Phosphodiesters serve as potentially tunable aglycones for fluoro sugar inactivators of retaining β-glycosidases

Article abstract of DOI:10.1039/c4ob00235k

2-Deoxy-2-fluoroglycosides bearing dibenzyl phosphate and phosphonate aglycones were synthesised and tested as covalent inactivators of several retaining α- and β-glycosidases. β-d-Gluco-, -manno- and -galacto-configured benzyl-benzylphosphonate derivatives efficiently inactivated β-gluco-, β-manno- and β-galactosidases, while α-gluco- and α-manno-configured phosphate and phosphonate derivatives served instead as slow substrates. This journal is the Partner Organisations 2014.

Full text of DOI:10.1039/c4ob00235k

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Phosphodiesters serve as potentially tunable
aglycones for fluoro sugar inactivators of retaining
β-glycosidases†
Cite this: Org. Biomol. Chem., 2014,
12, 2592
B. P. Rempel and S. G. Withers*
2-Deoxy-2-fluoroglycosides bearing dibenzyl phosphate and phosphonate aglycones were synthesised
and tested as covalent inactivators of several retaining α- and β-glycosidases. β-D-Gluco-, -manno- and
-galacto-configured benzyl-benzylphosphonate derivatives efficiently inactivated β-gluco-, β-manno-
and β-galactosidases, while α-gluco- and α-manno-configured phosphate and phosphonate derivatives
served instead as slow substrates.
Received 30th January 2014,
Accepted 28th February 2014
DOI: 10.1039/c4ob00235k
www.rsc.org/obc
Glycosidases play key roles in a number of diseases in in Scheme 1 for a retaining β-glycosidase.6 Both steps proceed
humans, including metabolic disorders such as Gaucher’s via oxocarbenium ion-like transition states. 2-Deoxy-2-fluoro-
disease,1 cancer,2 HIV/AIDS,3 Parkinson’s disease,1 and influ- glycosides bearing good leaving groups have been shown to
enza.4,5 Specific inhibitors of glycosidases are therefore of function as highly efficient and selective mechanism-based
interest as potential therapeutics, as well as tools for structural inactivators for a variety of retaining β-glycosidases.7 The pres-
and mechanistic characterisation of the enzymes themselves. ence of the highly electronegative fluorine atom on C2 of the
Retaining glycosidases are those that hydrolyse the glycosidic pyranosyl ring slows both steps of the reaction by destabilizing
bond with net retention of anomeric stereochemistry. Most both glycosylation and deglycosylation transition states. This
retaining glycosidases characterized to date employ a double- occurs through a combination of inductive effects and disrup-
displacement mechanism. In the first step, nucleophilic attack tion of critical active site hydrogen-bonding interactions with
of an active site carboxylate (Glu or Asp) on the sugar anomeric the hydroxyl normally found on C2 in the natural substrates.
centre, along with activation of the leaving group through The activated leaving group at the anomeric centre selectively
general acid catalysis, generates a covalent glycosyl-enzyme accelerates the glycosylation step, leading to an overall
intermediate in what is termed the glycosylation step. During accumulation of a highly stabilised covalent glycosyl-enzyme
the subsequent deglycosylation step, base-promoted attack of intermediate.8,9 This strategy is less successful for retaining
water at the anomeric centre completes the catalytic cycle to α-glycosidases: instead the 2-fluorosugars tend to act as slow
regenerate the free enzyme and the sugar residue with the substrates with a rate-limiting glycosylation step, most likely
same anomeric configuration as the substrate, as represented due to differences in the extent of positive charge development
on the anomeric centre vs. the ring oxygen atom in each
step.5 Consistent with this, 5-fluoroglycosyl fluorides, wherein
the fluorine substituent is close to the endocyclic oxygen,
do indeed serve as covalent inactivators of retaining
α-glycosidases.9,10
Kinetic studies with a series of aryl 2-deoxy-2-fluoroglyco-
sides revealed that the phenol pKa should be 5 or lower for the
reagents to function as efficient covalent inactivators.11 Indeed
the majority of reported fluorosugar inactivators either use flu-
oride or 2,-4-dinitrophenol aglycones, although chloride and
2,4,6-trinitrophenol aglycones were used as leaving groups for
Scheme 1 Mechanism for a typical retaining β-glycosidase.
