Labeling study of avidin by modular method for affinity labeling (MoAL)

Article abstract of DOI:10.1016/j.bmcl.2010.09.109

We studied the specific labeling of avidin with biotinylated modular ligand catalysts via MoAL, which we recently established. The labeling yield was found to depend on the linker length connecting the catalytic site to biotin in the modular ligand catalyst 1, and the maximum yield was obtained with 1d possessing octamethylene linker. The labeling reaction reached a maximum rate with only 4 equiv of the ligand catalyst. Presumably, all the subunits of avidin with homotetrameric structure formed a stable complex with 4 equiv of the catalyst because of the extremely high affinity. The ligand catalyst bound to avidin first catalyzed N-triazinylation of the -amino group of Lys111, and the resulting regenerated catalyst then catalyzed the reaction of Asp108 and CDMT.

Full text of DOI:10.1016/j.bmcl.2010.09.109

Bioorganic & Medicinal Chemistry Letters 20 (2010) 7050–7053
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Labeling study of avidin by modular method for affinity labeling (MoAL)
Shuichi Nakanishi a, Hiroyuki Tanaka a, Kazuhito Hioki b,c, Kohei Yamada a, Munetaka Kunishima a,b,
⇑
a Faculty of Pharmaceutical Sciences, Institute of Medical, Pharmaceutical, and Health Sciences, Kanazawa University, Kakuma-machi, Kanazawa 920-1192, Japan
b Cooperative Research Center of Life Sciences, Kobe Gakuin University, 1-1-3 Minatojima Chuo-ku, Kobe 655-8586, Japan
c Faculty of Pharmaceutical Sciences, Kobe Gakuin University, 1-1-3 Minatojima Chuo-ku, Kobe 655-8586, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 24 July 2010
Revised 16 September 2010
Accepted 21 September 2010
Available online 25 September 2010
We studied the specific labeling of avidin with biotinylated modular ligand catalysts via MoAL, which we
recently established. The labeling yield was found to depend on the linker length connecting the catalytic
site to biotin in the modular ligand catalyst 1, and the maximum yield was obtained with 1d possessing
octamethylene linker. The labeling reaction reached a maximum rate with only 4 equiv of the ligand cat-
alyst. Presumably, all the subunits of avidin with homotetrameric structure formed a stable complex with
4 equiv of the catalyst because of the extremely high affinity. The ligand catalyst bound to avidin first cat-
Keywords:
Affinity labeling
Amide-bond formation
Chemical modification of proteins
Avidin
alyzed N-triazinylation of the
e-amino group of Lys111, and the resulting regenerated catalyst then cat-
alyzed the reaction of Asp108 and CDMT.
Ó 2010 Elsevier Ltd. All rights reserved.
The elucidation of protein function is one of the most challeng-
ing and important issues in the post-genomic era. Affinity labeling
(AL), which is based on the inherent interaction between proteins
and their ligands, is a powerful tool for identifying target proteins
and analyzing their binding sites.1–5 In general, AL is performed
using an affinity probe that contains three essential elements in
a single molecule: a ligand with specific affinity for the target bio-
molecule; a labeling tag, such as an isotope or a fluorescent moiety,
to facilitate the isolation and identification of the labeled biomole-
cule; and a reacting group, such as a photophore, to form a cova-
lent bond at the binding site.2 The complicated multifunctional
structure of affinity probes sometimes makes their design and syn-
thesis difficult, and this may deter the wide application of AL for a
variety of protein/ligand systems.
To solve these problems, several useful methods for facilitating
the design and synthesis of affinity probes have been reported.4 Re-
cently, our laboratory has established a convenient modular meth-
od for affinity labeling (MoAL),5 which is based on a catalytic
amide-forming reaction proceeding in an aqueous solution by
using a combination of N,N-dimethylglycine (DMG) esters as a cat-
alyst and 2-chloro-4,6-dimethoxy-1,3,5-triazine (CDMT) as a reac-
tive molecule.6 The distinctive advantages of MoAL are as follows.
