Remarkable Reactivity Differences between Glucosides with Identical Leaving Groups

Article abstract of DOI:10.1021/jacs.7b09645

Two isomeric aryl 2-deoxy-2-fluoro-β-glucosides react with a β-glucosidase at rates differing by 10⁶-fold, despite the fact that they release the same aromatic aglycone. In contrast, the equivalent glucoside substrates react with essentially identical rate constants. Insight into the source of these surprising rate differences was obtained through a comprehensive study of the nonenzymatic (spontaneous) hydrolysis of these same substrates, wherein an approximate 10⁵-fold difference in rates was measured, clarifying that the differences were inherent rather than being due to specific interactions with the enzyme. The possibility that an alternate nucleophilic aryl substitution mechanism was responsible for the rapid reaction of the faster substrate was excluded through ¹⁸O-labeling studies. Further exploration of the origins of these rate differences involved analysis of X-ray crystal structures as well as quantum chemical calculations, which surprisingly revealed that ground state destabilization and transition state stabilizing effects contribute almost equally to the observed reactivity differences. These studies highlight the dangers of using simple reference equilibria such as pK_a values as measures of leaving group ability.

Full text of DOI:10.1021/jacs.7b09645

Article
Journal of the American
REMARKABLE
Chemical Society
is published by the
REACTIVITY
American Chemical
Society. 1155 Sixteenth
DIFFERENCES
Street N.W., Washington,
DC 20036
Subscriber access provided by La Trobe
Published by American
University Library
Chemical Society.
Copyright © American

BETWEEN
GLUCOSIDES
Journal of the American
WITH IDEis
N
pub
T
lis
I
he
ChemicalCSdobciety
A
y the
L
American Chemical
Society. 1155 Sixteenth
LEAVING GROUPS
Street N.W., Washington,
DC 20036
Subscriber access provided by La Trobe
Published by American
University Library
Chemical Society.
Copyright © American

Tianmeng Duo, Kyle Robinson,
Ian R. Greig, Hong-Ming
Journal of the American
Chen, Brian O Patrick,
Chemical Society
is published by the
and Stephen G. Withers
American Chemical
Society. 1155 Sixteenth
J. Am. Chem. Soc., Just Accepted
Street N.W., Washington,
Manuscript • DOI: 10.1021/
DC 20036
Subscriber access provided by La Trobe
Published by American
University Library
Chemical Society.
Copyright © American

jacs.7b09645 • Publication
Date (Web): 16 Oct 2017
Journal of the American
Downloaded from http://
Chemical Society
pubs.acs.org on October 16, 2017
is published by the
American Chemical
Society. 1155 Sixteenth
Street N.W., Washington,
DC 20036
Subscriber access provided by La Trobe
Published by American
University Library
Chemical Society.
Copyright © American

Just Accepted
Journal of the American
“Just Accepted” manuscripts have been pe
Chemical Society
online prior to technical editing, formatting f
is published by the
Society providesAm“Jeruicsant CAhecmciecapl ted” as a f
Society. 1155 Sixteenth
dissemination of scientific material as soon
Street N.W., Washington,
DC 20036
appear in full in PDF format accompanied by
Subscriber access provided by La Trobe
Published by American
University Library
Chemical Society.
Copyright © American

fully peer reviewed, but should not be consid
readers and citable by the Digital Object Ide
Journal of the American
to authors. Therefore, the “Just Accepted”
in the journal. AftCehremaicmal Saoncuiestycript is technic
is published by the
American Chemical
Accepted” Web site and published as an AS
Society. 1155 Sixteenth
changes to the manuscript text and/or gra
Street N.W., Washington,
DC 20036
Subscriber access provided by La Trobe
Published by American
University Library
Chemical Society.
Copyright © American

and ethical guidelines that apply to the jo
or consequences arising from the use of in
Journal of the American
Chemical Society
is published by the
American Chemical
Society. 1155 Sixteenth
Street N.W., Washington,
DC 20036
Subscriber access provided by La Trobe
Published by American
University Library
Chemical Society.
Copyright © American

PaJgoeu1rnoafl 2o3f the American Chemical Society
khydr (prox:dist)
=
DDAOprox-Glc
DDAOdist-Glc
~56,000 : 1
1
2OH
OH
Cl
Cl
O
O
O
HO
HO
Time
HO
HO
O
O
O
OH
3
4
OH
N
Cl
DDAO
N
Cl
Cl
ACS Paragon Plus Environment
O
O
5
6
7
khydr
=
khydr
=
0.46 h-1
8.2 x 10-6 h-1
Cl
N
+ Glucose

