A mechanism-based affinity-labeling agent for possible use in isolating N-acetylglucosaminidase

Article abstract of DOI:10.1016/S0960-894X(01)00300-6

We have prepared several mechanism-based affinity-labeling agents for possible use in isolating N-acetylglucosaminidase, in which an N-acetylglucosamine is linked to an o-monofluoro- or difluoro-methyl phenoxy glycoside with or without a cleavable disulfide group in the tether to biotin.

Full text of DOI:10.1016/S0960-894X(01)00300-6

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1769–1773
A Mechanism-Based Affinity-Labeling Agent for Possible
Use in Isolating N-Acetylglucosaminidase
Mie Ichikawa and Yoshitaka Ichikawa*
Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA
Received 16 March 2001; accepted 24 April 2001
Abstract—We have prepared several mechanism-based affinity-labeling agents for possible use in isolating N-acetylglucosaminidase,
in which an N-acetylglucosamine is linked to an o-monofluoro- or difluoro-methyl phenoxy glycoside with or without a cleavable
disulfide group in the tether to biotin. # 2001 Elsevier Science Ltd. All rights reserved.
Glycosidases hydrolyze glycoconjugates, including gly-
coproteins, glycolipids, DNA, RNA, glycosamino-
glycans, and peptidoglycans, into sugars and aglycons
and consequently play critical roles in carbohydrate
metabolism,1,2 DNA repair mechanisms,3 and cell phy-
siology.4 Therefore, it is important to identify and
characterize the specific enzymes involved in order to
understand their direct roles in such biological events.
Although several affinity-labeling agents have been
developed (sugars with an epoxy-5 or a-haloketo-con-
taining aglycon6 or fluorinated sugars)7 for active-site
mapping, only a few iminosugar-based inhibitors with a
photoaffinity-labeling group have been used in attempts
to identify specific glycosidases on SDS-PAGE.8,9
of its hydrophobic nature the acetophenone-based
compound tended to exhibit nonspecific interaction
with other co-existing proteins that may bind to the
GlcNAc moiety), and subsequently excised and
sequenced the radioactive bands from SDS-PAGE gel
(G. Parker and Y. Ichikawa, unpublished).13 Dis-
appointingly we were unable to identify the enzyme
sequence. We reasoned that, because there were a large
number of other proteins comigrating with the radioactive
O-GlcNAc’ase on the SDS-PAGE gel, we might have
sequenced one of them instead of the target enzyme. We
therefore decided to develop a procedure that allows us
to effectively isolate/purify a specific glycosidase.
O-Linked N-acetylglucosaminidase (O-GlcNAc’ase)
cleaves the O-glycosidic linkage between GlcNAc and
Ser/Thr residues of proteins to liberate the free OH
group of the Ser/Thr residue for subsequent phosphoryl-
ation by a kinase.10 This on–off glycosylation of Ser/Thr
residues, and therefore the activity of the enzyme, is
thought to be responsible for regulating the biological
function of many nuclear and cytosolic proteins.11
However, this important enzyme has not yet been com-
pletely isolated or characterized because of the lack of a
suitable affinity-labeling agent.12 We had previously
prepared several benzophenone-based GlcNAc thioglyco-
sides and examined the photolabeling of O-GlcNAc’ase
partially purified from rat spleen.13 After introducing a
radioactive marker on the photoaffinity-labeled pro-
teins, we observed ꢀ5 positive bands on an SDS-PAGE
gel together with several minor bands (probably because
Janda and co-workers have reported the use of a biotin-
conjugated mechanism-based tagging agent in the iso-
lation of an antibody with a b-galactosidase-like activity
from a phage library.14 Because no standard protocol
has yet been established for isolating a specific glycosi-
dase, we decided to evaluate and further develop such a
mechanism-based tagging agent as a tool for isolating
glycosidases, including O-GlcNAc’ase.
