One-pot: N -glycosylation remodeling of IgG with non-natural sialylglycopeptides enables glycosite-specific and dual-payload antibody-drug conjugates

Organic &

Biomolecular Chemistry

View Article Online

View Journal

PAPER

One-pot N-glycosylation remodeling of IgG with

non-natural sialylglycopeptides enables

glycosite-speciﬁc and dual-payload

antibody–drug conjugates†

Cite this: DOI: 10.1039/c6ob01751g

Feng Tang,^a,bYang Yang,^a,cYubo Tang,^a,dShuai Tang,^aLiyun Yang,^a,b

Bingyang Sun,^a,cBofeng Jiang,^a,cJinhua Dong,^dHong Liu,^aMin Huang,^a

Mei-Yu Geng^aand Wei Huang*^a,c

Chemoenzymatic transglycosylation catalyzed by endo-S mutants is a powerful tool for in vitro glyco-

engineering of therapeutic antibodies. In this paper, we report a one-pot chemoenzymatic synthesis of

glycoengineered Herceptin using an egg-yolk sialylglycopeptide (SGP) substrate. Combining this one-pot

strategy with novel non-natural SGP derivatives carrying azido or alkyne tags, glycosite-speciﬁc

conjugation was enabled for the development of new antibody–drug conjugates (ADCs). The site-speciﬁc

ADCs and semi-site-speciﬁc dual-drug ADCs were successfully achieved and characterized with

SDS-PAGE, intact antibody or ADC mass spectrometry analysis, and PNGase-F digestion analysis. Cancer

cell cytotoxicity assay revealed that small-molecule drug release of these ADCs relied on the cleavable

Val-Cit linker fragment embedded in the structure. These results represent a new approach for glycosite-

speciﬁc and dual-drug ADC design and rapid synthesis, and also provide the structural requirement for

their biologic activities.

Received 11th August 2016,

Accepted 2nd September 2016

DOI: 10.1039/c6ob01751g

www.rsc.org/obc

domain is involved in recruitment of immune cells by binding

to Fc receptors on the immune cell surface, and helps direct

Introduction

Therapeutic IgG-type monoclonal antibodies (mAbs) have the immune response^5–9including phagocytosis, immune cell

attracted continuous research attention during the past three activation, and cytokine stimulation towards the antigen-dis-

decades. Among the FDA approved bio-drugs, over one third is played targets. N-Glycosylation on Asn²⁹⁷of IgG Fc mediates Fc

mAb, applied in clinic treatments on cancer, autoimmune receptor interaction and regulates antibody-dependent eﬀector

diseases, and infectious diseases.^1–4IgG mAb consists of two functions.^9–14Core-fucosylation of Fc N-glycans hindered the

heavy chains and two light chains, forming two Fab (antigen binding with an FcγIIIa receptor and the lack of core-fucose

binding) domains and one Fc (crystallizable) domain. The Fc would dramatically enhance antibody-dependent cell-mediated

cytotoxicity (ADCC).^12,13,15Crystallization studies on the

Fc–FcγIIIa complex revealed a unique carbohydrate–carbo-

^aCAS Key Laboratory of Receptor Research, CAS Center for Excellence in Molecular

hydrate interaction between afucosylated N-glycans of Fc and

Cell Science, Shanghai Institute of Materia Medica, Chinese Academy of Sciences,

the FcγIIIa glycans that could be blocked by the core-fucose of

555 Zuchongzhi Road, Pudong, Shanghai, China 201203.

Fc N-glycans.¹²Galactose on the non-reducing end of

E-mail: huangwei@simm.ac.cn; Fax: +86-21-50807088;

Fc N-glycans improves the ADCC and complement-dependent

cytotoxicity (CDC) activity of IgG antibodies.^16,17It is reported

that the terminal sialic acid moieties of Fc N-glycans are

crucial for the anti-inflammatory activity of intravenous

immunoglobulin (IVIg).^14,18–20N-Glycosylation also aﬀects

the pharmacokinetics and immunogenicity of therapeutic

mAbs.^21–24

Massive data have demonstrated the significant modulation

eﬀect of N-glycans on IgG functions, therefore glyco-

engineering technologies have been developed and applied in

structural optimization on IgG glycans for better therapeutic

Tel: +86-21-20231000 ext. 2517

^bUniversity of Chinese Academy of Sciences, No. 19A Yuquan Road, Beijing 100049,

China

^ciHuman Institute, ShanghaiTech University, 99 Haike Road, Pudong, Shanghai,

201210 China

^dKey Laboratory of Structure-Based Drug Design and Discovery, Ministry of

Education, Shenyang Pharmaceutical University, Shenyang, 110016, China

†Electronic supplementary information (ESI) available: Detailed structures of

10a–d, 21a–d, payloads, and other small molecules; schemes of PNGase-F diges-

tion and ADC conjugation; SDS-PAGE and LC-MS of 7a–d, 14a–e, T-DM1,

T-MMAE, and other small molecules; synthetic procedures of 10a–d, 21a–d, pay-

loads, and other small molecules; NMR spectra of all intermediates. See DOI:

10.1039/c6ob01751g

This journal is © The Royal Society of Chemistry 2016

Org. Biomol. Chem.

View Article Online

Paper

Organic & Biomolecular Chemistry

eﬃcacy. In a review on glycoengineered mAbs,²¹the clinic the galactosylated IgG N-glycan and conjugate a cytotoxic

study status, glycoengineering technologies, and the func- agent (doxorubicin) on the pre-assembled azido group via

tional properties of over 15 glycoengineered IgG drug candi- the “click chemistry” of cycloaddition reaction. Zhou et. al.⁴⁷

dates were well summarized. The common technology to reported a similar strategy to attach natural sialic acid onto

control IgG glycosylation is to interfere with the biosynthetic Fc glycan and followed with NaIO₄oxidation to give an alde-

pathway of N-glycosylation in particular expression hyde tag on the N-glycan for site-specific conjugation. The

systems,^15,25–29combined with glycosyltransferase modulation chemical oxidation of an IgG antibody in this method is a very

such as knock-out of the fucosyltransferase FUT8 for tricky step that usually causes oxidative side reaction and

elimination of the core-fucose^28,29and up-regulation of bisect- protein denaturing.

A

galatosyltransferase mutant was

ing GlcNAc transferase GnT-III,³⁰or in vitro enhanced employed for glycosite-specific conjugation⁴⁸by transfering

galactylation³¹and sialylation³²to obtain the desired the keto-tagged⁴⁹or azido-tagged⁵⁰galactose onto the de-

N-glycans. Recently, several research groups have reported a galatosylated G0F glycoform or deglycosylated Fucα1,6GlcNAc

chemoenzymatic method for eﬃcient glycosylation remodeling disaccharide form of IgG.⁵¹These methods demonstrated a

of IgG N-glycans with a mutated endo-N-acetylglucosaminidase new avenue of glycosite-specific ADC development with a

from Streptococcus pyogenes (endo-S).^16,33,34This method precise structure and DAR.

features with two key steps: (1) trimming of natural glyco-

The glycosyltransferase-catalyzed glycosite-specific conju-

forms, mainly containing G0F, G1F, and G2F, with wild-type gation with a non-natural UDP-monosaccharide substrate was

endo-S to cleave the glycosidic bond of a GlcNAcβ1,4GlcNAc always restrained on the sugar structures with a narrow sub-

motif and give the deglycosylated IgG with the innermost strate specificity.⁵²To obtain structural diversity for the screen-

Fucα1,6GlcNAc disaccharide or further defucosylated GlcNAc ing and development of an ADC drug candidate, we sought to

monosaccharide; (2) endo-S mutant-catalyzed chemoenzymatic take advantage of endo-S for IgG N-glycosylation remodeling

transglycosylation of a chemically synthetic N-glycan oxazoline with various non-natural glycan substrates. Wang L. X. and

as the donor substrate onto Fucα1,6GlcNAc-IgG or GlcNAc-IgG, co-workers reported¹⁶that endo-S D233Q/A could construct IgG

giving a glycoengineered antibody with a well-defined N-glycan bearing biantennary complex-type N-glycans in a wide range

structure. The substrate specificity of endo-S covers a quite including sialylated, galactosylated, fucosylated, bisecting-

range of biantennary complex type, mannose type (5 mannose GlcNAc, and azido-tagged Man3GlcNAc cores. Wong C. H’s

moieties or less) and hybrid type N-glycans, but endo-S recent work³⁴has also indicated the wide substrate specificity

restrains on its acceptor substrate specifically for the IgG of this enzyme on glycan structures. In the current work, we

Fc domain that makes it a perfect enzyme for IgG in vitro employed a sialylglycopeptide extracted from egg yolks as the

glycoengineering.

precursor^53,54and synthesized a series of azido- or alkyne-

Antibody–drug conjugates (ADCs) covalently connect a tagged sialylglycopeptides for IgG glycosylation remodeling.

small-molecule drug onto an antibody via a suitable linker An eﬃcient one-pot method for the assembly of chemical

fragment. Through a high targeting eﬀect based on antibody– handles on IgG Fc N-glycans was achieved by a combination of

antigen specific recognition, the loaded drug was delivered three steps in one reaction cell: (1) release of the free glycans

into the antigen-displayed target cells by endocytosis.³⁵ADCs from these non-natural sialylglycoppeptide derivatives with

have shown potent therapeutic eﬃcacy in clinic studies on endo-N-acetylglucosaminidase from Mucor hiemalis (endo-M);

treatment of various tumors.³⁵Antitumor ADCs require the (2) conversion of the free glycans to their oxazoline forms as

antibody with specific recognition on the over-expressed the enzyme substrate; (3) transfer of the glycan oxazolines to

tumor biomarker in high binding aﬃnity, and a highly potent Fucα1,6GlcNAc-IgG catalyzed by endo-S D233Q. Very recently,

payload such as the cytotoxic agents DM1 and MMAE.^36,37The during the preparation of this manuscript, we noticed that

linker is also very important for remaining stable in blood flow Davis B. G. and his co-workers have reported⁵⁵another series

while being rapidly released after being delivered into the of non-natural N-glycans for IgG glycoengineering via a similar

tumor cells.³⁸The launched ADC drugs or ADC candidates in endo-S-catalyzed method but with diﬀerent glycan derivatives

clinic trials mainly conjugated at the Lys or Cys amino acids of that further demonstrated the power of chemoenzymatic IgG

the antibodies via a random way. The potential problem of glycosylation engineering and broad glycan substrate

random conjugation is apparent on drug quality control, recognition of endo-S on various non-natural structures.

precise drug-antibody ratio (DAR) control, bio-activity repeat- Furthermore, with the glycoengineered IgG carrying chemical

ability, etc. Site-specific conjugation of ADCs has become the handles, we performed the site-specific conjugation of small-

hot topic in this field.^39–41The approaches via incorporation molecule drugs via various linker fragments. The antitumor

of free Cys or non-natural amino acids at the specific sites of assay results indicated that a suitable cleavable linker structure

an antibody sequence for payload conjugation have been is extremely important for the activity of glycosite-specific ADCs.

reported.^42–45Besides these technologies, site-specific conju- We also provide a novel strategy for dual-drug ADC^56,57design by

gation on IgG Fc N-glycans provided an alternative strategy and connection of one kind of small-molecule drug on the glycans

attracted research interest. Boons G. J. established an elegant and another diﬀerent drug on the Lys residues via random con-

method⁴⁶for glycosite-specific ADC preparation using sialyl- jugation. The dual-drug ADC design enables employment of

transferase to attach an azido-tagged sialic acid moiety onto dual cytotoxic mechanisms such as simultaneous anti-tublin

Org. Biomol. Chem.

This journal is © The Royal Society of Chemistry 2016

View Article Online

Organic & Biomolecular Chemistry

Paper

and DNA damaging, that could be potentially useful for better was extended to prepare other bi-antennary N-glycan oxazolines

eﬃcacy or for drug-resistance cancer treatment.

3–5 by combined steps of exo-glycosidase trimming during the

procedure (Scheme 1). Neuraminidase, β-galactosidase, and β-N-

acetylglucosaminidase (GlcNAcase) were employed to remove

the non-reducing end saccharides of the released N-glycan from

SGP in one reaction cell and followed with oxazoline formation

reaction to give the respective purified Gal₂GlcNAc₂Man₃GlcNAc

oxazoline (3), GlcNAc₂Man₃GlcNAc oxazoline (4), and

Man₃GlcNAc oxazoline (5). High-performance anion-exchange

chromatography with pulsed amperometric detection

(HPAEC-PAD) monitoring on each step of enzymatic deglycosy-

lation and oxazoline formation was performed to ensure the

complete sugar truncation for homogeneous glycan oxazoline

production (see the ESI†).

The rapid semi-synthesis of N-glycan oxazolines 2–5 by this

one-pot approach provided us the donor substrate pool for

chemoenzymatic glycosylation remodeling of IgG following

the reported procedure.¹⁶We chose commercial mAb drug

Herceptin as the model and removed its major glycoforms of

G0F, G1F, and G2F by wild-type endo-S to give Herceptin-

Fucα1,6GlcNAc (6) as the acceptor substrate. Chemoenzymatic

transglycosylation catalyzed by endo-S D233Q using substrate

6 and oxazolines 2–5 oﬀered glycoengineered Herceptin 7a–d

bearing homogeneous glycoform S2G2F (7a), G2F (7b), G0F

(7c), and M3F (7d). The products were characterized with

SDS-PAGE and LC-MS analysis of the intact IgG (Fig. S1–S5†).

In Table 1, the deconvolution MS data of intact 7a–d are sum-

marized and indicate the high accordance of measured and

calculated mass.

Results and discussion

One-pot strategy for chemoenzymatic IgG glycosylation

remodeling

endo-Glycosidase-catalyzed chemoenzymatic synthesis was

recently established for preparation of homogeneous

N-glycopeptides,

glycoproteins,

and

other

glycoconjugates.^16,58–66In this approach, N-glycan oxazolines

were used as the glycosylation donor substrates of wild-type or

mutated endo-glycosidases to perform transglycosylation onto

a GlcNAc-containing acceptor substrate. Herein, we sought to

develop a one-pot strategy for rapid synthesis of N-glycan

oxazolines and for glycoengineering of antibody substrates.

As shown in Scheme 1, we started with the one-pot semi-

synthesis of various bi-antennary N-glycan oxazolines using

the egg-yolk sialylglycopeptide (SGP, 1)^53,54as the precursor.

SGP

from Mucor hiemalis (endo-M) to cleave the glycosidic bond

of GlcNAcβ1,4GlcNAc motif and release the glycan,

was

treated

with

endo-N-acetylglucosaminidase

a

Sia₂Gal₂GlcNAc₂Man₃GlcNAc, from the hexapeptide. The

in situ glycan was then converted to oxazoline 2 in the presence

of 2-chloro-1,3-dimethylimidazolinium chloride (DMC) and

triethylamine following Shoda’s method.⁶⁷The crude oxazo-

line from one-pot synthesis was then subjected to solid-phase

extraction (SPE) via a porous graphitic carbon (PGC) column

and gel-filtration via a Sephadex G-25 column to give the

purified product. Application of this one-pot semi-synthesis

The successful one-pot semi-synthesis of N-glycan oxazo-

lines encouraged us to further develop a one-pot method for

Scheme 1 Conditions: (a) endo-M, phosphate buﬀer (PB), pH 6.5; (b) Neuraminidase, PB, pH 5.0; (c) β-galactosidase, PB, pH 6.5; (d) β-GlcNAcase,

PB, pH 6.5; (e) DMC/Et₃N, ice bath; (f) Herceptin-Fucα1,6GlcNAc (6), endo-S D233Q, Tris buﬀer, pH 7.4.