Journal of the American Chemical Society
Page 2 of 23
1
1
2
3
4
REMARKABLE REACTIVITY DIFFERENCES BETWEEN GLUCOSIDES WITH IDENTICAL LEAVING GROUPS
5
6
7
Tianmeng Duo, Kyle Robinson, Ian R. Greig, Hong-Ming Chen, Brian O. Patrick and Stephen G. Withers*
8
9
Department of Chemistry
University of British Columbia
2036 Main Mall
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Vancouver, B.C.
Canada V6T 1Z1
*E-mail address: withers@chem.ubc.ca
ABSTRACT
Two isomeric aryl 2-deoxy-2-fluoro--glucosides react with a -glucosidase at rates differing by 106 fold,
despite the fact that they release the same aromatic aglycone. In contrast the equivalent glucoside
substrates react with essentially identical rate constants. Insight into the source of these surprising rate
differences was obtained through a comprehensive study of the non-enzymatic (spontaneous)
hydrolysis of these same substrates, wherein an approximate 105 fold difference in rates was measured,
clarifying that the differences were inherent, rather than being due to specific interactions with the
enzyme. The possibility that an alternate nucleophilic aryl substitution mechanism was responsible for
the rapid reaction of the faster substrate was excluded through 18O-labelling studies. Further exploration
of the origins of these rate differences involved analysis of X-ray crystal structures as well as quantum
chemical calculations, which surprisingly revealed that ground state destabilization and transition state
stabilizing effects contribute almost equally to the observed reactivity differences. These studies
ACS Paragon Plus Environment

Page 3 of 23
Journal of the American Chemical Society
2
1
2
3
4
5
6
highlight the dangers of using simple reference equilibria such as pKa values, as measures of leaving
group ability.
7
8
9
INTRODUCTION
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Much of our understanding of the mechanisms of enzymatic and non-enzymatic glycoside hydrolysis has
come from kinetic studies of aryl glycosides.1 The ability to manipulate reactivity through the
incorporation of different substituents into the phenolic moiety allows the construction of Hammett and
Brønsted relationships, which in turn inform on rate-limiting steps and transition state structures.2-4 In
such studies leaving group ability is generally quantified through Hammett σ values, or more intuitively
through the pKa values of the corresponding phenols. In this way, spontaneous hydrolysis of glycosides
under neutral to acidic conditions has been shown to proceed via a dissociative mechanism involving a
late, oxocarbenium ion-like transition state.5-10 Likewise, mechanisms of the vast majority of
glycosidases feature highly developed oxocarbenium ion-like transition states, with departure of the
less-acidic aglycones assisted by acid catalysis from suitably positioned carboxylic acid residues in the
active site.11-13 This is true both for inverting glycosidases, which effect catalysis through a single
displacement mechanism, and for retaining glycosidases, which employ a double-displacement
mechanism via two oxocarbenium ion-like transition states.11 Accordingly, Brønsted plots of log kcat or
log kcat/Km versus aglycone pKa typically feature negative slopes with βlg values often approaching -1,
indicating substantial negative charge development on the glycosidic oxygen at the reaction transition
state.2, 14 This is also true for spontaneous (non-enzymatic) aryl glycoside hydrolysis since lg values
determined in four independent, spontaneous hexopyranoside hydrolysis studies15 fall in a tightly
clustered range with an average lg = 1.25 ± 0.02. The increase in rate constant with decreasing phenol
pKa is typically ascribed to increased transition state stabilisation.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 23
3
1
2
3
4
5
6
7
8
9
As part of a recent study aimed at generating active site titrating agents for retaining glycosidases we
synthesised 2-deoxy-2-fluoro-glycosides bearing fluorogenic leaving groups of low pKa.16-17 The fluorine
at C-2 inductively destabilises both transition states, slowing both the formation and the hydrolysis of
the glycosyl enzyme intermediate. Incorporation of a good leaving group of low pKa ensures that
intermediate formation is fast and thus that the glycosyl enzyme accumulates, with stoichiometric
release of the fluorophore. One of the most sensitive such reagents we synthesized was the DDAO-
glucoside 1 (Figure 1), which includes a dichloro-dimethylacridinone (DDAO) leaving group having a pKa
value of 5.3.16 This fluorogenic glycoside proved very useful for selective quantitation of retaining -
glucosidases down to sub-nanomolar levels according to the process shown in Figure 2.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Reaction of 1 with the Agrobacterium sp. -glucosidase (Abg) was very fast with a second order rate
constant of ki/Ki = 1.38 × 106 mM-1 min-1, which is equivalent to 2.2 × 107 M-1 s-1 thus approaching
diffusion control.16 Indeed, stopped-flow techniques were needed to measure this value. Synthesis of 1
had been achieved by coupling DDAO with the protected 2-fluoroglycosyl bromide, yielding both 1 and
its regioisomer 2, in which the aglycone is attached by the other oxygen of the aryl moiety in its
alternative tautomeric form, thus will release the same phenolic leaving group. Henceforth this will be
referred to as the DDAOdist glycoside wherein the two chlorines are distal to the sugar. Likewise, 1 will
be referred to as the DDAOprox glycoside where the two chlorines are proximal to the sugar. These two
isomers have distinctly different absorption properties, where 1 appears yellow in colour
(max = 403 nm) and 2 appears red (max = 460 nm). Under the synthetic conditions employed they were
formed in a ratio of 85:15 (1:2). It was therefore of interest to explore the utility of the distal isomer as a
potential mechanism-based inhibitor, but surprisingly, when tested as a titrating agent for Abg, the
distal isomer 2 proved to react at a rate over 106 times lower than does 1, despite the fact that both
ACS Paragon Plus Environment

