The high performance of 3XFLAG for target purification of a bioactive metabolite: A tag combined with a highly effective linker structure

Article abstract of DOI:10.1016/j.bmcl.2011.01.038

Affinity purification using immunoprecipitation (IP) is an extremely useful method for target profiling of bioactive natural products. We examined IP purification of CMetE, which is a molecular target for potassium isolespedezate (1), a leaf-opening factor of Cassia plant. We studied IP efficiency using a panel of FLAG-connected molecular probes (2-8), including probes with varying structures and lengths of the linker moiety. The results suggest that not only the length, but the chemical nature of the linker moiety, strongly affect the IP efficiency. 3XFLAG, a tag combined with a linker moiety of charged amino acids, gave the best results and was most useful for IP purification of the molecular target.

Full text of DOI:10.1016/j.bmcl.2011.01.038

Bioorganic & Medicinal Chemistry Letters 21 (2011) 1359–1362
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The high performance of 3XFLAG for target purification of a bioactive
metabolite: A tag combined with a highly effective linker structure
⇑
Minoru Ueda , Yoshiyuki Manabe, Makoto Mukai
Department of Chemistry, Tohoku University, 6-3 Aramaki-Aza Aoba, Sendai 980-8578, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Affinity purification using immunoprecipitation (IP) is an extremely useful method for target profiling of
bioactive natural products. We examined IP purification of CMetE, which is a molecular target for potas-
sium isolespedezate (1), a leaf-opening factor of Cassia plant. We studied IP efficiency using a panel of
FLAG-connected molecular probes (2–8), including probes with varying structures and lengths of the lin-
ker moiety. The results suggest that not only the length, but the chemical nature of the linker moiety,
strongly affect the IP efficiency. 3XFLAG, a tag combined with a linker moiety of charged amino acids,
gave the best results and was most useful for IP purification of the molecular target.
Received 22 December 2010
Revised 8 January 2011
Accepted 11 January 2011
Available online 14 January 2011
Keywords:
Target protein
Immunoprecipitation
Molecular probe
Linker
Ó 2011 Elsevier Ltd. All rights reserved.
Target identification of bioactive natural products that induce
physiologically intriguing phenomenon would launch a new field
of chemical biology due to the fact that most targets remain un-
linkers, as well as one without a linker (2). Additionally, 3XFLAG,
which has high affinity for an anti-FLAG antibody and is widely
used as an elution ligand in IP purification, was examined as a
tag combined with a linker moiety in probe 8.
1
known. Chemical tagging combined with immunoprecipitation
(
IP) is a most promising method to purify molecular targets of a nat-
Syntheses of these probes were carried out as shown in
Schemes 1, 2, S1 and S2 (for S1 and S2, see Supplementary data).
Coupling with the FLAG tag was performed on solid beads. ‘Click-
ural product. Recently, we succeeded in the extremely effective IP
purification of a target protein for isolespedezate (1), which is the
2
4
leaf-opening factor of Cassia plants. The identified molecular target,
CMetE (Cassia 5-methyltetrahydropteroyltriglutamate-homocyste-
ine-methyltrans-ferase), was purified by using 3XFLAG (DYKDHDG-
DYKDHDIDYKDDDDK) with no non-specific bindings. In this Letter,
we examine the effectiveness of a panel of linkers varying in struc-
ture and length, and report that both the structural nature and the
length of the linker moiety are controlling factors for successful IP
purification.
The panel of FLAG probes used in the IP experiment is shown in
Figure 1. It is generally accepted that the role of the linker is to pro-
ject the large molecular tag away from the pharmacophore, which
is essential for binding to a specific target. Thus, the longer the lin-
made’ synthesis of each probe rendered the preparation of the pa-
nel compounds quite easy (Scheme 2). For both glycyl and PEG
linkers, the solubility in aqueous solution was gradually decreased
according to the addition of linker units.
After chemical tagging of the live Cassia motor cells with each
probe, the cells were disrupted and applied to IP using beads bear-
ing the anti-FLAG antibody according to the procedure in Ref. 2
(Fig. 2). The protein trapped on the beads was eluted using 3XFLAG
as a free ligand. The results are summarized in Figure 3. The effec-
tiveness of the IP purification was assessed by the ratio of CMetE in
the supernatant to CMetE trapped on the IP beads. This ratio anal-
ysis enables the performance of the linker to be assessed regardless
of the difference in chemical yield of the affinity labeling among
the different chemical probes.
As shown in Table 1, longer glycyl linkers provided more effective
trapping of CMetE on IP beads. Increasing the number of glycyl units,
for example, zero in 2 (sup/ppt = 8:2), three in 3 (sup/ppt = 6:4), six
in 4 (sup/ppt = 4:6), and 15 in 5 (sup/ppt = 1:9), improved the IP effi-
ciency. However, increasing the number of PEG units did not im-
prove the IP result, for example, zero in 2 (sup/ppt = 8/2), six in 6
3
ker, the more effective the projection. Tanaka et al. reported that a
long hydrophilic linker is indispensable for the reduction of non-
specific bindings and successful IP purification of a molecular tar-
get. We examined hydrophilic glycyl and polyethyleneglycol (PEG)
linkers. The use of an octapeptidyl FLAG tag also served to increase
hydrophilicity of the probes. We prepared a panel consisting of
probes possessing a FLAG-tag connected to triglycyl (3), hexaglycyl
(4), pentadecaglycyl (5), hexa-PEG (6), and octadeca-PEG (7)
(
sup/ppt = 7/3), and 18in 7 (NDin both pptand sup). It is noteworthy
that little improvement was observed by the insertion of a hexa-PEG
linker (cf. 2 and 6). This suggested that the long PEG moiety did not
⇑
Corresponding author.
E-mail address: ueda@m.tohoku.ac.jp (M. Ueda).
0
960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.01.038