2-deoxy-2,2-difluoroglycoside inactivators of retaining α-glyco-
sidases.12 Further, Overkleeft and co-workers recently
Department of Chemistry, University of British Columbia, Vancouver, B.C. V6 T 1Z1,
Canada. E-mail: withers@chem.ubc.ca
†Electronic supplementary information (ESI) available: Synthetic methods and
employed an N-phenyl trifluoroacetimidate as leaving group
on a 2-fluorosugar inactivator of human β-glucocerebrosi-
characterization of new compounds, in vitro characterization of enzymatic inhi-
bition or inactivation experiments. See DOI: 10.1039/c4ob00235k
dase.13 Interestingly, the natural aglycones were also reported
2592 | Org. Biomol. Chem., 2014, 12, 2592–2595
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
to serve as sufficiently reactive leaving groups for 2-deoxy-2-
Probes 1–5 were synthesized from their precursor 2-deoxy-2-
fluoroglycosides when the rate-limiting step for that natural fluoroglycopyranosides, as outlined in the ESI (see Schemes
substrate/enzyme pair was indeed the deglycosylation step.14,15 S1–S4†). Note that compounds 1, 2 and 4 were prepared and
In those cases the enzyme recruits transition state-stabilising evaluated as diastereomeric mixtures at phosphorus. While
interactions between the natural leaving group and the agly- the yields were low for the production of the per-O-acetylated
cone sub-site to accelerate the glycosylation step.
precursors to 3–5, modest amounts could be purified after
With that idea in mind we recently reported the synthesis selective decomposition by exploiting the greater stability of
and testing of dialkyl phosphate and phosphonate-containing the α-anomer relative to the readily available β-anomer.
2-fluorosugar inactivators of human β-glucocerebrosidase that
Compounds 1 and 2 were tested as covalent inactivators of
exploit the preference of that enzyme for binding hydrophobic retaining β-glycosidases by measuring residual enzyme activity
aglycones. The best three inactivators, which inactivate GCase as a function of time upon incubation with a range of concen-
1000–4000 times faster than does 2-fluoroglucosyl fluoride, trations of 1 or 2. Data yielded pseudo-first-order inactivation
were those with di-O-octyl phosphate, di-O-benzyl phosphate, rate constants at each inhibitor concentration. Fitting of these
or benzyl benzylphosphonate leaving groups, with the latter values to a Michaelis–Menten equation yielded values of ki, Ki,
displaying the best balance between hydrolytic stability and and ki/Ki. Kinetic parameters for these compounds and the
enzymatic reactivity.16 Indeed the di-O-alkyl phosphate deriva- gluco version synthesised previously13 as inactivators of the
tives underwent spontaneous hydrolysis with a half life of relatively broad specificity retaining β-glucosidase from Agro-
around 15 minutes, while the dibenzylphosphonate derivative bacterium sp. (Abg)17 are shown in Table 1, along with pre-
was considerably more stable with a half-life around 10 hours.