Page 5 of 23
Journal of the American Chemical Society
4
1
2
3
4
5
6
7
8
9
release the same leaving group. This report describes a detailed experimental and computational
analysis of the underlying basis for this reactivity difference, through an experimental and
computational exploration of the reactivity of the corresponding (non-fluorinated) glucosides towards
both enzymatic and non-enzymatic hydrolysis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
The two isomeric acridinoyl 2-deoxy-2-fluoro glycosides 1 and 2 were synthesised as described
previously16 and characterised as described in Supplementary Information. Incubation of the
Agrobacterium sp -glucosidase (Abg) with the distal isomer (2) of the acridinoyl -2-deoxy-2-fluoro-
glucoside resulted in only extremely slow inactivation of the enzyme activity compared to the rapid
inactivation afforded by 1. Much higher concentrations (60-120 µM vs. 0.1 M) as well as longer
reaction times, of up to 60 minutes (vs. 10 s), to obtain complete inactivation of Abg. Indeed, since 1 and
2 were products of the same synthetic reaction there was a concern that even this slow observed
inactivation might arise largely – or perhaps only – from small amounts of contaminating 1 at levels too
low to be detected by NMR but, given the reactivity difference, still sufficient to effect inactivation. To
address this concern a 100 µM solution of 2 was treated with a 300 nM solution of Abg as a “reagent” to
purify 2 by destruction of potentially contaminating 1, and the remaining enzymatic activity was
monitored for 60 minutes until a low, constant rate was observed. The sample was then loaded onto a
C18 SepPak reverse phase column to remove the Abg and any released DDAO. Purified 2 still inactivated
Abg, but with an apparent rate constant that was 2-fold lower than that seen prior to purification,
confirming the initial presence and subsequent removal of contaminating 1. Kinetic parameters for this
purified sample of 2 were determined by conventional inactivation kinetic analysis and found to be
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 23
5
1
2
3
4
5
6
ki = 4.8 x 10-2 min-1, Ki = 100 µM and ki/Ki = 0.5 mM-1 min-1. Thus, remarkably, the second order rate
constants for inactivation by 1 and 2 differ by over six orders of magnitude!
7
8
9
Such a huge difference in inactivation behaviour was particularly surprising since the aglycones released
by 1 and 2 are identical, as is seen in Figure 2. Further, since glycosidase transition states are late, it is
expected that there should be almost full charge development on the aglycone, and electron
reorganisation therein via resonance should be complete. A first concern was that the assumed
structures for 1 and/or 2 were incorrect, despite our detailed NMR characterisation. To explore this
possibility we set up crystallization trials for 1 and 2 and for DDAO glycosides of other 2-fluorosugars,
and were successful in establishing accurate x-ray structures for protected forms of the two galacto
derivatives, 3 and 4 as shown in Figure 3. These structures fully confirmed the presumed connectivity of
the aglycone to the sugar moiety in each case. Importantly 3 was also shown to be a potent inactivator
of Abg (which has both β-glucosidase and -galactosidase activities14), reacting with a rate constant of
ki/Ki = 6180 mM-1min-1. Meanwhile, 4 was a very weak inactivator, requiring 1000-fold higher
concentrations than were employed for 3, and even then reacting with a rate constant of only 2 min-1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In order to determine whether these differences in inactivation rate constant are also reflected in
differing hydrolytic rate constants for the corresponding glucoside substrates, the DDAO glucosides 5
and 6 were synthesized via standard Koenigs-Knorr chemistry followed by deprotection using
HCl/methanol. Kinetic parameters for hydrolysis of 5 and 6 by Abg were then determined and,
surprisingly, very similar values of kcat, Km and kcat/Km were found, as seen in Table 1 and Supplemental
Table 1. Thus, while the enzyme discriminates by approximately 106-fold between the pairs of DDAO 2-
fluoroglycosides 1 / 2 and 3 / 4, there is essentially no discrimination between the pair of glucoside
substrates. Importantly however, the second order rate constants for the two substrates (kcat/Km) are
both very large and, at ~ 2 x 107 M-1 s-1, similar to those previously measured for hydrolysis of 2,4-
ACS Paragon Plus Environment