1
360
M. Ueda et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1359–1362
Figure 1. Structures of molecular probes used in the IP purification of CMetE.
Scheme 1. Preparation of a panel of alkyne-connected FLAG tags bearing various linkers. (A) Syntheses of alkyne units bearing various linkers; (B) solid phase synthesis of
FLAG-connected alkyne units.

M. Ueda et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1359–1362
1361
Scheme 2. Preparation of a panel of probes bearing various linkers.
work as well as the glycyl linker. Moreover, the effectiveness of the
linker cannot be explained only by its length since the lengths of
the hexa-PEG linker in 6 and the hexaglycyl linker in 4 are almost
the same. However, a distinct difference was found between the
hexaglycyl (4: sup/ppt = 4:6) and hexa-PEG (6: sup/ppt = 7:3)
probes. The superiority of the glycyl linker over the PEG linker is also
confirmed by this result, suggesting that the chemical nature of the
linker structure is also important for high performance.
5
The calculated log P values of probes of various linker moieties
6
(
26–31) are shown in Table 2. The log D value was also examined
for31inwhichmanyionizedfunctionalitiesareembedded. Compar-
ison of calculated log P values for the series of glycyl-linked probes
(
26: log P = À1.9; 27: log P = À3.6; 28: log P = À8.5) suggested that
longer linkers were more hydrophobic. Thus, the projection worked
well and longer linkers gave better IP results. The extremely high
performance of 3XFLAG can be explained by the log D value of 31
(
31: log D = À10.2) instead of the log P value (31: log P = À6.9), be-
cause log D gives a better prediction of hydrophobicity for highly
ionized compounds. On the other hand, the calculated log P value
6
for the ineffective hexa-PEG 29 (log P = À1.3) was as high as triglycyl
2
6 (log P = À1.9), and much higher than that of hexaglycyl 27
(
log P = À3.6), whereas little difference in IP results was observed
among them. Additionally, CMetE could not be recovered in either
sup or ppt with octadeca-PEG 30 (log P = À3.9) possessing a longer
linker, and a substantial amountof nonspecific bindingwas obtained
Figure 2. Procedures for IP purification of CMetE from motor cells of Cassia
octadentalis.
(
Fig. 3), even though the log P value of 30 (log P = À3.9) was as low as
that of hexaglycyl 27 (log P = À3.6), which gave a moderate IP result
sup/ppt = 4/6). In general, the increase in nonspecific binding of
(
7
proteins can be attributed to the increase in the hydrophobicity.
However, these results revealed that hydrophobicity alone cannot
explain the performance of the linkers.
The aforementioned results suggested that the long linker moi-
ety with its low log P value does not work well and can also detract
from the selectivity of IP purification due to the increase in non-
specific bindings. This could imply that the long and comparatively
hydrophobic linker structure predominantly adopts a folded con-
formation, and thus cannot separate both ends effectively. The
poor IP result using probe 7 could be attributed to such nature of
long PEG linker. Charged oligopeptides, such as 3XFLAG, which
possesses as many as 11 aspartate residues in the 22 amino acid
sequence, are expected to be useful in improving both efficiency
and selectivity in IP purification because they can predominantly
adopt extended conformations in aqueous solution due to the elec-
tronic repulsion between charged aspartate residues. Thus, signif-
icant improvement was achieved when 3XFLAG was used. 3XFLAG
is a large molecular tag combined with a linker unit, and is ex-
pected to be of high affinity to anti-FLAG antibodies immobilized
on IP beads. The ratio of CMetE in sup to CMetE on IP beads was
as high as 1/99. Considering that the length of 3XFLAG is similar
to that of the pentadecaglycyl FLAG, the extremely high
performance of 3XFLAG could be attributed to higher affinity to
the anti-FLAG antibody as well as to the difference in structures be-
Figure 3. SDS–PAGE analysis of proteins trapped on IP beads using probes 2 and 7.