A possible mechanism for such tagging (affinity label-
ing) or inactivation of an enzyme, mediated by a 2-halo-
benzyl-phenol group, has been well documented.15,16 In
*Corresponding author. Tel.: +1-410-614-2892; fax: +1-410-955-
3023; e-mail: ichikawa@jhmi.edu
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00300-6

1770
M. Ichikawa, Y. Ichikawa / Bioorg. Med. Chem. Lett. 11 (2001) 1769–1773
our case, when an enzyme cleaves a b-GlcNAc linkage
(stage I!II), a reactive methide is generated with a loss
of HF. The methide, at stage II, serves as a Michael
acceptor to react with a potential nucleophile (i.e., the
carboxylate of Asp or Gln) inside or near the active site
to form a covalent linkage between the enzyme and the
released aryl moiety (stage III), which could be easily
isolated by streptavidin-based affinity column chroma-
tography.
with different tethers for comparison: agent A19 pos-
sesses a stable alkyl chain-based tether, and agent B19
carries, on the other hand, a cleavable disulfide group in
the tether. These were added by reacting 6 with com-
mercially available biotin LCLC NHS20 and biotin dis-
ulfide NHS,20 respectively. In addition, we also
prepared another disulfide-containing biotin conjugate
C, which lacks N-acetylglucosaminyl moiety, by treating
a commercially available Biotin-LC NHS20 with cysta-
mine in order to see a stability of the disulfide group.
Synthesis of Difluoromethyl Aryl ꢀ-N-Acetyl-D-gluco-
saminide (Scheme 1)
Affinity-Labeling Experiments
The basic strategy for synthesis of the tagging agents
was similar to that by Janda14 and Danzin.16 N-Acetyl-
glucosaminyl chloride 117 was treated with 2-hydroxy-5-
A 50% aq DMSO solution of each agent A–D was
added, to a final concentration of 2 mM (4% DMSO
final concentration), to a sodium cacodylate buffer
(100 mL, 50 mM, pH 6.5) containing a partially purified
O-GlcNAc’ase,10 EDTA (10 mM), Triton X-100 (1%),
and NaCl (200 mM), and the solution was incubated for
2 h at 37 ꢁC. An aliquot (10 mL) was taken from each
incubation mixture and solubilized in SDS sample buf-
fer (under nonreducing conditions). The remaining
mixture was treated with pretreated streptavidin-agar-
ose beads (40 mL, Pierce) overnight at 4 ꢁC. The beads
were washed three times with a 100 mM Tris pH 7.0
solution containing 2.5 M KCl and 1% Triton X-100
(the suspension was centrifuged and the supernatant
was removed by pipeting), and boiled in SDS sample
nitro-benzaldehyde
butylammonium bromide in a mixture of 5% NaOH
2
in the presence of tetra-
to give the b-glycoside 319 in 20–30%
16,18
and CH2Cl2
yield. Treatment of 3 with diethylaminosulfurtrifluoride
(DAST) gave a difluoromethyl aryl derivative 419 in
60% yield. The nitro group of 4 was reduced to an
amino group, which was coupled with N-Boc-6-amino-
hexanoic acid in the presence of 1-(3-dimethylamino-
propyl)-3-ethylcarbodiimide hydrochloride (EDAC) to
give 519 in 74% overall yield. Acidic treatment of 5
generated the free amino group that was used to prepare
the biotin conjugates. We have prepared two agents
Scheme 1. Synthesis of difluoromethyl aryl-containing tagging agents. Reagents and conditions: (a) Bu4NBr/5% NaOH–CH2Cl2 (24%); (b) DAST/
CH2Cl2 (60%); (c) (i) H2/Pd–C/EtOAc; (ii) 6-N-Boc-aminohexanoic acid/EDAC/THF (74% overall); (d) CF3CO2H–CH2Cl2 (1:4)/0 ꢁC; (e) (i) biotin
LCLC NHS/Et3N/DMF; (ii) EtOH–H2O–Et3N (4:1:1)/rt; (f) (i) biotin disulfide N-hydroxysuccinimide ester/Et3N/DMF; (ii) EtOH–H2O–Et3N
(4:1:1)/rt; (g) cystamine/Et3N/DMF.

