thr145 to the catalytic
Tyr and Lys and the NADPH coenzyme may mean that any changes to
this residue can significantly affect any excited state interactions be-
tween the Tyr, Lys, NADPH and Pchlide substrate. Although the
Thr230 is unlikely to directly participate in any excited state interac-
tions it maystill play an important role in maintaining the active site ge-
ometry to allow photochemistry to proceed efficiently as replacement
with a bulkier Phe residue abolishes the photochemical step.
Concluding remarks
We have now highlighted a number of key residues in the active site
of POR that are important for coenzyme and substrate binding, as well
as the excited state processes required for catalysis. The role of these
residues have been verified by a new structural model for T. elongatus
POR, which supports all of the findings from the binding, multiple turn-
over and single turnover laser studies on the active site variants. In the
absence of any crystal structure this work will provide the basis for fu-
ture functional studies of this key light-activated enzyme.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jphotobiol.2016.05.029.
242 B.R.K. Menon et al. / Journal of Photochemistry & Photobiology, B: Biology 161 (2016) 236–243

References
[1] R.D. Willows, Biosynthesis of chlorophylls from protoporphyrin IX, Nat. Prod. Rep.
20 (2003) 327–341.
[2] D.J. Heyes, C.N. Hunter, Making light work of enzyme catalysis: protochlorophyllide
oxidoreductase, Trends Biochem. Sci. 30 (2005) 642–649.
[3] C. Reinbothe, M. El Bakkouri, F. Buhr, N. Muraki, J. Nomata, G. Kurisu, Y. Fujita, S.
Reinbothe, Chlorophyll biosynthesis: spotlight on protochlorophyllide reduction,
Trends Plant Sci. 15 (2010) 614–624.
[4] N.S. Scrutton, M.L. Groot, D.J. Heyes, Excited state dynamics and catalytic mecha-
nism of the light-driven enzyme protochlorophyllide oxidoreductase, Phys. Chem.
Chem. Phys. 14 (2012) 8818–8824.
[5] M. Gabruk, B. Mysliwia-Kurdziel, Light-dependent protochlorophyllide oxidoreduc-
tase: phylogeny, regulation, and catalytic properties, Biochemistry 54 (2015)
5255–5262.
[6] N. Muraki,J. Nomata, K. Ebata, T. Mizoguchi, T. Shiba, H. Tamiaki, G. Kurisu, Y. Fujita,
X-ray crystal structure of the light-independent protochlorophyllide reductase, Na-
ture 465 (2010) 110–114.
[7] M.J. Bröcker, S. Schomburg, D.W. Heinz, D. Jahn, W.D. Schubert, J. Moser, Crystal
structure of the nitrogenase-like dark operative protochlorophyllide oxidoreductase
catalytic complex (ChlN/ChlB)
2
, J. Biol. Chem. 285 (2010) 27336–27345.
[8] D.J. Heyes, C. Levy, M.Sakuma, D.L. Robertson, N.S. Scrutton, A twin-track approach
has optimized proton and hydride transfer by dynamically coupled tunneling dur-
ing the evolution of protochlorophyllide oxidoreductase, J. Biol. Chem. 286 (2011)
11849–11854.
[9] D.J. Heyes, P. Heathcote, S.E.J. Rigby, M.A. Palacios, R. van Grondelle, C.N. Hunter,The
first catalytic step of the light-driven enzyme protochlorophyllide oxidoreductase
proceeds via a charge transfer complex, J. Biol. Chem. 281 (2006) 26847–26853.
[10] B.R.K. Menon, P.A. Davison, C.N. Hunter, N.S. Scrutton, D.J. Heyes, Mutagenesis alters
the catalytic mechanism of the light-driven enzyme protochlorophyllide oxidore-
ductase, J. Biol. Chem. 285 (2010) 2113–2119.
[11] B.R.K. Menon, J.P. Waltho, N.S. Scrutton, D.J. Heyes, Cryogenic and laser photoexcita-
tion studies identify multiple roles for active site residues in the light-driven en-
zyme protochlorophyllide oxidoreductase, J. Biol. Chem. 284 (2009) 18160–18166.
[12] D.J. Heyes, C.N. Hunter, Site-directed mutagenesis of Tyr-189 and Lys-193 in
NADPH: protochlorophyllide oxidoreductase from Synechocystis, Biochem. Soc.
Trans. 30 (2002) 601–604.
[13] B. Dietzek, W. Kiefer, G. Hermann, J. Popp, A.Schmitt, Solvent effects on the excited-
state processes of protochlorophyllide: a femtosecond time-resolved absorption
study, J. Phys. Chem. B 110 (2006) 4399–4406.
