Dynamic chirality determines critical roles for bioluminescence in symplectin-dehydrocoelenterazine system

Article abstract of DOI:10.1002/asia.201100089

Symplectin is a photoprotein containing the dehydrocoelenterazine (DCL) chromophore, which links to a cysteine residue through a covalent bond with the emission of blue light. This study focuses on the stereochemical process of the emerging stereogenic ce

Full text of DOI:10.1002/asia.201100089

FULL PAPERS
DOI: 10.1002/asia.201100089
Dynamic Chirality Determines Critical Roles for Bioluminescence in
Symplectin–Dehydrocoelenterazine System
Vorawan Kongjinda,[a, b] Yosuke Nakashima,[b] Naoki Tani,[b] Masaki Kuse,[c]
Toshio Nishikawa,[b] Chin-Hui Yu,[a] Nobuyuki Harada,[d] and Minoru Isobe*[a]
Dedicated to Professor Eun Lee on the occasion of his retirement and 65th birthday
Abstract: Symplectin is a photoprotein
containing the dehydrocoelenterazine
(DCL) chromophore, which links to a
cysteine residue through a covalent
bond with the emission of blue light.
This study focuses on the stereochemi-
cal process of the emerging stereogenic
centers. Two isomeric fluorinated DCL
analogs (2,4-diF- and 2,6-diF-DCL)
were employed owing to their different
bioluminescence activities, these being
200% and 20% compared to natural
DCL, respectively. Each of these diF-
DCLs was found to exchange with the
natural DCL in symplectin at pH 6.0.
The emerging stereogenic carbons
were racemic at the binding sites.
Changing the pH of this storage form
to the proteinꢀs optimum solubility pH
(pH 7.8) resulted in 2,4-diF-DCL-
bound symplectin luminescence, and
the spent solutions were then analyzed
DCLs moved to the active site at
pH 7.8, a change in the chirality with
the 390-Cys residue resulted. Model ex-
periments using l-cysteine-containing
CGLK-peptide supported two diaste-
reoisomers from each diF-DCL. The
significant difference in the lumines-
cence from these two chromophores is
attributed to a plausible mechanism in-
cluding the dynamically variable ste-
reogenic center emerging at the storage
and then the active site on the sym-
plectin. It is concluded that such dy-
namic chirality plays a significant role
in the symplectoteuthis biolumines-
cence.
and
coelenteramide-390-CGLK-pep-
tide and coelenteramine were detected
after a peptidase digestion. The same
analysis of the 2,6-diF-DCL-bound
symplectin, on the other hand, afforded
coelenteramine only but no coelentera-
mide. When the racemic storage diF-
Keywords: chirality
dichroism coelenterazine
cysteine · proteins
·
circular
·
·
l-
Introduction
[a] V. Kongjinda, Prof. C.-H. Yu, Prof. M. Isobe
Department of Chemistry
Developments in structural biology have revealed the details
of protein–substrate interactions through X-ray crystallo-
graphic analysis of the co-crystals. As an example of X-ray
structural analysis, Shimomura has reported that the coelen-
terazine (1) substrate changes to its hydroperoxide having a
2-(S)-configuration in the photoprotein aequorin system.[1] It
further changes in the presence of Ca2+ ions to a dioxeta-
none intermediate,[2] and then produces luminescence.[3–6]
The coelenterazine hydroperoxide selectively oxidizes the
cysteine-bound dithiothreitol moieties nearby in this protein
molecule.[7] The hydroperoxide in aequorin remains stable
before adding calcium ions, but usually the peroxide and di-
oxetanone are very unstable without the protein and decom-
pose spontaneously with emission of light. The luminous
peroxides were chemically synthesized by Usami and Isobe
National Tsing Hua University
101 Sec. 2 Kuang Fu Road, Hsinchu, 30013 (Taiwan)
Fax : (+886)3-573-6494
E-mail: minoru@mx.nthu.edu.tw
[b] V. Kongjinda, Y. Nakashima, Dr. N. Tani, Prof. T. Nishikawa
Lab. of Organic Chemistry
School of Bioagricultural Sciences
Nagoya University
Furocho, Chikusa, Nagoya 464-8601 (Japan)
[c] Dr. M. Kuse
Research Center for Material Sciences
Nagoya University
Furocho, Chikusa, Nagoya 464-8601 (Japan)
[d] Prof. N. Harada
Institute of Multidisciplinary Research for Advanced Materials
Tohoku University
Katahira, Sendai, 980-8577 (Japan)
and co-workers through photooxygenation of the coelenter-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/asia.201100089.
[8,9]
azine analogs at À788C in trifluoroethanol solvent.
They
2080
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 2080 – 2091

