Glycoconjugate Oxime Formation Catalyzed at Neutral pH: Mechanistic Insights and Applications of 1,4-Diaminobenzene as a Superior Catalyst for Complex Carbohydrates

Article abstract of DOI:10.1021/acs.bioconjchem.8b00019

The reaction of unprotected carbohydrates with aminooxy reagents to provide oximes is a key method for the construction of glycoconjugates. Aniline and derivatives serve as organocatalysts for the formation of oximes from simple aldehydes, and we have previously reported that aniline also catalyzes the formation of oximes from the more complex aldehydes, carbohydrates. Here, we present a comprehensive study of the effect of aniline analogues on the formation of carbohydrate oximes and related glycoconjugates depending on organocatalyst structure, pH, nucleophile, and carbohydrate, covering more than 150 different reaction conditions. The observed superiority of the 1,4-diaminobenzene (PDA) catalyst at neutral pH is rationalized by NMR analyses and DFT studies of reaction intermediates. Carbohydrate oxime formation at pH 7 is demonstrated by the formation of a bioactive glycoconjugate from a labile, decorated octasaccharide originating from exopolysaccharides of the soil bacterium Mesorhizobium loti. This study of glycoconjugate formation includes the first direct comparison of aniline-catalyzed reaction rates and equilibrium constants for different classes of nucleophiles, including primary oxyamines, secondary N-alkyl oxyamines, as well as aryl and arylsulfonyl hydrazides. We identified 1,4-diaminobenzene as a superior catalyst for the construction of oxime-linked glycoconjugates under mild conditions.

Full text of DOI:10.1021/acs.bioconjchem.8b00019

Subscriber access provided by MT ROYAL COLLEGE
Article
Glycoconjugate Oxime Formation Catalyzed at Neutral
pH: Mechanistic Insights and Applications of 1,4-
Diaminobenzene as Superior Catalyst for Complex Carbohydrates
Mads Østergaard, Niels Johan Christensen, Christian T. Hjuler, Knud J. Jensen, and Mikkel B. Thygesen
Bioconjugate Chem., Just Accepted Manuscript • DOI: 10.1021/acs.bioconjchem.8b00019 • Publication Date (Web): 13 Feb 2018
Downloaded from http://pubs.acs.org on February 20, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Bioconjugate Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
Glycoconjugate Oxime Formation Catalyzed at
Neutral pH: Mechanistic Insights and Applications
of 1,4ꢀDiaminobenzene as Superior Catalyst for
Complex Carbohydrates
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
Mads Østergaard, Niels Johan Christensen, Christian T. Hjuler, Knud J. Jensen* and Mikkel B.
Thygesen*
Department of Chemistry, Faculty of Science, University of Copenhagen, Thorvaldsensvej 40,
DKꢀ1871 Frederiksberg C, Denmark
Correspondence to: kjj@chem.ku.dk and mbth@chem.ku.dk
Abstract
The reaction of unprotected carbohydrates with aminooxy reagents to provide oximes is a key
method for the construction of glycoconjugates. Aniline and derivatives serve as organocatalysts
for the formation of oximes from simple aldehydes, and we have previously reported that aniline
also catalyzes the formation of oximes from the more complex aldehydes, carbohydrates. Here
we present a comprehensive study of the effect of aniline analogs on formation of carbohydrate
oximes and related glycoconjugates, depending on organocatalyst structure, pH, nucleophile, and
carbohydrate, covering more than 150 different reaction conditions. The observed superiority of
ACS Paragon Plus Environment
1

Bioconjugate Chemistry
Page 2 of 45
1
2
3
4
5
6
7
8
9
the 1,4ꢀdiaminobenzene (PDA) catalyst at neutral pH is rationalized by NMR analyses and DFT
studies of reaction intermediates. Carbohydrate oxime formation at pH 7 is demonstrated by the
formation of a bioactive glycoconjugate from a labile, decorated octasaccharide originating from
exopolysaccharides of the soil bacterium Mesorhizobium loti. This study of glycoconjugate
formation includes the first direct comparison of anilineꢀcatalyzed reaction rates and equilibrium
constants for different classes of nucleophiles, including primary oxyamines, secondary Nꢀalkyl
oxyamines, as well as aryl and arylsulfonyl hydrazides. We identified 1,4ꢀdiaminobenzene as a
superior catalyst for the construction of oximeꢀlinked glycoconjugates under mild conditions.
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
Introduction
The classical condensation reactions of oxyamines and hydrazides with aldehydes or ketones
have become powerful methods for modifying or linking biomacromolecules.1ꢀ2 Among the
advantages of this approach is the fact that oxime and hydrazone formation occur in aqueous
medium with a high degree of chemoselectivity with unprotected biomolecules, for example
peptides, proteins, carbohydrates or oligonucleotides. It is well established that anilines act as
nucleophilic catalysts for these reactions, enhancing reaction rates by several orders of
magnitude.3 The key step in the catalytic pathway is the formation of protonated iminium
species.4 While imines are generally considered less reactive than their parent carbonyl
compounds towards nucleophilic addition, protonated imines are several orders of magnitude
more reactive.4 The use of aniline catalysts has broadened the scope of these reactions towards
milder conditions, more suitable for conjugation of biomolecules.5ꢀ6
ACS Paragon Plus Environment
2

Page 3 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
In 2010, we reported the extension of nucleophilic catalysis to a different class of aldehydes,
reducing carbohydrates.7 Designed glycoconjugates are important molecules, for example as
drugs, or in biosensors, as well as for the study of fundamental aspects of glycobiology.8ꢀ11 Novel
methods for glycoconjugate synthesis from unprotected carbohydrates is a rapidly developing
field.12 Reducing carbohydrates act as electrophiles in oxime formation, and the masking of the
aldehydo/keto species as hemiꢀacetal/ketal moieties causes a lowering of reaction rates (Fig. 1).7
Catalysis of carbohydrate oxime formation is complicated by the fact that protonated iminium
intermediates may be trapped reversibly7 as cyclic Nꢀarylꢀglycosylamines by intramolecular
additionꢀelimination reactions (Fig. 1). Catalytic rate enhancements (kcat/kuncat) are in the order of
20ꢀfold for glycosyl (aldose) oximes at high catalyst concentration (100 mM) and pH 4.5.7 Such
rate enhancements are particularly important for preparation of glycoconjugates from
carbohydrates of low reactivity, for example oligosaccharides having 2ꢀacetamidoꢀ2ꢀ
deoxyglucose (GlcNAc) at the reducing end. Catalyzed oxime formation has enabled the
synthesis of a range of complex oligoꢀ and polysaccharide conjugate oximes at pH 4ꢀ5,7, 13ꢀ19
exemplified by the labeling of complex influenza antigen triꢀ, tetraꢀ and pentasaccharides.20
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
3

