Preparation of glycoconjugates by dialkyl squarate chemistry revisited

Article abstract of DOI:10.1016/j.carres.2007.10.015

The methyl 6-hydroxyhexanoyl glycoside of lactose was treated with each of 1,2-diaminoethane or hydrazine hydrate, and the corresponding amino amide 4 and acyl hydrazide 13, were treated with each of squaric acid dimethyl, diethyl, dibutyl, and didecyl esters. The monoesters were conjugated to bovine serum albumin (BSA) at different concentrations of hapten using 0.05 and 0.5 M pH 9 borate buffer. Maximum loading was achieved faster, and the conjugation efficiency was higher, when the conjugation was conducted at higher concentrations of both hapten and buffer. Conjugations involving haptens 14-17 prepared from hydrazide 13 were generally slower and less efficient than those with compounds 5-8, which were made from amino amide 4. Maintaining pH 9 during conjugation was found to be the most important factor in ensuring that the conjugation was a fast, highly efficient, and reproducible process. When the pH of the conjugation mixture fell during the reaction, resulting in decreased reaction rate or even termination of the conjugation process, the normal course of the conjugation process could be restored by addition of buffer salts. Hydrolysis studies with monoesters formed from amino amide 4 under conjugation conditions showed that decyl ester 8 was the most stable and that the methyl compound 5 was the one most readily hydrolyzed. The stability of monoesters prepared from hydrazide 13 was similar and comparable to the decyl ester prepared from 4. No definite advantage was found for the use of any of the four dialkyl squarate reagents (methyl-, ethyl-, butyl-, and decyl-) for conversion of carbohydrate derivatives to species amenable for conjugation. Nevertheless, dimethyl squarate seemed to be the most convenient reagent because it is a crystalline, easy to handle, and commercially available material with very good reactivity.

Full text of DOI:10.1016/j.carres.2007.10.015

Available online at www.sciencedirect.com
Carbohydrate Research 343 (2008) 196–210
Preparation of glycoconjugates by dialkyl squarate
chemistry revisitedI
´ˇ*
Shu-jie Hou, Rina Saksena and Pavol Kovac
NIDDK, LBC, National Institutes of Health, Bethesda, MD 20892-0815, USA
Received 8 August 2007; received in revised form 12 October 2007; accepted 22 October 2007
Available online 26 October 2007
Abstract—The methyl 6-hydroxyhexanoyl glycoside of lactose was treated with each of 1,2-diaminoethane or hydrazine hydrate,
and the corresponding amino amide 4 and acyl hydrazide 13, were treated with each of squaric acid dimethyl, diethyl, dibutyl,
and didecyl esters. The monoesters were conjugated to bovine serum albumin (BSA) at different concentrations of hapten using
0.05 and 0.5 M pH 9 borate buffer. Maximum loading was achieved faster, and the conjugation efficiency was higher, when the con-
jugation was conducted at higher concentrations of both hapten and buffer. Conjugations involving haptens 14–17 prepared from
hydrazide 13 were generally slower and less efficient than those with compounds 5–8, which were made from amino amide 4. Main-
taining pH 9 during conjugation was found to be the most important factor in ensuring that the conjugation was a fast, highly effi-
cient, and reproducible process. When the pH of the conjugation mixture fell during the reaction, resulting in decreased reaction rate
or even termination of the conjugation process, the normal course of the conjugation process could be restored by addition of buffer
salts. Hydrolysis studies with monoesters formed from amino amide 4 under conjugation conditions showed that decyl ester 8 was
the most stable and that the methyl compound 5 was the one most readily hydrolyzed. The stability of monoesters prepared from
hydrazide 13 was similar and comparable to the decyl ester prepared from 4. No definite advantage was found for the use of any of
the four dialkyl squarate reagents (methyl-, ethyl-, butyl-, and decyl-) for conversion of carbohydrate derivatives to species amenable
for conjugation. Nevertheless, dimethyl squarate seemed to be the most convenient reagent because it is a crystalline, easy to handle,
and commercially available material with very good reactivity.
Published by Elsevier Ltd.
Keywords: Conjugate; Neoglycoconjugate; Squaric acid diester; 5-Methoxycarbonyl-b-lactoside; SELDI TOF-MS
1. Introduction
to develop new, highly efficacious methods, or improve
reaction efficiency of existing methods for conjugation,
so that the labor-intensive, synthetic oligosaccharides
can be used more efficiently. Earlier work involving
squaric acid diesters was done with dimethyl8,9 and
diethyl esters,5 and more recently Bergh et al.10
described the use of didecyl squarate. The popularity
of this class of reagents in synthesizing glycoconjugates
increased after Kamath et al.11 reported conjugation
of oligosaccharides to protein on a very small scale using
squaric acid amide ethyl esters, and showed that the car-
bohydrate–protein ratio (loading) in glycoconjugates
thus formed could be objectively established by MAL-
DI-TOF mass spectrometry.
Synthetic oligosaccharides and their conjugates to other
molecules are indispensable probes in the life sciences.1–3
One of the methods that allow conjugation by a single-
point attachment4 to form well-defined neoglycoconju-
gates without cross-linking is based on squaric acid
chemistry.5–7 In general, conjugations are often carried
out with a large excess of synthetic oligosaccharides to
ensure reasonable reaction rates. Thus, there is a need
q Presented as a plenary presentation at the 7th International Meeting
of the Portuguese Carbohydrate Group (GLUPOR 7), Oieras,
Portugal, September 12–15, 2007.
Corresponding author. Tel.: +1 301 496 3569; fax: +1 301 480
5703; e-mail: kpn@helix.nih.gov
We have studied this conjugation method in detail,12
and have also developed a protocol that allows the
*
0008-6215/$ - see front matter Published by Elsevier Ltd.
doi:10.1016/j.carres.2007.10.015

S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
197
one-pot preparation of a series of neoglycoconjugates
with predetermined carbohydrate–protein ratios.13 An
attractive feature of conjugation by squaric acid chemis-
try is that the excess hapten used at the onset of the con-
jugation that remained unchanged during the reaction
can be recovered.10,14–16 Izumi et al.17 compared the effi-
ciency of conjugation of the ethyl squarate derivative of
mannose to BSA with that by six other, unrelated
reagents and found the protocol involving squaric acid
chemistry as one of the three most potent methods.
The diethyl ester has been the most popular reagent,
although squaric acid methyl15,18 and decyl10 diesters
have also been used in squaric acid ester-mediated con-
jugations, and dibutyl squarate is commercially avail-
able. There is no obvious rational for this practice, as
no systematic study has been carried out to establish
advantage of that reagent to form neoglycoconjugates.
Based on previous observations,9,19 Kamath et al.
suggested11 that it should be possible to prepare squaric
acid derivatives of saccharides from hydrazides and, in-
deed, linker-equipped carbohydrates have been conju-
gated to proteins in this way.15,18 We have not found20
a significant difference in immunogenicity of conjugates
from the hexasaccharide fragment of Vibrio cholerae
O:1, serotype Ogawa that was made by squaric acid
chemistry involving either 1,2-diaminoethane or hydra-
zine hydrate. This finding, showing that both types of
squaric acid derivatives can be useful in conjugate vac-
cine development, creates a need to evaluate their conju-
gation properties in more detail.
Data on hydrolysis of some individual squaric acid
diesters are available,6,8 but a stability study with amide
esters derived from carbohydrates under the conditions
of conjugation to proteins has not been undertaken.
Clearly, there is a need for systematic studies in these
areas to establish general guidelines for efficient conju-
gation of carbohydrates to proteins by this method.
The objective of this work was to evaluate the stability
of amide monoesters made from amine amides (com-
pounds 5–8) and hydrazides (compounds 14–17) under
identical conditions and to determine the efficiency of
conjugation to a model carrier (BSA) to give conjugates
9–12 and 18–21 (Scheme 1). Aside from obvious rea-
sons, information about the rate of hydrolysis of the
squaric acid monoesters at pH 9 is important, for exam-
ple, when poorly reactive carriers are being conjugated,
as it can indicate the time at which point continuing con-
jugation becomes futile because virtually no active ester
would be present in the reaction mixture.
jugations to BSA performed in different laboratories
cannot be readily compared because the commercial
source and history of isolation and purification of the
carrier are often not disclosed and, as we show here, this
can have a profound effect on the outcome of conjuga-
tion. A comparison of the outcome of earlier conjuga-
tions with more recent ones can be misleading because
in the more distant past, carbohydrate–protein ratios
in neoglycoconjugates were determined by various col-
orimetric methods and not by mass spectrometry. The
reliability of calibration curves required by colorimetry
is questionable when they are to be used to analyze mix-
tures of oligosaccharides that mimic the structure of rare
or unstable sugar-containing bacterial polysaccharides
with variable amounts of protein carriers. In addition
to comparing rates and efficiency of conjugation of
squaric acid derivatives made from the amino amide 4
and hydrazide 13, compounds 5–8 versus 14–17, we have
evaluated the importance of concentration and excess of
hapten as well as of the buffer concentration upon the
conjugation reaction.
2. Results and discussion
b-Lactose octaacetate (1) was converted (Scheme 1),
through 5-methoxycarbonylpentyl b-lactoside (3), to
squaric acid monoesters 5–8 and 14–17 and conjugated
to a model carrier bovine serum albumin (BSA). Com-
pound 3 was prepared by boron trifluoride etherate-
mediated glycosidation of 121,22 with methyl 6-hydroxy-
hexanoate (24),23 to give b-glycoside 2 in 60–70% yield.
In addition, small amounts of the 2I-hydroxy derivative
(27), as well as 6-O-acetyl derivative of 24, compound
25, were isolated as side products. Analogous com-
pounds were isolated from stannic tetrachloride cata-
lyzed glycosylation of 8-ethoxycarbonyloctanol with a
number of oligosaccharides.24 Formation of such by
products can be expected based on the mechanism25–28
of glycosylation with donors having a participating
group at position 2 vicinal to the anomeric position
´
involved in the glycosidation. Deacetylation (Zemplen)
RO(CH2)5COOCH3
R
H
24
Ac
25
OR1
OR1
R1O
R1O
O
O
Numerous data on conjugation from different labora-
tories are available but not readily comparable because
it appears that no two laboratories have conjugated
saccharides to proteins under identical, or at least very
similar, conditions (the same linker and carrier,
concentration of reactants and buffers, and initial carbo-
hydrate–protein ratio). Furthermore, the results of con-
O
R1O
OR1
RO
O(CH2)5COOCH3
R
Ac
H
R1
Ac
Ac
H
26
27
28
H