characterized the unstable structures using the 100% 13C-en-
riched samples by means of low-temperature 13C NMR and
FT-IR, which provided direct assignments for the short life-
time dioxetanone and hydroperoxide structures.[8,9] Many
marine bioluminescence systems, particularly in Aequorin,[3]
Oplophorus,[10,11] Watasenia,[12–14] Renilla,[15] Pholasin,[16,17]
and Obelin,[18] use coelenterazine 1 (CL; Figure 1) as the lu-
minous substrate for light emission in different photopro-
teins.[19]
In 1981, Tsuji found its membrane-bound photoprotein re-
sponded to monovalent cations, such as K+ or Na+, acting
as a trigger.[20] In 1993, Takahashi and Isobe reported that
this squid used dehydrocoelenterazine (DCL) 2 (Figure 1)
as a substrate instead of coelenterazine.[21,22] The Symplecto-
teuthis bioluminescence system is the first example using
oxidized coelenterazine DCL by covalently binding to the
photoprotein. Kuse recently reported a second example in
Pholasin.[23,24]
Symplectoteuthis oualaniensis (Tobi-Ika in Japanese) is
collected in Okinawa, and was used in this study. It is an
oceanic squid emitting blue light from the photogenic organ.
In 2002, Fujii, Isobe, and co-workers elucidated that a
60 kDa photoprotein, named symplectin,[25] is responsible
for the luminescence of S. oualaniensis. Symplectin contains
11 cysteine residues among 501 amino acids, and allows
DCL 2 binding through a covalent thioether bond as 3
(Figure 1). Binding-model studies, using 100% 13C-enriched
DCL and dithiothreitol (DTT), were also published in 1998
to reveal that the bound form of the chromophore exists in
equilibrium with its dissociated form.[26] Studies using a lu-
minous mono-fluorinated DCL analog, and subsequent de-
tection of the CGLK-chromo-peptide after trypsin digestion,
led us to conclude that the active site-thiol belonged to 390-
Cys.[27] Symplectin is not soluble in plain water, but it is only
soluble in a buffer at pH between 5–9 (optimum 7.8) con-
taining KCl salt of 0.6m or higher concentration.
Results and Discussion
Cysteine Analysis of Symplectin
The amino acid sequence of symplectin was reported in
2008.[27] The S S/SH analysis of symplectin was further im-
À
plemented about the eleven cysteine residues according to
an iodoacetamide protocol. The modified protein was di-
gested by the endopeptidase Lysine-C at pH 8 for 4 hours,
and then the mixture was subjected to capillary liquid chro-
matography-mass spectrometry (LC-MS) analysis (see de-
tails in the Supporting Information, sections S4 and S13),
and the results are shown in Figure 2. The five cysteine resi-
dues exist in free SH form locating at 196-, 339-, 344-, 345-,
Figure 1. Chromophores in symplectin bioluminescence.
Abstract in Japanese:
À
and 390-Cys, while three pairs of cystines are in S S form at
92/110, 129/137, and 380/385. Other modifications after
Lysine-C endopeptidase digestion using the capillary LC-MS
method are also shown in Figure 2. As was previously re-
ported for a sulfur-containing photoprotein, two methio-
nines (181 and 484) were found in the oxide form.[28]
Preparation of Difluorinated Dehydrocoelenterazine
Analogs
The symplectin luminescence system incorporates chromo-
phore DCL to its aposymplectin by Michael addition to
form the photoprotein with a thioether covalent bond (com-
pound 3 in Figure 1). In this case, the equilibrium of addi-
tion/elimination depends on the pH and electronic nature of
the substituent of the chromophore; namely, the phenyl ring
at the 2a-position of the chromophore.
Chem. Asian J. 2011, 6, 2080 – 2091
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
2081

M. Isobe et al.
FULL PAPERS
Figure 2. Amino acid sequence and cysteine/cystine analysis.
An ortho- or para-fluoro analog recorded good results in
the active site studies.[27] We became interested in more elec-
tron-deficient analogs to serve as substrates for this study.
Among several compounds we have prepared in preliminary
experiments, it was suggested that the 2,4-difluoro-dehydro-
coelenterazine (2,4-DiF-DCL, 4, Figure 3) analog indicated
a much higher bioluminescence activity than DCL. So here
we describe the synthesis of 4 and its position isomer 2,6-
diF-DCL (5, Figure 3). The preparation of 4 and 5 was ach-
ieved from difluorophenylacetic acids 8a and 8b in eight
steps in 14.6% and 25.5% overall yield, respectively, as
shown in Scheme 1, and the experimental details are shown
in the Supporting Information (sections S5–S13).
Equilibrium of Dehydrocoelenterazine Analogs 2,4- and 2,6-
DiF-DCL with l-Cysteine
In Figures 4–6, the UV/Vis spectra of 4 and 5 are compared
with those from their respective
adducts
6 and 7 (Figure 3),
when mixed with varying
amounts of l-cysteine (from 0
to10 equiv, as indicated in the
figures) at pH 3.2, 6.0, and 8.0,
respectively. These pH values
were selected from previous re-
sults with natural DCL 2; thus,
pH 3: adducts being stable,
pH 6.0: slow equilibration, and
pH 8: proteinꢀs optimum pH is
7.8.[22] The arrows indicate an
increase or decrease of the ab-
sorbance peaks. All of the free
DCL derivatives 2, 4, and 5, are
reddish in color (abs=550 nm)
and non-fluorescent. On the
other hand, all the adducts 3, 6,
and 7 changed to yellow in
color (abs=450 nm) and emit a
green fluorescence (GF) at
520 nm.
Under these conditions using
a methanol-buffer solvent, 2,6-
Figure 3. Two difluorinated dehydrocoelenterazine analogs, which behave differently on symplectin for lumi-
nescence.
2082
www.chemasianj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 2080 – 2091