Bioconjugate Chemistry
Page 4 of 45
1
2
3
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
Figure 1. Anilineꢀcatalyzed glycoconjugate formation from αꢀnucleophiles (R1ZꢀNHR2)
proceeds through a protonated iminium intermediate. Low degree of reversibility in additionꢀ
elimination steps is indicated by single arrows. Z: Heteroatom, O (hydroxylamines) or NH
(hydrazines).
Recent efforts towards optimizing nucleophilic catalysts with respect to efficiency at neutral
pH, have built on the early discovery of 4ꢀmethoxyaniline (paraꢀanisidine, PAN)3 as a useful
catalyst for peptide5 and carbohydrate oxime7 formation at pH 7. This led to the discovery of 2ꢀ
aminoꢀ5ꢀmethoxybenzoic acid (5ꢀmethoxyanthranilic acid, PANꢀC) and 3,5ꢀdiaminobenzoic acid
(MDAꢀC) in model hydrazone reactions,21 as well as 1,4ꢀdiaminobenzene22 (paraꢀ
phenylenediamine, PDA) for protein modification by oxime formation, each with 1–2 orders of
magnitude rate enhancements over aniline (Chart 1). Recently, Langenhan et al.23 compared a
panel of anilines for glycoconjugate formation with NꢀbenzylꢀNꢀmethoxyamine and found PANꢀ
C as the preferred catalyst at pH 4.5. However, this catalyst was found to be ineffective at neutral
pH.
ACS Paragon Plus Environment
4

Page 5 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
Chart 1. Aniline catalysts, including acronyms, used in this study and their associated pKa
7
8
9
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
In the present study, we have examined commercially available aniline derivatives from the
aboveꢀmentioned new generation24ꢀ25 of nucleophilic catalysts (Chart 1) with carbohydrates as
electrophiles in order to optimize conditions for glycoconjugate formation towards neutral pH.
This is important because certain glycosidic linkages are prone to hydrolyze even under mildly
acidic conditions, for example uronides (e.g., hyaluronic acids), ulosonic acids (e.g., sialic acids
in mucins, or KDOs), and more.12 Furthermore, complex carbohydrate ligands often carry
‘decorations’ in the form of various substituents, for example sulfates, carbamates, acetates, or
pyruvate ketals, which are highly susceptible to cleavage or rearrangement reactions under acidic
conditions. Such substituents are often crucial for the biological activity of carbohydrate ligands
in receptor binding events.18 Neutral conditions are often also preferable during protein glycoꢀ
modification and in dynamic covalent chemistry based on oxime or hydrazone chemistry. In the
present study, we find that PDA is a superior catalyst for aldose oxime formation at pH 7, and
report NMR studies and DFT calculations to rationalize the extraordinary catalytic effect of
PDA.
ACS Paragon Plus Environment
5

Bioconjugate Chemistry
Page 6 of 45
1
2
3
4
Results and Discussion
5
6
7
8
Anilineꢀbased catalysis of reactions between aldehydes or ketones and αꢀnucleophiles rely on the
formation of the intermediate imine with aniline being kinetically rapid, but thermodynamically
unfavourable relative to the desired product.4 In general, intermediate imine species do not
accumulate, or accumulate to a lesser extent,21 but once protonated they react rapidly with the αꢀ
nucleophile to yield the products. When carbohydrates are used as electrophiles, reactivity
towards αꢀnucleophiles differ markedly between monosaccharides as a result of different rates of
pyranose ringꢀopening. The rate of formation for the reactive aldehyde species is generally
governed by the thermodynamic stabilities of the pyranoses. Furthermore, various substitutions
to the pyranose ring influence the rate of oxime formation. In particular, the 2ꢀacetamidoꢀ2ꢀ
deoxy and 6ꢀcarboxy substitutions are known to decrease or increase reaction rates,
respectively,7 and could potentially influence catalysis. These considerations led us to first study
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
the catalytic effect for the three monosaccharides,
Dꢀglucose (Glc), 2ꢀacetamidoꢀ2ꢀdeoxyꢀDꢀ
glucose (GlcNAc), and ꢀglucuronic acid (GlcA), in the pH range 3ꢀ7 for five different anilines
D
(supporting information, Fig. S1), yielding products 1ꢀ11 (Chart 2).
ACS Paragon Plus Environment
6

Page 7 of 45
Bioconjugate Chemistry
1
2
3
4
Chart 2. Products of model glycoconjugation reactions compared in this study.
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
Glycoconjugate formation rates for different nucleophiles. We initially examined reactions of
the three monosaccharides with the structurally related nucleophiles, primary Oꢀ
benzylhydroxylamine (oxime products 1ꢀ3) and secondary, Nꢀsubstituted OꢀbenzylꢀNꢀ
methylhydroxylꢀamine (oxyamine products 4ꢀ6) and NꢀbenzylꢀOꢀmethylhydroxylamine
(oxyamine products 7ꢀ9) in the presence or absence of aniline. Importantly, the reactions of these
secondary hydroxylamine nucleophiles proceed through a cationic intermediate,9 formally a
quaternized oxime intermediate, which is unable to lose a proton in contrast to the primary
hydroxylamine (Fig. 1). To the best of our knowledge, no direct comparison of glycoconjugation
reaction rates and equilibrium constants between these different types of nucleophiles has been
reported in the literature.
We aimed to compare these different reactions and conditions, which displayed a rate span of
> 10,000ꢀfold between fast and slow reactions, as well as an equilibrium constant span of > 100ꢀ
ACS Paragon Plus Environment
7

Bioconjugate Chemistry
Page 8 of 45
1
2
3
4
5
6
7
8
9
fold for different nucleophiles and pH values. For this we chose data treatment for secondꢀorder
reversible reactions (Equation 1, and supporting information, supplementary methods). The
derived kinetic rate equation was able to provide unique rate constants, k2, for (i) reactions
performed under different starting conditions, as well as (ii) reactions that proceeded to an
equilibrium position rather than to full conversion (reversible reactions); supporting information,
Figs. S10ꢀS52. In the following, k2 values denote the intrinsic secondꢀorder rate constant for the
uncatalyzed equilibrium reactions, whereas k2,app values denote the apparent k2 for the
corresponding catalyzed reaction at a nucleophile/carbohydrate/catalyst ratio of 1:10:10 (Table
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
Table 1. Rate and equilibrium constants for different nucleophiles with D
ꢀglucose.a
Nucleophile
pH
Keq
(M-1)
800
18
k2 (no cat.)
(M-1h-1)b
12.5 ± 0.09
0.8 ± 0.01
0.2 ± 0.01
2.5 ± 0.05
1.4 ± 0.02
0.1 ± 0.01
NDe
k2,app (AN) Product
(M-1h-1)c
BnOꢀNH2
4
4
4
4
4
7
7
7
Oxime 1
43.6 ± 0.27
7.6 ± 0.18
2.7 ± 0.03
20.3 ± 0.80
3.8 ± 0.05
0.8 ± 0.07
0.1 ± 0.01
BnOꢀNHMe
MeOꢀNHBn
BzNHꢀNH2
TsNHꢀNH2
BnOꢀNH2
Oxyamine 4
Oxyamine 7
Hydrazine 10
Hydrazine 11
Oxime 1
18
44
29
3,600
∼40d
∼40d
BnOꢀNHMe
MeOꢀNHBn
Oxyamine 4
NDe
NDe Oxyamine 7
b
aConditions: 50 mM citrate buffer (pH 4.0), or 50 mM phosphate buffer (pH 7.0). Secondꢀ
c
order rate constant in the absence of catalyst ± standard deviation for the fit (σ). Apparent
d
secondꢀorder rate constant ± σ for nucleophile/
D
ꢀglucose/aniline ratio 1:10:10. The reaction did
ACS Paragon Plus Environment
8

