A synthetic GFP-like chromophore undergoes base-catalyzed autoxidation into acylimine red form

Article abstract of DOI:10.1021/jo200150b

Fluorescent proteins are widely used in modern experimental biology, but much controversy exists regarding details of maturation of different types of their chromophores. Here we studied possible mechanisms of DsRed-type red chromophore formation using sy

Full text of DOI:10.1021/jo200150b

ARTICLE
pubs.acs.org/joc
A Synthetic GFP-like Chromophore Undergoes Base-Catalyzed
Autoxidation into Acylimine Red Form
Pavel E. Ivashkin, Konstantin A. Lukyanov, Sergey Lukyanov, and Ilia V. Yampolsky*
Institute of Bioorganic Chemistry, Russian Academy of Sciences, Miklukho-Maklaya 16/10, 117997 Moscow, Russia
S Supporting Information
b
ABSTRACT: Fluorescent proteins are widely used in modern
experimental biology, but much controversy exists regarding details
of maturation of different types of their chromophores. Here we
studied possible mechanisms of DsRed-type red chromophore for-
mation using synthetic biomimetic GFP-like chromophores, bearing
an acylamino substituent, corresponding to an amino acid residue at
position 65. We have shown these model compounds to readily react
with molecular oxygen to produce a highly unstable DsRed-like
acylimine, isolated in the form of stable derivatives. Under the same
aerobic conditions an unusual red-shifted imide chromophore—a
product of 4-electron oxidation of Gly65 residue—is formed. Our
data showed that GFP chromophore is prone to autoxidation at
position 65 CR by its chemical nature with basic conditions being the
only key factor required.
’ INTRODUCTION
can be further modified within some proteins to produce
additional types of chromophores. For example, chromoprotein
asFP595 from the sea anemone Anemonia sulcata possesses a
protein backbone break just before its chromophore that con-
tains a keto group conjugated to the GFP-like core.5,6 Appar-
ently, this structure is formed by hydrolysis of the acylimino
group.
Green Fluorescent Protein (GFP) from jellyfish Aequorea
victoria and numerous homologue proteins from various marine
animals attract a keen interest due to their intensive use as
genetically encoded fluorescent labels in living systems.1 The
most intriguing feature of GFP-like proteins is self-catalyzed
formation of their chromophores. During this multistep process
three internal amino acids of fluorescent proteins (Ser65-Tyr66-
Gly67 in GFP) undergo chemical modifications resulting in bi- or
tricyclic structures with an extended system of conjugated double
bonds capable of absorbing and emitting visible light.
In GFP, chromophore is formed by cyclization of the protein
backbone (carbonyl group of Ser65 reacts with amide nitrogen of
Gly67 with formation of 5-membered heterocycle) and oxidation
of the CRꢀCβ bond of Tyr66. Chromophore formation is a fully
self-catalyzed process, no external enzymes and cofactors are
required, except molecular oxygen. These reactions result in a so-
called GFP-like chromophore: 5-(4-hydroxybenzylidene)-3,5-
dihydro-4H-imidazol-4-one (Figure 1). Structural studies re-
vealed that all known natural proteins of the GFP family contain
the GFP-like chromophore as a core structure. A number of
further modifications of the GFP-like core were found in red-
shifted proteins.2 The majority of natural red fluorescent proteins
and chromoproteins carry the so-called DsRed-like chromo-
phore, which includes the acylimine moiety formed by dehy-
drogenation of the CRꢀN bond of a residue at position 65 (here
and below numbering is according to GFP) by molecular
oxygen.3,4 The acylimine group of the DsRed-like chromophore
is stable only within the protein and readily undergoes hydrolysis
after protein denaturation.3 Moreover, this highly reactive group
Much controversy exists regarding details of both GFP-like
and DsRed-like chromophore formation. A “conventional” ma-
turation scheme of DsRed-like proteins, which was generally
accepted for many years, postulates autoxidation of GFP-like
intermediate into final DsRed-like chromophore (Figure 1,
upper pathway).4 On the contrary, the recent studies by Subach
et al.7 and Strack et al.8 give evidence for an unexpected blue-
shifted intermediate containing an N-acylamino group and a
saturated Tyr66 Cβ atom (Figure 1, bottom pathway). In these
studies the GFP-like chromophore was found to be a byproduct,
not an intermediate on the way to the fully mature red protein.
The existing experimental data suggest that both mechanisms
of RFP maturation are independently realized in Nature in
separate evolutionary branches of fluorescent proteins. Some
proteins (including DsRed, hcCP, cgCP, TagRFP, etc.) maturate
via a blue-shifted intermediate,7,8 while others, e.g., z2FP574 and
asFP595, undergo maturation with a green GFP-like form clearly
being an intermediate.9,10
Here we studied one of the possible mechanisms of DsRed-like
chromophore formation using chemical synthesis of biomimetic
Received: January 27, 2011
Published: March 11, 2011
r
2011 American Chemical Society
2782
dx.doi.org/10.1021/jo200150b J. Org. Chem. 2011, 76, 2782–2791
|

The Journal of Organic Chemistry
ARTICLE
Figure 1. Two routes of DsRed-like chromophore formation.
Figure 2. Synthesis of model GFP-like chromophores 5aꢀc. Reagents and conditions: (a) SOCl2, MeOH; (b) AcNHCHRCO2H/DCC/HOBt; (c)
MeNH2; (d) (EtCO)2O/ZnCl2 or AcBr/Ac2O; (e) H2, Pd/C; (f) K2CO3/DMF.
t
GFP-like chromophores bearing an R-acetylamino-substituent,
which models an amino acid at position 65 responsible for
formation of the DsRed-like acylimine group. These model
chromophores were found to undergo green-to-red oxidative
conversion upon contact with molecular oxygen in basic condi-
tions. This reaction leads to the trapped forms of DsRed-like
acylimine and also to a novel chromophore arising from the
4-electron oxidation.