Page 7 of 23
Journal of the American Chemical Society
6
1
2
3
4
5
6
7
8
9
dinitrophenyl -glucoside and 2,4,6-trichlorophenyl -glucoside by Abg.14 In that previous study, the
Brønsted plot of log (kcat/Km) vs. aglycone pKa had shown pronounced downward curvature and this was
attributed as being most likely due to diffusion control becoming rate-limiting for substrates with the
lowest aglycone pKa values. Since 5 and 6 have similar kcat/Km values it is probable that diffusion control
is rate-limiting for both these substrates, hence the lack of observed kinetic discrimination. Indeed, even
the non-chlorinated glucosyl acridinone DAO glucoside 7, (DAO pKa = 7.0) is hydrolysed by Abg at similar
rates, approximating diffusion control (Table 1, Supp Table 1). By contrast, within the 2-fluoroglycoside
series, only the proximal derivative 1 has a rate constant in that range, while the distal derivative 2 has a
much slower, rate-limiting, chemical step. The inductive destabilisation afforded by the fluorine has
brought rates down to below diffusion control.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A key question then was whether the surprising differences in inactivation rates seen between 1
and 2 (but masked for substrates 5, 6 and 7) were due to differences in enzyme recognition and
catalysis, or whether they had their origins in fundamental reactivity differences. To address this
question we measured rates of non-enzymatic hydrolysis for 5 and 6 at 50 °C, but otherwise under the
conditions used in the enzymatic studies (pH 7, phosphate buffer). DDAOprox glucoside 5 hydrolysed
completely within 10 hours, allowing determination of its solvolytic rate constant by direct fit of the
time course data to a first order expression, ksolv = 0.46 h-1. Hydrolysis of 6 and 7, by contrast, was
extremely slow. Thus, their respective rate constants were determined by measurement of initial rates
at a series of substrate concentrations. Since possible contamination by small amounts of 5 could
confound initial rate measurements, aglycone release was monitored for over 14 hours, during which
any contaminating 5 would have hydrolysed (t = 1.9 h). Results are shown in Table 1, alongside a
½
literature value for hydrolysis of 2,4-dinitrophenyl -glucoside.18 As is evident, under these conditions 5
hydrolyses nearly 1,000 fold more rapidly than 6 and 7, and at approximately the same rate as 2,4-
dinitrophenyl -glucoside. Clearly, placement of the two chlorine atoms proximal to the glycosidic bond
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 23
7
1
2
3
4
5
6
has major consequences, thus the differences in inactivation rate observed with Abg have their origins,
at least partially, in fundamental reactivity differences.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Since these hydrolytic rate constants may well be differentially pH sensitive – thereby containing
variable contributions from various pathways – a quantitative comparison required a more careful
study. Accordingly, rate constants for spontaneous hydrolysis of 5 and 6 at a series of pH values ranging
from approximately pH 1 to pH 11 were measured and the data and resultant plots are shown in Figure
4 and Supplemental Table 3. As can be seen, rate constants for hydrolysis of 5 are larger than those for 6
at each of the pH values studied. The profile observed for 5 includes a pH-independent region from
approximately pH 4 to 8.5 with flanking pH-dependent arms in which rate constants increase as pH
moves away from neutrality. As shown previously for solvolysis of 2,4-dinitrophenyl -glucoside, this pH-
independent region corresponds to simple heterolysis of the glycosidic bond while the flanking limbs
contain additional pathways involving acid or base assistance.18-19 The pH profile for hydrolysis of 6 is
similar, but with a narrower pH-independent regime from approximately pH 4 to 5. The presence of this
pH-independent region for both substrates allows direct comparison of rate constants for simple
heterolysis of the glycosidic bond in the two cases: an approximate 105 fold difference in rate constants
is seen. This enormous difference closely reflects the large (~106 fold) difference in rates of Abg
inactivation by the pair of 2-fluoroglucosides. Thus it appears that a large portion of the difference in
enzymatic inactivation rates resides in fundamental reactivity differences between the two isomers
rather than being due to different substrate-specific interactions with the enzyme.
One possible explanation for this vast difference in fundamental reactivity, despite their
apparently very similar transition states, is that a different mechanism is followed in the two cases. The
most probable alternative would be one involving ipso-attack of water on carbon-1 of the aryl moiety of
ACS Paragon Plus Environment