Page 9 of 45
Bioconjugate Chemistry
1
2
3
4
5
e
not reach completion during the experiment. Rate constant below 0.01; See supporting
information, Tables S1 and S2 for complete data set.
6
7
Fast reactions were performed with 1 mM nucleophile, whereas slower reactions were
8
9
performed with at 3 or 5 mM nucleophile. Our data show that uncatalyzed reactions of Oꢀbenzylꢀ
Nꢀmethylhydroxylꢀamine were approximately 20ꢀfold slower than reactions with Oꢀ
benzylhydroxylamine. Reactions performed in the presence of aniline showed a 4ꢀ15ꢀfold
enhancement of rate constants across pH 4ꢀ7 (supporting information, Table S1). Interestingly,
we found that the rate constants for the OꢀbenzylꢀNꢀmethylhydroxylamine nucleophile were
more than 3ꢀfold that of the isomeric NꢀbenzylꢀOꢀmethylhydroxylamine (Table 1). This
discrepancy is most likely a result of steric hindrance from the benzyl substituent, which is
located closer to the reactive nitrogen in the latter case. Our comparison with hydrazone
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
formation of benzhydrazide, or tosylhydrazide with Dꢀglucose (products 10 and 11, respectively)
showed that both of these nucleophiles reacted 2ꢀ10ꢀfold faster than both secondary
hydroxylamines. The equilibrium constants, Keq, for the different conjugates generally increased
at higher pH (supporting information, Table S2). Furthermore, Keq was increased by more than
one order of magnitude for reactions using the primary aminooxy nucleophile relative to all other
nucleophiles studied. The exceptionally high Keq for oxime reactions with GlcA indicate that
these oxime bonds are stable down to pH ∼ 3. The low Keq values for reactions using Nꢀ
alkylhydoxylamine or hydrazideꢀtype αꢀnucleophiles imply that very high (>0.3ꢀ0.5 M)
concentrations of either nucleophile or carbohydrate are required for effective conversion in
these cases, whereas oxime formation can be performed using low mM reactant concentrations
(Table 1).
ACS Paragon Plus Environment
9

Bioconjugate Chemistry
Page 10 of 45
1
2
3
4
5
6
7
8
9
Catalysts for aldose oxime formation towards neutral conditions. Next, we focused on oxime
formation with variation in the aniline catalyst structure. Oxime formation with the three
monosaccharides (1ꢀ3) was studied in the pH range 3ꢀ7 for the five different anilines,26 and
second order rate constants for reactions with an equimolar catalyst to carbohydrate ratio were
determined (supporting information, Table S1). We found that the highest reaction rates for all
three monosaccharides were observed at pH 4.0. For Glc and GlcNAc, the PANꢀC catalysed
reaction was the fastest, while the anilineꢀcatalysed reaction was the fastest for GlcA. At pH 3.0,
we observed a decrease in reaction rate relative to pH 4.0, as well as a ∼ 10ꢀfold decrease in the
equilibrium constant. Generally, a decrease in reaction rate and increase in the equilibrium
constant was observed as the pH increased from pH 4. At pH 5, PANꢀC remained the most
effective catalyst in reactions with Glc and GlcNAc, while PDA was superior in the reaction with
GlcA. At pH 6 and 7, PDA provided the fastest reaction rates for all carbohydrates tested. PANꢀ
C, which has a carboxylic acid moiety in the ortho position,21 was a superior catalyst at lower
pH, except in reactions with GlcA, which carries a carboxylic acid moiety. At higher pH, PDA
was a superior catalyst for all three monosaccharides.
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
For each carbohydrate, we evaluated the performance of the five catalysts as a function of pH
by considering the reaction rate of the catalyzed reaction relative to the uncatalyzed reaction (krel
= k2,app/k2, Fig. 2). The relative catalytic rate enhancement, krel, generally increased with
increasing pH, that is, the advantage of using a nucleophilic catalyst becomes greater at higher
pH values. At pH 4, little variance was seen between the different catalysts, while at pH 6 and 7,
PDA greatly outperformed the other catalysts tested. The PDAꢀcatalyzed reaction at pH 7 offered
a 34ꢀfold increase in reaction rate for Glc, 50ꢀfold for GlcA, and 75ꢀfold for GlcNAc relative to
the uncatalyzed reactions. The catalysts showed similar trends in catalytic rate enhancement
ACS Paragon Plus Environment
10

Page 11 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
across the pH values for Glc and GlcA. At pH 6 and 7, PAN was a better catalyst than PANꢀC in
reactions with Glc and GlcA. This relationship was however reversed in reactions with GlcNAc,
for which PANꢀC was a better catalyst than PAN across all pH values. Interestingly, PANꢀC was
an outstanding catalyst at pH 5 in reactions with GlcNAc. In addition, the increase in the
catalytic efficiency of PDA from pH 6 to pH 7 was more prominent for GlcNAc, than for the
other carbohydrates. The trends in catalytic rate enhancement by PDA as a function of pH seen
here resemble those reported by Baca and coꢀworkers22 for aldehydes introduced on proteins,
except that we observed an increase in the catalytic rate enhancement by PDA as the pH
increased from 6 to 7.
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
c
80
60
40
20
0
80
60
40
20
0
80
60
40
20
0
1
2
3
pH 4
pH 5
pH 6
pH 7
pH 4
pH 5
pH 6
pH 7
pH 4
pH 5
pH 6
pH 7
Figure 2. Catalytic rate enhancements (krel = k2,app/k2, i.e. fold rate increase over the uncatalyzed
reaction) for the five catalyst, aniline (red), MDAꢀC (orange), PAN (yellow), PANꢀC (green),
and PDA (blue), for Oꢀbenzyl oxime formation with Glc (a), GlcNAc (b), and GlcA (c) across
pH 4ꢀ7. Error bars indicate standard deviations for the ratio of individual fits, (k2,app±σ)/(k2±σ).
The structure of the oxime products are indicated in each panel.
ACS Paragon Plus Environment
11

Bioconjugate Chemistry
Page 12 of 45
1
2
3
4
5
6
7
8
9
Catalytic profile of 1,4-diaminobenzene. The effective catalyst must encompass two opposing
chemical characteristics: (i) It should be a good nucleophile for a fast kinetic reaction with the
aldehyde, and (ii) it should provide a thermodynamically instable imine for substitution by the
hydroxylamine. The latter implies that it should also be a good leaving group. For anilines,
nucleophilicity as well as basicity of the intermediate imine are generally proportional to the
aniline pKa value, while leaving group capability is inversely proportional to this pKa value.
Thus, a pKaꢀdependent pH optimum exists for each catalyst at which point nucleophile and
leaving group effects are balanced. PAN and PANꢀC both have an optimum range, with respect
to catalytic efficiency, peaking at ∼ pH 6 (Fig. 2). In reactions catalyzed by PDA, we observed
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
that the catalytic efficiency was markedly higher at pH 7, as well as at pH ≤ pKa, compared with
the other catalysts tested. The pKa value of PDA is slightly higher than that of the other catalysts,
but its catalytic efficiency at pH∼pKa was dramatically increased compared to the other catalysts.
PDA is a unique catalyst in that it contains a paraꢀamino group with acid/base properties. It
has been suggested that this second amino group statistically increases the likelihood of imine
formation.27 However, our results contradict this hypothesis, as we found that the catalytic
efficiency was significantly more than twoꢀfold compared to aniline. Furthermore, it has been
reported that 4ꢀ(methylamino)aniline (NꢀmethylꢀPDA), which can only form the imine at the
primary amine position, has a catalytic effect similar to PDA in oxime formation.22
In search of an alternative explanation for the enhanced catalytic effect of PDA we compared
Nꢀarylꢀglucosylamine formation at equilibrium for aniline, PANꢀC, and PDA, by 13C NMR
spectroscopy, as determined from the characteristic Cꢀ1 anomeric peaks for the αꢀ and βꢀ
pyranosyl forms (Fig. 3, and supporting information, Tables S3ꢀS5). Interestingly, the
glucosylamine levels for PANꢀC were significantly lower than for aniline, indicating a
ACS Paragon Plus Environment
12