M. Ichikawa, Y. Ichikawa / Bioorg. Med. Chem. Lett. 11 (2001) 1769–1773
1771
buffer (under non-reducing conditions). The eluted pro-
teins were separated by SDS-PAGE (10% gel) and
electrophoretically transferred to nitrocellulose (Fig. 1).
The protein bands were visualized with alkaline phos-
phatase-streptavidin and BCIP/NBT. In separate
experiments, when the enzyme preparation was treated
with agents A, B, and C and O-deacetylated forms of 4,
5, 8, and 9, we did not observe the apparent time-
dependent inactivation of the enzymes (O-GlcNAc’ase
activity), which may indicate the labeling takes place
outside the enzyme active site.14 When the affinity-
labeling experiments were conducted in the presence of
a thio-glycoside derivative 11 (see ref 13), which had
been identified as a potent inhibitor of O-GlcNAc’ase
(Ki=50ꢂ5 mM), the tagging agents A, B, and D were
not hydrolyzed (spectrophotometric and TLC analysis),
and consequently, no protein bands were observed in
SDS-PAGE by the Western blotting.
conjugate might have readily become ionized to facil-
itate the formation of an active methide if a fluorine
atom still remains (I!II!III, Fig. 2), which would
react with water to form a hemiacetal (III!IV). The
hemiacetal would easily undergo formation of the stable
aldehyde, with a concomitant loss of the once-hooked
enzyme (IV!V).
In the case of agent B with a disulfide group in the
tether, we observed a pattern of protein bands before
the affinity purification that was similar to that obtained
with agent A (lane 3, Fig. 1) and one major band
retained after the affinity purification (lane 4). To our
surprise, however, these protein bands in lane 3 were
also seen when we used agent C, a disulfide-containing
biotin conjugate without a GlcNAc aryl moiety (lane 5).
We therefore speculated that a sulfide-exchange reaction
might have taken place during the incubation (Fig. 3).
The enzyme preparation had other proteins present,
including GlcNAc-binding and free cysteine-containing
proteins such as VI in Fig. 3. In this situation, when
the tagging agent binds to a GlcNAc-binding protein, a
free cysteine near the binding site reacts with the di-
sulfide group to cause a sulfide-exchange reaction
(VII).22 In path a (Fig. 3), a cysteine reacts with a sulfur
atom close to the biotin to form a new disulfide bond,
with a biotin with a shorter tether length, which is visi-
ble after Western blotting; however, this biotin-con-
jugate cannot bind to streptavidin-agarose beads
probably because only a small portion of agent C
undergoes the disulfide-exchange reaction to form biotin-
conjugates or the tether may be too short to be retained
with streptavidin beads.23 On the other hand, if the
cysteine reacts with a sulfur atom away from the biotin,
An affinity tagging experiment in which a partially pur-
ified O-GlcNAc’ase from rat spleen was treated with the
difluoromethyl aryl alkyl-tethered tagging agent A
showed that several protein bands were tagged by the
biotin (lane 1); however, these bands stained very
weakly and were not detected after the attempted affi-
nity purification probably because only a small fraction
of the proteins were conjugated with the biotin (lane 2).
A possible explanation for this low tagging efficiency
might involve the difluoromethyl group. After the
covalent linkage was formed between the tagging agent
A and the enzyme, these conjugates were still present in
the incubation mixture for 2 h in the presence of 4%
DMSO which helped solubilize the tagging agent.
Under such conditions,21 the phenol group of the
Figure 1. SDS-PAGE of the affinity-tagging experiments. After incubation with each biotin-conjugated tagging agent (2 mM of A, B, C, or D), the
samples were separated by SDS-PAGE before (lanes 1, 3, 5, and 7) and after affinity purification with streptavidin-agarose beads (lanes 2, 4, 6, and
8). The biotinylated protein bands were visualized with streptavidin-alkaline phosphatase and BCIP/NBT.
