A strategy for functional proteomic analysis of glycosidase activity from cell lysates

Article abstract of DOI:10.1002/anie.200454235

A multipurpose flag: The inactivator 1 covalently labels the catalytic nucleophiles of retaining β-glycosidases to form species 2. The small azide group allows labeling of enzymes with sterically congested active sites. Staudinger ligation of 2 with phosp

Full text of DOI:10.1002/anie.200454235

Communications
Enzyme Labels
A Strategy for Functional Proteomic Analysis of
Glycosidase Activity from Cell Lysates**
David J. Vocadlo* and Carolyn R. Bertozzi*
The glycosidases are an extremely large class of hydrolases
that are found throughout nature in organisms ranging from
bacteria to humans and have been classified into a number of
families on the basis of structural similarity.[1] These enzymes
have central roles in both prokaryotic and eukaryotic cellular
metabolism and their dysfunction can have deleterious
effects. In light of their biological significance, the rapid
detection of proteins with glycosidase activity in organisms is
of considerable interest. Activity-dependent labeling of
enzymes in complex mixtures is a central challenge of the
field of functional proteomics.[2] Anumber of approaches
have been developed for labeling and identifying proteins on
the basis of their activity through conjugation of active-site-
directed inactivators to biochemical probes.[3–5] These meth-
ods have been used to probe protease and other enzyme
activities both in cells and in vitro.[6,7] For many glycosidases,
however, such a strategy is significantly complicated by their
active site architecture. Unlike proteases, which commonly
have a cleft-like active site, exo-glycosidases typically have a
pocket-shaped active site[8] in which extensive interactions
between substrate and enzyme leave little space for append-
ing biochemical probes. Attempts to develop activity-based
covalent labeling reagents for glycosidases that can label
these enzymes when in complex mixtures have not yet proven
[*] Prof. D. J. Vocadlo,[+] Prof. C. R. Bertozzi
Center for New Direction in Organic Synthesis
Howard Hughes Medical Institute
Departments of Chemistry and Molecular and Cell Biology
University of California
Room 419, Latimer Hall, Berkeley, CA 94720-1460 (USA)
Fax: (+1)510-643-2628
E-mail: dvocadlo@sfu.ca
bertozzi@cchem.berkeley.edu
[+] Current Address:
Department of Chemistry
Simon Fraser University
8888 University Boulevard, Burnaby
BritishColumbia, V5A 1A6 (Canada)
Fax: (+1)604-291-3765
[**] The authors thank S. G. Withers for samples of recombinant
Agrobacterium sp. b-glucosidase and Xanthomonas manihotis b-
galactosidase, and H. C. Hang for useful discussions. D.J.V. thanks
the Canadian Institutes of Health Research for a fellowship. The
authors thank A. Debowski for technical assistance. This research
was supported by a grant to C.R.B. from the National Institutes of
Health(Grant no. GM066047), a President’s ResearchGrant from
Simon Fraser University, and a grant to D.J.V from the Natural
Sciences and Engineering ResearchCouncil of Canada. The center
for New Directions in Organic Synthesis is funded by Bristol–Myers
Squibb as a supporting member and Novartis as a sponsoring
member.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
5338
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200454235
Angew. Chem. Int. Ed. 2004, 43, 5338 –5342

Angewandte
Chemie
successful.[9,10] We felt that attempts to develop
an activity-based probe for targeting glycosi-
dases, particularly exo-glycosidases, would ben-
efit from consideration of the catalytic mech-
anism of these enzymes.
The great majority of retaining b-glycosi-
dases use a catalytic mechanism involving the
formation and subsequent breakdown of a
covalent glycosyl–enzyme intermediate.[11]
These steps involve two carefully positioned
carboxylate residues, one of which functions as
a general acid/base catalyst, whilst the other
acts as a catalytic nucleophile. Akey point is
that both steps involve electrophilic migration
of the anomeric center and oxocarbenium ion-
like transition states.[12,13] Fluorosugars have
been designed that act as mechanism-based
inactivators of retaining b-glycosidases and
function by trapping the covalent glycosyl–
enzyme intermediate by destabilizing these
transition states.[11] These inactivators have
facilitated identification of the enzymic nucle-
ophiles of a number of retaining b-glycosi-
dases.[11,12,14,15] We envisioned that these com-
pounds could be modified for use in functional
Figure 1. The activity-based labeling approach and Staudinger ligation. Glycosidase labeling was
performed bothwithpurified Escherichia coli LacZ and withcell lysates from E. coli. Samples were
incubated with6-azido-2,6-dideoxy-2-fluoro- b-d-galactosyl fluoride (6Az2FGalF), which resulted in
formation of a covalent glycosyl–enzyme adduct on the active site nucleophile. Staudinger ligation
of the azido group with a phosphine–FLAG (DYKDDDDK) probe resulted in specific labeling.