O-benzylated precursor 1 with acetylamino acid (Gly, Phe, Leu)
under standard DCC/HOBt coupling conditions followed by
reaction with excess methylamine afforded the desired β-hydro-
xytyrosine derivative 2 in high yield. The latter was O-acylated and
debenzylated under standard catalytic hydrogenation condi-
tions. The resulting 3 was subjected to base-catalyzed elimination/
cyclization reaction. In the case of 3c this reaction was clearly stepwise
with a dehydrotyrosine derivative 4c formed at the first step. For the
synthesis of 5a (Figure 2) a strictly anaerobic environment is
required. Even traces of oxygen significantly reduce the yield of
the target product (see below). In the case of tert-leucine derivatives
peptide coupling and cyclization steps are significantly slowed
down, presumably due to the bulkiness of the tert-butyl group.
’ RESULTS
Synthesis of r-Acylamino-Substituted 4-(4-Hydroxy-
benzylidene)imidazolin-5-ones. N-Acylation of a common
2783
dx.doi.org/10.1021/jo200150b |J. Org. Chem. 2011, 76, 2782–2791

The Journal of Organic Chemistry
ARTICLE
Figure 3. Autoxidation of Gly-derived GFP-like chromophore 5a.
autoxidation greatly depends on the amino acid residue, decreas-
ing in the following order: Gly-derivative 5a (half-life <1 min at
100 °C in DMF/Cs2CO3), Phe-derivative 5b (half-life about 3 h
under the same conditions), and tert-leucine analogue 5c (almost
inert).
Autoxidation of 5a leads to two products 6 and 7 (Figure 3),
their ratio depending on the base used, temperature, and the
presence of moisture. Thus, in DMF containing NEtiPr2,
Cs2CO3, and NBu4F the 6:7 ratio was 3:1, 1:1.5, and 1:2,
respectively. At elevated temperatures aldehyde 8 appeared in
the reaction mixture along with 6 and 7. The separate experiment
where pure 6 was heated in basic DMF showed 8 to result from
decomposition of 6. In strictly anhydrous conditions the yield of
7 in autoxidation of 5a remained unchanged, whereas only traces
of 6 were observed.
Purified 6 remained unchanged when subjected to the same
autoxidation conditions suggesting that two independent path-
ways lead from 5a to 6 and 7 (Figure 3). Another alternative is
the autoxidation of unhydrated acylimine intermediate but it can
be ruled out since autoxidation should be much slower than
hydration.
Autoxidation of Phe-derivative 5b leads to a tautomer of
DsRed-like chromophore 9 (mixture of cis- and trans-isomers)
(Figure 4).
tert-Leucine derivative 5c is almost inert toward autoxidation
even under prolonged boiling in oxygenated DMF/Cs2CO3.
Chromophore6 10 was studied under the same autoxidation
conditions as compounds 5aꢀc. The rate of autoxidation was
very low (half-life approximately 6 h at 100 °C in DMF/
Cs2CO3), consistent with reported preparation of 10 not
requiring anaerobic environment. The two main products of
autoxidation reaction were oxo-derivative6 11 and hydantoine
12 (Figure 4).
Figure 4. Autoxidation of phenylalanine-derived GFP-like chromo-
phore 5b and ethyl-substituted chromophore 10.
Details of the autoxidation study of 10 are given in the
Supporting Information.
Spectral Properties of 5aꢀc, 6, 7, and 9. As expected, both 5
and 6 displayed spectral behavior similar to that of the simplest
2-methyl-substituted chromophore11 (Table 1). In contrast,
Autoxidation of r-Acylamino-Substituted Chromophores
5aꢀc and 2-Ethyl-Substituted Chromophore 10. Chromo-
phores 5aꢀc undergo autoxidation in the presence of bases. The
half-life of 5a in THF at room temperature decreases from more
than 8 h to 20 min upon addition of 0.01 M NEt3. The rate of
2784
dx.doi.org/10.1021/jo200150b |J. Org. Chem. 2011, 76, 2782–2791

The Journal of Organic Chemistry
ARTICLE
Table 1. Spectral Properties of 5aꢀc, 6, 7, and 9
neutral forma
anionic formb
abs max, nm
(ε, Mꢀ1 cmꢀ1
abs max, nm
(ε, Mꢀ1 cmꢀ1
)
emission
max, nm
compd
)
5, 6
7
376
492
n.d.
603
618
419 (48000)
577 (77000)
9
400 (25000)
514 (46000)
a Measured in 10 mM AcOH in DMF. b Measured in 5 mM Cs2CO3
in DMF.
absorption maxima of the chromophores 7 and 9 showed a
strong red shift for both neutral and anionic states (Table 1
and Figure 5). Interestingly, absorption maximum of chromo-
phore 9 (514 nm in basic DMF) is strongly blue-shifted with
respect to the corresponding Kaede-like chromophore12 with-
out an N-acylamino group (553 nm). This fact is consistent
with the lowered polarization of the chromophore due to the
cross-conjugation with the acylamino group. Anionic forms
of 7 and 9 exhibited weak red fluorescence (fluorescence
quantum yield ∼5 ꢁ 10ꢀ4). A considerable difference be-
tween absorbance and excitation maxima of 9 in basic DMF
was observed (Figure 5B). A similar peculiarity was documen-
ted for the close structural analogue of 9—Kaede-like
chromophore12 FYG—as well as for some mutant fluorescent
proteins.13
Figure 5. Absorbance and fluorescence spectra of chromophores 7 (A)
and 9 (B): normalized absorbance in acidic DMF (blue lines) and basic
DMF (magenta lines), excitation in basic DMF (green lines), and
emission in basic DMF (red lines)
’ DISCUSSION
Synthesis of GFP-like Substrates for Autoxidation. Known
methods for the synthesis of imidazolones14 include base-cata-
lyzed cyclization of dehydrotyrosine derivatives,11 reaction of
amidines15,16 or imidates17 with suitable 1,2-dielectrophiles, intra-
molecular aza-Wittig reaction of R-azidoimides,18 and cross-
coupling of boronic acids with thioimidazolone.19 We developed
a synthetic approach to previously unknown R-acetylamino-
substituted GFP-like chromophores 5 based on cyclization of
dehydrotyrosine derivatives 4 (Figure 2). We found that Erlen-
meyer azlactonization usually used for synthesis of dehydrotyr-
osine derivatives does not tolerate the R-acetylamino group
(presumably due to the rearrangement reported elsewhere20).