(1) Since the three essential elements of affinity probes are com-
pletely separated into individual module molecules (a modular li-
gand catalyst, a labeling module, and CDMT as a reactive module),
design and synthesis of the simplified modules are easy. In fact,
modular ligand catalysts can be readily prepared from ligands via
the introduction of a DMG group as a catalytic site, and a variety
of commercially available amines can be employed as labeling
modules. (2) Proteins can be labeled specifically by only mixing
them with these modules. The carboxyl group of Asp or Glu resi-
dues near the ligand-binding site can be specifically activated by
the ligand catalysts into acyloxytriazines, which then undergo cou-
pling with the amino group of labeling modules to give amide
(Fig. 1). Thus, the site specificity of labeling should depend on
the types of ligand catalysts but not on that of labeling modules.
In our previous report on MoAL using the avidin–biotin system,
we revealed that the labeling reaction proceeded at only Asp108
among three acidic amino acid residues (Asp105, Asp108, and
Asp109) included in the helical region between Ser104 and
Arg114, which is located in the vicinity of the biotin binding site.
In this Letter, we report further studies on avidin labeling with a
biotinylated modular ligand catalyst by MoAL.
First, we examined the effect of the linker length connecting
biotin and the catalytic site (DMG) on the labeling yield. Modular
ligand catalysts (1a and 1c–1e) possessing linkers of various
lengths were synthesized according to the previously reported
method for synthesizing 1b (Fig. 2). Avidin was labeled by simply
mixing it with the ligand catalyst, CBA (Cascade BlueÒ ethylenedi-
amine, trisodium salt) as a labeling module, and CDMT as a reac-
tive module. As shown in Table 1, the highest labeling yield
(53%) per monomeric subunit was observed with ligand catalyst
1d. Since avidin is a homotetrameric glycoprotein and each of its
subunits can equally bind to biotin,7 the labeling yield per protein
should be 210%.8
Interestingly, LC–ESI–MS/MS analysis indicated that the label-
ing reaction proceeded at Asp108 with every ligand catalyst in
⇑
Corresponding author. Tel./fax: +81 76 264 6201.
E-mail address: kunisima@p.kanazawa-u.ac.jp (M. Kunishima).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.09.109

S. Nakanishi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 7050–7053
7051
Figure 1. Specific labeling of proteins by MoAL.
Figure 2. Chemical structure of module molecules.
conducted using 4-bromobenzylamine (BBA) instead of CBA as a
labeling module because BBA facilitated the MS analysis of the
labeled amino acid.5,9 We should note that although the X-ray
structure of avidin (PBD ID: 1LEL) suggests that the length of linker
in 1a is shorter than the distance between the biotin binding site
and Asp108, labeling at Asp108 is also observed with 1a. As the li-
gand catalyst is tightly affixed to the biotin binding site of avidin
because of the high affinity between biotin and avidin (vide infra),
the activated triazino group bound to the catalytic site cannot
reach Asp108. Thus, alternatively, the amino acid undergoing
labeling would need to approach the triazino group at the binding
site, indicating that the conformation of avidin would move during
the labeling reaction. This may indicate that the linker length of a
ligand catalyst used in MoAL need not be designed strictly.
Table 1
Effect of linker length on labeling yield of avidin
Entry
Ligand catalyst
L (Å)
Yield (%)
1
2
3
4
5
1a
1b
1c
1d
1e
25.8
27.0
28.2
36.6
33.0
25
29
36
53
52
The labeling reaction was performed using avidin (20 lM), 1 (80 lM), CBA
(1.6 mM), and CDMT (1.6 mM) in phosphate buffer (pH 8.0) for 8 h at rt. The
labeling yield was determined by UV–vis spectral analysis of labeled avidin,5 and
represents the yield per monomeric subunit.
spite of the different linker lengths, and we did not observe label-
ing at Asp105 or Asp109. The labeling reaction for MS analyses was
Furthermore, we carried out MS analysis in detail to confirm the
site specificity of the amino acid labeling of avidin by MoAL. The

7052
S. Nakanishi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 7050–7053
labeled avidin obtained from 1b and 4-(N,N-dimethylaminosulfo-
nyl)-7-(2-aminoethylamino)-2,1,3-benzoxadiazole (DBD-ED) as a
labeling module as well as the above avidin proteins labeled using
1 and BBA were analyzed. We detected 10 acidic amino acids (both
Asp and Glu, except for Asp13 and Glu128) among 12 acidic amino
acids contained in the avidin monomeric subunit in the peptide
fragments obtained by tryptic digestion. As a result, no signal indi-
cating the introduction of DBD-ED or BBA to the 9 acidic amino
acids (other than Asp108) was detected. Therefore, we can con-
clude that the specific labeling of avidin occurred selectively at
Asp108 in MoAL.
concentration of employed avidin, the reaction of the ligand cata-
lyst binding to avidin with CDMT becomes slow.