198
S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
OAc
OAc
AcO
AcO
O
O
OAc
O
AcO
OAc
OAc
1
OAc
OH
OAc
OH
AcO
AcO
O
O
O(CH2)5COOCH3
O(CH2)5COOCH3
O
AcO
OAc
OAc
2
HO
HO
O
O
O
HO
OH
OH
3
OH
OH
OH
HO
OH
HO
HO
O
O
O
O
O(CH2)5CONH(CH2)2NHR
O
O(CH2)5CONHNHR
O
HO
HO
HO
OH
OH
OH
R
OH
R
H
H
4
13
R
R
R
O
O
O
O
O
O
O
O
O
O
OH
OMe
NH
OMe
O
BSA
BSA
BSA
BSA
BSA
BSA
BSA
BSA
22
18.0; 16.3; 7.7; 15.5;
4.8; 5.3; 45.9; 55.6
5
9
14
18 NH
14.2; 9.3; 8.5
R
O
O
O
O
O
O
O
O
O
O
O
OEt
O
OH
OEt
O
10NH
15.6
13.5
14.5
NH
17.4; 16.3,
6
7
8
23
19
15
16
O
O
O
O
OBu
O
NH
OBu
O
NH
19.0; 15.6
11
20
O
O
O
O
O
O
ODec
NH
ODec
NH
18.1; 16.5; 16.7
12
17
21
Scheme 1.
of 2 gave crystalline, linker-equipped lactoside 3, which,
in turn, was converted to the corresponding amino
amide 4 and hydrazide 13 by reaction with 1,2-diamino-
ethane and hydrazine hydrate, respectively. Each of
products 4 and 13 was then treated with dimethyl,
diethyl, dibutyl, and didecyl10 squarates to give mono-
alkyl squarates (squarate monoesters) 5–8 and 14–17.
The conversion of 3 with 1,2-diaminoethane was much
slower than that with hydrazine hydrate. To be able to
follow the hydrolysis of squaric acid monoesters confi-
dently by HPLC, acids 22 and 23 were also prepared.
Conjugation of carbohydrates to proteins by squaric
acid chemistry with different squaric acid reagents has
been carried out at different, largely arbitrarily chosen
conditions, namely, different concentration of hapten
and buffer and initial hapten–carrier ratio. We have
observed some irregularities in the outcome of our pre-
vious conjugations, such as occasional poor reproduc-
ibility,29 faster conjugation of a higher compared to a
lower oligosaccharide13,30 or very different reaction rates
of conjugation of the same size of oligosaccharides.31
We could offer no other explanation for these irregular-
ities except that, perhaps, conjugation was affected by
factors of which we were unaware.
Our past conjugations were performed using 0.05 M
borate buffer, and the same buffer was used for the first
conjugation experiment in this series (see Section 3.8.1).
We have previously shown12 that conjugation efficiency
increased with hapten concentrations between 5 and
25 mM. We hypothesized that using still higher concen-
trations might further increase the rate and efficiency
and, therefore, the first conjugation with methyl squa-
rate 5 was performed at a hapten concentration of
40 mM in 0.05 M borate buffer. The reaction was car-
ried out at a molar ratio 5/BSA = 20 and the progress
of the conjugation was monitored by SELDI TOF-
MS. This showed that only a small fraction of the hap-
ten used at the onset of the conjugation was attached to

S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
199
Table 1. Conjugation of squaric acid monoesters 5–8 made from amino amide 4 and monoesters 14–17 made from hydrazide 13 to BSAa
Hapten/conjugateb
Concentration hapten
(mM)/buffer (M)
Reaction
time (h)
Hapten/BSA
(M/M (n)b)
Yieldc
(%)
Conjugation
efficiency
5/9
40/0.5
1
2
3
6
8
8.9
13.5
15.4
17.9
18.0
91
92
89
90
82
39
5/9
5/9
5/9
4/0.5
1
3
6
8
24
4.0
8.6
12.0
13.6
16.3
0.4/0.5
4/0.05
1
3
8
24
48
1.5
3.2
3.8
6.8
7.7
1
3
3.0
6.4
6
9.5
7
8
27
48
10.1
10.5
14.3
15.5
91
78
6/10
40/0.5
1
2
3
6
7
7.8
10.3
12.3
15.3
15.7
16.0
17.4
8
24
86
92
88
93
66
93
87
82
95
78
91
83
6/10
7/11
7/11
8/12
8/12
4/0.5
40/0.5
4/0.5
40/0.5
4/0.5
1
3
8
24
48
72
2.4
4.7
8.4
13.2
15.5
16.3
1
2
3
7
8
7.4
10.2
12.3
15.2
16.1
19.0
24
1
3
6
8
24
48
2.4
4.9
8.1
9.2
13.9
15.6
1
2
3
6
8
8.8
10.5
13.2
14.4
15.2
17.6
24
1
2
3
6
8
9.7
12.3
13.6
15.5
16.5
16.5
24
(continued on next page)

200
S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
Table 1 (continued)
Hapten/conjugateb
Concentration hapten
(mM)/buffer (M)
Reaction
time (h)
Hapten/BSA
(M/M (n)b)
Yieldc
(%)
Conjugation
efficiency
8/12
0.4/0.5
1
2
3
6
8
8.5
11.0
11.1
14.1
16.1
16.7
24
91
90
84
71
14/18
14/18
40/0.5
4/0.5
1
3
8
24
48
3.1
5.7
8.2
12.5
14.2
1
3
6
0.6
1.2
1.6
2.5
4.7
7.1
8.5
9.3
8
24
48
96
120
92
47
14/18
4/0.05
1
3
8
24
48
0.6
1.5
2.8
4.3
5.6
7.5
7.4
8.0
8.5
96
120
144
168
92
43
15/19
40/0.5
1
2
0.6
2.0
3
2.1
6
3.9
8
4.6
24
48
72
120
9.2
12.2
13.9
15.6
80
85
91
78
68
73
16/20
40/0.5
1
2
3
6
8
0.6
1.3
2.1
4.2
5.4
24
72
10.8
13.5
17/21
40/0.5
1
2
3
2.9
4.3
5.3
7
7.3
24
48
72
11.5
13.9
14.5
a All conjugations were carried out at hapten–BSA ratio of 20:1 and ambient temperature (22–24 ꢀC) at pH 9 (Buffer B) using BSA (Sigma Cat.
Number A-0281) as carrier. The final time noted is the one when the hapten–carrier ratio leveled off; Progress of conjugation was monitored by
SELDI TOF-MS.
b See Scheme 1.
c The amount of material used up for analysis during monitoring of the conjugation has been accounted for.
BSA (see Section 3.8.1), and that the reaction virtually
stopped after 4 h. A check with Hydrion MicroFine
pH Paper showed the pH of the mixture to be
7.2 0.2. Considering the previous16 results of stability,