Direct N-acylation of O-protected ΔTyr is known to be
inefficient,21 so we used a readily available synthetic equivalent22
of ΔTyr 1. Contrary to the previous results23 peptide 2 under-
went fragmentation leading to 4-benzyloxybenzaldehyde at the
attempt at acylation in basic conditions. However, we were able
to perform acylation of 2 in the presence of a Lewis acid
((EtCO)2O/ZnCl2 or AcBr/Ac2O). Potassium carbonate was
used at a cyclization stage instead of Cs2CO3 (proposed earlier24)
because of complicated separation of the highly polar product 5a
from cesium salts. Inert atmosphere is required for the preparation
of 5a as it reacts with oxygen readily.
be another key factor determining the autoxidation rate, con-
sistent with the manifold rate decrease from 5a to 5c. The
facilitating and directing effect of the N-acylamino group on
autoxidation is evident from comparison of chromophores 5 with
ethyl-substituted chromophore 10, which is autoxidized very
slowly, producing a different set of products (see Figure S1 in the
Supporting Information for details).
On the basis of the above data we propose R-hydroperoxide 13
(first suggested for the DsRed maturation mechanism4) to be a
common intermediate in all autoxidation reactions observed
(Figure 6). Its fate is determined by the nature of the amino
acid residue. Thus, both 13a and 13b undergo hydrogen
peroxide elimination leading to the highly unstable DsRed-like
acylimines 14a and 14b. In the case of Phe-derived acylimine 14b
the likely path for its stabilization is tautomerization into
acylenamine 9, whereas the Gly-derived acylimine 14a is stabi-
lized by addition of a water molecule resulting in hyroxylamide 6.
For Gly-derived hydroperoxide 13a, an additional decomposi-
tion path arises due to the absence of the side chain in the Gly
residue. Cleavage of the OꢀO bond results in a 4-electron
oxidation product 7 that is formed independently from a
2-electron oxidation product 6. The 6:7 ratio is thought to
depend on relative rates of NH vs CH deprotonation in 13a.
These facts taken together imply that the 2R-position of the
chromophore is in fact the site of oxygenation.
Autoxidation Mechanism. We found that the autoxidation
rate greatly depends on the basicity of the media and on the
nature of the amino acid side chain. In neutral and acidic media
(AcOH or TFA in CH2Cl2, THF, or DMF) 5a is subject to very
slow autoxidation. Addition of bases leads to a dramatic accel-
eration of the reaction. Increasing the concentration of the base
as well as its strength results in further enhancement of the
reaction rate. The bulkiness of the amino acid side chain seems to
Isolation of acylimines 14 corresponding to the DsRed-like
chromophore is complicated by their high electrophilicity. Our
attempts to prepare sterically hindered and thus less reactive
2785
dx.doi.org/10.1021/jo200150b |J. Org. Chem. 2011, 76, 2782–2791

The Journal of Organic Chemistry
ARTICLE
Figure 6. A proposed mechanism of autoxidation of GFP-like chromophores 5a and 5b leading to stabilized forms of DsRed-like acylimine 6 and 9 and
to an unexpected imide 7.
tert-butyl-substituted acylimine were unsuccessful because tert-
Leu derivative 5c is almost inert toward oxygen (higher tem-
peratures result in decomposition not associated with R-auto-
xidation). Low reactivity of 5c is consistent with both the
bulkiness and the electron-donating effect of the tert-butyl group.
Relevance to Fluorescent Proteins. Our data for the first
time show that the GFP chromophore is prone to autoxidation at
position 65 CR by its chemical nature with basic conditions being
the only key factor required. It should be emphasized, that the
conditions of this autoxidation closely resemble native condi-
tions of maturation of FPs. Indeed, dimethylformamide is known
to mimic a polyamide protein shell, and in the case of nonbulky
chromophore analogues the autoxidation process took place
effectively at room temperature with a half-life of 20 min, which
is comparable with or even faster than maturation of red
fluorescent proteins.
Therefore, in red FPs maturing through a GFP-like inter-
mediate (such as z2FP574 and asFP595) the role of protein
environment can be essentially basic catalysis. Residues acting as
a general base in the vicinity of the chromophore may be crucial
for red chromophore formation. Notably, our data allow us to
propose a new mechanism of z2FP574 maturation. The fully
mature z2FP574 carries DsRed-type chromophore formed by
Asp65-Tyr66-Gly67 in which Asp65 is decarboxylated.25,26 It was
demonstrated that z2FP574 maturation proceeds through the
GFP-like green intermediate and Asp65 is essential for this
green-to-red conversion.9 To explain the observed interconnec-
tions between formation of acylimine and Asp65 decarboxyla-
tion, Pakhomov and Martynov suggested a “coupled oxidationꢀ
decarboxylation reaction”.9 However, in our opinion, green-to-
red maturation of z2FP574 evidently consists of at least two
consequent irreversible reactions; thus, decarboxylation cannot
drive the preceding oxidation. In the present paper we show that
oxidation at position 65 requires deprotonation of CR. There-
fore, we suggest that the Asp65 side chain carboxylate can be a
base facilitating proton abstraction from CR-65 (Figure 7).
Further oxidation results in β-imino carboxylic acid, which is
known to readily undergo decarboxylation. Interestingly, muta-
tion of D65E in z2FP574 results in only partial inhibition of
green-to-red maturation and complete inhibition of decarboxyla-
tion,9 which is consistent with our model but inconsistent with
the “coupled oxidationꢀdecarboxylation reaction”.
There are several examples in the literature where introduction
of the COOꢀ group (which can act as a general base) in the
vicinity of CR-65 resulted in formation of a red chromophore. A
single substitution Asn65Asp converted purely green FP zFP506
into a dual-color form where roughly one-third of the protein
becomes red fluorescent.9 Also, it was demonstrated that only
Glu and Asp residues at position 65 lead to the appearance of red
fluorescence in yellow FP zFP538.27
We have found a novel imide-substituted chromophore aris-
ing from 4-electron oxidation of the glycine-derived GFP-like
2786
dx.doi.org/10.1021/jo200150b |J. Org. Chem. 2011, 76, 2782–2791

The Journal of Organic Chemistry
ARTICLE
Figure 7. Plausible mechanism of chromophore maturation in z2FP574.
predecessor. In this regard the Gly-Tyr-Gly chromophore-form-
ing sequence is promising as it could produce the novel type of
red fluorescent proteins possessing an imide-derived chromo-
phore. This chromophore is not yet found in known fluorescent
proteins. Its extended conjugated π-system together with the
electron-withdrawing effect of imide substituent make it strongly
red-shifted putting it together with known chromophores of
DsRed, Kaede, or asFP595 proteins.6,28 Its formation within the
protein could presumably proceed under the same conditions as
common maturation of DsRed-like chromophores. Fine tuning
of the basic residues near the R-position of Gly65 may be
required to enable this pathway.