Dynamic Chirality Determines Critical Roles for Bioluminescence
Incorporation and
Luminescence of DiF-DCL
with Aposymplectin
The capacity and rate of the
binding of the aposymplectin to
4 or 5 were different as well
when monitored by the increase
in green fluorescence at 520 nm
at pH 5.6 (Figure S1 in the Sup-
porting Information). The in-
crease in fluorescence upon
binding to 5 was faster and
greater than that observed with
4 (Figure S1 in the Supporting
Information). After these bind-
Scheme 1. Synthetic route for di-fluoro-DCLs (4 and 5).
Figure 5. UV/Vis spectra of diF-DCL analogs in the presence of various
amounts of l-cysteine at pH 6.0; A) 2,4-diF-DCL 4 and 6; B) 2,6-diF-
DCL 5 and 7. Buffer components are 10 mM KCl, 12 mM AcOH/12 mM
AcONa in 68% MeOH/H2O.
Figure 4. UV/Vis spectra of diF-DCL analogs in the presence of various
amounts of l-cysteine at pH 3.2; A) 2,4-diF-DCL 4 and 6; B) 2,6-diF-
DCL 5 and 7. Buffer components are 17.6 mM KCl, 10 mM citric acid in
68% MeOH/H2O.
ing experiments, the pH of the final mixture was raised from
5.6 to 7.8, and the green fluorescence (GF) intensity de-
creased over 10 minutes (see Figures S2–S3 in the Support-
ing Information). At the same time, the bioluminescence
was observed, and the light yields were integrated for 5 mi-
nutes. The experiment with 4 emitted 180% more light than
that with natural DCL (2), while 5 emitted less light than
diF-DCL tends to exist as the adduct form 7 rather than in
equilibrium with 5. The 2,4-diF-DCL may be insoluble at
pH 8.0 without cysteine.
Chem. Asian J. 2011, 6, 2080 – 2091
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
2083