S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
201
which we largely confirmed (see Figs. 3 and 4), and the
loading achieved, most of the hapten used at the onset of
the reaction must have been unchanged after 25 h. Thus,
the lowered pH of the mixture, which caused the termi-
nation of conjugation, could not be due to the small
amount of acid that could have been formed as a result
of ester hydrolysis. When a small amount of buffer salts
was added, the pH raised to ꢀ8.8 0.2, and after addi-
tional 12 h the carbohydrate–protein ratio was 18:1, as
determined by SELDI TOF-MS analysis. When the
reaction was repeated using 0.5 M buffer, the latter load-
ing was reached after 8 h (Table 1). This confirmed the
hypothesis formulated above, that is, that increasing
the concentration of the hapten increases the rate and
efficiency of conjugations by squaric acid chemistry.
The cause of the observed lowering of the pH of the
conjugation mixture was found when a sample of BSA
used in the preliminary experiment (533 mg, Sigma
Cat. No. A-4503, purified32) or another sample of com-
mercially available BSA (Sigma Cat. Number A-0281)
was dissolved in pH 9 buffer (0.05 M, 4 mL, to mimic
solution concentration of the conjugation described
above); the pH of the solutions was found to be 7.46
and 8.0, respectively. On the other hand, the pH of solu-
tions of these proteins prepared in the same way in
0.5 M buffer was lowered from 9.0 by only ꢀ0.2 pH
units. It appears that when highly concentrated solutions
of BSA are prepared in low concentration buffers, the pH
of these solutions is lowered depending on the mode of
preparation and purification of the protein. The same,
with similar consequences concerning conjugation,
may be the case with other carriers.
the conjugation in more concentrated, higher capacity,
0.5 M buffer. The use of still more concentrated buffer
(1 M; this experiment is not described in the Experimen-
tal) did not provide additional benefit for the outcome
of conjugation. Concentration of the buffer and its
capacity, an important variable in application of squaric
acid chemistry, has not been duly emphasized, and
information that would make it possible to identify tech-
nical specifications of buffers is often not available from
the description of conjugation experiments. On the other
hand, the improved efficiency of the conjugations car-
ried out in buffers slightly more basic than pH 9 has
been noted,14,33,34 despite pH 9.0 being generally consid-
ered optimal for squaric acid chemistry conjugations.5,11
Because the lowered pH of the reaction media following
solubilization of carriers, or other concrete reasons for
the use of more basic buffers, had not been mentioned
in that context, the more basic media used for successful
conjugation in those situations must have serendipi-
Table 2. Reproducibility of conjugation to BSAa
Hapten
Reaction time (h)
Hapten/BSA
Run 1
Run 2
8b
1
2
4
6
8
24
8.8
10.5
13.2
14.4
15.2
17.6
8.2
10.4
12.5
14.4
14.5
18.1
30c
0.5
1
1.5
2
2.5
4
6
1.4
1.9
2.9
2.9
3.9
4.8
5.0
1.7
2.3
2.8
3.4
3.8
4.4
4.9
Unfortunately, it is not customary to report detailed
information concerning the nature and purification of
the BSA used, although this protein in its many, but sel-
dom specified forms, is the most commonly used carrier
in synthesizing experimental tools for the life sciences.
The above results prompted us to ensure that the pH
of reaction mixtures in further experiments persisted
close to 9.0 during the whole duration of the conjuga-
tion. This was most conveniently realized by conducting
a Unless stated otherwise, all conjugations were carried out at ambient
temperature (22–24 ꢀC) in 0.5 M pH 9 buffer, at a hapten concen-
tration of 40 mM.
b Initial hapten–BSA ratio, 20:1.
c Initial hapten–BSA ratio, 6:1.
Table 3. Conjugations of the di-(5), tri-(29), tetra-(30), and hexasaccharide (31) to BSA at different initial hapten–BSA ratiosa
Hapten/conjugate
Initial hapten/BSA
Reaction time (h)
Hapten/BSA in product
Yield (%)
Conjugation efficiency (%)
5/9
5/9
5/9
5/9
29/32
30/33
31/34
31/34
31/34
6:1
2
4
72
72
8
8
5
2.5
48
5.3
4.8
45.9
55.6
5.3
4.9
5.4
5.4
12
88
93
88
91
80
85
91
88
80
76.5
46
88
81.6
77
6:1b
60:1
120:1
6:1
6:1
7:1
15:1
15:1
88
NDc
Irrelevantc
80
a Conjugation carried out at 22–24 ꢀC and a hapten concentration of 40 mM in 0.5 M buffer.
b Hapten concentration, 4 mM.
c Two conjugates prepared in one pot (see text); ND: not determined.

202
S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
tously compensated for the acidity of the carrier, which
then resulted in conjugation proceeding normally.
Based on the observations described above, a series of
conjugations of squarates 5–8 and 14–17 to BSA were
carried out at a hapten concentration of 40 mM, buffer
concentration of 0.5 and 0.05 M, and an initial carbo-
hydrate protein ratio of 20:1 (Tables 1–3, Figs. 1–3).
Because conjugations have also been carried out success-
fully at much lower concentrations of hapten,7,10,17,35,36
we also performed some conjugations at hapten concen-
trations of 4 and 0.4 mM (Table 1, Fig. 2). We ensured
that the pH of the reaction medium in all conjugations
persisted between 8.5 and 9.0. The objective was to com-
pare conjugations using different types of alkyl squarate
(made from amino amide or hydrazide) as the reactive
species, and to examine the effect of buffer and hapten
concentration on the rate and efficiency of conjugation.
Aided by the results of the study, we aimed to identify
structural features in tethers that make conjugation
more efficient. When applied in future conjugations of
different haptens and carriers, the best conditions found
here with BSA and haptens derived from lactose may
have to be adjusted, to account for reactivity of other
reagents and the desired hapten–carrier ratios.
final loading than those with haptens 14–17 made from
hydrazide 13. That squarates prepared from the hydr-
azide react with lesser efficiency manifest itself most
clearly at low concentrations of hapten in a low concen-
tration buffer (Table 1). At 40 mM concentration of
hapten, the four squarates within the two groups (5–8
and 14–17) reacted at a comparable rate and, when
the reaction within the two groups was complete, affor-
ded two groups of products (9–12 and 18–21) with sim-
ilar hapten–BSA ratios (Table 1, Fig. 1). Conjugation at
three concentrations of hapten (40, 4, and 0.4 mM) was
compared with methyl (5) and decyl (8) esters. The reac-
tion rate and the final loading were much more concen-
tration dependent with the methyl derivative 5 than with
its decyl counterpart 8 (Table 1, Fig. 2), which showed
practically no dependence on concentration.
We have previously12 investigated the effect of concen-
tration of hapten (c = 5–25 mM) upon conjugation effi-
ciency with an ethyl squarate derivative. When 0.05 M
borate buffer was used, the conjugation was substan-
tially more efficient when carried out at c = 15 mM
(66%) than at c = 5 mM (45.6%). A further increase in
hapten concentration was much less beneficial (67.6%
at 20 mM), and the conjugation efficiency actually
decreased to 63.4% at a concentration of 25 mM. We did
not ascribe any significance to this moderate decrease
in efficiency at the time when those studies were carried
At a hapten concentration of 40 mM (Table 1, Fig. 1),
conjugations involving haptens 5–8 made from amino
amide 4 were faster and afforded conjugates with higher
20
19
18
17
16
15
14
13
12
11
10
9
5
6
7
8
8
14
15
16
17
7
6
5
4
3
2
1
0
0
10
20
30
40
50
60
70
80
90
100
110
120
130
Time [h]
Figure 1. Conjugation of squaric acid monoesters 5–8 made from amino amide 4 and squaric acid monoesters 14–17 made from hydrazide 13 and
BSA. All conjugations were done in Buffer B at hapten concentration of 40 mM.

S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
203
20.0
17.5
15.0
12.5
10.0
7.5
5
, 40 mM
5, 4 mM
5, 0.4 mM
8
8
8
, 40 mM
, 4 mM
, 0.4 mM
5.0
2.5
0.0
0
5
10
15
20
25
30
35
40
45
50
55
Time [h]
Figure 2. Conjugation of squaric acid monoesters 5 and 8 made from amino amide 4 and BSA in Buffer B at different concentrations of hapten.
110
100
90
80
70
60
50
16
17
40
30
20
10
0
14
8
15
7
6
5
0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20
Time [Days]
Figure 3. Hydrolysis of squaric acid monoesters 5–8 made from amino amide 4 and squaric acid monoesters 14–17 made from hydrazide 13 at pH 9
Buffer B.
out. However, in view of the now observed lowering of
the pH of the reaction medium upon solubilization of
BSA in 0.05 M buffer (see above), it is reasonable to
conclude, in retrospect, that in that past conjugation
at 25 mM concentration, which was also conducted in
a 0.05 M buffer, the buffer may have just started to
break at that hapten concentration. At those conditions,
that is, when only a relatively small volume of buffer was
used (because the concentration of the hapten was rela-
tively high), the buffer capacity was low because only a
small amount of buffer salts was present, and the rate
and efficiency of conjugation decreased when the reac-
tion medium became less basic. Support for this conclu-
sion was found when the conjugation of 5 was
conducted in the same buffer but at a lower concentra-
tion of hapten (4 mM). Here, a proportionally larger
volume of buffer was used, which contained a larger
total amount of salts—in fact the same amount of salts
as when the conjugation is done in 0.5 M buffer at a hap-
ten concentration of 40 mM—and the buffer had, there-
fore, a larger capacity. The conjugation was somewhat
slower, as could be expected due to the lower concentra-