1
362
M. Ueda et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1359–1362
Table 1
Comparison of efficiencies in IP purification by a panel of probes (2–8)
Linker + Tag
Bands @ 83 kDa
MetEsup/ppt
Sup
Beads
Sup
Beads
-
-
-
-
-
-
-
FLAG (2)
8
2
(Gly)
(Gly)
3
-FLAG (3)
-FLAG (4)
6
4
6
4
6
(Gly)15-FLAG (5)
(PEG) -FLAG (6)
1
9
6
7
3
(PEG)18-FLAG (7)
3ÂFLAG (8)
—
<1
—
>99
Table 2
Calculated log P (log D) values for various linker structures
Acknowledgments
(
26–31) which correspond to linker moieties of probes 3–8
This work was supported by Grants-in-Aid for Scientific Re-
search from MEXT (No. 20310128), the Asahi-Glass Foundation,
the TOREY Foundation, and the Naito Foundation. We also thank
the Experimental Station for Medicinal Plant Studies, Tohoku Uni-
versity, for cultivation of plants.
*
Linker
log P (or log D )
Ac-(Gly)
Ac-(Gly)
3
-OMe (26)
-OMe (27)
À1.9
6
À3.6
Ac-(Gly)15-OMe (28)
Ac-(PEG) -OMe (29)
Ac-[(PEG) -OMe (30)
Ac-DYKDHDGDYKDHDI-OMe
À8.5
6
À1.3
]
6 3
À3.9
À6.9 (À10.2)
*
(31)
Supplementary data
*
log D was calculated for 31.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.01.038. These data
include Schemes S1 and S2, experimental method, and HRMS as
well as LC–MS/MS data of alkyne units 18–22, 28, 30, and probes
tween the 15-glycyl and aspartate-rich 14-aminoacyl (DYKDHDG-
DYKDHDI) moieties.
2–8.
Our present studies suggest that the nature of the linker struc-
ture strongly affects the IP purification of molecular targets for a
small bioactive molecule. The use of good linkers improves the effi-
ciency and selectivity in IP purification by decreasing nonspecific
binding of proteins, which often causes significant obstacles for
target identification. The performance of the linker was not only
determined by its length but also by its nature, including hydro-
phobicity and conformation in an aqueous solution. Our study
using a panel of probes with various linkers concluded that the
References and notes
1.
(a) Jessani, N.; Cravatt, B. F. Curr. Opin. Chem. Biol. 2004, 8, 54; (b) Evans, M. J.;
Cravatt, B. F. Chem. Rev. 2006, 106, 3279.
2
3
4
.
.
.
Manabe, Y.; Mukai, M.; Ito, S.; Kato, N.; Ueda, M. Chem. Commun. 2010, 46, 469.
Tamura, T.; Terada, T.; Tanaka, A. Bioconjug. Chem. 2003, 14, 1222.
(a) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int. Ed. 2001, 40, 2004;
(b) Tornøe, C. W.; Christensen, C.; Meldal, M. J. Org. Chem. 2002, 67, 3057.
(a) Tekto, I. V.; Tanchuk, V. Y.; Villa, A. E. P. J. Chem. Inf. Comput. Sci. 2001, 41,
5.
6.
7.
3
XFLAG as a tag combined with a linker provided the best result.
1407; (b) Tekto, I. V.; Tanchuk, V. Y. J. Chem. Inf. Comput. Sci. 2002, 42, 1136.
To the best of our knowledge, this study is the first example to
use 3XFLAG as a tag for affinity purification of a target protein
for a small molecule. This result will assure the promising perfor-
mance of 3XFLAG as a high-performance molecular tag for IP
purification of targets for small bioactive molecules.
(a) Tekto, I. V.; Poda, G. I. J. Med. Chem. 2004, 47, 5601; (b) Tekto, I. V.; Bruneau, P.
J. Pharma. Sci. 2004, 93, 3103.
Hofstee, B. H. Biochem. Biophys. Res. Commun. 1973, 53, 1137.