Figure 2. A possible explanation for disappearance of the protein bands after affinity purification.

1772
M. Ichikawa, Y. Ichikawa / Bioorg. Med. Chem. Lett. 11 (2001) 1769–1773
Figure 3. A possible explanation for protein tagging via a disulfide group.
Scheme 2. Synthesis of a monofluoromethyl aryl-containing tagging agent. Reagents and conditions: (a) NaBH3CN/THF/aq HCl; (b) DAST/
CH2Cl2 (60%); (c) (i) H2/Pd–C/EtOAc; (ii) 6-N-Boc-aminohexanoic acid/EDAC/THF (74% overall); (d) CF3CO2H–CH2Cl2 (1:4)/0 ꢁC; (e) (i) biotin
LCLC NHS/Et3N/DMF; (ii) EtOH–H2O–Et3N (4:1:1)/rt.
References and Notes
as in path b, the GlcNAc-aryl group becomes attached to
the protein which is not visible after Western blotting.
1. Herscovics, A. Biochim. Biophys. Acta 1996, 1473, 96.
2. Kolter, T.; Sandhoff, K. Angew. Chem., Int. Ed. 1999, 38,
1532.
3. (a) McCullough, A. K.; Dadson, M. L.; Lloyd, R. S. Annu.
Rev. Biochem. 1999, 68, 255. (b) Lindahl, T.; Wood, R. D.
Science 1999, 286, 1897.
Synthesis of Another Tagging Agent with a Mono-
fluoromethyl Aryl Group
4. Dempsey, L. A.; Brunn, G. J.; Platt, J. L. Trends Biol. Sci.
2000, 25, 349.
5. Thomas, E. W.; McKelvy, J. F.; Sharon, N. Nature 1969,
222, 485.
6. Howard, S.; Withers, S. G. Biochemistry 1998, 37, 3858.
7. (a) Withers, S. G.; Street, I. P.; Bird, P.; Dolphin, D. H. J.
Am. Chem. Soc. 1987, 109, 7530. (b) Withers, S. G.; Street, I. P.
J. Am. Chem. Soc. 1988, 110, 8551.
8. Kuhn, C.-S.; Lehmann, J.; Sandhoff, K. Bioconjugate J.
1992, 3, 230.
Because we hypothesized that a monofluoromethyl
group would prevent such ‘release’ of the already tagged
enzyme as depicted in Fig. 2, we decided to prepare
another type of tagging agent with a monofluoromethyl
aryl group (Scheme 2). The aldehyde 2 was subjected to
reduction with NaBH3CN16 to convert it to the corre-
sponding alcohol 7. The rest of the synthetic strategy
was almost identical to that for the synthesis of a
difluoromethyl derivative, agent A.
9. Liessem, B.; Glombitza, G. J.; Knoll, F.; Lehman, J.; Kel-
lerman, J.; Lottspeich, F.; Sandhoff, K. J. Biol. Chem. 1995,
270, 23693.
10. Dong, D. L.-Y.; Hart, G. W. J. Biol. Chem. 1994, 269,
19321.
We then performed affinity tagging experiments with
the monofluoromethyl aryl agent D. When the enzyme
was treated with tagging agent D and analyzed by
SDS-PAGE, one major protein band was visible (lane
7, Fig. 1) which was retained after affinity purifica-
tion (lane 8). Its molecular weight was about 51,000
which is consistent with the reported value of O-
GlcNAc’ase from rat spleen.10 We are now in the pro-
cess of isolating this protein band and microsequencing
it.
11. Snow, D. M.; Hart, G. W. Int. Rev. Cytol. 1998, 181, 43.
12. While we were studying the enzyme from rat spleen, the
Hart group reported the first cloning of O-GlcNAc’ase from
human brain; see: Gao, Y.; Wells, L.; Comer, F. I.; Parker, G.