Our results shed light on the evolution of fluorescent proteins.
It was recognized that the color diversity of GFP-like proteins
originated independently within different lineages.29 In particu-
lar, multiple independent appearance of RFPs and chromopro-
teins with the DsRed-like chromophore provides an intriguing
example of convergent evolution at the molecular level. Ease of
autoxidation at position 65 adjacent to the GFP-like chromo-
phore helps to explain this “mystery”, showing that green-to-red
maturation is not as complex a process as was previously thought.
Finally, unexpectedly high reactivity of CR-65 toward oxygen
brings up a paradoxical question: Why not all fluorescent
proteins are red? Indeed, all green FPs contain GFP-like chro-
mophore with a residue at position 65, which is prone to
oxidation. One would expect at least slow conversion of GFPs
into the red state, especially when position 65 is occupied by a
small residue (e.g., Ser65 in A. victoria GFP, Gly65 in copepod29
and lancelet30 green FPs). However, it is known that even
prolonged (for years) storage of green FPs does not result in
the appearance of red fluorescence. It seems that only the
absence of strong basic groups around position 65 ensures the
stability of green chromophore in GFPs. Thus, a novel approach
to generate RFPs from GFPs—introduction of basic residues in
proximity to position 65—is emerging. It might be an important
step toward the rational design of new fluorescent proteins for
the needs of particular applications.
SOCl2 (0.1 mol, 7.8 mL) was added dropwise to the stirred suspen-
sion of 2-amino-3-(4-(benzyloxy)phenyl)-3-hydroxypropanoic acid
(47.3 mmol, 13.6 g) in 80 mL of MeOH at ꢀ10 °C. The resulting
yellowish solution was refluxed for 2 h. The solvent was evaporated and
the residue was suspended in 200 mL of H2O and thoroughly washed
with EtOAc to remove colored byproduct. The suspension was covered
with 150 mL of EtOAc and 100 mL of 10% aqueous Na2CO3 was added
with stirring. The aqueous phase was then extracted twice with 60 mL of
EtOAc. The combined extracts were washed with brine, dried over
Na2SO4, and evaporated. The residue was subjected to the column
chromatography (CHCl3ꢀEtOH 10:1) to give 2 (11.1 g, 78%, 3.2:1
mixture of rel-(2S,3R) and rel-(2S,3S) isomers by NMR) as a colorless
waxy solid. Analytical data were consistent with the literature.32 No
separation of isomers is needed for the further steps.
2-(2-Acetamidoacetamido)-3-(4-(benzyloxy)phenyl)-3-
hydroxy-N-methylpropanamide (2a).
’ EXPERIMENTAL SECTION
1
A 1 M solution of DCC in THF (12 mmol, 12 mL) was added at 0 °C
to the stirred solution of acetylglycine (12 mmol, 1.40 g), NEtiPr2
(12 mmol, 2.09 mL), and HOBt (12 mmol, 1.62 g) in 30 mL of DMF.
After the mixture was stirred at ambient temperature for 1 h a solution of
2 (11 mmol, 3.31 g) in 10 mL of DMF was added and the mixture was
left overnight. After addition of 1 mL of AcOH the mixture was stirred
for 20 min and filtered, then the filtrate was evaporated in vacuo. The
residue was dissolved in 120 mL of EtOAc, washed with water, then 0.5
M HCl, saturated NaHCO3, and brine, and finally dried over Na2SO4.
After evaporation of EtOAc the crude methyl ester (almost pure by TLC
in EtOAcꢀEtOH 2:1) was redissolved in MeCN (60 mL) and aqueous
MeNH2 (40%, 60 mmol, 5.17 mL) was added. The mixture was warmed
Assignments of H and 13C NMR peaks were made based on a
combination of COSY, HSQC, and HMBC spectra.
2-Amino-3-(4-(benzyloxy)phenyl)-3-hydroxypropanoic acid,22 3,3-
dimethyl-2-(N-acetylamino)butanoic acid,31 and 3-methyl-2-ethyl-
5-(4-hydroxybenzylidene)dihydroimidazole-5-one6 (10) were prepared
according to the known procedures.
Chromophores 8 and 11 were identical by Rf, UVꢀvis, and 1H NMR
to the samples prepared independently according to the known
procedures.6,12
Methyl 2-Amino-3-(4-(benzyloxy)phenyl)-3-hydroxypro-
panoate (1).
2787
dx.doi.org/10.1021/jo200150b |J. Org. Chem. 2011, 76, 2782–2791

The Journal of Organic Chemistry
ARTICLE
to 40 °C until the reaction was complete by TLC (ca. 3 h). During this
time a white precipitate is formed that was found to be pure 2a (3.03 g,
69% yield for two steps). The mother solution was evaporated and the
oily residue was subjected to column chromatography (EtOAcꢀEtOH
3:1) to give an additional 0.76 g (19%) of 2a. Total yield: 3.79 g (88% for
two steps).
J = 8.6 Hz, H5), 7.21 (1H, q, J = 4.6 Hz, H14), 7.27 (2H, d, J = 8.6 Hz,
H6), 7.30ꢀ7.43 (5H, m, H1, H2, H3), 7.73 (1H, d, J = 6.8 Hz, H12), 7.87
(1H, d, J = 8.8 Hz, H10).
HRMS (ESI) calcd for C25H34N3O5 456.2499, found 456.2475.
2-(2-Acetamidoacetamido)-1-(4-hydroxyphenyl)-3-(methyl-
amino)-3-oxopropyl Propionate (3a).