viously determined values for other fluorosugars. First order
Since the alkylated phosphate leaving groups are highly fits are presented as Fig. S1–S2 in the ESI.† The time-depen-
activated we were interested in exploring how broadly this new dent loss of enzyme activity measured from treating Abg with
class of activated fluorosugar can be used as covalent inactiva- either 1 or 2 was too rapid to permit sampling at concen-
tors of retaining glycosidases, even in the absence of specific trations approaching saturation, so only ki/Ki values could be
aglycone site interactions. These compounds are of particular measured. Both 1 (ki/Ki = 23.9 min−1 mM−1) and 2 (ki/Ki =
interest since their physical, thus pharmacokinetic, properties 3.0 min−1 mM−1) functioned as rapid, time-dependent inacti-
can be modulated by modifying the alkyl substituent without vators of Abg, with values bracketing those for the gluco- and
substantial effect on inherent reactivity of the sugar moiety. To manno-analogues.11,18 This makes the manno-configured com-
this end, probes 1–5 (Scheme 2) were synthesized and tested pound 1 the second most efficient fluorosugar inactivator of Abg
as substrates/inactivators for a variety of retaining α- and β-gly- reported to date, with only 2,4-dinitrophenyl 2-deoxy-2-fluoro-
cosidases. We hypothesized that the α-configured probes
would function as slow substrates for the appropriate retaining
α-glycosidases without any accumulation of a covalent glycosyl-
Table 1 Kinetic parameters for selected fluorosugars as inactivators
enzyme intermediate, since it is known that their covalent
2-deoxy-2-fluoroglycosyl-enzyme intermediates are rapidly
turned over.9,10 For the β-configured probes, the dibenzyl
phosphonate aglycone was chosen as it had been shown
to be relatively stable towards spontaneous hydrolysis while
maintaining reactivity with the enzyme.16 Dibenzyl phosphate
and phosphonate derivatives were chosen for the α-configured
probes based on their relative synthetic accessibility.
of Abg
ki
Ki
(mM)
ki/Ki
Compound
(min−1
)
(min−1 mM−1
)
5.9
0.40
14.8
—
—
11
5.6
1.2
4.7
—
—
23.9
2.6
3.2
0.81
3.0
—
—
Scheme 2 Compounds tested as covalent inactivators of retaining
β-glycosidases (1 and 2) and as slow substrates for retaining α-glyco-
sidases (3–5).
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2592–2595 | 2593

View Article Online
Paper
β-D-glucopyranoside (ki/Ki = 500 min−1 mM−1 19
Organic & Biomolecular Chemistry
)
being a better were each shown to function as covalent inactivators for their
inactivator. The phosphonate derivative of the galacto-sugar, 2, cognate β-glucosidase, β-mannosidase, and β-galactosidase. In
was also a better inactivator than the corresponding fluoride.11,18 addition, all three compounds functioned as inactivators of a
Compound 1 was also tested as a covalent inactivator of the single broad-spectrum β-glycosidase (Abg). These results
retaining β-mannosidase from Cellulomonas fimi (Man2A)20 demonstrate the broad applicability of this new class of com-
yielding inactivation parameters of ki = 0.52 min−1, Ki
=
pounds as covalent inactivators for retaining β-glycosidases. In
3.7 mM, and ki/Ki = 0.14 min−1 mM−1. Interestingly this is general the inactivators were of comparable reactivity to analo-
almost a 10-fold worse inactivator of Man2A than is 2-deoxy-2- gous inactivators bearing fluoride or dinitrophenolate agly-
fluoro-β-D-mannopyranosyl fluoride (ki = 0.57 min−1, Ki
=
cones, consistent with their roughly comparable chemical
0.41 mM, and ki/Ki = 1.4 min−1 mM−1),20 due almost entirely reactivities. As anticipated, none of the alpha-linked versions,
to an increase in Ki. This suggests that the benzyl benzylphos- 2-deoxy-2-fluoro-α-D-glucopyranosyl dibenzyl phosphate 3,
phonate aglycone is less well accommodated in the active site 2-deoxy-2-fluoro-α-D-mannopyranosyl dibenzyl phosphate 4, or
than is the much smaller fluoride aglycone. However, there 2-deoxy-2-fluoro-α-D-mannopyranosyl benzyl-benzylphosphonate
does not appear to be a significant effect on the energies of 5 functioned as covalent inactivators of their cognate α-glyco-
the two transition states relative to their respective ground sidases. While both the dibenzyl phosphate derivatives were
states, as the ki values are extremely close in magnitude.