Page 13 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
destabilizing effect on the imine with PANꢀC, as also reported previously as a key catalytic
driving force.21 Surprisingly, the equilibrium glucosylamine levels for PDA were dramatically
larger than for both aniline and PANꢀC (Fig. 3). From this data we conclude that the PDA
catalyst provides increased stabilization of the intermediate imine species as compared to aniline
and PANꢀC. At neutral conditions, however, more than half of the PDA catalyst will exist in the
uncharged state in which it will be a significantly stronger nucleophile than in the
monoprotonated form (pKa of doubly protonated PDA = 3.1), leading to an increase in the rate of
hemiaminal formation and in imine stabilization. By NMR, we determined the pKa value for the
distal amino group of the PDA glycosylamine (pKa = 6.1, Fig. 4, and supporting information,
Fig. S53). Clearly, the presence of the glucosyl moiety did not significantly alter the pKa of the
distal amino group for this intermediate. Protonation of the distal amino group of corresponding
PDAꢀimine species would presumably lead to destabilization of the imine and enhanced leaving
group properties during substitution by the αꢀnucleophile. Hence, at pH 7, PDA would be
capable of playing the dual role of a strong nucleophile and a good leaving group through proton
exchange reactions. We hypothesized that this could be an important mechanism for enhancing
the catalytic efficiency of PDA at neutral pH.
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
40
30
20
10
0
AN
PANꢀC
PDA
pH 4
pH 5
pH 6
pH 7
ACS Paragon Plus Environment
13

Bioconjugate Chemistry
Page 14 of 45
1
2
3
4
Figure 3. Equilibrium levels of glucosylamines for the catalysts, aniline (red), PANꢀC (green),
5
6
and PDA (blue) in the pH range 4ꢀ7. Determined from the labeled Cꢀ1 signals of Nꢀarylꢀ
D
ꢀ[1ꢀ
7
8
9
13C]glucopyranosylamine in equilibrated solutions of
each) by 13C NMR.
D
ꢀ[1ꢀ13C]glucose and anilines (50 mM
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
85.5
85.0
84.5
84.0
83.5
pKa= 6.1
2
3
4
5
6
7
8
9
10
pH
Figure 4. Determination of pKa value for the distal amino group of Nꢀ(4ꢀaminophenyl)ꢀβꢀ
13C]glucopyranosylamine by 13C NMR in 50 mM aqueous citrate (ꢀ) or phosphate (ꢁ) buffers.
ꢀ[1ꢀ13C]glucosylamine (), lacking the distal amino
Dꢀ[1ꢀ
Corresponding values for Nꢀphenylꢀβꢀ
D
group, showed no pH dependency. * = 13C label.
Interestingly, small amounts of 1,4ꢀbis(glucosylamino)ꢀbenzene were observed in 13C NMR
experiments for PDA, demonstrating that the two amino groups may act in tandem (supporting
information, Fig. S53). However, the amounts of this species relative to the monoꢀglucosylamine
did not reach levels (∼ 3 % at pH 6, supporting information, Table S5) that could justify a
stoichiometric advantage over monoꢀaminobenzenes, such as aniline.
Various other anilines containing a second amino group may undergo proton exchange
reactions, or form bisꢀsubstituted intermediates, as described for PDA above. This includes
MDAꢀC, as well as 1,2ꢀdiaminobenzene (orthoꢀphenylenediamine, ODA), and 1,3ꢀ
ACS Paragon Plus Environment
14

Page 15 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
diaminobenzene (metaꢀphenylenediamine, MDA), which we also tested. However, these three
anilines were found to be inferior catalysts to PDA at pH 7 (supporting information, Table S1).
One explanation could be that MDA and MDAꢀC are poorer nucleophiles than catalysts with a
paraꢀdonating group, and presumably react less rapidly with the aldose, while ODA has a
markedly lower pKa value than PDA, indicating extensive proton sharing of the two proximate
amino groups. Furthermore, the second amino group of ODA may react intramolecularly with
the iminium intermediate thereby lowering the rate of oxime formation.
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
Mechanistic studies. In order to gain a better understanding of the superior catalytic properties
of the PDA catalyst, we sought to analyze reactive intermediates during imine formation in the
13
absence of αꢀnucleophile by CꢀNMR, using [1ꢀ13C]ꢀlabelled
Dꢀglucose and aniline or PDA.
While Nꢀarylꢀglucosylamines were readily detected (Fig. 3 and 4), we were surprisingly unable
to detect any imine, iminium ion, or hemiaminal species even at very high (1 M) reagent
concentrations (data not shown). On this basis, we estimate the concentration of these species to
be below 0.005% of the glucose concentration.28 The fact that neutral imine species were not
detected indicates that aldose imine protonation is highly dynamic or is more prominent than
with other aliphatic or aromatic imines, which are easily detected.21
We also examined the ‘lag phase’ (sigmoidal profile) for product formation observed by HPLC
with several catalysts, including PDA, in catalyzed oxime formation at pH 7 (supporting
information, Figs. S16, S24, and S26). Here, the rate of product formation was not highest at the
start of reaction, but rather after a period of 1ꢀ4 h. This ‘lag phase’, which has also been reported
previously,21 is in the present case presumably due to a buildꢀup of a reservoir of glycosylamine
species available for substitution by the αꢀnucleophile. A kinetic study of the rate of oxime
ACS Paragon Plus Environment
15

Bioconjugate Chemistry
Page 16 of 45
1
2
3
4
formation from preꢀequilibrated solutions of
D
ꢀ[1ꢀ13C]ꢀglucose and catalysts, including aniline,
5
6
7
8
PANꢀC, and PDA, by NMR at pH 6.0, confirmed that a significant amount of the preꢀformed
glucosylamine rapidly disappeared upon addition of the nucleophile (supporting information,
Fig. S54). This reveals that, under the conditions of the experiment, the initial rate for conversion
of glycosylamine to imine (Fig. 5c) was faster than that of formation of glycosylamine from
catalyst and carbohydrate. These findings indicate that the predominant reaction pathway at
neutral pH for catalyzed glycoconjugate oxime formation proceeds via imine formation from
ringꢀopening of glycosylamines.
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
HO HO
HO
HO HO
HO
HO
HO
HO
H
N
HO
HO
OH
CH3
n
H
N
O
OH
H3C
H
H3C
H
H
n
O
OH
H
H
H
H
N
N
R
HO
H
HO
H
H
n H2
O
ꢀn H2O
H
H
H
H
n H2O
ꢀn H2O
HO
N
N
H
H
+
H2O
R
R
R
R
R
Ia: R=H
TS_Ia-d
IIa-d
Ib: R=NH2+
I'a: R=H
TS_I'a-c
II'a-c
Ic: R=NH3
I'b: R=NH2+
Id: R=NHꢀ
ꢀDꢀglucosyl
I'c: R=NH3
β
c
d
Figure 5. Calculated reaction free energy profiles as a function of distal Rꢀgroup protonation
+
state (ꢀNH2, ꢀNH3 ) for imine formation by hemiaminal dehydration (a, b), or pyranosylamine
ringꢀopening (c, d). Proton transfers were considered for n=0ꢀ2 participating water molecules (a,
ACS Paragon Plus Environment
16