1H NMR (800 MHz, DMSO-d6, major isomer/rotamer): δ 1.84
(s, 3H, H13), 2.58 (3H, d, J = 4.7 Hz, H15), 3.60 (1H, dd, J = 5.6 and
16.4 Hz, H11), 3.77 (1H, dd, J = 5.6 and 16.4 Hz, H11), 4.27 (1H, dd,
J = 3.2 and 8.8 Hz, H9), 5.02 (1H, t, J = 3.2 and 4.6 Hz, H7), 5.06 (2H,
s, H4), 5.60 (1H, d, J = 4.6 Hz, H8), 6.91 (2H, d, J = 8.6 Hz, H5), 7.25
(2H, d, 8.6 Hz, H6), 7.31ꢀ7.44 (5H, 3 m, H1, H2, H3), 7.67 (1H, q,
J = 4.6 Hz, H14), 7.74 (1H, d, J = 8.8 Hz, H10), 8.07 (1H, t, J =
5.6 Hz, H12).
HRMS (ESI) calcd for C21H26N3O5 400.1873, found 400.1844.
2-(2-Acetamido-3-phenylpropanamido)-3-(4-(benzyloxy)-
phenyl)-3-hydroxy-N-methylpropanamide (2b).
A mixture of 2a (3.45 mmol, 1.38 g, mixture of isomers), propionic
anhydride (15 mL), and ZnCl2 (1 mmol, 136 mg) was heated to 90 °C
with stirring until the solution became clear and no starting material
was detected by TLC (total time approximately 4 h). The excess
propionic anhydride and propionic acid were evaporated in vacuo at
70 °C and the crude product was dissolved in 50 mL of EtOH and
filtered through a pad of SiO2. To that solution were added 10% Pd/C
(200 mg) and AcOH (5 mL) and the resulting mixture was hydro-
genated at ambient temperature for several hours. When the reaction
was complete by TLC (ca. 4 h) the mixture was filtered through the
paper and evaporated. The product was purified by the column
chromatography (CHCl3ꢀEtOH 85:15). Yield: 0.91 g (72% for two
steps).
1H NMR (700 MHz, D2O, major isomer/rotamer): δ 1.24 (3H, t, J =
7.13 Hz, H11), 2.18 (3H, s, H9), 3.08 (3H, s, H13), 3.71 (2H, q, J = 7.13
Hz, H10), 4.34 (2H, s, H7), 6.92 (2H, d, J = 8.7 Hz, H2), 6.95 (1H, s, H4),
7.86 (2H, d, J = 8.7 Hz, H3).
This compound was prepared similarly to 2a. The product was
purified by the column chromatography (CHCl3ꢀEtOH 85:15). Yield:
68% (2 steps).
2-(2-Acetamido-3-phenylpropanamido)-1-(4-hydroxy-
phenyl)-3-(methylamino)-3-oxopropyl Propionate (3b).
1H NMR (700 MHz, DMSO-d6, major isomer/rotamer): δ 1.72 (3H,
s, H13), 2.56 (3H, d, J = 4.7 Hz, H15), 2.66 (1H, dd, J = 10.3 and 13.9 Hz,
H16), 2.87 (1H, dd, J = 4.9 and 13.9 Hz, H16), 4.24 (1H, dd, J = 2.3 and
8.6 Hz, H9), 4.59 (1H, m, H11), 5.05 (3H, m, H4 and H7), 5.61 (1H, d,
J = 4.4 Hz, H8), 6.89 (2H, d, J = 8.8 Hz, H5), 7.05 (1H, d, J = 7.1 Hz, H),
7.18ꢀ7.41 (5H, m, H1, H2, and H3), 7.22 (2H, d, J = 8.8 Hz, H6),
7.53 (1H, q, J = 4.7 Hz, H14), 7.65 (1H, d, J = 8.6 Hz, H10), 8.09 (1H, d,
J = 8.3 Hz, H12).
HRMS (ESI) calcd for C28H32N3O5 490.2342, found 490.2325.
2-Acetamido-N-(1-(4-(benzyloxy)phenyl)-1-hydroxy-3-
(methylamino)-3-oxopropan-2-yl)-3,3-dimethylbutana-
mide (2c).
AcBr (2.0 mmol, 0.15 mL in 1.5 mL of Ac2O) was added to the
solution of 2b (2.0 mmol, 1.0 g) in CH2Cl2 (10 mL). After 1 h (100%
conversion by TLC CHCl3ꢀEtOH 9:1) the reaction mixture was
poured into EtOAc (50 mL), washed with saturated NaHCO3
solution and brine, and dried over Na2SO4. After evaporation in
vacuo the residue (pure acetylation product by TLC) was resus-
pended in 30 mL of MeOH. To this solution were added 150 mg of
Pd/C, 100 μL of TFA, and the resulting mixture was hydrogenated at
ambient pressure until 100% conversion of the substrate by TLC
(CHCl3ꢀEtOH 7:1). The clear solution was filtered through the
paper and evaporated to dryness. The product was purified by the
column chromatography (CHCl3ꢀEtOH 8:1). Yield: 0.65 g (74% for
two steps).
1H NMR (800 MHz, DMSO-d6, major isomer/rotamer): δ 1.91 and
2.05 (6H, 2s, H11 and H16), 2.63 (3H, d, J = 4.9 Hz, H7), 2.85 and 3.00
(2H, 2 m, H12), 4.63 (m, H9), 4.74 (1H, dd, J = 6.1 and 8.8 Hz, H5), 6.15
(1H, d, J = 6.1 Hz, H4), 6.71ꢀ7.26 (8H, m, H3, H10, H13, H14, and H15),
6.74 (2H, d, J = 8.6 Hz, H2), 6.84 (1H, q, J = 4.9 Hz, H6), 7.09 (1H, d, J =
8.8 Hz, H8).
This compound was prepared similarly to 2a, except the mixture was
stirred for 2 h before addition of amine and for 24 h after that. Product
was purified by the column chromatography (CHCl3ꢀEtOH 10:1).
Yield: 73% (for two steps, mixture of isomers).
