Active/inactive dual-probe system for selective photoaffinity labeling of small molecule-binding proteins

Article abstract of DOI:10.1002/asia.201200085

Two are better than one: A new approach to selective photoaffinity labeling is described in which a bioactive probe is used in combination with its inactive analog as a scavenger of nonspecific proteins. Copyright

Full text of DOI:10.1002/asia.201200085

DOI: 10.1002/asia.201200085
Active/Inactive Dual-Probe System for Selective Photoaffinity Labeling of
Small Molecule-Binding Proteins
Kaori Sakurai,* Masaki Tawa, Ayumi Okada, Rika Yamada, Noriyuki Sato,
Masahiro Inahara, and Maia Inoue[a]
Selective detection is a critical first step in the identifica-
tion of small molecule-binding proteins from a complex mix-
ture of cellular proteins.[1,2] Photoaffinity labeling (PAL)
offers a potentially ideal strategy for isolating direct binding
proteins from the proteome, especially when the structural
details or chemical reactivities of the proteins at the binding
sites are completely unknown.[3,4] PAL involves the use of
bioactive small-molecule ligands derivatized with a photo-
reactive group and a reporter group.[3e–f] Upon photoactiva-
tion, a highly reactive intermediate generated from a probe
can react with a variety of amino acid residues of a protein
in a distant-dependent manner. The unique reactivity pro-
vides the basis for affinity-selective crosslinking of the small
molecule–protein complexes. Once crosslinked, the chemi-
cally stable complexes become amenable to various bio-
chemical analyses such as proteolytic digestion followed by
identification of the binding proteins or the binding sites
within the binding proteins by mass spectrometry (MS).[3d–f]
However, the general utility of PAL in the discovery of
binding proteins has often been limited in reality largely due
to low selectivity and low crosslinking yields.[3,5] These prob-
lems are particularly pronounced in cases where the binding
affinity is modest (Kd >mM) or the binding proteins are not
abundant in the cell.[6] The development of PAL methods to
date has been aimed primarily toward increasing the reactiv-
ity rather than the selectivity of photoactivatable functionali-
ties.[3,7] Herein, we report a novel strategy to control the se-
lectivity of PAL by simultaneously using a bioactive PAL
probe and its inactive analog. We believe that it represents
a step toward expanding the scope of PAL especially in the
search of small molecule-binding proteins with modest affin-
ity, which has traditionally been difficult.
well as various other nonspecific proteins, both of which are
then photocrosslinked through a fast intramolecular reaction
(Figure 1, upper path). The formation of nonspecific com-
plexes brings about two undesirable effects in PAL reac-
tions. First, it leads to multiple reaction products, complicat-
ing subsequent product analysis. Second, sequestering of the
probe by nonspecific proteins decreases the effective con-
centration of the probe. We hypothesized that the probe
could be tuned to selectively react with its specific binding
proteins if the formation of undesired nonspecific complexes
is disfavored by an inhibitor prior to the PAL reaction step.
Small molecules are known to nonspecifically bind to non-
polar patches of protein surfaces through hydrophobic inter-
action.[8] The strength of hydrophobic interaction is consid-
ered roughly proportional to the buried surface area of the
bound complex. The overall magnitude of the nonspecific
small molecule–protein interaction typically follows a simple
absorption isotherm as there are numerous nonspecific bind-
ing sites on proteins.[8a] Our approach therefore is to intro-
duce a biologically inactive structural analog, which would
mimic the active PAL probe in the nonspecific protein bind-
ing property (Figure 1, lower path ). Excess amounts of an
inactive PAL probe would predominantly form complexes
and subsequently react with nonspecific proteins as a scav-
enger,[9] while the free active probe would be fully available
to complex with its specific binding protein.
To explore this new PAL approach, we employed benze-
nesulfonamide as a model bioactive small-molecule ligand
and studied the reactivity of its derivatives as PAL probes.