J.; Hart, G. W. J. Biol. Chem. e-publication 2001, Jan 8.
13. A photoaffinity-labeling agent 11 was prepared by react-
ing a 1-thio-GlcNAc (Horton, D.; Wolfrom, M. L. J. Org.
Chem. 1962, 27, 1794. Ohnishi, Y.; Ichikawa, M.; Ichikawa, Y.
Bioorg. Med. Chem. Lett. 2000, 10, 1289) and 4-(N-mal-
eimido)benzophenone (Sigma) followed by O-deacetylation.
The enzyme preparation was pretreated with UDP-Gal and
b1,4-galactosyltransferase, which was then photoaffinity-
labeled with the agent 11 by 365 nm (X), and treated with
UDP-[14C]Gal and b1,4-galactosyltransferase in order to
introduce radioactivity onto the photoaffinity-labeled protein
(Roquemore, E. P.; Chou, T.-Y.; Hart, G. W. Methods
Enzymol. 1994, 230, 433) (XI). The radioactive protein bands
were excised from the SDS-PAGE gel and microsequenced.
Acknowledgements
The authors thank Drs. Glendon Parker and Gerald
Hart at the Johns Hopkins University School of Medi-
cine for providing us with a partially purified enzyme.
Support from the National Institutes of Health (GM
52324) and a Johns Hopkins Institutional Research
Grant is greatly appreciated.

M. Ichikawa, Y. Ichikawa / Bioorg. Med. Chem. Lett. 11 (2001) 1769–1773
1773
14. Janda, K. D.; Lo, L.-C.; Lo, C.-H. L.; Sim, M.-M.;
Wang, R.; Wong, C.-H.; Lerner, R. A. Science 1997, 275,
945.
d, J=9.00 Hz), 7.47 (1H, br s), 7.71 (1H, d, J=9.60 Hz), and
8.05 (1H, br s).
Agent A: H NMR (500 MHz, DMSO-d6) d 1.14–1.64 (m),
1
15. Silverman, R. Mechanism-Based Enzyme Inactivation:
Chemistry and Enzymology; CRC: Boca Raton, FL, 1988.
16. Halazy, S.; Berges, V.; Ehrhard, A.; Danzin, C. Bioorg.
Chem. 1990, 18, 330.
17. (a) Horton, D. Org. Synth. 1966, 46, 1. (b) Horton, D.
Meth. Carbohydr. Chem. 1972, 6, 282.
1.81 (3H, s, NHAc), 3.71–3.76 (2H, m), 4.10–4.15 (1H, m),
4.28–4.33 (1H, m), 4.80 (1H, d, J=9.00 Hz), 6.37 (1H, br s),
6.43 (1H, br s), 6.92 (1H, t, J=55.5 Hz, –CF2H), 7.17 (1H, d,
J=9.00 Hz), 7.61–7.86 (4H, m), and 9.98 (0.8H, s). Agent B:
1H NMR (300 MHz, DMSO-d6) d 1.15–1.65 (15H, m), 1.81
(3H, s, NHAc), 3.69–3.80 (2H, m), 4.10–4.15 (1H, m), 4.27–
4.34 (1H, m), 4.63 (1H, br t, J=6.30 Hz), 4.81 (1H, d,
J=8.70 Hz, H-1), 5.11 (1H, d, J=5.40 Hz), 5.16 (1H, d,
J=5.10 Hz), 6.35 (1H, br s), 6.41 (1H, br s), 6.92 (1H, t,
J=55.5 Hz, –CF2H), 7.19 (1H, d, J=9.00 Hz), 7.63 (1H, br d,
J=9.60 Hz), 7.70–8.02 (5H, m), and 9.96 (0.6H, s). Agent D:
1H NMR (500 MHz, DMSO-d6) d 1.81 (3H, s, NHAc), 3.70–
3.77 (2H, m), 4.10–4.14 (1H, m), 4.29–4.32 (1H, m), 4.77 (1H,
d, J=8.00 Hz), 5.31 (2H, dd, J=1.50, 47.50 Hz, –CFH2), 6.37
(1H, s), 7.10 (1H, d, J=9.00 Hz), 7.49–7.82 (4H, m), and 9.86
(0.8H, s).