1H NMR (700 MHz, CDCl3, major isomer/rotamer): δ 0.67 (9H, s,
H16), 1.89 (3H, s, H13), 2.58 (3H, d, J = 4.6 Hz, H15), 4.05 (1H, d, J = 6.8
Hz, H11), 4.30 (1H, dd, J = 2.9 and 8.8 Hz, H9), 5.07 (2 H, m, H4), 5.16
(1H, dd, 2.9 and 5.3 Hz, H7), 5.51 (1H, d, J = 5.3 Hz, H8), 6.87 (2H, d,
2-(2-Acetamido-3,3-dimethylbutanamido)-1-(4-(benzyl-
oxy)phenyl)-3-(methylamino)-3-oxopropyl Propionate (3c).
2788
dx.doi.org/10.1021/jo200150b |J. Org. Chem. 2011, 76, 2782–2791

The Journal of Organic Chemistry
ARTICLE
13
HSQC 1Hꢀ C (800 MHz, DMSO-d6, Z-isomer): 22.7 (H8), 26.8
(H5), 37.7 (H9), 48.0 (H6), 116.1 (H2), 126.6 (H12), 127.8 (H4),
128.4ꢀ129.7 (H11ꢀH12), 135.0 (H3).
HRMS (ESI) calcd for C21H22N3O3 364.1661, found 364.1683.
(Z)-N-(1-(4-(4-Hydroxybenzylidene)-1-methyl-5-oxo-4,5-di-
hydro-1H-imidazol-2-yl)-2,2-dimethylpropyl)acetamide (5c).
1H NMR (700 MHz, CDCl3, major isomer/rotamer): δ 0.86 (9H, s,
H15), 0.98 (3H, t, J = 7.5 Hz, H9), 1.87 (3H, s, H14), 2.27 (2H, dq, J = 2.4
and 7.5 Hz, H8), 2.42 (3H, d, J = 4.7 Hz, H17), 4.33 (1H, d, J = 9.5 Hz,
H12), 4.70 (1H, dd, J = 6.9 and 8.8 Hz, H10), 5.06 (2H, s, H4), 6.85 (1H,
d, J = 6.9 Hz, H7), 6.92 (2H, d, J = 8.8 Hz, H5), 7.19 (2H, d, J = 8.8 Hz,
H6), 7.33ꢀ7.43 (5H, m, H1, H2, H3), 7.73 (1H, d, J = 9.5 Hz, H13), 7.78
(1H, q, J = 4.7 Hz, H16), 8.07 (1H, d, J = 8.8 Hz, H11).
This compound was prepared similarly to 5a, but the temperature was
140 °C and the reaction was clearly stepwise by TLC with the chromo-
phore formed at the second step.
1H NMR (700 MHz, CDCl3, Z-isomer): δ 1.10 (9H, s, H9), 2.08
(3H, s, H8), 3.26 (3H, s, H5), 4.94 (1H, d, J = 9.25 Hz, H6), 6.66 (1H, d,
J = 9.25 Hz, H7), 6.86 (2H, d, J = 8.6 Hz, H2), 7.12 (1H, s, H4), 7.99 (2H,
J = 8.6 Hz, H3), 8.26 (1H, br, H1).
ESI-MS: m/z 512 ([M þ H]þ), 438 ([M ꢀ C2H5CO2H þ H]þ),
534 ([M þ Na]þ).
N-((4-(4-Hydroxybenzylidene)-1-methyl-5-oxo-4,5-dihy-
dro-1H-imidazol-2-yl)methyl)acetamide (5a).
HRMS (ESI) calcd for C18H24N3O3 330.1818, found 330.1826.
General Procedure for Autoxidation of 5aꢀc. Base (triethyl-
amine, DBU, AcONa, or Cs2CO3) was added to a 0.1 M solution of 5aꢀc
in CH2Cl2, THF, or DMF and oxygen gas was slowly bubbled through the
reaction mixture at constant temperature. Formation of products was
determined by UVꢀvis, NMR spectroscopy, and TLC.
N-(Hydroxy(4-(4-hydroxybenzylidene)-1-methyl-5-oxo-
4,5-dihydro-1H-imidazol-2-yl)methyl)acetamide (6).
4a (0.60 mmol, 220 mg), K2CO3 (1.28 mmol, 176 mg), and DMF
(20 mL) together with a stirring bar were placed into the flask equipped
with a septum. Argon was bubbled through the solution via needle for 15
min. The mixture was heated to 110 °C and stirred at this temperature
for 40 min. After cooling aqueous NH4Cl (1.5 mmol, 80 mg, 5 mL of
H2O) was added and the resulting orange solution was evaporated in
vacuo at 60 °C (the color changes to yellow as neutralization of
chromophore occurs). The residue was extracted with EtOAcꢀEtOH
(20:1) and the extract was subjected to column chromatography
(EtOAcꢀEtOH 15:1). Yield: 121 mg (74%).
A solution of 5a in wet THF containing 10 mM NEt3 is oxygenated
(see General Procedure) at 30 °C for ca. 45 min. When the consumption
of 5a approaches 90% by TLC the mixture is evaporated in vacuo and
subjected to column chromatography (EtOAcꢀEtOH 10:1). Hydro-
xyamide 6 and imide 7 are formed in these conditions in ca. 40% and
25% yields, respectively.
1H NMR (800 MHz, DMSO-d6, Z-isomer): δ 1.94 (3H, s, H8), 3.08
(3H, s, H5), 4.32 (2H, d, J = 5.6 Hz, H6), 6.85 (2H, d, J = 8.8 Hz, H2),
6.99 (1H, s, H4), 8.11 (2H, d, J = 8.8 Hz, H3), 8.41 (1H, t, J = 5.6 Hz, H7),
10.16 (1H, br, H1).
1H NMR (800 MHz, DMSO-d6, Z-isomer): δ 1.95 (3H, s, H8), 3.14
(3H, s, H5), 6.19 (1H, dd, J = 6.1 and 8.6 Hz, H6), 6.78 (1H, d, 6.1 Hz,
H9), 6.84 (2H, d, J = 8.8 Hz, H2), 7.06 (1H, s, H4), 8.15 (2H, d, J = 8.8
Hz, H3), 8.81 (1H, d, J = 8.6 Hz, H7), 10.19 (1H, br, H1).
13
HSQC 1Hꢀ C (800 MHz, DMSO-d6): 23.4 (H8), 26.5 (H5), 37.5
(H6), 116.0 (H2), 127.3 (H4), 134.7 (H3).