This journal is © The Royal Society of Chemistry 2016

Org. Biomol. Chem.

View Article Online

Paper

Organic & Biomolecular Chemistry

Table 1 Glycoforms and MS data of glycoengineered Herceptin

Entry

Compound

Glycoforms

Calculated mass (Dalton)

Found mass (Dalton)

1

2

3

4

5

6

7

8

9

10

6

7a

7b

7c

Fucα1,6GlcNAc

S2G2F

G2F

145806.06

149809.45

148645.06

147996.85

147184.54

149897.47

149897.62

150181.62

150001.49

149773.41

145805.35

149810.52

148656.89

147997.27

147185.75

149890.43

149906.31

150191.82

150002.00

149772.38

G0F

7d

Man3GlcNAc2F

Azido-oxime-S2G2F

Azido-amine-S2G2F

Azido-amide-oxime-S2G2F

Alkyne-amide-oxime-S2G2F

Alkyne-oxime-S2G2F

14a

14b

14c

14d

14e

chemoenzymatic transglycosylation using in situ oxazolines. after endo-M digestion did not aﬀect the chemoenzymatic reac-

As shown in Scheme 1, in situ oxazoline 2 obtained from SGP tion because the endo-S and its mutants specifically catalyze

by successive endo-M digestion and DMC/Et₃N treatment was the IgG Fc fragment other than other GlcNAc-containing sub-

directly incubated with Herceptin-Fucα1,6GlcNAc (6) and endo- strates.¹⁶DMC as a coupling reagent raises concern of chemi-

S D233Q in a Tris buﬀer (pH 7.4). SDS-PAGE monitoring cal side-reactions, but in the step of oxazoline formation the

(Fig. 1A) implicated the eﬃcient chemoenzymatic reaction majority of DMC was converted to its inactive derivative 1,3-

giving product 7a in nearly complete conversion rate within dimethyl-2-imidazolidinone (DMI) in NMR monitoring as

30 min. LC-MS analysis (Fig. 1B) of 7a via this one-pot pro- reported.^67,69In our one-pot procedure, there was no cross-

cedure also indicated the homogeneous S2G2F glycoform of linking side-product observed and adding of glycine as the

the product in both multiple charged and deconvoluted MS scavenger for DMC into the reaction did not exhibit a diﬀerence

spectra. This result demonstrated that the one-pot chemo- in the reaction results. Also, the activity of the endo-S mutant

enzymatic synthesis of glycoengineered IgG could be success- was not aﬀected in the presence of DMI and triethylamine salts.

fully and eﬃciently achieved despite the presence of endo-M, The data clearly showed that the one-pot strategy for the chemo-

GlcNAc-hexapeptide, DMC derivative, and triethylamine salts enzymatic synthesis of glycoengineered IgG was compatible

in the reaction buﬀer. In this case, endo-M does not cleave the with the reagents used for oxazoline preparation and a single-

fucosylated N-glycan⁶⁸and thus would not cause the reverse step purification of the product via a protein A aﬃnity column

deglycosylation of glycoengineered IgG. Additionally, we could easily aﬀord the pure glycoengineered IgG.

noticed that the hydrolytic activity of endo-M was completely

The non-enzymatic attachment of N-glycan oxazoline onto

lost after adding DMC/Et₃N and barely recovered after the next- the protein occurred in previous reports^55,69especially in the

step, which involved neutralization of the in situ oxazoline case of high-concentration substrates. Therefore, we investi-

with Tris buﬀer. The GlcNAc-hexapeptide released from SGP gated the non-enzymatic glycation reaction between in situ oxa-

zoline 2 and acceptor 6. Oxazoline 2 in 1.0, 1.5, 2.5, and

5.0 mM was incubated with 6 in the absence of the endo-S

mutant and monitored by SDS-PAGE after 1 h, 2 h, 3 h, 4 h,

and 12 h (Fig. S6†). After 12 h, the non-enzymatic glycation

side-products were observed in the reaction under all con-

ditions, however, the higher concentration vial exhibited much

more glycation product. Within 3 h, reactions with 1.0, 1.5,

and 2.5 mM of 2 caused very few non-enzymatic glycation

products. Based on this result, we optimized the one-pot chemo-

enzymatic reaction conditions using oxazoline in 1.5 mM

and performing transglycosylation for 1–2 h to avoid the non-

enzymatic side-reactions. The glycoengineered Herceptin

products were achieved in high yield under these conditions

and the products were further characterized by LC-MS analysis

after PNGase-F digestion to ensure the right linkage of target

products as discussed below (Fig. 2, 5, and Scheme S1†).

Fig. 1 SDS-PAGE and LC-MS proﬁles of glycoengineered Herceptin 7a

via one-pot chemoenzymatic synthesis. (A) SDS-PAGE analysis of one-

pot chemoenzymatic transglycosylation of Herceptin-Fucα1,6GlcNAc (6)

IgG glycoengineering with non-natural N-glycan substrates

and in situ oxazoline 2 catalyzed by endo-S D233Q. Lane 0: marker,

Lane 1: Herceptin-Fucα1,6GlcNAc (6), Lane 2–5: one-pot reaction moni-

toring after 0.5 h, 1 h, 2 h, and 3 h. (B) LC-MS proﬁle of intact glycoengi-

neered Herceptin 7a. The multiple charged m/z data are labeled with

charge numbers. The deconvolution mass spectrum is shown in the

embedded box.

The eﬃcient one-pot chemoenzymatic synthesis of glycoengi-

neered IgG enables rapid glycosylation remodeling of mAb

using the N-glycan precursor SGP directly. With this estab-

lished approach, we sought to introduce chemical tags on

Org. Biomol. Chem.

This journal is © The Royal Society of Chemistry 2016

View Article Online

Organic & Biomolecular Chemistry

Paper

Fig. 2 SDS-PAGE and LC-MS characterization on glycoengineered Herceptin 14a–e bearing non-natural N-glycans. (A) SDS-PAGE analysis of 14a–e.

Lane 0: marker, Lane 1: commercial Herceptin, Lane 2: Herceptin-Fucα1,6GlcNAc (6), Lane 3–7, 14a–e; (B) PNGase-F digestion of 14a–e. Lane 0:

marker, Lane 1: commercial Herceptin, Lane 2, 4, 6, 8, 10: 14a–e, Lane 3, 5, 7, 9, 11: 14a–e after PNGase-F digestion; (C) LC-MS proﬁles of 14a. The

multiple charged m/z data are labeled with charge numbers. The deconvolution mass spectrum is shown in the embedded box. (D) LC-MS proﬁles

of Herceptin–Asp²⁹⁷resulted from PNGase-F-treated 14a. (E) MS proﬁle of the released glycan 28 from 14a by PNGase-F digestion. The ion ﬂow

chromatography of the PNGase-F digested sample is shown in the embedded box and the peak of 28 is marked with an arrow.

IgG glycans by non-natural glycan substrates for site-specific

Next, we applied the one-pot method for chemoenzymatic

conjugation. To test this idea, we synthesized a series of modi- transglycosylation of Herceptin with these non-natural glycan

fied SGPs as the precursors of glycan donors. As shown in substrates. The in situ oxazoline 13 derived from azido-SGP

Scheme 2, SGP (1) was treated with NaIO₄to oxidize the triol 11a was directly incubated with Herceptin acceptor 6 and the

of sialic acid motifs giving the CHO-SGP (8). With the aldehyde endo-S mutant in a Tris buﬀer. SDS-PAGE monitoring and

function group, azido or alkyne tags were easily assembled LC-MS detection (Fig. S7†) indicated that product 14a was

onto 8 via reductive amidation with amines (9a–b) or oxime obtained eﬃciently in a quantitative conversion rate after one

formation with hydroxylamines (10a–d). Three azido-SGPs hour. To test whether the non-enzymatic glycation took place

(11a–c) and three alkyne-SGPs (11d–f) were prepared for simultaneously, we performed glycan analysis of 14a by

further glycoengineering of Herceptin. Characterization of PNGase-F digestion. Chemoenzymatic transglycosylation cata-

11a–f with NMR, MS, and HPLC is provided in the ESI.†

lyzed by the endo-S mutant assembles the N-glycan onto the

To investigate the substrate tolerance of the endo-S enzyme first GlcNAc on IgG Fc Asn²⁹⁷via a beta 1,4 linked glycosidic

on these sialyl modifications, we chose the azido-SGP 11a as bond, whereas the non-enzymatic glycation takes place ran-

the starting material and synthesized the corresponding domly on the amino acids of protein. PNGase-F cleaves the

N-glycan oxazoline 13 by a similar procedure described above. amide bond between the nitrogen of the N-glycan reducing

endo-M digestion of 11a released the azido N-glycan 12 which end and the side-chain carbonyl of glycosite Asn, and then

was purified by semi-preparative HPLC. 12 was converted to its releases the intact N-glycan and converts the Asn to Asp

oxazoline form 13 and subjected to gel-filtration purification (Scheme S1†). Therefore, the enzymatic product will be degly-

through a Sephadex G-25 column. The assignments on the cosylated by PNGase-F and the non-enzymatic side-product

NMR spectra of 12 and 13 are available in the Experimental will not. As shown in SDS-PAGE of Fig. 2B, the heavy-chain

section and ESI.† With azido N-glycan oxazoline 13 in hand, band of PNGase-F-treated 14a totally shifted to a downside

we performed the chemoenzymatic transglycosylation with position suggesting the complete release of N-glycan, and

Herceptin-Fucα1,6GlcNAc (6) catalyzed by endo-S D233Q. The there was no non-enzymatic glycation observed. Furthermore,

result indicated that the endo-S mutant recognized the azido- LC-MS analysis of the resulting Herceptin–Asp²⁹⁷(Fig. 2C) and

tagged oxazoline 13 as the same as the native N-glycan oxazo- azido-N-glycan 28 (Fig. 2D) clearly verified the enzymatic

line 2 and aﬀorded the azido-Herceptin 14a in a quantitative product that linked the N-glycan onto the GlcNAc motif of

yield within 1 h. MS analysis of intact 14a showed the decon- 6 but not onto the amino acid residues.

voluted mass of 149 890.43 (Fig. 2B) which was in agreement

with the calculated mass.

The successful one-pot glycoengineering of Herceptin with

azido-tagged SGP demonstrated the broad substrate specificity

This journal is © The Royal Society of Chemistry 2016

Org. Biomol. Chem.

View Article Online

Paper

Organic & Biomolecular Chemistry

Scheme 2 Conditions: (a) NaIO₄, ice bath; (b) 3-azido-1-propanamine (9a) or propargylamine (9b)/NaCNBH₃, phosphate buﬀer, pH 6.0;

(c) hydroxylamine 10a–d (see the ESI†), PB, pH 7.5; (d) endo-M, PB, pH 6.5; (e) DMC/Et₃N; (f) 6, endo-S D233Q, Tris buﬀer, pH 7.4.

of the endo-S mutant with extensive tolerance on sialyl modi- fluorescent tag Cy5 (19) to give the corresponding DBCO-

fication of biantennary complex-type N-glycans. We then tested tagged compounds 21a–d (see the Experimental section and

the other four non-natural N-glycan substrates 11b–11e ESI† for chemistry details). Using the azido-Herceptin (14a)

bearing diverse linker fragments for one-pot glycoengineering. and DBCO-MCC-DM1 (21a) as the substrates, we performed

The SDS-PAGE monitoring and LC-MS analysis on intact glyco- the copper-free click reaction in a phosphate buﬀer (pH 7.5)

engineered products are illustrated in the ESI (Fig. S8–S11†). containing 10% DMSO and synthesized the site-specific ADC

The SDS-PAGE of purified products 14b–e is summarized in 22a. The cycloaddition conjugation was accomplished eﬃcien-

Fig. 2A, and PNGase-F digestion of each glycoengineered tly as monitored by SDS-PAGE and LC-MS (Fig. 3A and 4A).

Herceptin indicated no non-enzymatic side-product in SDS-PAGE analysis of 22a showed its heavy-chain band in a

SDS-PAGE detection (Fig. 2B). All five non-natural N-glycan higher position than 14a but light-chain bands of both

substrates suggested similar eﬃciency in one-pot glyco- samples remained in the same size, suggesting that the gained

engineering of Herceptin. With this method, we can easily mass by toxin conjugation was site-specific on the heavy chain.

label the N-glycan of IgG for site-specific conjugation.

LC-MS verified the successful conjugation of totally four

DM1 moieties on the four azido groups of two Fc N-glycans.

Glycosite-specific ADCs with diverse linker fragments were

synthesized following the same strategy. Azido-oxime-

Glycosite-specific antibody–drug conjugations via the

tagged-glycoforms of IgG

Azido or alkyne-tagged Herceptin 14a–e enable the small- Herceptin (14a) and azido-amine-Herceptin (14b) were conju-

molecule conjugation via the azide–alkyne Huisgen cyclo- gated with other two DBCO-tagged toxins 21b and 21c bearing

addition, also known as “click chemistry”. We initiated the a cleavable Val-Cit dipeptide and/or a poly(ethylamine) (PEG)

conjugation with copper-catalyzed click reaction using alkyne- motif. Corresponding ADCs 22b–d were obtained and the

Herceptin 14d–e and azido-propanamine (9a) as the model increased mass by conjugation was confirmed by SDS-PAGE

materials. Under catalytic conditions of CuSO₄, L-ascorbic acid, (Fig. 3B and C) and LC-MS (Fig. 4B–E). The MS data and linker

and ligand BTTAA,⁷⁰the click reaction did not proceed information of these site-specific ADCs are also summarized

smoothly. Precipitation of the IgG protein was observed in the in Table 2. To validate the site-specific conjugation, 22a–c were

reaction system and LC-MS monitoring showed a low conver- treated with PNGase-F for deglycosylation. Fig. 5A shows the

sion rate. Then, we turned to the strain-promoted copper-free heavy-chain bands of all three ADCs reduced to the lower posi-

approach using cyclooctyne reagents^71–73for enhanced click tion as the same level as deglycosylated commercial Herceptin.

reactions. Dibenzocyclooctyne (DBCO) was assembled on LC-MS measurements on the released glycan–drug complex 29

small molecules such as toxin DM1 (15) and MMAE (16), or displayed a series of charged m/z data in high accordance with

Org. Biomol. Chem.