ð1Þ
The relationship between the labeling yield and the amount of
ligand catalyst 1d was linear and reached a plateau at 4 equiv of
1d. We did not observe any significant increase in the yield up to
16 equiv of 1d (Fig. 3). In contrast, the labeling yield increased in
a linear fashion with the amount of N,N-dimethylglycine ethyl
ester (DMGE), albeit in a low yield even though it exceeded
4 equiv. In general, in enzyme-like reactions involving the forma-
tion of an enzyme–substrate complex (ES complex), the reaction
rate increases with increasing concentration of the complex. How-
ever, when all enzymes are converted into ES complexes by satura-
tion at high concentrations of the substrate, the reaction rate
reaches a maximum. From the observed results, we may conclude
that all four subunits constituting the avidin molecule are con-
verted into the complexes with only 4 equiv of ligand catalyst 1d
because of the large binding affinity between avidin and biotin
(Ka = 1015 Mꢀ1).10 In other words, the biotinylated catalysts would
bind to avidin in an irreversible fashion. The ligand catalyst added
exceeding 4 equiv did not bind to avidin, and therefore, it could not
contribute to labeling.11
In order to examine whether the binding of the ligand catalysts
with avidin was irreversible during the labeling reaction, we em-
ployed two different procedures for the labeling reaction. First, avi-
din was treated with 4 equiv of biotin for 30 min at rt followed by
the addition of the same amount of ligand catalyst 1d, and the
resulting mixture was subjected to the labeling reaction under
the standard conditions (Eq. (1)). As a result, the labeling yield
dropped to only 2%. On the other hand, when avidin was pre-trea-
ted with 4 equiv of 1d for 30 min followed by addition of the same
amount of biotin and then labeling reaction was carried out
(Eq. (2)), the observed labeling yield was 53%, the same as that in
Table 1, entry 4. These results clearly indicate that avidin interacts
with 1d as well as biotin in an irreversible fashion under the reac-
tion conditions, and that the resulting avidin modified at Asp108
still retains the large binding affinity for 1d. Because of the low
ð2Þ
Thus, we expected that the labeling reaction would proceed
rapidly in a high yield when the ligand catalyst prospectively
incorporating a 1,3,5-triazino group (ligand condensing reagent
2d), which can be prepared from 1d and 2,4-dimethoxy-6-trif-
luoromethanesulfonyloxy-1,3,5-triazine,12 is directly treated with
avidin. However, the reaction of avidin with 2d and CBA at rt for
8 h actually resulted in only a 5% labeling yield (Eq. (3)). The label-
ing yield improved to 42% by further treatment of the resulting
mixture with CDMT for another 8 h (Eq. (4)). On the basis of above
MS analysis, we concluded that the triazino group of 2d bound to
the biotin binding site preferably reacted with Lys111 over
Asp108, and therefore, the labeling yield decreased. After transfer-
ring the triazino group to Lys111, the N,N-dimethylamino group of
the regenerated 1d, tightly bound to the avidin molecule, could
then catalyze the conversion of the carboxylate of Asp108 into a
triazino ester, and therefore, the labeling of avidin proceeded only
in the presence of CDMT. The observed preference for the amine is
in marked contrast to that using 4-(4,6-dimethoxy-1,3,5-triazin-2-
yl)-4-methyl-morpholinium chloride (DMT-MM) and other ammo-
niotriazines, in which the triazino group reacts exclusively with
carboxylic acids over amines.13 Since Asp108 and Lys111 are
known to form a salt bridge in the X-ray structure of avidin,14
the observed unusual amine selectivity with 2d in the present sys-
tem may be attributed to the specific interaction between the two
residues. In fact, it was reported that Lys111 readily underwent
intermolecular arylation with 2,4-dinitro-1-fluorobenzene.14b
ð3Þ
ð4Þ
60
50
40
30
20
10
0
In the cyclotransferase model involving a tertiary amine cata-
lyst coupled to 18-crown-6, which is based on a similar mechanism
as that of MoAL, 96% of the substrate selectivity with 78% yield was
realized on the basis of a binding constant of ꢁ104 Mꢀ1 between
18-crown-6 and the primary ammonium in MeOH.15 In marked
contrast, in spite of the extremely high affinity between avidin
and biotin, the observed labeling yield was not as high as we had
expected. In the case of 18-crown-6 system, all the catalyst mole-
cules can work repetitively for the reaction and contribute to
increasing the yield because of the reversible formation of the
complex between the catalyst and substrates. In the present study,
on the other hand, the ligand catalysts 1 do not work very
efficiently presumably due to both the irreversible binding of 1
to avidin and the preferential reactivity of Lys111 over Asp108.
