Blue fluorescent amino acids as in vivo building blocks for proteins

Article abstract of DOI:10.1002/cbic.200900651

In vivo expression of colored proteins without post-translational modification or chemical functionalization is highly desired for protein studies and cell biology. Cell-permeable tryptophan analogues, such as azatryptophans, have proved to be almost ideal isosteric substitutes for natural tryptophan in cellular proteins. Their unique spectral features, such as markedly redshifted fluorescence, are transmitted into protein structures upon incorporation. Among the azaindoles under study (2-, 4-, 5-, 6-, and 7-azaindole) 4-azaindole has exhibited the largest Stokes shift (～130 nm) in steady-state fluorescence measurements. It is also highly biocompatible and as 4-azatryptophan it can be translated into target protein sequences. However, its quantum yield and fluorescence intensity are still significantly lower when compared with natural indole/tryptophan. Since azatryptophans are hydrophilic, their presence in the hydrophobic core of proteins could be harmful. In order to overcome these limitations we have performed nitrogen methylation of azaindoles and generated mono- and dimethylated azaindoles. Some of these methyl derivatives retain the pronounced red shift present in the parent 4-azaindole, but with much higher fluorescence intensity (reaching the level of indole/tryptophan). Therefore, the blue fluorescence of azaindole-containing proteins could be further enhanced by the use of methylated analogues. Further substitution of any azaindole ring with either endo- or exocyclic nitrogen will not yield a spectral fluorescence maximum shift beyond 450 nm under steady-state conditions in the physiological milieu. However, green fluorescence is a special feature of tautomeric species of azaindoles in various nonaqueous solvents. Thus, the design or evolution of the protein interior combined with the incorporation of these azaindoles might lead to the generation of specific chromophore microenvironments that facilitate tautomeric or protonated/deprotoned states associated with green fluorescence.

Full text of DOI:10.1002/cbic.200900651

DOI: 10.1002/cbic.200900651
Blue Fluorescent Amino Acids as In Vivo Building Blocks
for Proteins
Lars Merkel,[a] Michael G. Hoesl,[a] Marcel Albrecht,[b] Andreas Schmidt,[b] and
Nediljko Budisa*[a]
In vivo expression of colored proteins without post-translation-
al modification or chemical functionalization is highly desired
for protein studies and cell biology. Cell-permeable tryptophan
analogues, such as azatryptophans, have proved to be almost
ideal isosteric substitutes for natural tryptophan in cellular pro-
teins. Their unique spectral features, such as markedly red-
shifted fluorescence, are transmitted into protein structures
upon incorporation. Among the azaindoles under study (2-, 4-,
5-, 6-, and 7-azaindole) 4-azaindole has exhibited the largest
Stokes shift (~130 nm) in steady-state fluorescence measure-
ments. It is also highly biocompatible and as 4-azatryptophan
it can be translated into target protein sequences. However, its
quantum yield and fluorescence intensity are still significantly
lower when compared with natural indole/tryptophan. Since
azatryptophans are hydrophilic, their presence in the hydro-
phobic core of proteins could be harmful. In order to over-
come these limitations we have performed nitrogen methyla-
tion of azaindoles and generated mono- and dimethylated
azaindoles. Some of these methyl derivatives retain the pro-
nounced red shift present in the parent 4-azaindole, but with
much higher fluorescence intensity (reaching the level of
indole/tryptophan). Therefore, the blue fluorescence of azain-
dole-containing proteins could be further enhanced by the use
of methylated analogues. Further substitution of any azaindole
ring with either endo- or exocyclic nitrogen will not yield a
spectral fluorescence maximum shift beyond 450 nm under
steady-state conditions in the physiological milieu. However,
green fluorescence is a special feature of tautomeric species of
azaindoles in various nonaqueous solvents. Thus, the design or
evolution of the protein interior combined with the incorpora-
tion of these azaindoles might lead to the generation of specif-
ic chromophore microenvironments that facilitate tautomeric
or protonated/deprotoned states associated with green fluo-
rescence.
Introduction
Accurate assessment of protein localization is crucial for com-
plete understanding of its in vivo function. Most native chro-
mophore-free proteins are colorless, which prevents studies of
their location and dynamics deep within living tissues by using
standard fluorescence microscopy. Genetically encoded fluores-
cent tags, such as fluorescent proteins (FPs), are widely used
as visualization tools for such purposes.[1] The green fluores-
cent protein (GFP) from Aquorea victoria and other FPs recently
became the most popular protein localization tools.[2] They
proved to be useful markers to monitor and study protein
turnover, transport, and molecular interactions by using tech-
niques such as FRET, fluorescence lifetime imaging, bimolecular
fluorescence complementation, fluorescence recovery after
photobleaching or photoactivation.[3]
of fluorophores like coumarines or FlAsH, although cell-perme-
able, is restricted since it requires the tailoring of the target se-
quences prior to standard labeling[7] or coupled enzymatic re-
actions.[8] In order to overcome these limitations, our long-
term goal is to develop procedures for the direct cotranslation-
al generation of colored proteins in living cells without the
need for further post-translational modification or chemical
functionalization by externally added reagents. In attempts to
visualize natural chromophore-free proteins and even portions
of the whole proteome it is highly desirable to generate color
without structural and functional perturbations of target pro-
teins.[9] As shown in Scheme 1 and Figure 1 tryptophan ana-
logues, such as 4-azatryptophan (4AW), are very promising
candidates for such minimal chromophores that can be used
in imaging studies. Indeed, they are cell permeable and have
proved to be almost ideal isosteric substitutes for natural tryp-
tophan in cellular proteins. Their unique spectral features, such
However, the principal disadvantage of these genetically en-
coded tags is that their fusion with target proteins might gen-
erate experimental artifacts.[4] For example, fusion constructs
might alter the location or introduce oligomer formation of
the native protein.[5] Therefore, complementary approaches are
often necessary in order to gain a more accurate picture of
particular in vivo protein functionalities. On the other hand, re-
cently developed bioorthogonal transformations require the
presence of a functional group in the protein and an addition
of exogenous synthetic probes.[6] In this way, such ligation
rather mimics a post-translational modification. Even the use
[a] Dr. L. Merkel, M. G. Hoesl, Dr. N. Budisa
Molecular Biotechnology, Max Planck Institute of Biochemistry
Am Klopferspitz 18, 82152 Martinsried (Germany)
Fax: (+49)8578-3557
E-mail: budisa@biochem.mpg.de
[b] M. Albrecht, Prof. Dr. A. Schmidt
Institute of Organic Chemistry, Clausthal University of Technology
Leibnizstr. 6, 38678 Clausthal-Zellerfeld (Germany)
ChemBioChem 2010, 11, 305 – 314
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
305

N. Budisa et al.
Scheme 1. Comparison of fluorescent probes and spectral properties. Commonly used imaging tags and fluorescent probes that are relevant for protein stud-
ies. The illustration should approximate the real size differences of the chemical components. From the biggest to the smallest compound, 1: member of the
GFP family, EGFP; 2: FlAsH (4’,5’-bis(1,3,2-dithioarsolan-2-yl)fluorescein)); 3: representative of the coumarin family, 3-triazol-coumarin; 4: 4-azatryptophan; 5:
tryptophan; and 6: tyrosine. All GFP derived fluorescent proteins as well as FlAsH, require an engineering of the target amino acid sequence. Moreover, GFPs
are relatively large in size, which can cause unpredictable perturbations within the studied system. The FlAsH method is based on the subsequent addition of
arsenic compounds, which are toxic to many cells. In contrast, 4-azatryptophan is structurally very similar to W, that is, represents only a CH!N substitution.