Page 17 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
c). Resulting energy profiles are shown for n=0 (b, d). Blue arrows indicate lowered barriers for
distally protonated PDA intermediates.
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
To investigate our hypothesis for the catalytic mechanism of PDA at neutral pH, we performed
DFT calculations enabling comparison of key reaction barriers for the neutral and the distally
+
protonated form of PDA (Fig. 5 and 6, R=NH2, and NH3 , respectively). Reference calculations
were done for aniline (Fig. 5 and 6, R= H). In these calculations we used
Dꢀglucose in ringꢀ
opening steps (Fig. 5c). Truncated electrophiles and αꢀnucleophiles, acetaldehyde and
methoxyamine, respectively, were used in all other reaction steps (Fig. 5a, 6a). The DFT results
reported here for aniline closely match the results reported for the same model compounds at
nearly the same level of theory by Kirmizialtin et al.29 Our inclusion of the D3 empirical
dispersion correction30 to B3LYP may account for the small discrepancies at some points in the
reaction free energy profile.
DFT calculated free energies were determined for a series of important reaction steps. Proton
transfers were generally calculated in the absence of explicit water, with reference calculations
performed with one or two water molecules acting as proton shuttles (supporting information,
Figs. S57 and S58). Calculated free energies for the dehydration of the protonated Nꢀarylꢀ
hemiaminals (Fig. 5aꢀb), and the analogous ringꢀopening of the Nꢀarylꢀglucosylamines (Fig. 5cꢀ
d), showed that the barriers were approximately 2ꢀ2.5 kcal/mol higher for ringꢀopening (TS_I’)
compared to dehydration (TS_I). Otherwise, the relative differences in barriers between aniline,
PDA, and distally protonated PDA intermediates were highly similar for dehydration and ringꢀ
opening.
ACS Paragon Plus Environment
17

Bioconjugate Chemistry
Page 18 of 45
1
2
3
4
5
6
7
8
9
As seen in Fig. 5 (and supporting information, Fig. S55), distally protonated PDA gave
noticeable lower barriers for dehydration of ꢁG‡ = 30.3 kcal/mol (TS_Ic) relative to both aniline
and PDA with identical barriers of ꢁG‡ = 32.4 kcal/mol (TS_Ia and TS_Ib). For ringꢀopening,
the corresponding barrier for distally protonated PDA was ꢁG‡ = 32.5 kcal/mol (TS_I’c)
compared to aniline and PDA with similar barriers of ꢁG‡ = 34.2 and 35.0 kcal/mol (TS_I’a and
TS_I’b, respectively). Based on the Eyring equation,31 the lower barrier for dehydration for
distally protonated PDA corresponds roughly to 35ꢀfold and 40ꢀfold rate enhancements relative
to aniline and PDA, respectively. For ringꢀopening, the corresponding rate enhancements for
distally protonated PDA are 18ꢀfold and 68ꢀfold relative to aniline and PDA, respectively.
Dehydration in the presence of a distal glucosylamino group was calculated in order to
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
13
investigate any catalytic role of the bisꢀsubstituted PDA observed in C NMR experiments (Fig.
5a). The calculated barrier for this species (TS_Id, ꢁG‡ = 33.0 kcal/mol) was higher than that of
any of the other intermediates. Thus, the presence of 1,4ꢀbis(glycosylamino)benzenes are not
expected to contribute significantly to the catalytic effect of PDA.
Comparing the barriers (ꢁG‡ values) across catalysts for reactions steps in imine to oxime
conversion shows that PDA gave the smallest barrier to proton transfer of 22.8 kcal/mol
(TS_IIa-c, Fig. 6), followed by aniline (24.0 kcal/mol) and distally protonated PDA (25.8
kcal/mol). Thus, a 3.0 kcal/mol penalty was observed for distal protonation of the PDA
intermediate in this step. The reverse order holds for the barriers for the catalyst leaving group
step, i.e. departure of distally protonated PDA (4.9 kcal/mol), aniline (6.0 kcal/mol), and PDA
(7.1 kcal/mol); supporting information, Fig. S56.
ACS Paragon Plus Environment
18

Page 19 of 45
Bioconjugate Chemistry
1
2
3
4
a
5
6
7
8
H3C
H3C
O
H3C
O
H
H
O
H
N
H
N
H
δ+
N
N
OH
H
H3C
H
H3C
H
H3C
H
H
H
H
N
H
n
n H2O
ꢀn H2O
N
δ+
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
R
R
R
IIIa: R=H
TS_IIa-c
IVa-c
IIIb: R= NH2+
IIIc: R= NH3
b
Figure 6. Calculated reaction free energy profiles as a function of distal Rꢀgroup protonation
+
state (ꢀNH2, ꢀNH3 ) for the key proton transfer during imine to oxime conversion. Proton
transfers were considered for n=0ꢀ2 participating water molecules (a). Resulting energy profiles
are shown for n=0 (b). The blue arrow indicates a lowered barrier for the neutral PDA
intermediate in this step.
Overall, the calculated energies are consistent with the qualitative trends expected from the
+
inductive effects of the catalyst substituents: Relative to aniline, electron withdrawal by ꢀNH3 in
distally protonated PDA should destabilize transition states where the transferred proton is
located close to the ring, while electron donation by ꢀNH2 in PDA should stabilize the
corresponding transition states. The same inductive effects are expected to increase and decrease
the barrier to catalyst expulsion in PDA and distally protonated PDA, respectively: The breaking
ACS Paragon Plus Environment
19

Bioconjugate Chemistry
Page 20 of 45
1
2
3
4
CꢀN bond is strengthened by electron donation for R=NH2, and weakened by electron
5
6
+
withdrawal by R=NH3 in the distally protonated PDA.
7
8
9
Collectively, these results support the hypothesis, that the pair, PDA and distally protonated
PDA, provides a catalytic advantage by being able to switch distal protonation state at different
intermediates and before key barriers in the reaction path (Fig. 7).
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 7. Switching of distal protonation state (blue: neutral, and red: protonated) for important
intermediates in PDAꢀcatalyzed carbohydrate oxime formation. The fastest route, based on our
DFT calculations, is indicated by the grey arrow. Intermediates are shown in the charged forms
preceding key reaction steps.
Oxime formation with bioactive octasaccharide at pH 7. To study the applicability of PDA as
an efficient catalyst for oxime formation at pH 7, we first investigated different catalyst
concentrations. Formation of the Glc oxime at pH 7 under PDA catalysis showed a
concentrationꢀdependent increase in the rate of product formation with increasing PDA
ACS Paragon Plus Environment
20