Benzenesulfonamide is a potent inhibitor of human carbonic
anhydrase II (hCAII; IC50 =0.27 nm[10a]), a ubiquitous cyto-
solic protein, which catalyzes the reversible hydration of
CO2. The protein–ligand interaction has been well character-
ized biochemically and crystallographically.[10,11] To detect
the binding protein of benzenesulfonamide by PAL, we de-
signed a trifunctional probe 1 based on an l-lysine scaffold
that contains a benzenesulfonamide moiety as a protein-
binding ligand, benzophenone as a photoactivatable group,
and biotin as a reporter group,[12] which allows the detection
of protein–photoadducts (Figure 2). Compound 3 possessing
the ligand group but lacking the biotin reporter group was
designed as a positive control. As a biologically inactive
analog of the benzenesulfonamide group, the 4-methoxyben-
zoyl group was chosen on the basis of its similar shape and
surface area. Compounds 2 and 4 represent inactive analogs
In a conventional PAL reaction with a mixture of cellular
proteins, a probe displaying a bioactive ligand can form
equilibrium complexes with its specific binding proteins as
[a] Prof. K. Sakurai, M. Tawa,+ A. Okada,+ R. Yamada, N. Sato,
M. Inahara, M. Inoue
Department of Biotechnology and Life Science
Tokyo University of Agriculture and Technology
2-24-16 Naka-cho, Koganei-shi, Tokyo 184-8588 (Japan)
Fax : (+81)42-388-7374
E-mail: sakuraik@cc.tuat.ac.jp
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/asia.201200085.
Chem. Asian J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
&
&
&
These are not the final page numbers! ÞÞ

COMMUNICATION
overall
yield
of
38%
(Scheme 1). Analogs 2–4 were
obtained in two steps from the
intermediate 7. We evaluated
the protein-binding activity of
compounds 1–4 by performing
an enzyme activity inhibition
assay for hCAII.[8a] Whereas 2
and
4
showed no activity
(IC50 >1 mm) as expected, the
PAL probes 1 and 3, which
were designed to bind hCAII,
displayed IC50 values of 0.26 mm
and 0.23 mm, respectively. Thus,
the incoporation of the photo-
activatable group or the biotin
reporter group into the probe
did not interfere with the small
molecule–protein interaction.
PAL reactions were conduct-
ed to first evaluate the reactivi-
Figure 1. PAL reaction in a cell lysate with a) traditional PAL probe and b) dual-probe PAL system. The bind-
ing protein of a ligand is shown in magenta. Magenta circle, reporter group; gray circle: non-reporter group.
ty of compounds
toward proteins. Typically, reac-
tions were initiated by UV irra-
1 and 2
diation (l=365 nm) of the reactants on ice; subsequently,
the reaction mixture was separated by SDS-PAGE. The pho-
toadducts derived from 1 and 2 as biotin-labeled proteins
were detected by using fluorophore-conjugated Streptavidin
and analyzed by using a fluorescence imager. On the other
hand, the photoadducts of 3 or 4 without the biotin group
did not give rise to fluorescent signals. The efficiency of
each reaction was evaluated by the determining the intensity
of fluorescence signals corresponding to the quantity of the
protein photoadducts in the gels. When 1 was reacted with
its known binding protein, hCAII, a photoadduct (12) of
a single fluorescent band was generated in a dose-dependent
manner for 1, with EC50 =0.19 mm (Figure 3a).[13] While the
active analog 3 inhibited the generation of the photoadduct
12 with an IC50 of 13 mm (Figure 3b), the inactive analog 4
displayed no inhibitory activity, even at a concentration of
1 mm. These results confirmed that the PAL reaction be-
tween 1 and hCAII is based on specific binding. To assess if
1 reacts with other proteins in a nonspecific fashion, it was
also reacted with bovine serum albumin (BSA), which is not
known to bind benzenesulfonamide but known to nonspecif-
ically bind various hydrophobic small molecules.[14] The re-
action generated the photoadduct 13; with an increase in
the concentration of 1, increasing amounts of 13 were
formed with no saturation observed up to 1 mm of 1 (ex-
trapolated to EC50 =2.9 mm; Figure 3c), which is characteris-
tic for a phenomenon driven by nonspecific binding.[8a,15]
Compound 2 also showed a dose-dependent reactivity with
hCAII and BSA, with EC50 values of respectively 1.7 mm and
0.84 mm (Figure S1 in the Supporting Information), thus indi-
cating that the apparent binding affinities of 2 for both pro-
teins are relatively high. To test the hypothesis that 1 and its
inactive probe 4 compete for similar nonspecific binding
Figure 2. Structures of the active PAL probe 1 and the inactive probes 2–
4 and their inhibitory activity against hCAII (mM). N. D., no activity de-
tected.