Furthermore, it comprises the largest Stokes shift of all depicted imaging probes. The details of its spectroscopic properties are provided in ref. [10]; R
denotes H2CCH(NH2)COOH.
Figure 1. Molecular structure and fluorescence of free and protein incorporated azaindoles/azatryptophans. Left-hand panel: comparison of the fluorescence
emission profiles of indole and azaindoldes. Middle panel: chemical structures of indole and azaindoles, which are the chromophores of the related trypto-
phans. Ribbon plot of human annexin A5 (side view) with W187 buried in the hydrophobic pocket at the convex side of the molecule. Right-hand panel:
SDS-PAGE profiles of purified annexin A5 and its variants. Proteins were expressed either in the presence of tryptophan, 4-azatryptophan (4AW), 5-azatrypto-
phan (5AW) or 7-azatryptophan (7AW). The upper part shows the gel upon UV exposure and the lower part after subsequent Coomassie blue staining. Note
that simple UV irradiation reveals protein autofluorescence; M: molecular weight marker.
as markedly red-shifted fluorescence, charge conductivity and
pH sensitivity, are transmitted into the protein structure upon
incorporation.[10]
would be biocompatible and well incorporated into the target
protein(s) by the endogenous translational apparatus. Their
chemical structures in the context of an engineered or evolved
surrounding protein matrix endow the resulting biologically
active structures with unique spectral properties. Beside spec-
troscopy, other features gained in this way should be generally
useful for biologically inspired research, innovations and appli-
cations.
In this context, the major challenge for the synthesis chemis-
try of indole combined with protein engineering would be to
identify possible indole-based chromophores suitable for serv-
ing as imaging tools. They should feature a significant shift in
fluorescence towards the red spectral region combined with
improved quantum yields. In an ideal case, such chromophores
306
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 305 – 314

Blue Fluorescent Amino Acids
features a unique principle of photophysics that causes a
Stokes shift of about 100 nm.[20] Similarly, incorporation of
para-aminophenylalanine (pAmF) into GFP yields a Stokes shift
of almost 70 nm.[21–22] Thus, pioneering works on generation of
novel GFP protein classes, offer unprecedented opportunities
to understand and manipulate the relationship between pro-
tein structure and its spectroscopic function. These efforts will
certainly find more recognition and applications in future re-
search.
Tryptophan As a Target for the Engineering of
Protein Fluorescence
Tryptophan residues play crucial roles in protein stability and
folding and numerous cellular functions, such as aromatic in-
teractions with DNA bases. Since W is the main source of fluo-
rescence and absorbance in proteins it was recognized very
early to be a useful intrinsic probe for structure and function
studies of proteins and enzymes. For example, the fluores-
cence maximum of most globular proteins is changed from
330 to 350 nm when the protein is denatured by urea.[11] This
takes place upon solvent exposure of W residues, which are
normally buried in the protein interior. In the latter case their
maxima is close to 330 nm whereas the exposed residue
maxima is near 350 nm. However, these spectral features are
not suited for the majority of current imaging methods. The re-
placement of W residues by other canonical aromatic amino
acids (phenylalanine, tyrosine, histidine) does not improve the
protein’s spectral features and might even perturb its structure
or function due to differences in shape, size and even charge
between W and these amino acids.
Photophysics of 7-Azatryptophan and Its Me-
thylated Derivatives
During the first decades of studies with azatryptophan (AW)
the photophysics of 7-azatryptophan (7AW) and its chromo-
phoric side chain 7-azaindole (7AI) were in the focus of various
researches. This included parallel studies of their properties in
solution as well as studies of their spectral properties in pro-
tein structures. Recently, this research was extended to other
azaindole isomers, such as 4-azaindole (4AI), 5-azaindole (5AI)
and 6-azaindole (6AI). Before testing their suitability for protein
engineering it is essential to understand the basic principles
behind their unique photophysics. All these photophysical
studies were mainly performed with indole and azaindoles to
avoid complications by the charged (zwitterionic) nature of the
related amino acid. For that reason we also focused on the
azaindoles and their derivatives rather than related amino
acids. Nonetheless, due to the better solubility in water we
used W as a reference sample in all measurements, since its
differences to free indole are well documented (e.g., red shift
of 10 nm in fluorescence without changes in intensity).
On the other hand, W’s regular appearance and its low natu-
ral abundance make it an almost irreplaceable reporter and a
very attractive target for engineering in protein science. The
expanded genetic code opens up possibilities for a rational
manipulation of W fluorescence. It exploits the diverse and rich
chemistry of indole that offers numerous analogues/surrogates
with predefined spectral properties.[12] The dominance of W in
spectral properties of proteins is of great advantage in identify-
ing introduced modifications since they are easily observable.
In other words, spectroscopic properties delivered by chromo-
phoric amino acids often emerge as intrinsic features of substi-
tuted target proteins. In this way, several isosteric tryptophan
analogues, such as hydroxyl-, methyl-, amino- and fluorotrypto-
phans, have been incorporated into the sequences of model
proteins endowing them with enhanced or even novel spectral
features.[13–15]
Generally, in the structure of all azaindoles two tautomers
can be evidenced. For 7-azaindole these tautomers are labeled
7AI and 7AI-T (Scheme 2A). In the ground electronic state the
7AI form is energetically favored and is the only thermally ac-
cessible form.[23] Conversely, in the excited state the tautomeric
form 7AI-T is more stable.[23] This phenomenon is known as ex-
cited state proton transfer (ESPT) and leads to a more stable
tautomer after light exposure. The same holds true for the
other azaindoles.[24]
Even before the genetic code was deciphered Pardee et al.
