Chemoenzymatic synthesis and structural characterization of 2-O-sulfated glucuronic acid-containing heparan sulfate hexasaccharides

Article abstract of DOI:10.1093/glycob/cwu032

Heparan sulfate and heparin are highly sulfated polysaccharides that consist of a repeating disaccharide unit of glucosamine and glucuronic or iduronic acid. The 2-O-sulfated iduronic acid (IdoA2S) residue is commonly found in heparan sulfate and heparin; however, 2-O-sulfated glucuronic acid (GlcA2S) is a less abundant monosaccharide (～<5% of total saccharides). Here, we report the synthesis of three GlcA2S-containing hexasaccharides using a chemoenzymatic approach. For comparison purposes, additional IdoA2S-containing hexasaccharides were synthesized. Nuclear magnetic resonance analyses were performed to obtain full chemical shift assignments for the GlcA2S- and IdoA2S-hexasaccharides. These data show that GlcA2S is a more structurally rigid saccharide residue than IdoA2S. The antithrombin (AT) binding affinities of a GlcA2S- and an IdoA2S-hexasaccharide were determined by affinity co-electrophoresis. In contrast to IdoA2S-hexasaccharides, the GlcA2S-hexasaccharide does not bind to AT, confirming that the presence of IdoA2S is critically important for the anticoagulant activity. The availability of pure synthetic GlcA2S-containing oligosaccharides will allow the investigation of the structure and activity relationships of individual sites in heparin or heparan sulfate. The Author 2014.

Full text of DOI:10.1093/glycob/cwu032

Glycobiology Advance Access published April 25, 2014
1
Chemoenzymatic synthesis and structural characterization of 2-O-sulfated glucuronic acid
containing heparan sulfate hexasaccharides
Po-Hung Hsieh1, Yongmei Xu1, David A Keire2 and Jian Liu1,*
1Division of Chemical Biology and Medicinal Chemistry, Eshelman School of Pharmacy,
University of North Carolina, Chapel Hill, NC 27599
2Food & Drug Administration, CDER (Center for Drug Evaluation and Research), Division of
Pharmaceutical Analysis, 1114 Market Street, St. Louis, MO 63101
*To whom correspondence should be addressed: Rm 303, Beard Hall, University of North
Carolina, Chapel Hill, NC 27599. Tel.: 919-843-6511; E-mail: jian_liu@unc.edu
© The Author 2014. Published by Oxford University Press. All rights reserved. For permissions,
please e‐mail: journals.permissions@oup.com

2
Abstract
Heparan sulfate and heparin are highly sulfated polysaccharides that consist of a repeating
disaccharide unit of glucosamine and glucuronic or iduronic acid. The 2-O-sulfated iduronic
acid (IdoA2S) residue is commonly found in heparan sulfate and heparin; however, 2-O-sulfated
glucuronic acid (GlcA2S) is a less abundant monosaccharide (~<5% of total saccharides). Here,
we report the synthesis of three GlcA2S-containing hexasaccharides using a chemoenzymatic
approach. For comparison purposes, additional IdoA2S-containing hexasaccharides were
synthesized. NMR analyses were performed to obtain full chemical shift assignments for the
GlcA2S- and IdoA2S-hexasaccharides. These data show that GlcA2S is a more structurally rigid
saccharide residue than IdoA2S. The antithrombin binding affinities of a GlcA2S-
hexasaccharide and an IdoA2S-hexasaccharide were determined by affinity co-electrophoresis.
In contrast to IdoA2S-hexasaccharides, the GlcA2S-hexasaccharide does not bind to
antithrombin, confirming that the presence of IdoA2S is critically important for the anticoagulant
activity. The availability of pure synthetic GlcA2S-containing oligosaccharides will allow the
investigation of the structure and activity relationships of individual sites in heparin or heparan
sulfate.

3
Introduction
Proteoglycans mediate many protein functions and physiological processes ranging from
cell development, differentiation, inflammation, viral or parasite infection, metastasis to cancer
(Bishop, Schuksz et al. 2007). Proteoglycans consist of a core protein and various
glycosaminoglycans (GAG) side chains. Although the amino acid sequence of the core protein
is encoded by a specific gene, the GAG side chains are generated by a non-template enzymatic
process (Feyerabend, Li et al. 2006, Carlsson, Presto et al. 2008, Victor, Nguyen et al. 2009).
Heparin and heparan sulfate (HS) are members of this polysulfated GAG family.
Heparin, a widely used clinical anticoagulant since the 1930’s, was the first biopolymeric
drug and is one of the few carbohydrate based drugs (Linhardt 2003, Liu, Zhang et al. 2009).
Because of the heterogeneity of heparin sequences, understanding the contribution of a specific
sulfated carbohydrate sequence for a given biological activity remains challenging. Heparin is a
mixture of polysaccharides and has high polydispersity and polyheterogeneity (Linhardt 2003) as
well as flexible structural characteristics (Liu, Zhang et al. 2009). In addition, heparin is a highly
sulfated polysaccharide with an average molecular weight of ~17 kDa (Sommers, Ye et al. 2011)
and with a range of chain lengths from 5,000 to 50,000 Da (Ahsan, Jeske et al. 1995). Heparin
has repeating disaccharide units consisting of uronic acid (L-iduronic acid, IdoA, or D-glucuronic
acid, GlcA) linked 1 → 4 to D-glucosamine (GlcN) (Casu 1985), where the glucosamine residues
can be either N-acetylated (GlcNAc) or sulfated (GlcNS). The GlcN monosaccharide can also
contain O-sulfo groups, including 6-O- and 3-O-sulfation, while the GlcA and IdoA residues can
be 2-O-sulfated. Heparan sulfate (HS) is structurally analogous to heparin. HS has a similar

4
disaccharide repeating structure to heparin with some noticeable differences. Generally, the HS
chains are longer with an average MW of 30 kDa (Griffin, Linhardt et al. 1995), and HS contains
a higher GlcA content and less sulfation than heparin (Linhardt 2003). In addition, HS has more
diverse sulfated saccharide sequences and domain structures compared with heparin.
Reliable structural characterization techniques are critically important to safeguard the
quality of heparin drugs. In 2008, batches of contaminated heparin were discovered in US which
drew significant public attention because the contaminated drug killed 94 patients in US
(McMahon, Pratt et al. 2010). Nuclear Magnetic Resonance (NMR) techniques were employed
and rapidly identified the nature of the contaminant to be over sulfated chondroitin sulfate. This
demonstrated the important role of NMR measurements in heparin quality control (Liu, Zhang et
al. 2009). NMR data provides structural information that cannot be obtained by other methods,
specifically; epimer recognition, anomeric configuration identification, sulfated position
determination, and conformation and secondary structure determination. One limitation in
utilizing NMR to analyze the structure of heparin or HS is that these species are a mixture of
compounds and pure oligosaccharide standards have not been available. Although certain
oligosaccharides have been isolated from partially depolymerized heparin or HS, obtaining
sufficient amounts of those highly purified oligosaccharides is a difficult and time consuming
process.
Recently, a new method using recombinant HS biosynthetic enzymes to prepare
structurally homogeneous HS oligosaccharides was developed (Xu, Masuko et al. 2011). This
method allows the synthesis of a series of structurally homogeneous heparin or HS
oligosaccharides at the scale of tens to hundreds of milligram with high yield. Access to highly
purified HS oligosaccharides allows for full-scale analyses of these complex molecules by NMR.

5
Notably these synthetic compounds enable the analysis of structurally “rare” oligosaccharides
which are difficult to isolate in sufficient quantity for NMR analysis. For example, the -GlcA2S-
residue represents only 1-5% of total uronic acid in heparin isolated from natural sources
(Guerrini and Torri 2008).
Here, to address this knowledge gap of the structure characteristics of these potentially
important oligosaccharide species, we report the NMR analysis of a series of GlcA2S-containing
hexasaccharides with different sulfation patterns. The GlcA2S oligosaccharides were prepared
by a chemoenzymatic approach. The GlcA2S NMR data was compared with the data obtained
from IdoA2S-containing oligosaccharides synthesized by the same method. The data confirm
previous assignments for the NMR signals of GlcA2S saccharides (Guerrini and Torri 2008).
Results
Chemoenzymatic synthesis of GlcA2S- and IdoA2S-containing hexasaccharides
Total of eight hexasaccharide constructs 1 to 8 were synthesized in this study using a
chemoenzymatic approach (Fig 1). The synthesis was initiated from a commercially available
monosaccharide, 1-O-(para-nitrophenyl) glucuronide (GlcA-pnp), requiring elongation steps,
sulfation and epimerization steps (Supplementary Fig S1) (Xu, Masuko et al. 2011). In the
synthesis, a hexasaccharide with a structure of GlcNTFA-GlcA-GlcNS-GlcA-GlcNS-GlcA-pnp
is a crucial intermediate (Supplementary Fig S1). The N-trifluoroacetyl group of this
hexasaccharide intermediate was hydrolyzed to yield construct 1 by treatment with lithium
hydroxide. Then this hexasaccharide intermediate was treated with a mixture of C5-epi and 2-
OST to yield a mixture of GlcA2S- and IdoA2S-intermediate (Supplementary Fig S1). An
anion-exchange HPLC method was used to isolate and purify these two products with a ratio of

6
1:9 (Fig 2A). After the purification, a GlcA2S- intermediate was obtained with the purity greater
than 93% as determined by anion-exchange HPLC (Fig 2B). Similarly, an IdoA2S-intermediate
was obtained with the purity greater than 96% (Fig 2C). The GlcA2S-intermediate was further
converted to construct 2 by detriflouroacetylation under an alkaline condition. Additional
modifications were made, including N-sulfation and 6-O-sulfation, yielded to construct 3 and 4,
respectively. Similarly, the IdoA2S-intermediate was transformed to construct 5, 6 and 7. 3-O-
sulfation was carried out on construct 7 to form construct 8. The purity analysis of each
synthesized oligosaccharide and mass spectrometric analyses are shown in the Supplementary
data (Supplementary Fig S2, and S3).
NMR analysis of hexasaccharides
The availability of high purity structurally defined hexasaccharides provided us sufficient
quantities for NMR analyses. The 1H-NMR spectra of GlcA-containing hexasaccharide
(construct 1) and GlcA2S-containing hexasaccharides, including construct 2, 3 and 4, are shown
in Fig 3; the 1H-NMR spectra of IdoA2S-containing hexasaccharides, including construct 5 to 8,
are shown in Fig 4. The proton signals from the saccharide residues are highly overlapped and
this makes the chemical shift assignment difficult. Therefore, 2D-homonuclear and -
heteronuclear correlation spectroscopy analyses (1H-1H COSY, TOCSY, 1H-13C HSQC, and
HMBC) were conducted; these analyses only partially resolved the overlapped signals. Thus, 2D
1H-13C HSQC-TOCSY NMR experiments (Fig 5) were employed to resolve these overlapped
signals. The full proton and carbon chemical shift assignments for constructs 1-8 are listed in
Table 1 and 2.

7
Characterization of GlcA2S-containing oligosaccharides
Here, we use construct 3 (with the structure of GlcNS-GlcA-GlcNS-GlcA2S-GlcNS-
GlcA-pnp) to demonstrate how the signals of GlcA2S were identified in the synthetic
oligosaccharides. In the 1H-NMR spectrum of construct 3, the signals from anomeric protons of
the three GlcNS units are located around 5.6 ppm and the signals from GlcA/GlcA2S are located
around 4.6 ppm (Fig 3). The ~5.3 ppm shift of I-1 is attributed to the effect of the nearby electron
withdrawing group, para-nitrophenyl (pnp). The 3JHH coupling constants of I-1, III-1, and V-1
were ~8 Hz (consistent with β linkages), and II-1, IV-1, and VI-1 were ~4 Hz (consistent with α
linkages). Other saccharide ring proton signals were highly overlapped and resonated between
3.2 and 4.1 ppm. The increased shift dispersion in 2D 1H-13C HSQC-TOCSY NMR, fully
resolved these signals so that assignments could be made (Fig 5). In addition, longer 1H-1H
COSY and 1H-13C HSQC experiments were run with more points in the indirect dimension for
better digital resolution which allowed more accurate identification of the proton and carbon
assignments (Table 1). The assignments for other GlcA2S-containing hexasaccharides were
conducted in a similar manner.
The 2D-1H-13C HMBC experiment confirmed the glycosidic linkages between each
saccharide unit through 3J correlations between the anomeric II-1, III-1, IV-1, V-1 or VI-1
protons show connectivity to the I-4, II-4, III-4, IV-4 or V-4 carbons, respectively (Fig 6). These
connectivities allowed sequencing of the hexasaccharides. In addition, the anomeric I-1 proton
shows a 3JCH-correlation with the quaternary carbon of the pnp tag. Furthermore, the II-6, IV-6
or VI-6 protons shows through-bond connectivity to II-4, IV-4 or VI-4 carbons, respectively,
supporting the assignments of glucosamine residues. The other GlcA2S-containing
hexasaccharides were characterized by following the same approach. In addition to 1H-1H COSY

8
and 1H-13C HSQC experiments, we focused on using 2D 1H-13C HSQC-TOCSY NMR to
identify the residue spin systems and 2D 1H-13C HMBC NMR to confirm the linkages (Fig 3,
Table 1).
Comparing the NMR spectra of construct 1 with construct 2 provides evidence of the
influence of 2-O-sulfation on the neighboring residues (Fig 3). The signals of III-1, III-2 and III-
3 from construct 2 are shifted +0.17, +0.73 and +0.17 ppm (to 4.70, 4.11, and 4.02 ppm),
respectively, due to the attached 2-O-sulfo group. Notably, the shift of proton II-1 (-0.12 ppm),
II-6b (+0.11 ppm), and VI-2 (-0.15 ppm) are also sensitive to 2-O-sulfation even though they are
distant from the site of substitution. The chemical shift changes observed for the II residue are
attributed to inter-residue influences between this additional 2-O-sulfo group and cationic
groups. The effect on VI-2 may be due to the interactions between the positive charge from the
unsubstituted amino group to the nearby 2-O-sulfo group. The ~8 Hz axial-axial proton coupling
constants of the III residue remains unchanged indicating that the GlcA2S conformation remains
predominately in the 4C1 form in Construct 2 and is the same as GlcA in Construct 1 (Sattelle,
Hansen et al. 2010).
To reveal the structural flexibility of GlcA2S and IdoA2S residues, we compared the
proton spectra of GlcA2S-containing hexasaccharides with IdoA2S-containing hexasaccharides,
focusing on the signals of the Residue III. The measurements of 3JHH coupling constant reflect
the change of saccharide conformation. The coupling constant (doublet, 8.0 Hz) of III-1 signal of
GlcA2S in construct 2 remains same as that in construct 3 and 4, suggesting that the
conformation of GlcA2S is relatively unaffected by the extra sulfo groups (Fig 3). This result is
consistent with that GlcA2S predominately maintain the 4C1 conformation as GlcA in construct 1
(Sattelle, Hansen et al. 2010). We observed that broad signals for IdoA2S (residue III) (Fig 4).

9
The addition of EDTA significantly sharpened the IdoA2S signals (Supplementary Fig S5). The
spectra showed the coupling constants of III-1 signals are altered in different IdoA2S-containing
constructs. Similarly, the coupling constants of III-5 signals are also sensitive to different
sulfation patterns, suggesting that conformational distribution of IdoA2S is affected by different
sulfation pattern. For example, the 3JHH coupling constants of III-1 (2.2 Hz) and III-5 (2.2 Hz)
signal in construct 5 are increased to (3.7 Hz) and (3.1 Hz) in construct 8 due to the presence of
additional sulfo groups. These changes are attributed to greater structural flexibility of IdoA2S
residue compared to GlcA2S in these constructs. This result is consistent with the previously
reported that IdoA2S equilibrates between 2S0 (skew boat) and 1C4 (chair) conformers (Ferro,
Provasoli et al. 1990, Sattelle, Hansen et al. 2010). Although the addition of EDTA did not
considerably change the chemical shifts of IdoA2S-containing hexasaccharides, some changes
were noted for construct 5. In construct 5 the F-2 (VI-2) proton signal shifted by 0.19 ppm with
addition of 3 mM EDTA under the same conditions (pD 7.0, 25°C). Similarly, the signal of F-
1(VI-1) and F-3 (VI-3) were altered as well (Supplementary Table S1).
The presence of GlcA2S or IdoA2S also impacted the chemical environment of
neighboring saccharide residues. We compared the signals from each proton of GlcA2S- and
IdoA2S-containing hexasaccharides (Table 3). As expected, signals from Residue III displayed
the largest difference in chemical shifts. Outside Residue III, significant differences in the
chemical shifts were observed for the protons of Residue II (II-1, II-3, II-6), and protons of
Residue IV (IV-1 and IV-5) (Table 3). For example, comparing the IV-1 proton in construct 2
and 5, the signals are shifted by 0.34 ppm. Similarly, between construct 3 and 6 the IV-1 proton
was shifted by 0.32 ppm and 0.22 ppm between construct 4 and 7, respectively. These data
suggest the presence of GlcA2S or IdoA2S has different effects on the chemical environment of

10
neighboring residues. Beyond Residue II and IV, chemical shift differences between GlcA2S-
and IdoA2S-hexasaccharides diminished significantly. Some differences in chemical shifts were
observed from Residue VI, an N-unsubstituted glucosamine residue in construct 2 and 5,
suggesting that the presence of a primary amino group (-NH2) of this end residue is sensitive to
GlcA2S vs IdoA2S residues.
Intramolecular effects by N-sulfation, 6-O-sulfation, and 3-O-sulfation
Comparing the chemical shifts of the signals from hexasaccharides bearing N-
unsubstituted versus N-sulfated Residue VI suggests the intramolecular interaction between the
N-sulfo groups and the GlcA residue at its reducing end. The chemical shifts of the V-3 signals
in construct 2 and 3 are shifted by +0.08 ppm; similarly, the signals of the V-3 protons in
construct 5 and 6 are shifted +0.10 ppm (Table 1 and 2). These effects may be attributed to the
function of the acidic sulfo group masking the basic amino group on the VI residue and altering
the chemical environment around the V residue. These observations are consistent with
molecular dynamics simulation results that showed an interaction between the N-sulfo group of
Residue VI and the V-3 position of Residue V as proposed by Langeslay and colleagues for
fondaparinux (Langeslay, Young et al. 2012). Another interesting observation is the impact of
N-sulfation on the chemical shift of VI-2 signals. In GlcA2S-containing hexasaccharides, N-
sulfation shifts the VI-2 signal to higher ppm (+0.06 ppm) due to the effects of de-shielding (less
local electron density). By contrast, in IdoA2S-containg hexasaccharides, the N-sulfation shifts
the VI-2 signal to lower ppm (-0.07 ppm) due to increased shielding (more electron density).
Such different effects suggest more complicated through space interactions between the
saccharide residues, which could involve in the participation of IdoA2S or GlcA2S residue;
however, the available techniques do not permit us to dissect these interactions.

11
As expected, 6-O-sulfation clearly shifts the proton resonances of Residues II, IV and VI
in both GlcA2S-containing and IdoA2S-containing hexasaccharides (Table 1 and 2). The
presence of 6-O-sulfation also affects the chemical environments of the neighboring residues.
For instance, Residue V in construct 4 is flanked by two GlcNS6S resides (Residue IV and VI),
the signal of V-1 is shifted by +0.08 ppm in comparison with this signal in construct 3.
Similarly, a chemical shift difference for the V-1 proton was also observed (+0.07 ppm) between
construct 6 and 7 (Table 1 and 2). These data suggest that the impact of 6-O-sulfation on the
neighboring residues can be observed in both GlcA2S-containing and IdoA2S-containing
hexasaccharides.
The availability of 3-O-sulfated construct 8 and non-3-O-sulfated construct 7
differentiates the impact of 3-O-sulfation on the distribution of IdoA2S residue conformers. The
3-O-sulfation on Residue IV alters the conformation of its reducing-end IdoA2S residue
(Residue III) as demonstrated by changes in the coupling constant of IdoA2S signals. For
example, the 3JHH coupling constants of III-1 and III-5 are larger in construct 8 (3.7 Hz, 3.1 Hz)
than in construct 7 (3.0 Hz, 2.7 Hz) (Supplementary Fig S5), suggesting an increased proportion
of 2S0 conformation for Residue III in construct 8. Indeed, Guerrini and colleagues demonstrated
that 3-O-sulfated glucosamine residue influences the IdoA2S residue conformational equilibrium
between 1C4 and 2S0 (Guerrini, Elli et al. 2013).
Contribution of GlcA2S and IdoA2S to the binding affinity of antithrombin
Heparin binds antithrombin to exert its anticoagulant activity. Within heparin a unique
pentasaccharide sequence, -GlcNS6S-GlcA-GlcNS3S6S-IdoA2S-GlcNS6S-, has been shown to
display high binding affinity. Previous work has suggested that IdoA2S in this pentasaccharide

12
domain contributes to the binding of antithrombin (Das, Mallet et al. 2001). To confirm this
conclusion, the binding affinity to antithrombin of the IdoA2S versus GlcA2S constructs was
determined. Both 3-O-[35S]sulfated GlcA2S-containing hexasaccharide and IdoA2S-containing
hexasaccharide were prepared for the analysis. The result showed that IdoA2S-containing
hexasaccharide binds to antithrombin with a dissociation constant (Kd) of 12 ± 3 nM, whereas no
binding affinity was measurable for GlcA2S-containing hexasaccharide. Our result confirmed
the essential role of IdoA2S in antithrombin binding.
Discussion
The GlcA2S residue is present in low abundance in heparin, and thus, very limited information
about the role of these rare saccharide residues has been obtained. In this work, we demonstrate
the synthesis and structural characterization of highly pure oligosaccharides containing GlcA2S.
The synthesis of three GlcA2S containing hexasaccharides was completed through a
chemoenzymatic approach. The substrate specificity of 2-OST modifies only one of the two
GlcA in the hexasaccharides, allowing us to position of GlcA2S residue at the desired site
(Supplemetary Fig S1). Although the synthesis of a number of HS hexasaccharides using a
chemical and chemoenzymatic approach has been reported (Petitou and van Boeckel 1992, Liu,
Xu et al. 2010), the synthesis of the GlcA2S-containing hexasaccharide variants described in this
research is novel.
The chemoenzymatic approach allows the construction of diverse hexasaccharide analogs
containing antithrombin-binding domain to confirm the contribution of IdoA2S or GlcA2S
residues to the binding affinity to antithrombin. Our data suggest that GlcA2S-containing
hexasaccharide does not bind to antithrombin despite the fact it carries 3-O- and 6-O-sulfo

13
groups, confirming the previously reported essential role of IdoA2S for high affinity binding to
antithrombin (Jin, Abrahams et al. 1997, Hricovini, Guerrini et al. 2001, Guerrini, Elli et al.
2010).
The NMR structural analysis data are consistent with the hypothesis that GlcA2S is a
relatively structural rigid unit in heparin and HS compared to IdoA2S. The 3JHH coupling
constants of GlcA ring protons in all constructs are about 8.0 Hz as reported previously as 4C1
form (Sattelle, Hansen et al. 2010). In our studies, the coupling constant of GlcA2S ring protons
in the synthesized hexasaccharide show equal values, supporting the conformation of GlcA2S
remains predominately in the 4C1 form as well. In contrast, the protons of IdoA2S display
smaller and various coupling constants in the constructs, consistent with the fact that the
conformations of the IdoA2S residue are in the equilibrium between 2S0 and 1C4 forms.
The 2S0 conformation of IdoA2S is known to contribute to antithrombin binding affinity
(Hricovini, Guerrini et al. 2001, Guerrini, Guglieri et al. 2008, Guerrini, Elli et al. 2010). Indeed,
IdoA2S-containing hexasaccharide displayed high affinity with a Kd value of 12 nM, while no
detectable antithrombin-binding affinity for GlcA2S-containing hexasaccharide was observed.
Importantly, the conformation of the IdoA2S residue can be affected by the protein binding as
well as interactions with neighboring sulfated saccharide residues. For example, binding of
fibroblast growth factor 1 (FGF1) requires a 1C4 form of IdoA2S that is flanked by GlcNS6S
residues (Faham, Hileman et al. 1996, DiGabriele, Lax et al. 1998, Kreuger, Salmivirta et al.
2001); however, in the absence of a nearby GlcNS6S residue, IdoA2S can rearrange its
conformation from a 2S0 to a 1C4 form which binds FGF2 (Guglier, Hricovini et al. 2008, Zhang,
McCallum et al. 2008). The IdoA2S conformation flexibility may play a role in orienting the
oligosaccharide for increased binding affinity to a specific protein.

14
Using highly purified GlcA2S- and IdoA2S-containing hexasaccharides synthesized by
the chemoenzymatic approach, we successful fully assign all the proton and carbon chemical
shifts of GlcA2S and IdoA2S-containing oligosaccharides through 1D and 2D NMR
experiments. We compared the assignments obtained from our study with the previously
reported values. The proton assignments of GlcA2S signals in this study are similar to the
previous reports, where the GlcA2S residue is flanked by GlcNS6S and GlcNS residue in a
tetrasaccharide (Yamada, Murakami et al. 1995, Limtiaco, Beni et al. 2011, Ozug, Wudyka et al.
2012, Guerrini, Elli et al. 2013). The anomeric carbon assignment of GlcA2S was also reported
(Guerrini and Torri 2008), and our result is consistent with the reported values. In those previous
reports, only partial assignments were achieved for GlcA2S residues. Here, we provide the full
assignments for the GlcA2S residue within different sulfated saccharide contexts. The
assignments for IdoA2S are consistent with previous reports; however an IdoA2S residue
flanked by two GlcNS3S6S residues show significant differences in the proton and carbon
assignments (Guerrini, Elli et al. 2013). The GlcNS3S6S residues may participate in the change
of chemical environment and conformation equilibrium of neighboring IdoA2S residue, causing
significant changes in chemical shifts.
Taking advantage of these chemical shift assignments, the presence of GlcA2S can be
easily distinguished from mixture of GlcA2S- and IdoA2S-containing oligosaccharides within
NMR spectra. 2D 1H-13C HSQC NMR is a powerful tool for fingerprint the GlcA2S residue, and
the 1H-13C correlation of III-1 and III-2 position signals from this residue is easily observable
(Supplemetary Fig S4). The cross peak of GlcA2S anomeric signal is also consistent to the signal
observed in enoxaparin (Guerrini and Torri 2008) or porcine intestinal heparin (Yamada,
Murakami et al. 1995), demonstrating the capability of using the 2D-NMR fingerprint method to

15
confirm the authenticity of heparin drugs. Very interestingly, the anomeric signals of GlcA2S
were reportedly difficult to identify in unfractionated heparin and low-molecular weight
heparins; however, such signals are visible in enoxaparin, a low-molecular weight heparin
manufactured by depolymerized unfractionated heparin under basic conditions (Guerrini and
Torri 2008, Keire, Buhse et al. 2013).
Since the 2008 contaminated heparin crisis, NMR analysis of heparin drug has become
part of the heparin sodium US Pharmacopeia monograph. Because heparin is a heterogeneous
polysaccharide mixture, the assignment of all of the signals is difficult. To this end, acquiring the
NMR spectra of highly purified heparin oligosaccharide standards provides a map for assigning
highly complex heparin spectra, especially those signals from rare components. Our study
should assist NMR analysis of heparin to safeguard the heparin drug.
Materials and methods
Preparation of enzymes and cofactors
E. coli was used for expression of all enzymes: KfiA, pmHS2, NST, C5-epi, 2-OST, 6-OST-1, 6-
OST-3, and 3-OST-1 for chemoenzymatic synthesis. Purification was completed by appropriate
affinity chromatography as described previously (Xu, Moon et al. 2008, Liu, Xu et al. 2010). 3’-
phosphoadenosine 5’-phophosulfate (PAPS) is the sulfo donor, was prepared from sodium
sulfate and ATP utilizing ATP-sulfurylase and adenosine phosphokinase (Zhou, Chandarajoti et
al. 2011). UDP-GlcNTFA was prepared from glucosamine, followed by reaction with S-
ethyltrifluorothioacetate and protocol (Liu, Xu et al. 2010). GlcNTFA was converted to
GlcNTFA-1-phosphate by N-acetylhexasamine-1-kinase (Zhao, Guan et al. 2010) and
subsequently transformed to UDP-GlcNTFA using glucosamine-1-phosphate

16
acetyltransferase/N-acetylglucosamine-1-phosphate uridyltransferase (GlmU) (Liu, Xu et al.
2010).
Synthesis of hexasaccharide intermediate
The elongation started from a commercial material, GlcA-pnp, followed by elongation of
GlcNTFA three times, elongation of GlcA two times, and one detrifluoroacetylation and one N-
sulfation to make hexasaccharide intermediates (Supplementary Fig S1). To introduce a
GlcNTFA residue, GlcA-pnp (1.2 mM) was incubated with KfiA (20 µg/ml) in a buffer
containing Tris (25 mM, pH 7.5), MnCl2 (15 mM) and UDP-GlcNTFA (1.5 mM), at room
temperature overnight. To introduce a GlcA residue, disaccharide substrate, GlcNTFA-GlcA-
pnp (1.2 mM), was incubated with pmHS2 (20 µg/ml) in a buffer containing Tris (25 mM, pH
7.5), MnCl2 (15 mM) and UDP-GlcA (1.5 mM), at room temperature overnight. C18 column
(0.75 × 20 cm; Biotage) was used for purification with gradient elution (0-100% acetonitrile in
H2O, 0.1% TFA, 2 ml/min in 60 min). The eluted products were monitored by UV absorbance at
310 nm and characterized by ESI-MS. The additional GlcNTFA and GlcA elongation was
repeated under the same condition. Detrifluoroacetylation was completed using 0.1M LiOH at
0°C for two hours. The products were analysis by PAMN-HPLC (polyamine-based high
performance liquid chromatography) and electrospray ionization mass spectrometry (ESI-MS).
Afterwards, N-sulfation was performed within pH 7.0 MES (50 mM), N-sulfotransferase (10
µg/ml), and PAPS (0.5 mM, 1.5 equivalent of substrate amount) at 37°C overnight. The sulfated
products were purified by Q-Sepharose column (15 × 200 mm, GE Health Sciences) with linear
gradient elution (20–100% 1 M NaCl in 20 mM NaOAc, pH 5.0, flow rate 2 ml/min) and
followed by dialysis to obtain highly pure hexasaccharide intermediates.

17
Synthesis of oligosaccharide constructs 1 to 8
Detrifluoroacetylation of hexasaccharide intermediate could get construct 1. Construct 2 and 5
were acquired by epimerization and 2-O-sulfation within the solution contained MES (50 mM,
pH 7.0), CaCl2 (2 mM), C5-epi (10 µg/ml), 2-OST (10 µg/ml), PAPS (0.2 mM), and
hexasaccharide intermediate (0.13 mM) at 37oC overnight. Following the procedure of N-
sulfation as described above could synthesize construct 3 and 6. Construct 4 was obtained by 6-
O-sulfation using PAPS (800 µM), MES (50 mM, pH 7.0), 6-OST-1 (0.6 mg/ml), 6-OST-3 (0.6
mg/ml) overnight at 37°C. Construct 7 was obtained by 6-O-sulfation using PAPS (400 µM),
MES (50 mM, pH 7.0), 6-OST-1 (0.1 mg/ml), 6-OST-3 (0.1 mg/ml) overnight at 37°C. 3-O-
sulfation was performed to synthesize construct 8 under the condition of MES (50 mM, pH 7.0),
MnCl2 (10 mM), MgCl2 (5 mM), 3-OST-1 (0.03 µg/ml), PAPS (600 µM) at 37oC overnight.
The purification of construct 1 to 8 is essentially the same as hexasaccharide intermediate as
described above.
Structural characterization of oligosaccharides
The purity of products was analyzed by polyamine-based anion exchange (PAMN)-HPLC or
DEAE-HPLC (Fig 2, S2 and S3). The elution conditions were described in previous literature
(Liu, Xu et al. 2010).
Mass spectrometric analyses were performed on all of the intermediates and constructs by
ESI-MS (Thermo LCQ-Deca) with direct sample infusion from a syringe pump (Harvard
Apparatus). Each sample was diluted in 1:9 MeOH/H2O and analyzed in negative ionization
mode (3.5 KV, 150 °C). Xcalibur 1.3 was utilized to acquire and process full scanned MS with 1
X 107 automatic gain control.

18
The structure of the oligosaccharides constructs and intermediates were analyzed by NMR
experiments, including 1D- (“s2pul” pulse sequence 1H or 13C), 2D- (1H-1H “gCOSY” pulse
sequence COSY, “TOCSY” pulse sequence TOCSY, 1H-13C “gHSQC” pulse sequence HSQC,
“gHMBC” pulse sequence HMBC, and “gHSQCTOXY” pulse sequence HSQC-TOCSY) NMR.
NMR experiments were performed at 298 K on a Varian Inova 500 MHz spectrometer equipped
with 5mm triple resonance XYZ or broadband PFG probe and processed by VnmrJ 2.2D
software. Chemical shifts are referenced to external 2,2-dimethyl-2-silapentane-5-sulfonate
sodium salt (DSS, Sigma, Co.). Deuterated EDTA (Sigma, Co.) was added to resolve the
coupling constants of IdoA2S signals. Samples (5.0 to 10.0 mg) were each dissolved in 0.5 mL
D2O (99.994%, Sigma, Co.) and lyophilized three times to remove the exchangeable protons.
The samples were re-dissolved in 0.5 mL D2O and transferred to NMR microtubes (OD 5 mm,
Norrell). 1D 1H-NMR experiments were performed with 256 scans and an acquisition time of
768 msec. 1D 13C-NMR experiments were performed with 40000 scans, 1.0 sec relaxation delay,
and an acquisition time of 1000 msec. 2D (1H-1H COSY, TOCSY, 1H-13C HSQC, HSQC-
TOCSY) spectra were recorded with GARP carbon decoupling, a 1.5 sec relaxation delay, a 204
msec acquisition time with 500 increments for 48 scans. Forty eight steady state scans were used
prior to the start of acquisition. 2048 points were collected in f2. Gradients were used for
coherence transfer selection with sensitivity enhancement in the HSQC experiment. 13C
transmitter offset was set at 77.0 ppm. The echo-anti-echo phase cycling scheme was used. The
polarization transfer delay was set with a 1JC-H coupling value of 147Hz. 2D 1H-13C HMBC
experiments were performed with 72 scans, 1.5 sec relaxation delay, and a 204 msec acquisition
time. The delay for evolution of long-range couplings was set with Jlr of 7.6 Hz.

19
Determination of the binding affinity of oligosaccharides to antithrombin (AT)
35S-labeled 3-O-sulfated constructs were prepared via 3-O-sulfation as described previously,
using [35S]PAPS (0.5 nmol, specific activity 2.2 × 104 cpm/pmol) and oligosaccharide precursor
(5 µg) in total 500 l at 37°C overnight. The products were purified by DEAE-HPLC column.
Affinity co-electrophoresis (Lee and Lander 1991) was used for measurement of dissociation
constant (Kd) of oligosaccharides. 35S-labeled 3-O-sulfated constructs (1500-2500 cpm) were
loaded for lanes with AT concentration at 0, 8, 16, 32, 60, 120, 250, 500, and 1000 nM. The
agarose gel was run with 300 mA electrophoresis for two hours, and was dried at room
temperature overnight. The dried gel was analyzed by PhosphoImager (Amersham Biosciences,
Storm 860). The retardation coefficient was calculated at R = (M0-M)/M0, where M0 is the
mobility of the oligosaccharide flow through in the lane without AT, and M is the mobility of the
sample in an individual lane. The retardation coefficient was then plotted against the retardation
coefficient divided by its respective AT concentration. The constant -1/Kd is the slope of the
trendline.

20
References
Ahsan A, Jeske W, Hoppensteadt D, Lormeau JC, Wolf H, Fareed J. 1995. Molecular profiling
and weight determination of heparins and depolymerized heparins. J Pharm Sci. 84(6): 724-727.
Bishop JR, Schuksz M, Esko JD. 2007. Heparan sulphate proteoglycans fine-tune mammalian
physiology. Nature. 446(7139): 1030-1037.
Carlsson P, Presto J, Spillmann D, Lindahl U, Kjellen L. 2008. Heparin/heparan sulfate
biosynthesis: processive formation of N-sulfated domains. J Biol Chem. 283(29): 20008-20014.
Casu B. 1985. Structure and biological activity of heparin. Adv Carbohydr Chem Biochem. 43:
51-134.
Das SK, Mallet JM, Esnault J, Driguez PA, Duchaussoy P, Sizun P, Herault JP, Herbert JM,
Petitou M, Sinay P. 2001. Synthesis of conformationally locked L-iduronic acid derivatives:
direct evidence for a critical role of the skew-boat 2S0 conformer in the activation of
antithrombin by heparin. Chemistry. 7(22): 4821-4834.
DiGabriele AD, Lax I, Chen DI, Svahn CM, Jaye M, Schlessinger J, Hendrickson WA. 1998.
Structure of a heparin-linked biologically active dimer of fibroblast growth factor. Nature.
393(6687): 812-817.
Faham S, Hileman RE, Fromm JR, Linhardt RJ, Rees DC. 1996. Heparin structure and
interactions with basic fibroblast growth factor. Science. 271(5252): 1116-1120.
Ferro DR, Provasoli A, Ragazzi M, Casu B, Torri G, Bossennec V, Perly B, Sinay P, Petitou M,
Choay J. 1990. Conformer populations of L-iduronic acid residues in glycosaminoglycan
sequences. Carbohydr Res. 195(2): 157-167.
Feyerabend TB, Li JP, Lindahl U, Rodewald HR. 2006. Heparan sulfate C5-epimerase is
essential for heparin biosynthesis in mast cells. Nat Chem Biol. 2(4): 195-196.
Griffin CC, Linhardt RJ, Van Gorp CL, Toida T, Hileman RE, Schubert RL, 2nd, Brown SE.
1995. Isolation and characterization of heparan sulfate from crude porcine intestinal mucosal
peptidoglycan heparin. Carbohydr Res. 276(1): 183-197.
Guerrini M, Elli S, Gaudesi D, Torri G, Casu B, Mourier P, Herman F, Boudier C, Lorenz M,
Viskov C. 2010. Effects on molecular conformation and anticoagulant activities of 1,6-
anhydrosugars at the reducing terminal of antithrombin-binding octasaccharides isolated from
low-molecular-weight heparin enoxaparin. J Med Chem. 53(22): 8030-8040.
Guerrini M, Elli S, Mourier P, Rudd TR, Gaudesi D, Casu B, Boudier C, Torri G, Viskov C.
2013. An unusual antithrombin-binding heparin octasaccharide with an additional 3-O-sulfated
glucosamine in the active pentasaccharide sequence. Biochem J. 449(2): 343-351.
Guerrini M, Guglieri S, Casu B, Torri G, Mourier P, Boudier C, Viskov C. 2008. Antithrombin-
binding octasaccharides and role of extensions of the active pentasaccharide sequence in the
specificity and strength of interaction. Evidence for very high affinity induced by an unusual
glucuronic acid residue. J Biol Chem. 283(39): 26662-26675.
Guerrini M, Torri G (2008). NMR Spectroscopy in Pharmaceutical Analysis. Holzgrabe U, Diehl
B and Wawer I: 407-428.
Guglier S, Hricovini M, Raman R, Polito L, Torri G, Casu B, Sasisekharan R, Guerrini M. 2008.
Minimum FGF2 binding structural requirements of heparin and heparan sulfate oligosaccharides
as determined by NMR spectroscopy. Biochemistry. 47(52): 13862-13869.

21
Hricovini M, Guerrini M, Bisio A, Torri G, Petitou M, Casu B. 2001. Conformation of heparin
pentasaccharide bound to antithrombin III. Biochem J. 359(Pt 2): 265-272.
Jin L, Abrahams JP, Skinner R, Petitou M, Pike RN, Carrell RW. 1997. The anticoagulant
activation of antithrombin by heparin. Proc Natl Acad Sci U S A. 94(26): 14683-14688.
Keire DA, Buhse LF, Al-Hakim A. 2013. Characterization of currently marketed heparin
products: composition analysis by 2D-NMR. Analytical Methods. 5(12): 2984-2994.
Kreuger J, Salmivirta M, Sturiale L, Gimenez-Gallego G, Lindahl U. 2001. Sequence analysis of
heparan sulfate epitopes with graded affinities for fibroblast growth factors 1 and 2. J Biol Chem.
276(33): 30744-30752.
Langeslay DJ, Young RP, Beni S, Beecher CN, Mueller LJ, Larive CK. 2012. Sulfamate proton
solvent exchange in heparin oligosaccharides: evidence for a persistent hydrogen bond in the
antithrombin-binding pentasaccharide Arixtra. Glycobiology. 22(9): 1173-1182.
Lee MK, Lander AD. 1991. Analysis of affinity and structural selectivity in the binding of
proteins to glycosaminoglycans: development of a sensitive electrophoretic approach. Proc Natl
Acad Sci U S A. 88(7): 2768-2772.
Limtiaco JF, Beni S, Jones CJ, Langeslay DJ, Larive CK. 2011. The efficient structure
elucidation of minor components in heparin digests using microcoil NMR. Carbohydr Res.
346(14): 2244-2254.
Linhardt RJ. 2003. 2003 Claude S. Hudson Award address in carbohydrate chemistry. Heparin:
structure and activity. J Med Chem. 46(13): 2551-2564.
Liu H, Zhang Z, Linhardt RJ. 2009. Lessons learned from the contamination of heparin. Nat
Prod Rep. 26(3): 313-321.
Liu R, Xu Y, Chen M, Weiwer M, Zhou X, Bridges AS, DeAngelis PL, Zhang Q, Linhardt RJ,
Liu J. 2010. Chemoenzymatic design of heparan sulfate oligosaccharides. J Biol Chem. 285(44):
34240-34249.
McMahon AW, Pratt RG, Hammad TA, Kozlowski S, Zhou E, Lu S, Kulick CG, Mallick T, Dal
Pan G. 2010. Description of hypersensitivity adverse events following administration of heparin
that was potentially contaminated with oversulfated chondroitin sulfate in early 2008.
Pharmacoepidemiol Drug Saf. 19(9): 921-933.
Ozug J, Wudyka S, Gunay NS, Beccati D, Lansing J, Wang J, Capila I, Shriver Z, Kaundinya
GV. 2012. Structural elucidation of the tetrasaccharide pool in enoxaparin sodium. Anal Bioanal
Chem. 403(9): 2733-2744.
Petitou M, van Boeckel CA. 1992. Chemical synthesis of heparin fragments and analogues.
Fortschr Chem Org Naturst. 60: 143-210.
Sattelle BM, Hansen SU, Gardiner J, Almond A. 2010. Free energy landscapes of iduronic acid
and related monosaccharides. J Am Chem Soc. 132(38): 13132-13134.
Sommers CD, Ye H, Kolinski RE, Nasr M, Buhse LF, Al-Hakim A, Keire DA. 2011.
Characterization of currently marketed heparin products: analysis of molecular weight and
heparinase-I digest patterns. Anal Bioanal Chem. 401(8): 2445-2454.
Victor XV, Nguyen TK, Ethirajan M, Tran VM, Nguyen KV, Kuberan B. 2009. Investigating the
elusive mechanism of glycosaminoglycan biosynthesis. J Biol Chem. 284(38): 25842-25853.
Xu D, Moon AF, Song D, Pedersen LC, Liu J. 2008. Engineering sulfotransferases to modify
heparan sulfate. Nat Chem Biol. 4(3): 200-202.
Xu Y, Masuko S, Takieddin M, Xu H, Liu R, Jing J, Mousa SA, Linhardt RJ, Liu J. 2011.
Chemoenzymatic synthesis of homogeneous ultralow molecular weight heparins. Science.
334(6055): 498-501.

22
Yamada S, Murakami T, Tsuda H, Yoshida K, Sugahara K. 1995. Isolation of the Porcine
Heparin Tetrasaccharides with Glucuronate 2-O-Sulfate - Heparinase Cleaves Glucuronate 2-O-
Sulfate-Containing Disaccharides in Highly Sulfated Blocks in Heparin. Journal of Biological
Chemistry. 270(15): 8696-8705.
Yamada S, Murakami T, Tsuda H, Yoshida K, Sugahara K. 1995. Isolation of the porcine
heparin tetrasaccharides with glucuronate 2-O-sulfate. Heparinase cleaves glucuronate 2-O-
sulfate-containing disaccharides in highly sulfated blocks in heparin. J Biol Chem. 270(15):
8696-8705.
Zhang Z, McCallum SA, Xie J, Nieto L, Corzana F, Jimenez-Barbero J, Chen M, Liu J, Linhardt
RJ. 2008. Solution structures of chemoenzymatically synthesized heparin and its precursors. J
Am Chem Soc. 130(39): 12998-13007.
Zhao G, Guan W, Cai L, Wang PG. 2010. Enzymatic route to preparative-scale synthesis of
UDP-GlcNAc/GalNAc, their analogues and GDP-fucose. Nat Protoc. 5(4): 636-646.
Zhou X, Chandarajoti K, Pham TQ, Liu R, Liu J. 2011. Expression of heparan sulfate
sulfotransferases in Kluyveromyces lactis and preparation of 3'-phosphoadenosine-5'-
phosphosulfate. Glycobiology. 21(6): 771-780.

23
Figure Legends
Figure 1. The synthetic hexasaccharide constructs 1 to 8. A total of eight synthetic
hexasaccharides were prepared in this study.
Figure 2. Two major products of epimerization and 2-O-sulfation. Panel A shows the anion-
exchange HPLC chromatogram after epimerization/2-O-sulfation of hexasaccharide
intermediate. The product contains 10% GlcA2S-intermediate and 90% IdoA2S-intermediate.
Panel B and C show the anion-exchange HPLC chromatogram of GlcA2S-intermediate and
IdoA2S-intermediate after purification. The purity is greater than 93% and 96%, respectively.
The chemical structure of GlcA2S-intermediate and IdoA2S-intermediate are shown on right of
Panel B and C.
Figure 3. 1H-NMR spectra of construct 1, 2, 3 and 4. Panel A, B, C and D show the 1H-NMR
spectrum of construct 1, 2, 3 and 4, respectively. The signals of anomeric protons and some non-
overlapping protons are indicated. The chemical structure of construct is shown on top of each
panel. The 1H-NMR spectra represent the range between 3.1 and 5.7 ppm.
Figure 4. 1H-NMR spectra of construct 5, 6, 7 and 8. Panel A, B, C and D show the 1H-NMR
spectrum of construct 5, 6, 7 and 8, respectively. The signals of anomeric protons and some non-
overlapping protons are indicated. The chemical structure of construct is shown on top of each
panel. The 1H-NMR spectra represent the range between 3.1 and 5.7 ppm.
Figure 5. 2D 1H-13C HSQC-TOCSY NMR spectrum of construct 3. 2D 1H-13C HSQC-TOCSY
NMR experiment was employed to characterize the structure of construct 3. The signals of
protons are indicated. The signals are connected with lines identifying correlated protons on each
saccharide residue. The chemical structure of construct 3 is shown on top of the Figure.

24
Figure 6. 1H-13C HMBC key correlations of construct 3. 2D 1H-13C HMBC NMR experiment
was employed to confirm the glycosidic linkages between each saccharide unit of construct 3.
The signals of anomeric protons are indicated. The peaks are connected with lines identifying
correlated carbon and proton. The chemical structure of construct 3 is shown on top of Figure.
Some other key 3JCH-correlations are shown as dashed lines on the chemical structure.

Table 1. 1H/13C NMR chemical shift assignments (in p.p.m.) of construct 1 – 4 (25°C, D2O, pD 7.0)
Construct 1 (GlcNH2-GlcA-GlcNS-GlcA-GlcNS-GlcA-pnp)
1
2
3
4
5
6a
6b
I
5.29/102.0
3.69/75.1
3.28/60.4
3.38/75.4
3.25/60.4
3.40/75.9
3.30/56.9
3.97/78.6
3.69/72.1
3.85/78.7
3.68/72.1
3.75/78.6
3.84/72.1
3.89/79.0
3.69/80.6
3.79/79.0
3.68/80.5
3.82/78.7
3.47/71.8
4.04/79.1
3.81/73.3
3.84/78.7
3.79/73.2
3.87/78.6
3.71/75.1
II 5.63/99.9
III 4.53/104.8
IV 5.60/99.9
3.80/62.1
3.80/62.1
3.79/62.4
3.85/62.1
3.85/62.1
3.80/62.4
V
4.50/104.8
VI 5.68/97.9
Construct 2 (GlcNH2-GlcA-GlcNS-GlcA2S-GlcNS-GlcA-pnp)
1
2
3
4
5
6a
6b
I
5.27/102.0
3.70/75.1
3.25/60.6
4.11/82.5
3.22/60.3
3.39/75.9
3.15/57.2
3.94/78.2
3.73/72.2
4.02/77.9
3.67/72.4
3.75/78.6
3.76/73.3
3.88/81.0
3.65/81.9
3.84/78.9
3.67/80.6
3.77/78.7
3.44/72.0
4.00/79.8
3.82/73.4
3.85/79.0
3.81/73.3
3.83/78.6
3.73/75.0
II 5.51/100.7
III 4.70/103.3
IV 5.64/99.9
3.82/62.1
3.84/62.1
3.78/62.6
3.96/62.1
3.85/62.1
3.79/62.6
V
4.49/104.9
VI 5.61/98.9
Construct 3 (GlcNS-GlcA-GlcNS-GlcA2S-GlcNS-GlcA-pnp)
1
2
3
4
5
6a
6b
I
5.31/102.0
3.70/75.0
3.26/60.5
4.12/82.4
3.23/60.4
3.39/75.5
3.21/60.8
3.95/78.1
3.73/72.0
4.01/77.7
3.67/72.3
3.83/78.6
3.58/73.9
3.89/80.5
3.66/81.7
3.88/78.8
3.67/80.6
3.79/79.0
3.45/72.3
4.11/78.9
3.78/73.5
3.90/78.6
3.77/73.4
3.86/78.8
3.66/74.4
II 5.54/100.7
III 4.71/103.2
IV 5.62/100.1
3.81/62.0
3.83/62.1
3.76/62.8
3.96/62.0
3.86/62.1
3.78/62.8
V
4.52/104.9
VI 5.61/100.1
Construct 4 (GlcNS6S-GlcA-GlcNS6S-GlcA2S-GlcNS6S-GlcA-pnp)
1
2
3
4
5
6a
6b
I
5.31/102.0
3.72/75.1
3.30/60.5
4.14/82.4
3.25/60.2
3.38/75.6
3.24/60.7
3.96/78.2
3.73/72.0
3.99/77.8
3.69/72.3
3.85/78.7
3.61/73.8
3.91/81.4
3.73/80.3
3.86/79.3
3.73/79.7
3.81/79.6
3.56/71.7
4.12/79.2
3.98/71.7
3.87/79.1
3.99/71.5
3.87/79.0
3.86/72.5
II 5.51/101.0
III 4.76/102.7
IV 5.62/100.2
4.18/68.5
4.19/68.5
4.15/69.0
4.54/68.5
4.43/68.5
4.33/69.0
V
4.60/104.6
VI 5.59/100.3

Table 2. 1H/13C NMR chemical shift assignments (in p.p.m.) of construct 5 – 8 (25°C, D2O, pD 7.0)
Construct 5 (GlcNH2-GlcA-GlcNS-IdoA2S-GlcNS-GlcA-pnp)
1
2
3
4
5
6a
6b
I
5.27/102.0
3.69/75.1
3.25/61.0
4.30/77.6
3.23/60.5
3.41/75.9
3.29/56.9
3.95/78.5
3.65/72.5
4.22/70.7
3.71/72.2
3.74/78.6
3.83/72.2
3.87/79.8
3.70/79.7
4.03/78.4
3.66/80.5
3.83/78.7
3.47/71.9
4.00/79.6
3.79/73.8
4.82/71.1
3.90/73.3
3.83/78.8
3.72/75.0
II 5.58/100.3
III 5.23/101.8
IV 5.30/99.8
3.83/62.4
3.84/62.3
3.78/62.5
3.84/62.4
3.86/62.3
3.79/62.5
V
4.50/104.8
VI 5.68/97.9
Construct 6 (GlcNS-GlcA-GlcNS-IdoA2S-GlcNS-GlcA-pnp)
1
2
3
4
5
6a
6b
I
5.28/102.0
3.70/75.1
3.26/61.0
4.32/77.2
3.25/60.4
3.39/75.5
3.22/60.8
3.96/78.5
3.66/72.4
4.24/70.3
3.71/72.1
3.84/78.6
3.59/73.9
3.88/79.7
3.70/79.8
4.03/78.4
3.69/80.6
3.78/78.9
3.45/72.4
4.02/79.4
3.80/73.8
4.92/70.7
3.85/73.4
3.82/79.1
3.68/74.4
II 5.59/100.3
III 5.26/101.8
IV 5.30/100.0
3.84/62.3
3.87/62.1
3.77/62.9
3.84/62.3
3.87/62.1
3.77/62.9
V
4.53/104.9
VI 5.62/100.0
Construct 7 (GlcNS6S-GlcA-GlcNS6S-IdoA2S-GlcNS6S-GlcA-pnp)
1
2
3
4
5
6a
6b
I
5.29/102.0
3.71/75.1
3.30/60.7
4.32/78.3
3.28/60.2
3.38/75.5
3.25/60.6
3.95/78.5
3.67/72.4
4.22/71.5
3.69/72.1
3.84/78.6
3.61/73.8
3.90/79.7
3.77/78.5
4.09/78.8
3.74/79.7
3.78/79.4
3.56/71.7
4.04/79.3
3.95/71.9
4.88/71.7
4.05/71.6
3.82/79.0
3.87/72.5
II 5.60/100.3
III 5.24/101.8
IV 5.40/99.6
4.22/69.0
4.22/68.6
4.15/69.0
4.38/69.0
4.44/68.6
4.33/69.0
V
4.60/104.5
VI 5.61/100.3
Construct 8 (GlcNS6S-GlcA-GlcNS3S6S-IdoA2S-GlcNS6S-GlcA-pnp)
1
2
3
4
5
6a
6b
I
5.28/102.0
3.70/75.1
3.29/60.8
4.31/79.6
3.43/59.4
3.39/75.5
3.23/60.7
3.96/78.5
3.67/72.4
4.17/72.8
4.37/78.8
3.81/78.2
3.60/73.9
3.90/80.0
3.78/79.7
4.14/78.5
3.96/75.7
3.82/79.6
3.56/71.7
4.02/79.4
3.96/72.0
4.83/72.8
4.14/72.3
3.82/78.8
3.86/72.5
II 5.58/100.4
III 5.23/102.2
IV 5.49/98.9
4.22/69.0
4.24/68.7
4.14/69.0
4.43/69.0
4.46/68.7
4.34/69.0
V
4.61/103.9
VI 5.60/100.3

Table 3. 1H NMR chemical shift difference (Δδ) by subtracting GlcA2S- from IdoA2S-containing
hexasaccharides†
Construct 2 vs. construct 5
1
2
3
4
5
6a
6b
I
II [+0.07]
III [+0.53]
IV [-0.34]
V
[-0.08]
[+0.20]
[-0.12]
[+0.19]
[+0.14]
[+0.19]
[+0.97]
[+0.09]
VI [+0.07]
[+0.07]
Construct 3 vs. construct 6
1
2
3
4
5
6a
6b
I
[-0.09]
II [+0.05]
III [+0.55]
IV [-0.32]
V
[-0.07]
[+0.23]
[-0.12]
[+0.20]
[+0.15]
[+1.02]
[+0.08]
VI
Construct 4 vs. construct 7
1
2
3
4
5
6a
6b
I
[-0.08]
II [+0.09]
III [+0.48]
IV [-0.22]
V
[-0.06]
[+0.23]
[-0.16]
[+0.18]
[+0.23]
[+1.01]
[+0.06]
VI
† Listed values represent chemical shift difference ≥ 0.05 ppm

Figure 1. The synthetic hexasaccharide constructs 1 to 8
a t T O B B E k o n o m i v e T e k n o l o j i h U t t n p i : v / / e g r l s y i c t e o s b i . o n x f M o r a d D y j o 7 w u , r n n 2 l a 0 o l 1 a s 4 d . o e r d g / f r o m

A
90%
10%
IdoA2S‐intermediate
GlcA2S‐intermediate
GlcA2S‐intermediate
B
C
IdoA2S‐intermediate
Retention Time (min)
Figure 2. Two major products of epimerization and 2‐O‐sulfation
a t T O B B E k o n o m i v e T e k n o l o j i h U t t n p i : v / / e g r l s y i c t e o s b i . o n x f M o r a d D y j o 7 w u , r n n 2 l a 0 o l 1 a s 4 d . o e r d g / f r o m

VI3
III3
VI
V
IV
III
II
I
A
V2 III2
II1
III1V1
I3
I1
I5
VI1 IV1
VI4
VI2 II2 IV2
B
VI3
III3
I5 I3
I1
VI1
IV1
II1
III1
V1
VI4 V2
III2
II2 IV2 VI2
C
I3
IV1 VI1
I5
III2
I1
VI2
IV2
VI3
III1
V1
V2
VI4
III3
II1
II2
D
I3
I1
III2I5
IV2 VI2
II2
VI4
VI3
IV1 VI1
II1
V1
III3
III1
II6
VI6
V2
IV6
Figure 3. 1H‐NMR spectra of construct 1, 2, 3 and 4
a t T O B B E k o n o m i v e T e k n o l o j i h U t t n p i : v / / e g r l s y i c t e o s b i . o n x f M o r a d D y j o 7 w u , r n n 2 l a 0 o l 1 a s 4 d . o e r d g / f r o m