proteomics analysis of retaining b-glycosidases. This applica-
tion requires adornment of the fluorosugars with a detectable
tag.
To evaluate the consequences of substituting the 6-
hydroxy group of a b-galactoside with an azide moiety, we
synthesized the chromogenic substrate para-nitrophenyl 6-
azido-6-deoxy-b-d-galactoside (pNP6AzGal, 3, Scheme 1).
Although exo-glycosidases often exhibit some promiscu-
ity in their substrate specificity, their pocket-shaped active
sites preclude elaboration of the glycone moiety with large
affinity tags such as biotin. We felt that a fine balance could be
struck between the tolerance of these enzymes for small
modifications and the requirement for a detectable probe by
using the small, uniquely reactive azide moiety. This bioor-
thogonal functional group can be elaborated by Staudinger
ligation[16] or by the CuI-catalyzed [3+2] Huisgen cycloaddi-
tion reported by Rostovtsev et al.[17] Herein, we present a
strategy for the detection of retaining glycosidases in complex
mixtures by using azide-functionalized 2-fluorosugars in
conjunction with Staudinger ligation. An outline of the
strategy is shown in Figure 1, with Escherichia coli b-
galactosidase (LacZ) as a target protein. We chose LacZ to
demonstrate the strategy because this enzyme has previously
been inactivated with 2-fluorogalactosides.[18] Ahigh-resolu-
tion crystal structure of the resulting covalent complex has
been reported. The structure reveals a demanding steric
environment surrounding the 6-hydroxy group of the gal-
actose moiety, which should provide a stringent test of our
strategy.[19] We chose the 6-position as the azide substitution
site for synthetic ease and because the possibility of free
rotation around the C-6 methylene group may permit the
azide moiety to adopt a favorable orientation within the
enzyme active site. Furthermore, kinetic studies of LacZ with
modified substrates have shown that the 6-hydroxy group is
not essential for catalysis.[20] There is precedent for galacto-
sides with substitutions at the 6-position serving as substrates
for several different galactosidases,[21] and other glycosidases
also tolerate substitutions at this position.[22]
Scheme 1. a) TsCl, py, 59%; b) NaN3, DMF, 808C, 87%. py, pyridine;
Ts, tosyl; DMF, dimethylformamide (All experimental procedures are
described in the Supporting Information).
The ratio of the second-order rate constants [(Vmax/KM)OH
/
(Vmax/KM)N ] of the LacZ-catalyzed hydrolysis of pNP6AzGal
3
and that of the 6-hydroxy compound para-nitrophenyl b-d-
galactoside (pNPGal, 1, Scheme 1) was determined to be
12500. Substitution of the 6-hydroxy group with an azide
moiety therefore results in an energetic penalty of 5.6 kcal-
molÀ1 for this catalytic process. McCarter and Withers have
shown that substitution of the 6-hydroxy group with a fluorine
or hydrogen atom results in an energetic penalty of approx-
imately 4.0 kcalmolÀ1 for LacZ-catalyzed hydrolysis of aryl
galactosidases, presumably as a result of a loss of favorable
interactions within the enzyme active site.[20] The residual 1.6
kcalmolÀ1 difference between these energy penalties prob-
ably stems from a small steric penalty associated with the
greater bulk of the azido group compared to the hydroxy
group. The sterically confined active site of LacZ does not
prevent turnover of compounds bearing a 6-azido group and
Angew. Chem. Int. Ed. 2004, 43, 5338 –5342
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5339

Communications
we were optimistic that inactivators bearing this modification
would label the enzyme.