ˇ
as well as Brawerman and Ycas reported experiments on the
incorporation of 7-azatryptophan (7AW) and 2-azatryptophan
(2AW) into the proteome of Escherichia coli and its phage.[16–17]
A decade later Schlesinger reported an alkaline phosphatase to
be the first enzyme fully substituted with azatryptophans in
E. coli cells.[18] Szabo, Ross and their co-workers further extend-
ed this repertoire by introducing 5-hydroxytryptophan as a
useful intrinsic fluorescent probe.[14–15] Its excitation is ~20 nm
red shifted compared to that of W; this enables selective exci-
tation in the range of 315 to 320 nm.[14] On the other hand, 4-
hydroxytryptophan exhibited a blue-shifted absorbance and
fluorescence.[19]
One of the most striking features of 7AI is its extreme sensi-
tivity to the nature of the solvent microenvironment.[25] For ex-
ample, in nonpolar solvents dimerization of 7AI takes place
and besides the usual violet emission maximum (lmaxem
390 nm) a second peak in the green spectral region (lmaxem
~
~
500 nm) emerges.[26] The appearance of the second peak is at-
tributed to the formation of a 7AI dimer and cooperative
double-proton transfer in the excited state (ESDPT). This yields
the tautomeric form (7AI-T) for both dimer molecules
(Scheme 2B).[25–27]
One of the most recent advances represents the incorpora-
tion of aminotryptophan (AmW) analogues that offer the possi-
bility to design protein-based sensors[19] as well as novel
classes of autofluorescent proteins.[20] Incorporation of 4-ami-
notryptophan (4AmW) into cyan fluorescence protein (CFP)
substantially shifts its fluorescence emission to longer wave-
lengths and yields a gold fluorescence protein (GdFP). GdFP
In linear alcohols ESDPT also occurs.[25,27] The formation of a
7AI-C-like cyclic intermediate (Scheme 2C) facilitates the inter-
actions of 7AI with the solvent molecules, and again results in
the more stable tautomeric form.[27] However, in water only a
single and smooth fluorescent band is detectable for 7AI with
a maximum at 386 nm.[27] This observation can be explained
by taking into account that most of the solute molecules are
ChemBioChem 2010, 11, 305 – 314
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
307

N. Budisa et al.
Scheme 2. Schematic representation of 4- and 7-azaindoles with related tautomeric structures as well as the methylated derivatives. Methylation can be used
as a chemical tool to arrest one particular tautomeric state in order to study its optical properties. A) Upon excited state proton transfer (ESPT) of 7-azaindole
(7AI) its tautomer (7AI-T) is formed.[23] B) Excited state double proton transfer (ESDPT) in apolar solvents enables tautomerization after dimerization.[26] C) Com-
pound 7AI and its interactions with water. Note that the 7AI-CT state has a characteristic fluorescence in the green region whereas the protonated form (7AI-
TH+) exhibits a blue fluorescence. On the right the blocked species (7AI-B1 and 7AI-B2) dominating in bulk water are depicted. The B1 represents hydrogen
bonding by two independent solvent molecules whereas B2 resembles a cyclic hydrogen bonding system. In polar solvents (e.g., ethanol) the 7AI-CT state
can be observed.[25] D) Chemical structures of 1-methyl-4-azaindole (4M4AI), 4-methyl-azaindole (4M4AI), 1,4-dimethyl-4-azaindolium iodide (1,4DAI), 1-methyl-
7-azaindole (1M7AI), 7-methyl-7azaindole (7M7AI), 1,7-dimethyl-7-azaindolium iodide (1,7M7AI; iodide as counter ion is excluded from the formulas in case of
the dimethylated derivatives), 4-aminoindole (4AmI) and 4-amino-7-azaindole (4Am7AI). For a detailed discussion see the text.
inappropriately solvated. Only a small portion of molecules are
able to form the cyclic hydrogen-bonded complex 7AI-C with
subsequent ESDPT and formation of 7AI-CT (lmaxem ~500 nm,
Scheme 2C). Unfortunately, the latter species will be rapidly
protonated to 7AI-TH+ (Scheme 2C). This cationic species has
a fluorescence emission maximum of approximately 440 nm.[27]
The majority of 7AI molecules in bulk water establish hydro-
gen bonds with two different water molecules due to the high
geometry of the water molecule itself as well as to the highly
ordered water network in bulk solution.[9] This leads to so-
called blocked species as depicted in Scheme 2C (7AI-B1 and
7AI-B2). There is no rapid conversion of the blocked species to
one of the others on the time scale of their fluorescence life-
times.[27] Thus, in aqueous solvents, 7AW fluorescence is
strongly quenched, which limits its general use as a fluoro-
phore.[28]
In order to arrest the tautomeric form in a stable state in
water, selective methylation of either the pyrrole (N1) or pyri-
dine nitrogen (N7) was performed.[23,27,29] The attachment of a
methyl group to N7 ended in the formation of 7-methyl-7-
azaindole (7M7AI, Scheme 2D). Compound 7M7AI exhibits a
fluorescence emission maximum at 442 nm, whereas at pH
values above 10 the emission maximum is shifted to
510 nm.[27] Thus, high pH values facilitate deprotonation result-
ing in the green fluorescence of 7M7AI (lmax =510 nm). Un-
fortunately, 7M7AI is very unstable[23,29] and unattractive for
spectroscopic research. Namely, it has only a very low fluores-
cence intensity with negligible quantum yields (50-times lower
when compared with 7AI).[27]
Conversely, the methylation of N1 leading to 1M7AI
(Scheme 2D) suppresses the major nonradiative pathways in
the excited state determined by the N1 proton and its
308
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 305 – 314

Blue Fluorescent Amino Acids
interactions with the solvent.[25,27] Although this is reflected in
a relatively long fluorescence lifetime the red shift is rather
moderate (lmaxem =386 nm, Figure 2). In the literature there are
also reports about slightly higher values for the emission maxi-
mum (396–405 nm).[25,27] However, it is reasonable to conclude
that 1M7AI reflects the blocked species of 7AI (Scheme 2C and
D). Nevertheless, because of its improved fluorescence intensi-
ty and quantum yield it is a suitable optical probe. First reports
of its use in a biological context by, for example, incorporation
of 1-methyl-7-azatryptophan into small peptides are already
published.[25]
Interestingly, simultaneous methylation of both N1 and N7
gives 1,7-dimethyl-7-azaindolium iodide (1,7M7AI, Scheme 2D;
see the Experimental Section) a species that emits blue light at
approximately 450 nm. However, fluorescence intensity is even
lower than that of 7AI (Figure 2). Thus, the properties of 1M7AI
(high fluorescence intensity and quantum yield) and 7M7AI
(pronounced red shift) could not be combined in a way that
desirable spectral features are displayed. In addition, the
charged nature of this molecule could cause difficulties for its
possible biological applications. For example, this could be a
major obstacle for efficient cellular uptake through transporter
systems as well as for recognition and tRNA charging reac-
tions.
4-, 5- and 6-Azatryptophan
Only recently photophysical studies of 4AI, 5AI and 6AI have
also been reported.[24,30] Both, experiments and theoretical
studies show unambiguously that the stability of the N1H tau-
tomers and the NnH tautomers (n=4, 5, 6) are similar to that
of 7AI. However, the mechanism of tautomerization of 4AI, 5AI
and 6AI differs from that of 7AI due to the geometric arrange-
ment of the pyridine and pyrrole nitrogens of the respective
azaindole systems.[24] Here, excitation is thought to be subse-
quently followed by Nn (n=4, 5, 6) abstracting a proton of a
solvent molecule. Only after that the deprotonation of the N1
takes place.