Page 21 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
concentration in the range of 10ꢀ200 mM (supporting information, Fig. S18), similar to that
previously observed for aniline at pH 4.5.7 The obtained k2,app value of 25.7 Mꢀ1hꢀ1 at 100 mM
PDA corresponds to a 250ꢀfold rate enhancement over the uncatalyzed reaction at pH 7. At very
high PDA concentrations (≥ 200 mM), however, the reaction equilibrium for product formation
was shifted slightly, indicating that PDA no longer behaved as a true catalyst under these
conditions.
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
Finally, the optimized conditions for oxime conjugation were applied to the synthesis of a
glycoconjugate of a complex and chemically sensitive oligosaccharide. Here, we utilized the
exceptionally mild conditions, pH 7, enabled by the PDA catalyst. We recently reported on the
regulatory role of exopolysaccharide (EPS) fragments, in the form of bioactive
18
octasaccharides.15,
Through covalent conjugation of isolated EPS octasaccharides to
biosensors,19 it was shown that the exopolysaccharide transmembrane receptor kinase (EPR3)
from the plant Lotus japonicus binds bacterial EPS from the soil bacterium Mesorhizobium loti
directly, and disruption of this perception causes a suppression of infection thread development,
which in turn inhibits formation of infected root nodules and nitrogen fixation. Direct
carbohydrateꢀprotein interaction between EPS octasaccharide and EPR3, as well as kinetic
measurements, were ultimately determined using biolayer interferometry (BLI). The purified
EPS octasaccharide18 contains two uronic acid units and is decorated with a nonꢀstoichiometric
acetylation pattern, which is believed to be important for correct recognition by the EPR3
receptor. As the complex carbohydrate was available only in minute quantities, we performed
these experiments with a nucleophile/carbohydrate ratio of 2:1 in the absence or presence of 50
mM aniline catalyst. The octasaccharide was coupled via oxime formation to a heteroꢀ
bifunctional OEG linker (supporting information, Fig. S59),19, 15 which we have used previously
ACS Paragon Plus Environment
21

Bioconjugate Chemistry
Page 22 of 45
1
2
3
4
5
6
7
8
9
for further conjugation reactions after unmasking of a trityl (Trt)ꢀprotected thiol group.19 The
oxime reactions were performed with PDA at pH 7, and with PANꢀC at pH 4.5, with aniline as a
control. Ample formation of the desired oxime product (Fig. 8a) was observed with PANꢀC
relative to aniline at pH 4.5 (Fig 8b). Remarkably, the PDAꢀcatalyzed reaction at pH 7
approached the rate of the anilineꢀcatalyzed reaction at pH 4.5. Near quantitative conversion
could be reached after 2 days at pH 7 at room temperature using PDA as catalyst under the
experimental conditions (Fig. 8b).
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
pH 4.5
pH 7.0
100
80
60
40
20
0
100
80
60
40
20
0
0
6
12 18 24 30 36 42
Time (h)
0
6
12 18 24 30 36 42
Time (h)
Figure 8. Oxime formation with EPS octasaccharide. Structure of resulting glycoconjugate
oxime (a). HPLC data showing oxime formation at pH 4.5 (b), or pH 7.0 (b) in the absence of
catalyst (), or catalyzed by aniline (ꢀ), PANꢀC (ꢀ), or PDA (ꢁ). Trt= trityl.
ACS Paragon Plus Environment
22

Page 23 of 45
Bioconjugate Chemistry
1
2
3
4
Conclusions
5
6
7
8
In summary, we have conducted a systematic study of glycoconjugate formation with different
nucleophiles, carbohydrate electrophiles, and aniline catalyst, and with variation of the pH of the
reactions. By application of secondꢀorder kinetic data treatment for equilibrium reactions, rate
constants for the individual reactions, which could be compared across all of the different
reaction conditions, were obtained. Primary aminooxy nucleophiles, which provide
glycoconjugate oximes, were clearly superior to all other nucleophiles, both in terms of reaction
rates and equilibrium constants for product formation.
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
For glycoconjugate oxime formation, five catalysts, including aniline, pꢀanisidine (PAN), 5ꢀ
methoxyanthranilic acid (PANꢀC), 3,5ꢀdiaminobenzoic acid (MDAꢀC), and 1,4ꢀdiaminobenzene
(PDA), were meticulously compared at pH 4ꢀ7. Surprisingly, MDAꢀC did not show any
significant improvement over aniline across pH 4ꢀ7 for carbohydrate oxime formation. PANꢀC
was generally a superior catalyst at pH 4ꢀ5, but at pH 7 it was in many cases not better than
PAN, indicating a decline in the effect of the orthoꢀcarboxy group with increasing pH for
carbohydrate electrophiles.
PDA was superior to all other catalysts at pH 7 for all tested carbohydrates. At elevated PDA
concentrations, rate enhancements of up to 250ꢀfold over the uncatalyzed reaction could be
13
reached at pH 7. Based on the experimental data, as well as mechanistic studies by CꢀNMR
spectroscopy and DFT calculations, we propose that switching of protonation state for the distal
amino group of covalent PDAꢀcarbohydrate intermediates provides a catalytic advantage by
altering the electronic substituent effects along the reaction pathway.
In a demonstration of the potential of the optimized conditions, they were applied for oxime
coupling of a bioactive octasaccharide, derived from exopolysaccharides of nitrogenꢀfixing
ACS Paragon Plus Environment
23

Bioconjugate Chemistry
Page 24 of 45
1
2
3
4
5
6
7
8
bacteria, at pH 7. Importantly, the PDAꢀcatalyzed reaction at pH 7 approached that of an anilineꢀ
catalyzed control reaction at pH 4.5. Near quantitative conversion could be reached after 2 days
at pH 7 at room temperature using PDA as catalyst with only 5 mM octasaccharide.
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
Experimental Procedures
Materials and methods. All chemicals were purchased from SigmaꢀAldrich (Denmark) and
used as received. MilliQ water was used for aqueous preparations. HPLC measurements were
carried out on a Dionex UltiMate 3000 HPLC system, using a C18 column (FEF Chemicals
Daiso (ODDMS) 10 µm, 200 Å, 250 × 20 mm or Phenomenex Gemini 5 µm, 110 Å, 50 × 4.6
mm). Samples were eluted by applying a gradient of CH3CN in H2O containing 0.1 % HCOOH
or TFA with a flow rate of 1 mL/min for 15 minutes. Highꢀresolution mass spectrometric (HRꢀ
MS) analyses were carried out on a Dionex UltiMate 3000 UHPLC+ focused system equipped
with a Bruker Impact HD UHRꢀQqTOF mass spectrometer. NMR spectra were recorded on a
Bruker Avance 300 MHz equipped with a BBO probe or a Bruker 500 MHz instrument equipped
with a cryoprobe at 300 K. Chemical shift values were referenced to the residual solvent signal.
Assignments were aided by H,H COSY, gHSQC, and APT experiments. Data analysis, fitting
and plotting was performed using GraphPad software (GraphPad Prism version 6.07 for
Windows, San Diego, CA, USA).
O-Benzyl-N-methylhydroxylamine. The compound, previously described by Gudmundsdottir
et al.,32 was synthesized by a modified procedure. Briefly, Oꢀbenzylhydroxylamine
hydrochloride (2.0 g, 12.5 mmol) was dissolved in 1 M NaOAc (50 mL), and pH adjusted to 4.6.
Formaldehyde (37 % w/w in H2O, 2.8 mL, 37.6 mmol) was added, and the solution was stirred
for 24 h at 25 °C. The solution was extracted twice with CH2Cl2 (50 mL), and the combined
ACS Paragon Plus Environment
24