M. Isobe et al.
FULL PAPERS
aposymplectin with 5 suggests not only that 5 produced a
non-luminescent adduct, but also its inhibitory action in the
photoprotein. Although approximately 40% of the light
yield is still observed, this might arise from the DCL re-
maining in the apoprotein or 2,6-diF-DCL (5) itself.
Molecular Shape Difference of DCL
With respect to the decrease of 520 nm fluorescence of pho-
toprotein symplectin in accordance with time after changing
the buffer from pH 5.6 to pH 7.8, both the 2,4-diF-DCL and
2,6-diF-DCL lose GF almost at the same rate (Figure S2 in
the Supporting Information). One of the factors making the
difference between the binding and bioluminescence abili-
ties of 4 and 5 might be a result of the different physico-
chemical properties of these two chromophores. Figure 8 in-
dicates two kinds of dipole–dipole interactions between the
two isomeric fluorine atoms and the carbonyl group, thus, 4
(2,4-diF, b+a) is assumed to be smaller than 5 (2,6-diF,
c+a). In addition, ortho-diF of the latter has a higher steric
interaction with the imidazolone nitrogen atom. Therefore
the conformation of these two aromatic ring planes may be
twisted differently. In fact, the dihedral angles of energy-op-
timized analogs were calculated to be 37.88 for 4 and 55.98
for 5 (Figure 8), while that of natural DCL 2 is 39.58. Only
2,6-diF-DCL shows a different molecular shape from natural
DCL. This difference could be caused by the different bind-
ing to the active site in symplectin.
Figure 6. UV/Vis spectra of diF-DCL analogs in the presence of various
amounts of l-cysteine at pH 8.0; A) 2,4-diF-DCL 4 and 6; B) 2,6-diF-
DCL 5 and 7. Buffer components are 2 mM Tris, 2 mM Tris-HCl in 68%
MeOH/H2O.
Double Incubation of DCL Analogs to Symplectin
The results of the luminescent assay (Figure 7) and confor-
mational analysis (Figure 8) led us to design the following
double incubation assay using systematic combinations of
the chromophores. Namely, aposymplectin was first incubat-
ed with one chromophore for 20 minutes at pH 5.6
(1st reconstitution), then, with a second or the same chromo-
phore for another 20 minutes (2nd reconstitution). The inte-
grated amounts of light emission at pH 7.8 after each of
these double incubations are shown in Figure 9.
the control (Figure 7). In both cases, the GF decreases
within 10 minutes at pH 7.8. While the absolute fluorescence
value of the control GF varies, owing to different amounts
of natural DCL (2) remaining in the dimethyl sulfoxide
“ACHTUNGTRENNUNG(DMSO)”-experiment in each aposymplectin preparation,
the trends in the binding of these three compounds in each
set of experiments are the same and are highly reproducible.
The lack of luminescence produced upon the incubation of
As shown in Figure 9, the double incubation experiments
with DCL (2) and 2,4-diF-DCL (4) either in the first and/or
second round (Figure 9, entries 3, 5, 7) always yielded more
than 200% light amounts relative to the control (Figure 9,
entries 1, 2). In contrast, using 2,6-diF-DCL (5) in either of
the double incubations (Figure 9, entries 4, 6, 8, 9, and 10)
resulted in very little light yields (compared with the 20–
45% in Figure 9, entry 2). This implies that 2,6-diF-DCL (5)
serves as an inhibitor and does not exhibit a biolumines-
cence ability. On the contrary, 5 shows approximately a two
times higher chemiluminescence ability than 4, when each
of them was bound to a model peptide, glutathione (gGlu-
Cys-Gly) and was subjected to tBuOK in DMSO. Both of
them were converted into coelenteramide analogs (18 and
19; X=S-(gGlu)Cys
difference between 4 and 5 in the bioluminescence with
(Gly))[26,29] (data not shown). Thus the
ACHTUNGTRENNUNG
Figure 7. Bioluminescence ability after reconstituted symplectin with nat-
ural DCL 2, 2,4-diF-DCL 4, and 2,6-diF-DCL 5. The total light yield is in
arbitrary units, but comparable with Figure 9 and Figure 14.
2084
www.chemasianj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 2080 – 2091

Dynamic Chirality Determines Critical Roles for Bioluminescence
Product Analysis of the DCL
Analogs after the
Bioluminescence
Both of the isomeric diF-DCL
analogs 4 and 5 incorporated to
aposymplectin at pH 5.6 (Fig-
ure S1 in the Supporting Infor-
mation) and were similarly con-
sumed at pH 7.8 (Figure S3 in
the Supporting Information),
while 4 shows bioluminescence
but 5 does not (Figures 7 and
9). Next, we checked the prod-
ucts after the pH 7.8 condition
in order to answer the question
whether the non-luminescent
2,6-diF-DCL analog would pro-
vide the corresponding coelen-
teramide analog 19 or not.
After the luminescent condi-
tions, both of the spent solu-
tions from the experiments re-
Figure 8. Dipole–dipole interaction of 4 and 5 with the two aromatic planes twisted by 388 and 568, respective-
ly.
lating to Figures 7 and
9
showed blue fluorescence. Of
the possible products, as shown
symplectin is assumed to be attributed to their interacting or
binding modes with symplectin.
in Scheme 2, the analogs coelenteramide 17, 18, and 19
should show similar fluorescence maxima; thus, FLmax at
around 420 nm and coelenteramine 23 at FLmax 410 nm.
However, the differentiation of these molecules was not so
Figure 9. Bioluminescence at pH 7.8 of samples from doubly reconstituted symplectin with DCL and/or diF-DCL analogs at pH 5.6 for 20 min in each in-
cubation.
Chem. Asian J. 2011, 6, 2080 – 2091
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
2085