The absorbance of 6AI is red shifted to 320 nm (325 nm for
the corresponding tryptophan derivative 6AW).[30] At physio-
logical pH this chromophore is protonated (lmaxem =380 nm)
since its pKa value is 8 whereas above pH 10 fluorescence can
Figure 2. Absorption (left) and emission (right) profiles of tryptophan and related mono- and dimethylated 4- and 7-azaindoles. Upper graphs: the spectral
properties of W and 1-methyl-4-azaindole (1M4AI), 4-methyl-4-azaindole (4M4AI) and 1-methyl-7-azaindole (1M7AI) are depicted. Note the additional absorp-
tion maxima for 4M4AI absent in W or 1M4AI. Furthermore, the emission profiles clearly show a more pronounced red shift and higher intensity for 4M4AI
compared to 1M4AI. Lower graphs: the spectral properties of W and 1,4-dimethyl-4-azaindolium iodide (1,4M4AI) and 1,7-dimethyl-7-azaindolium iodide
(1,7M7AI). Again, the absorption profile of the 4-azaindole derivative shows a second excitation maximum comparable to that of 4M4AI. The emission profile
shows almost no detectable fluorescence for 1,7M7AI whereas 1,4M4AI exhibits a fluorescence profile similar to that of 4M4AI.
ChemBioChem 2010, 11, 305 – 314
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
309

N. Budisa et al.
be detected in the blue spectral region (lmaxem =440 nm). This
is attributed to the excited-state tautomer N6H.[30]
for selective excitation of related proteins in biological sam-
ples. Its fluorescence intensity as well as emission maximum
(lmaxem =414 nm, Figure 2) are comparable to 4AI (Figure 1).
Expectedly, at high pH values 4M4AI’s emission profile is red
shifted to 480 nm.[24]
In comparison with the traditionally used 7AI, the 6AI chro-
mophore has the advantage of a slight red shift in absorption
(30 nm) and much higher fluorescence intensity. In protein sci-
ence such an optical probe is especially interesting for study-
ing protein–protein interactions in the presence of native tryp-
tophan side chains. Its practical use was already shown by in-
corporation and subsequent spectral investigation of calmodu-
lin.[30] It is worth to note, that this chromophore is less biocom-
patible than 7AI. For example, its translation into protein
sequences is not efficient and difficult to reproduce.[10,30]
Similar to the above-mentioned examples, it was found that
in case of 4AI and 5AI excited-state tautomerizaton occurs in
the presence of strong bases (pH>10). Under these basic con-
ditions the chromophore emission maxima shift to 480 and
450 nm, respectively.[24] In neutral aqueous solution both 4AI
and 5AI possess fluorescence emission maxima at 418 and
410 nm (Figure 1).[24] The pronounced red shift and the mark-
edly higher fluorescence intensity of 4AI compared to 7AI and
5AI (Figure 1) makes it a powerful tool for protein fluorescence
design. Recently, we could confirm these assumptions experi-
mentally by converting the colorless protein annexin A5 into
its blue fluorescent counterpart by W!4AW substitution.[10]
In order to investigate closely the tautomeric species of 4AI
we and others[24] methylated 4AI both at N1 and N4. 1-Methyl-
4-azaindole (1M4AI, Scheme 2D) has an absorption profile sim-
ilar to that of the parent compound 4AI, but its fluorescence is
red shifted by approximately 20 nm (lmax =438 nm). Notewor-
thy, the fluorescence intensity is almost doubled and reaches
the level of I/W (Figures 1 and 3). Assuming a similar biocom-
patibility of 4AI and 1M4AI, the incorporation of 1M4AI into
protein sequences could provide structures with a more in-
tense blue fluorescence than already described for 4AI-contain-
ing proteins.[10]
To the best of our knowledge there are no reports about
chemical synthesis of dimethylated 4-azaindole. Therefore, we
chemically
synthesized
1,4-dimethyl-4-azaindole
iodide
(1,4M4AI; see the Experimental Section). As described above,
dimethylation of 7AI almost extinguished its fluorescence
(Figure 2). Conversely, the 1,4M4AI emission spectrum is
almost identical with that of 1M4AI (lmaxem =435 nm). Also,
fluorescence intensities were not significantly changed com-
pared to 1M4AI (Figure 2). Moreover, it features two absorption
bands (290 and 338 nm) suitable for selective excitation of bio-
logical samples. In this way, 1,4M4AI as a chromophore with
potential biological use combines two highly desired features:
selective excitation wavelength (338 nm) and red shifted fluo-
rescence (lmaxem =435 nm, Stokes shift ~100 nm). However,
due to the cationic nature of 1,4M4AI its biocompatibility
should be investigated especially with respect to intracellular
accumulation.
In summary, among all the above described azaindole deriv-
atives 1M4AI seems to be the most promising chromophore
for the engineering of protein fluorescence. Namely, it com-
bines a very pronounced red shift in fluorescence emission
(Stokes shift of 150 nm) and good fluorescence intensities.
Attributes and Perspectives of Indole-Based
Chromophores As Possible Imagining Tools
In any chromophore structure, it is generally expected that an
increase in the size of delocalized p system results in an in-
crease in the excitation and emission wavelength.[31] The best
documented example among natural FPs is DsRed from Disco-
soma striata. Here, the basic chromophore structure of GFP is
further augmented by an extension of the p-bonding system
by additional conjugation through the protein backbone (i.e.,
Ca-N-bond of neighboring residue).[32] Any attempt to use
In contrast, the regioisomer 4-methyl-4-azaindole (4M4AI,
Scheme 2D) exhibits two absorbance maxima at 284 and
332 nm. Although, it has lower intensities compared to I/W
(Figure 2) the second absorption shoulder is especially useful
Figure 3. Absorption (left) and emission profiles (right) of 4-aminoindole (4AmI) and 4-amino-7-azaindole (4Am7AI). It is obvious that amination of 7AI pro-
vides two-fold increase in intensity and a considerable red shift (27 nm). In comparison with monomethylated counterparts this is a rather modest improve-
ment. But 4Am7AI could be an especially interesting candidate for insertion into the chromophores of FPs, such as ECFP.
310
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 305 – 314

Blue Fluorescent Amino Acids
azaindole-based chromophores capable of providing detecta-
ble emission in blue or even green regions, should consider
the possibilities offered by the chemistry of the indole ring.
The goal is to identify chemical substitutions that lead to sig-
nificant spectral red shifts, if possible without necessity for the
expansion of the aromatic system with a third ring. Until now,
the red shift of azaindoles under physiological conditions have
never extended beyond 450 nm.
the fluorescence of GFP is not an intrinsic property of the cy-
clyzed Ser-Tyr-Gly tripeptide since this sequence can be found
in a number of other proteins. However, its cyclization is possi-
ble only in the context of the GFP structure. For this reason,
neither is the Ser-Tyr-Gly tripeptide in the structure of proteins
other than GFP cyclized nor is the tyrosine side chain oxidized.