This journal is © The Royal Society of Chemistry 2016

View Article Online

Organic & Biomolecular Chemistry

Paper

Fig. 3 SDS-PAGE proﬁles of site-speciﬁc ADCs. (A) ADC 22a. Lane 0: marker, Lane 1: commercial Herceptin, Lane 2: Herceptin 6, Lane 3: Herceptin

14a, Lane 4: ADC 22a. (B) ADCs 22b and 22c. Lane 0: marker, Lane 1: commercial Herceptin, Lane 2: Herceptin 6, Lane 3: Herceptin 14a, Lane 4:

ADC 22b, Lane 5: ADC 22c. (C) ADCs 22d and 22e. Lane 0: marker, Lane 1: commercial Herceptin, Lane 2: Herceptin 6, Lane 3: Herceptin 14a, Lane 4:

ADC 22d, Lane 5: ADC 22e. (D) ADCs 23 and 24. Lane 0: marker, Lane 1: commercial Herceptin, Lane 2: Herceptin 6, Lane 3: Herceptin 14a, Lane 4:

ADC 23, Lane 5: ADC 24.

Fig. 4 Mass spectra of site-speciﬁc ADCs. (A) ADC 22a. The multiple charged m/z data are labeled with charge numbers. The deconvolution mass

spectrum is shown in the embedded box. (B) ADC 22b. (C) ADC 22c. (D) ADC 22d. (E) ADC 22e. (F) The deconvolution mass spectrum of Lys-conju-

gated ADC 23. The conjugation number of DM1 for each detected mass was assigned and marked with red-colored numbers. (G) The deconvolution

mass spectrum of dual-drug ADC 24. (H) The deconvolution mass spectrum of 25 bearing Lys-conjugated DM1 and glycan-labeled Cy5.

This journal is © The Royal Society of Chemistry 2016

Org. Biomol. Chem.

View Article Online

Paper

Organic & Biomolecular Chemistry

Table 2 Chemical proﬁles and antitumor activities of ADC samples

DAR^b

Found mass

EC₅₀

g

Entry

ADCs

Precursors

Glycosite linkers

Payload

1

2

3

4

5

6

7

8

9

T-DM1

T-MMAE

22a

22b

22c

22d

22e

MCC (on Lys)

DBCO-MCC

DBCO-PEG₄

DBCO-PEG₄-VC-PAB

DBCO-PEG₄

DBCO-PEG₄-VC-PAB

MCC (on Lys)

DBCO-PEG₄-VC-PAB

DM1

MMAE

DM1

MMAE

DM1

MMAE

DM1

3.2

3.5

4

Multi-mass^d

154847.29

155684.74

156659.32

155688.26

156643.96

Multi-mass^e

Multi-mass^f

0.06

0.15

>10

0.09

>10

ND

14a

14b

14a

23

4

3.8^c

4

3.8^c

2.6

3.8^c

23

ND

0.06

24^a

MMAE

^aDual-drug ADC 24 contains DM1 conjugated on Lys residues and MMAE conjugated on glycans with two DAR values. ^bThe DAR values are cal-

culated based on the intensity percentage of each found mass of ADC in the MS spectra. ^cDBCO-PEG4-VC-MMAE 21c used in preparation of

these ADCs contained a small quantity of the DBCO fragment without MMAE and caused the DAR reduction. ^dSee the ESI. ^eThe found multiple

mass in deconvolution data: 149907.63 (+0 DM1), 147719.04 (+1 DM1), 151823.00 (+2 DM1), 152786.70 (+3 DM1), 153734.55 (+4 DM1), 154696.10

(+5 DM1). ^fThe found multiple mass in deconvolution data: 157606.71 (+1 DM1), 158566.82 (+2 DM1), 159530.22 (+3 DM1), 160477.23 (+4 DM1),

161446.46 (+5 DM1). ^gThe EC₅₀values (μg mL⁻¹) of antitumor activity are calculated based on the cytotoxicity assay with the SK-BR-3 cell line.

ND: not detected.

Fig. 5 SDS-PAGE and LC-MS analysis of site-speciﬁc ADCs after PNGase-F digestion. (A) SDS-PAGE of ADCs 22a–c treated with PNGase-F. Lane 0:

marker; Lane 1: commercial Herceptin; Lane 2, 4, 6: 22a, 22b, 22c; Lane 3, 5, 7: 22a, 22b, 22c after PNGase-F digestion. (B) HR-MS proﬁle of the

released glycan-drug moiety 29 from ADC 22c.

the calculated molecular weight (Fig. 5B). These data solidly on the intensity percentage of each found mass in all deconvo-

demonstrated that the glycosite-specific conjugation method lution data. Dual-drug ADC 24 has two DARs: DAR of DM1 on

was successfully applied in ADC preparation with diverse Lys residues was 2.6 as the same as 23; DAR of MMAE on

linker structures.

glycans was 3.8 (Table 2).

Furthermore, combining this glycosite-specific conjugation

The azido-tagged IgG also enables site-specific labeling

and conventional random conjugation, we developed a dual- with fluorescent reagents. DBCO-Cy5 (21d) was linked to the

payload strategy for ADC design by assembly of one payload azido-glycan of 14a giving site-specific Cy5-labeled Herceptin

(MMAE) on glycans and another diﬀerent payload (DM1) on 22f (Fig. S12†). This label could be applied in conventional

Lys residues. As shown in Scheme 3, azido-Herceptin 14a was ADC as well. Random linked ADC 23 bearing azido moieties

treated with DM1-SMCC (20) bearing an active carboxylic succi- on glycans was labeled with 21d and aﬀorded site-specific

nimide for Lys conjugation, to give the random linked ADC 23 Cy5-labeled ADC 25 (Fig. 4H). Fluorescent labeling of IgG or

carrying azido-tagged glycans. Then, DBCO-PEG4-VC-MMAE ADC on N-glycans represents another functionalization appli-

(21c) was introduced on the azido groups of 23 via click reac- cation of chemoenzymatic glycoengineering with non-natural

tion and aﬀorded dual-drug ADC 24. The SDS-PAGE analysis glycan substrates.

(Fig. 3D) of 23 and 24 indicated the smeared bands of both

heavy chains and light chains because of the random conju-

Biological assays with site-specific conjugates

gation. The LC-MS spectra (Fig. 4F and G) of 23 and 24 dis-

played the mixture mass of ADC carrying diﬀerent numbers of

payloads. The drug–antibody ratio (DAR) was calculated based

The advantage of site-specific conjugation is to avoid the poss-

ible influence on the antigen-binding domain by random lig-

ation. We tested the antigen-binding activity of commercial

Org. Biomol. Chem.

This journal is © The Royal Society of Chemistry 2016

View Article Online

Organic & Biomolecular Chemistry

Paper

Scheme 3 Conditions: (a) DM1-SMCC (20), phosphate buﬀer, pH 7.5; (b) 21a–d, PB, pH 7.5; (c) 21c, PB, pH 7.5.

Fig. 6 Cell-based antigen binding properties of Cy5-labeled Herceptin. (A) SK-BR-3 cell-based ELISA assay of commercial Herceptin (red), site-

speciﬁc conjugate Herceptin-Cy5 (22f) (blue), and random conjugate Herceptin-Cy5 (26) (black); (B) SK-BR-3 cell binding assay with site-speciﬁc

conjugate Herceptin-Cy5 (22f) by using a confocal microscope. The nucleus is labeled in blue and Cy5-labeled Herceptin on the cell surface is dis-

played in red.

Herceptin, random-labeled Herceptin-Cy5 (26) and glycosite- site-specific labeling of IgG enabled by glycoengineering is an

specific Cy5-labeled Herceptin (22f) by cell-based ELISA. optimal strategy for imaging without aﬀecting the antigen

A HER2-positive breast cancer cell line, SK-BR-3, was coated on binding properties.

a plate and these three Herceptin samples were measured for

Cytotoxicity assay against the cancer cell line SK-BR-3 was

the binding aﬃnity against the coated target cell. As shown in carried out with glycosite-specific ADCs. For comparison, we

Fig. 6A, the KDs of commercial Herceptin, 22f, and 26 were also synthesized the random Lys conjugate T-DM1 and

1.2, 3.5, and 24 μg mL⁻¹respectively. Apparently, random con- T-MMAE (see the ESI†). As shown in Fig. 7 and Table 2, T-DM1

jugation caused significant reduction on antigen-binding and T-MMAE indicated potent antitumor activity with the EC₅₀

activity compared with the site-specific conjugation. values of 0.06 and 0.15 μg mL⁻¹, while ADCs 22a, 22b, and 22d

In addition, we examined the target cell attachment of site- did not aﬀect the cell proliferation besides the eﬃcacy of the

specific labeled Herceptin 22f by confocal microscopy antibody itself. However, ADCs 22c and 24 inhibited the

(Fig. 6B). SK-BR-3 cells were incubated with 22f at 4 °C for cancer cell growth as eﬃciently as T-DM1 with the EC₅₀values

30 min and then subjected to confocal imaging. Fluorescence of 0.09 and 0.06 μg mL⁻¹. These data elucidated the impor-

detection based on Cy5 label indicated that the antibody was tance of a linker fragment for glycosite-specific ADCs. Based

crowded on the cell surface. These data demonstrated that the on these results, the cleavable Val-Cit motif is crucial for

This journal is © The Royal Society of Chemistry 2016

Org. Biomol. Chem.

View Article Online

Paper

Organic & Biomolecular Chemistry

(see the ESI†). β-1,4-Galactosidase was purchased from New

England Biolabs (Ipswich, MA). Neuraminidase and β-N-acetyl-

glucosaminidase were purchased from Sigma-Aldrich

(Shanghai, China). Sialylglycopeptide (1) was extracted and

purified from egg yolk powder following the literature⁵⁴(see

the ESI†). 3-Azido-1-propanamine (9a), propargylamine (9b),

and HPLC grade acetonitrile was purchased from J&K

Chemical Ltd (Shanghai, China). Syntheses of hydroxylamine

compounds 10a–d are described in the ESI.† DM1 (15) and

MMAE (16) were purchased from Levena Biopharma (Nanjing,

China).

Dibenzocyclooctyne

(DBCO)-amine

(17)

and

DBCO-PEG₄-acid (18) were purchased from Sigma-Aldrich

(Shanghai, China). Sulfo-cyanine5 NHS ester (19) was

purchased from Little-PA Sciences Co., Ltd (Wuhan, China).

Syntheses of drug-linker intermediates 20, and 21a–d are

described in the ESI.† Porous graphite carbon (PGC) cartridges

(Hypercarb) were purchased from Thermo Fisher Scientific

(Beijing, China). Other chemical reagents and solvents were

purchased from Sinopharm Chemical Reagent Co. (Shanghai,

China) or Sigma-Aldrich (Shanghai, China) and used without

further purification. Nuclear magnetic resonance (NMR)

spectra were measured on a Varian-MERCURY Plus-400 or 500

instrument. ESI-HRMS spectra were measured on an Agilent

6230 LC-TOF MS spectrometer.

Fig. 7 Cytotoxicity assay of ADC samples on a HER2-positive cell line,

SK-BR-3.

site-specific ADCs but not for Lys-conjugated ADCs to release

the toxin after cell endocytosis. The endocytosed T-DM1 was

digested in lysosome to give DM1-MCC-Lys as the cytotoxic

agent. For glycan-conjugated ADCs, this mechanism does not

seem to be eﬀective probably due to the hindrance of large

N-glycan by releasing toxins. Even with a cleavable handle of

oxime, the ADCs 22a and 22b did not show antitumor

cytotoxicity. Whereas, 22c and 24 carrying the Val-Cit fragment

as a protease substrate successfully achieved cancer cell

inhibition. The glycosite-specific ADCs required a diﬀerent

releasing mechanism that could guide us for ADC design in

future. The dual-drug ADC 24 at a lower DM1 DAR (2.6) indi-

cated the same antitumor activity as T-DM1 (DAR 3.2) that may

be due to the partial contribution of MMAE to 24. Since DM1

and MMAE target the cancer cell tublin in a similar mechan-

ism, dual-drug ADCs with two diﬀerent toxic agents such as

combined anti-tublin and DNA-damaging toxins are more

attractive for future research and development.

High performance liquid chromatography (HPLC)

Method A: Analytical RP-HPLC was performed on a Beijing

ChuangXinTongHeng LC3000 (analytic) instrument with a C18

column (5 μm, 4.6 × 150 mm) at 40 °C. The column was eluted

with a linear gradient of 2–90% acetonitrile containing 0.1%

TFA for 30 min at a flow rate of 1 mL min⁻¹. Method B:

Analytical RP-HPLC was performed on

a

Beijing

ChuangXinTongHeng LC3000 (analytic) instrument with a C18

column (5 μm, 4.6 × 150 mm) at 40 °C. The column was eluted

with a linear gradient of 2–11% acetonitrile containing 0.1%

TFA for 10 min and then 11–90% acetonitrile for an additional

20 min at a flow rate of 1 mL min⁻¹. Method C: Analytical

RP-HPLC was performed on a Thermo ultimate 3000 instru-

ment with a column (5 μm, 4.6 × 250 mm) at 40 °C. The

column was eluted with a linear gradient of 2–90% acetonitrile

Conclusion

containing 0.1% TFA for 30 min at a flow rate of 1 mL min⁻¹

.

Method D: Preparative HPLC was performed on a Beijing

ChuangXinTongHeng LC3000 (preparative) instrument with a

preparative column (Waters, C18, OBD, 5 μm, 19 × 250 mm).

The column was eluted with a suitable gradient of aqueous

acetonitrile containing 0.1% TFA at a flow rate of 10 mL

An eﬃcient one-pot chemoenzymatic glycoengineering

approach was established for IgG Fc glycosylation remodeling

directly using an egg-yolk sialylglycopeptide (SGP) or its non-

natural derivatives carrying azido or alkyne tags. This approach

enables the rapid synthesis of functionalized therapeutic IgG

antibodies for glycosite-specific conjugation with small-

molecule drugs or fluorescent agents. Site-specific ADCs and

dual-drug ADCs were developed based on this strategy.

min⁻¹

.

Liquid chromatography mass spectrometry (LC-MS)

ESI-MS spectra were measured on an Agilent 6230 LC–TOF MS

spectrometer. The small molecules were analyzed using a

short guard column and eluted with 70% methanol containing

0.1% formic acid. The mass spectra of small molecules were

recorded in the mass range of 200–3000 or 600–2000 under a

Experimental section

Materials

endo-M, endo-S wild type, endo-S mutant D233Q and PNGase-F high resolution mass-spec mode (HRMS, standard 3200 m/z,

were expressed in E. coli following the reported procedure^16,654 GHz). Key source parameters: a drying nitrogen gas flow of

Org. Biomol. Chem.