with and without biotin reporter group and are expected to
bind proteins nonspecifically in a manner similar to 1. Probe
1 was synthesized from N-Boc-l-lysine in four steps with an
&
&
&2
&
www.chemasianj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 0000, 00, 0 – 0
ÝÝ These are not the final page numbers!

Selective Photoaffinity Labeling of Small Molecule-Binding Proteins
Scheme 1. Synthesis of the active PAL probe 1 and the inactive probes 2–4. Reagents and conditions: a) 6, 20% Na2CO3 aq., DMF, 77%; b) 8, DMT-
MM, MeOH, 86%; c) (i) 4m HCl-dioxane, CH2Cl2, (ii) 4-sulfamoylbenzoate N-hydroxysuccinimide ester, CH2Cl2, DIPEA, DMF, 55%; d) (i) TFA/
CH2Cl2 =1:2, (ii) 4-methoxybenzoyl chloride, DIPEA, DMF, 47%; e) 10, HATU, DIPEA, DMF, 84%; f) (i) 4m HCl-dioxane, CH2Cl2, (ii) 4-sulfamoylben-
zoic acid, HATU, DIPEA, DMF, 17%; g) (i) 4m HCl-dioxane, (ii) 4-methoxybenzoic acid, HATU, DIPEA, DMF, 60%.
Figure 3. a) PAL reactions of 1 at various concentrations with hCAII (1 mg, 1.7 mm). b) Effects of 3 on the PAL reaction of 1 (1 mm) with hCAII (1 mg,
1.7 mm). c) PAL reactions of 1 at various concentrations with BSA (5 mg, 3.8 mm); d) Effects of the inactive probe 4 on the PAL reaction of 1 (1 mm) with
BSA (5 mg, 3.7 mm). Fluorescence data were normalized to the largest value for each experiment.
Chem. Asian J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
3
&
&
&
These are not the final page numbers! ÞÞ

Kaori Sakurai et al.
COMMUNICATION
sites on a protein, we reacted 1 with BSA in the presence of
varying amounts of 4. Significantly, we observed a decrease
in the formation of photoadduct 13 at increasing concentra-
tions of 4 (IC50 =170 mm; Figure 3d). This result indicated
that BSA preferentially bound and reacted with the excess
reactant 4 over 1. Therefore, it raised the possibility that
a high selectivity of the active PAL probe could be achieved
even in a protein mixture by simply disfavoring the forma-
tion of undesired complexes through addition of an inactive
PAL probe.
We next performed the PAL reaction of 1 in mixtures of
hCAII and BSA. As indicators of effective PAL reactions,
we employed the selectivity for the outcome of each reac-
tion, which is calculated from the following equation: selec-
tivity (%)=(fluorescent intensity of 12/total fluorescent in-
tensity)ꢁ100. When the initial ratio of hCAII to BSA was
decreased from 10:1 to 1:100, the PAL probe 1 generated 12
with a decreasing selectivity (from 90% to 37%, see Fig-
ure 4a). Thus, probe 1 reacted with abundant BSA more ef-
ficiently than with its specific binding protein hCAII, consis-
tent with the typical outcome of PAL experiments under
conventional conditions. Under the same reaction condi-
tions, 2 gave a photoadduct with BSA (17) as the major
product reflecting the initial ratio of the two proteins (Fig-
ure S1c in the Supporting Information). When 1 was reacted
with a mixture of hCAII and BSA at the ratio of 1:100 in
the presence of various amounts of 4, we found that the
highest selectivity of 65% was achieved with 1000-fold
excess of 4 relative to 1. Therefore, the inactive probe 4 se-
lectively inhibited the formation of 13 in a dose-dependent
manner. Due to the limited solubility of 4 in aqueous solu-
tion, we were unable to explore its ability at much higher
concentrations to improve the selectivity for the active
probe 1.