Thus, the green fluorescence of GFP is a reporter of the inter-
actions between the chromophore and its surrounding protein
matrix.
The substitution of endocyclic ꢀCH groups of indole with ni-
trogen leads to changes in the photophysics of the whole aro-
matic system.[33] Furthermore, similar effects can be achieved
by introducing electron-donating amino and hydroxy groups
at different positions of the indole moiety. Both substitutions
induce an intramolecular charge transfer that is sensitive to pH
changes.[19] The DNA/RNA purine nucleotides adenine and gua-
nine and their analogues (such as 6-aminopurine) harbor a
combination of both endo- and exocyclic nitrogen substitu-
tions. The lone electron pairs of these nitrogen atoms are re-
sponsible for unique features like charge transfer, the fluores-
cence dependence of the solvent and pH value.[19] However,
additional endocyclic nitrogen atoms in the azaindole system
do not contribute to the extension of the p-electron system
with their lone electron pair. Accordingly, such substances
tend to be more basic.
Thus, the engineering of these FPs always includes two ap-
proaches: 1) covalent modification of the chromophore by
classical site-directed mutagenesis (e.g., Ser65Thr), and 2) ma-
nipulation of electrostatic interactions between the chromo-
phore and its surrounding protein matrix. For example, ECFP
(enhanced cyan FP) was generated by random mutagenesis.
This allowed for identification of the structure with a series of
mutations in the neighborhood of the chromophore responsi-
ble for greatly improved quantum yield and fluorescence life-
time. On the other hand, in EGFP covalent modification of the
chromophore (Ser!Thr) favors its anionic form (lmaxem
=
509 nm) responsible for strong green fluorescence especially
attractive for numerous practical applications in molecular and
cellular biology. The most recent examples include the conver-
sion of a nonfluorescent chromoprotein into a fluorescent pro-
tein by semirandom mutagenesis.[38]
Similar to 7AI in apolar solutions, aminoindoles also possess
dual fluorescence emission with one peak centered around
350 nm and the other at around 520 nm.[34] The second spec-
tral peak is normally not visible in water for the following
reason. Water as a solvent acts as proton donor in the So state
while the intramolecular charge transfer from the amino group
to the indole is most effective in the S1 excited state. This also
explains the decrease in the relative fluorescence intensity of
aminoindoles when compared to the natural indole. In other
words, interactions such as hydrogen bonding of water mole-
cules to the aminoindole in the S1 state cause enhancement of
nonradiative decay processes and explain the absence of the
second fluorescence band in aqueous buffered solutions. How-
ever, in particular water-free milieus, such as membrane bilay-
ers or apolar protein interiors, these substances might be con-
ceived as dual-fluorescence protein-based sensors.
However, the basic advantage of GFP-like chromophores is
that their excitation wavelengths are usually above 350 nm,
whereas azaindole-based chromophores are usually excitable
below this value. Future work should clarify whether it is possi-
ble to circumvent this problem by extensive modifications of
azaindole chromophores (e.g., introduction of novel substitu-
ents like cyano, nitro, nitroso, imidazole, pyrrole, etc.).
From a general point of view, suitable protein microenviron-
ments might also be able to stabilize particular tautomeric
states of incorporated amino acid like chromophores. In other
words, these microenvironments could lead to the stabilization
of fluorescence species not assessable in bulk water. For the
evolution of such protein environments classical protein engi-
neering (i.e., site directed mutagenesis) and evolutionary meth-
ods (i.e., guided evolution) should be combined with an ex-
panded genetic code. In addition, the newly evolved indole
analogues could also be introduced to the chromophores of
FPs to expand their spectral properties. In this context a suita-
ble candidate for further improvement of the spectral proper-
ties could, for example, be 4-amino-7-azaindole (Figure 3).
Taken together, all the mentioned azaindole derivatives could
be solely used as chromophores or as part of post-translation-
ally generated chromophores in FPs.
In protein fluorescence design it might be of outstanding
importance to generate water-free protein interiors (i.e., cavi-
ties) especially suited for the accommodation of amino- or
azaindoles and their derivatives. Suitable arrangements of
amino acid side chains in such microenvironments could es-
tablish interactions that are, for example, able to stabilize
green emitting chromophore species. This concept is indeed
conceivable, if one considers, for example, the fluorescence of
GFP. The chromophore (4-(p-hydroxybenzilidene)imidazolid-5-
one) of GFP is completely encoded in its amino acid sequence
and autocatalytically formed by the post-translational reaction
between the side chains of residues 65–67 with molecular
oxygen as the only externally required component.[35] The
unique absorbance and fluorescence properties of GFP are not
an inherent property of the isolated chromophore, which is
nonfluorescent in water.[36] Its fluorescence is possible only in
the context of the properly folded protein.[37] In other words,
Methodological Issues
There are a few crucial methodological issues that should be
accounted for in order to employ the methylated azaindoles
for protein incorporation. Firstly, these chromophores should
be chemically synthesized as amino acids capable of
mimicking protein building blocks. Secondly, the biocompati-
bility of these compounds should be established. Thirdly, given
ChemBioChem 2010, 11, 305 – 314
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
311

N. Budisa et al.
that cotranslational incorporation is possible, engineering of
the protein target itself would be most probably necessary as
discussed above. Finally, the influence of these chromophores
on the protein structural integrity upon insertion into hydro-
phobic interior should be considered as well. For example, we
found that both 4-aminoindole as well as 4-azaindole—as W
analogues—deliver hydrophilicity into the apolar protein interi-
or, which impairs folding cooperativity and stability by disrup-
tion of the interaction network within the core.[39] In future
studies, the use of methylated azaindoles might be helpful to
circumvent these problems.
that support the proposal that such reprogrammed protein
translation in mammalian cells is possible.[44] It is usually assist-
ed by the supplementation of cells with plasmids encoding
modified translation components (aminoacyl-tRNA synthetases,
tRNAs, etc.) that are compatible with the endogenous transla-
tional apparatus of the host. In the coming years, we anticipate
significant progress in the development of reprogrammed
translation with an expanded genetic code for use in mamma-
lian cells.