Page 25 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
organic phases were washed with brine (50 mL), dried over Na2SO4 and concentrated under
reduced pressure. The resulting liquid was dissolved in glacial acetic acid (50 mL), and
NaBH3CN (1.77 g, 28.2 mmol) was added. The solution was stirred for 1 h at 25 °C, and acetic
acid was removed under reduced pressure. Sat. NaHCO3 (50 mL) was added, and the solution
was extracted twice with CH2Cl2 (50 mL), and the combined organic phases were then extracted
twice with 0.1 M HCl (50 mL). The combined aqueous phases were washed 3 times with CH2Cl2
(50 mL), basified by dropwise addition of 4 M NaOH and extracted twice with CH2Cl2 (50 mL).
The combined organic phases were dried over Na2SO4 and carefully concentrated under reduced
pressure (note that the product as the free base is volatile) to give the pure product (1.11 g, 65%)
as a colorless liquid. The analytical data were in agreement with previously published data.32ꢀ33
N-Benzyl-O-methylhydroxylamine. The compound, previously described by Gantt et al.,34 was
synthesized by a modified procedure. Briefly, methoxyamine hydrochloride (5.0 g, 59.9 mmol)
and benzaldehyde (4.36 mL, 42.8 mmol) were dissolved in 60 mL CH2Cl2, and pyridine (7.58
mL, 94.1 mmol) was added. The solution was stirred overnight at 25 °C. 200 mL CH2Cl2 was
added, and the organic phase was washed with 1 M HCl (100 mL) and brine (100 mL), dried
over Na2SO4, and concentrated under reduced pressure. The resulting liquid was mixed with
glacial acetic acid (30 mL), and NaBH3CN (5.37 g, 85.5 mmol) was added. The solution was
stirred for 1 h at 25 °C, and acetic acid was removed under reduced pressure. The residue was
dissolved in 0.05 M HCl (50 mL) and washed 3 times with CH2Cl2 (50 mL). The aqueous
solution was basified (pH > 7.0) by addition of sat. Na2CO3, and extracted three times with
CH2Cl2 (50 mL). The combined organic phases were washed with sat. Na2CO3 (100 mL) and
brine (100 mL), dried over Na2SO4, and carefully concentrated under reduced pressure (note that
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
25

Bioconjugate Chemistry
Page 26 of 45
1
2
3
4
5
6
the product as the free base is volatile) to give the pure product (2.36 g, 40%) as a colorless
liquid. The analytical data were in agreement with previously published data.34
7
8
D
-Glucose O-benzyloxime (1). The analytical data were in agreement with previously published
9
data.7 HRꢀMS (ESIꢀTOF): calcd. for C13H19NO6, [M+H]+: m/z 286.1285; found: 286.1285.
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
2-Acetamido-2-deoxy-
D
-glucose O-benzyloxime (2). HRꢀMS (ESIꢀTOF): calcd. for
C15H22N2O6, [M+H]+: m/z 327.1551; found: 327.1552. H NMR (300 MHz, CD3OD, (E)ꢀ
oxime/(Z)ꢀoxime/βꢀpyranosyloxyamine tautomeric ratio 7:2:1): δ 7.50 (d, J=5.9 Hz, 0.7H; (E)ꢀ
oxime Hꢀ1), 7.42ꢀ7.23 (m, 5H; Ph), 6.76 (d, J=6.3 Hz, 0.2H; (Z)ꢀoxime Hꢀ1), 5.17 (dd, J=5.8 and
6.0 Hz, 0.2H; (Z)ꢀoxime Hꢀ2), 5.12 (s, 0.4H; (Z)ꢀoxime CH2Ph), 5.05 (s, 1.4H; (E)ꢀoxime
CH2Ph), 4.76ꢀ4.62 (m, 0.8H; (E)ꢀoxime Hꢀ2 + βꢀpyranosyloxyamine CH2Ph), 4.24 (d, J=9.7 Hz,
0.1H; βꢀpyranosyloxyamine Hꢀ1), 4.12 (dd, J=2.6 and 5.5 Hz, 0.2H; (Z)ꢀoxime Hꢀ3), 4.02 (dd,
J=2.0 and 6.7 Hz, 0.7H; (E)ꢀoxime Hꢀ3), 3.88ꢀ3.25 (m, 4.2H), 1.99ꢀ1.95 (3s, 3H; NAc). 13C
NMR (75 MHz, CD3OD, (E)ꢀoxime/(Z)ꢀoxime/βꢀpyranosyloxyamine tautomeric ratio 7:2:1): δ
173.3 (C=O), 151.4 ((Z)ꢀoxime Cꢀ1), 150.0 ((E)ꢀoxime Cꢀ1), 139.1 (Ph ipso), 130.1 (Ph), 129.4
(Ph), 129.3 (Ph), 129.3 (Ph), 129.1 (Ph), 128.8 (Ph), 91.4 (βꢀpyranosyloxyamine Cꢀ1), 79.3, 78.1
(βꢀpyranosyloxyamine CH2Ph), 77.2 ((Z)ꢀoxime CH2Ph), 77.0, 76.9 ((E)ꢀoxime CH2Ph), 73.4,
73.0, 72.9, 72.2, 71.9, 71.4, 70.6, 64.9, 64.8, 63.0, 53.7 (βꢀpyranosyloxyamine Cꢀ2), 53.5 ((E)ꢀ
oxime Cꢀ2), 50.6 ((Z)ꢀoxime Cꢀ2), 23.0 (βꢀpyranosyloxyamine NAc), 22.7 ((E)ꢀoxime NAc),
22.6 ((Z)ꢀoxime NAc).
1
6-Oxo-D
-Glucose O-benzyloxime (3). HRꢀMS (ESIꢀTOF): calcd. for C13H17NO7, [M+H]+: m/z
300.1078; found: 300.1077. 1H NMR (300 MHz, CD3OD, (E)ꢀoxime with trace (Z)ꢀoxime (<5%,
not assigned)): δ 7.56 (d, J=5.1 Hz, 1H; Hꢀ1), 7.37ꢀ7.25 (m, 5H; Ph), 5.09 (s, 2H; CH2Ph), 4.59
(dd, J=5.1 and 8.1 Hz, 1H; Hꢀ2), 4.50 (d, J=4.6 Hz, 1H; Hꢀ5), 4.43 (dd, J=2.9 and 8.1 Hz, 1H; Hꢀ
ACS Paragon Plus Environment
26