This journal is © The Royal Society of Chemistry 2016

View Article Online

Organic & Biomolecular Chemistry

Paper

11 L min⁻¹; a nebulizer pressure of 40 psi; a gas temperature CHO-SGP (8)

of 350 °C; a fragmenter voltage of 175 V; a skimmer voltage of

To a stirred solution of SGP 1 (100 mg) in PB (0.2 M, pH 7.1,

5 mL) in an ice bath, aqueous sodium periodate (30 mM,

5 mL) was added. The reaction mixture was stirred and

protected from light for 15 min and was then subjected to gel-

filtration via a Sephadex G-25 column. The column was eluted

with water and the fractions containing the product were com-

bined and lyophilized to give 8 as a white powder (90.0 mg,

yield 94%). ¹H NMR (500 MHz, deuterium oxide) δ 5.05 (1H, s,

H1c), 4.96 (1H, d, J = 9.7 Hz, H1a′), 4.88 (3H, m, H1c′, H of

CHO –C(H)–(OH)₂), 4.70 (1H, s, H1b), 4.60 (1H, t, J = 6.7 Hz,

Asn Hα), 4.53 (3H, d, J = 7.4 Hz, H1a, H1d, H1d′), 4.39 (1H, d,

J = 3.5 Hz, Thr Hα), 4.37 (1H, d, J = 2.5 Hz, H1e), 4.35–4.27

(3H, m, H1e′, Thr Hβ, Lys Hα), 4.24 (1H, q, J = 7.1 Hz, Ala Hα),

4.17 (1H, s, H2b), 4.12 (1H, d, J = 3.4 Hz, H2c), 4.08 (1H, d, J =

7.5 Hz, Val Hα), 4.03 (1H, d, H2c′), 3.99 (3H, t, J = 6.4 Hz, Lys

Hα, 2H), 2.92 (m, 4H, Lys Hε), 2.82–2.59 (2H, m, Asn Hβ), 2.51

(2H, dd, J = 12.7, 4.2 Hz, H3f_eq, H3f′_eq), 2.09–1.88 (19H, m,

6 × 3 Ac, Val Hβ), 1.82 (6H, m, Lys Hβ, H3f_ax, H3f′_ax), 1.75–1.56

(4H, m, J = 7.5 Hz, Lys Hδ), 1.36 (4H, m, Lys Hγ), 1.30 (3H, d,

J = 7.5 Hz, Ala Hβ), 1.12 (3H, J = 6.5 Hz, Thr Hγ), 0.89 (6H, d,

7.0 Hz, Val Hγ). HSQC ((¹H, 500 MHz)/(¹³C, 126 MHz), deuter-

ium oxide) 5.05/100.10 (H1c/C1c), 4.96/78.91 (H1a′/C1a′),

4.88/89.17 (H1g/C1g), 4.87/97.65 (H1c′/C1c′), 4.70/101.14

(H1b/C1b), 4.59/50.77 (Hα Asn/Cα Asn), 4.53/100.41 (H1a/C1a,

H1d/C1d, H1d′/C1d′), 4.38/58.86 (Hα Thr/Cα Thr), 4.37/104.22

(H1e/C1e, H1e′/C1e′), 4.33/54.21 (Hα Lys/Cα Lys), 4.31/67.76

(Hβ Thr/Cβ Thr), 4.23/50.13 (Hα Ala/Cα Ala), 4.17/71.26

(H2b/C2b), 4.11/77.18 (H2c/C2c), 4.07/60.15 (Hα Val/Cα Val),

4.03/76.82 (H2c′/C2c′), 3.99/53.60 (Hα Lys/Cα Lys), 2.92/39.73

(Hε Lys/Cε Lys), 2.66/37.45 (Hβ Asn/Cβ Asn), 2.51/40.00

(H3f/C3f, H3f′/C3f′), 2.00/31.01 (Hβ Val/Cβ Val), 1.83/31.01

(Hβ Lys/Cβ Lys), 1.69/39.96 (H3f/C3f, H3f′/C3f′), 1.61/26.93

(Hδ Lys/Cδ Lys), 1.37/22.00 (Hγ Lys/Cγ Lys), 1.30/17.37 (Hβ Ala/

Cβ Ala), 1.12/19.63 (Hγ Thr/Cγ Thr), 0.89/19.33 (Hγ Val/Cγ Val).

HRMS, calculated for C₁₀₈H₁₇₇N₁₅O₆₆[M + 3H]³⁺914.3730,

found 914.3734; [M + H₂O + 3H]³⁺920.3765, found 920.3787.

65 V; and a capillary voltage of 4000 V. The antibodies and

ADCs were measured with an Agilent C-18 column (3.5 μm,

50 × 2.1 mm) at 55 °C. The column was eluted with an isocratic

gradient of 5% acetonitrile (Buﬀer B) and 95% water contain-

ing 0.1% formic acid (Buﬀer A) in the first 10 min, a linear

gradient of 5–95% acetonitrile for an additional 10 min and an

isocratic gradient of 95% acetonitrile for another 10 min at a

flow rate of 0.2 mL min⁻¹. The mass spectra of antibodies were

performed under the extended mass range mode (high 20 000

m/z, 1 GHz) and the data were collected in the mass range of

800–5000. Key source parameters: a drying nitrogen gas flow of

11 L min⁻¹; a nebulizer pressure of 60 psi; a gas temperature

of 350 °C; a fragmenter voltage of 400 V; a skimmer voltage of

65 V; and a capillary voltage of 5000 V. The multiple charged

peaks of the antibody were deconvoluted using the Agilent

MassHunter Bioconfirm software (deconvolution for protein,

Agilent technology) with the deconvolution range from

100 kDa to 200 kDa; other parameters were set at default

values for protein deconvolution. The TOF was calibrated over

the range 0–5000 m/z using Agilent ESI calibration mix

solution before analysis. The peak of MS 922 is the internal

standard for calibration.

High-performance anion exchange chromatography with

pulsed amperometric detection (HPAEC-PAD)

The HPAEC-PAD analysis was performed on an ICS 5000⁺

system (Thermo Scientific) with a PA100 anion exchange

column (4 × 250 mm) and a guard column (4 × 50 mm) at

40 °C. The column was eluted with 100 mM sodium hydrate

solution (Sigma) for the first 2 min and a linear of 0–150 mM

sodium acetate (Sigma) in 100 mM sodium hydrate for an

additional 20 min at a flow rate of 1 mL min⁻¹

.

General procedure for one-pot semi-synthesis of N-glycan

oxazoline substrates (2–5)

General Procedure for synthesis of non-natural oxime-SGPs

(11a, 11c, 11d, and 11e)

A solution of sialylglycopeptide 1 (100 mg) in phosphate buﬀer

(PB, 50 mM, pH 6.25, 1 mL) was incubated with endo-M

(0.5 mg) at 30 °C and monitored by RP-HPLC (Method B). The To a solution of 8 (10 mg, 3.65 μmol, 1.0 eq.) in PB (50 mM,

sialoglycan in the reaction mixture was then transformed into pH 7.5, 100 μL) was added 10a–d (14.6 μmol, 4.0 eq.). The

sialoglycan-oxazoline 2 with the addition of 2-chloro-1,3-di- mixture was stirred at room temperature for 2 hours and moni-

methylimidazolidinium chloride (DMC, 15 eq.) and trimethyl- tored by RP-HPLC (Method B). The residue was subjected to

amine (Et₃N, 45 eq.). For exoglycosidase digestion, the reaction semi-preparative HPLC purification to give the oxime SGPs

mixture was incubated with additional neuraminidase and/or 11a, 11c, 11d, and 11e respectively.

β

1,4-galactosidase and/or β-N-acetylglucosaminidase. The

Azido oxime SGP (11a). Yield 92.3%. ¹H NMR (500 MHz,

reaction was monitored by HPAEC until complete deglycosyla- deuterium oxide) δ 7.41 (1.6H, dd, J = 7.2, 3.3 Hz, H1 of

tion, then DMC (15 eq.) and Et₃N (45 eq.) was added. oxime), 6.85 (0.4H, dd, J = 6.7, 2.6 Hz, H1 of oxime), 5.07 (1H,

Oxazolines 3, 4 and 5 were obtained after solid-phase extrac- s, H1c), 4.96 (1H, d, J = 9.6 Hz, H1a′), 4.96 (0.4 H, H6f, H6f′),

tion via a porous graphite carbon (PGC) column and gel- 4.88 (1H, s, H1c′), 4.76 (1H, s, H1b), 4.59 (1H, t, Asn Hα), 4.53

filtration via a Sephadex G-25 column. The NMR spectra of the (3H, d, J = 6.7 Hz, H1a, H1d. H1d′), 4.39 (1H, d, J = 3.5 Hz, Thr

oxazolines synthesized from the one-pot strategy were consist- Hα), 4.37 (1H, d, J = 2.5 Hz, H1e), 4.35 (1H, d, J = 3 Hz, H1e′),

ent with the reported data.^16,54,61NMR and HPAEC-PAD 4.34–4.28 (2H, m, Lys1 Hα, Thr Hβ), 4.24 (1H, q, J = 7 Hz, Ala

spectra are available in the ESI.†

Hα), 4.17 (5H, m, H2b, H2 of oxime), 4.17 (1.6H, m, H6f, H6f′),

This journal is © The Royal Society of Chemistry 2016

Org. Biomol. Chem.

View Article Online

Paper

Organic & Biomolecular Chemistry

4.12 (1H, d, J = 2 Hz, H2c), 4.08 (1H, d, J = 8 Hz, Val Hα), 4.04 (Hγ Thr/Cγ Thr), 0.86/17.86 (Hγ Val/Cγ Val). HRMS, calculated

(1H, d, J = 2 Hz, H2c′), 3.99 (1H, t, J = 6.5 Hz, Lys2 Hα), 2.92 for C₁₁₈H₁₉₅N₂₅O₆₈[M + 3H]³⁺1017.7601, found 1017.7606;

(4H, q, J = 7.5 Hz, Lys Hε), 2.81–2.61 (2H, m, Asn Hβ), [M + 2H]²⁺1526.1363, found 1526.1346.

2.61–2.51 (2H, dd, J = 12.5 Hz, 2.5 Hz, H3f_eq, H3f′_eq), 2.03–1.88

Alkyne oxime SGP (11d). Yield 88%. ¹H NMR (500 MHz, deu-

(19H, m, 6 × 3 Ac, Val Hβ), 1.87–1.66 (6H, m, Lys Hβ, H3f_ax, terium oxide) δ 7.49 (1.7H, dd, J = 7.0, 4.1 Hz, H1g), 6.90 (0.3H,

H3f′_ax), 1.66 (4H, m, J = 7.5 Hz, Lys Hδ), 1.43–1.32 (4H, m, Lys dd, J = 7.0, 3.9 Hz, H1g), 5.07 (1.3H, s, H1c, H6f, H6f′), 4.96

Hγ), 1.3 (3H, d, J = 7.5 Hz, Ala Hβ), 1.13 (3H, J = 6.5 Hz, Thr (1H, d, J = 9.6 Hz, H1a′), 4.88 (1H, s, H1c′), 4.67 (1H, s, H1b),

Hγ), 0.89 (6H, d, 7.0 Hz, Val Hγ). HSQC ((¹H, 500 MHz)/ 4.60 (1H, t, J = 6.5 Hz, 1H, Asn Hα), 4.51 (7H, m, H1a, H1d,

(¹³C, 126 MHz), deuterium oxide) 7.41/149.38 (H1g/C1g), 6.85/ H1d′, H2g), 4.40 (1H, d, Thr Hα), 4.38 (2H, dd, H1e, H1e′),

149.11 (H1g/C1g), 5.06/99.43 (H1c/C1c), 4.96/78.27 (H1a′/C1a′), 4.33–4.30 (2H, m, Lys Hα, Thr Hβ), 4.24 (1H, q, J = 7.2 Hz, Ala

4.95/67.43 (H6f/C6f, H6f′/C6f′), 4.87/96.94 (H1c′/C1c′), 4.69/ Hα), 4.18 (2.7H, m, H1b, H6f, H6f′), 4.12 (1H, d, J = 2.5 Hz,

100.39 (H1b/C1b), 4.59/49.87 (Hα Asn/Cα Asn), 4.52/99.60 H2c), 4.06 (1H, d, J = 7.6 Hz, Val Hα), 4.03 (1H, d, J = 2 Hz,

(H1a/C1a, H1d/C1d, H1d′/C1d′), 4.39/57.76 (Hα Thr/Cα Thr), H2c′), 3.98 (1H, t, J = 6.7 Hz, Lys Hα), 2.92 (q, J = 7.5 Hz, 4H,

4.36/103.72 (H1e/C1e, H1e′/C1e′), 4.33/53.57 (Hα Lys/Cα Lys), Lys Hε), 2.81–2.61 (ddd, J = 62.3, 16.4, 6.7 Hz, 2H, Asn Hβ),

4.32/67.08 (Hβ Thr/Cβ Thr), 4.23/49.46 (Hα Ala/Cα Ala), 4.17/ 2.60–2.51 (3H, m, H3f_eq, H3f′_eq, H6 g), 2.05–1.87 (19H, m, Ac,

72.10 (H2b/C2b, H2g/C2g, H6f/C6f, H6f′/C6f′), 4.12/76.41 Val Hβ), 1.87–1.66 (6H, m, Lys Hβ, H3f_ax, H3f′_ax), 1.62 (4H, m,

(H2c/C2c), 4.07/59.53 (Hα Val/Cα Val), 4.04/76.23 (H2c′/C2c′), J = 7.5 Hz, Lya Hδ), 1.38 (4H, d, J = 5.9 Hz, Lys Hγ), 1.30 (3H, d,

3.99/52.63 (Hα Lys/Cα Lys), 3.46/49.76 (H3 g/C3 g), 2.92/39.12 J = 7.2 Hz, Ala Hβ), 1.12 (3H, d, J = 6.4 Hz, Thr Hγ), 0.89 (6H, d,

(Hε Lys/Cε Lys), 2.64/36.66 (Hβ Asn/Cβ Asn), 2.58/39.18 J = 6.7 Hz, Val Hγ). HSQC ((¹H, 500 MHz)/(¹³C, 126 MHz), deu-

(H3f/C3f, H3f′/C3f′), 1.97/30.60 (Hβ Val/Cβ Val), 1.81/39.05 terium oxide) 7.49/150.71 (H1g/C1g), 6.90/150.28 (H1g/C1g),

(H3f/C3f, H3f′/C3f′), 1.77/30.39 (Hβ Lys/Cβ Lys), 1.62/26.13 (Hδ 5.06/99.53 (H1c/C1c), 5.03/67.24 (H6f/C6f, H6f′/C6f′), 4.96/

Lys/Cδ Lys), 1.37/21.38 (Hγ Lys/Cγ Lys), 1.34/16.54 (Hβ Ala/Cβ 78.25 (H1a′/C1a′), 4.88/96.92 (H1c′/C1c′), 4.70/100.20

Ala), 1.16/18.82 (Hγ Thr/Cγ Thr), 0.85/17.73 (Hγ Val/Cγ Val). (H1b/C1b), 4.60/49.87 (Hα Asn/Cα Asn), 4.53/99.93 (H1a/C1a,

HRMS, calculated for C₁₁₂H₁₈₅N₂₃O₆₆[M + 3H]³⁺970.4020, H1d/C1d, H1d′/C1d′), 4.51/72.14 (H2g/C2g), 4.39/58.01

found 970.3925. [M + 2H]²⁺1455.0991, found 1455.0845.