Synthetic Biology: from DNA Mimicry to Mo-
lecular Switches, Recognition, Novel Biomateri-
als and Base Pairs
For successful in vivo protein biosynthesis with noncanonical
amino acids the following conditions must be fulfilled: 1) effec-
tive uptake/import of the noncanonical amino acid into the
cell, 2) its intracellular metabolic and chemical stability and ac-
cumulation at levels sufficient for activation and tRNA amino-
acylation, and 3) ribosomal translation into the nascent poly-
peptide chain in response to a sense or stop codon. To date,
there are no studies about possible intracellular uptake and
transport of methylated azatryptophans although there are
plenty of data about carbon methylated tryptophan deriva-
tives.[40] In spite of the fact that azaindoles/azatryptophans ful-
fill all the above listed criteria their methylated derivatives pos-
sess no protonated pyrrole nitrogen. As a consequence they
are not substrates of Trp synthase (Trp metabolism) or for tryp-
tophanyl-tRNA synthetase (Trp translation), that is, they are
bioorthogonal in the cellular milieu. This was confirmed by
studies of indole isosters in which the pyrrole nitrogen was re-
placed with other heteroatoms (e.g., sulfur or oxygen).[41]
Therefore, incorporation of methylated azatryptophans would
require the installation of novel aminoacylation pathways as
well as codon reassignment in the genetic code. Interestingly,
W analogues in which the pyrrole nitrogen is replaced with
other heteroatoms are found to be substrates for engineered
Phe tRNA synthetase.[42]
The slightly basic indole moiety of the canonical amino acid W
substituted by endocyclic nitrogen acquires stronger basicity.
Nucleobases and amino acids are strictly separated building
blocks in the synthesis of biopolymers, such as DNA/RNA on
the one hand and polypeptides on the other hand. Current de-
velopments in the field of code engineering by introducing
amino- and azaindole-based amino acids as protein building
blocks can also be seen as a bridge between these two worlds.
This will further assist novel research in biologically-inspired
material science.
DNA nucleobases exhibit a combination of both exocyclic
and endocyclic nitrogen atoms facilitating charge transfer and
basicity. These features play a key role in biological recognition
as well as in numerous intra- and intermolecular interactions.
All these features could be transferred to the level of proteins
by introduction of various nitrogen substituted indole ana-
logues. Azaindoles are indeed similar to the nucleic acid purine
bases since they contain pyridine nitrogen with a lone electron
pair capable of hydrogen bonding and excited-state tautomeri-
zation. Subsequently, as protein building blocks, azaindole and
its derivatives can decorate them with new functions and
properties Nature has exclusively reserved for nucleic acids.
This might be a crucial feature for the design of future protein-
based optoelectronic data storage devices or molecular wires
for information transduction.[45]
Currently existing approaches for in vivo incorporation of
noncanonical amino acids into proteins are normally per-
formed in bacterial hosts. Their transmission to eukaryotic cells
is not a trivial task. The use of stop codon suppression is addi-
tionally limited by read-through efficiency (usually ~40%)[43]
and permissiveness of the particular sequence positions (i.e.,
not all positions in the sequence are equally well suited for la-
beling). Last but not least, evolved enzymes for amino acid rec-
ognition and cognate tRNA have poor catalytic performance
and selectivity.[43] The other approach, known as selective pres-
sure incorporation (SPI)[43] relies on the relaxed substrate spe-
cificity of aminoacyl-tRNA synthetases as the crucial enzymes
in the interpretation of the genetic code. Thereby, the activa-
tion and transfer onto cognate tRNAs of a variety of structural-
ly and chemically similar substrate analogues is possible with-
out the necessity for extensive host cell engineering. This
straightforward in vivo approach is especially suited for multi-
site substitutions in a target protein sequence.
DNA mimicry is the property of some proteins to mimic
DNA structures.[46] They normally interact with enzymes that
bind DNA, such as restriction and repair enzymes, gyrase or
nucleosomal and nucleotide-associated proteins (molecular
switches). They also include repressor or activator proteins,
which can either block the access of the polymerase to DNA or
enhance enzyme binding to DNA. DNA mimicry might be fur-
ther developed if specific, for example, purine-mimicking W an-
alogues can be incorporated into proteins. Molecular recogni-
tion designed in this way opens up possibilities to develop
novel generations of molecular switches for tight control and
regulation of enzyme interactions with genomic DNA. In other
words, novel synthetic biology of molecular recognition and
molecular switching at the genome level might be established.
Finally, azaindole-based substances can be especially useful
for genetic code expansion by generation of noncanonical
base pairs. Thereby, a third, and therefore noncanonical, base
pair could be designed.[26] It is expected that this is stable in
The biggest challenge will certainly be to enable broader in
vivo use of these novel chromophores as a marker in cell biol-
ogy. This is currently limited due to the necessity of special in-
corporation protocols. Fortunately, there is already solid data
312
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 305 – 314

Blue Fluorescent Amino Acids
anhydrous toluene (10 mL). The mixture was stirred under nitrogen
at reflux temperature for 3 h. After being cooled, the solvent was
removed and the remaining white solid was dried in vacuo. The
crude product was used for the next reaction without further
purification. A mixture of ethyl acetate (25 mL) and concentrated
hydrochloric acid (10 mL) were added to the methylated salt. After
being stirred at room temperature for 18 h a concentrated aque-
ous solution of potassium hydroxide was added (pH 13) and the
mixture was extracted with dichloromethane (4ꢁ100 mL). Drying
with magnesium sulfate and evaporation of the solvent gave
duplex DNA and can be replicated by appropriate polymerases
with high efficiency and selectivity. This would lead to a greatly
expanded coding capacity of the genetic code.[47] These novel
noncanonical pairings could be based on sterical complemen-
tarity or hydrophobic interactions or on hydrogen bonding
through their tautomeric forms. For example, indole-like bases,
such as 7-azaindole or benzotriazole, could be good promising
substrates for incorporation into nucleic acids based on shape
complementarity.[48,49]
1
450 mg (90%) of an orange-colored oil. H NMR (400 MHz, CDCl3):
d=8.28 (d, J=1.8 Hz, 1H), 8.18 (d, J=7.7 Hz, 1H), 7.57 (d, J=
6.2 Hz, 1H), 6.94 (dd, J=7.7, 6.2 Hz, 1H), 6.44 (dd, J=1.8, 0.9 Hz,
1H), 4.15 (s, 3H) ppm.
Experimental Section
All chemicals were purchased from Sigma–Aldrich, unless other-
wise stated; 4AI and 7AI were purchased from Molekula and Bio-
synth. 1-Methyl-7-azaindole and 1-methyl-4-azaindole were synthe-
sized as described.[45,50] The protection of 4AI was performed as de-
scribed.[51] For the deprotection the literature procedure was
used.[52]
UV absorbance and fluorescence measurements: We acquired eM
of indole and azaindoles by UV spectra measurement with the UV/
Vis spectrometer lambda 19 (PerkinElmer and Life Sciences). Sam-
ples were measured in Tris-HCl buffer (100 mm, pH 7.5). Fluores-
cence spectra in a range of 300–500 nm were recorded with the lu-
minescence spectrometer LS50B (PerkinElmer and Life Sciences) at
208C at the determined maximum absorbance wavelengths as fol-
lows: W (278 nm), 4AI (288 nm), 1M4AI (289 nm), 4M4AI (285 nm),
1,4M4AI (290 nm), 7AI (288 nm), 1M7AI (289 nm), 1,7M7AI
(290 nm), 4AmI (270 nm), 4Am4AI (300 nm). Different excitation/
emission slits were used for the different substances measured be-
cause the fluorescence emission of some derivatives could not be
measured in a reasonable dilution at slit 2.5/2.5. Therefore, in case
of 4AI, 7AI, 4AmI and 4Am4AI the slits were set to 5/5. By co-
measuring with tryptophan as standard, measurements at different
slits and concentrations were later normalized with respect to each
other.