0
2
4
6
8
10 12 14 16 18
equiv. ratio of catalyst to avidin
Figure 3. Correlation between the amount of tert-amine catalyst and labeling yield.
The labeling reaction was performed using 1, avidin (20 M), CBA (1.6 mM), and
CDMT (1.6 mM) in phosphate buffer (pH 8.0) for 8 h at rt, and the concentration of 1
l
was changed from 20 lM to 320 lM. The labeling yield was determined by UV–vis
spectral analysis of labeled avidin,5 and represents the yield per monomeric
subunit. The labeling yield was plotted against the amount of the catalyst in equiv
(N: modular ligand catalyst 1d; d: DMGE).
Since the ligand catalyst 1 (80 lM) should immediately bind to avi-

S. Nakanishi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 7050–7053
7053
2. (a) Kotzyba-Hibert, F.; Kapter, I.; Goeldner, M. Angew. Chem., Int. Ed. Engl. 1995,
34, 1296; (b) Fleming, S. A. Tetrahedron 1995, 51, 12479; (c) Doman, G.;
Prestwich, G. D. Trends Biotechnol. 2000, 18, 64; (d) Campbell, D. A.;
Szardenings, A. K. Curr. Opin. Chem. Biol. 2003, 7, 296; (e) Robinette, D.;
Neamati, N.; Tomer, K. B.; Borchers, C. H. Expert Rev. Proteomics 2006, 3, 399; (f)
Vodovozova, E. L. Biochemistry (Moscow) 2007, 72, 1; (g) Hashimoto, M.;
Hatanaka, Y. Eur. J. Org. Chem. 2008, 2513; (h) Leslie, B. J.; Hergenrother, P. J.
Chem. Soc. Rev. 2008, 37, 1347.
3. (a) Levitsky, K.; Ciolli, C. J.; Belshaw, P. J. Org. Lett. 2003, 5, 693; (b) Chen, G.;
Heim, A.; Riether, D.; Yee, D.; Milgrom, Y.; Gawinowicz, M. A.; Sames, D. J. Am.
Chem. Soc. 2003, 125, 8130; (c) Levitsky, K.; Boersma, M. D.; Ciolli, C. J.;
Belshaw, P. J. ChemBioChem 2005, 6, 890; (d) Krusemark, C. J.; Belshaw, P. J. Org.
Biomol. Chem. 2007, 5, 2201; (e) Tanaka, K.; Fujii, Y.; Fukase, K. ChemBioChem
2008, 9, 2392.
4. (a) Tsukiji, S.; Miyagawa, M.; Takaoka, Y.; Tamura, T.; Hamachi, I. Nat. Chem.
Biol. 2009, 5, 341; (b) Koshi, Y.; Nakata, E.; Miyagawa, M.; Tsukiji, S.; Ogawa, T.;
Hamachi, I. J. Am. Chem. Soc. 2008, 130, 245; (c) Wakabayashi, H.; Miyagawa,
M.; Koshi, Y.; Takaoka, Y.; Tsukiji, S.; Hamachi, I. Chem. Asian J. 2008, 3, 1134;
(d) Mayer, T.; Maier, M. E. Eur. J. Org. Chem. 2007, 4711; (e) Hosoya, T.;
Hiramatsu, T.; Ikemoto, T.; Nakanishi, M.; Aoyama, H.; Hosoya, A.; Iwata, T.;
Maruyama, K.; Endo, M.; Suzuki, M. Org. Biomol. Chem. 2004, 2, 637; (f) Kan, T.;
Kita, Y.; Tominari, Y.; Morohashi, Y.; Natsugari, H.; Tomita, T.; Iwatsubo, T.;
Fukuyama, T. Chem. Commun. 2003, 2244.
din prior to reaction with CDMT to give 2, the actual concentration
of 1 could be as low as that of avidin itself (20 M). Thus, the rate
l
of generation of 2 from 1 at the biotin binding site of avidin should
much decrease. In addition, this slow step is necessary to repeat at
least twice because the intended labeling of Asp108 takes place
only after the reaction of 2 with Lys111. Decomposition of the
ligand catalyst by the hydrolysis or N-demethylation of 2,13 or
the hydrolysis of the ester group of 116 during the slow labeling
reaction might be responsible for decreasing the labeling yields.