To gain insight into the structural basis of the nonspecific
PAL reaction, we also evaluated the effect of substructures
of the inactive probe (Figure 5, 4, 7, and 14) in the PAL re-
action of 1 with a 1:100 mixture of hCAII and BSA. As
shown in Figure 5, compound 7 also exhibited an inhibitory
effect on the nonspecific reactivity of 1 with BSA; however,
its efficiency was weaker than that of 4. Lack of competitive
inhibition by 14 confirmed that the nonspecific PAL reac-
tions by 1 involve nonspecific protein binding events and do
not occur via simple intermolecular reactions. From these
data, all parts of the inactive probe 4 were found necessary
to maximally compete with 1. Therefore, our data indicate
that designing an appropriate analog is important for achiev-
ing PAL detection of specific binding proteins with a high
selectivity. Nevertheless, when a structurally close inactive
ligand is not readily available, using a simple scaffold struc-
ture such as 7 may sufficiently improve the selectivity of
PAL reactions.
An ideal PAL probe should react selectively with specific
binding proteins in the presence of a complex array of cellu-
lar proteins in a reaction solution. With the conventional
PAL method, when 1 was reacted with hCAII in the pres-
ence of increasing amounts of proteins from HeLa cell ly-
sates, an increasing amount of photoadducts with various
proteins was also formed apart from the desired photoad-
duct 12 (Figure 6a). Due to the overlap between fluorescent
bands for 12 and other photoadducts in the SDS-PAGE gel,
the selectivity of product formation was not assessed quanti-
tatively in these cases. When the inactive probe 4 was added
at 1000-fold excess over the active probe 1 to the reaction
mixture of hCAII and lysate proteins (ratio of 1:100), the
formation of undesired photoadducts was visibly inhibited,
as observed in the SDS-PAGE gel (Figure 6b). This showed
that the inactive probe 4 served effectively as a mimic of
Figure 4. a) Effects of BSA on the PAL reactions of 1 (1 mm) with hCAII
(1 mg). b) Effects of the inactive probe 4 on the PAL reactions of
1 (1 mm) with hCAII (1 mg) and BSA (100 mg).
Figure 5. Effects of inactive analogs 4, 7, 14, or 15 (1 mm) on the PAL reactions of 1 (1 mm) with hCAII (1 mg) and BSA (100 mg).
&
&
&4
&
www.chemasianj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 0000, 00, 0 – 0
ÝÝ These are not the final page numbers!

Selective Photoaffinity Labeling of Small Molecule-Binding Proteins
[2] R. E. Kleiner, C. E. Dumelin, G. C. Tiu, K. Sakurai, D. R. Liu, J.
Am. Chem. Soc. 2010, 132, 11779–11791.
[3] a) A. Kawamura, D. M. Mihai, Methods Mol. Biol. 2012, 803, 65–
75; b) Y. Hatanaka, Y. Sadakane, Curr. Top. Med. Chem. 2002, 2,
271–288; c) Y. Tanaka, M. R. Bond, J. J. Kohler, Mol. BioSyst. 2008,
4, 473–480; d) F. Kotzyba-Hibert, I. Kapfer, M. Goeldner, Angew.