(Hα Thr/Cα Thr), 4.37/103.63 (H1e/C1e, H1e′/C1e′), 4.33/53.61

Azido amide oxime SGP (11c). Yield 94.2%. ¹H NMR (Hα Lys/Cα Lys), 4.32/67.03 (Hβ Thr/Cβ Thr), 4.24/49.50

(500 MHz, deuterium oxide) δ 7.48 (1.6H, dd, J = 7.0, 3.7 Hz, (Hα Ala/Cα Ala), 4.18/72.67 (H2b/C2b, H6f/C6f, H6f′/C6f′), 4.12/

H1g), 6.89 (0.4H, dd, J = 7.0, 2.5 Hz, H1g), 5.05 (1.4H, s, H1c, 76.11 (H2c/C2c), 4.08/59.45 (Hα Val/Cα Val), 4.04/76.17 (H2c′/

H6f, H6f′), 4.95 (1H, d, J = 9.7 Hz, H1a′), 4.87 (1H, s, H1c′), C2c′), 3.99/52.61 (Hα Lys/Cα Lys), 3.98/28.50 (H4 g/C4 g), 2.92/

4.67 (1H, s, H1b), 4.60 (1H, t, Asn Hα), 4.52 (3H, d, J = 7.1 Hz, 38.96 (Hε Lys/Cε Lys), 2.76/36.56 (Hβ Asn/Cβ Asn), 2.58/39.41

H1a, H1d, H1d′), 4.49–4.41 (4H, m, H2g), 4.38 (1H, d, Thr Hα), (H3f/C3f, H3f′/C3f′), 2.56/72.13 (H6 g/C6 g), 2.03/30.15 (Hβ Val/

4.36 (2H, dd, H1e, H1e′), 4.33–4.29 (2H, m, Lys Hα, Thr Hβ), Cβ Val), 1.86/30.31 (Hβ Lys/Cβ Lys), 1.79/39.19 (H3f/C3f, H3f′/

4.24 (1H, q, J = 7.2 Hz, Ala Hα), 4.18 (2.6H, m, H1b, H6f, H6f′), C3f′), 1.62/26.04 (Hδ Lys/Cδ Lys), 1.37/21.44 (Hγ Lys/Cγ Lys),

4.11 (1H, d, J = 2.5 Hz, H2c), 4.06 (1H, d, J = 7.6 Hz, Val Hα), 1.34/16.69 (Hβ Ala/Cβ Ala), 1.16/18.82 (Hγ Thr/Cγ Thr), 0.92/

4.03 (1H, d, J = 2 Hz, H2c′), 3.98 (1H, t, J = 6.7 Hz, Lys Hα), 17.91 (Hγ Val/Cγ Val). HRMS, calculated for C₁₁₈H₁₈₉N₁₉O₆₈

3.37–3.15 (8H, m, H4 g, H6 g), 2.91 (4H, q, J = 7.6 Hz, Lys Hε), [M + 3H]³⁺987.7383, found 987.7380; [M + 2H]²⁺1481.1036,

2.70 (2H, ddd, J = 62.2, 16.4, 6.7 Hz, Asn Hβ), 2.56 (2H, dd, J = found 1481.1070.

11.0 Hz, H3f_eq, H3f′_eq), 1.80 (4H, m, Lys Hβ)1.72 (6H, m, H5g,

Alkyne amide oxime SGP (11e). Yield 86.6%. ¹H NMR

H3f_ax, H3f′_ax), 1.43–1.31 (4H, m, Lys Hγ), 1.29 (3H, d, Ala Hβ), (500 MHz, deuterium oxide) δ 7.39 (1.6H, dd, J = 7.2, 3.3 Hz,

1.11 (3H, d, J = 6.4 Hz, Thr Hγ), 0.88 (6H, d, J = 6.7 Hz, Val Hγ). H1 of oxime), 6.85 (0.4H, dd, J = 6.7, 2.6 Hz, H1 of oxime), 5.07

HSQC ((¹H, 500 MHz)/(¹³C, 126 MHz), deuterium oxide) 7.49/ (1H, s, H1c), 4.96 (1.4H, d, J = 9.6 Hz, H1a′, H6f/f′), 4.88 (1H, s,

150.77 (H1g/C1g), 6.90/150.47 (H1g/C1g), 5.06/99.43 (H1c/C1c), H1c′), 4.76 (1H, s, H1b), 4.61 (5H, m, H2g, Asn Hα), 4.54 (3H,

5.05/67.26 (H6f/C6f, H6f′/C6f′), 4.96/78.16 (H1a′/C1a′), 4.88/ d, J = 6.7 Hz, H1a, H1d. H1d′), 4.39 (1H, d, J = 3.5 Hz, Thr Hα),

97.00 (H1c′/C1c′), 4.70/100.37 (H1b/C1b), 4.60/49.86 (Hα Asn/ 4.37 (1H, d, J = 2.5 Hz, H1e), 4.35 (1H, d, J = 3 Hz, H1e′),

Cα Asn), 4.53/99.51 (H1a/C1a, H1d/C1d, H1d′/C1d′), 4.48/72.14 4.34–4.28 (2H, m, Lys1 Hα, Thr Hβ), 4.24 (1H, q, J = 7 Hz, Ala

(H2g/C2g), 4.39/57.99 (Hα Thr/Cα Thr), 4.37/103.58 (H1e/C1e, Hα), 4.18 (1H, s, H2b), 4.16 (1.6H, dd, J = 10.0 Hz, 7.5 Hz,

H1e′/C1e′), 4.33/53.62 (Hα Lys/Cα Lys), 4.32/67.14 (Hβ Thr/Cβ H6f/f′), 4.12 (1H, d, J = 2 Hz, H2c), 4.08 (1H, d, J = 8 Hz, Val

Thr), 4.24/49.53 (Hα Ala/Cα Ala), 4.18/72.35 (H2b/C2b, Hα), 4.04 (1H, d, J = 2 Hz, H2c′) 3.99 (1H, t, J = 6.5 Hz, Lys2

H6f/C6f, H6f′/C6f′), 4.12/76.11 (H2c/C2c), 4.08/59.56 (Hα Val/ Hα), 2.92 (4H, q, J = 7.5 Hz, Lys Hε), 2.86 (2H, q, H3 g),

Cα Val), 4.04/76.01 (H2c′/C2c′), 4.00/52.70 (Hα Lys/Cα Lys), 2.81–2.61 (2H, m, Asn Hβ), 2.61–2.51 (2H, dd, J = 12.5 Hz,

3.32/36.54 (H4 g/C4 g), 3.31/48.73 (H6 g/C6 g), 3.23/36.58 2.5 Hz, H3f_eq, H3f′_eq), 2.03–1.88 (19H, m, 6 × 3 Ac, Val Hβ),

(H4 g/C4 g), 2.93/39.11 (Hε Lys/Cε Lys), 2.77/36.47 (Hβ Asn/Cβ 1.87–1.66 (6H, m, Lys Hβ, H3f_ax, H3f′_ax), 1.66 (4H, m, J =

Asn), 2.58/39.32 (H3f/C3f, H3f′/C3f′), 2.04/29.84 (Hβ Val/Cβ 7.5 Hz, Lya Hδ), 1.43–1.32 (4H, m, Lys Hγ), 1.3 (3H, d, J =

Val), 1.87/30.45 (Hβ Lys/Cβ Lys), 1.78/39.39 (H3f/C3f, H3f′/ 7.5 Hz, Ala Hβ), 1.13 (3H, J = 6.5 Hz, Thr Hγ), 0.89 (6H, d,

C3f′), 1.71/27.53 (H5g/C5g), 1.63/26.18 (Hδ Lys/Cδ Lys), 1.37/ 7.0 Hz, Val Hγ). HSQC ((¹H, 500 MHz)/(¹³C, 126 MHz),

21.65 (Hγ Lys/Cγ Lys), 1.34/16.85 (Hβ Ala/Cβ Ala), 1.09/18.70 deuterium oxide) 7.41/150.05 (H1g/C1g), 6.88/149.59

Org. Biomol. Chem.

This journal is © The Royal Society of Chemistry 2016

View Article Online

Organic & Biomolecular Chemistry

Paper

(H1g/C1g), 5.09/99.34 (H1c/C1c), 4.98/78.34 (H1a′/C1a′), 4.97/ (Hγ Val/Cγ Val). HRMS, calculated for C₁₁₄H₁₉₃N₂₃O₆₄

67.43 (H6f/C6f, H6f′/C6f′), 4.91/96.89 (H1c′/C1c′), 4.72/100.33 [M + 4H]⁴⁺728.0717, found 728.0717; [M + 3H]³⁺970.4263,

(H1b/C1b), 4.63/61.46 (H2g/C2g), 4.62/49.85 (Hα Asn/Cα Asn), found 970.4254; [M + 2H]²⁺1455.1356, found 1455.1381.

4.55/99.74 (H1a/C1a, H1d/C1d, H1d′/C1d′), 4.42/58.25 (Hα Thr/

Alkyne amine SGP (11f). (8.5 mg, 3.0 μmol, yield 82.2%).

Cα Thr), 4.39/103.91 (H1e/C1e, H1e′/C1e′), 4.35/53.53 (Hα Lys/ ¹H NMR (500 MHz, deuterium oxide) δ 5.15 (1H, s, H1c), 5.05

Cα Lys), 4.34/67.11 (Hβ Thr/Cβ Thr), 4.26/49.43 (Hα Ala/Cα (1H, d, J = 9.6 Hz, H1a′), 4.95 (1H, s, H1c′), 4.84 (1H, s, H1b),

Ala), 4.20/70.52 (H2b/C2b, H6f/C6f, H6f′/C6f′), 4.14/76.54 4.67 (1H, t, J = 6.7 Hz, Asn Hα), 4.62 (3H, dd, J = 6.4, 5.1 Hz,

(H2c/C2c), 4.10/59.64 (Hα Val/Cα Val), 4.06/76.25 (H2c′/C2c′), H1a, H1d, H1d′), 4.46 (2H, dd, J = 7 Hz, H1e, H1e′), 4.41 (1H, t,

4.02/52.83 (Hα Lys/Cα Lys), 2.94/39.16 (Hε Lys/Cε Lys), 2.89/ Lys1 Hα), 4.31 (1H, q, J = 7.2 Hz, Ala Hα), 4.26 (1H, s, H2b),

76.16 (H4 g/C4 g), 2.78/37.03 (Hβ Asn/Cβ Asn), 2.58/39.45 4.22 (1H, m, Thr Hβ) 4.21 (1H, m, Thr Hα), 4.17–4.10 (3H, m,

(H3f/C3f, H3f′/C3f′), 2.01/30.19 (Hβ Val/Cβ Val), 1.84/30.59 (Hβ H2c, Val Hα, H2c′), 4.01 (4H, H2g), 3.43–3.16 (4H, m, H1g),

Lys/Cβ Lys), 1.79/39.37 (H3f/C3f, H3f′/C3f′), 1.65/26.40 (Hδ Lys/ 3.01 (6H, m, Lys Hε, H4g), 2.93–2.69 (2H, m, Asn Hβ),

Cδ Lys), 1.43/21.61 (Hγ Lys/Cγ Lys), 1.32/16.78 (Hβ Ala/Cβ Ala), 2.69–2.59 (2H, dd, J = 12.5 Hz, 2.5 Hz, H3f_eq, H3f′_eq), 2.15ii–

1.14/19.15 (Hγ Thr/Cγ Thr), 0.91/18.15 (Hγ Val/Cγ Val). HRMS, 1.90 (16iH, m, 5 × 3 Ac, Val Hβ), 1.97–1.81 (7H, m, Lys Hβ, Ac)

calculated for C₁₁₄H₁₈₃N₁₇O₆₆[M + 3H]³⁺949.7240, found 1.80–1.64 (6H, m, Lys Hβ, H3f_ax, H3f′_ax), 1.66 (4H, m, J =

949.7164; [M + 2H]²⁺1424.0821, found 1424.0687.

7.5 Hz, Lys Hδ), 1.43–1.32 (4H, m, Lys Hγ), 1.3 (3H, d, J =

Synthesis of non-natural amine-SGPs (11b and 11f). 10 mg 7.5 Hz, Ala Hβ), 1.13 (3H, J = 6.5 Hz, Thr Hγ), 0.89 (6H, d,

CHO-SGP 8 (3.65 μmol, 1.0 eq.) was dissolved in methane/PB 7.0 Hz, Val Hγ). HSQC ((¹H, 500 MHz)/(¹³C, 126 MHz), deuter-

(50 mM, pH 6.25) 1 : 1 solution. Then, 3-azido-1-propanamine ium oxide) 5.06/100.47 (H1c/C1c), 4.96/79.04 (H1a′/C1a′), 4.86/

9a or propargylamine 9b (30.0 eq.), NaCNBH₃(4.6 mg, 97.82 (H1c′/C1c′), 4.70/101.42 (H1b/C1b), 4.59/50.87 (Hα Asn/

73 μmol, 20.0 eq.) was added to the mixture in an ice bath. Cα Asn), 4.53/100.62 (H1a/C1a, H1d/C1d, H1d′/C1d′), 4.37/

After 3 hours when RP-HPLC (Method B) indicated the 103.51 (H1e/C1e, H1e′/C1e′), 4.32/54.42 (Hα Lys/Cα Lys), 4.31/

completion of the reaction, the mixture was subjected to semi- 67.02 (Hβ Thr/Cβ Thr), 4.22/50.36 (Hα Ala/Cα Ala), 4.17/71.01

preparative HPLC purification (Method D). The fractions con- (H2b/C2b), 4.12/77.38 (H2c/C2c), 4.06/60.90 (Hα Val/Cα Val),

taining the products was combined and lyophilized to obtain 4.04/77.09 (H2c′/C2c′), 3.92/53.74 (Hα Lys/Cα Lys), 3.91/37.48

11b and 11f as a white powder.

(H2g/C2g), 3.28/48.52 (H1g/C1g), 2.90/39.89 (Hε Lys/Cε Lys),

Azido amine SGP (11b). (9.0 mg, yield 84.8%). ¹H NMR 2.79/39.26 (Hβ Asn/Cβ Asn), 2.60/40.92 (H3f/C3f, H3f′/C3f′),

(500 MHz, deuterium oxide) δ 5.03 (1H, s, H1c), 4.93 (1H, d, J = 1.97/30.09 (Hβ Val/Cβ Val), 1.79/31.46 (Hβ Lys/Cβ Lys), 1.69/

9.6 Hz, H1a′), 4.83 (1H, s, H1c′), 4.70 (1H, s, H1b), 4.56 (1H, t, 41.18 (H3f/C3f, H3f′/C3f′), 1.65/27.10 (Hδ Lys/Cδ Lys), 1.35/

J = 6.7 Hz, Asn Hα), 4.50 (3H, dd, J = 6.4, 5.1 Hz, H1a, H1d, 22.40 (Hγ Lys/Cγ Lys), 1.30/17.64 (Hβ Ala/Cβ Ala), 1.09/20.15

H1d′), 4.37 (1H, d, J = 3.5 Hz, Thr Hα), 4.35 (1H, d, J = 2.5 Hz, (Hγ Thr/Cγ Thr), 0.89/19.16 (Hγ Val/Cγ Val). HRMS, calculated

H1e), 4.34 (1H, d, J = 3 Hz, H1e′), 4.30 (2H, m, Lys1 Hα, Thr for C₁₁₄H₁₈₇N₁₇O₆₄[M + 4H]⁴⁺705.5554, found 705.5570;

Hβ), 4.21 (1H, q, J = 7.2 Hz, Ala Hα), 4.14 (1H, s, H2b), 4.09 [M + 3H]³⁺940.4045, found 940.4066; [M + 2H]²⁺1410.1029,

(1H, d, J = 2.7 Hz, H2c), 4.05 (1H, d, J = 7.6 Hz, Val Hα), 4.01 found 1410.1140.