204
S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
tion of hapten (Table 1), but the product obtained
was very similar to the one formed in 0.5 M buffer
(Table 1). Monitoring the pH of the reaction mix-
tures showed that the pH persisted during both
conjugations.
when the desired, low carbohydrate–protein ratio has
been reached.
The high reaction rate and efficiency of conjugation
observed when using 0.5 M pH 9 buffer at an initial hap-
ten–BSA ratio 20:1 prompted us to examine, with differ-
ent oligosaccharides, the minimum excess of hapten
necessary to attach ꢀ5 hapten residues per carrier. This
loading was targeted because a conjugate from hexasac-
charide 3120 (Scheme 2) and BSA with a loading ꢀ4.5
was previously37 found to be a more potent immunogen
for anti V. cholerae O:1, serotype Ogawa antibodies
than similar conjugates containing ꢀ10 or ꢀ15 hapten
residues/BSA. In addition to hapten 5, conjugation of
the tri- (29),13 tetra- (30)13 and hexasaccharide (31) frag-
ments of the O-PS of V. cholerae O:1, serotype Ogawa
was also examined. The original preparations13 of
conjugates from BSA and the Ogawa tri- and the
tetrasaccharides were aimed at preparations of three
conjugates in one pot, with loadings of 5, 10, and 15,
The results of conjugations of methyl squarates 5 and
14 at different concentrations of buffer (0.5 and 0.05 M,
that is, the buffer capacity in both cases was sufficient to
keep the pH close to 9) allowed us to conclude (Table 1)
that the concentration of buffer is another variable that
affects the conjugation process. This finding widens the
choices of conditions suitable for a specific conjugation.
There are situations when it may be beneficial to con-
duct conjugations at a lower concentration of haptens
or buffers, or both, or not with the most powerful
reagent. Such is the case, for example, when a series of
conjugates are being prepared in one pot,13 and some
conjugates in the series are targeted to have low loading,
because a slower reaction may be easier to terminate
OMe
OMe
H3C
O
H3C
O
O
O
NH
HO
HO
HO
HO
HO
HO
NH
HO
HO
HO
HO
HO
HO
O
O
H3C
O
H3C
O
O
O
NH
HO
NH
HO
O
O
H3C
O
H3C
O
O
O
O
O
O
O
NH
NH
HO
HO
HO
O(CH2)5CONH(CH2)2NH
HO
OEt
O(CH2)5CONH(CH2)2NH
NH
BSA
5.3
29
32
OMe
OMe
O
H3C
O
H3C
O
O
NH
HO
HO
HO
HO
HO
NH
HO
HO
HO
HO
HO
HO
HO
O
O
H3C
O
H3C
O
O
O
NH
HO
NH
HO
O
2
O
2
H3C
O
H3C
O
O
O
O
O
O
O
NH
HO
NH
HO
HO
O(CH2)5CONH(CH2)2NH
HO
OEt
O(CH2)5CONH(CH2)2NH
NH
BSA
4.9
30
33
OMe
OMe
O
H3C
O
H3C
O
O
NH
HO
HO
HO
HO
HO
HO
NH
HO
HO
HO
HO
HO
HO
HO
HO
O
O
H3C
O
H3C
O
O
O
NH
HO
NH
HO
O
4
O
4
H3C
O
H3C
O
O
O
O
O
O
O
NH
NH
HO
HO
O(CH2)5CONH(CH2)2NH
OEt
O(CH2)5CONH(CH2)2NH
NH
BSA
5.4, 12
31
34
Scheme 2.

S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
205
hence the high initial hapten–BSA ratio of 75:1. Under
those conditions (0.05 M pH 9 buffer, 15 mM concentra-
tion of hapten), the first desired loading was reached in
both cases after 3 h. As we show here, when the 0.5 M
buffer was used at the hapten concentration of 40 mM,
a conjugate with loading ꢀ5 could be obtained within an
acceptable time of 8 h (Table 3), while only slight, 20%
excess of hapten was used (initial hapten–BSA ratio, 6:1).
Because conjugation of larger oligosaccharides is a
slower process compared to that of smaller molecules,
the bulkier hexasaccharide 31 was conjugated in 0.5 M
buffer at the initial hapten–BSA ratio of 7:1. The desired
conjugate with ꢀ5 haptens attached to BSA was
obtained after 5 h (Table 3), when the reaction was vir-
tually complete. For comparison, the conjugation was
also carried out in 0.5 M buffer, at the hapten–BSA ratio
of 15:1 as in the past.20 At that time, the conjugation
was carried out in 0.05 M buffer at a hapten concentra-
tion of 20 mM, and five hapten residues were attached
after 24 h. The increasing molecular mass of the conju-
gate leveled off after 14 days with 11 mol of hapten/
BSA. In the present experiment, five hapten residues/
BSA were attached after 2.5 h, and the increasing hap-
ten/BSA ratio leveled off after 2 days, when the conju-
gate formed showed hapten/BSA ratio of 12.
the same number of secondary amino groups, which
could further react.6 Thus, it should be theoretically pos-
sible to attach many more than 59 hapten residues/BSA.
Conjugations with disaccharide 5 were carried out at the
initial hapten–BSA ratios 60:1 and 120:1. The products
formed after 3 days, when the loading no longer in-
creased, contained (Table 3) on average, ꢀ46 and ꢀ56
hapten residues/BSA, respectively. Experiments aimed
at determining if any tertiary amines had been formed
are in progress. Elsewhere,18 human serum albumin
(HSA, 58 lysine residues) was treated with a methyl
squarate monoester prepared from D-glucose, at the
initial hapten–HSA ratio of 197:1. The hapten was
equipped with the same linker as the one present in 5,
and 32 sugar residues were attached after 7 days (conju-
gation efficiency 17%). More recently, Izumi17 conju-
gated to BSA ethyl squarate derivative of D-mannose
equipped with a different spacer at a hapten–BSA ratio
of 60 and 120, and obtained conjugates with ꢀ25 and
ꢀ39 hapten residues, respectively, after 5 days (conjuga-
tion efficiencies, 41.6% and 32.5%).
The results of the above experiments showed that con-
jugates could be prepared very efficiently by applying
the new protocol (Section 3.8.2), with an excess of la-
bor-intensive synthetic oligosaccharides as low as that
used in organic synthesis of small molecular mass
substances (ꢀ20%). Thus, recovery of the unchanged
hapten from the conjugation reaction15,16 becomes
virtually a non-issue.
To refine and expand on our previous fragmentary
data29 regarding the stability of squarate monoesters ob-
tained by TLC analysis, the hydrolysis of 5–8 and 14–17
under the conjugation conditions but in the absence of
The efficiency of the new protocol (Section 3.8.2) was
further tested and verified when hapten 5 was conju-
gated to BSA under conditions when the amount of hap-
ten used was not the limiting factor. There are 59 lysine
residues38 in BSA, each bearing a terminal amino group
but not all are equally accessible to chemical transfor-
mations. If all engaged in the formation of a conjugate,
the primary amino groups would be transformed into
8, 4.0 mM, Buffer C
7, 4.0 mM, Buffer C
6, 4.0 mM, Buffer C
5, 4.0 mM, Buffer C
5, 0.4 mM, Buffer B
100
90
80
70
60
50
40
30
20
10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Time [Days]
Figure 4. Hydrolysis of squaric acid monoesters 5–8 made from amino amide 4 in Buffer C at a hapten concentration of 4 mM with that of 5 in Buffer
B at a hapten concentration of 0.4 mM.