We prepared 6-azido-2,6-dideoxy-2-fluoro-b-d-galactosyl
fluoride (6Az2FgalF, 10, Scheme 2) as an activity-based probe
for b-galactosides. Incubation of purified LacZ with various
concentrations of 6Az2FGalF revealed a concentration and
time-dependent inactivation of the enzyme that could be fit to
a single exponential decay function (data not shown). From
the data shown in Figure 2, we determined the apparent
second-order rate constant of the inactivation process to be
0.2mÀ1 minÀ1. This result shows that 6Az2FGalF inactivates
LacZ on a timescale appropriate for practical applications
and suggests that this compound could be used to detect b-
galactosidase activity. To explore this possibility we incubated
purified LacZ with 6Az2FGalF overnight such that less than
10% residual activity remained after incubation. The sample
was dialyzed to remove excess inactivator and the protein was
denatured under mild reducing conditions at pH 3.5 so that
both the acylal ester linkage and the azide moiety remained
stable. Treatment of the sample with phosphine–FLAG[16]
followed by Western blot analysis revealed that the lower
limit for the detection of LacZ after all sample handling losses
was approximately 50 ng (Figure 3a). Treatment of a sample
of inactivated, phosphine–FLAG-labeled enzyme with 5%
ammonium hydroxide resulted in complete abrogation of the
FLAG-associated signal (data not shown), which is consistent
with aminolysis of the covalent acylal 6Az2FGal–enzyme
intermediate.
To extend this strategy to the analysis of glycosidase
activity from cell lysates, we examined the inactivation of
LacZ in cultures of E. coli K-12, either induced with IPTG or
not induced. Analysis of the crude cell lysates by SDS-PAGE
followed by Coomassie staining (Figure 3d) or by Western
blotting (Figure 3b), and analysis of galactosidase activity
with a standard colorimetric assay revealed the expected
marked difference in production of LacZ between cultures
that were induced and those that were not (0.6% LacZ
expression in uninduced cells as compared to that in induced
cells, taken as 100%). The cell lysates were then incubated
with the activity-based probe 6Az2FGalF and labeled with
phosphine–FLAG as described above for purified LacZ. As
shown in Figure 3c, the FLAG epitope was only observed for
samples that had been induced to express LacZ and treated
with 6Az2FGalF. These results indicate that both the
inactivation of LacZ using 6Az2FGalF and the subsequent
Figure 2. Activity of E. coli LacZ with(a) pNPGal ( 1) and (b) pNP6Az-
Gal (3). Enzyme assays were carried out at 378C in phosphate-buffered
saline at a range of substrate concentrations and in the presence of
15% DMF. c) Plot of the rate constants for the inactivation of purified
E. coli LacZ versus the concentration of the inactivator (6Az2FGalF).
Residual enzyme activity was monitored as a function of time by peri-
odically injecting an aliquot of the inactivation mixture into a vessel
containing 6.5 mm pNPGal. The inactivation process followed pseudo-
first-order kinetics. The progress of the enzymatic reaction was fol-
lowed continuously by monitoring the release of para-nitrophenolate
anions (absorbance at 400 nm).
Staudinger ligation are highly specific, and neither
treatment independently results in the labeling of
proteins from cell lysates.
To demonstrate the versatility of these reagents
we applied our labeling strategy to five other
retaining b-glycosidases reported to have b-galacto-
sidase activity. Two of these were recombinant
bacterial enzymes, Agrobacterium sp. b-glucosidase
(Abg)[23] and Xanthomonas manihotis b-galactosi-
dase (Xbg).[24] The other three enzymes (a plant b-
glucosidase from sweet almond (Sabg),[25] a fungal
Scheme 2. a) i) Tetrabutylammonium hydrogen sulfate, CH2Cl2, p-ClSPh, 1m NaOH,
ii) NaOMe, MeOH, iii) Amberlyst IR-20 H+, 67%; b) TsCl, py; c) NaN3, DMF, 808C, 56%;
d) Ac2O, py, 82%; e) i) N-bromosuccinimide, acetone, H2O, ii) diethylaminosulfur trifluoride,
tetrahydrofuran, CH2Cl2, À408C to RT; f) i) NaOMe, MeOH, ii) Amberlyst IR-20 H+, 58%
over two steps.