Page 9 of 23
Journal of the American Chemical Society
8
1
2
3
4
5
6
7
8
9
the proximally substituted derivative 5, since this carbon, situated between the two chlorines, is very
electron deficient. To investigate this possibility, 5 was incubated in H218O phosphate buffer at 50 °C and
the composition of the hydrolysis products over time was determined to explore possible 18O
incorporation into the DDAO product. The resulting products – 16O-DDAO, single 18O-substituted
DDA(18O) and double 18O-substituted DDA(18O)2 – were monitored by mass spectrometry and their ratios
were calculated (taking into account the chlorine isotopes). The relative concentration of each species as
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a function of time was then determined by substitution into a first order equation representing
decay, y = A푒푘
푡, using the rate constant of khydr = 0.46 h-1 measured previously. This value
ℎ푦푑푟
represents the combined rates of hydrolysis from attack at both the anomeric carbon and the C-1 aryl
carbon.
As is shown in Figure 5, the analysis of the resultant data is rendered somewhat more complex by the
possibility of “wash-in” of 18O into the carbonyl group of the DDAO moiety, either into the glycoside 5 or
into the released DDAO. To simplify the analysis, rates of these processes were measured in simple
parallel experiments, yielding a rate constant for “wash-in” of 18O into DDAO and the DDAOprox-
glucoside 1 of k5 = 0.02 h-1, and k2 = 0.23 h-1 respectively.
The resulting percentages as a function of time are shown in Figure 6. These progress curves were fit
globally using the KinTek Explorer software20-21 (Supp Info Table 2), revealing that the ipso attack
mechanism contributes a maximum of 24% to the overall rate of hydrolysis, which is clearly not
sufficient to explain the enormous rate differences observed between 5 and 6. These results therefore
confirm that, at pH 7, no significant change of mechanism has occurred between substrates and that the
primary pathway for hydrolysis of 5 is that of the originally anticipated heterolysis of the glycosidic
bond. These enormous (>105 fold) differences in reactivity of two isomeric glycosides that hydrolyse
through very similar dissociative transition states were very surprising and are the underlying reason for
the differences in glycosidase inactivation efficiency observed for 1 and 2.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 23
9
1
2
3
4
5
6
7
8
9
Given the essentially electronically-identical transition state structures, some of the difference
in reactivity must have its origins in ground state structural differences, be they steric effects arising
from the relatively bulky ortho-chloro substituents, or electronic due to the inductive effects of the two
halogens. Closer inspection of the X-ray crystal structures obtained for the per-O-acetylated versions of
3 and 4 (S3 and S4) reveals somewhat different conformations of the two glycosides. The plane of the
aromatic ring is tilted 50o away from the plane of the sugar ring in S4, while in S3, with its proximal
chlorine substituents, the aromatic ring is almost perpendicular to the sugar ring plane (75o). This
displacement appears to be a response to steric constraints imposed by these halogens. Perhaps more
importantly, there are small but very significant differences in the glycosidic bond lengths in the two
cases, with that for S3 being 0.009 Å longer than that for S4. This longer bond length for S3 is consistent
with inductive effects that favour charge development on the glycosidic oxygen in the early stages of
ionisation/heterolysis of the glycosidic bond towards the oxocarbenium ion/phenolate pair. Such
behaviour has been investigated and quantitated in some detail by Kirby et al. through determination of
x-ray crystal structures of several series of axial and equatorial aryl acetals bearing different aryl
substituents within both a tetrahydropyranyl series and a glycosyl series.18, 22-24 Within each series Kirby
et al. observed a linear correlation of the acetal carbon-aryl oxygen bond length with the pKa of the
phenol from which it had been derived. The sensitivity (slope) of that correlation varied for each system,
being higher for axial than equatorial systems, consistent with expectations of stereoelectronic theory.22
The sensitivity was decreased by electronegative substituents alpha to the acetal centre, with -
glucosides (OH at C2) showing a smaller dependence (0.00285 Å / pKa) than the equatorial
tetrahydropyranyl (H at C2) system (0.00476 Å / pKa).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The most directly comparable data from the Kirby studies are for the (equatorial) -glucosides. Applying
the value of bond length vs. pKa dependence observed in that case to the bond length difference
measured between S3 and S4 (0.009 Å) suggests a difference in effective pKa of the phenols of 1 and 2 of
ACS Paragon Plus Environment