Syntheses
1-Methyl-7-azaindole (1M7AI): 7-Azaindole (472 mg, 4 mmol) and
sodium hydride (192 mg, 4.8 mmol; 60% in oil) were cooled to
08C under a nitrogen atmosphere. Then anhydrous dimethylforma-
mide (10 mL) was added and the mixture was stirred until the evo-
lution of hydrogen was complete. After the dropwise addition of
iodomethane (0.25 mL, 4 mmol) the reaction was allowed to warm
to room temperature. After being stirred at this temperature for
3 h, the solvent was removed and the residue was purified by flash
chromatography (PE:EE:DCM, 8:1.8:0.2) to give 494 mg (94%) of a
light yellow oil. 1H NMR (400 MHz, CDCl3): d=08.34 (dd, J=4.7,
1.5 Hz, 1H), 7.90 (dd, J=7.8, 1.5 Hz, 1H), 7.18 (d, J=3.5 Hz, 1H),
7.05 (dd, J=7.8, 4.7 Hz, 1H), 6.45 (d, J=3.5 Hz, 1H), 3.900 (s, 3H)
ppm. The NMR data are in agreement with those reported in the
literature.[50]
Excitation spectra were corrected automatically by the instrument
for artifacts originating from the xenon lamp (energy output is
wavelength and time dependent) and from the excitation mono-
chromator (efficiency is wavelength dependent). Emission spectra
were corrected with respect to artifacts originating from the emis-
sion monochromator (efficiency is wavelength dependent) and the
sample multiplier (sensitivity is wavelength dependent). Correction
files for emission spectra are yearly updated and provided by the
manufacturer of the instrument (PerkinElmer and Life Sciences). All
measurements were made with constant detector current.
1-Methyl-4-azaindole (1M4AI): 4-Azaindole (354 mg, 3 mmol) and
sodium hydride (144 mg, 3.6 mmol; 60% in oil) were cooled to
08C under a nitrogen atmosphere. Then dry dimethylformamide
(10 mL) was added and the mixture was stirred until the evolution
of hydrogen was complete. After the dropwise addition of iodome-
thane (0.19 mL, 3 mmol) the reaction was allowed to warm to
room temperature. After being stirred at this temperature for 3 h,
the solvent was removed and the residue was purified by flash
chromatography (PE:EE, 1:1) to give 330 mg (83%) of a light
brown oil. 1H NMR (400 MHz, CDCl3): d=8.46 (dd, J=4.6, 1.2 Hz,
1H), 7.62 (d, J=8.2 Hz, 1H), 7.32 (d, J=3.2 Hz, 1H), 7.12 (dd, J=
8.2, 4.6 Hz, 1H), 6.69 (dd, J=3.2, 0.8 Hz, 1H), 3.81 (s, 3H) ppm.
Acknowledgements
We thank Prof. Luis Moroder for critical reading of the manu-
script and Dr. Sandra Lepthien for the protein gels in Figure 1.
We are also grateful to Dr. Stephan Uebel from the Microchemis-
try Core Facility of the MPI for provided us with instrumentation
for spectroscopic measurements. This work was supported by the
BioFuture program of the Federal Ministry of Science and Educa-
tion (BMBF) of Germany.
tert-Butyl-4-azaindole-1-carboxylate:
N,N-Dimethylamino-pyridine
(49 mg, 0.4 mmol) and di-tert-butyl dicarbonate (1.05 g, 4.8 mmol)
were added to a solution of 7-azaindole (472 mg, 4 mmol) in anhy-
drous acetonitrile (10 mL). After being stirred under nitrogen at
room temperature for 3 h, the solvent was removed and the resi-
due was purified by flash chromatography (PE:EE, 1:1) to give
1
874 mg (quant.) of a white solid; m.p. 69–718C. H NMR (400 MHz,
Keywords: charge transfer · green fluorescence · proteins ·
tautomerism · tryptophan
CDCl3): d=8.34 (dd, J=4.7, 1.3 Hz, 1H), 7.90 (d, J=6.8 Hz, 1H),
7.83 (d, J=3.7 Hz, 1H), 7.05 (dd, J=8.3, 4.7 Hz, 1H), 6.78 (d, J=
3.7 Hz, 1H), 1.68 (s, 9H) ppm. 13C NMR (100 MHz, CDCl3): d=149.1,
148.8, 145.5, 128.9, 122.2, 118.7, 108.3, 84.5, 28.1 ppm. The NMR
data are in agreement to those reported in the literature.[53]
[1] I. M. Dobbie, N. F. Lowndes, K. F. Sullivan in Autofluorescent proteins,
Vol. 85 (Ed.: K. F. Sullivan), Academic Press, London, 2008, pp. 1–22.
[2] N. C. Shaner, G. H. Patterson, M. W. Davidson, J. Cell Sci. 2007, 120,
4247–4260.
4-Methyl-4-azaindole (4M4AI): Dimethylsulfate was added to a solu-
tion of tert-butyl-4-azaindole-1-carboxylate (775 mg, 3.55 mmol) in
ChemBioChem 2010, 11, 305 – 314
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
313

N. Budisa et al.
[3] M. Fernandez-Suarez, A. Y. Ting, Nat. Rev. Mol. Cell. Biol. 2008, 9, 929–
943.
[30] S. M. Twine, L. Murphy, R. S. Phillips, P. Callis, M. T. Cash, A. G. Szabo, J.
Phys. Chem. B. 2003, 107, 637–645.
[4] I. Moore, A. Murphy, Plant Cell 2009, 21, 1632–1636.
[5] E. L. Snapp, Trends Cell Biol. 2009, 19, 649–655.
[6] J. A. Prescher, C. R. Bertozzi, Nat. Chem. Biol. 2005, 1, 13–21.
[7] S. R. Adams, R. E. Campbell, L. A. Gross, B. R. Martin, G. K. Walkup, Y. Yao,
J. Llopis, R. Y. Tsien, J. Am. Chem. Soc. 2002, 124, 6063–6076.
[8] M. Sunbul, J. Yin, Org. Biomol. Chem. 2009, 7, 3361–3371.
[9] A. R. Katritzky, T. Narindoshvili, Org. Biomol. Chem. 2009, 7, 627–634.
[10] S. Lepthien, M. G. Hoesl, L. Merkel, N. Budisa, Proc. Natl. Acad. Sci. USA
2008, 105, 16095–16100.
[31] Y. A. Labas, N. G. Gurskaya, Y. G. Yanushevich, A. F. Fradkov, K. A. Lukya-
nov, S. A. Lukyanov, M. V. Matz, Proc. Natl. Acad. Sci. USA 2002, 99,
4256–4261.