b-galactosidase from Aspergillus oryzae (Aobg),[21]
and the yeast b-galactosidase preparation from
Kluveromyces lactis (Kbg) known as Lactozym[26]
)
5340
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 5338 –5342

Angewandte
Chemie
Figure 4. Labeling of five b-retaining glycosidases by using 6Az2FGalF
in conjunction withStaudinger ligation. a) Western blot of samples of
Agrobacterium sp. b-glucosidase (Lane 1), Xanthomonas manihotis b-
galactosidase (Lane 2), Aspergillus oryzae b-galactosidase (Lane 3), Klu-
veromyces lactis b-galactosidase (Lane 4), and sweet almond b-glucosi-
dase (Lane 5). After inactivation, the samples were labeled with phos-
phine–FLAG and then analyzed by Western blotting using anti-FLAG–
HRP. b) SDS-PAGE analysis of the samples shown in (a). Molecular
weight standards are shown on the right.
Figure 3. a) Western blot showing the detection limit of 6Az2FGalF,
used in conjunction withStaudinger ligation. NL, no 6Az2FGalF; HI,
heat-inactivated LacZ (100 ng); 0, 0 ng LacZ; 100, 100 ng LacZ; 75,
75 ng LacZ; 50, 50 ng LacZ; 25, 25 ng LacZ. After inactivation, the
samples were labeled with phosphine–FLAG and analyzed by the West-
ern blot technique using anti-FLAG–HRP (HRP, horseradish perox-
idase). b) SDS-PAGE analysis of cell lysates from cultures of E. coli K-
12, either induced with isopropyl-b-d-thiogalactopyranoside (IPTG;
0.1 mm) or not induced. c) Western blot treated withanti-FLAG–HRP
to detect LacZ in the lysates of cells grown in the presence or absence
of IPTG. Samples were treated or untreated with6Az2FGalF and
reacted with phosphine–FLAG. d) Western blot analysis of LacZ levels
in the lysates of cells grown in the presence or absence of IPTG. The
blot was probed by treatment witha mouse anti-LacZ monoclonal anti-
body followed by an anti-mouse IgG–HRP conjugate. e) The Western
blot shown in (d) stripped and probed with anti-FLAG–HRP.
anticipate that the strategy will find broad utility in proteomic
analysis of these enzymes in prokaryotic and eukaryotic
proteomes. The azide group might also be useful as a chemical
tag within inactivators of other enzymes from entirely differ-
ent families with sterically confining active sites.
Received: March 11, 2004
Revised: June 28, 2004 [Z54235]
Keywords: carbohydrates · glycosidases · mechanism-based
.
inactivators · proteomics · Staudinger ligation
are commercially available. In each case, we observed specific
labeling of the enzymes (Figure 4a) in a manner dependent
on inactivation with 6Az2FGalF, just as observed for LacZ
(Figure 3). Of the six enzymes studied, two are from Family 1
(Abg and Sabg), two from Family 2 (LacZ and Kbg), and two
from Family 35 (Xbg and Aobg) of the glycoside hydrolases.[1]
Members of the same glycoside hydrolase family have been
shown to have similar protein folds and active site architec-
tures, and to effect catalysis through similar transition
states.[27–29] The ability of our reagents to label enzymes
from different glycoside hydrolase families suggests that this
strategy will be applicable to many families of retaining
glycosidases.
[1] B. Henrissat, G. Davies, Curr. Opin. Struct. Biol. 1997, 7, 637 –
644.
[2] G. C. Adam, E. J. Sorensen, B. F. Cravatt, Mol. Cell. Proteomics
2002, 1, 781 – 790.
[3] T. Nazif, M. Bogyo, Proc. Natl. Acad. Sci. USA 2001, 98, 2967 –
2972.
[4] D. Leung, C. Hardouin, D. L. Boger, B. F. Cravatt, Nat.
Biotechnol. 2003, 21, 687 – 691.
[5] D. Greenbaum, K. F. Medzihradszky, A. Burlingame, M. Bogyo,
Chem. Biol. 2000, 7, 569 – 581.
[6] A. E. Speers, G. C. Adam, B. F. Cravatt, J. Am. Chem. Soc. 2003,
125, 4686 – 4687.