(1H, d, J = 3.5 Hz, H2c′), 4.00–3.90 (3H, m, Lys Hα, H6f, H6f′),

Synthesis of azido oxime sialoglycan 12. Azido oxime SGP

3.24–3.04 (8H, m, −2 × –CH₂–NH–CH₂–), 2.92 (4H, q, J = 11a (8.7 mg, 3.0 μmol) was incubated with endo-M (43.5 μg) in

7.5 Hz, Lys Hε), 2.81–2.61 (2H, m, Asn Hβ), 2.61–2.51 (2H, dd, PB (50 mM pH 6.5, 87 μL) at 30 °C and monitored by HPLC.

J = 12.5 Hz, 2.5 Hz, H3f_eq, H3f′_eq), 2.03–1.90 (19H, m, 6 × 3 Ac, The residue was subjected to semi-preparative HPLC purifi-

Val Hβ), 1.90–1.84 (4H, m, H3 g), 1.84–1.65 (6H, m, Lys Hβ, cation to give the pure azido oxime sialo glycan 12 as a white

1

H3f_ax, H3f′_ax), 1.59 (4H, m, J = 7.5 Hz, Lya Hδ), 1.40–1.29 (4H, powder (5.5 mg, yield 88.87%). H NMR (500 MHz, deuterium

m, Lys Hγ), 1.27 (3H, d, J = 7.5 Hz, Ala Hβ), 1.10 (3H, J = oxide) δ 7.39 (dd, J = 7.2, 4.8 Hz, 1.3H, –NvCH–), 6.83 (dd, J =

6.5 Hz, Thr Hγ), 0.86 (6H, d, 7.0 Hz, Val Hγ). HSQC ((¹H, 6.6, 4.0 Hz, 0.7H, –NvCH–), 5.11 (d, J = 3.1 Hz, 0.8H, H1a(α)),

500 MHz)/(¹³C, 126 MHz), deuterium oxide) 5.05/99.53 5.05 (s, 1H, H1c), 4.94 (dd, J = 9.0, 6.6 Hz, 0.7H, H6f/f′), 4.86 (s,

(H1c/C1c), 4.96/78.22 (H1a′/C1a′), 4.85/96.87 (H1c′/C1c′), 4.69/ 1H, H1c′), 4.50 (d, J = 7.3 Hz, 2H, H1d/d′), 4.33 (d, J = 7.9 Hz,

100.38 (H1b/C1b), 4.59/49.84 (Hα Asn/Cα Asn), 4.53/99.58 2H, H1e/e′), 4.22–4.12 (m, 6.3H, H2b, –CH2–O–NvC, H6f/f′),

(H1a/C1a, H1d/C1d, H1d′/C1d′), 4.38/58.04 (Hα Thr/Cα Thr), 4.09 (s, 1H, H2c), 4.02 (s, 1H, H2c′), 3.92–3.76 (m, 14H, H2a,

4.37/103.51 (H1e/C1e, H1e′/C1e′), 4.32/53.51 (Hα Lys/Cα Lys), H6′b, H3c, H3c′, H6′c, H6′c′, H6d, H6d′, H6′d, H6′d′, H4e,

4.31/67.02 (Hβ Thr/Cβ Thr), 4.23/49.46 (Hα Ala/Cα Ala), 4.17/ H4e′, H6′e, H6′e′), 3.68 (m, 20H), 3.58–3.38 (m, 21H), 1.99–1.87

70.16 (H2b/C2b), 4.11/76.42 (H2c/C2c), 4.07/59.39 (Hα Val/Cα (m, 15H), 2.55 (d, J = 11.8 Hz, 2H, H3f_eq/f′_eq), 2.00–1.85 (m,

Val), 4.03/76.11 (H2c′/C2c′), 3.99/52.73 (Hα Lys/Cα Lys), 3.43/ 15H, Ac), 1.77 (t,

J = 11.4 Hz, 2H, H3f_ax/f′_ax). HSQC

48.19 (H4g/C4g), 3.21/48.59 (H2g/C2g), 3.16/46.14 (H1g/C1g), ((¹H, 500 MHz)/(¹³C, 126 MHz), deuterium oxide) 7.39/149.30

2.91/39.01 (Hε Lys/Cε Lys), 2.76/36.86 (Hβ Asn/Cβ Asn), 2.59/ (H1g/C1g), 6.83/149.30 (H1g/C1g), 5.11/90.40 (H1aα/C1aα),

39.43 (H3f/C3f, H3f′/C3f′), 1.98/30.12 (Hβ Val/Cβ Val), 1.80/ 5.05/99.56 (H1c/C1c), 4.95/67.23 (H6f/C6f, H6f′/C6f′), 4.86/

30.30 (Hβ Lys/Cβ Lys), 1.75/39.63 (H3f/C3f, H3f′/C3f′), 1.59/ 96.91 (H1c′/C1c′), 4.68/100.36 (H1b/C1b), 4.62/94.73 (H1aβ/

26.28 (Hδ Lys/Cδ Lys), 1.36/21.42 (Hγ Lys/Cγ Lys), 1.29/16.67 C1aβ), 4.50/99.20 (H1d/C1d, H1d′/C1d′), 4.34/103.64 (H1e/C1e,

(Hβ Ala/Cβ Ala), 1.12/19.03 (Hγ Thr/Cγ Thr), 0.88/18.21 H1e′/C1e′), 4.16/72.65 (H2b/C2b, H2g/C2g, H6f/C6f, H6f′/C6f′),

This journal is © The Royal Society of Chemistry 2016

Org. Biomol. Chem.

View Article Online

Paper

Organic & Biomolecular Chemistry

4.10/76.23 (H2c/C2c), 4.03/76.06 (H2c′/C2c′), 3.44/49.89 a solution of 6 (1 mg) in Tris buﬀer (50 mM, pH 7.4, 200 μL)

(H3 g/C3 g), 2.54/39.18 (H3f/C3f, H3f′/C3f′), 1.77/38.84 were added the purified oxazolines (2–5, 13) (1.5 mM, 30 eq.)

(H3f/C3f, H3f′/C3f′). HRMS, calculated for C₇₆H₁₂₁N₁₃O₅₃and endo-S D233Q (30 μg). The reaction mixture was incubated

[M + 2H]²⁺1032.8664, found [M + 2H]²⁺1032.8665, [M + 3H]³⁺at 30 °C for 1–2 hours. SDS-PAGE and LC-MS monitoring indi-

688.9136, found 688.9182.

cated the complete reaction. The reaction mixture was immedi-

Synthesis of oxime sialoglycan oxazoline 13. 5.0 mg ately subjected to aﬃnity chromatography via protein A resin

(2.42 μmol, 1.0 eq) 12, 34.0 μL ddH₂O, 11.0 mg (108.9 μmol, (100 μL) following the above procedure to give the

15.0 μL, 45.0 eq.) Et₃N and 6.2 mg (36.3 μmol, 15.0 eq.) DMC corresponding glycoengineered Herceptin 7a–d and 14a.

was added to a tube successively in an ice bath for 1 hour after LC-MS deconvolution data: calculated for S2G2F-Herceptin

which the proton NMR indicated the complete transformation 7a 149809.45, found 149810.52. Calculated for G2F-Herceptin

of the starting material to the corresponding oxazoline with an 7b 148645.06, found 148656.89. Calculated for G0F-Herceptin

azido functional group. The solution was desalted by using a 7c

147996.85,

found

147997.27.

Calculated

for

Sephadex G-25 column, which was eluted with water contain- Man3GlcNAc2F-Herceptin 7d 147184.54, found 147185.75.

ing 0.2% triethylamine. The product was combined and lyo- Calculated for azido-oxime-S2G2F-Herceptin 14a 149897.47,

1

philized to give 13 as a white powder (4.5 mg, 91%). H NMR found 149890.43 (see the ESI† for more details).

(500 MHz, deuterium oxide) δ 7.37 (d, J = 7.2 Hz, 1.6H, H1g),

One-pot chemoenzymatic synthesis of glycoengineered

6.80 (d, J = 6.6 Hz, 0.4H, H1g), 6.01 (d, J = 7.2 Hz, 1H, H1 of Herceptin with 6 and in situ oxazolines. SGP (1) or SGP non-

oxazoline), 5.11–4.96 (m, 1.4H, H1c, H6f, H6f′), 4.89 (s, 1H, natural derivatives (11a–e, 10 mg) were incubated with endo-M

H1c′), 4.66 (s, 1H, H1b), 4.53 (dd, J = 5.7 Hz, 2H, H1d, H1d′), (50 μg) in PB (50 mM, pH 6.25, 50 μL) at 30 °C. After

4.35 (d, J = 6.6 Hz, 2H, H1e, H1e′), 4.30 (s, 1H, H3a), 4.26–4.18 3–4 hours, RP-HPLC (Method B) monitoring indicated the

(dd, J = 9.7 Hz, 7.2 Hz, 1.6H, H6f, H6f′), 4.15 (t, J = 4.6 Hz, 4H, complete hydrolysis of GlcNAc-peptide from the glycopeptide.

H2g), 4.11 (m, 1H, H2a), 4.09 (m, 1H, H2b), 2.55 (d, J = 13.2 Then the mixture was diluted to 50 mM with water, and DMC

Hz, 2H, H3f_eq, H3f′_eq), 2.13–1.83 (m, 15H, Ac), 1.67 (t, 2H, (15 eq.) and Et₃N (40 eq.) was added. The solution was put on

H3f_ax, H3f′_ax). HSQC ((¹H, 500 MHz)/(¹³C, 126 MHz), deuter- an ice bath for 1–2 hours. Then, the reaction mixture was

ium oxide) 7.39/150.81 (H1g/C1g), 6.82/150.84 (H1g/C1g), 6.02/ further diluted into 2 mL Tris buﬀer (50 mM, pH 7.4) contain-

100.66 (H1a/C1a oxa), 5.07/100.38 (H1c/C1c), 4.90/97.41 (H1c′/ ing 6 (10 mg) and endo-S D233Q (300 μg). The residue was

C1c′), 4.67/102.13 (H1b/C1b), 4.54/100.31 (H1d/C1d, H1d′/C1d′), incubated at 30 °C for 1–2 hours, monitored by SDS-PAGE and

4.36/104.59 (H1e/C1e, H1e′/C1e′), 4.32/70.15 (H3a/C3a, oxa), 4.23/ LC-MS. After aﬃnity chromatography via protein A resin,

73.66 (H6f/C6f, H6f′/C6f′), 4.17/73.32 (H2g/C2g), 4.13/66.06 glycoengineered Herceptin 7a and 14a–e were obtained.

(H2a/C2a, oxa), 4.12/77.28 (H2c/C2c), 4.09/77.06 (H2c′/C2c′), 4.08/ LC-MS deconvolution data: calculated for azido-oxime-

71.19 (H2b/C2b), 3.47/50.52 (H3 g/C3 g), 2.57/41.16 (H3f/C3f, S2G2F-Herceptin 14a 149897.47, found 149909.91. Calculated

H3f′/C3f′), 1.68/41.06 (H3f/C3f, H3f′/C3f′). ESI-HRMS, calculated for azido-amine-S2G2F-Herceptin 14b 149897.62, found

for C₇₆H₁₁₉N₁₃O₅₂[M + 2H]²⁺1023.8612, found [M + 2H]²⁺149906.31. Calculated for azido-amide-oxime-S2G2F-Herceptin

1023.8557, [M + H + Na]²⁺1034.8521, found 1034.8508, [M + 14c 150181.62, found 150191.82. Calculated for alkyne-amide-

2Na]²⁺1045.8431, found 1045.8471.

oxime-S2G2F-Herceptin 14d 150001.49, found 150002.00.

Synthesis of Herceptin-Fucα1,6GlcNAc (6). A solution of Calculated for alkyne-oxime-S2G2F-Herceptin 14e 149773.41,

commercial Herceptin (38 mg) in PB buﬀer (pH 7.5) was incu- found 149772.38. (see the ESI† for more details).

bated with wild-type endo-S (150 μg) at 37 °C for 2 hours. The

Synthesis of glycosite-specific ADCs 22a–e. The azido

SDS-PAGE analysis showed the complete deglycosylation of Herceptin 14a or 14b (10 mg) was incubated with 21a–c

Herceptin. The antibody was loaded to protein A-agarose resin (130 μM, 20 eq.) in PB (50 mM, pH 7.5, 10 mL) containing 10%

(3 mL), which was pre-washed with glycine–HCl (100 mM, pH DMSO at 30 °C for 4 h and monitored by SDS-PAGE and

2.5, 25 mL) and pre-equilibrated with PB (50 mM, pH 8.0, LC-MS. After the conjugation, the reaction mixture was

50 mL). The column was washed with PB (50 mM, pH 8.0) subjected to aﬃnity chromatography via protein A resin follow-

50 mL and glycine–HCl (20 mM, pH 5.0, 30 mL) successively ing the above procedure. Fractions containing the products

and the bound antibody was eluted with glycine–HCl were combined to give 22a–e. LC-MS deconvolution data: 22a,

(100 mM, pH 2.5) followed by neutralization to ∼pH 7.5 with 154847.29; 22b, 155684.74; 22c, 156659.32; 22d, 155688.26;

glycine–HCl (1 M, pH 8.8) immediately. The fractions contain- 22e, 156643.96.

ing the deglycosylated Herceptin 6 was combined and concen-

Synthesis of ADC 23. A solution of azido tagged Herceptin

trated by centrifugal filtration through a 10 kDa cut-oﬀ 14a (1.0 mg mL⁻¹5 mg, 1.0 eq.) and DM1-SMCC (20) (100 μM,

membrane. And the concentration of the product was 15.0 eq.) in a PB (50 mM, pH 7.5) containing 10% DMSO was

measured by using a Bicinchoninic Acid (BCA) Kit for protein incubated at 37 °C for 1 hour. LC-MS indicated the DAR of 2.6.

determination following the manufacture’s protocol. LC-MS The mixture was immediately subjected to aﬃnity chromato-

deconvolution data: calculated for Herceptin-Fucα1,6GlcNAc graphy via a protein A resin column to give ADC 23. LC-MS

(6), 145806.06, found, 145805.35.

deconvolution data: 149907.63 (+0 DM1), 147719.04 (+1 DM1),

endo-S D233Q-catalyzed chemoenzymatic synthesis of 151823.00 (+2 DM1), 152786.70 (+3 DM1), 153734.55 (+4 DM1),

glycoengineered Herceptins with 6 and purified oxazolines. To 154696.10 (+5 DM1).

Org. Biomol. Chem.

View Article Online

Organic & Biomolecular Chemistry

Paper

Synthesis of dual-drug ADC 24.