206
S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
BSA was monitored by HPLC. In Buffer B (Section 3.1)
at a 40 mM hapten concentration, alkyl squarates pre-
pared from hydrazide 13 (squarates 14–17) were more
stable than compounds prepared from amino amide 5,
except for the decyl compound 8, whose stability was
similar to those of 14–17. The hydrolysis of 5–8 followed
a similar pattern in Buffer C at 4 mM hapten concentra-
tion. Because of their poor reactivity at low concentra-
tions of hapten, the hydrolysis of squarates made from
hydrazide 13 was not examined with the latter condi-
tions. The relative propensity of monoalkyl squarates
to hydrolysis is shown in Figures 3 and 4. The relatively
high rate of hydrolysis of squarate 5 is particularly note-
worthy, and it explains the low efficiency of conjugation
at these conditions (Table 1). Nevertheless, because at
the conditions of the new protocol almost all hapten is
conjugated within 24 h, the commercially available,
crystalline squaric acid dimethyl ester is a very useful
reagent to transform carbohydrates into derivatives
amenable to conjugation.
follow unequivocally from the mode of synthesis,
NMR data and m/z values found in their mass spectra,
and their purity was verified by TLC, HPLC, and NMR
spectroscopy. HPLC was performed with Agilent 1100
Series Chromatography System, using a 150 · 4.5 mm
column with Ultra IBD RP packing, particle size,
5 lm (Restek Corporation, Bellefonte, PA 16823). A
mobile phase of 30% MeOH–70% 10 mM monobasic
potassium phosphate (1.0 mL/min) was used for quanti-
tative analysis of methyl and butyl squarates 5, 7, 14,
and 16. For ethyl and decyl squarates (6, 8, 15, and
17), 10% MeCN–90% 10 mM monobasic potassium
phosphate (1.0 mL/min) was used as the mobile phase.
Compounds were detected at 280 nm. BSA was pur-
chased from Sigma Chemical Company: A. Cat. Num-
ber A-0281; B. Fraction V, Sigma Cat. No. A-4503,
from which fatty acids were removed by charcoal treat-
ment.32 Buffers used were as follows. A, BuffAR pH 7.0
Reference solution (Mallincrodt, Cat. No. 0031-04),
concentrated to 1/5 volume; B, 0.5 M borate buffer pH
9, made in house (1 L) from boric acid (30.9 g), KCl
(26.1 g), and KOH (8.42 g), and final adjustment to
pH 9.0 by addition of solid KOH; C, 0.05 M pH 9.0
buffer (RICCA chemical company, Cat. No. 1590-16;
information concerning the concentration of this buffer
was provided by the company’s technical assistance
service). Buffer salts were obtained by freeze-drying
Buffer B. It was experimentally determined that 27 mL
of Buffer A had to be added to 2.5 mL Buffer B to
bring pH to 7. Carbohydrate–protein conjugates were
rid of low-molecular mass materials by filtration
through either Amicon Ultra-15 Centrifugal Filter
(Milipore Corporation) or similar Vivaspin 15 R (Sarto-
rius Group) devices.31 Measurement of pH was done
with Corning pH Meter, Model 440 and a semi-micro
refillable glass electrode. For quick check of pH of
reaction mixtures run in a very small volume (75–
150 lL), samples (ꢀ0.5–1 lL) were withdrawn with a
glass capillary and applied onto small strips of Hydrion
MicroFine pH Paper (Micro Essential Laboratory, Inc.,
Cat. No. MF-1616 and MF-1608), whose accuracy is
described as 0.2–0.3 pH units. Dimethyl, diethyl, and
dibutyl squarates were purchased from Aldrich Chemi-
cal Company. Didecyl squarate was prepared as
described.10 Aqueous solutions were made using HPLC
grade water. Solutions in organic solvents were dried
with anhydrous Na2SO4, and concentrated at 40 ꢀC/
2 kPa.
3. Experimental
3.1. General methods
Unless stated otherwise, optical rotations were mea-
sured at ambient temperature with a Perkin–Elmer
automatic polarimeter, Model 341. All reactions were
monitored by thin-layer chromatography (TLC) on Sil-
ica Gel 60 coated glass slides. Column chromatography
was performed by elution from columns of silica gel with
the CombiFlash Companion Chromatograph (Isco,
Inc.). Solvent mixtures for TLC were more polar than
those used for preparative separation. Nuclear Magnetic
Resonance (NMR) spectra were measured at 300 MHz
(1H) and 75 MHz (13C) with a Varian Gemini or Varian
Mercury spectrometer, or at 600 MHz (1H) and
150 MHz (13C) with a Bruker Avance 600 spectrometer.
Assignments of NMR signals were made by homo-
nuclear and heteronuclear 2-dimensional correlation
spectroscopy, run with the software supplied with the
spectrometers. When reporting assignment of NMR sig-
nals, nuclei associated with the spacer (linker) are
denoted with a prime. Sugar residues in oligosaccharides
are serially numbered, beginning with the one bearing
the aglycone, and are identified by a Roman numeral
superscript in listings of signal assignments. Liquid
chromatography–electron spray-ionization mass spec-
trometry (ESI-MS) was performed with a Hewlett–
Packard 1100 MSD spectrometer. Attempts have been
made to obtain correct combustion analysis data for
all new compounds. However, some compounds tenaci-
ously retained traces of solvents, despite exhaustive
drying, and analytical figures for carbon could not be
obtained within 0.4%. Structures of these compounds
3.2. 2,3,4,6-Tetra-O-acetyl-b-D-galactopyranosyl-(1!4)-
1,2,3,6-tetra-O-acetyl-b-D-glucopyranose (1)
This compound was prepared as described.21,22 Traces
of the a anomer, which was impractical to remove by
multiple recrystallization, were removed by chromato-
graphy on silica gel, mp 91–95 ꢀC (CH2Cl2–MeOH),