M. Isobe et al.
FULL PAPERS
peak corresponding to 19 was
found from the experiments de-
rived from 5 either by FL or
MS, but coelenteramine 23 was
clearly detected as the only flu-
orescent product.
LC-CD-UV Analysis of the
DiF-DCL Analogs with l-
Cysteine and CGLK-Peptides
We now became interested in
the absolute configuration that
emerged on the chromophore
as a result of the temporary ste-
reogenic carbon atom attached
to an S-atom from l-cysteine or
a cysteine residue of symplec-
tin. In our previous work, DCL
was proven to bind at the 390-
Cys of symplectin as the active
center for the biolumines-
cence.[27] Judging from the re-
constitution experiments (Fig-
ures 7 and 9), this binding
seems to be more dynamic in
that the chromophores ex-
change in the photoprotein.
The methodology for determin-
ing the absolute configura-
Scheme 2. Structures of dehydrocoelenterazine (DCL) and its analogs and their possible changes on protein
surface.
easy owing to the similar fluorescence at pH 7.8. Fortunate-
ly, coelenteramide showed a fluorescence-shift at alkaline
pH in DMSO solvent to around 530 nm. This FL-shift
method was, in fact, applied to the spent solutions from 4
and 5 by addition of 1N NaOH and DMSO and the differ-
ence in fluorescence spectra was measured (for details, see
Figures S4 and S5 in the Supporting Information). It was
concluded that 4 showed a significant alkaline shift but none
was observed for 5. Thus, 5 would directly yield 23, and
would never provide coelenteramide 19.
Further product analysis of 18 but not 19 was performed
by capillary LC-MS as follows. Each of the spent solutions
from symplectin and diF-DCL analogs 4 and 5 were digest-
ed with endopeptidase Lys-C at pH 8 for 4 hours or over-
night. The resultant hydrolysates were analyzed with a
house-assembled non-split-flow capillary LC-ESI-IT-MS
(electrospray ionization-ion trap-mass spectrometry) experi-
ment equipped with a capillary FL-detector, and the results
are shown in Figure 10. From the digested peptide mixture
derived from 4, we observed 2,4-diF-coelenteramide-S-
CGLK 18, (FL 420 nm, X=S-Cys390-Gly391-Leu392-Lys393; m/
z: 849.3, at retention time (r.t.)=46.6 min) together with
coelenteramine 23 (m/z: 278.2, r.t.=50.6 min). These frag-
ment ions and the assignments are illustrated in Figure 10.
These experiments confirmed our previous results that the
active site cystine being 390-Cys with mono-fluoro-DCL 20
and 22.[27] On the other hand, no analogous chromo-peptide
tions[30] usually requires exciton coupling, X-ray crystallogra-
phy, or the 1H NMR anisotropy effect. However, these
methods are very limited for cases where the temporary
chirality changes dynamically taking place on the protein
surface need to be determined, as in the current case. None
of these methods are applicable to pico-mol scale experi-
ments. In addition, the current case includes the exchanging
of chromophores on account of the addition–elimination
equilibrium at the binding site and moving from storage site
to active site by dynamically changing the chirality. In the
current research, we focused on the significant role of the
stereochemistry of the chromophore determining the lumi-
nescent activity (Scheme 2).
As model experiments for chirality studies, DCL 4 was
converted to the l-cysteine-containing CGLK peptide (X=
À
S-Cys-Gly-Leu-Lys in Scheme 2) by addition of 1.2 equiva-
À
lents of the peptide and the diastereomeric mixture 6 (X=
SCH2CH2CGLK) was immediately analyzed by LC-CD
equipped with a reversed phase ODS column (4.6 mmfꢂ
25 cm) under acidic conditions (35% CH3CN/H2O contain-
ing 0.1% TFA, approximately pH 3) using a flow-cell UV-
CD detector. The results are shown in Figure 11, where the
first elution peak 1 at 20.2 minutes shows negative CD chro-
matographed at 310 nm (A, C), while the second elution
peak 2 at 22.9 minutes shows a positive signal. From the UV
data of Figure 11B,D, the same amount of two diastereo-
mers were monitored. The flow-stop-scanning of CD and
2086
www.chemasianj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 2080 – 2091