In summary, we have developed a strategy for activity-
based labeling of retaining glycosidases by using the azide
group as a sterically unobtrusive chemical tag. The high
selectivity of both the inactivation with fluorosugars and the
Staudinger ligation with phosphine probes allows detection of
glycosidases in complex mixtures and the strategy can be used
for profiling these enzyme activities in cell lysates. We have
demonstrated that the approach can be used to tag several
glycosidases from different glycoside hydrolase families. We
[7] H. Ovaa, P. F. van Swieten, B. M. Kessler, M. A. Leeuwenburgh,
E. Fiebiger, A. M. C. H. van den Nieuwendijk, P. J. Galardy,
G. A. van der Marel, H. L. Ploegh, H. S. Overkleeft, Angew.
Chem. 2003, 115, 3754 – 3757; Angew. Chem. Int. Ed. 2003, 42,
3626 – 3629.
[8] G. J. Davies, B. Henrissat, Biochem. Soc. Trans. 2002, 30, 291 –
297.
[9] C.-S. Tsai, Y.-K. Li, L.-C. Lo, Org. Lett. 2002, 4, 3607 – 3610.
[10] M. Ichikawa, Y. Ichikawa, Bioorg. Med. Chem. Lett. 2001, 11,
1769 – 1773.
Angew. Chem. Int. Ed. 2004, 43, 5338 –5342
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5341

Communications
[11] J. Wicki, D. R. Rose, S. G. Withers, Methods Enzymol. 2002, 354,
84 – 105.
[12] D. J. Vocadlo, G. J. Davies, R. Laine, S. G. Withers, Nature 2001,
412, 835 – 838.
[13] G. J. Davies, L. Mackenzie, A. Varrot, M. Dauter, A. M.
Brzozowski, M. Schulein, S. G. Withers, Biochemistry 1998, 37,
11707 – 11713.
[14] D. J. Vocadlo, S. G. Withers, Methods Mol. Biol. 2000, 146, 203 –
222.
[15] S. G. Withers, K. Rupitz, I. P. Street, J. Biol. Chem. 1988, 263,
7929 – 7932.
[16] E. Saxon, C. R. Bertozzi, Science 2000, 287, 2007 – 2010.
[17] V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless,
Angew. Chem. 2002, 114, 2708 – 2711; Angew. Chem. Int. Ed.
2002, 41, 2596 – 2599.
[18] J. C. Gebler, R. Aebersold, S. G. Withers, J. Biol. Chem. 1992,
267, 11126 – 11130.
[19] D. H. Juers, T. D. Heightman, A. Vasella, J. D. McCarter, L.
Mackenzie, S. G. Withers, B. W. Matthews, Biochemistry 2001,
40, 14781 – 14794.
[20] J. D. McCarter, M. J. Adam, S. G. Withers, Biochem. J. 1992, 286,
721 – 727.
[21] U. Grabowska, D. A. MacManus, K. Biggadike, M. I. Bird, S.
Davies, T. Gallagher, L. D. Hall, E. N. Vulfson, Carbohydr. Res.
1997, 305, 351 – 361.
[22] D. A. MacManus, U. Grabowska, K. Biggadike, M. I. Bird, S.
Davies, E. N. Vulfson, T. Gallagher, J. Chem. Soc. Perkin Trans. 1
1999, 295 – 305.
[23] J. B. Kempton, S. G. Withers, Biochemistry 1992, 31, 9961 – 9969.
[24] J. E. Blanchard, L. Gal, S. He, J. Foisy, R. A. Warren, S. G.
Withers, Carbohydr. Res. 2001, 333, 7 – 17.
[25] S. He, S. G. Withers, J. Biol. Chem. 1997, 272, 24864 – 24867.
[26] T. Maugard, D. Gaunt, M. D. Legoy, T. Besson, Biotechnol. Lett.
2003, 25, 623 – 629.
[27] G. J. Davies, M. L. Sinnott, S. G. Withers, in Comprehensive
Biological Catalysis, Vol. 1 (Ed.: M. L. Sinnott), Academic Press,
London, 1997, pp. 119 – 208.
[28] J. Gebler, N. R. Gilkes, M. Claeyssens, D. B. Wilson, P. Beguin,
W. W. Wakarchuk, D. G. Kilburn, R. C. Miller, Jr., R. A.
Warren, S. G. Withers, J. Biol. Chem. 1992, 267, 12559 – 12561.
[29] M. W. Bauer, R. M. Kelly, Biochemistry 1998, 37, 17170 – 17178.
5342
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 5338 –5342