S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
207
[a]D ꢁ4.3 (c 6, CHCl3). The compound is polymor-
phous. Lit.21 mp 140–143 ꢀC, [a]D ꢁ4.0 (c 3.6, CDCl3),
J3,4 = 3.8 Hz, J4,5 = 1.0 Hz, H-4II), 5.21 (t, 1H,
J = 9.5 Hz, H-3I), 5.12 (dd, 1H, J1,2 = 8.0 Hz,
J2,3 = 10.5 Hz, H-2II), 4.96 (dd, 1H, J3,4 = 3.4 Hz,
J2,3 = 10.4 Hz, H-3II), 4.83 (d, 1H J1,2 = 3.9 Hz, H-1I),
4.50 (d, 1H, J1,2 = 7.9 Hz, H-1II), 4.42–4.39 (dd, 1H,
1
lit.22 mp 90 ꢀC, [a]D ꢁ4.7 (c 10.3, CHCl3). H NMR
data agreed with those reported.21 13C NMR (CDCl3,
150 MHz):
d
100.9 (C-1II), 91.5 (C-1I), 75.6
(C-4I), 73.5 (C-5I), 72.6 (C-3I), 70.9 (C-3II), 70.7 (C-
5II), 70.5 (C-2I), 69.0 (C-2II), 66.6 (C-4II), 61.7 (C-6I),
60.8 (C-6II).
J5,6 = 2.1 Hz, J6a,6b = 11.9 Hz, H-6II), 4.19–4.16 (dd,
a
1H, J5,6 = 6.0 Hz, J6a,6b = 11.2 Hz, H-6I ), 4.14–4.11
a
(dd, 1H, J5,6 = 5.0 Hz, J6a,6b = 12.0 Hz, H-6II), 4.09–
b
4.06 (dd, 1H, J5,6 = 7.7Hz, J6a,6b = 11.1Hz, H-6I ),
b
3.3. 5-Methoxycarbonylpentyl 2,3,4,6-tetra-O-acetyl-b-D-
galactopyranosyl-(1!4)-2,3,6-tri-O-acetyl-b-D-glucopy-
ranoside (2)
3.89–3.85 (m, 2H, H-5I and H-5II), 3.74–3.70 (m, 1H,
H-10 ), 3.67 (s, 3H, OCH3), 3.66 (t, 1H, J = 10.4Hz,
a
H-4I), 3.56–3.52 (m, 1H, H-2I, shifts after acetylation
to d 4.78), 3.48–3.44 (m, 1H, H-10b), 2.35 (t, 2H, H-50),
2.16 (s, 3H, COCH3), 2.12 (2 s, 6H, 2 COCH3), 2.06
(s, 6H, 2COCH3), 1.96 (s, 3H, COCH3), 1.69–1.61 (m,
4H, H-20 and H-40), 1.40–1.36 (m, 2H, H-30); 13C
NMR (CDCl3): d 173.9, 170.6, 170.3, 170.2, 170.1,
170.0, 169.0, 101.7 (C-1II), 98.0 (C-1I), 76.2 (C-4I), 73.4
(C-3I), 71.2 (C-2I), 71.0 (C-3II), 70.4 (C-5I), 69.1
(C-2II), 68.3 (2 C, overlap, C-5II and C-10), 66.5
(C-4II), 62.1 (C-6II), 60.7 (C-6I), 51.4 (OCH3), 33.7 (C-
50), 28.8 (C-20), 25.4 (C-30), 24.4 (C-40), 21.0, 20.7,
20.5, 20.4; TOF-MS: 740.3 [M+NH4]+, 745.2
[M+Na]+, 761.2 [M+K]+.
Methyl 6-hydroxyhexanoate23 (745 mg, 5.10 mmol) and
BF3ÆEt2O (0.5 mL, 2.08 mmol) was added to a solution
of 1 (1.38 g, 2.04 mmol) in CH2Cl2 (10 mL) and the mix-
ture was stirred at room temperature overnight, when
TLC showed that the reaction was virtually complete.
After neutralization with Et3N (2 mL) and concentra-
tion, chromatography (4:1 hexane–acetone) gave first
methyl 6-O-acetylhexanoate 25 (0.12 g, 13%). 1H
NMR (CDCl3): d 4.06 (t, 2H, J = 7.0 Hz, H-1), 3.67
(s, 3H, OCH3), 2.32 (t, 2H, J = 7.4 Hz, H-5), 1.68–
1.62 (m, 4H, H-2 and H-4), 1.39 (m, 2H, H-3); 13C
NMR (CDCl3): d 173.7 (CO), 170.9 (CO), 64.0 (C-1),
51.3 (OCH3), 33.7(C-5), 28.1(C-2), 25.3 (C-3), 24.3 (C-
4), 20.8 (CH3CO); TOF-MS: 189 [M+H]+, 129
[MꢁHOAc]+.
3.4. 5-Methoxycarbonylpentyl b-D-galactopyranosyl-
(1!4)-a-D-glucopyranoside (28) and 5-methoxycarbonyl-
pentyl 2,3,4,6-tetra-O-acetyl-b-D-galactopyranosyl-
(1!4)-a-D-2,3,6-tri-O-acetyl-a-D-glucopyranoside (26)
Eluted next was the title compound 2 (970 mg, 62%),
1
[a]D ꢁ12.7 (c 1.1, CHCl3). H NMR (CDCl3): d 5.34
(dd, 1H, J3,4 = 3.5 Hz, J4,5 = 1.1 Hz, H-4II), 5.19 (t,
1H, J = 9.3 Hz, H-3I), 5.10 (dd, 1H, J1,2 = 8.0 Hz,
J2,3 = 10.4 Hz, H-2II), 4.96 (dd, 1H, J3,4 = 3.5 Hz, H-
3II), 4.87 (dd, 1H, J1,2 = 8.0 Hz, J2,3 = 9.6 Hz, H-2I),
4.50 (d, 1H, J1,2 = 7.9 Hz, H-1I), 4.47 (dd, 1H,
For characterization, a small amount of compound 27
was treated with NaOMe in MeOH to give 28, mp
1
149–150 ꢀC (from MeOH); [a]D +87.6 (c 0.9, H2O); H
NMR (D2O): d 4.90 (d, 1H, J1,2 = 3.9 Hz, H-1I), 4.34
(d, 1H, J1,2 = 7.8 Hz, H-1II), 3.92 (dd, 1H, J3,4
3.4 Hz, J4,5 = 0.8 Hz, H-4II), 3.90 (dd, 1H, J5,6
=
=
J5,6 = 2.1 Hz, J6a,6b = 12.0 Hz, H-6I ), 4.45 (d, 1H,
a
J1,2 = 8.0 Hz, H-1II), 4.14–4.07 (m, 3H, H-6I , H-6aII;b),
2.4 Hz, J6a,6b = 12.2 Hz, H-6Ia), 3.84 (dd, 1H,
b
3.88 (m, 1H, H-5II), 3.83 (m, 1H, H-1a0 ), 3.79 (t, 1H,
J = 9.3 Hz, H-4I), 3.66 (s, 3H, OCH3), 3.60 (m, 1H,
H-5I), 3.45 (m, 1H, H-10b), 2.29 (t, J = 7.6 Hz, 2H,
H-50), 2.17, 2.15, 2.12, 2.06, 2.04, 2.03, 1.96 (7s, 21H,
7COCH3), 1.64–1.54 (m, 4H, H-20 and H-40), 1.35 (m,
2H, H-30); 13C NMR (CDCl3): d 173.8, 170.2, 170.1,
169.9, 169.8, 169.6, 169.4, 168.9, 100.9 (C-1II), 100.4
(C-1I), 76.1 (C-4I), 72.6 (C-3I), 72.4 (C-5I), 71.5 (C-2I),
70.8 (C-3II), 70.5 (C-5II), 69.6 (C-10), 68.9 (C-2II), 66.4
(C-4II), 61.9 (C-6I), 60.6 (C-6II), 51.3 (OCH3), 33.7
(C-50), 28.9 (C-20), 25.2 (C-30), 24.4 (C-40), 20.7, 20.6,
20.5, 20.4 (3 C), 20.3; TOF-MS: [M+Na]+ calcd for
C33H48O20Na, 787.2637; found, 787.2640. Anal. Calcd
for C33H48O20: C, 51.83; H, 6.33. Found: C, 52.08; H,
6.38.
J5,6 = 4.6 Hz, J6a,6b = 12.2 Hz, H-6I ), 3.81 (dd, 1H,
b
J2,3 = 9.8 Hz, J3,4 = 8.8 Hz, H-3I), 3.80–3.75 (m, 3H,
H-5I, H-6II ), 3.74–3.71 (m, 2H, H-5I, H-10a), 3.66 (dd,
a;b
1H, J2,3 = 10.0 Hz, J3,4 = 3.4 Hz, H-3II), 3.64 (dd, 1H,
J3,4 = 8.8 Hz, J4,5 = 10.0 Hz, H-4I), 3.59 (dd, 1H,
J1,2 = 3.9 Hz, J2,3 = 9.8 Hz, H-2I), 3.54 (dd, 1H, J1,2
=
7.8 Hz, J2,3 = 10.0 Hz, H-2II), 3.52 (m, 1H, H-10 ), 2.41
b
(t, 2H, J = 7.3 Hz, H-50), 1.66–1.61 (m, 4H, H-20 and
H-40), 1.39 (m, 2H, H-30); 13C NMR (D2O): d
180.4 (CO), 105.6 (C-1II), 100.6 (C-1I), 81.2 (C-4I),
78.1 (C-5II), 75.3 (C-3II), 74.5 (C-3I), 73.8 (2 C, C-2I
and C-2II), 73.4 (C-5I), 71.3 (C-4II), 70.9 (C-10),
63.8 (C-6II), 62.7 (C-6I), 54.9 (OCH3), 36.4 (C-50), 31.0
(C-20), 27.7 (C-30), 26.9 (C-40); TOF-MS: [M+Na]+
calcd for C19H34O13Na, 493.1897; found, 493.1901.
Anal. Calcd for C19H34O13: C, 48.51; H, 7.28. Found:
C, 48.46; H, 7.24.
Continued elution gave the amorphous 5-methoxycar-
bonylpentyl
2,3,4,6-tetra-O-acetyl-b-D-galactopyran-
osyl-(1!4)-3,6-di-O-acetyl-a-D-glucopyranoside (27, 170
A small amount of 28 was treated with acetic anhy-
dride in pyridine to give amorphous 26, [a]D +7.3 (c
mg, 12%): 1H NMR (CDCl3): d 5.35 (dd, 1H,