15
HSQC 1Hꢀ N (800 MHz, DMSO-d6): cross-peak at 8.41 (H7).
13
HSQC 1Hꢀ C (800 MHz, DMSO-d6): 23.4 (H8), 27.3 (H5), 68.9
HRMS (ESI) calcd for C14H16N3O3 274.1192, found 274.1167.
(Z)-N-(1-(4-(4-Hydroxybenzylidene)-1-methyl-5-oxo-4,5-
dihydro-1H-imidazol-2-yl)-2-phenylethyl)acetamide (5b).
(H6), 116.0 (H2), 128.7 (H4), 135.2 (H3).
15
HSQC 1Hꢀ N (800 MHz, DMSO-d6): cross-peak at 8.82 (H7).
HRMS (ESI) calcd for C14H16N3O4 290.1141, found 290.1125.
N-Acetyl-4-(4-hydroxybenzylidene)-1-methyl-5-oxo-4,5-
dihydro-1H-imidazole-2-carboxamide (7).
1H NMR (700 MHz, DMSO-d6, Z-isomer): δ 1.78 (3H, s, H8), 2.97
(3H, s, H5), 3.07 (1H, dd, J = 8.6 and 13.7 Hz, H9), 3.29 (1H, dd, J = 6.4
and 13.9 Hz, H9), 5.09 (1H, ddd, J = 6.4, 8.3, and 8.6 Hz, H6), 6.86 (2H,
d, J = 8.8 Hz, H2), 7.00 (1H, s, H4), 7.19ꢀ7.31 (5H, m, H10, H11, and
H12), 8.13 (2H, d, J = 8.8 Hz, H3), 8.53 (1H, d, J = 8.6 Hz, H7), 10.17
(1H, s, H1).
Imide 7 is formed along with hydroxyamide 6 during autoxidation of
5a in 25% yield (see above).
Also, imide 7 can be obtained by oxidation of 5a or 6 with selenium
dioxide by using the following procedure. Selenium dioxide (0.6 mmol,
2789
dx.doi.org/10.1021/jo200150b |J. Org. Chem. 2011, 76, 2782–2791

The Journal of Organic Chemistry
ARTICLE
67 mg) was added to a stirred solution of 5a or 6 (0.40 mmol) in dioxane
(4 mL). After the mixture was stirred at room temperature for 40 min the
solvent was evaporated and the product was purified by column chroma-
tography (EtOAc). Yields are 75% and 60% from 5a and 6, respectively.
1H NMR (700 MHz, CDCl3): δ 2.61 (3H, s, H7), 3.56 (3H, s, H5),
6.96 (2H, d, J = 8.7 Hz, H2), 7.45 (1H, s, H4), 8.09 (2H, J = 8.7 Hz, H3),
9.67 (1H, br, H1).
material is available free of charge via the Internet at http://pubs.
acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: ivyamp@gmail.com.
13C NMR (200 MHz, DMSO-d6): δ 25.9, 28.6, 116.7, 125.2, 134.4,
134.8, 136.6, 152.0, 158.7, 162.0, 169.9, 171.5.
HSQC 1Hꢀ C (700 MHz, DMSO-d6): 25.9 (H7), 28.6 (H5), 116.7
13
’ ACKNOWLEDGMENT
(H2), 134.8 (H4), 136.7 (H3).
This work was supported by the Molecular and Cell Biology
program of the Russian Academy of Sciences, grant 16.740.11.
0367 from the Ministry of Education and Science of the Russian
Federation, and Russian Foundation for Basic Research grant 11-
04-01765-a.
HRMS (ESI) calcd for C14H14N3O4 288.0984, found 288.0983.
N-(1-(4-(4-Hydroxybenzylidene)-1-methyl-5-oxo-4,5-di-
hydro-1H-imidazole-2-yl)-2-phenylvinyl)acetamide (9).
’ REFERENCES
(1) Chudakov, D. M.; Lukyanov, S.; Lukyanov, K. A. Trends Biotechnol.
2005, 23, 605.
(2) Pakhomov, A. A.; Martynov, V. I. Chem. Biol. 2008, 15, 755.
(3) Gross, L. A.; Baird, G. S.; Hoffman, R. C.; Baldridge, K. K.; Tsien,
R. Y. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 11990.
Dry dioxigen gas was passed through the mixture of 5b (0.30 mmol,
110 mg), Cs2CO3 (0.2mmol, 65mg), andDMF(5mL) at115°Cfor45min
(almost total conversion of 5b). The mixture was poured into EtOAcꢀ0.1 M
HCl, then the aqueous layer was extracted with additional portions of EtOAc.
The combined extracts were washed with water and brine and dried over
Na2SO4. After evaporation the only colored product was purified by chroma-
tography twice (column chromatography: hexaneꢀEtOAc 1:3, then preparative
TLC: pure EtOAc). Yield: 21 mg (20%) as a mixture of inseparable isomers.
1H NMR (800 MHz, DMSO-d6, major isomer/rotamer): δ 2.06
(3 H, s, H7), 3.07 (3H, s, H5), 6.84 (2H, d, J = 8.8 Hz, H2), 6.90 (1H, s,
H8), 7.01 (1H, s, H4), 7.40ꢀ7.66 (5H, m, H9, H10, and H11), 8.10 (2H,
d, J = 8.8 Hz, H3), 10.06 (1H, br, H1).
(4) Yarbrough, D.; Wachter, R. M.; Kallio, K.; Matz, M. V.; Re-
mington, S. J. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 462.
(5) Quillin, M. L.; Anstrom, D. M.; Shu, X.; O’Leary, S.; Kallio, K.;
Chudakov, D. M.; Remington, S. J. Biochemistry 2005, 44, 5774.
(6) Yampolsky, I. V.; Remington, S. J.; Martynov, V. I.; Potapov, V. K.;
Lukyanov, S.; Lukyanov, K. A. Biochemistry 2005, 44, 5788.
(7) Subach, O. M.; Malashkevich, V. N.; Zencheck, W. D.; Morozova,
K. S.; Piatkevich, K. D.; Almo, S. C.; Verkhusha, V. V. Chem. Biol. 2010,17,333.
(8) Strack, R. L.; Strongin, D. E.; Mets, L.; Glick, B. S.; Keenan, R. J.