[14] B. Dietzek, W. Kiefer, A. Yartsev, V. Sundstrom, P. Schellenberg, P. Grigaravicius, G.
Hermann, J. Popp, M. Schmitt, The excited-state chemistry of protochlorophyllide
a: a time-resolved fluorescence study, ChemPhysChem 7 (2006) 1727–1733.
[15] D.J. Heyes, M. Sakuma, S.P. de Visser, N.S. Scrutton, Nuclear quantum tunneling in
the light-activated enzyme protochlorophyllide oxidoreductase, J. Biol. Chem. 284
(2009) 3762–3767.
[16] D.J. Heyes, C.N. Hunter, I.H.M. van Stokkum, R. van Grondelle, M.L. Groot, Ultrafast
enzymatic reaction dynamics in protochlorophyllide oxidoreductase (vol 10, pg
491, 2003), Nat. Struct. Biol. 10 (2003).
[17] D.J. Heyes, A.V. Ruban, H.M. Wilks, C.N. Hunter, Enzymology below200 K: the kinet-
ics and thermodynamics of the photochemistry catalyzed by protochlorophyllide
oxidoreductase, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 11145–11150.
[18] D.J. Heyes, S.J.O.Hardman, D. Mansell,J.M. Gardiner, N.S.Scrutton, Mechanistic reap-
praisal of early stage photochemistry in the light-driven enzyme protochlorophyllide
oxidoreductase, PLoS ONE 7 (2012).
[19] A. Garrone, N. Archipowa, P.F. Zipfel, G. Hermann, B. Dietzek, Plant
protochlorophyllide oxidoreductases A and B: catalytic efficiency and initial reac-
tion steps, J. Biol. Chem. 290 (2015) 28530–28539.
[20] D.J. Heyes, S.J.O.Hardman, T.M. Hedison, R. Hoeven, G.M. Greetham, M. Towrie, N.S.
Scrutton, Excited-state charge separation in the photochemical mechanism of the
light-driven enzyme protochlorophyllide oxidoreductase, Angew. Chem. Int. Ed.
54 (2015) 1512–1515.
[21] R. Hoeven,S.J.O. Hardman, D.J. Heyes, N.S. Scrutton, Cross-species analysis of protein
dynamics associated with hydride and proton transfer in the catalytic cycle of the
light-driven enzyme protochlorophyllide oxidoreductase, Biochemistry 55 (2016)
903–913.
[22] D.J. Heyes, M. Sakuma, N.S. Scrutton, Solvent-slaved protein motions accompany
proton but not hydride tunneling in light-activated protochlorophyllide oxidore-
ductase, Angew. Chem. Int. Ed. 48 (2009) 3850–3853.
[23] D.J. Heyes, M. Sakuma, N.S. Scrutton, Laser excitation studies of the product release
steps in the catalytic cycle of the light-driven enzyme, protochlorophyllide oxidore-
ductase, J. Biol. Chem. 282 (2007) 32015–32020.
[24] D.J. Heyes, B.R.K. Menon, M. Sakuma, N.S. Scrutton, Rate-limiting conformational
changes control the formation of the ternary enzyme–substrate complex in the
light-driven enzyme protochlorophyllide oxidoreductase (POR), Biochemistry 47
(2008) 10991–10998.
[25] M.E. Baker, Protochlorophyllide reductase is homologous to human carbonyl reduc-
tase and pig 20-beta-hydroxysteroid dehydrogenase, Biochem. J. 300 (1994)
605–607.
[26] N. Lebedev, M.P. Timko, POR structural domains important for the enzyme activity
in R-capsulatus complementation system, Photosynth. Res. 74 (2002) 153–163.
[27] H.E. Townley, R.B. Sessions, A.R. Clarke, T.R. Dafforn, W.T. Griffiths,
Protochlorophyllide oxidoreductase: a homology model examined by site-directed
mutagenesis, Proteins 44 (2001) 329–335.
[28] C. Dahlin, H. Aronsson, H.M. Wilks, N. Lebedev, C. Sundqvist, M.P. Timko, The role of
protein surface charge in catalytic activity and chloroplast membrane association of
the pea NADPH: protochlorophyllide oxidoreductase (POR) as revealed by alanine
scanning mutagenesis, Plant Mol. Biol. 39 (1999) 309–323.
[29] F. Buhr, M. El Bakkouri, O. Valdez, S. Pollmann, N. Lebedev, S. Reinbothe, C.
Reinbothe, Photoprotective role of NADPH: protochlorophyllide oxidoreductase A,
Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 12629–12634.