208
S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
1.7, CHCl3); 1H NMR (CDCl3): d 5.46 (dd, 1H,
J2,3 = 10.1 Hz, J3,4 = 9.3 Hz, H-3I), 5.34 (dd, 1H,
J3,4 = 3.5 Hz, J4,5 = 1.0 Hz, H-4II), 5.11 (dd, 1H,
J1,2 = 7.9 Hz, J2,3 = 10.4 Hz, H-2II), 4.97–4.94 (m, 2H,
3.6. (2-Aminoethylamido)carbonylpentyl b-D-galacto-
pyranosyl-(1!4)-b-D-glucopyranoside (4)
A solution of compound 3 (8.08 g, 17.2 mmol) in ethyl-
enediamine (35 mL, 518 mmol) was stirred at 50 ꢀC with
the exclusion of atmospheric moisture and CO2. The
mixture, which shortly became pale yellow, was stirred
at the same temperature for 24 h and concentrated with
co-evaporation of water, to remove ethylenediamine.
The residue was chromatographed (6:2:1 MeOH–
DCM–NH4OH), to give hemihydrate of 4 (7.11 g,
83.5%), mp 168–170 ꢀC (from MeOH); [a]D ꢁ2.6 (c
H-1I and H-3II), 4.78 (dd, 1H, J1,2 = 3.7 Hz, J2,3
=
10.3 Hz, H-2I), 4.49 (d, 1H, J1,2 = 8.0 Hz, H-1II), 4.45–
4.42 (dd, J5,6 = 2.1 Hz, J6a,6b = 12.0 Hz, H-6I ), 4.16–
a
4.12 (m, 2H, H-6Ib and H-6IaI), 4.08 (dd, 1H J =
7.5 Hz, J = 11.1 Hz, H-6IbI), 3.92–3.90 (m, 1H, H-5I),
3.88–3.86 (m, 1H, H-5II), 3.72 (t, J = 9.0 Hz, 1H,
H-4I), 3.68–3.64 (m, 4H, OCH3 and H-10 ), 3.40–3.37
a
(m, 1H, H-10b), 2.32 (t, 2H, J = 7.2 Hz, H-50), 2.15 (s,
3H, COCH3), 2.12 (s, 3H, COCH3), 2.06 (s, 3H,
COCH3), 2.05 (2 s, 6H, 2COCH3), 2.04 (s, 3H, COCH3),
1.96 (s, 3H, COCH3), 1.67–1.59 (m, 4H, H-20 and H-40),
1.40–1.36 (m, 2H, H-30); 13C NMR (CDCl3): d 173.9,
170.4, 170.3, 170.2, 170.1, 170.0, 169.4, 169.0, 101.1
(C-1II), 95.5 (C-1I), 76.6 (C-4I), 71.1 (C-2I), 71.0
(C-3II), 70.5 (C-5II), 69.9 (C-3I), 69.1 (C-2II), 68.2
(C-10), 68.0 (C-5I), 66.5 (C-4II), 61.9 (C-6I), 60.7
(C-6II), 51.4 (OCH3), 33.8 (C-50), 28.8 (C-20), 25.4
(C-30), 24.5 (C-40), 20.8 (2C), 20.6, 20.5 (2C), 20.4;
TOF-MS: [M+NH4]+ calcd for C33H52NO20, 782.3083;
found, 782.3072. Anal. Calcd for C33H48O20: C, 51.83;
H, 6.33. Found: C, 52.10; H, 6.54.
1
1.1, H2O); H NMR (D2O): d 4.46 (d, 1H, J1,2 7.9 Hz,
H-1I), 4.43 (d, 1H, J1,2 = 7.8 Hz, H-1II), 3.96 (dd, 1H,
J5,6 = 2.2 Hz, J6a,6b = 12.3 Hz, H-6I ), 3.92–3.88 (m,
a
2H, H-10a and H-4II), 3.80–3.72 (m, 3H, H-6bI , H-6IaI,
H-6bII), 3.71 (m, 1H, H-5II), 3.68–3.61 (m, 4H, H-10b,
H-3I, H-4I, H-3II), 3.59–3.56 (m, 1H, H-5I), 3.53 (dd,
1H, J1,2 = 7.8 Hz, J2,3 = 9.9 Hz, H-2II), 3.29 (dd, 1H,
J1,2 = 8.0 Hz, J2,3 = 9.4 Hz, H-2I), 3.27–3.22 (m, 2H,
H-60), 2.75 (m, 2H, H-70), 2.25 (t, 2H, J = 7.4 Hz, H-
50), 1.65–1.58 (m, 4H, H-20 and H-40), 1.35 (m, 2H, H-
30); 13C NMR (D2O): d 105.7 (C-1II), 104.8 (C-1I),
81.2 (C-4I), 78.1 (C-5II), 77.5 (C-5I), 77.2 (C-3I), 75.6
(C-2I), 75.3 (C-3II), 73.7 (C-2II), 73.1 (C-10), 71.3
(C-4II), 63.8 (C-6II), 62.9 (C-6I), 43.9 (C-60), 42.6
(C-70), 38.5 (C-50), 31.1 (C-20), 27.7 (C-40), 27.3 (C-30);
TOF-MS: [M+H]+ calcd for C20H39N2O12, 499.249;
found, 499.250. Anal. Calcd for C19H34O13Æ0.5H2O: C,
47.33; H, 7.75. Found: C, 47.54; H, 7.75.
3.5. 5-Methoxycarbonylpentyl b-D-galactopyranosyl-
(1!4)-b-D-glucopyranoside (3)
A few drops of NaOMe in MeOH (1 M) was added to a
solution of 2 (16.59 g, 21.7 mmol) in a mixture of MeOH
(300 mL) and CH2Cl2 (50 mL) until strong alkalinity
and the mixture was stirred at room temperature for
3 h, when TLC showed that the reaction was complete.
After neutralization with Amberite IR-120 (H+) resin,
filtration, and concentration, the residue was chromato-
graphed (4:1:0.1 CH2Cl2–MeOH–H2O) to give 3 (8.74 g,
86%) as a hemihydrate, mp 159–160 ꢀC (from MeOH);
3.7. 1-[(2-Aminoethylamido)carbonylpentyl b-D-galacto-
pyranosyl-(1!4)-b-D-glucopyranoside]-2-methoxycyclo-
butene-3,4-dione (5)
A solution of compound 4 (580 mg, 1.16 mmol) and
3,4-dimethoxy-3-cyclobutene-1,2-dione (247 mg, 1.74
mmol) in Buffer A (6 mL) was stirred at room tempera-
ture for 3 h, when TLC showed that almost all 4 was
consumed. The mixture was concentrated and the resi-
due was chromatographed (8:1 CH2Cl2–MeOH) to give
5 (540 mg, 76%): 1H NMR (D2O): d 4.45 (d, 1H,
J1,2 = 8.1 Hz, H-1I), 4.43 (d, 1H, J1,2 = 7.9 Hz, H-1II),
4.37 and 4.34 (2s, 3H, H-80), 3.96 (dd, 1H,
1
[a]D ꢁ3.5 (c 1, H2O); H NMR (D2O): d 4.46 (d, 1H,
J1,2 = 8.0 Hz, H-1I), 4.43 (d, 1H, J1,2 = 7.8 Hz, H-1II),
3.95 (dd, 1H, J5,6 = 2.2 Hz, J6a,6b = 12.3 Hz, H-6I ),
a
3.91–3.89 (m, 2H, H-1a0 and H-4II), 3.80–3.72 (m, 3H,
H-6bI , H-6aII, H-6IbI), 3.71 (m, 1H, H-5II), 3.67 (s, 3H,
OCH3), 3.66–3.61 (m, 4H, H-10 , H-3I, H-4I, H-3II),
b
3.56 (m, 1H, H-5I), 3.52 (dd, 1H, J1,2 = 7.8 Hz,
J2,3 = 9.9 Hz, H-2II), 3.28 (m, 1H, H-2I), 2,39 (t, 2H,
J= 7.4 Hz, H-50), 1.64–1.59 (m, 4H, H-20 and H-40),
1.37 (m, 2H, H-30); 13C NMR (D2O): d 180.4
(CO), 105.7 (C-1II), 104.8 (C-1I), 81.2 (C-4I), 78.1
(C-5II), 77.5 (C-5I), 77.2 (C-3I), 75.6 (C-2I), 75.3
(C-3II), 73.7 (C-2II), 73.1 (C-10), 71.3 (C-4II), 63.8
(C-6II), 62.9 (C-6I), 54.8 (OCH3), 36.3 (C-50), 31.1
(C-20), 27.4 (C-40), 26.8 (C-30); TOF-MS: [M+H]+ calcd
for C19H35O13, 471.208; found, 471.207. Anal. Calcd for
C19H34O13Æ0.5H2O: C, 47.59; H, 7.36. Found: C, 47.60;
H, 7.39.
J5,6 = 2.2 Hz, J6a,6b = 12.2 Hz, H-6I ), 3.91 (d, 1H,
a
J3,4 = 3.5 Hz, H-4II), 3.87 (m, 1H, H-10 ), 3.81–3.72 (m,
a
3H, H-6Ib, H-6IaI, H-6bII), 3.73–3.69 (m, 2H, H-5II and
H-60a), 3.66–3.61 (m, 4H, H-10b, H-3I, H-4I, H-3II),
3.59–3.56 (m, 2H, H-5I and H-60b), 3.53 (dd, 1H,
J1,2 = 7.8 Hz, J2,3 = 10.0 Hz, H-2II), 3.40 (m, 2H,
H-70), 3.30–3.27 (m, 1H, H-2I), 2.21 (m, 2H, H-50),
1.63–1.51 (m, 4H, H-20 and H-40), 1.35–1.28 (m, 2H,
H-30); 13C NMR (D2O): d 105.7 (C-1II), 104.8 (C-1I),
81.2 (C-4I), 78.1 (C-5II), 77.5 (C-5I), 77.2 (C-3I), 75.6
(C-2I), 75.3 (C-3II), 73.7 (C-2II), 73.0 (C-10), 71.3
(C-4II), 63.8 (C-6II), 63.7 (C-80), 62.9 (C-6I), 46.7