Page 11 of 23
Journal of the American Chemical Society
10
1
2
3
4
5
6
7
8
9
a little over 3 units. However, this is undoubtedly an underestimate given the abovementioned
dependence of this value on electronegative substituents and the presence, in 1 and 2, of a fluorine
substituent at C-2. Based upon the I values for H, OH and F of 0, 0.25 and 0.50, substantial inductive
effects could be expected, as was seen in our previous data on the pH-independent heterolysis of 2,4-
dinitrophenyl glycosides, where the solvolysis rate constant correlated strongly ( = -9)with the I value
of the 2-substituent.18 This would suggest a difference of effective pKa of the phenols in the two isomers
of at least 3 to 4 units. Using the average βlg value of -1.25 ± 0.02 determined within four independent
studies of spontaneous hexopyranoside hydrolysis15 this pKa difference should translate into a rate
difference of approximately 104 fold. Thus it would seem that much of the rate difference may be
explained by different effective leaving group pKa values.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To probe the relative contributions of transition state and ground state energetic effects,
computational models for the spontaneous hydrolysis of the proximal and distal glucosides 5 and 6 were
constructed using the SMD/M06-2X/6-31+G(d,p) quantum chemical methods (see Supporting
Information). These models predicted that the ground state structure of 5 was indeed 5.3 kcal mol-1 less
stable than the ground state structure of 6, as suggested from the measured bond length data above.
Somewhat surprisingly, they also suggested that the transition state structure of 5 was 4.7 kcal mol-1
more stable than the corresponding structure for 6. These differences in energy are reflected in key
transition state structure bond lengths: the anomeric carbon – leaving group oxygen bond length for
DDAOprox-Glc 5 is 2.276 Å, with an associated anomeric carbon – nucleophilic water oxygen bond
length of 2.596 Å, while the C–Olg bond length for DDAOdist-Glc 6 is 2.360 Å with an associated C–Owater
bond length of 2.529 Å. These observed bond lengths are consistent with DDAOprox being a better
leaving group than DDAOdist, yielding an earlier transition state, i.e. both bond fission and bond
formation are less far advanced in the transition state for DDAOprox than for DDAOdist due to the
better stabilisation of the charge at the leaving group oxygen. In addition to these differences in key
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 23
11
1
2
3
4
5
6
7
8
9
transition state bond lengths, the conformational searches carried out in determining the lowest energy
ground state and transition state structures (supplementary info) reveal that the dispositions of the
aglycone with respect to the glycone in both ground state and transition state structures are quite
distinct for DDAOprox and DDAOdist. The models, in combination with transition state theory, predict
that 5 would be spontaneously hydrolyzed some 6 × 106-fold faster than 6 at the experimental
temperature of 50 °C. These results are consistent with the experimentally observed difference in
spontaneous rates of hydrolysis and, perhaps surprisingly, indicate that both ground state and transition
state energy differences contribute almost equally to the overall difference in rates of hydrolysis. The
primary source of these effects is likely electronic, due to through-bond inductive effects from the
proximal chlorines, though steric contributions to the ground state destabilisation cannot be discounted
at this stage.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSIONS
The surprisingly large differences in reaction rate of two isomeric glycosides that release the
same aglycone were explored. Through the use of a pseudo-symmetric aglycone the resonance
contributions to transition state stabilisation in the reactions of the two glycosides are made largely
equivalent, such that rate differences arise primarily from inductive or steric effects on the ground state.
These effects are surprisingly large, and do not truly reflect the measured pKa values of DDAO (5.3) and
DAO (7.0); rather, they must reflect values of effective pKa at the reaction transition state. The study
therefore serves as a reminder of the limitations (hazards) of relating effects on rate to thermodynamic
parameters such as pKa values.
MATERIALS AND METHODS
All experimental protocols plus product characterisations are provided as Supplementary
Information.
ACS Paragon Plus Environment