[30] M. Gabruk, J. Grzyb, J. Kruk, B. Mysliwa-Kurdziel, Light-dependent and light-inde-
pendent protochlorophyllide oxidoreductases share similar sequence motifs —in
silico studies, Photosynthetica 50 (2012) 529–540.
[31] C. Reinbothe, A. Lepinat, M. Deckers, E. Beck, S. Reinbothe, The extra loop
distinguishing POR from the structurally related short-chain alcohol dehydroge-
nases is dispensable for pigment binding but needed for the assembly of light-har-
vesting POR-protochlorophyllide complex, J. Biol. Chem. 278 (2003) 816–822.
[32] K. Arnold, L. Bordoli, J. Kopp, T. Schwede, The SWISS-MODEL workspace: a web-
based environment for protein structure homology modelling, Bioinformatics 22
(2006) 195–201.
[33] Hospital, A., Andrio, P., Fenollosa, C., Cicin-Sain, D., Orozco, M. & Gelpí, J.L. (2012)
MDWeb and MDMoby: an integrated web-based platform for molecular dynamics
simulation, Bioinformatics 2012, 28(9):1278–1279.
[34] P. Benkert, S.C.E.Tosatto, D.Schomburg, QMEAN: a comprehensive scoring function
for model quality assessment.", Proteins: Struct., Funct., Bioinf. 71 (1) (2008)
261–277.
[35] A.S. Konagurthu, C.F. Reboul, J.W. Schmidberger, J.A. Irving, A.M. Lesk, P.J. Stuckey,
J.C. Whisstock, A.M. Buckle, MUSTANG-MR structural sieving server: applications
in protein structural analysis and crystallography, PLoS ONE 5 (2010).
[36] D.J. Heyes, C.N. Hunter, Identification and characterization of the product release
steps within the catalytic cycle of protochlorophyllide oxidoreductase, Biochemistry
43 (2004) 8265–8271.
[37] D.J. Heyes, A.V. Ruban, C.N. Hunter, Protochlorophyllide oxidoreductase: spectro-
scopic characterization of the dark reactions, Biochemistry 42 (2003) 523–528.
[38] M.J. McFarlane, C.N. Hunter, D.J. Heyes, Kinetic characterisation of the light-driven
protochlorophyllide oxidoreductase (POR) from Thermosynechococcus elongatus,
Photochem. Photobiol. Sci. 4 (2005) 1055–1059.
[39] J.J. Snellenburg, S.P. Laptenok, R. Seger, K.M. Mullen, I.H.M. Stokkun, Glotaran: a
Java-based graphical user interface for the R package TIMP, J. Stat. Softw. 49
(2012) 1–22.
[40] J.E. Bray, B.D. Marsden, U. Oppermann, The human short-chain dehydrogenase/re-
ductase (SDR) superfamily: a bioinformatics summary, Chem. Biol. Interact. 178
(2009) 99–109.
[41] N. Tanaka, T. Nonaka, M. Nakanishi, Y. Deyashiki, A. Hara, Y.Mitsui, Crystal structure
of the ternary complex of mouse lung carbonyl reductase at 1.8 angstrom resolu-
tion: the structural origin of coenzyme specificity in the short-chain dehydroge-
nase/reductase family, Structures 4 (1996) 33–45.
[42] N. Tanaka, T. Nonaka, K.T. Nakamura, A. Hara, SDR: structure, mechanism ofaction,
and substrate recognition, Curr. Org. Chem. 5 (2001) 89–111.
[43] U. Oppermann, C. Filling, M. Hult, N. Shafqat, X.Q. Wu, M. Lindh, J. Shafqat, E.
Nordling, Y. Kallberg, B. Persson, H. Jornvall, Short-chain dehydrogenases/reduc-
tases (SDR): the 2002 update, Chem. Biol. Interact. 143 (2003) 247–253.
[44] K. Kavanagh, H. Jornvall, B. Persson, U. Oppermann, The SDR superfamily: functional
and structural diversity within a family of metabolic and regulatory enzymes, Cell.
Mol. Life Sci. 65 (2008) 3895–3906.
[45] Y. Kallberg, B. Persson, Prediction of coenzyme specificity in dehydrogenases/reduc-
tases —a hidden Markov model-based method and its application on complete ge-
nomes, FEBS J. 273 (2006) 1177–1184.
[46] C. Filling, K.D. Berndt, J. Benach, S. Knapp, T. Prozorovski, E. Nordling, R. Ladenstein,
H. Jornvall, U. Oppermann, Critical residues for structure and catalysis in short-chain
dehydrogenases/reductases, J. Biol. Chem. 277 (2002) 25677–25684.
243B.R.K. Menon et al. / Journal of Photochemistry & Photobiology, B: Biology 161 (2016) 236–243