18. Kleine, H. P.; Weinberg, D. V.; Kaufman, R. J.; Sidhu,
R. S. Carbohydr. Res. 1985, 142, 333.
19. Compound 3: 1H NMR (300 MHz, DMSO-d6) d 1.77,
1.97, 2.008, 2.010 (3H each, s, 3ꢃOAc, NHAc), 4.14–4.35 (4H,
m), 4.98 (1H, t, J=9.60 Hz, H-4), 5.27 (1H, J=10.2 Hz, H-3),
5.58 (1H, d, J=8.40 Hz, H-1), 7.60 (1H, d, J=9.30 Hz, aro-
matic), 8.10 (1H, d, J=9.00 Hz, NH), 8.39 (1H, d, J=2.70 Hz,
aromatic), 8.52 (1H, dd, J=2.70, 9.30 Hz, aromatic), and
10.17 (1H, s, aldehydro).
1
Compound 4: H NMR (300 MHz, DMSO-d6) d 1.94, 2.06,
2.08, 2.10 (3H each, s, 3ꢃOAc, NHAc), 3.98 (1H, ddd,
J=2.70, 5.40, 9.90 Hz), 4.18–4.32 (3H, m), 5.16 (1H, t,
J=9.30 Hz), 5.34–5.43 (2H, m), 5.91 (1H, d, J=8.70 Hz, NH),
6.88 (1H, t, J=54.6 Hz, CF2H), 7.22 (1H, br s), 8.29 (1H, dd,
J=1.80, 9.30 Hz, aromatic), and 8.45 (1H, dd, J=1.80,
2.10 Hz, aromatic).
20. These biotin conjugates are: biotin LCLC NHS, (6-
biotinamidocaproylamido)caproic acid N-hydroxysuccinimide
ester from Sigma B3295; biotin disulfide NHS, biotin disulfide
N-hydroxysuccinimide ester from Sigma B4531; biotin-LC
NHS, biotinamidocaproate N-hydroxysuccinimide ester from
Sigma B 2643.
21. Bogardus, J. B.; Higuchi, T. J. Pharm. Sci. 1982, 71, 729.
22. A disulfide-exchange reaction has been utilized to discover
a specific ligand with a tether containing a disulfide group for
a receptor with an engineered cysteine; see: Erlanson, D. A.;
Braisted, A. C.; Raphael, D. R.; Stroud, R. M.; Gordon,
E. M.; Wells, J. A. Proc. Natl. Acad. Soc. U.S.A. 2000, 97, 9367.
23. Redeuilh, G.; Secco, C.; Baulieu, E.-E. J. Biol. Chem.
1985, 260, 3996.
1
Compound 5: H NMR (300 MHz, DMSO-d6) d 1.25–1.49
(4H, m), 1.43 (9H, s, NHtBoc), 1.94 (3H, s), 2.05 (6H, br s),
2.08 (3H, s), 2.31–2.37 (2H, m), 3.05–3.12 (2H, –CH2–
NHBoc), 3.83 (1H, ddd, J=2.10, 5.10, 9.90 Hz, H-5), 4.16
(1H, dd, J=2.40, 12.3 Hz, H-6a), 4.24–4.34 (2H, m), 4.66 (1H,
br s), 5.03 (1H, d, J=8.40 Hz, H-1), 5.12 (1H, J=9.30 Hz, H-
4), 5.32 (1H, dd, J=9.30, 10.50 Hz, H-3), 6.20 (1H, d,
J=9.00 Hz, NH), 6.79 (1H, t, J=54.9 Hz, –CF2H), 7.02 (1H,