Page 13 of 23
Journal of the American Chemical Society
12
1
2
3
4
Acknowledgments
5
6
7
8
9
We thank the Natural Sciences and Engineering Research Council of Canada for financial support of this
work and numerous physical organic colleagues for their comments. Computing facilities and quantum
chemical software were provided by WestGrid (www.westgrid.ca) and Compute Canada Calcul Canada
(www.computecanada.ca).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 23
13
1
2
3
4
5
6
Figure 1.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 15 of 23
Journal of the American Chemical Society
14
1
2
3
4
5
6
Figure 2. Inactivation of Abg by DDAO 2-fluoroglucosides
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 23
15
1
2
3
4
5
6
7
Figure 3.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
ACS Paragon Plus Environment

Page 17 of 23
Journal of the American Chemical Society
16
1
2
3
4
Figure 4.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
~105
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 23
17
1
2
3
4
Figure 5.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 19 of 23
Journal of the American Chemical Society
18
1
2
3
4
Figure 6.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 23
19
1
2
3
4
Figure captions
5
6
7
8
Figure 1. Aryl glycosides (1-7) synthesized in this study. DDAO-proximal glycosides (1,3,5) have the Cl
atoms of DDAO adjacent to the anomeric center, while DDAO-distal glucosides (2,4,6) have the Cl atoms
on the opposite side of the aglycone. DDAO was dechlorinated and coupled with the glucosyl bromide to
form DAO-glucoside after deprotection (7).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Mechanistic depiction of -glucosidase-mediated hydrolysis of DDAO-2FGlc. DDAOprox (1) and
DDAOdist (2) 2F-glucosides are regioisomers that, when cleaved, release the same aglycone (DDAO).
Figure 3. a) Crystal structure of per-O-acetylated DDAOprox-2F-Gal (S3) b) Crystal structure of per-O-
acetylated DDAOdist-2F-Gal (S4) Colour code: gray – carbon; red - oxygen; blue - nitrogen; yellow -
fluorine; green - chlorine. Hydrogens have been omitted for clarity.
Figure 4. pH-Dependent hydrolysis of DDAO glucosides (5,6). Logarithm of spontaneous hydrolysis rate
(50 mM buffer, 50 oC) versus pH of buffer. At neutral pH comparison of hydrolytic rates show 5 cleaves
>105-fold faster than regioisomer 6.
Figure 5. Potential breakdown pathways of DDAOprox-Glc (5) in heavy water (H218O) at 50 oC. A to B (k1)
occurs via ‘wash-in’ of water to the distal carbonyl of the aglycone; A to C (k2) and B to D (k4) occurs via
classical heterolysis of water to the anomeric carbon of glucose; A to D (k3) and B to E (k7) hydrolysis
occurs via attack of water on the ipso carbon of the aglycone; C to D (k5) and D to E (k6) occurs via ‘wash-
in’ of water to the distal carbonyl of DDAO. Product m/z values of A – E were monitored over time via
mass-spectrometry.
Figure 6. Monitoring DDAOprox-Glc cleavage over time in H218O at 50 oC. Inset model - A = DDAOprox-
Glc; B = DDA18Oprox-Glc; C = DDAO; D = DDA18O; E = DDA(18O)2. Best fit curves are generated by KinTek
Explorer20 to determine rates of each pathway. Constraints employed to fitting parameters: k2 = k4; k3 =
k7; k6 = k5/2; k2 + k3 = 0.46 h-1.
ACS Paragon Plus Environment