A

mixture of 23 twice, HRP-labeled goat anti-rabbit IgG (H + L) Ab (Yeason)

(0.5 mg mL⁻¹, 2 mg, 1.0 eq.) and DBCO-PEG₄-VC-PAB-MMAE was added to the wells, and left at 4 °C for 30 min. After ade-

(21c, 66.67 μM, 20.0 eq.) was incubated at 30 °C. The quate washing, 50 mL tetramethylbenzidine (TMB) reagent

SDS-PAGE and ESI-LCMS indicated complete conjugation. The was added as the substrate for HRP, and the reaction was

solution was subjected to protein A resin purification to give stopped with 2N H₂SO₄. Finally, absorbance was measured at

the dual-drug ADC 24. LCMS deconvolution data: 157606.71 450 nm. The binding curves were analyzed by GraphPad

(+1 DM1), 158566.82 (+2 DM1), 159530.22 (+3 DM1), 160477.23 software.

(+4 DM1), 161446.46 (+5 DM1).

Fluorescence cell imaging. SK-BR-3 cells were cultured

Synthesis of glycosite-specific Cy5-labeled Herceptin 22f. directly on slides with a Chamber Slide System (Nunc) for 16 h

The azido tagged Herceptin 14a (0.5 mg mL⁻¹, 1 mg) was incu- at 37 °C. Cells were washed twice with PBS and blocked using

bated with DBCO-Cy5 (21d, 61 μg, 10.0 eq.) in PB (50 mM, pH 1% BSA/PBS at room temperature for 30 min. 20 μg mL⁻¹of

7.5) at 30 °C and monitored by SDS-PAGE or LCMS. The reac- Cy5-Herceptin 22f was added to the wells and incubated for

tion buﬀer was subjected to protein A column purification to 30 min at 4 °C. After washing with 0.05% Tween 20/PBS (pH

give the Cy5 labeled Herceptin 22f. LC-MS deconvolution data: 7.4), the cells were fixed with 4% paraformaldehyde/4%

153572.33.

sucrose/PBS solution at room temperature for 20 min. Hoechst

(Sigma) staining was used for localization of nuclei at room

Synthesis of Cy5-labeled ADC 25. ADC 23 (0.5 mg mL⁻¹

,

1 mg) was incubated with DBCO-Cy5 21d (61 μg, 10.0 eq.) at temperature for 5 min. Extensive washing was done prior to

30 °C and monitored by SDS-PAGE or LCMS. The reaction confocal microscopy (Olympus) detection.

buﬀer was subjected to protein A column purification to give

Cell cytotoxicity. The MTT (3-(4,5-dimethyl-2-thiazolyl)-2,5-

the Cy5 labeled ADC 25. LC-MS deconvolution data: 153561.22 diphenyl-2-H-tetrazolium bromide) method was used to

(+0 DM1), 154520.53 (+1 DM1), 155476.70 (+2 DM1), 156431.15 measure the in vitro eﬃcacy of the ADCs. SK-Br-3 cells

(+3 DM1), 157393.21 (+4 DM1), 158349.16 (+5 DM1), 159310.31 were cultured in RPMI 1640 GlutaMAX (Invitrogen) sup-

(+6 DM1).

plemented with 10% fetal bovine serum (FBS) and were

PNGase-F digestion analysis of 14a and ADC 22c. The site- planted into 96-well plates with 8000 cells per well. These

specific transglycosylation and conjugation of 14a and 22c plates were incubated overnight at 37 °C and 5% CO₂. The

were analyzed after PNGase-F deglycosylation. Azido-tagged small molecules, DM1 and MMAE, were diluted 3 fold with the

Herceptin 14a (5.0 mg mL⁻¹

, 100 μg) was incubated culture solution from an initial concentration of 100 nM to the

with PNGaseF (0.1 mg mL⁻¹) overnight at 37 °C to give final concentration of 0.41 nM. And ten-fold dilution

Herceptin–Asp²⁹⁷and released glycan 28. SDS-PAGE indicated was applied to the ADC samples from 10 μg mL⁻¹to

that an unnatural glycan was released from the antibody. 0.0001 μg mL⁻¹. The samples were added to three wells with

The reaction solution was subjected to LC-MS analysis. every single concentration and the cells were cultured at 37 °C

Herceptin–Asp²⁹⁷

,

LC-MS deconvolution data: calculated and 5% CO₂for three days before the addition of MTT solu-

145109.75, found 145107.65. Released glycan 28, HRMS: tion. 10% SDS solution was added to the cell culture solution

calculated for C₉₀H₁₄₄N₁₄O₆₂[M + 3H]³⁺805.2927, found which had been incubated with MTT for 4 h, to dissolve the

805.2936; [M + 2H]²⁺1207.4351, found 1207.4365. With a formazan. Optical density (OD) was measured at 570 nm using

similar procedure, ADC 22c was incubated with PNGase-F over- an Epoch (BioTek) after incubation overnight at 37 °C and 5%

night at 37 °C to give deglycosylated Herceptin–Asp²⁹⁷and CO₂. Finally, the EC₅₀values and the cell viability curve were

released glycan–drug complex 29. SDS-PAGE and LCMS indi- calculated by GraphPad software.

cated that the unnatural glycan attached with MMAE was

released from the antibody. The released glycan–drug complex

29, HRMS: calculated for

827.1222, found 827.1241; [M

C

₂₇₀H₄₀₈N₃₈O₁₀₀[M

+

7H]⁷⁺

6H]⁶⁺964.8080, found

964.9725. [M + 5H]⁵⁺1157.5680, found 1157.5654; [M + 4H]⁴⁺

1446.7080, found 1446.6993. [M + 3H]³⁺1928.6081, found

1928.6228.

Acknowledgements

This work was supported by the National Natural Science

SK-BR-3 cell-based ELISA assay. SK-BR-3 cells were cultured Foundation of China (NNSFC, No. 21372238 and 21572244),

in a 96-well plate (1 × 10⁵cells per well) for 16 h at 37 °C. Cells the SIMM Institute Fund (CASIMM0120153004), and the

were washed twice with PBS, and fixed with 4% paraformalde- “Personalized Medicines—Molecular Signature-based Drug

hyde/PBS buﬀer (pH 7.4). Commercial Herceptin, random Discovery and Development”, Strategic Priority Research

labeled Cy5-Herceptin (26) and glycosite-specific labeled Cy5- Program of the Chinese Academy of Sciences, Grant No.

Herceptin (22f) in 1% BSA/PBS were added to the wells and XDA12020311. We thank the mass spectrometry facility of the

incubated for 30 min at 4 °C. Unbound proteins were removed iHuman Institute for providing us the LC-MS and HPAEC

and washed twice with 0.05% Tween 20/PBS (pH 7.4). instruments. We thank Dr Raymond Stevens, Dr Zhijie Liu, Dr

Subsequently, the cells were blocked with 1% BSA/PBS at room Houchao Tao, and other colleagues in iHuman for helpful

temperature for 30 min, and then incubated with rabbit anti- discussion. W. H. thanks his former supervisor Dr Lai-Xi Wang

human IgG1 Fc mAb (Sigma) for 60 min at 4 °C. After washing at the University of Maryland for his kind mentoring.

Org. Biomol. Chem.

View Article Online

Paper

Organic & Biomolecular Chemistry

14 R. M. Anthony, F. Wermeling and J. V. Ravetch, Novel roles

for the IgG Fc glycan, Ann. N. Y. Acad. Sci., 2012, 1253,

170–180.

References

1 S. Aggarwal, What’s fueling the biotech engine-2010 to

2011, Nat. Biotechnol., 2011, 29(12), 1083–1089.

15 R. L. Shields, J. Lai, R. Keck, L. Y. O’Connell, K. Hong,

Y. G. Meng, S. H. Weikert and L. G. Presta, Lack of fucose

on human IgG1 N-linked oligosaccharide improves binding

to human Fcgamma RIII and antibody-dependent cellular

toxicity, J. Biol. Chem., 2002, 277(30), 26733–26740.

16 W. Huang, J. Giddens, S. Q. Fan, C. Toonstra and

L. X. Wang, Chemoenzymatic glycoengineering of intact

IgG antibodies for gain of functions, J. Am. Chem. Soc.,

2012, 134(29), 12308–12318.

2 G. J. Weiner, Building better monoclonal antibody-based

therapeutics, Nat. Rev. Cancer, 2015, 15(6), 361–370.

3 S. Salemi, M. Markovic, G. Martini and R. D’Amelio, The

expanding role of therapeutic antibodies, Int. Rev.

Immunol., 2015, 34(3), 202–264.

4 A. Zumla, M. Rao, R. S. Wallis, S. H. Kaufmann,

R. Rustomjee, P. Mwaba, C. Vilaplana, D. Yeboah-Manu,

J. Chakaya, G. Ippolito, E. Azhar, M. Hoelscher and

M. Maeurer, Host-directed therapies for infectious dis- 17 M. J. Gramer, J. J. Eckblad, R. Donahue, J. Brown,

eases: current status, recent progress, and future prospects,

Lancet Infect. Dis., 2016, 16(4), e47–e63.

5 P. Sondermann and D. E. Szymkowski, Harnessing Fc

receptor biology in the design of therapeutic antibodies,

Curr. Opin. Immunol., 2016, 40, 78–87.

C. Shultz, K. Vickerman, P. Priem, E. T. Van den Bremer,

J. Gerritsen and P. H. Van Berkel, Modulation of antibody

galactosylation through feeding of uridine, manganese

chloride, and galactose, Biotechnol. Bioeng., 2011, 108(7),

1591–1602.

6 D. J. DiLillo and J. V. Ravetch, Fc-Receptor Interactions 18 R. M. Anthony, F. Nimmerjahn, D. J. Ashline,

Regulate Both Cytotoxic and Immunomodulatory

Therapeutic Antibody Eﬀector Functions, Cancer Immunol.

Res., 2015, 3(7), 704–713.

V. N. Reinhold, J. C. Paulson and J. V. Ravetch,

Recapitulation of IVIG anti-inflammatory activity with a

recombinant IgG Fc, Science, 2008, 320(5874), 373–376.

7 A. M. Brandsma, S. R. Jacobino, S. Meyer, T. Ten Broeke 19 R. M. Anthony, F. Wermeling, M. C. Karlsson and

and J. H. Leusen, Fc receptor inside-out signaling and poss-

ible impact on antibody therapy, Immunol. Rev., 2015,

268(1), 74–87.

J. V. Ravetch, Identification of a receptor required for the

anti-inflammatory activity of IVIG, Proc. Natl. Acad.

Sci. U. S. A., 2008, 105(50), 19571–19578.

8 L. T. Vogelpoel, D. L. Baeten, E. C. De Jong and J. Den 20 Y. Kaneko, F. Nimmerjahn and J. V. Ravetch, Anti-inflam-

Dunnen, Control of cytokine production by human fc

gamma receptors: implications for pathogen defense and

autoimmunity, Front. Immunol., 2015, 6, 79.

9 A. Pincetic, S. Bournazos, D. J. DiLillo, J. Maamary,

T. T. Wang, R. Dahan, B. M. Fiebiger and J. V. Ravetch,

matory activity of immunoglobulin G resulting from Fc

sialylation, Science, 2006, 313(5787), 670–673.

21 A. Beck and J. M. Reichert, Marketing approval of mogamu-

lizumab: a triumph for glyco-engineering, mAbs, 2012, 4(4),

419–425.

Type I and type II Fc receptors regulate innate and adaptive 22 A. M. Goetze, Y. D. Liu, Z. Zhang, B. Shah, E. Lee,

immunity, Nat. Immunol., 2014, 15(8), 707–716.

10 T. S. Raju, Terminal sugars of Fc glycans influence anti-

body eﬀector functions of IgGs, Curr. Opin. Immunol., 2008,

20(4), 471–478.

11 R. M. Anthony and F. Nimmerjahn, The role of diﬀerential

IgG glycosylation in the interaction of antibodies with

FcgammaRs in vivo, Curr. Opin. Organ Transplant., 2011,

16(1), 7–14.

P. V. Bondarenko and G. C. Flynn, High-mannose glycans on

the Fc region of therapeutic IgG antibodies increase serum

clearance in humans, Glycobiology, 2011, 21(7), 949–959.

23 L. Alessandri, D. Ouellette, A. Acquah, M. Rieser,

D. Leblond, M. Saltarelli, C. Radziejewski, T. Fujimori and

I. Correia, Increased serum clearance of oligomannose

species present on a human IgG1 molecule, mAbs, 2012,

4(4), 509–520.

12 C. Ferrara, S. Grau, C. Jager, P. Sondermann, P. Brunker, 24 L. Liu, Antibody glycosylation and its impact on the phar-

I. Waldhauer, M. Hennig, A. Ruf, A. C. Rufer,

M. Stihle, P. Umana and J. Benz, Unique carbohydrate-

carbohydrate interactions are required for high aﬃnity

macokinetics and pharmacodynamics of monoclonal anti-

bodies and Fc-fusion proteins, J. Pharm. Sci., 2015, 104(6),

1866–1884.

binding between FcgammaRIII and antibodies lacking core 25 A. Beck, O. Cochet and T. Wurch, GlycoFi’s technology to

fucose, Proc. Natl. Acad. Sci. U. S. A., 2011, 108(31), 12669–

control the glycosylation of recombinant therapeutic pro-

12674.

teins, Expert Opin. Drug Discovery, 2010, 5(1), 95–111.

13 T. Shinkawa, K. Nakamura, N. Yamane, E. Shoji-Hosaka, 26 H. Li, N. Sethuraman, T. A. Stadheim, D. Zha, B. Prinz,

Y. Kanda, M. Sakurada, K. Uchida, H. Anazawa, M. Satoh,

M. Yamasaki, N. Hanai and K. Shitara, The absence of

fucose but not the presence of galactose or bisecting

N-acetylglucosamine of human IgG1 complex-type oligosac-

charides shows the critical role of enhancing antibody-

dependent cellular cytotoxicity, J. Biol. Chem., 2003, 278(5),

3466–3473.

N. Ballew, P. Bobrowicz, B. K. Choi, W. J. Cook, M. Cukan,

N. R. Houston-Cummings, R. Davidson, B. Gong,

S. R. Hamilton, J. P. Hoopes, Y. Jiang, N. Kim,

R. Mansfield, J. H. Nett, S. Rios, R. Strawbridge, S. Wildt

and T. U. Gerngross, Optimization of humanized IgGs in

glycoengineered Pichia pastoris, Nat. Biotechnol., 2006,

24(2), 210–215.

Org. Biomol. Chem.

View Article Online

Organic & Biomolecular Chemistry

Paper

27 M. Cerutti and J. Golay, Lepidopteran cells, an alternative 40 Q. Zhou and J. Kim, Advances in the Development of Site-

for the production of recombinant antibodies?, mAbs,

2012, 4(3), 294–309.

Specific Antibody-Drug Conjugation, Adv. Anticancer Agents

Med. Chem., 2015, 15(7), 828–836.

28 N. Yamane-Ohnuki and M. Satoh, Production of thera- 41 P. Akkapeddi, S. A. Azizi, A. M. Freedy, P. M. S. D. Cal,

peutic antibodies with controlled fucosylation, mAbs, 2009,

1(3), 230–236.