Page 27 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
3), 3.35 (dd, J=2.9 Hz and 4.6 Hz, 1H; Hꢀ4). 13C NMR (75 MHz, CD3OD, (E)ꢀoxime with trace
(Z)ꢀoxime (<5%, not assigned)) δ 177.7 (C=O), 149.5 (Cꢀ1), 139.0 (Ph ipso), 129.4 (Ph), 129.3
(Ph), 128.9 (Ph), 82.6 (Cꢀ3), 77.1 (CH2Ph), 72.1 (Cꢀ5), 70.9 (Cꢀ4), 68.8 (Cꢀ2).
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
O-Benzyl-N-(D-glucosyl)-N-(methyl)hydroxylamine (4). The analytical data were in agreement
with previously published data.32ꢀ33 HRꢀMS (ESIꢀTOF): calcd. for C14H21NO6, [M+H]+: m/z
300.1442; found: 300.1442.
O-Benzyl-N-(2-acetamido-2-deoxy-D-glucosyl)-N-(methyl)hydroxylamine (5). The analytical
data were in agreement with previously published data.32 HRꢀMS (ESIꢀTOF): calcd. for
C16H24N2O6, [M+H]+: m/z 341.1707; found: 341.1707.
O-Benzyl-N-(6-oxo-
D
-glucosyl)-N-(methyl)hydroxylamine (6). HRꢀMS (ESIꢀTOF): calcd. for
1
C14H19NO7, [M+H]+: m/z 314.1234; found: 314.1233. H NMR (300 MHz, CD3OD, >80% βꢀ
oxyamine, other forms not assigned): δ 7.43ꢀ7.24 (m, 5H; Ph), 4.77 (s, 2H; CH2Ph), 4.07 (d,
13
J=8.4 Hz, 1H; Hꢀ1), 3.61ꢀ3.36 (m, 4H; Hꢀ2, Hꢀ3, Hꢀ4, Hꢀ5), 2.74 (s, 3H; NCH3). C NMR (75
MHz, CD3OD, >80% βꢀoxyamine, other forms not assigned) δ 176.7 (C=O), 138.5 (Ph ipso),
130.3 (Ph), 129.3 (Ph), 129.1 (Ph), 95.4 (Cꢀ1), 79.0 (Cꢀ3), 77.9 (Cꢀ5), 76.4 (CH2Ph), 73.4 (Cꢀ4),
71.5 (Cꢀ2), 39.4 (NCH3).
N-Benzyl-N-(D-glucosyl)-O-(methyl)hydroxylamine (7). The analytical data were in agreement
with previously published data.35 HRꢀMS (ESIꢀTOF): calcd. for C14H21NO6, [M+H]+: m/z
300.1442; found: 300.1441.
N-Benzyl-N-(2-acetamido-2-deoxy-D-glucosyl)-O-(methyl)hydroxylamine (8). The analytical
data were in agreement with previously published data.23 HRꢀMS (ESIꢀTOF): calcd. for
C16H24N2O6, [M+H]+: m/z 341.1707; found: 341.1705.
ACS Paragon Plus Environment
27

Bioconjugate Chemistry
Page 28 of 45
1
2
3
4
N-Benzyl-N-(6-oxo-
D
-glucosyl)-O-(methyl)hydroxylamine (9). HRꢀMS (ESIꢀTOF): calcd. for
5
6
7
8
C14H19NO7, [M+H]+: m/z 314.1234; found: 314.1233. H NMR (300 MHz, CD3OD, >80% βꢀ
oxyamine, other forms not assigned): δ 7.49ꢀ7.22 (m, 5H; Ph), 4.22 and 4.07 (2d, J=12.8 Hz, 2H;
CH2Ph), 4.01 (d, J=8.9 Hz, 1H; Hꢀ1), 3.75 (t, J=8.9 Hz, 1H; Hꢀ2), 3.54 (d, J=9.2 Hz, 1H; Hꢀ5),
3.46 (t, J=8.9 Hz, 1H; Hꢀ4), 3.40 (s, 3H; OMe), 3.37 (t, partly overlapped, J=8.8 Hz, 1H; Hꢀ3).
13C NMR (75 MHz, CD3OD, >80% βꢀoxyamine, other forms not assigned) δ 76.2 (C=O), 138.7
(Ph ipso), 131.1 (Ph), 129.2 (Ph), 128.4 (Ph), 93.4 (Cꢀ1), 79.1 (Cꢀ3), 78.0 (Cꢀ5), 73.3 (Cꢀ4), 71.3
(Cꢀ2), 62.6 (OMe), 57.6 (CH2Ph).
1
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
N-Benzoyl-N’-(D-glucosyl)hydrazine (10). The analytical data were in agreement with
previously published data.36 HRꢀMS (ESIꢀTOF): calcd. for C13H18N2O6, [M+H]+: m/z 299.1238;
found: 299.1239.
N-(D-Glucosyl)-N’-(4-methylbenzenesulfonyl)hydrazine (11). The analytical data were in
agreement with previously published data.32 HRꢀMS (ESIꢀTOF): calcd. for C13H20N2O7S,
[M+H]+: m/z 349.1064; found: 349.1064.
Exopolysaccharide octaose conjugation. A purified exopolysaccharideꢀderived octaose (EPS
octasaccharide)18, 37 from the soil bacterium Mesorhizobium loti was used to assess the feasibility
of using 1,4ꢀdiaminobenzene as a catalyst of oxime formation at pH 7.0. A heteroꢀbifunctional
aminooxyꢀtetra(ethylene glycol)ꢀthiol linker (OEG linker)19 was used as the nucleophile.
Separate solutions of the EPS octasaccharide (2.2 mg, 1.5 ꢀmol), and the OEG linker (1.4 mg,
3.0 ꢀmol) were prepared in MeCNꢀH2O (v/v 1:1 ratio) at 20 mM and 40 mM, respectively.
These solutions were combined (v/v 1:1 ratio) and diluted with buffer solution (100 mM sodium
acetate (pH 4.5), or 100 mM NaH2PO4 (pH 7.0), both MeCNꢀH2O (v/v 1:1 ratio)) to produce a
final concentration of 5 mM EPS octasaccharide, 10 mM OEG linker, 50 mM buffer, MeCNꢀ
ACS Paragon Plus Environment
28

Page 29 of 45
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
H2O (v/v 1:1 ratio). For catalyzed reactions, the buffers were prepared with the further addition
of 100 mM aniline, PANꢀC, or PDA catalyst. Final concentrations were thus 5 mM EPS
octasaccharide, 10 mM OEG linker, 50 mM buffer, 50 mM catalyst, MeCNꢀH2O (v/v 1:1 ratio).
Product formation was monitored by HPLC. Samples were injected 15 minutes after mixing of
reactants, and every 2 hours thereafter for 26ꢀ58 hours depending on the reaction rate. The
conversion was determined from the absorbance at 254 nm by relating twice (due to the 2:1 ratio
of nucleophile to EPS octasaccharide) the integrated absorbance of the product to the sum of the
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
integrated
absorbance
of
product
and
OEG
linker
(i.e.,
conversion
=
2*Absproduct/(Absproduct+Abslinker)). The oxime conjugate product had HRꢀMS (ESIꢀTOF),
triacetylated oxime conjugate: calcd. for C80H113NO48S, [M+H]+: m/z 1888.6226; found:
1888.6297, and calcd. for C61H99NO48S, [Mꢀtrityl+2H]+: m/z 1646.5130; found: 1646.5084. HRꢀ
MS (ESIꢀTOF), diacetylated oxime conjugate: calcd. for C78H111NO47S, [M+H]+: m/z
1846.6120; found: 1846.6165, and calcd. for C59H97NO47S, [Mꢀtrityl+2H]+: m/z 1604.5024;
found: 1604.5013.
Kinetic HPLC Measurements. The rate of glycoconjugate formation was determined by HPLC,
by calculating the integrated absorption of the product peak relative to the total integrated
absorption of nucleophile and product peaks at 215 nm. As the extinction coefficient at 215 nm
of the products differed from that of the corresponding nucleophiles, the absorption integrals
were corrected to match the absorption at 254 nm. Generally, a volume of 10 µL was injected
every 30 minutes for 13 hours. Solutions containing αꢀnucleophile, carbohydrate electrophile
and aniline derivative catalyst were prepared individually in 50 mM buffer solutions. The pH and
volume of each solution was adjusted individually to reach the desired conditions immediately
upon mixing. Reactions were started, after prior equilibration of the carbohydrate solution for at
ACS Paragon Plus Environment
29