Dynamic Chirality Determines Critical Roles for Bioluminescence
with negative first and positive second Cotton effect on elu-
tion peak 1. The reversely positive first and negative second
Cotton effect is seen on the elution peak 2. These Cotton ef-
fects correspond to the emerging stereogenic carbons on
À
chromophore 6. Similarly those spectra of
7 (X= S-
CGLK), which were prepared from 2,6-diF-DCL 5, are illus-
trated in Figure 12A–D. Simpler examples from l-cysteine
adducts of 2, 4, and 5 are shown in the Supporting Informa-
tion (Figure S8–S10, respectively). The delta epsilon values
De of these CD data were calculated to be 5–8, and are
tabulated in Table S1 in the Supporting Information.
All of these samples (3, 6, and 7) showed a similar ten-
dency in the chromatograms and CD spectra. Among the
three adducts 3, 6, and 7, 2,6-diF-DCL/l-cysteine adduct 7
seems the most stable judging from the higher peak height
ratios of elution peak 2 to peak 1. The different stability
may be attributed to the electron-withdrawing nature by flu-
orine atoms as discussed above. So the CD data was accu-
rately measured with 2,6-diF-DCL/l-Cys adduct using 2,6-
diF-CL (7, X=H) having e=16900. The small values of
delta epsilon De from the CD spectra and Cotton peak
found at 310 nm and around 260 nm are so small that the
absolute configuration is unable to be predicted by using the
exciton chirality method. It is, however, suggested that the
same dominant stereogenic center is observed when the
chromophore binds with the l-cysteine residue of an amino
acid or peptide model as summarized in Table S1 in the Sup-
porting Information. It is clear that the chirality comes from
the stereogenic carbon carrying the sulfur atom, so determi-
nation of this chirality requires further comparison with au-
thentic sample preparations in a future work.
Molecular Mechanic Calculation of 2,4-DiF-DCL- and 2,6-
DiF-DCL-Cysteine Adducts
In order to compare the molecular shape difference of DCL
analogs, as shown in Figure 8, here we have performed mo-
lecular mechanic calculations of diF-DCL analogs and l-cys-
teine adducts for determination of the global minimum con-
formation. To make the calculation easier, we simplified the
À
substituents to CH3 groups instead of CH2Ph and the CH3
À
SH (S Me) instead of cysteine adduct as the 2,4-diF-DCL
model compound 24 (2R)- and 2,6-diF-DCL model 25 (2S)-
configuration of the isomeric diF-DCL chromophores 4 and
5, respectively (Figure 13).
The calculations above are focusing on the stereogenic
center. Full geometry optimizations and relaxed potential
energy surface scans were performed by means of hybrid
density functional theory B3LYP with 6-31G(d) basis sets
using the Gaussian 09 program package.[31,32] The conforma-
tional analysis for the torsional energy curve was computed
Figure 10. Product analysis of the 2,4-diF-DCL after bioluminescence and
À
by varying C1-C2-C3-C4 dihedral angle q around the C2
symplectin endopeptidase Lys-C digest, using capillary HPLC-ESI-IT-
MS.
C3 bond (as shown in the Supporting Information, Figur-
es S11–S12) stepwise by 7.58 for (R)-2,4-DCL 24 and (S)-
2,6-DCL 25 from the global minima. To obtain local minima
and transition states on the scanned curve, the structural
points close to the top and the bottom in the scanned tor-
UV spectra at each peak top are also seen in Figure 11C, in
which the Cotton effect is seen around 310 nm and 250 nm
Chem. Asian J. 2011, 6, 2080 – 2091
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
2087

M. Isobe et al.
FULL PAPERS
Figure 11. LC-CD analysis of authentic chromopeptide, CGLK-2,4-diF-DCL-(l)-cysteine diastereomeric adducts; A) CD chromatogram at l=310 nm,
B) UV chromatogram l=310 nm, C) CD spectra, and D) UV spectra. Chromatogram and spectra were separated with ODS-5 column, 4.6 mm-i.d.ꢂ
250 mm using 35% CH3CN/H2O containing 0.1% TFA as a mobile phase at 0.5 mLminÀ1. The sample was applied in the amount of 2 mL at 0.2 mM
(0.4 nmol).
Figure 12. LC-CD analysis of chromopeptide, CGLK-2,6-diF-DCL-(L)-cysteine diastereomeric adducts; A) CD chromatogram at l=310 nm, B) UV
chromatogram l=310 nm, C) CD spectra, and D) UV spectra. Chromatogram and spectra were separated with ODS-5 column, 4.6 mm-i.d.ꢂ250 mm
using 35% CH3CN/H2O containing 0.1% TFA as a mobile phase at 0.5 mLminÀ1. The sample was applied in the amount of 3 mL at 0.5 mM (1.5 nmol).
sional curve were further fully optimized. Frequency calcula-
tions were employed to verify stationary points to be
minima with all real frequencies and transition states with
only one imaginary frequency. The conformation around the
2088
www.chemasianj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 2080 – 2091