hydrolysed as slow substrates, 2-deoxy-2-fluoro-α-D-mannopyra-
Compound 2 proved to be a useful inactivator of the retain- nosyl benzyl-benzylphosphonate 5 did not bind to the jack
ing β-galactosidase from Escherichia coli (Lac-Z), with kinetic bean α-mannosidase at any concentration tested. In general,
parameters of ki = 0.14 min−1, Ki = 0.058 mM, and ki/Ki = the use of substituted phosphonate aglycones as leaving
2.5 min−1 mM−1. It is therefore a tighter binding inactivator groups in fluorosugar-based inactivators of β-glycosidases
than is 2-deoxy-2-fluoro-β-D-galactopyranosyl fluoride (ki
=
yields reagents of generally comparable reactivity to the com-
13.2 min−1, Ki = 1.3 mM, and ki/Ki = 10.2 min−1 mM−1),21 but monly employed fluoride aglycone. However the alkyl substitu-
undergoes reaction almost 10 fold slower. Fluoride thus ents employed on the phosphonate may be chosen to either
appears to be the aglycone that is best accommodated in the modulate pharmacokinetic behaviour of the inhibitors or to
glycosylation transition state, as can be seen from a compari- harness specific interactions with the enzyme’s aglycone
son of the relative ki values, while the benzyl benzylphospho- subsite to improve both efficiency and selectivity.
nate aglycone appears to be the one that binds tightest to the
enzyme in the ground state. This finding is consistent with the
fact that Lac-Z is known to have a relatively open active site
Notes and references
that can accommodate both of the natural substrates, allo-
lactose and lactose.22
1 V. Cullen, P. Sardi, J. Ng, Y. H. Xu, Y. Sun, J. J. Tomlinson,
P. Kolodziej, I. Kahn, P. Saftig, J. Woulfe, J. C. Rochet,
M. A. Glicksman, S. H. Cheng, G. A. Grabowski,
L. S. Shihabuddin and M. G. Schlossmacher, Ann. Neurol.,
2011, 69, 940–953.
2 S. Gerber-Lemaire and L. Juillerat-Jeanneret, Mini-Rev. Med.
Chem., 2006, 6, 1043–1052.
3 T. Kajimoto and M. Node, Curr. Top. Med. Chem., 2009, 9,
13–33.
4 M. Vonitzstein, W. Y. Wu, G. B. Kok, M. S. Pegg,
J. C. Dyason, B. Jin, T. V. Phan, M. L. Smythe, H. F. White,
S. W. Oliver, P. M. Colman, J. N. Varghese, D. M. Ryan,
J. M. Woods, R. C. Bethell, V. J. Hotham, J. M. Cameron
and C. R. Penn, Nature, 1993, 363, 418–423.
Probe 3 was incubated with yeast α-glucosidase (Yag) to test
whether it functioned as an inactivator, or as a slow substrate.
When residual enzyme activity was measured as a function of
time, no time-dependent decrease in activity was observed
either at 37 °C or 4 °C, demonstrating that no substantial
accumulation of the covalent intermediate was occurring. A
TLC analysis of the reaction products from incubation of Yag
and 3 showed the enzyme-dependent formation of 2-deoxy-2-
fluoro-D-glucose, confirming that 3 was indeed acting as a sub-
strate. By testing 3 and 4-nitrophenyl α-D-glucopyranoside23 as
competitive substrates for Yag under steady state conditions,
an apparent K′ value for 3 (corresponding to the Km value for 3
i
as a substrate) of 0.7 mM was obtained (see Fig. S6†).
Probes 4 and 5 were separately evaluated as substrates for
Jack bean α-mannosidase (JBAM). Neither behaved as a
covalent inactivator at either 37 °C or 4 °C when enzyme
activity was monitored as a function of incubation time.