S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
209
(C-60), 42.0 (C-70), 38.5 (C-50), 31.1 (C-20), 27.8 (C-40),
27.4 (C-30); TOF-MS: [M+Na]+ calcd for C25H40N2O12-
Na, 631.231; found, 631.232.
(0.0003 mmol) and 60 mg (0009 mmol), respectively]
at hapten–BSA ratio of 20:1 following the general
(suggested) protocol. For results, see Table 2.
3.8. Conjugation of squaric acid monoesters 5–8 and
14–17 to BSA
3.9. Hydrolysis of squarate monoesters
Solutions of compounds 5–8 and 14–17 in Buffer B or C
were kept at room temperature at hapten concentration
of 40, 4, and 0.4 mM, respectively. The progress of
hydrolysis was followed by HPLC, and results were cal-
culated using calibration curves, which were constructed
with the aid of standard solutions of pure substances.
For results, see Figures 3 and 4.
3.8.1. Preliminary conjugation experiment. Compound
5 (3.7 mg, 0.006 mmol) and BSA (Sigma A-4503, puri-
fied,32 20 mg, 0.0003 mmol) were allowed to react at a
hapten concentration of 40 mM, using 0.05 M pH 9
buffer (buffer C, 150 lL). The reaction was monitored
by SELDI TOF-MS, which showed that after 1, 4,
and 25 h the hapten–BSA ratios were 1.3, 3.3, and 3.5,
respectively. A small amount of buffer salts was added
to raise the pH from 7.2 0.2 to ꢀ8.8 0.2 and, after
an additional 12 h, the carbohydrate–protein ratio was
18:1, as showed by SELDI TOF-MS analysis. When
the reaction was repeated using buffer B, the same load-
ing was reached after 8 h (Table 1).
Acknowledgments
This research was supported by the Intramural Research
Program of the NIH, NIDDK. The authors wish to
thank Rick Lake of Restek Corporation for his interest
in this work, and for expert advice during the quantita-
tive analysis by HPLC.
3.8.2. General (suggested) conjugation protocol. [The
protocol described here for conjugation of 5 and BSA
at hapten concentration of 40 mmol in Buffer B (0.5 M
borate buffer, pH 9.0) should be adjusted as required,
depending on the nature and concentration of hapten,
carrier, and buffer. Conjugates listed in Table 1 have
been prepared in a similar way]. With the aid of Buffer
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.carres.
2007.10.015.
B
(3 · 50 lL), squaric acid derivative
5 (3.7 mg,
0.006 mm) was transferred into a glass vial containing
BSA (20 mg, 0.0003 mmol). The reaction was gently stir-
red and periodically monitored by SELDI TOF-MS.
When the increasing molecular mass leveled off, the mix-
ture was transferred into a centrifugal filter device and
processed to remove low-molecular mass material. A
minimum of 8 washes with aq 10 mM (NH4)2CO3 were
applied, to ensure complete solvent exchange. Freeze-
drying afforded conjugates as white solids. To determine
yields, materials thus obtained were further dried at
35 ꢀC for 48 h/133 Pa. For results, see Discussion,
Tables 1 and 3 and Figures 1 and 2.
References
1. Synthetic Oligosaccharides. Indispensable Probes in the Life
´ˇ
Sciences; Kovac, P., Ed.; ACS Symposium Series 560;
American Chemical Society: Washington, DC, 1994.
2. Lee, Y. C.; Lee, R. T. In Synthetic Glycoconjugates; Allen,
H. J., Kisailus, E. C., Eds.; Marcel Dekker: New York,
1992; pp 121–165.
3. Galonic, D. P.; Gin, D. Y. Nature 2007, 446, 1000–
1007.
4. Dick, W. E., Jr.; Beurret, M.. In Glycoconjugates of
Bacterial Carbohydrate Antigens; Cruse, J. M., Lewis, R.
E., Jr., Eds.; Krager: Basel, 1989; Vol. 10, pp 48–
114.
When two conjugates were made in one pot13 (as in
the conversion 31!34, Table 3), 90% of the reaction
mixture was withdrawn when the first target hapten–
BSA ratio was reached, and the remaining mixture
was stirred and periodically monitored until the mole-
cular mass of the conjugate being formed no longer
increased, as determined by SELDI TOF-MS. The
withdrawn portion of the mixture was diluted with 11
volumes of Buffer A, and processed as described above.
5. Tietze, L. F.; Arlt, M.; Beller, M.; Glusenkamp, K.-H.;
¨
Ja¨hde, E.; Rajewsky, M. F. Chem. Ber. 1991, 124, 1215–
1221.
6. Glusenkamp, K.-H.; Drosdziok, W.; Eberle, G.; Ja¨hde, E.;
¨
Rajewsky, M. F. Z. Naturforsch., C: Biosci. 1991, 46, 498–
501.
7. Tietze, L. F.; Schro¨ter, C.; Gabius, S.; Brinck, U.;
Goerlach-Graw, A.; Gabius, H.-J. Bioconjugate Chem.
1991, 2, 148–153.
8. Cohen, S.; Cohen, S. G. J. Am. Chem. Soc. 1966, 88, 1533–
1536.
3.8.3. Reproducibility of conjugation. Hapten 8 (4.42
mg, 0.006 mmol) was treated with BSA (20 mg,
0.0003 mmol) in Buffer B at a hapten concentration of
40 mM. Hapten 30 (1.9 mg, 0.0018 mmol and 5.7 mg,
0.0054 mmol) was treated in Buffer B with BSA [20 mg
9. Grunefeld, J.; Bredhauer, G.; Zinner, G. Arch. Pharm.
¨
(Weinheim) 1985, 318, 984–988.
10. Bergh, A.; Magnusson, B.-G.; Ohlsson, J.; Wellmar, U.;
Nilsson, U. J. Glycoconjugate J. 2001, 18, 615–621.

210
S. Hou et al. / Carbohydrate Research 343 (2008) 196–210
11. Kamath, V. P.; Diedrich, P.; Hindsgaul, O. Glycoconjugate
J. 1996, 13, 315–319.
26. Bochkov, A. F.; Zaikov, G. E. Chemistry of the Glyco-
sidic Bond: Formation and Cleavage. In Chemistry of the
Glycosidic Bond: Formation and Cleavage; Pergamon
Press: Oxford, 1979.
´ˇ
12. Zhang, J.; Yergey, A.; Kowalak, J.; Kovac, P. Carbohydr.
Res. 1998, 313, 15–20.
´ˇ
13. Saksena, R.; Ma, X.; Kovac, P. Carbohydr. Res. 2003, 338,
27. Banoub, J.; Bundle, D. R. Can. J. Chem. 1979, 57, 2091–
2097.
2591–2603.
´ˇ
14. Auzanneau, F.-I.; Pinto, M. Bioorg. Med. Chem. 1996, 4,
28. Ziegler, T.; Kovac, P.; Glaudemans, C. P. J. Liebigs Ann.
2003–2010.
Chem. 1990, 613–615.
´ˇ
´ˇ
15. Chernyak, A.; Karavanov, A.; Ogawa, Y.; Kovac, P.
29. Saksena, R.; Chernyak, A.; Karavanov, A.; Kovac, P. In
Carbohydr. Res. 2001, 330, 479–486.
Methods in Enzymology; Lee, Y. C., Lee, R., Eds.;
´ˇ
16. Saksena, R.; Chernyak, A.; Poirot, E.; Kovac, P. Methods
Academic Press, 2003; Vol. 362, pp 125–139.
´ˇ
Enzymol. 2003, 362, 140–160.
30. Ma, X.; Saksena, R.; Chernyak, A.; Kovac, P. Org.
17. Izumi, M.; Okumura, S.; Yuasa, H.; Hashimoto, H. J.
Carbohydr. Chem. 2003, 22, 317–329.
18. Pozsgay, V.; Dubois, E.; Pannell, L. J. Org. Chem. 1997,
62, 2832–2846.
19. Seitz, G.; Morck, H. Arch. Pharm. 1972, 305, 614–
618.
Biomol. Chem. 2003, 1, 775–784.
´ˇ
31. Saksena, R.; Adamo, R.; Kovac, P. Bioorg. Med. Chem.
2007, 15, 4283–4310.
32. Chen, R. F. J. Biol. Chem. 1967, 242, 173–181.
33. Chernyak, A.; Oscarson, S.; Turek, D. Carbohydr. Res.
2000, 329, 309–316.
´ˇ
20. Saksena, R.; Zhang, J.; Kovac, P. Tetrahedron: Asym-
34. Lefeber, D. J.; Kamerling, J. P.; Vliegenthart, J. F. G.
Chem.-Eur. J. 2001, 7, 4411–4421.
metry 2005, 16, 187–197.
21. Hronowski, L. J. J.; Szarek, W. A.; Hay, G. W.; Krebs,
35. Ellis, L. A.; Mc Vay, C. S.; Probert, M. A.; Zhang, Z.;
Bundle, D. R.; Appleton, J. A. Glycobiology 1997, 7, 383–
390.
A.; Depew, W. T. Carbohydr. Res. 1989, 190, 203–
218.
22. Hudson, C. S.; Johnson, J. M. J. Am. Chem. Soc. 1915, 37,
36. Mawas, F.; Niggemann, J.; Jones, C.; Corbel, M. J.;
Kamerling, J. P.; Vliegenthart, J. F. G. Infect. Immun.
2002, 5107–5114.
1270–1275.
23. Solladie, G.; Ziani-Cherif, C. J. Org. Chem. 1993, 58,
2181–2185.
37. Chernyak, A.; Kondo, S.; Wade, T. K.; Meeks, M. D.;
´ˇ
24. Banoub, J.; Bundle, D. Can. J. Chem. 1979, 57, 2085–
Alzari, P. M.; Fournier, J.-M.; Taylor, R. K.; Kovac, P.;
2090.
Wade, W. F. J. Infect. Dis. 2002, 185, 950–962.
38. Hirayama, K.; Akashi, S.; Furuya, M.; Fukuhara, K.-I.
Biochem. Biophys. Res. Commun. 1990, 173, 639–646.
25. Bochkov, A. F.; Betaneli, V. I.; Kochetkov, N. K. Bioorg.
Khim. 1977, 3, 39–45.