[32] M. A. Wall, M. Socolich, R. Ranganathan, Nat. Struct. Biol. 2000, 7, 1133–
1138.
[33] P.-T. Chou, C.-Y. Wei, G.-R. Wu, W.-S. Chen, J. Am. Chem. Soc. 1999, 121,
12186–12187.
[34] H. K. Sinha, S. K. Dogra, M. Krishnamurthy, Bull. Chem. Soc. Jap. 1987,
60, 4401–4407.
[11] J. R. Lakowitz, Protein Fluorescence, Kluwer Academic/Plenum Publisher,
New York, 1999.
[35] C. W. Cody, D. C. Prasher, W. M. Westler, F. G. Prendergast, W. W. Ward,
Biochemistry 1993, 32, 1212–1218.
[12] N. Budisa, P. Paramita Pal, Biol. Chem. 2004, 385, 893–904.
[13] S. Lepthien, B. Wiltschi, B. Bolic, N. Budisa, Appl. Microbiol. Biotechnol.
2006, 73, 740–754.
[14] J. B. A. Ross, A. G. Szabo, C. W. V. Hogue, Methods Enzymol. 1997, 278,
151–190.
[15] J. B. A. Ross, E. Rusinova, L. A. Luck, K. W. Rousslang in Topics in Fluores-
cence Spectroscopy, Vol. 6 (Ed.: J. R. Lakowitz), Plenum, New York, 2000,
pp. 17–42.
[36] H. Niwa, S. Inouye, T. Hirano, T. Matsuno, S. Kojima, M. Kubota, M.
Ohashi, F. I. Tsuji, Proc. Natl. Acad. Sci. USA 1996, 93, 13617–13622.
[37] Y. Nishiuchi, T. Inui, H. Nishio, J. Bodi, T. Kimura, F. I. Tsuji, S. Sakakibara,
Proc. Natl. Acad. Sci. USA 1998, 95, 13549–13554.
[38] T. Kogure, S. Karasawa, T. Araki, K. Saito, M. Kinjo, A. Miyawaki, Nat. Bio-
technol. 2006, 24, 577–581.
[39] M. Rubini, S. Lepthien, R. Golbik, N. Budisa, Biochim. Biophys. Acta 2006,
1764, 1147–1158.
[16] A. B. Pardee, V. G. Shore, L. S. Prestidge, Biochim. Biophys. Acta 1956, 21,
406–407.
[40] K. D. Brown, J. Bacteriol. 1970, 104, 177–188.
[41] N. Budisa, S. Alefelder, J. H. Bae, R. Golbik, C. Minks, R. Huber, L. Morod-
er, Protein Sci. 2001, 10, 1281–1292.
[42] I. Kwon, D. A. Tirrell, J. Am. Chem. Soc. 2007, 129, 10431–10437.
[43] N. Budisa, Angew. Chem. 2004, 116, 6586–6624; Angew. Chem. Int. Ed.
2004, 43, 6426–6463.
[44] W. S. Liu, A. Brock, S. Chen, S. B. Chen, P. G. Schultz, Nat. Methods 2007,
4, 239–244.
[45] V. Shafirovich, N. E. Geacintov in Charge Transfer in DNA (Ed.: H. A. Wa-
genknecht), Wiley-VCH, Weinheim, 2005, pp. 175–193.
[46] D. T. Dryden, M. R. Tock, Biochem. Soc. Trans. 2006, 34, 317–319.
[47] E. Szathmary, Nat. Rev. Genet. 2003, 4, 995–1001.
[48] I. Hirao, T. Ohtsuki, T. Fujiwara, T. Mitsui, T. Yokogawa, T. Okuni, H. Na-
kayama, K. Takio, T. Yabuki, T. Kigawa, K. Kodama, K. Nishikawa, S. Yo-
koyama, Nat. Biotechnol. 2002, 20, 177–182.
[17] G. Brawerman, M. Ycas, Arch. Biochem. Biophys. 1957, 68, 112–117.
[18] S. Schlesinger, J. Biol. Chem. 1968, 243, 3877–3883.
[19] N. Budisa, M. Rubini, J. H. Bae, E. Weyher, W. Wenger, R. Golbik, R.
Huber, L. Moroder, Angew. Chem. 2002, 114, 4238–4242; Angew. Chem.
Int. Ed. 2002, 41, 4066–4069.
[20] J. H. Bae, M. Rubini, G. Jung, G. Wiegand, M. H. J. Seifert, M. K. Azim, J.-
S. Kim, A. Zumbusch, T. A. Holak, L. Moroder, R. Huber, N. Budisa, J. Mol.
Biol. 2003, 328, 1071–1081.
[21] L. Wang, J. M. Xie, A. A. Deniz, P. G. Schultz, J. Org. Chem. 2003, 68,
174–176.
[22] D. Kajihara, T. Hohsaka, M. Sisido, Protein Eng. Des. Sel. 2005, 18, 273–
278.
[23] C. F. Chapman, M. Maroncelli, J. Phys. Chem. 1992, 96, 8430–8441.
[24] M. T. Cash, P. R. Schreiner, R. S. Phillips, Org. Biomol. Chem. 2005, 3,
3701–3706.
[25] A. V. Smirnov, D. S. English, R. L. Rich, J. Lane, L. Teyton, A. W. Schwa-
bacher, S. Luo, R. W. Thornburg, J. W. Petrich, J. Phys. Chem. B 1997,
101, 2758–2769.
[26] C. A. Taylor, M. A. El-Bayoumi, M. Kasha, Proc. Natl. Acad. Sci. USA 1969,
66, 253–260.
[27] Y. Chen, R. L. Rich, F. Gai, J. W. Petrich, J. Phys. Chem. 1993, 97, 1770–
1780.
[49] D. E. Bergstrom, Chem. Biol. 2004, 11, 18–20.
[50] B. S. Lane, D. Sames, Org. Lett. 2004, 6, 2897–2900.
[51] L. Grehn, U. Ragnarsson, Angew. Chem. 1984, 96, 291–292; Angew.
Chem. Int. Ed. Engl. 1984, 23, 296–301.
[52] C. M. Coleman, D. F. O’Shea, J. Am. Chem. Soc. 2003, 125, 4054–4055.
[53] K. Yamawaki, T. Nomura, T. Yasukata, N. Tanimoto, K. Uotani, H. Miwa, Y.
Yamano, K. Takeda, Y. Nishitani, Bioorg. Med. Chem. 2008, 16, 1632–
1647.
[28] R. L. Rich, Y. Chen, D. Neven, M. Nꢂgrerie, F. Gai, J. W. Petrich, J. Phys.
Chem. 1993, 97, 1781–1788.
Received: October 23, 2009
[29] J. Waluk, B. Pakula, S. J. Komorowski, J. Photochem. 1987, 39, 49–58.
Published online on January 7, 2010
314
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 305 – 314