Dynamic Chirality Determines Critical Roles for Bioluminescence
6.0 mm i.d. for separating the free chromophore) and the
same CD/UV/FL detectors as previous cases to detect ap-
proximately 1.5 nmol, estimated from Figures 11 and 12.
The results are shown in Figure 15. The absorption and FL
intensities suggested that even though nearly all equal
amounts of DCL 2 analogs 4 and 5 are injected (Figure 15
entries 2–5), the CD detector at 310 nm only shows an in-
crease of the negative signal of 30% with 2,4-diF-DCL 4
(Figure 15, entry 2). This negative CD cotton signal clearly
showed the similarity of the result from peak 1 with l-cys-
teine containing the peptide model experiments.
Figure 13. Chemical representations of the simplified structures of the
diF-DCL analogs and l-cysteine adducts for determination of the global
minimum conformation. See the Supporting Information, Figure S13 for
more details on the conformational calculations.
stereogenic carbon of 6 and 7 (both X=l-Cys) are conclud-
ed to be similar having the heterocyclic aromatic ring and
di-fluorophenyl ring located in a dihedral angle with C2’(R)-
configuration as negative as À598 and +1538 twisting by ap-
proximately 608 counter-clockwise; while C2’(S)-configura-
tion as negative as À1588 twisting by approximately +208
clockwise. These calculations, however, do not lead to a con-
clusion for the absolute configuration. It should be men-
tioned that the rotational restriction of the 2,6-diF-DCL-Cys
adduct 7 is only one conformation compared with 6 having
two conformations in about 1:1 population.
LC-CD Analysis of Symplectin and DCL and its Analogs
To compare these results from l-Cys adducts to the sym-
plectin-bound chromophore, we prepared symplectin sam-
ples in a similar fashion as for the experiments relating to
Figure 9. In this case, the samples were incubated only for a
very short time (2 min) and then subjected to the corre-
sponding bioluminescence measurements. The results are
summarized in Figure 14.
Figure 15. LC-CD-UV-FL analysis of symplectin exchanged and re-ex-
changed with DCL analogs for 2 min at pH 7.8; A) CD chromatogram at
l=310 nm at pH 6, B) UV chromatogram at l=310 nm, and C) FL chro-
matogram at 520 nm.
The same set of 2 minute experiments were implemented
before injecting into the HPLC equipped with a very short
GPC column (TSK-gel guard column SWXL 40 mmꢂ
Conclusions
When the two isomeric diF-DCL analogs were added to
aposymplectin, a significant increase in green fluorescence
(GF) intensity was observed at 520 nm at pH 5.6 (see the
Supporting Information, Figure S1), and GF decreased by
increasing the pH to 7.8 (see the Supporting Information,
Figure S3) by a similar degree for both the 2,4- and 2,6-diF-
DCL samples (see the Supporting Information, Figure S1).
The LC-CD signal increase was observed only in the 2,4-diF
case (Figure 15). As summarized in Scheme 3, this implies
À
that the DCLs are bound with Cys SH residues at the bind-
ing sites (A) largely as a racemic mixture at pH 5.6–6.0. At
pH 7.8, the chromophores are then transferred from the
storage site cysteines to the active site 390-Cys to form the
intermediate (B). The chromophore 2,4-diF-DCL binds
largely in one stereochemistry (either R- or S-configuration,
Figure 14. Bioluminescence activity of symplectin re-exchanged with
DCL and diF-DCL analogs after each 2 min incubation. A test of DMSO
corresponds to the control for the activity of the symplectin solution.
Chem. Asian J. 2011, 6, 2080 – 2091
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
2089

M. Isobe et al.
FULL PAPERS
Scheme 3. Postulated mechanism of Symplectoteuthis bioluminescence.
À
e.g., B1) and then it is further oxidized to the hydroperox-
ides by photoprotein and active oxygen. In this oxidation, it
is likely to result in the peroxi-carbon atom in either one of
the R- or S-configurations.[9,33]
anymore owing to the damage of 390-Cys SH. This damage
might make the second run of the photoprotein inefficient,
so it looks like inhibition. The chirality at the binding site is
racemic judging from the lack of increase of the CD but
there is an increase in green fluorescence (see the Support-
ing Information, Figure S1), and then the chromophores
move to the active site at pH 7.8 and stay for a short time at
a different diastereoisomer before the next oxidation step
(B to C). Thus, such dynamism of the chirality in B–E plays
a crucial role for determining the bioluminescence ability al-
though the starting material and final product have no ste-
reogenic center. Other photoproteins might include such a
temporary and dynamic chirality mechanism, the studies of
which are still in progress[34,35] as well as to be continued in-
cluding new bioanalytical and substrate design.[34]
On the contrary, 2,6-diF-DCL binds more tightly with
À
Cys SH, and moves to the active center resulting presuma-
bly in a different configuration (e.g., S-B), and then forms
the hydroperoxide as well. The stereoisomer in this case is
hypothesized to be a diastereoisomer (e.g., SS, C2), which is
further reacted intramolecularly between the hydroperoxide
and the sulfide (of 390-cysteine) to give sulfoxide-aminal
(D2) and then further to an imino-acylamide.[9,21] Subse-
quently, it is directly hydrolyzed without luminescence to
aminopyrazine (F=23) without going through coelentera-
mide, and the photoprotein cannot accept the next substrate
2090
www.chemasianj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 2080 – 2091