J. Am. Chem. Soc. 2010, 132, 8496.
(9) Pakhomov, A. A.; Martynov, V. I. Biochemistry 2007, 46, 11528.
(10) Pakhomov, A.; Tretyakova, Y.; Martynov, V. Russ. J. Bioorg.
Chem. 2010, 36, 109.
(11) Kojima, S.; Ohkawa, H.; Hirano, T.; Maki, S.; Niwa, H.; Ohashi, M.;
Inouye, S.; Tsuji, F. I. Tetrahedron Lett. 1998, 39, 5239.
(12) Yampolsky, I. V.; Kislukhin, A. A.; Amatov, T. T.; Shcherbo, D.;
Potapov, V. K.; Lukyanov, S.; Lukyanov, K. A. Bioorg. Chem. 2008,
36, 96.
1
13
Characteristic HSQC Hꢀ C cross-peaks: 2.06ꢀ22.8, 3.07ꢀ28.1,
6.90ꢀ127.2, 6.84ꢀ116.2, 7.01ꢀ127.4, 8.10ꢀ134.8.
1
13
Characteristic HMBC Hꢀ C cross-peaks: 2.06ꢀ169.7; 3.07ꢀ160.5
and 170.4; 6.90ꢀ130.0 and 160.5; 8.10ꢀ116.2, 127.4, 134.7, and 160.1.
HRMS (ESI) calcd for C21H20N3O4 362.1505, found 362.1492.
5-(4-Hydroxybenzylidene)-3-methylimidazolidine-2,4-
dione (12).
(13) Bulina, M.; Chudakov, D.; Mudrik, N.; Lukyanov, K. BMC
Biochem. 2002, 3, 7.
(14) Ivashkin, P.; Yampolsky, I.; Lukyanov, K. Russ. J. Bioorg. Chem.
2009, 35, 652.
(15) Devasia, G. M. Tetrahedron Lett. 1976, 17, 571.
(16) Ekeley, J. B.; Ronzio, A. R. J. Am. Chem. Soc. 1935, 57, 1353.
(17) Kidwai, A. R.; Devasia, G. M. J. Org. Chem. 1962, 27, 4527.
(18) Wu, L.; Burgess, K. J. Am. Chem. Soc. 2008, 130, 4089.
(19) Oumouch, S.; Bourotte, M.; Schmitt, M.; Bourguignon, J.-J.
Synthesis 2005, 25.
(20) Boyd, V. L.; Bozzini, M.; Guga, P. J.; DeFranco, R. J.; Yuan, P.-
M.; Loudon, G. M.; Nguyen, D. J. Org. Chem. 1995, 60, 2581.
(21) Shin, C.-g.; Yonezawa, Y.; Obara, T.; Nishio, H. Bull. Chem. Soc.
Jpn. 1988, 61, 885.
(22) Bolhofer, W. A. J. Am. Chem. Soc. 1954, 76, 1322.
(23) Shigematsu, N.; Kayakiri, N.; Okada, S.; Tanaka, H. Bull. Chem.
Soc. Jpn. 1997, 45, 236.
(24) Yampolsky, I. V.; Balashova, T. A.; Lukyanov, K. A. Biochemistry
2009, 48, 8077.
(25) Pletneva, N.; Pletnev, S.; Tikhonova, T.; Popov, V.; Martynov, V.;
Pletnev, V. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2006, 62, 527.
(26) Pletneva, N.; Pletnev, V.; Tikhonova, T.; Pakhomov, A. A.;
Popov, V.; Martynov, V. I.; Wlodawer, A.; Dauter, Z.; Pletnev, S. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 2007, 63, 1082.
(27) Remington, S. J.; Wachter, R. M.; Yarbrough, D. K.; Branchaud, B.;
Anderson, D. C.; Kallio, K.; Lukyanov, K. A. Biochemistry 2005, 44, 202.
Et-HBI 10 was added to a saturated solution of Cs2CO3 in DMF and
the resulting mixture was oxygenated at 80 °C for 3 h. The mixture was
partitioned between EtOAc and 0.1 M HCl, then the aqueous layer was
extracted twice with EtOAc. The combined extracts were washed with
water and brine, dried over Na2SO4, and evaporated. Column chroma-
tography (CHCl3ꢀEtOH 10:1) gave 113 (21%) and 12 (48%)
1H NMR (700 MHz, CDCl3): δ 2.96 (3H, s, H5), 6.47 (1H, s, H4),
6.81 (2H, d, J = 8.5 Hz, H2), 7.50 (2H, d, J = 8.5 Hz, H3), 9.86 (1H, s, H6),
10.52 (1H, br, H1).
HRMS (ESI) calcd for C11H11N2O3 219.0770, found 219.0756.
’ ASSOCIATED CONTENT
S
Supporting Information. Autoxidation study of 2-ethyl-
b
substituted GFP-chromophore, 1D and 2D NMR spectra. This
2790
dx.doi.org/10.1021/jo200150b |J. Org. Chem. 2011, 76, 2782–2791

The Journal of Organic Chemistry
ARTICLE
(28) Mizuno, H.;Mal, T. K.;Tong, K. I.;Ando, R.;Furuta, T.;Ikura, M.;
Miyawaki, A. Mol. Cell 2003, 12, 1051.
(29) Shagin, D. A.; Barsova, E. V.; Yanushevich, Y. G.; Fradkov, A. F.;
Lukyanov, K. A.; Labas, Y. A.; Semenova, T. N.; Ugalde, J. A.; Meyers, A.;
Nunez, J. M.; Widder, E. A.; Lukyanov, S. A.; Matz, M. V. Mol. Biol. Evol.
2004, 21, 841.
(30) Bomati, E. K.; Manning, G.; Deheyn, D. D. BMC Evol Biol.
2009, 9, 77.
(31) Jaeger, D. A.; Broadhurst, M. D.; Cram, D. J. J. Am. Chem. Soc.
1979, 101, 717.
(32) Boger, D. L.; Zhou, J.; Borzilleri, R. M.; Nukui, S.; Castle, S. L.
J. Org. Chem. 1997, 62, 2054.
2791
dx.doi.org/10.1021/jo200150b |J. Org. Chem. 2011, 76, 2782–2791