Thus, the too strong binding ability of biotin with avidin does not
allow replacement of 1 bound to avidin with free 2, and therefore,
would be paradoxically responsible for the suppression rather than
the promotion of the labeling reaction. After all, the labeling yield
increased with both the increasing concentration of CDMT
(3.2 mM, y. 63%) or a prolonged reaction time (24 h, y. 77%).
In summary, we revealed the details of the specific labeling of
avidin by MoAL. The labeling reaction occurred specifically at
Asp108 at a significant level even if the linker lengths of the ligand
catalysts differed. This may indicate that the linker length of the li-
gand catalyst does not need to be strictly designed. MoAL would be
useful not only in searching for target proteins but also in the con-
formational analysis of the proteins.
5. Kunishima, M.; Nakanishi, S.; Nishida, J.; Tanaka, H.; Morisaki, D.; Hioki, K.;
Nomoto, H. Chem. Commun. 2009, 5597.
6. (a) Kunishima, M.; Yoshimura, K.; Morigaki, H.; Kawamata, R.; Terao, K.; Tani, S.
J. Am. Chem. Soc. 2001, 123, 10760; (b) Kunishima, M.; Imada, H.; Kikuchi, K.;
Hioki, K.; Nishida, J.; Tani, S. Angew. Chem., Int. Ed. 2005, 44, 7254.
7. Livnah, O.; Bayer, E. A.; Wilchek, M.; Sussman, J. L. Proc. Natl. Acad. Sci. U.S.A.
1993, 90, 5076.
Acknowledgment
8. The labeling yield shown throughout the Letter represents that of a monomeric
subunit.
9. (a) Sinz, A. Angew. Chem., Int. Ed. 2007, 46, 660; (b) Lamos, S. M.; Krusemark, C.
J.; McGee, C. J.; Scalf, M.; Smith, L. M.; Belshaw, P. J. Angew. Chem., Int. Ed. 2006,
45, 4329.
This work was supported partially by Grant-in-Aid for Science
Research (No. 20390007) from the Ministry of Education, Culture,
Sports, Science and Technology, Japan.
10. Green, N. M. Biochem. J. 1963, 89, 585.
11. We ignored the rate of the intermolecular labeling reaction by free ligand
catalysts because it was remarkably slow under the low concentration
conditions we employed.
Supplementary data
12. Kunishima, M.; Hioki, K.; Wada, A.; Kobayashi, H.; Tani, S. Tetrahedron Lett.
2002, 43, 3323.
13. Kunishima, M.; Kawachi, C.; Morita, J.; Terao, K.; Iwasaki, F.; Tani, S.
Tetrahedron 1999, 55, 13159.
14. (a) Publiese, L.; Coda, A.; Malcovati, M.; Bolognesi, M. J. Mol. Biol. 1993, 231,
698; (b) Gitlin, G.; Bayer, E. A.; Wilchek, M. Biochem. J. 1987, 242, 923.
15. Kunishima, M.; Moriya, T.; Morita, J.; Ikuta, T.; Hioki, K.; Tani, S. Angew. Chem.,
Int. Ed. 2006, 45, 1252.
16. It is reported that avidin catalyzes hydrolysis of biotin p-nitrophenyl ester at
the binding site, see: Huberman, T.; Eisenberg-Domovich, Y.; Gitlin, G.; Kulik,
T.; Bayer, E. A.; Wilchek, M.; Livnah, O. J. Biol. Chem. 2001, 276, 32031.
Supplementary data (general methods, experimental and ana-
lytical data for the synthesized compounds) associated with this
article can be found, in the online version, at doi:10.1016/
j.bmcl.2010.09.109.
References and notes
1. Jakoby, W. B.; Colowick, S. P.; Wilchek, M. Methods Enzymol. 1977, 46, 3.