Page 21 of 23
Journal of the American Chemical Society
20
1
2
3
4
Table 1. Kinetic parameters for Abg-catalysed, and spontaneous hydrolysis of aryl glycosides
5
6
7
8
Compound
2F-DDAOprox-Glc (1)16
2F-DDAOdist-Glc (2)
2F-DDAOprox-Gal (3)
2F-DDAOdist-Gal (4)
DDAOprox-Glc (5)
DDAOdist-Glc (6)
DAO-Glc (7)
ki/Ki (mM-1 min-1)
kcat/Km (s-1 mM-1)
Spontaneous hydrolysis (s-1)
1.3 x 106
-
-
-
-
-
-
-
-
0.5
6180
<20
-
-
-
-
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
28000
17000
17000
5960
1.27 x 10-4 (pH 5) 50 oC
2.29 x 10-9 (pH 5) 50 oC
2.54 x 10-8 (pH 5) 50 oC
5.58 × 10-6 (pH 6.5) 37 oC
2,4-dinitrophenyl β-
glucoside18
2,4,6-trichlorophenyl β-
glucoside14
-
26000
-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 23
21
1
2
3
4
5
6
References
7
8
9
1.
2.
3.
Capon, B. Chem. Rev. 1969, 69 (4), 407-&.
Greig, I. R. Chem. Soc. Rev. 2010, 39 (6), 2272-2301.
Jaffé, H. H. Chem. Rev. 1953, 53 (2), 191-261.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Johnson, C. D., The Hammett Equation. Cambridge University Press, New Yor: 1973.
Allen, F. H.; Kirby, A. J. J. Am. Chem. Soc. 1984, 106 (21), 6197-6200.
Overend, W. G.; Rees, C. W.; Sequeira, J. S. J. Chem. Soc. 1962, (0), 3429-3440.
Amyes, T. L.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111 (20), 7888-7900.
Banait, N. S.; Jencks, W. P. J. Am. Chem. Soc. 1991, 113 (21), 7951-7958.
Bennet, A. J.; Sinnott, M. L. J. Am. Chem. Soc. 1986, 108 (23), 7287-7294.
Zhu, J.; Bennet, A. J. J. Am. Chem. Soc. 1998, 120 (16), 3887-3893.
Sinnott, M. L., The Chemistry of Enzyme Action. Elsevier: New York, 1984.
Sinnott, M. L. Chem. Rev. 1990, 90 (7), 1171-1202.
Vocadlo, D. J.; Davies, G. J. Curr. Opin. Chem. Biol. 2008, 12 (5), 539-555.
Kempton, J. B.; Withers, S. G. Biochemistry 1992, 31 (41), 9961-9969.
Richard, J. P.; Westerfeld, J. G.; Lin, S. Biochemistry 1995, 34 (37), 11703-11712.
Duo, T.; Goddard-Borger, E. D.; Withers, S. G. Chem. Commun. 2014, 50 (66), 9379-9382.
Gao, Z. Z.; Niikura, M.; Withers, S. G. Angew. Chemie Int. Ed. 2017, 56 (22), 6112-6116.
Namchuk, M. N.; McCarter, J. D.; Becalski, A.; Andrews, T.; Withers, S. G. J. Am. Chem. Soc. 2000,
122 (7), 1270-1277.
19.
20.
21.
22.
6200.
23.
24.
Cocker, D.; Sinnott, M. L. J. Chem. Soc.-Perkin 2 1975, (13), 1391-1395.
Johnson, K. A.; Simpson, Z. B.; Blom, T. Anal. Biochem. 2009, 387 (1), 20-29.
Johnson, K. A.; Simpson, Z. B.; Blom, T. Anal. Biochem. 2009, 387 (1), 30-41.
Briggs, A. J.; Glenn, R.; Jones, P. G.; Kirby, A. J.; Ramaswamy, P. J. Am. Chem. Soc. 1984, 106,
Jones, P. G.; Kirby, A. J. J. Am. Chem. Soc. 1984, 106, 6207.
Deslongchamps, P.; Dory, Y. L.; Li, S. G. Can. J. Chem. 1994, 72, 2021.
ACS Paragon Plus Environment

Page 23 of 23
Journal of the American Chemical Society
22
1
2
3
4
TOC Graphic
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment