A Synthetic MUC1 Glycopeptide Bearing βGalNAc-Thr as a Tn Antigen Isomer Induces the Production of Antibodies against Tumor Cells

Article abstract of DOI:10.1002/cbic.201600473

This study presents the synthesis of the novel protected O-glycosylated amino acid derivatives 1 and 2, containing βGalNAc-SerOBn and βGalNAc-ThrOBn units, respectively, as mimetics of the natural Tn antigen (αGalNAc-Ser/Thr), along with the solid-phase assembly of the glycopeptides NHAcSer-Ala-Pro-Asp-Thr[αGalNAc]-Arg-Pro-Ala-Pro-Gly-BSA (3-BSA) and NHAcSer-Ala-Pro-Asp-Thr[βGalNAc]-Arg-Pro-Ala-Pro-Gly-BSA (4-BSA), bearing αGalNAc-Thr or βGalNAc-Thr units, respectively, as mimetics of MUC1 tumor mucin glycoproteins. According to ELISA tests, immunizations of mice with βGalNAc-glycopeptide 4-BSA induced higher sera titers (1:320 000) than immunizations with αGalNAc-glycopeptide 3-BSA (1:40 000). Likewise, flow cytometry assays showed higher capacity of the obtained anti-glycopeptide 4-BSA antibodies to recognize MCF-7 tumor cells. Cross-recognition between immunopurified anti-βGalNAc antibodies and αGalNAc-glycopeptide and vice versa was also verified. Lastly, molecular dynamics simulations and surface plasmon resonance (SPR) showed that βGalNAc-glycopeptide 4 can interact with a model antitumor monoclonal antibody (SM3). Taken together, these data highlight the improved immunogenicity of the unnatural glycopeptide 4-BSA, bearing βGalNAc-Thr as Tn antigen isomer.

Full text of DOI:10.1002/cbic.201600473

Accepted Article
Title: Synthetic MUC1 glycopeptide bearing βGalNAc-Thr as Tn
antigen isomer induces antibodies production against tumor cells
Authors: Vanessa Leiria Campo, Thalita Riul, Leandro Bortot, Maristela
Martins-Teixeira, Marcelo Marchiori, Emanuela Iaccarino,
Menotti Ruvo, Marcelo Dias-Baruffi, and Ivone Carvalho
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemBioChem 10.1002/cbic.201600473
Link to VoR: http://dx.doi.org/10.1002/cbic.201600473
A Journal of
www.chembiochem.org

10.1002/cbic.201600473
ChemBioChem
Synthetic MUC1 glycopeptide bearing βGalNAc-Thr as Tn antigen isomer induces
antibodies production against tumor cells
Vanessa Leiria Campo,*1a Thalita B. Riul,a Leandro Oliveira Bortot,a Maristela B. Martins-Teixeira,a
Marcelo Fiori Marchiori,a Emanuela Iaccarino,b,c Menotti Ruvo,b Marcelo Dias-Baruffia and Ivone
Carvalhoa
a. Faculdade de Ciências Farmacêuticas de Ribeirão Preto, USP, Av. Café S/N, CEP 14040-903,
Ribeirão Preto, SP, Brazil, E-mail: vlcampo@fcfrp.usp.br
b. Istituto di Biostrutture e Bioimmagini, CNR, via Mezzocannone, 16, 80134, Napoli,
Italy
c. Second University of Naples, via Vivaldi, 43, 81100, Caserta, Italy
Abstract
This study presents the synthesis of the novel protected O-glycosyl amino acids βGalNAc-SerOBn (1)
and βGalNAc-ThrOBn (2) as mimetics of the natural Tn antigen (αGalNAc-Ser/Thr), along with the
solid-phase assembly of the glycopeptides NHAcSer-Ala-Pro-Asp-Thr[αGalNAc]-Arg-Pro-Ala-Pro-Gly-
BSA (3–BSA) and NHAcSer-Ala-Pro-Asp-Thr[βGalNAc]-Arg-Pro-Ala-Pro-Gly-BSA (4–BSA), bearing
αGalNAc-Thr or βGalNAc-Thr, as mimetics of MUC1 tumor mucin glycoproteins. According to ELISA
tests, mice immunizations with βGalNAc-glycopeptide 4–BSA induced higher sera titers (1:320000) if
compared to immunizations with αGalNAc-glycopeptide 3–BSA (1:40000). Likewise, flow cytometry
assays showed higher capacity of the obtained anti-glycopeptide 4–BSA antibodies to recognize MCF-7
tumor cells. Cross recognition between immunopurified anti-βGalNAc antibodies and αGalNAc-
glycopeptide and vice versa was also verified. Lastly, molecular dynamics simulations and surface
plasmon resonance (SPR) showed that βGalNAc-glycopeptide 4 can interact with a model antitumor
monoclonal antibody (SM3). Taken together, these data highlight the improved immunogenicity of the
unnatural glycopeptide 4–BSA, bearing βGalNAc-Thr as Tn antigen isomer.
Keywords
MUC1-Mucins; Tumor; Tn antigen; Glycopeptides; Antibodies
1
Corresponding author. Tel.: +55 16 33154285; fax: +55 16 33154879; e-mail: vlcampo@fcfrp.usp.br
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
1. Introduction
Mucins are highly O-glycosylated (up to 80% of carbohydrates by weight) glycoproteins, containing
typical tandem repeat motifs rich in serine (Ser) and threonine (Thr), which represent potential sites for
abundant O-glycosylation. Characteristically, the sugar residue directly linked to these amino acids is 2-
acetamido-2-deoxy-α-D-galactopyranose (α-D-GalNAc).[1, 2] Mucins are present on cell surfaces of many
epithelial tissues, for instance, respiratory mucosa, reproductive system and gastrointestinal tract, among
others, where they exert essential roles of protection, hydration and lubrication. Additionally, apart from
these primary functions, mucins are also involved in complex biological processes, such as cell
differentiation, signaling, adhesion and migration, fertilization, infections and inflammation.[1, 3, 4] In
tumor-related mucins (e.g MUC1, MUC4 and MUC16), however, the glycosylation process is
deregulated, resulting in abnormal, incomplete and truncated glycans, where the exposure or
overexpression of O-α-D-GalNAc residue, also denominated Tn antigen when linked to the amino acid
Ser/Thr (αGalNAc-Ser/Thr), is a common feature.[3, 5, 6]
Among different classes of tumor mucins, MUC1 is the most extensively studied, as it is present on
nearly all epithelial tissues, being strongly overexpressed on epithelial tumor cells, where it is
characteristically changed if compared with the glycan profile of MUC1 on normal cells, particularly, in
7, 8]
the extracellular domain.[3,
Thus, a variable number (25–125) of tandem repeat sequences
(HGVTSAPDTRPAPGSTAPPA) can be observed, containing five potential O-glycosylation (O-
αGalNAc) sites (Scheme 1A).[7, 9] Due to its functional properties of adhesion, invasion and metastasis,
besides association to several types of cancer, such as breast, prostate, lung and pancreas, Tn antigen is
known as tumor associated carbohydrate antigen (TACA) and is considered a clinically relevant tumor
marker.[10] Thus, Tn antigen up-regulation acts as a signature associated with malignancy and, since its
occurrence in normal tissues is limited, it represents a potential target for development of synthetic
antitumor vaccines.[6, 11]
Accordingly, new synthetic methods for obtaining Tn antigen as well as its counterpart analogs have
been constantly investigated.[10] To date, several Tn antigen analogs have been reported for use as vaccine
13]
16]
constructs, for instance, C-glycosides,[12,
S-glycosides,[14] deoxyfluoro sugars[15,
and Tn
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
glycopeptides mimics containing non-natural amino acids.[17, 18] C-glycosides and S-glycosides are more
stable toward enzymatic and chemical hydrolysis than natural Tn, as they have the interglycosidic oxygen
replaced by methylene group or sulfur, respectively.[12, 14] Deoxyfluoro Tn analogs possessing fluorine
atom in substitution of a sugar hydroxy group are supposed to enhance both the immunogenicity and the
bioavailability of the antigen, considering the absence of fluorine in most organisms and its high bond
strength to carbon.[15] The use of non-natural amino acids, such as α-methylserine or those containing
long aliphatic side chains, has proved to be effective for obtaining non-natural Tn peptide-linked vaccines
18]
with improved immunogenicity.[17,
The application of Cu(I)-assisted 1,3-dipolar azide-alkyne
cycloaddition (CuAAC) reactions between propargylated α-GalNAc sugar and azido-functionalized
amino acids, or vice-versa, represents another strategy for synthesis of triazole-derived Tn
neoglycopeptide vaccines with advantageous physicochemical properties, inherent to the triazole
group.[19]
However, despite the vast repertoire of Tn antigen (αGalNAc-Ser/Thr) and related glycopeptides
analogs described so far, there are only few studies concerning the synthesis of its corresponding
stereoisomer (βGalNAc-Ser/Thr).[20] Furthermore, comparative immunological assays involving
glycopeptides bearing the distinct isomers have not been described yet. As judged by its structural
similarity to Tn antigen, differing only on the anomeric configuration, βGalNAc-Ser/Thr may be a
frontline building block for development of antitumor vaccines, depending on its efficiency for inducing
active immunity. In this regard, the crystal structure of the antitumor monoclonal antibody SM3 (SM3
mAb), able to recognize the fundamental epitope of the MUC1 peptide (SAPDTRPAP) and the
corresponding glycopeptide [SAPDT(O-αGalNAc)RPAP], may represent a model for theoretical
investigations of the βGalNAc-Ser/Thr core as a novel Tn mimic.[18, 21]
Therefore, this paper presents the synthesis of the novel protected O-glycosyl amino acids
βGalNAc-SerOBn (1) and βGalNAc-ThrOBn (2) as mimetics of the natural Tn antigen (αGalNAc-
Ser/Thr) (Scheme 1B), along with the solid-phase assembly of the MUC1 glycopeptides NHAcSer-Ala-
Pro-Asp-Thr[αGalNAc]-Arg-Pro-Ala-Pro-Gly-BSA (3) and NHAcSer-Ala-Pro-Asp-Thr[βGalNAc]-Arg-
Pro-Ala-Pro-Gly-BSA (4), bearing αGalNAc-Thr or βGalNAc-Thr (Scheme 1C). The immunological
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
assays comparing the antibodies production against tumor cells induced by glycopeptides 3-BSA and 4-
BSA, as well as their theoretical and experimental capacities to interact with a model antitumor
monoclonal antibody (SM3) by corresponding molecular dynamic simulations and surface plasmon
resonance (SPR) assays, will also be reported herein.
(SCHEME 1)
2. Results and Discussion
2.1. Synthesis
The target glycopeptides NHAcSer-Ala-Pro-Asp-Thr[αGalNAc]-Arg-Pro-Ala-Pro-Gly-BSA 3 and
NHAcSer-Ala-Pro-Asp-Thr[βGalNAc]-Arg-Pro-Ala-Pro-Gly-BSA 4 were obtained by solid-phase
synthesis, which involved the previous preparation of the glycosyl amino acids αGalNAc-ThrOH (Tn) (5)
and βGalNAc-ThrOH (6) as building blocks for obtaining corresponding tumor MUC1 fundamental
scaffold and its counterpart analog.
2.1.1. Synthesis of glycosyl amino acids
Owing to the polyfunctionality of carbohydrates and amino acids, the synthesis of glycosyl amino
acids demands the utilization of specific protecting groups, which must be suitably selected considering
the lability of the glycosidic bond under acidic conditions and the possibility of β-elimination reactions
under strong basic conditions.[22] Moreover, specific stereoselective methods are essential for synthesis of
distinct O-glycosyl amino acids isomers (α and β),[23] requiring the use of glycosyl donors containing
non-participating or participating groups at C-2 for synthesis of 1,2-cis (αGalNAc-Ser/Thr) or 1,2-trans
glycosides (βGalNAc-Ser/Thr), respectively.[24, 25] One of the most frequently employed strategy for
synthesis of glycosyl-amino acids relies on Koenigs-Knorr method,[26] which involves the use of glycosyl
halides, such as chlorides, bromides and iodides, as glycosyl donors, along with typically heavy metal
salts as catalysts.[27, 28] For instance, silver salts (Ag2CO3/AgClO4, AgOTf),[29] trimethylsilyl triflate
(TMSOTf),[30] dimethyl(methylthio)sulfonium triflate (DMTST)[31] and mercury salts (HgBr2, HgCl2,
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
Hg(CN)2)[32] are some representative catalysts used in Koenigs-Knorr type glycosylation reactions,[26]
since they are effective to assist leaving group displacement by complexation with anomeric halides.
In this context, for synthesis of the glycosyl amino acid αGalNAc-ThrOH (Tn) 5 we made use of
our established glycosylation method using HgBr2, by which glycosylation of FmocThr benzyl ester (7)
with the azido-derived glycosyl donor αGalN3Cl (8),[33] unable to exhibit neighboring group effect,
afforded exclusively the α isomer αGalN3-ThrOBn (9) (78%) (Scheme 1S).[34] Subsequently, the sugar
azido group of 9 was reductively acetylated using zinc powder in THF-acetic anhydride-acetic acid, and
its benzyl ester group was removed by standard hydrogenation reaction (10% Pd-C/H2, 1 h),[34] giving the
αGalNAc-ThrOH 5 (60%) to be utilized in the solid-phase synthesis of glycopeptide 3 (Scheme 1S).
Regarding the synthesis of β-linked GalNAc amino acids, Danishefsky et al. reported the synthesis of the
tumor associated Lewisy (Ley) pentasaccharides bearing both αGalNAc- and βGalNAc-Ser (anomeric
mixture α/β 3:1) by careful glycosylation conditions utilizing an azido-functionalized glycosyl donor and
TMSOTf as catalyst, followed by conversion to acetamides; however, similar glycosylation reaction with
threonine acceptor afforded only the α derivative.[35] In fact, our described HgBr2 catalyzed glycosylation
reaction of FmocSer benzyl ester (10) with αGalN3Cl 8 led to the predicted α glycoside (50%), along with
very low yield of βGalN3-SerOBn (4%), whereas none of the azido threonine-derived β-anomer could be
obtained by glycosylation of threonine acceptor 7 with the donor 8.[33] Thus, the fact that glycosylation
reactions utilizing the azido-donor 8 were ineffective to get both βGalNAc-Ser/Thr, prompted us to apply
an alternative route using the glycosyl donor αGalNAcCl (11), possessing an acetamide C-2 participating
group. Firstly, the synthesis of the donor 11 was carried out by peracetylation (Ac2O/ Py) of 2-amino-2-
deoxy-D-galactose hydrochloride (12), followed by treatment with acetyl chloride and HCl (g) for 18 h,
being obtained in 85% yield (Scheme 2).[25] This synthetic procedure differs from those previously
described conditions that utilize the titanium tetrachloride (TiCl4) as catalyst, affording 11 in similar
yield.[36] Subsequently, glycosylation reactions of FmocSer benzyl ester 10 or FmocThr benzyl ester 7
with the obtained αGalNAcCl 11, in the presence of 1,2-dichloroethane as solvent and HgBr2 as catalyst,
afforded exclusively the novel protected βGalNAc-SerOBn 1 (55%) or βGalNAc-ThrOBn 2 (52%) after
purification by column chromatography, highlighting the applicability of HgBr2 as a catalyst also for this
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
particular glycosyl donor 11 (Scheme 2). Noteworthy, these moderate yields are comparable to other
related β-glycosyl amino acids described in the literature.[37] The structures of the β isomers 1 and 2 were
confirmed by 1H NMR spectroscopy, being identified characteristic doublets at δ 4.60 (J1,2 8.3 Hz) and δ
4.65 (J1,2 8.3 Hz), confirming their β configuration. ESI-MS analysis also showed characteristic adducts
for 1 ([M+H]+ 747.2762) and 2 ([M+H]+ 761.2916). The glycosyl amino acid βGalNAc-ThrOBn 2 was
next submitted to hydrogenation reaction for removal of its benzyl ester group, affording βGalNAc-
ThrOH 6 (88%) as prompt intermediate for solid-phase synthesis of glycopeptide 4 (Scheme 2).
Regarding glycosyl amino acid 1, though this work did not involve its use for synthesis of glycopeptides
3 and 4, it can be a valuable building block for future synthesis of other MUC1 glycopeptide analogs with
distinct glycosylation patterns.
(SCHEME 2)
2.1.2. Solid-phase synthesis of glycopeptides 3 and 4
Moving towards the MUC1 glycopeptides 3 and 4, solid-phase glycopeptide synthesis (SPPS) was
employed, utilizing commercial Wang resin pre-loaded with Fmoc-Gly (0.65 mmol/g resin) as the solid
support and following the orthogonal Fmoc-based SPPS.[34, 38] Thus, after removal of N-Fmoc groups with
50% morpholine in DMF, the amino acids building blocks were coupled in the presence of the coupling
reagents PyBOP/HOBt and N,N-diisopropylethylamine (DIPEA), in DMF. Optimal coupling times
ranged from 6 h to 24 h for the incorporation of single amino acids, depending on preceding peptide
sequence. For incorporation of the more sterically and electronically demanding glycosyl amino acids 5
or 6 the coupling times typically ranged from 24 h to 30 h.
The synthesis of glycopeptides 3 and 4 was initiated by cleavage of the Fmoc group from Fmoc-
Gly-Wang resin, followed by nine rounds of coupling-deprotection with sequential incorporation of
Fmoc-ProOH, Fmoc-AlaOH, Fmoc-ProOH, Fmoc-Arg(Pbf)OH, αGalNAc-ThrOH 5 or βGalNAc-ThrOH
6, Fmoc-Asp(OtBu)OH, Fmoc-ProOH, Fmoc-AlaOH and Fmoc-SerOH, thus completing the preparation
of the protected, resin-bound glycopeptides (13) and (14) (Scheme 3). The obtained peptide coupling
efficiencies, as judged by measuring the released dibenzofulvene from the product, varied with the
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
position in the peptide chain: Pro1 (70%), Ala1 (80%), Pro2 (75%), Arg (72%), αGalNAc-ThrOH 5
39]
(52%), βGalNAc-ThrOH 6 (50%), Asp (60%), Pro3 (70%), Ala2 (85%) and Ser (80%).[28,
Subsequently, the obtained glycopeptides 13 and 14 were N-acetylated (Ac2O/ pyridine) for capping of
terminal NH2 and then cleaved from resin by treatment with aqueous TFA (80%), with concomitant
removal of Pbf and OtBu protecting groups of the corresponding Arg and Asp amino acids. Lastly,
removal of the sugar acetate protecting groups with catalytic NaOMe in MeOH gave crude glycopeptides
3 and 4, which were then analysed and purified by reversed-phase HPLC. As confirmed by NMR 1H and
ESI-MS analysis, the isolated peaks corresponded to glycopeptides 3 and 4, along with residual
1
glycopeptides containing Arg(Pbf) due to incomplete removal of this protecting group. Thus, NMR H
spectra of glycopeptides 3 (18%) and 4 (15%) showed for both compounds methylene and methyl signals
of amino acids between 3.0-1.12 ppm, 2 × NHAc, HβAsp signals between 2.82-2.70 ppm, as well as
characteristic signals for sugar H-2 of 3 (3.95 ppm, dd, J1,2 3.6 Hz, J2,3 10.8 Hz), and 4 (3.75 ppm, dd, J1,2
8.5 Hz, J2,3 10.9 Hz ). ESI-MS analysis also showed characteristic adducts for 3 (observed [M + 3NH4]+
1266.5413) and 4 (observed [M + 3NH4]+ 1266.5303).
Envisioning further use in immunization assays, glycopeptides 3 and 4 were conjugated to BSA
carrier protein. Thus, the coupling reactions of 3 and 4 to BSA were performed in aqueous Na2HPO4
buffer at pH 9.0, in the presence of the reagents EDCI and NHS (Scheme 3).[7, 40] The conjugated
glycopeptides were purified by ultrafiltration against deionized water using a membrane of 3 kDa. The
loading rates of the glycopeptides 3–BSA and 4–BSA conjugates were determined by MALDI-TOF mass
spectrometry, being found one or two molecules of 3 (67500.6) and 4 (68827.6), respectively, linked to
one BSA protein.
(SCHEME 3)
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
2.2. Immunological assays
2.2.1. ELISA assays
In order to verify the capacity of glycopeptides 3–BSA and 4–BSA conjugates to induce a humoral
immune response, each synthetic glycopeptide (10 μg) in combination with complete Freund’s adjuvant
(CFA) were injected into different groups of four Balb/c mice. After an interval of two weeks, the second
and third immunizations were performed with incomplete Freund’s adjuvant (IFA). Prior to the first
immunization and two weeks after the third immunization, blood was drawn from the eyelid vein of each
mouse, and the obtained sera were used for analysis of the induced antibodies by enzyme-linked
immunosorbent assay (ELISA).[41, 42] For comparative purposes, another group of four Balb/c mice was
immunized solely with BSA under the same conditions described for glycopeptides 3–BSA and 4–BSA,
with subsequent sera collection and analysis by ELISA. In the ELISA tests, the microtiter plates were
coated separately with glycopeptides 3–BSA and 4–BSA, as well as only with BSA as control. The sera
were titrated at increasing dilutions (1:2500 a 1:2560000) in a phosphate buffered saline (PBS) solution.
The antibodies concentrations were detected photometrically (450 nm) by using a secondary anti-mouse
antibody conjugated to horseradish peroxidase (HRP).[43] The titers of the ELISA tests were determined to
be approximately 1:40000 and 1:320000 for mice immunized with glycopeptides 3–BSA and 4–BSA,
respectively, and using these corresponding conjugates as coats. However, when coated exclusively with
BSA, lower corresponding titers of 1:20000 and 1:40000 were verified for mice immunized with
glycopeptides 3–BSA and 4–BSA, as expected. For immunizations with BSA and its use as coat the titer
of the ELISA test was close to 1:160000 (Figure 1). Thus, according to the obtained results, the employed
mice immunization protocol proved to be effective to induce the production of antibodies against the
antigenic structures of glycopeptides 3–BSA and 4–BSA related to MUC1 tumor mucins. Moreover, in
order to estimate the relative efficient immune reaction of anti-glycopeptide 3–BSA and anti-glycopeptide
4–BSA antibodies in the polyclonal sera from immunized mice, 1 µL of each serum was incubated with
PBS containing 3% BSA (m/v) for 1 h at room temperature, prior to additional ELISA assays (as
previously described). This step allowed the depletion of non-specific anti-BSA antibodies, resulting in a
decreased OD for all tested sera (p<0.05). However, this reduction was much less pronounced for
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
glycopeptide 4–BSA antisera (Figure S9), suggesting that the polyclonal serum from mice immunized
with glycopeptide 4–BSA is richer in specific antibodies that recognize glycopeptide 4 epitopes. In fact,
this result added to the higher sera titers induced by the glycopeptide 4–BSA, containing the glycosyl-
amino acid βGalNAc-ThrOH 6, if compared to the glycopeptide 3–BSA bearing the natural tumor antigen
αGalNAc-ThrOH 5 (Tn), highlights the increased immunological potential of this unnatural β-configured
building block, translating into an advantageous aspect for development of an antitumor vaccine, since it
is not present in normal and tumor cells found in mammals.
(FIGURE 1)
2.2.2. Flow cytometry assays
Flow cytometry assays were subsequently performed to evaluate the binding of the glycopeptides
3–BSA and 4–BSA induced antibodies to cells of epithelial breast tumor (MCF-7).[39, 42] Previous to these
assays, anti-MCF-7 sera was obtained for use as reference by immunization of four Balb/c mice with
MCF-7 cells, following the conditions cited above. Thus, MCF-7 cells were incubated with glycopeptides
3–BSA and 4–BSA induced antisera at increasing dilutions (1:400 a 1:8000), along with MCF-7 antisera
(1:1000). After washing, secondary antimouse-IgG antibodies from goat conjugated to a fluorescent label
(AlexaFluor 488) were added to the cells, which were then acquired and analyzed by flow cytometry. The
cells recognized by the antibodies from the mouse antisera showed fluorescence and were counted
separately (right area-P3). According to Figure 2A, cells treated with preimmune sera as negative control
did not show positive staining, appearing in the left area (P2). The MCF-7 cells that were incubated with
the corresponding MCF-7 antisera (1:1000) as positive control were all labeled and appeared in the right
area (Figures 2A, 3). When MCF-7 cells were incubated with glycopeptides 3–BSA and 4–BSA antisera,
they were somehow recognized by these antibodies, being verified for glycopeptide 3–BSA antisera
binding values (percentage of positive staining) of 37% (1:400 and 1:800), 15% (1:1000), 12% (1:2000),
8% (1:4000) and 6% (1:8000) (Figures 2B and 3), whereas glycopeptide 4–BSA antisera displayed higher
binding values at increased dilutions, as follows, 54% (1:400), 50% (1:800), 18% (1:1000), 16%
(1:2000), 15% (1:4000) and 10% (1:8000) (Figures 2C and 3).
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
Therefore, it is possible to assert that the obtained flow cytometry data using MCF-7 cells are in
accordance with ELISA results, strictly based on the synthetic glycopeptides 3–BSA and 4–BSA. In fact,
the observed binding capacities of glycopeptides 3–BSA and 4–BSA antisera towards MCF-7, mainly at
1:400 and 1:800 dilutions, are very significative, considering the high complexity of the natural MUC1
tumor mucins expressed by MCF-7 cells, whose essential scaffold is being mimicked by glycopeptides 3
and 4. Moreover, the higher binding to MCF-7 cells presented by glycopeptide 4–BSA antisera, if
compared to the glycopeptide 3–BSA antisera, reinforces the improved immunogenicity of the
glycopeptide 4–BSA, containing the unnatural glycosyl-amino acid βGalNAc-ThrOH 6.
(FIGURE 2)
(FIGURE 3)
2.2.3. Immunopurification of antisera and binding to glycopeptides 3 and 4
In order to better assess the recognition between αGalNAc- and βGalNAc-glycopeptides and their
corresponding anti-glycopeptides 3 and 4 antibodies, and to evaluate their cross recognition capacity,
antisera were immunopurified and evaluated by ELISA assays. Their immunopurifications were carried
out by initial immobilization of BSA-free glycopeptides 3 (αGalNAc) and 4 (βGalNAc) on the Sepharose
4B resin through their C-terminus after introduction of an hexamethylenediamine moiety, being their
couplings to the resin monitored via HPLC analysis of glycopeptides on the supernatants. By this method,
estimated incorporation values of 75% (about 1.3 mg) and 72% (about 1.3 mg) were verified for the
αGalNAc- and βGalNAc-glycopeptides, respectively. Anti-glycopeptides 3 and 4 antibodies were then
captured by the corresponding immobilized glycopeptides. After purification about 154 µg (0.22 mg/mL,
700 µL) antibodies were recovered from the anti-α serum and about 91 µg (0.13 mg/mL, 700 µL) from
the anti-β serum.
Subsequently, recognition between αGalNAc- and βGalNAc-glycopeptides and corresponding anti-
glycopeptides 3 and 4 antibodies, as well as the cross recognition between αGalNAc-glycopeptide and
anti-βGalNAc-glycopeptide antibodies and vice versa, were assessed by ELISA. In a preliminary
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
experiment performed by coating BSA-conjugated glycopeptides 3 and 4 the expected strong binding
(about 1.5 absorbance) was observed between immunopurified antibodies and the respective antigens
(Figure 4). Interestingly, also a remarkable recognition was observed between anti-αGalNAc antibodies
and βGalNAc-glycopeptide, and between anti-βGalNAc antibodies and αGalNAc-glycopeptide (on
average 1.0 and 0.8 absorbance, respectively), even at 0.1 µg/mL antibodies. These findings suggest that
antibodies raised with the βGalNAc-glycopeptide 4 can be recognized by the αGalNAc variant; therefore,
such antibodies have the potential to also recognize the natural Tn tumor antigen. Importantly, very weak
binding was recorded for the immunopurified anti-αGalNAc and anti-βGalNAc antibodies to immobilized
BSA, thus indicating that they selectively recognize the antigens and not the carrier used for
immunization (Figure 4). In addition, binding of immunopurified anti-αGalNAc and anti-βGalNAc
antibodies to asialo-submaxillary mucin, which commonly expresses the alpha-anomer, was also assessed
by ELISA assays, being verified that both anti-αGalNAc and anti-βGalNAc antibodies bound to asialo-
submaxillary mucin (See Supplementary Information, Figure S10).
(FIGURE 4)
2.3. Molecular dynamics simulations and surface plasmon resonance (SPR) assays with a model
antitumor monoclonal antibody
Considering the better immunogenicity observed for the glycopeptide 4–BSA, bearing the glycosyl-
amino acid βGalNAc-ThrOH 6, we asked whether or not it could interact with a model anti-MUC1 mAb,
specifically SM3, since it is the only anti-MUC1 mAb for which there is a crystallographic structure.
Accordingly, molecular dynamics simulations were carried out using the crystallographic complex
consisted of the SM3 mAb complexed with a nine-residue MUC-1 derived peptide (SAPDTRPAP)
(Figure 5A),[21] from which two distinct SM3-peptide complexes differing only by the O-glycosylation
type at the Thr residue (α or β GalNAc-Thr) were built, representing the glycopeptides 3 and 4 (Figure
5B-C).
(FIGURE 5)
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
According to molecular dynamics simulations, applying Chi1 dihedral angle of +60°, the
interactions between the SM3 mAb and the glycopeptides 3 (α) and 4 (β) does not show major deviations
from the crystallographic structure.[21] The carbohydrate moieties of both glycopeptides interact with SM3
via hydrogen bonds with Tyr-32L and Asn-53L (Figure 6A-B), suggesting that SM3 can also
accommodate the β-configured glycopeptide. This finding prompted us to further investigate the SM3-
αGalNAc 3 and SM3-βGalNAc 4 glycopeptides interactions experimentally.
(FIGURE 6)
Thus, SPR binding experiments of αGalNAc- and βGalNAc-glycopeptides were performed with
immobilized SM3 mAb. Firstly, αGalNAc-glycopeptide 3 was injected at increasing concentrations
starting from 50 nM. As shown in Figure 7A, classical binding and dissociation curves were observed but
no saturation was observed up to 2 µM. A rough estimation of the KD obtained by fitting the dissociation
and association curves between 50 and 400 nM gave a value of about 52 nM. In relation to βGalNAc-
glycopeptide 4, no binding was detected for concentrations below 500 nM (Figure 7B). Curves for the
concentrations between 500 nM and 2 µM were very noisy and the maximum RU signal was about 25%
that achieved with the αGalNAc-glycopeptide 3 at the same concentrations. A very rough estimation of
the KD obtained by fitting dissociation and association curves at 500 nM, 1 µM and 2 µM gave a value of
about 20 µM, which is quite similar to the reported binding affinity (KD 24 µM) between SM3 and the
unglycosylated peptide.[21] In fact, this βGalNAc- KD value is higher than that estimated with the
αGalNAc-glycopeptide variant, but it is still a representative binding to SM3 mAb.
(FIGURE 7)
3. Conclusions
In summary, we have synthesized the novel protected O-glycosyl amino acids βGalNAc-SerOBn 1
and βGalNAc-ThrOBn 2 as mimetics of the natural Tn antigen (αGalNAc-Ser/Thr), along with the solid-
phase assembly of the glycopeptides NHAcSer-Ala-Pro-Asp-Thr[αGalNAc]-Arg-Pro-Ala-Pro-Gly-BSA
(3–BSA) and NHAcSer-Ala-Pro-Asp-Thr[βGalNAc]-Arg-Pro-Ala-Pro-Gly-BSA (4–BSA), bearing
αGalNAc-Thr or βGalNAc-Thr, as mimetics of MUC1 tumor mucin glycoproteins. In this regard, HgBr2-
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
catalysed glycosylations, broadened applied to the glycosyl donor αGalNAcCl 11, afforded exclusively
the β-configured glycosyl amino acids 1 and 2.
According to ELISA tests, mice immunizations with βGalNAc-glycopeptide 4–BSA induced higher
sera titers (1:320000) if compared to immunizations with αGalNAc-glycopeptide 3–BSA (1:40000).
Likewise, flow cytometry assays showed higher capacity of the obtained anti-glycopeptide 4–BSA
antibodies to recognize MCF-7 tumor cells, as judged by the corresponding binding values of 37% and
54% verified for glycopeptides 3–BSA and 4–BSA antisera at 1:400 dilution. Cross recognition between
immunopurified anti-βGalNAc antibodies and αGalNAc-glycopeptide and vice versa was also verified.
Lastly, molecular dynamics simulations and surface plasmon resonance (SPR) showed that βGalNAc-
glycopeptide 4 can interact with a model antitumor monoclonal antibody (SM3). Overall, experimental
and theoretical data highlight the improved immunogenicity of the unnatural glycopeptide 4–BSA,
bearing βGalNAc-Thr as Tn antigen isomer, representing a frontline glycopeptide for development of
antitumor vaccines.
4. Experimental
4.1. General
All chemicals were purchased as reagent grade and used without further purification. Reactions
were monitored by thin layer chromatography (TLC) on 0.25 nm precoated silica gel plates (Whatman,
AL SIL G/UV, aluminium backing) with the indicated eluents. Compounds were visualized under UV
light (254 nm) and/or dipping in ethanol–sulfuric acid (95:5, v/v), followed by heating the plate for a few
minutes. Column chromatography was performed on silica gel 60 (Fluorochem, 35–70 mesh). Nuclear
magnetic resonance spectra were recorded on Bruker Advance DRX 300 (300 MHz), DPX 400 (400
MHz) or DPX 500 (500 MHz) spectrometers. Chemical shifts (δ) are given in parts per million downfield
from tetramethylsilane. Assignments were made with the aid of HMQC and COSY experiments.
Accurate mass electrospray ionization mass spectra (ESI-HRMS) were obtained using positive or
negative ionization modes on a Bruker Daltonics MicroOTOF II ESI-TOF mass spectrometer. MALDI-
TOF analyses were acquired using positive ionization mode on an UltrafleXtreme (Bruker, Billerica)
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
mass spectrometer. Absorbance of 96-well ELISA plates was measured at 450 nm using a
spectrophotometer SpectraMax - Molecular Device. FACS analysis were performed using a FACSCanto
II flow cytometer equipped with FACS Diva software (BD).
4.2. Synthesis
4.2.1. 2-Acetamide-3,4,6-tri-O-acetyl-2-deoxy-α-D-galactopyranosyl chloride 11
2-Amino-2-deoxy-D-galactose hydrochloride 12 (200 mg, 0.928 mmol) was treated with Ac2O (1.2
mL) and pyridine (2.0), and stirred at room temperature for 15 h. The obtained mixture was then extracted
with 10% HCl and saturated NaHCO3 and concentrated in vacuo, affording the peracetylated product 2-
acetamide-1,3,4,6-tri-O-acetyl-2-deoxy-α-D-galactopyranosyl as a pale viscous oil (350 mg, 0.89 mmol,
96%). The obtained compound (300 mg, 0.77 mmol) was treated with acetyl chloride (5 mL), previously
saturated with HCl gas, and then stirred at 40°C for 18 h. The obtained amber mixture was diluted in
dichloromethane, extracted with saturated NaHCO3, dried over MgSO4 and concentrated in vacuo. After
purification by silica gel column chromatography (EtOAc/Hex 7:3, v/v), the chloride 11 was obtained as
an yellowish solid in 85% yield (238.6 mg, 0.65 mmol). δH (CDCl3, 300 MHz) 6.26 (1H, d, J1,2 3.6 Hz,
H-1), 5.65 (1H, d, J 9.3 Hz, NHCO), 5.47 (1H, d, J 1.8 Hz, H-4), 5.29 (1H, dd, J3,4 3.3 Hz, J2,3 11.1 Hz,
H-3), 4.80 (1H, dt, J1,2 3.6 Hz, J2,3 11.2 Hz, H-2), 4.49 (1H, t, J 6.2 Hz, H-5), 4.19 (1H, dd, J5,6a 6.2 Hz,
J6a,6b 11.4 Hz, H-6a), 4.09 (1H, dd, J5,6b 6.7 Hz, J6a,6b 11.4 Hz, H-6b), 2.17, 2.06, 2.03, 2.0 (12H,4 s,
CH3CO).
4.2.2. General procedure for glycosylation reaction
A mixture of 2-acetamide-3,4,6-tri-O-acetyl-2-deoxy-α-D-galactopyranosyl chloride 11 (2 equiv.)
and N-(fluoren-9-ylmethoxycarbonyl)-serine/threonine benzyl ester 7/10 (1 equiv.), (prepared from
commercially available amino acid Fmoc-serine/threonine by treatment with caesium carbonate and
benzyl bromide in DMF), in 1,2-dichloroethane (2.0 mL) was refluxed for 20 to 22 h with mercuric
bromide (2.25 equiv.) and followed by TLC (EtOAc/hexane 7:3, v/v). The resulting amber mixture was
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
concentrated in vacuo and the residue was purified by a silica gel column chromatography
(EtOAc/hexane 7:3, v/v).
N-(Fluoren-9-ylmethoxycarbonyl)-(2-acetamido-2-deoxy-3,4,6-tri-O-acetyl-β-D-galactopyranosyl)-
L-serine benzyl ester 1
Following general glycosylation procedure, the reaction mixture of 11 (40 mg, 0.11 mmol) and 10
(22.5 mg, 0.054 mmol), in 1,2-dichloroethane (3.0 mL) was refluxed for 20 h with mercuric bromide (40
mg, 0.11 mmol), affording the β anomer 1 as an amorphous solid (22.4 mg, 0.029 mmol, 55%) after
purification by silica gel column chromatography (EtOAc/hexane 7:3, v/v). [a]D25 −23.2 (c 0.2 CHCl3). H
(CDCl3, 500 MHz) 7.70 (2H, d, J 7.0 Hz, CH Fmoc arom.), 7.59-7.54 (2H, m, CH Fmoc arom.), 7.35-
7.24 (9H, m, CH Bn arom., CH Fmoc arom.), 5.75 (1H, d, J 8.0 Hz, NH Ser), 5.27 (1H, d, J3,4 2.9 Hz, H-
4), 5.16-5.11 (3H, m, CH2 Bn, H-3), 4.60 (1H, d, J1,2 8.3 Hz, H-1), 4.49-4.44 (1H, m, CH Ser), 4.42-4.35
(2H, m, CH2 Fmoc), 4.21 (1H, dd, J 3.4 Hz, J 10.6 Hz, CH2a Ser), 4.18-4.13 (1H, m, CH Fmoc), 4.07-
4.01 (2H, m, H-6a, H-6b), 3.98-3.93 (1H, m, CH2b Ser), 3.84-3.81 (1H, m, H-2), 3.76 (1H, t, J 6,5 Hz, H-
5), 2.09, 2.07, 1.95 (12H, 3s, COCH3). C (CDCl3, 125 MHz) 129.0, 128.6, 128.2, 127.5, 125.5, 120.4
(CH Fmoc arom., CH Bn arom.), 101.4 (C-1), 70.9, 69.8, 69.0 (C-5, C-3, CH2 Ser), 67.7, 67.2, 66.7
(CH2Bn, CH2 Fmoc, C-4), 61.3 (C-6), 54.5 (CH Ser), 51.7 (C-2), 47.2 (CH Fmoc), 23.1 (NHCOCH3),
20.8 (OCOCH3). ESI-HRMS: calculated for C39H43N2O13 [M+H]+ 747.2760, found 747.2762.
N-(Fluoren-9-ylmethoxycarbonyl)-(2-acetamido-2-deoxy-3,4,6-tri-O-acetyl-β-D-galactopyranosyl)-
L-threonine benzyl ester 2
Following general glycosylation procedure, the reaction mixture of 11 (111 mg, 0.30 mmol) and 7
(64.6 mg, 0.15 mmol), in 1,2-dichloroethane (4.0 mL) was refluxed for 22 h with mercuric bromide
(109.8 mg, 0.30 mmol), affording the β anomer 2 as an amorphous solid (60.1 mg, 0.078 mmol, 52%)
25
after purification by silica gel column chromatography (EtOAc/hexane 7:3, v/v). [a]D −19.5 (c 0.2
CHCl3). H (CDCl3, 500 MHz) 7.69 (2H, d, J 7.5 Hz, CH Fmoc arom.), 7.57 (2H, d, J 7.3 Hz, CH Fmoc
arom.), 7.37-7.20 (9H, m, CH Bn arom., CH Fmoc arom.), 5.80 (1H, d, J 9.1 Hz, NH Thr), 5.57 (1H, d, J 8.1
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
Hz, NHAc), 5.23 (1H, d, J 3.1 Hz, H-4), 5.16 (1H, m, H-3), 5.13 (2H, AB, JAB 12.5 Hz, CH2 Bn), 4.57 (1H,
d, J1,2 8.3 Hz, H-1), 4.42-4.25 (4H, m, βCH Thr, αCH Thr, CH2 Fmoc), 4.17 (1H, t, J 7.0 Hz, CH Fmoc), 4.0
(1H, dd, J5,6 7.5 Hz, J6a,6b 11.4 Hz, H-6a), 3.95 (1H, dd, J5,6 6.5 Hz, J6a,6b 11.2 Hz, H-6b), 3.80 (1H, dd, J1,2
8.4 Hz, J2,310.6 Hz, H-2), 3.67 (1H, t, J 6.5 Hz, H-5), 2.06, 1.96, 1.93, 1.88 (12H, 4s, COCH3), 1.13 (3H, d, J
6.2 Hz, CH3 Thr). C (CDCl3, 125 MHz) 128.7, 128.3, 127.8, 127.8, 127.2, 125.2, 120.0 (CH Fmoc arom.,
CH Bn arom.), 99.0 (C-1), 74.8 (βCH Thr), 70.6 (C-5), 69.0 (C-3), 67.4 (CH2Bn), 66.9 (CH2 Fmoc), 66.5
(C-4), 61.1 (C-6), 58.7 (αCH Thr), 51.8 (C-2), 46.9 (CH Fmoc), 22.7 (NCOCH3), 20.7, 20.6, 20.5
(OCOCH3), 17.3 (CH3 Thr). ESI-HRMS: calculated for C40H45N2O13 [M+H]+ 761.2916, found 761.2916.
N-(Fluoren-9-ylmethoxycarbonyl)-(2-acetamido-2-deoxy-3,4,6-tri-O-acetyl-β-D-galactopyranosyl)-
L-threonine 6
A solution of N-(fluoren-9-ylmethoxycarbonyl)-O-(2-acetamido-2-deoxy-3,4,6-tri-O-acetyl-β-D-
galactopyranosyl)-L threonine benzyl ester 2 (60.1 mg, 0.078 mmol) in MeOH (2.5 mL) was treated with
glacial AcOH (0.25 mL) and 10% Pd/C (30 mg), being stirred under H2 (~1.5 atm) for 1 h. The reaction
mixture was then filtered through Celite, concentrated in vacuo and purified by column chromatography
(DCM-MeOH 9:1 v/v), affording the product 6 as an amorphous solid (48 mg, 0.078 mmol, 88%). H
(CD3OD, 400 MHz) 7.69 (2H, d, J 7.5, CH Fmoc arom.), 7.58 (2H, t, J 7.1, CH Fmoc arom.), 7.30-7.19
(4H, m, CH Fmoc arom.), 5.22 (1H, d, J 2.5 Hz, H-4), 4.99 (1H, dd, J3,4 3.1, J2,3 11.2, H-3), 4.52 (1H, d, J
8.3 Hz, H-1), 4.30-4.21 (3H, m, βCH Thr, αCH Thr, H-2), 4.13 (1H, t, J 7.0 Hz, CH Fmoc), 4.09-3.90 (4H,
CH2 Fmoc, H-6a, H-6b, H-5), 1.97, 1.91, 1.87, 1.84 (12H, 4s, COCH3), 1.09 (3H, d, J 6.0 Hz, CH3 Thr). C
(CDCl3, 100 MHz) 126.3-119.1 (CH Fmoc arom., CH Bn arom.), 99.5 (C-1), 75.7 (βCH Thr), 70.7 (C-3),
69.8(C-5), 66.8 (CH2 Fmoc), 66.6 (C-4), 66.2 (C-2), 60.9 (C-6), 58.7 (αCH Thr), 46.7 (CH Fmoc), 21.3-
19.4 (OCOCH3), 16.4 (CH3 Thr). ESI-HRMS: calculated for C33H38N2O13 Na+ [M+Na]+ 693.2266, found
693.2269.
4.2.3. General method for glycopeptide synthesis
The glycopeptides 3 and 4 were assembled manually using a fritted glass reaction vessel with
nitrogen purging for effective mixing. Pre-loaded FmocGly-Wang resin (typically 100 mg) was swollen
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
in CH2Cl2 for 2 h and then washed with DMF (3×). Deprotection of the N-α-Fmoc group was carried out
using 50% morpholine–DMF followed by filtration and washing with DMF (3x). Coupling reactions were
performed with 2.0 mol equiv of Fmoc-amino acid and coupling agents (PyBOP and HOBt) dissolved in
DMF, and the base DIPEA. Coupling times were variable, ranging from 6 h to 24 h for the amino acids
and between 24 h and 30 h for the glycosyl-amino acids 5 and 6. The reaction mixtures were filtered after
each coupling and the resin washed three times with DMF, CH2Cl2 and MeOH. After drying in vacuo,
small aliquots of resin (1 mg) were treated with 20% morpholine in DMF for removal of NFmoc group
and consequent generation of dibenzofulvene product which absorb UV strongly (290 nm), offering
potential for monitoring of coupling reactions by spectrophotometer. After removal of the NFmoc group
of the last amino acid, the glycopeptides were N-acetylated with Py and Ac2O and then cleaved from resin
in the presence of aqueous 80% TFA. After cleavage was complete, the solutions were filtered to remove
the cleaved resin, concentrated and then treated with 1M sodium methoxide solution in MeOH for
removal of the sugar acetate protecting groups. The purification of the obtained glycopeptides 3 and 4
was carried out on Shimadzu Shim-PaK HPLC equipment using a C18 reversed-phase column Shim-
PacK CLC-ODS (M) column (250 x 4.6 mm), under gradient conditions (A: 0.1% TFA/ H2O, B: MeCN,
0-70% B, within 30 min.), at a flow rate of 1.0 mL min.-1, with UV detection (224 nm). Under these
conditions the retention times for glycopeptides 3 and 4 were 25.8 min and 25.3 min, respectively.
NHAcSer-Ala-Pro-Asp-Thr[αGalNAc]-Arg-Pro-Ala-Pro-Gly-OH 3
Using the standard peptide synthesis procedure outlined, followed by purification by reversed-phase
HPLC, glycopeptide 3 was obtained as an amorphous solid (15.0 mg, 0.012 mmol, 18 %). tR (224 nm) =
25.8 min; H (D2O, 400 MHz) 4.52-4.48 (2H, m, HαThr, HαSer), 4.41-4.38 (2H, m, HαPro, HαArg), 4.30 (2H,
m, 2HαAla), 4.20 (1H, m, HβThr), 3.95 (1H, dd, J1,2 3.6 Hz, J2,3 10.8 Hz, H-2), 3.87 (1H, t, J 6.2 Hz, H-5),
3.83-3.61 (11H, m, 2H-6, H-3, H-4, 3HδPro, CH2Gly, 2HβSer), 3.54-3.50 (3H, m, HδPro), 3.07 (1H, t, J 7.3 Hz,
HδArg), 2.82 (1H, dd, J 6.8 Hz, J 17.0 Hz, HβAsp), 2.70 (1H, dd, J 6.7 Hz, J 17.0 Hz, HβAsp), 2.18-2.10 (3H,
m, HβPro),1.95, 1.93 (6H, 2 s, 2 x NHCOCH3), 1.84-1.69 (6H, m, HβPro, HγPro), 1.35 (4H, d, J 5.5 Hz,
2HβArg, 2HγArg), 1.25 (6H, d, J 7.3 Hz, CH3 Ala), 1.12 (3H, d, J 6.2 Hz, CH3 Thr). ESI-HRMS: calcd. for
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
+
C50H80N14O21 3NH4 [M + 3NH4]+: 1266.6637, found: 1266.5413. Residual glycopeptide containing
Arg(Pbf) - ESI-HRMS: calcd. for C63H96N14O24S [M + Na]+: 1487.6335, found: 1487.6091.
NHAcSer-Ala-Pro-Asp-Thr[βGalNAc]-Arg-Pro-Ala-Pro-Gly-OH 4
Using the standard peptide synthesis procedure outlined, followed by purification by reversed-phase
HPLC, glycopeptide 4 was obtained as an amorphous solid (13.0 mg, 0.010 mmol, 15 %). tR (224 nm) =
25.3 min;H (D2O, 400 MHz) 4.47-4.39 (2H, m, HαThr, HαSer), 4.35-4.33 (2H, m, HαPro, HαArg), 4.31-4.28
(2H, m, 2HαAla), 4.13-4.10 (1H, m, HβThr), , 3.88-3.77 (4H, H-5, 3HδPro), 3.75 (dd, J1,2 8.5 Hz, J2,3 10.9 Hz,
H-2), 3.70-3.56 (8H, 2H-6, H-3, H-4, CH2Gly, 2HβSer), 3.53-3.49 (3H, m, HδPro), 3.09 (1H, t, J 7.3 Hz,
,
HδArg), 2.81 (1H, dd, J 6.9 Hz, J 16.8 Hz, HβAsp), 2.73 (1H, dd, J 7.3 Hz, J 17.1 Hz, HβAsp), 2.18-2.10 (3H,
m, HβPro), 1.96, 1.93 (6H, 2 s, 2 x NHCOCH3), 1.83-1.69 (6H, m, HβPro, HγPro), 1.35 (4H, d, J 5.5 Hz,
2HβArg, 2HγArg), 1.25 (6H, m, CH3 Ala), 1.00 (3H, d, J 6.3 Hz, CH3 Thr). ESI-HRMS: calcd. for
+
C50H80N14O21 3NH4 [M + 3NH4]+: 1266.6637, found: 1266.5303. Residual glycopeptide containing
Arg(Pbf) - ESI-HRMS: calcd. for C63H96N14O24S [M + Na]+: 1487.6335, found: 1488.6071.
4.2.4. General procedure for coupling of glycopeptides 3 and 4 to BSA
Glycopeptides 3 and 4 (2.5 mg, 2 mol, 45 eq.) were dissolved in water (200 μL) and treated with
the coupling reagents EDCI (380 μg, 2 mol) and HOBt (270 μg, 2 mol). The reaction mixtures were
allowed to stand at room temperature with occasional gentle mixing and after 50 min they were added to
a solution of BSA (1.5 mg) in Na2HPO4 buffer (20 mM, pH 9.0, 200 μL), previously precooled to 4°C.
The solutions were then reacted for 48 h at room temperature and subsequently purified by ultrafiltration
(AMICON ULTRA-0.5-MWCO 3 kDa). After lyophilization the obtained conjugates (approximately 7.0
mg) were analyzed by MALDI-TOF MS, being verified for glycopeptides 3- and 4-BSA corresponding
peaks of mass 67500.6 and 68827.6, consistent with 1 and 2 epitopes per BSA.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
4.3. Immunological assays
4.3.1. Mice Immunization
A solution of 10 μg of BSA (control) or glycopeptides 3 and 4–BSA conjugates in PBS, in
combination with complete Freund’s adjuvant (CFA), was injected into 4 different Balb/c mice. After an
interval of two weeks, the second and third immunizations were performed utilizing an emulsion with
incomplete Freund’s adjuvant (IFA). The immunization protocol for MCF-7 cells was similar to that used
for glycopeptides 3 and 4–BSA, with utilization of 1 x 106 of lysed MCF-7cells, obtained by extraction
with cell lysis buffer. Preimmune sera were collected from all mice previously to immunization.
4.3.2. Enzyme-Linked Immunosorbent Assay (ELISA)
96-Well ELISA plates were coated with glycopeptides 3 or 4–BSA in phosphate buffer solution
(PBS), or only with BSA (control), and incubated overnight at 4 °C. The plates were then washed with
PBS buffer and blocked with 3% gelatin in 0.05% Tween PBS buffer for 2 h at 37 °C. After washing with
0.05% Tween PBS buffer, the plates were incubated with different dilutions of preimmune and immune
sera from all mice (ranging from 1:2500 to 1:2560000) in 0,05% Tween PBS for 2 h at 37 °C.
Subsequently, the wells were washed and incubated with horseradish peroxidase conjugated to rabbit
anti-mouse IgG Ab for 1 h at 37 °C. The plates were then washed and the substrate solution of 3,3′,5,5′-
Tetramethylbenzidine (TMB) and H2O2 were added. After 10 min the solution of 2N sulfuric acid was
added and the absorbance of each well was measured at 450 nm using a spectrophotometer (SpectraMax -
Molecular Device).
4.3.3. Flow cytometry analysis
Freshly harvested MCF-7 cells were washed with cold PBS and incubated with FcR blocking
solution for 1 hour in ice bath. Then, cells were washed thrice (300 x g, 5 min.), and separated in flow
cytometry tubes (1x106 cells per tube) for subsequent incubation with different dilutions of preimmune
and immune sera from mice immunized with glycopeptide 3- and 4-BSA (ranging from 1:400 to 1:8000)
in cold PBS and ice bath. Control cells were incubated with sera from those mice immunized with MCF-7
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
extract. After 1 h, cells were washed three times and goat antimouse-IgG antibodies labeled with
fluorescent dye (AlexaFluor 488) were added to the cells. After staining, cells were analysed by flow
cytometry with a FACSCanto flow cytometer equipped with FACSDiva software (BD Biosciences).
4.3.4. Immunopurification of antisera and binding to glycopeptides 3 and 4
Immobilization of peptides on Sepharose 4B resin
αGalNAc-glycopeptide 3 (1.7 mg, 1.4 μmol) and βGalNAc-glycopeptide 4 (1.8 mg, 1.5 μmol) were
coupled to an NHS-activated Sepharose 4B Fast flow resin (NHS: N-hydroxy-succinimmide) to perform
immunopurification of anti-glycopeptides 3 and 4 antibodies. The two glycopeptides had no free amines
for direct coupling, but they contained one C-terminal carboxyl group. We therefore pre-treated the resin
in order to introduce an amino group and then conjugate glycopeptides 3 and 4 through a canonical
EDC/NHS chemistry. The resin had a nominal substitution of 16 to 23 μmol NHS/mL drained medium.
Two 200 μL resin aliquots (about 3 – 8 μmol NHS groups) were washed twice with cold 1 mM HCl and
then with cold water. Solvents were removed by centrifugation. Resins were then treated with
hexamethylenediamine by suspending them in 500 μL of 1.0 M reagent in 50 mM phosphate buffer pH
8.0 and left under stirring overnight at 4 °C. Resins were then washed extensively with phosphate buffer.
Glycopeptides 3 and 4 were preactivated by dissolution in 600 μL of 0.05M NHS/0.2M EDC in water.
After 30 minutes solutions were transferred onto the two separate resins. Couplings of the NHS-
preactivated glycopeptides to the amine groups on the resin were allowed to proceed overnight and were
monitored by RP-HPLC by taking resin samples (50 μL) at t=0 and after the overnight coupling. For the
HPLC analysis, resin samples were centrifuged at 13000g and the supernatants analyzed on a C18
monolithic ONYX column 50 x 2 mm ID operated at 0.6 mL/min, applying a gradient of CH3CN, 0.1%
TFA from 1% to 70% in 15 minutes. The resin was finally washed twice with NaAc 0.1 M pH 5.0 and
PBS pH 7.5.
Antibodies Immunopurification
Resins were equilibrated in PBS pH 7.5. Anti-αGalNAc-glycopeptide 3 and anti-βGalNAc-
glycopeptide 4 sera (500 μL) were incubated with the corresponding resin-immobilized glycopeptides for
16 h under gentle shaking at room temperature. Supernatants were then recovered and resins were washed
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
5 times with PBS pH 7.5. Resins were then treated with 350 μL 10 mM NaOH for 10 minutes under
gentle shaking. Supernatants were removed by centrifugation and immediately neutralized with 2M TRIS
pH 7.0. The presence of antibodies was assessed by SDS-PAGE analysis of the bound fraction. Protein
concentration was assessed by the Bradford method.
ELISA assays
ELISA was used to monitor the binding of immunopurified antibodies to the corresponding
immunogens (αGalNAc-glycopeptide 3 and βGalNAc-glycopeptide 4) and cross recognition of anti-
αGalNAc-glycopeptide antibodies to βGalNAc-glycopeptide and viceversa. The glycopeptides 3 and 4
were used as BSA-conjugates to improve plate adsorption. Initial binding assays were performed by
coating BSA-glycopeptides 3 and 4 on microtiter plates at 10 µg/mL, 1 µg/mL and 0.1 µg/mL dissolved
in PBS (100 µL, overnight, triplicate wells). To exclude the immunopurified antibodies that recognize
BSA used as carrier for immunization, wells were coated with 10 µg/mL BSA and probed with the two
different anti-glycopeptide 3 and anti-glycopeptide 4 antibodies. After extensive washing with PBS, wells
were blocked by incubation with 300 µL of 3% BSA w/v for 3h at room temperature. Immunopurified
anti-αGalNAc and βGalNAc-glycopeptide antibodies were then dispensed in the wells both at 1 µg/mL
and incubated for 1h at 37 °C (100 µL). After washing, wells were treated with goat anti-mouse-HRP
1:1000 (1 µg/mL) for 1h at 37 °C (100 µL). After washing, wells were treated with o-phenylenediamine
dihydrochloride 0.4 mg/mL in 0.1 M citric acid and 0.2 M Na2HPO4 buffer (pH 4.8) and adding 0.2
μL/mL of 30% H2O2 as substrate (100 µL). The chromogenic reaction was stopped adding 50 µL 2.5 M
H2SO4 and absorbance at 490 nm was read using an EnSpire multiwell reader (Perkin Elmer). The
obtained data were averaged, blanks subtracted and plotted using GraphPad vers. 4.
4.4. Surface plasmon resonance assays (SPR)
SPR assays were performed using the anti-MUC1 SM3 monoclonal antibody (ABCAM,
concentration 1 mg/mL) and the two synthetic glycopeptides 3 and 4. All binding assays were performed
on a Biacore 3000 instrument (GE Healthcare). A CM5 sensor chip was functionalized with the anti-
MUC1 SM3 monoclonal antibody running the Wizard method and applying a standard amine coupling
chemistry.[44] Firstly, the sensor chip surface was activated with EDC/NHS. Then SM3 at 10 μg/mL was
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
injected (70 μL) in NaAc 10 mM pH 4.0. When ligand immobilization was complete, 1M ethanolamine,
pH 8.5 was passed over the chip. After a washing pulse with NaOH 10 mM and baseline equilibration, we
estimated an immobilization level of about 7000 Response Units (RUs). The reference channel was
deactivated by treatment with 1 M ethanolamine after the NHS/EDC activation. Signals from this channel
were subtracted from those of the sample channel. All binding assays were conducted at 25 °C, at a
constant flow rate of 20 μL/min using HBS-EP as running buffer. The samples of glycopeptides 3 and 4
were prepared in HBS-EP buffer and a total volume of 60μL of each analyte solution was injected onto
the surface at increasing concentrations providing a dose-response signal. Overall binding events were
monitored for 550 sec; the association step for 180 sec while the dissociation step for 370 sec and after
each injection the chip surface was regenerated with NaOH using a range of concentrations comprised
between 2 μM and 0.05 μM. Fitting of binding data was performed by using BIAevaluation software
(version 4.1), adopting the 1:1 Langmuir binding model to extrapolate kinetic and thermodynamic
parameters. ka stands for association constant, kd stands for dissociation constant and KD stands for
apparent affinity constant. The final dissociation constant value (KD) reported for each analyte was the
average between the three KD values extrapolated from every channel.
4.5. Molecular dynamics simulations
All simulations were performed with the GROMACS 5.0.4 simulation package[45] using the
AMBER99SB-ILDN forcefield[46] combined with the Glycam06 parameters for carbohydrates and
glycoproteins[47] and TIP3P water model using PBC in a triclinic box in which the ionic strength was set
to 150 mM with Na+ and Cl-. The time-step value was 2 fs with water and non-water bonds constrained
by the LINCS and SETTLE algorithms, respectively. All intermolecular interactions were treated
explicitly inside 1 nm cutoffs and the long-range electrostatic interactions were treated by PME.
The initial states of systems were according to the crystallographic structure of the SM3 antibody
complexing the MUC1 9-mer peptide (PDB ID 1SM3).[21] The glycosylated peptides were built with the
GLYCAM-Web server (http://glycam.org/). The dihedral angle χ1 of the glycosylated Thr residue was +60°
as in the crystallographic structure[21] and other published reports. [20, 48]
The fully solvated, electrically neutral systems at the desired ionic strength were energy minimized
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
using the steepest descent algorithm. During the equilibration phase, temperature was set to 309 K by the
v-rescale thermostat and pressure to 1 bar by the Berendsen barostat. Temperature and pressure were kept
at the same values during production runs by the Nosé-Hoover thermostat and Parrinello-Rahman
barostat. Total simulation time was 1 ns for the equilibration phase and 100 ns for the production phase
for each system.
Images were generated with PyMOL. The hardware used for preparative and equilibration steps of
the molecular dynamics simulations was composed by virtual machines hosted at the University of São
Paulo cloud computing infrastructure. A total of 4 virtual machines, each with 8 Intel Xeon cores at 2.40
GHz and 32 GB RAM, were used. The production step were performed with the DAVinCI supercomputer
located at Rice University (Texas-USA), through a scientific collaboration agreement with the University
of São Paulo (São Paulo-Brazil). For each simulation, 36 processing cores at 2.83 GHz and 6 nVIDIA
Tesla M2050 GPUs were used.
1
Supporting Information Available: The Supporting Information encloses the H NMR, ESI-MS/
MALDI-TOF analyses and chromatograms of the main compounds, as well as complementary figures
related to biological assays.
Acknowledgements
We acknowledge the financial support and fellowships from FAPESP (Fundação de Amparo à
Pesquisa do Estado de São Paulo – Brazil, Process N° 2012/19390-0) and CAPES (Coordenação de
Aperfeiçoamento de Pessoal de Nível Superior). We are grateful to Dra. Lílian Cataldi Rodrigues for
technical assistance in some biological assays, to Vinícius Palaretti for NMR analyses, to José Carlos
Tomaz for ESI-MS analyses, to José César Rosa for MALDI-TOF analyses, to Prof. Antonio Caliri for
discussions about the molecular dynamics simulations, and to the Data Analysis and Visualization
CyberInfrastructure funded by NSF (grant OCI-0959097) for supporting the molecular dynamics studies.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
References
[1]
M. Andrianifahanana, N. Moniaux, S. K. Batra, Biochimica Et Biophysica Acta-Reviews on Cancer 2006, 1765, 189-
222.
[2]
[3]
C. A. Buscaglia, V. A. Campo, A. C. C. Frasch, J. M. Di Noia, Nature Reviews Microbiology 2006, 4, 229-236.
H. Cai, Z.-Y. Sun, Z.-H. Huang, L. Shi, Y.-F. Zhao, H. Kunz, Y.-M. Li, Chemistry-a European Journal 2013, 19,
1962-1970.
[4]
[5]
[6]
[7]
K. Agarwal, R. Kaul, M. Garg, A. Shajahan, S. K. Jha, S.-G. Sampathkumar, Journal of the American Chemical
Society 2013, 135, 14189-14197.
E. Osinaga, Iubmb Life 2007, 59, 269-273; C. Brocke, H. Kunz, Bioorganic & Medicinal Chemistry 2002, 10, 3085-
3112.
D. J. Gill, T. Keit Min, J. Chia, S. C. Wang, C. Steentoft, H. Clausen, E. A. Bard-Chapeau, F. A. Bard, Proceedings of
the National Academy of Sciences of the United States of America 2013, 110, E3152-E3161.
U. Westerlind, A. Hobel, N. Gaidzik, E. Schmitt, H. Kunz, Angewandte Chemie-International Edition 2008, 47, 7551-
7556.
[8]
[9]
G. Rivalland, B. Loveland, P. Mitchell, Expert Opinion on Biological Therapy 2015, 15, 1773-1787.
S. Dziadek, D. Kowalczyk, H. Kunz, Angewandte Chemie-International Edition 2005, 44, 7624-7630.
D. Y. Feng, A. S. Shaikh, F. S. Wang, Acs Chemical Biology 2016, 11, 850-863.
A. Holeman, P. H. Seeberger, Current Opinion in Biotechnology 2004, 15, 615-622.
D. Urban, T. Skrydstrup, J. M. Beau, Chemical Communications 1998, 955-956.
F. Peri, L. Cipolla, M. Rescigno, B. La Ferla, F. Nicotra, Bioconjugate Chemistry 2001, 12, 325-328.
D. A. Thayer, H. N. Yu, M. C. Galan, C. H. Wong, Angewandte Chemie-International Edition 2005, 44, 4596-4599.
A. Hoffmann-Roeder, M. Johannesy, Chemical Communications 2011, 47, 9903-9905.
S. Wagner, C. Mersch, A. Hoffmann-Roeder, Chemistry-a European Journal 2010, 16, 7319-7330; M. Johannes, T.
Oberbillig, A. Hoffmann-Roeder, Organic & Biomolecular Chemistry 2011, 9, 5541-5546.
S. J. Keding, A. Endo, S. J. Danishefsky, Tetrahedron 2003, 59, 7023-7031.
F. Corzana, J. H. Busto, F. Marcelo, M. Garcia de Luis, J. L. Asensio, S. Martin-Santamaria, Y. Saenz, C. Torres, J.
Jimenez-Barbero, A. Avenoza, J. M. Peregrina, Chemical Communications 2011, 47, 5319-5321.
Q. Wan, J. Chen, G. Chen, S. J. Danishefsky, Journal of Organic Chemistry 2006, 71, 8244-8249; N. Miller, G. M.
Williams, M. A. Brimble, Organic Letters 2009, 11, 2409-2412.
F. Corzana, J. H. Busto, G. Jimenez-Oses, M. G. de Luis, J. L. Asensio, J. Jimenez-Barbero, J. M. Peregrina, A.
Avenoza, Journal of the American Chemical Society 2007, 129, 9458-9467.
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
P. Dokurno, P. A. Bates, H. A. Band, L. M. D. Stewart, J. M. Lally, J. M. Burchell, J. Taylor-Papadimitriou, D. Snary,
M. J. E. Sternberg, P. S. Freemont, Journal of Molecular Biology 1998, 284, 713-728.
C.-H. Wong, Carbohydrate-based drug discovery, Wiley-VCH Verlag GmbH & Co. KgaA, Weinheim, 2003; G.
Arsequell, G. Valencia, Tetrahedron-Asymmetry 1997, 8, 2839-2876; Wakabaya.K, W. Pigman, Carbohydrate
Research 1974, 35, 3-14.
[23]
[24]
[25]
[26]
[27]
W. P. Stockl, H. Weidmann, Journal of Carbohydrate Chemistry 1989, 8, 169-198.
H. Paulsen, C. Kolar, W. Stenzel, Chemische Berichte-Recueil 1978, 111, 2358-2369.
R. L. B. J. N. Whistler, Methods in carbohydrate chemistry, Vol. 6, Academic, New York, 1972.
W. Koenigs, E. Knorr, Berichte Der Deutschen Chemischen Gesellschaft 1901, 34, 957-981.
A. F. G. Bongat, A. V. Demchenko, Carbohydrate Research 2007, 342, 374-406; R. M. van Well, K. P. R. Kartha, R.
A. Field, Journal of Carbohydrate Chemistry 2005, 24, 463-474; T. S. Kaerkkaeinen, K. P. R. Kartha, D. MacMillan,
R. A. Field, Carbohydrate Research 2008, 343, 1830-1834; J. C. Manimala, Z. T. Li, A. Jain, S. VedBrat, J. C.
Gildersleeve, Chembiochem 2005, 6, 2229-2241.
[28]
[29]
[30]
[31]
[32]
[33]
S. A. Mitchell, M. R. Pratt, V. J. Hruby, R. Polt, Journal of Organic Chemistry 2001, 66, 2327-2342.
E. Meinjohanns, M. Meldal, K. Bock, Tetrahedron Letters 1995, 36, 9205-9208.
U. K. Saha, R. R. Schmidt, Journal of the Chemical Society-Perkin Transactions 1 1997, 1855-1860.
M. Schultz, H. Kunz, Tetrahedron-Asymmetry 1993, 4, 1205-1220.
I. Carvalho, S. L. Scheuerl, K. P. R. Kartha, R. A. Field, Carbohydrate Research 2003, 338, 1039-1043.
M. B. Martins-Teixeira, V. L. Campo, M. Biondo, R. Sesti-Costa, Z. A. Carneiro, J. S. Silva, I. Carvalho, Bioorganic
& Medicinal Chemistry 2013, 21, 1978-1987.