Dynamic Chirality Determines Critical Roles for Bioluminescence
[15] S. Inouye, O. Shimomura, Biochem. Biophys. Res. Commun. 1997,
233, 349–353.
[16] P. A. Roberts, J. Knight, A. K. Campbell, Biochem. Soc. Trans. 1985,
13, 1139–1140.
Experimental Section
The Experimental Section for this work can be found in the Supporting
Information.
[17] T. Mꢃller, A. K. Campbell, J. Biolumin. Chemilumin. 1990, 5, 25–
30.
[18] E. V. Eremeeva, S. V. Markova, A. H. Westphal, J. W. G. Visser,
W. J. H. van Berkel, E. S. Vysotski, FEBS Lett. 2009, 583, 1939–
1944.
Acknowledgements
[19] O. Shimomura, S. Inoue, F. H. Johnson, Y. Haneda, Comp. Biochem.
Physiol. 1980, 65B, 435–437.
[20] F. I. Tsuji, G. B. Leisman, Proc. Natl. Acad. Sci. USA 1981, 78,
6719–6723.
[21] H. Takahashi, M. Isobe, Bioorg. Med. Chem. Lett. 1993, 3, 2647–
2652.
[22] H. Takahashi, M. Isobe, Chem. Lett. 1994, 843–846.
[23] M. Kuse, E. Tanaka, T. Nishikawa, Bioorg. Med. Chem. Lett. 2008,
18, 5657–5659.
[24] E. Tanaka, M. Kuse, T. Nishikawa, ChemBioChem 2009, 10, 2725–
2729.
[25] T. Fujii, J.-Y. Ahn, M. Kuse, H. Mori, T. Matsuda, M. Isobe, Bio-
chem. Biophys. Res. Commun. 2002, 293, 874–879.
[26] M. Isobe, M. Kuse, Y. Yasuda, H. Takahashi, Bioorg. Med. Chem.
Lett. 1998, 8, 2919–2924.
The authors are indebted to Nagoya University, MEXT-Japan, and JSPS,
as well as National Tsing Hua University and National Science Council
of Taiwan for financial support. Technical support by JACSO Co. Ltd.
and Toyo-Soda Co. Ltd. are also acknowledged. The authors are also
grateful to Japanese government scholarship (Monbukagakusho) for V.K.
to perform her PhD thesis work in the foreign student program.
[1] J. F. Head, S. Inouye, K. Teranishi, O. Shimomura, Nature 2000, 405,
372–376.
[2] F. McCapra, Endeavour 1973, 32, 139–U145.
[3] O. Shimomura, F. H. Johnson, Proc. Natl. Acad. Sci. USA 1978, 75,
2611–2615.
[4] O. Shimomura, B. Musicki, Y. Kishi, Biochem. J. 1988, 251, 405–
410.
[5] O. Shimomura, F. H. Johnson, Symp. Soc. Exp. Biol. 1976, 30, 41–
54.
[6] O. Shimomura, Biochem. Biophys. Res. Commun. 1995, 211, 359–
363.
[7] I. Doi, M. Kuse, T. Nishikawa, M. Isobe, Bioorg. Med. Chem. 2009,
17, 3399–3404.
[8] K. Usami, M. Isobe, Tetrahedron Lett. 1995, 36, 8613–8616.
[9] K. Usami, M. Isobe, Tetrahedron 1996, 52, 12061–12090.
[10] O. Shimomura, T. Masugi, F. H. Johnson, Y. Haneda, Biochemistry
1978, 17, 994–998.
[11] C. M. Thomson, P. J. Herring, A. K. Campbell, Mar. Biol. 1995, 124,
197–207.
[12] F. I. Tsuji, Biochim. Biophys. Acta Biomembr. 2002, 1564, 189–197.
[13] S. Inoue, H. Kakoi, T. Goto, Tetrahedron Lett. 1976, 17, 2971–2974.
[14] T. Goto, H. Iio, S. Inoue, H. Kakoi, Tetrahedron Lett. 1974, 15,
2321–2324.
[27] M. Isobe, M. Kuse, N. Tani, T. Fujii, T. Matsuda, Proc. Jpn. Acad.
Ser. B 2008, 84, 386–392.
À À À
À
À
À
[28] Cys S S CH2 CH(OH) CH=CH SOACTHNUGRTNEUNG(1 or 2)H (for more details;
see Ref. [7]).
[29] M. Kuse, M. Isobe, Tetrahedron 2000, 56, 2629–2639.
[30] N. Harada, Chirality 2008, 20, 691–723.
[31] A. D. Becke, J. Chem. Phys. 1993, 98.
[32] C. Lee, W. Yang; R. G. Parr, Phys. Rev. B 1988, 37, 785–789; R. G.
Parr, Phys. Rev. B 1988, 37, 785–789.
[33] H. Isobe, S. Yamanaka, S. Kuramitsu, K. Yamaguchi, J. Am. Chem.
Soc. 2008, 130, 132–149.
[34] M. O. Sydnes, M. Isobe, Chem. Asian J. 2010, 5, 410–420.
[35] W. T. Phakhodee, J. M. Jhou, N. Khunnawutmanotham, M. Isobe,
Tetrahedron 2011, 67, 1150–1157.
Received: January 31, 2011
Published online: June 7, 2011
Chem. Asian J. 2011, 6, 2080 – 2091
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
2091