Further, TLC analysis indicated that only 4 was hydrolysed by
the enzyme. Evaluation of 4 as a competitive substrate for the
known substrate 2,4-dinitrophenyl α-D-mannopyranoside24 for
5 J. H. Kim, R. Resende, T. Wennekes, H. M. Chen, N. Bance,
S. Buchini, A. G. Watts, P. Pilling, V. A. Streltsov, M. Petric,
R. Liggins, S. Barrett, J. L. McKimm-Breschkin, M. Niikura
and S. G. Withers, Science, 2013, 340, 71–75.
6 D. L. Zechel and S. G. Withers, Acc. Chem. Res., 2000, 33,
11–18.
7 B. P. Rempel and S. G. Withers, Glycobiology, 2008, 18,
570–586.
JBAM under steady state conditions gave an apparent K′ value
i
of ∼5 mM, indicating that 4 was a relatively poor substrate for
JBAM (see Fig. S8†).
In conclusion, 2-deoxy-2-fluoro-β-D-gluco-, -manno- and
-galactosides bearing benzyl-benzylphosphonate aglycones
8 J. Wicki, D. R. Rose and S. G. Withers, Methods in Enzy-
mology, Academic Press Inc., San Diego, 2002, pp. 84–105.
9 M. T. C. Walvoort, G. A. van der Marel, H. S. Overkleeft and
J. D. C. Codee, Chem. Sci., 2013, 4, 897–906.
2594 | Org. Biomol. Chem., 2014, 12, 2592–2595
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
10 R. M. Mosi and S. G. Withers, Methods in Enzymology, 17 J. B. Kempton and S. G. Withers, Biochemistry, 1992, 31,
Academic Press Inc., San Diego, 2002, pp. 64–84. 9961–9969.
11 I. P. Street, J. B. Kempton and S. G. Withers, Biochemistry, 18 S. G. Withers, K. Rupitz and I. P. Street, J. Biol. Chem.,
1992, 31, 9970–9978. 1988, 263, 7929–7932.
12 R. Zhang, J. D. McCarter, C. Braun, W. Yeung, G. D. Brayer 19 S. G. Withers, I. P. Street, P. Bird and D. H. Dolphin, J. Am.
and S. G. Withers, J. Org. Chem., 2008, 73, 3070–3077. Chem. Soc., 1987, 109, 7530–7531.
13 M. T. C. Walvoort, W. W. Kallemeijn, L. I. Willems, 20 D. Stoll, S. M. He, S. G. Withers and R. A. J. Warren,
M. D. Witte, J. Aerts, G. A. van der Marel, J. D. C. Codee Biochem. J., 2000, 351, 833–838.
and H. S. Overkleeft, Chem. Commun., 2012, 48, 10386– 21 J. C. Gebler, R. Aebersold and S. G. Withers, J. Biol. Chem.,
10388. 1992, 267, 11126–11130.
14 S. Cottaz, B. Henrissat and H. Driguez, Biochemistry, 1996, 22 D. H. Juers, T. D. Heightman, A. Vasella, J. D. McCarter,
35, 15256–15259.
L. Mackenzie, S. G. Withers and B. W. Matthews, Bio-
chemistry, 2001, 40, 14781–14794.
23 J. D. McCarter and S. G. Withers, J. Biol. Chem., 1996, 271,
6889–6894.
15 J. D. McCarter, W. Yeung, J. Chow, D. Dolphin and
S. G. Withers, J. Am. Chem. Soc., 1997, 119, 5792–5797.
16 B. P. Rempel, M. B. Tropak, D. J. Mahuran and
S. G. Withers, Angew. Chem., Int. Ed., 2011, 50, 10381– 24 S. Howard, S. M. He and S. G. Withers, J. Biol. Chem., 1998,
10383.
273, 2067–2072.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2592–2595 | 2595