[34]
[35]
V. L. Campo, T. B. Riul, I. Carvalho, M.-D. Baruffi, Chembiochem 2014, 15, 1495-1507.
P. W. Glunz, S. Hintermann, L. J. Williams, J. B. Schwarz, S. D. Kuduk, V. Kudryashov, K. O. Lloyd, S. J.
Danishefsky, Journal of the American Chemical Society 2000, 122, 7273-7279.
[36]
Z. Tarasiejska, R. W. Jeanloz, Journal of the American Chemical Society 1958, 80, 6325-6327; D. Stokmaier, O.
Khorev, B. Cutting, R. Born, D. Ricklin, T. O. G. Ernst, F. Boeni, K. Schwingruber, M. Gentner, M. Wittwer, M.
Spreafico, A. Vedani, S. Rabbani, O. Schwardt, B. Ernst, Bioorganic & Medicinal Chemistry 2009, 17, 7254-7264.
V. L. Campo, I. Carvalho, S. Allman, B. G. Davis, R. A. Field, Organic & Biomolecular Chemistry 2007, 5, 2645-
2657.
[37]
[38]
W. C. Chan, P. D. White, Fmoc solid phase peptide synthesis: a practical approach, Oxford University Press Inc.,
New York, 2000.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
[39]
A. Hoffmann-Roeder, A. Kaiser, S. Wagner, N. Gaidzik, D. Kowalczyk, U. Westerlind, B. Gerlitzki, E. Schmitt, H.
Kunz, Angewandte Chemie-International Edition 2010, 49, 8498-8503.
[40]
[41]
H. Kunz, Journal of Peptide Science 2003, 9, 563-573.
J. S. Plested, P. A. Coull, M. A. J. Gidney, in Haemophilus influenzae Protocols, Vol. 71 (Ed.: M. A. H. Herbert,
Derek W. Moxon, E. Richard), Humana Press, 2003, pp. 243-261.
[42]
[43]
[44]
[45]
[46]
[47]
[48]
H. Cai, Z.-Y. Sun, M.-S. Chen, Y.-F. Zhao, H. Kunz, Y.-M. Li, Angewandte Chemie-International Edition 2014, 53,
1699-1703.
A. Kurosaka, A. Yano, N. Itoh, Y. Kuroda, T. Nakagawa, T. Kawasaki, Journal of Biological Chemistry 1991, 266,
4168-4172.
A. Sandomenico, V. Severino, A. Chambery, A. Foca, G. Foca, C. Farina, M. Ruvo, Molecular Biotechnology 2013,
54, 983-995.
D. Van der Spoel, E. Lindahl, B. Hess, G. Groenhof, A. E. Mark, H. J. C. Berendsen, Journal of Computational
Chemistry 2005, 26, 1701-1718.
K. Lindorff-Larsen, S. Piana, K. Palmo, P. Maragakis, J. L. Klepeis, R. O. Dror, D. E. shaw, Proteins-Structure
Function and Bioinformatics 2010, 78, 1950-1958.
K. N. Kirschner, A. B. Yongye, S. M. Tschampel, J. Gonzalez-Outeirino, C. R. Daniels, B. L. Foley, R. J. Woods,
Journal of Computational Chemistry 2008, 29, 622-655.
D. Madariaga, N. Martinez-Saez, V. J. Somovilla, H. Coelho, J. Valero-Gonzalez, J. Castro-Lopez, J. L. Asensio, J.
Jimenez-Barbero, J. H. Busto, A. Avenoza, F. Marcelo, R. Hurtado-Guerrero, F. Corzana, J. M. Peregrina, Acs
Chemical Biology 2015, 10, 747-756; D. M. Coltart, A. K. Royyuru, L. J. Williams, P. W. Glunz, D. Sames, S. D.
Kuduk, J. B. Schwarz, X. T. Chen, S. J. Danishefsky, D. H. Live, Journal of the American Chemical Society 2002,
124, 9833-9844; U. Karsten, N. Serttas, H. Paulsen, A. Danielczyk, S. Goletz, Glycobiology 2004, 14, 681-692; L.
Kinarsky, G. Suryanarayanan, O. Prakash, H. Paulsen, H. Clausen, F. G. Hanisch, M. A. Hollingsworth, S. Sherman,
Glycobiology 2003, 13, 929-939; J. Schuman, A. P. Campbell, R. R. Koganty, B. M. Longenecker, Journal of Peptide
Research 2003, 61, 91-108; S. Dziadek, C. Griesinger, H. Kunz, U. M. Reinscheid, Chemistry-a European Journal
2006, 12, 4981-4993; D. Madariaga, N. Martinez-Saez, V. J. Somovilla, L. Garcia-Garcia, M. A. Berbis, J. Valero-
Gonzalez, S. Martin-Santamaria, R. Hurtado-Guerrero, J. L. Asensio, J. Jimenez-Barbero, A. Avenoza, J. H. Busto, F.
Corzana, J. M. Peregrina, Chemistry-a European Journal 2014, 20, 12616-12627; T. Matsushita, N. Ohyabu, N.
Fujitani, K. Naruchi, H. Shimizu, H. Hinou, S. I. Nishimura, Biochemistry 2013, 52, 402-414.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
FIGURES AND SCHEMES
Scheme 1. (A) Schematic representation of MUC1 tandem repeat sequence containing Tn antigen
(αGalNAc-Ser/Thr). (B) Chemical structures of the protected O-glycosyl amino acids βGalNAc-SerOBn
1 and βGalNAc-ThrOBn 2 as Tn antigen analogs. (C) Chemical structures of the target MUC1
glycopeptides 3–BSA and 4–BSA.
Scheme 2. Synthesis of protected O-glycosyl amino acids βGalNAc-SerOBn 1, βGalNAc-ThrOBn 2 and
βGalNAc-ThrOH 6. Reagents and conditions: (i) Ac2O/ pyridine; (ii). Acetyl chloride, HCl(g), 40°C, 18
h; (iii) HgBr2, 1,2-DCE, 90°C, 20 h; (iv) H2/Pd-C 10%, MeOH, 0.5 h.
Scheme 3. Solid phase synthesis of glycopeptides 3–BSA and 4–BSA. Reagents and conditions: i. 50%
morpholine in DMF; ii. Fmoc-ProOH, PyBOP, HOBt, DIPEA; iii. Fmoc-AlaOH, PyBOP, HOBt, DIPEA;
iv. Fmoc-Arg(Pbf)OH, PyBOP, HOBt, DIPEA; v. αGalNAc-ThrOH 5 or βGalNAc-ThrOH 6, PyBOP,
HOBt, DIPEA; vi. Fmoc-Asp(OtBu)OH, PyBOP, HOBt, DIPEA; vii. Fmoc-SerOH, PyBOP, HOBt,
DIPEA; viii. Ac2O/ pyridine; ix. TFA 80%; x. NaOMe, MeOH; xi. BSA, EDCI, NHS, Na2HPO4.
Figure 1. Sera titration from mice immunized with glycopeptides A) 3–BSA and B) 4–BSA by ELISA.
OD450nm: optical density at 450 nm. ■: Immunizations with glycopeptides 3–BSA and 4–BSA and coating
with corresponding glycopeptides 3–BSA and 4–BSA.▼: Immunizations with glycopeptides 3–BSA and
4–BSA and coating with BSA. ●: Immunizations with BSA and coating with BSA. ▲: Preimmune (PI)
sera and coating with corresponding glycopeptides 3–BSA and 4–BSA.
Figure 2. Flow cytometry analysis of the binding (percentages of positive staining) of glycopeptides 3–
BSA and 4–BSA antisera to MCF-7 cells. A) MCF-7 cells incubated with preimmune (PI) and anti-MCF-
7 sera (1:1000). B) MCF-7 cells incubated with glycopeptide 3–BSA antisera at increasing dilutions of
1:400 to 1:8000. C) MCF-7 cells incubated with glycopeptide 4–BSA antisera at increasing dilutions of
1:400 to 1:8000.
Figure 3. Percentages of positive staining, obtained by flow cytometry, of glycopeptides 3– and 4–BSA
antisera to MCF-7 cells. MCF-7 cells incubated with anti-MCF-7 sera (black bar) and preimmune (PI)
sera (white bars) at 1:1000 dilution, and with glycopeptide 3–BSA antisera (light grey bars) and
glycopeptide 4–BSA antisera (dark grey bars) at increasing dilutions of 1:400 to 1:8000.
Figure 4. Binding between anti-αGalNAc and anti-βGalNAc immunopurified antibodies to coated BSA-
conjugated glycopeptides 3 and 4, and BSA, at 10, 1 and 0.1 µg/mL.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
Figure 5. (A) Overall view of the simulated system. The SM3 monoclonal antibody chains A and B are
shown as green and cyan cartoons, respectively. The MUC1-derived glycopeptide is shown as spheres.
Yellow indicates a carbon atom, blue indicates nitrogen and red an oxygen. (B,C) Detailed view showing
the difference in the carbohydrate ring orientation between the α and βGalNAc glycopeptides 3 and 4
shown as sticks.
Figure 6. (A) Detailed view of the interaction between the carbohydrate moiety of the αGalNAc-
glycopeptide 3. Its carbohydrate moiety interacts with Tyr-32L and Asn-53L via hydrogen bonds. (B)
Detailed view of the interaction between the carbohydrate moiety of the βGalNAc-glycopeptide 4. Its
carbohydrate moiety also interacts with Tyr-32L and Asn-53L via hydrogen bonds.
Figure 7. Binding sensorgrams relative to the interactions of glycopeptide 3 (α) (A) and glycopeptide 4
(β) (B) with immobilized SM3 MAb. Glycopeptides were injected at increasing concentrations between
50 nM and 2 µM.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
SCHEME 1
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600473
ChemBioChem
SCHEME 2
This article is protected by copyright. All rights reserved.