Chem. 1995, 107, 1391–1408; Angew. Chem. Int. Ed. Engl. 1995, 34,
1296–1312; e) G. Dormꢃn, G. D. Prestwich, Trends Biotechnol. 2000,
18, 64–77; f) L. Dubinsky, B. P. Krom, M. M. Meijler, Bioorg. Med.
Chem. 2012, 20, 554–570.
[4] For recent examples regarding affinity-based protein profiling using
photoreactive groups, see: a) C. M. Salisbury, B. F. Cravatt, J. Am.
´
Chem. Soc. 2008, 130, 2184–2194; b) M. Fonovic, M. Bogyo, Exp.
Figure 6. a) In-gel western blot analysis of PAL reactions of 1 (1 mm) with
hCAII (1 mg) mixed with various amounts of HeLa cell lysate. b) In-gel
western blot analysis of PAL reactions of 1 with hCAII (1 mg) mixed with
HeLa cell lysate (100 mg) in the absence (lane 1) or presence (lane 2) of 4
(1 mm).
Rev. Proteomics 2008, 5, 721–730; c) H. Shi, C.-J. Zhang, G. Y. J.
Chen, S. Q. Yao, J. Am. Chem. Soc. 2012, 134, 3001–3014; d) J.
Lenger, F. Kaschani, T. Lenz, C. Dalhoff, J. G. Villamor, H. Kçster,
N. Sewald, R. A. L. van der Hoorn, Bioorg. Med. Chem. 2012, 20,
592–596.
[5] Enantiodifferential PAL is a promising approach to circumvent the
low selectivity issue by using an external negative control. See: a) Y.
Nakamura, R. Miyatake, M. Ueda, Angew. Chem. 2008, 120, 7399–
7402; Angew. Chem. Int. Ed. 2008, 47, 7289–7292; b) A. Kawamura,
S. Hindi, D. M. Mihai, L. James, O. Aminova, Bioorg. Med. Chem.
2008, 16, 8824–8829.
[6] By contrast, PAL probes with high affinity ligands could lead to
a high labeling selectivity even when the binding protein represents
less than 0.1% of total protein. See, for example: A. K. Hamouda,
X. Jin, M. Sanghvi, S. Srivastava, A. Pandhare, P. K. Duddempudi,
J. H. Steinbach, M. P. Blanton, Biochim. Biophys. Acta Biomembr.
2009, 1788, 1987–1995.
[7] a) M. Ueda, Y. Manabe, Y. Otsuka, N. Kanzawa, Chem. Asian J.
2011, 6, 3286–3297; b) B. Albertoni, J. S. Hannan, D. Ackermann,
A. Schmitz, M. Famulok, Chem. Commun. 2012, 48, 1272–1274.
[8] a) C. P. Woodbury, Introduction to Macromolecular Binding Equili-
bria, CRC Press, Boca Raton, 2008, Chap. 1; b) T. E. Creighton,
Proteins: Structures and Molecular Properties, 2nd ed., W. H. Free-
man & Co., New York, 1993, pp. 139–169.
[9] A. E. Ruoho, H. Kefer, P. E. Roeder, S. J. Singer, Proc. Natl. Acad.
Sci. USA 1973, 70, 2567–2571.
[10] a) Y. Takaoka, H. Tsutsumi, N. Kasagi, E. Nakata, I. Hamachi, J.
Am. Chem. Soc. 2006, 128, 3273–3280; b) G. Chen, A. Heim, D.
Riether, D. Yee, Y. Milgrom, M. A. Gawinowicz, D. Sames, J. Am.
Chem. Soc. 2003, 125, 8130–8133.
[11] a) V. M. Krishnamurthy, G. K. Kaufman, A. R. Urbach, I. Gitlin,
K. L. Gudiksen, D. B. Weibel, G. M. Whitesides, Chem. Rev. 2008,
108, 946–1051; b) S. Parkkila, H. Rajaniemi, A.-K. Parkkila, J.
Kivelꢄ, A. Waheed, S. Pastorekovꢃ, J. Pastorek, W. S. Sly, Proc. Natl.
Acad. Sci. USA 2000, 97, 2220–2224; c) A. Thiry, C. T. Supuran, B.
Masereel, J.-M. Dognꢅ, J. Med. Chem. 2008, 51, 3051–3056.
[12] B. A. Gilbert, R. R. Rando, J. Am. Chem. Soc. 1995, 117, 8061–
8066.
[13] The photoadduct 12 was generated only when the reaction mixture
was irradiated with UV light. An analysis of the PAL reaction time
showed that the photoadduct formation saturates beyond 30 min
(Figure S2 in the Supporting Information).
probe 1 in its nonspecific interaction with various cellular
proteins but not in the specific interaction with hCAII.
These results demonstrate that it is possible to tune the se-
lectivity of an active PAL probe in a complex protein mix-
ture by simultaneously employing an inactive probe as
a scavenger of nonspecific proteins.
In conclusion, we have described an active/inactive dual-
probe approach to control the selectivity of PAL reactions
toward detecting specific small molecule-binding proteins.
We demonstrated that our new PAL system is applicable to
levels of a binding protein in the cell lysate as low as 1%
(w/w). In addition, we found that a simple inactive analog
representing the scaffold moiety of the PAL probe can also
improve the labeling selectivity, which should be useful in
cases where inactive analog is not easily obtained. With judi-
cious choice of the inactive PAL probe, our method should
be generally applicable to a wide range of bioactive small
molecules for straightforward detection of their binding pro-
teins. It should be particularly useful for exploring yet un-
known small molecule–protein interactions with modest af-
finity, which have been previously difficult to detect.
Acknowledgements
We thank Kazuhiro Ogata for technical assistance. This work was funded
through the Tenure-Track program and the Career-Advance program of
the Womenꢂs Future Development Organization at TUAT, and the young
investigator award from HFSP (K.S.). A.O. is supported by the Ajinomo-
to scholarship.
[14] a) N. A. Avdulov, S. V. Chochina, V. A. Daragan, F. Shroeder, K. H.
Mayo, W. G. Wood, Biochemistry 1996, 35, 340–347; b) L. Fielding,
S. Rutherford, D. Fletcher, Magn. Reson. Chem. 2005, 43, 463–470.
[15] We excluded the possibility of intermolecular PAL reactions be-
cause no quenching was observed when radical quenchers such as
phenol (0.1m) were used as an additive and because the concentra-
tions of the reactants are beyond the regime for simple bimolecular
reactions.
Keywords: analog · ligand–protein interaction · nonspecific
binding · photoaffinity labeling · proteins
[1] a) B. Lomenick, R. W. Olsen, J. Huang, ACS Chem. Biol. 2011, 6,
34–46; b) S. Sato, A. Murata, T. Shirakawa, M. Uesugi, Chem. Biol.
2010, 17, 616–623; c) M. J. Evans, B. F. Cravatt, Chem. Rev. 2006,
106, 3279–3301; d) T. Bçttcher, M. Pitscheider, S. A. Sieber, Angew.
Chem. 2010, 122, 2740–2759; Angew. Chem. Int. Ed. 2010, 49, 2680–
2698 and references therein.
Received: January 25, 2012
Revised: March 2, 2012
Published online: && &&, 0000
Chem. Asian J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
5
&
&
&
These are not the final page numbers! ÞÞ

COMMUNICATION
Photoaffinity Labeling
Kaori Sakurai,* Masaki Tawa,
Ayumi Okada, Rika Yamada,
Noriyuki Sato, Masahiro Inahara,
Maia Inoue
&&&& — &&&&
Active/Inactive Dual-Probe System for
Selective Photoaffinity Labeling of
Small Molecule-Binding Proteins
Two are better than one: A new
approach to selective photoaffinity
labeling is described in which a bioac-
tive probe is used in combination with
its inactive analog as a scavenger of
nonspecific proteins.
&
&
&6
&
www.chemasianj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 0000, 00, 0 – 0
ÝÝ These are not the final page numbers!