29 N. Yamane-Ohnuki, S. Kinoshita, M. Inoue-Urakubo,

P. M. P. Gois and G. J. L. Bernardes, Construction of homo-

geneous antibody-drug conjugates using site-selective

protein chemistry, Chem. Sci., 2016, 7(5), 2954–2963.

M. Kusunoki, S. Iida, R. Nakano, M. Wakitani, R. Niwa, 42 J. Y. Axup, K. M. Bajjuri, M. Ritland, B. M. Hutchins,

M. Sakurada, K. Uchida, K. Shitara and M. Satoh,

Establishment of FUT8 knockout Chinese hamster ovary

cells: an ideal host cell line for producing completely defu-

cosylated antibodies with enhanced antibody-dependent

cellular cytotoxicity, Biotechnol. Bioeng., 2004, 87(5), 614–

622.

C. H. Kim, S. A. Kazane, R. Halder, J. S. Forsyth,

A. F. Santidrian, K. Stafin, Y. Lu, H. Tran, A. J. Seller,

S. L. Biroc, A. Szydlik, J. K. Pinkstaﬀ, F. Tian, S. C. Sinha,

B. Felding-Habermann, V. V. Smider and P. G. Schultz,

Synthesis of site-specific antibody-drug conjugates using

unnatural amino acids, Proc. Natl. Acad. Sci. U. S. A., 2012,

109(40), 16101–16106.

30 P. Umana, J. Jean-Mairet, R. Moudry, H. Amstutz and

J. E. Bailey, Engineered glycoforms of an antineuroblas- 43 F. Tian, Y. Lu, A. Manibusan, A. Sellers, H. Tran, Y. Sun,

toma IgG1 with optimized antibody-dependent cellular

cytotoxic activity, Nat. Biotechnol., 1999, 17(2), 176–180.

31 D. Warnock, X. Bai, K. Autote, J. Gonzales, K. Kinealy,

B. Yan, J. Qian, T. Stevenson, D. Zopf and R. J. Bayer, In

vitro galactosylation of human IgG at 1 kg scale using

recombinant galactosyltransferase, Biotechnol. Bioeng.,

2005, 92(7), 831–842.

T. Phuong, R. Barnett, B. Hehli, F. Song, M. J. DeGuzman,

S. Ensari, J. K. Pinkstaﬀ, L. M. Sullivan, S. L. Biroc, H. Cho,

P. G. Schultz, J. DiJoseph, M. Dougher, D. Ma, R. Dushin,

M. Leal, L. Tchistiakova, E. Feyfant, H. P. Gerber and

P. Sapra, A general approach to site-specific antibody drug

conjugates, Proc. Natl. Acad. Sci. U. S. A., 2014, 111(5),

1766–1771.

32 T. S. Raju, J. B. Briggs, S. M. Chamow, M. E. Winkler and 44 E. S. Zimmerman, T. H. Heibeck, A. Gill, X. Li, C. J. Murray,

A. J. Jones, Glycoengineering of therapeutic glycoproteins:

in vitro galactosylation and sialylation of glycoproteins

with terminal N-acetylglucosamine and galactose residues,

Biochemistry, 2001, 40(30), 8868–8876.

M. R. Madlansacay, C. Tran, N. T. Uter, G. Yin, P. J. Rivers,

A. Y. Yam, W. D. Wang, A. R. Steiner, S. U. Bajad, K. Penta,

W. Yang, T. J. Hallam, C. D. Thanos and A. K. Sato,

Production of site-specific antibody-drug conjugates using

optimized non-natural amino acids in a cell-free expression

system, Bioconjugate Chem., 2014, 25(2), 351–361.

33 J. J. Goodfellow, K. Baruah, K. Yamamoto, C. Bonomelli,

B. Krishna, D. J. Harvey, M. Crispin, C. N. Scanlan and

B. G. Davis, An endoglycosidase with alternative glycan 45 D. Shinmi, E. Taguchi, J. Iwano, T. Yamaguchi, K. Masuda,

specificity allows broadened glycoprotein remodelling,

J. Am. Chem. Soc., 2012, 134(19), 8030–8033.

34 C. W. Lin, M. H. Tsai, S. T. Li, T. I. Tsai, K. C. Chu,

Y. C. Liu, M. Y. Lai, C. Y. Wu, Y. C. Tseng, S. S. Shivatare,

J. Enokizono and Y. Shiraishi, One-step conjugation

method for site-specific antibody-drug conjugates through

reactive cysteine-engineered antibodies, Bioconjugate

Chem., 2016, 27(5), 1324–1331.

C. H. Wang, P. Chao, S. Y. Wang, H. W. Shih, Y. F. Zeng, 46 X. Li, T. Fang and G. J. Boons, Preparation of well-defined

T. H. You, J. Y. Liao, Y. C. Tu, Y. S. Lin, H. Y. Chuang,

C. L. Chen, C. S. Tsai, C. C. Huang, N. H. Lin,

C. Ma, C. Y. Wu and C. H. Wong, A common glycan struc-

antibody-drug conjugates through glycan remodeling and

strain-promoted azide–alkyne cycloadditions, Angew.

Chem., Int. Ed., 2014, 53(28), 7179–7182.

ture on immunoglobulin G for enhancement of eﬀector 47 Q. Zhou, J. E. Stefano, C. Manning, J. Kyazike, B. Chen,

functions, Proc. Natl. Acad. Sci. U. S. A., 2015, 112(34),

10611–10616.

35 I. Sassoon and V. Blanc, Antibody-drug conjugate (ADC)

clinical pipeline: a review, Methods Mol. Biol., 2013, 1045,

1–27.

D. A. Gianolio, A. Park, M. Busch, J. Bird, X. Zheng,

H. Simonds-Mannes, J. Kim, R. C. Gregory, R. J. Miller,

W. H. Brondyk, P. K. Dhal and C. Q. Pan, Site-specific anti-

body-drug

conjugation

through

glycoengineering,

Bioconjugate Chem., 2014, 25(3), 510–520.

36 P. J. Carter and P. D. Senter, Antibody-drug conjugates for 48 E. Boeggeman, B. Ramakrishnan, M. Pasek, M. Manzoni,

cancer therapy, Cancer J., 2008, 14(3), 154–169.

37 P. M. Drake and D. Rabuka, An emerging playbook for anti-

body-drug conjugates: lessons from the laboratory and

clinic suggest a strategy for improving eﬃcacy and safety,

Curr. Opin. Chem. Biol., 2015, 28, 174–180.

A. Puri, K. H. Loomis, T. J. Waybright and P. K. Qasba, Site

specific conjugation of fluoroprobes to the remodeled Fc

N-glycans of monoclonal antibodies using mutant glycosyl-

transferases: application for cell surface antigen detection,

Bioconjugate Chem., 2009, 20(6), 1228–1236.

38 N. Jain, S. W. Smith, S. Ghone and B. Tomczuk, Current 49 Z. Zhu, B. Ramakrishnan, J. Li, Y. Wang, Y. Feng,

ADC Linker Chemistry, Pharm. Res., 2015, 32(11), 3526–

3540.

39 C. R. Behrens and B. Liu, Methods for site-specific drug

conjugation to antibodies, mAbs, 2014, 6(1), 46–53.

P. Prabakaran, S. Colantonio, M. A. Dyba, P. K. Qasba and

D. S. Dimitrov, Site-specific antibody-drug conjugation

through an engineered glycotransferase and a chemically

reactive sugar, mAbs, 2014, 6(5), 1190–1200.

Org. Biomol. Chem.

View Article Online

Paper

Organic & Biomolecular Chemistry

50 B. M. Zeglis, C. B. Davis, R. Aggeler, H. C. Kang, A. Chen, 62 W. Huang, H. Ochiai, X. Zhang and L. X. Wang,

B. J. Agnew and J. S. Lewis, Enzyme-mediated methodology

for the site-specific radiolabeling of antibodies based on

catalyst-free click chemistry, Bioconjugate Chem., 2013,

24(6), 1057–1067.

Introducing N-glycans into natural products through a

chemoenzymatic approach, Carbohydr. Res., 2008, 343(17),

2903–2913.

63 H. Ochiai, W. Huang and L. X. Wang, Expeditious chemo-

enzymatic synthesis of homogeneous N-glycoproteins car-

rying defined oligosaccharide ligands, J. Am. Chem. Soc.,

2008, 130(41), 13790–13803.

51 R. van Geel, M. A. Wijdeven, R. Heesbeen, J. M. Verkade,

A. A. Wasiel, S. S. Van Berkel and F. L. Van Delft,

Chemoenzymatic Conjugation of Toxic Payloads to the

Globally Conserved N-Glycan of Native mAbs Provides 64 T. W. Rising, C. D. Heidecke, J. W. Moir, Z. Ling and

Homogeneous and Highly Eﬃcacious Antibody-Drug

Conjugates, Bioconjugate Chem., 2015, 26(11), 2233–2242.

52 P. Tumbale, H. Jamaluddin, N. Thiyagarajan, K. R. Acharya

A. J. Fairbanks, Endohexosaminidase-catalysed glycosylation

with oxazoline donors: fine tuning of catalytic eﬃciency and

reversibility, Chemistry, 2008, 14(21), 6444–6464.

and K. Brew, Screening a limited structure-based library 65 M. Umekawa, W. Huang, B. Li, K. Fujita, H. Ashida,

identifies UDP-GalNAc-specific mutants of alpha-1,3-galacto-

syltransferase, Glycobiology, 2008, 18(12), 1036–1043.

53 A. Seko, M. Koketsu, M. Nishizono, Y. Enoki,

H. R. Ibrahim, L. R. Juneja, M. Kim and T. Yamamoto,

L. X. Wang and K. Yamamoto, Mutants of Mucor hiemalis

endo-beta-N-acetylglucosaminidase show enhanced trans-

glycosylation and glycosynthase-like activities, J. Biol.

Chem., 2008, 283(8), 4469–4479.

Occurrence of a sialylglycopeptide and free sialylglycans in 66 W. Huang, C. Li, B. Li, M. Umekawa, K. Yamamoto,

hen’s egg yolk, Biochim. Biophys. Acta, 1997, 1335(1–2),

23–32.

54 B. Sun, W. Bao, X. Tian, M. Li, H. Liu, J. Dong and

W. Huang, A simplified procedure for gram-scale pro-

X. Zhang and L. X. Wang, Glycosynthases enable a highly

eﬃcient chemoenzymatic synthesis of N-glycoproteins car-

rying intact natural N-glycans, J. Am. Chem. Soc., 2009,

131(6), 2214–2223.

duction of sialylglycopeptide (SGP) from egg yolks and sub- 67 M. Noguchi, T. Tanaka, H. Gyakushi, A. Kobayashi and

sequent semi-synthesis of Man3GlcNAc oxazoline,

Carbohydr. Res., 2014, 396, 62–69.

S. Shoda, Eﬃcient Synthesis of Sugar Oxazolines from

Unprotected N-Acetyl-2-amino Sugars by Using

55 T. B. Parsons, W. B. Struwe, J. Gault, K. Yamamoto,

T. A. Taylor, R. Raj, K. Wals, S. Mohammed,

Chloroformamidinium Reagent in Water, J. Org. Chem.,

2009, 74(5), 2210–2212.

C. V. Robinson, J. L. Benesch and B. G. Davis, Optimal 68 K. Yamamoto, S. Kadowaki, M. Fujisaki, H. Kumagai and

Synthetic Glycosylation of a Therapeutic Antibody, Angew.

Chem., Int. Ed., 2016, 55(7), 2361–2367.

56 S. Puthenveetil, S. Musto, F. Loganzo, L. N. Tumey,

C. J. O’Donnell and E. Graziani, Development of Solid-

T. Tochikura, Novel Specificities of Mucor Hiemalis Endo-

Beta-N-Acetylglucosaminidase Acting Complex Asparagine-

Linked Oligosaccharides, Biosci., Biotechnol., Biochem.,

1994, 58(1), 72–77.

Phase Site-Specific Conjugation and Its Application toward 69 W. Huang, Q. A. Yang, M. Umekawa, K. Yamamoto and

Generation of Dual Labeled Antibody and Fab Drug

Conjugates, Bioconjugate Chem., 2016, 27(4), 1030–1039.

57 X. L. Li, J. T. Patterson, M. Sarkar, L. Pedzisa,

T. Kodadek, W. R. Roush and C. Rader, Site-Specific Dual

Antibody Conjugation via Engineered Cysteine and

L.

X.

Wang,

Arthrobacter

Endo-beta-N-Acetyl

glucosaminidase Shows Transglycosylation Activity on

Complex-Type N-Glycan Oxazolines: One-Pot Conversion of

Ribonuclease

B

to

Sialylated

Ribonuclease

C,

ChemBioChem, 2010, 11(10), 1350–1355.

Selenocysteine Residues, Bioconjugate Chem., 2015, 26(11), 70 C. Besanceney-Webler, H. Jiang, T. Q. Zheng, L. Feng,

2243–2248.

D. S. Del Amo, W. Wang, L. M. Klivansky, F. L. Marlow,

Y. Liu and P. Wu, Increasing the Eﬃcacy of Bioorthogonal

Click Reactions for Bioconjugation: A Comparative Study,

Angew. Chem., Int. Ed., 2011, 50(35), 8051–8056.

58 B. Li, Y. Zeng, S. Hauser, H. Song and L. X. Wang, Highly

eﬃcient endoglycosidase-catalyzed synthesis of glycopep-

tides using oligosaccharide oxazolines as donor substrates,

J. Am. Chem. Soc., 2005, 127(27), 9692–9693.

59 H. Li, B. Li, H. Song, L. Breydo, I. V. Baskakov and

L. X. Wang, Chemoenzymatic synthesis of HIV-1

V3 glycopeptides carrying two N-glycans and eﬀects of

71 N. J. Agard, J. A. Prescher and C. R. Bertozzi, A strain-pro-

moted [3+2] azide-alkyne cycloaddition for covalent modifi-

cation of blomolecules in living systems, J. Am. Chem. Soc.,

2004, 126(46), 15046–15047.

glycosylation on the peptide domain, J. Org. Chem., 2005, 72 J. M. Baskin, J. A. Prescher, S. T. Laughlin, N. J. Agard,

70(24), 9990–9996.

P. V. Chang, I. A. Miller, A. Lo, J. A. Codelli and

C. R. Bertozzi, Copper-free click chemistry for dynamic in

vivo imaging, Proc. Natl. Acad. Sci. U. S. A., 2007, 104(43),

16793–16797.

60 B. Li, H. Song, S. Hauser and L. X. Wang, A highly eﬃcient

chemoenzymatic approach toward glycoprotein synthesis,

Org. Lett., 2006, 8(14), 3081–3084.

61 T. W. Rising, T. D. Claridge, N. Davies, D. P. Gamblin, 73 X. H. Ning, J. Guo, M. A. Wolfert and G. J. Boons,

J. W. Moir and A. J. Fairbanks, Synthesis of N-glycan oxazo-

lines: donors for endohexosaminidase catalysed glycosyla-

tion, Carbohydr. Res., 2006, 341(10), 1574–1596.

Visualizing metabolically labeled glycoconjugates of living

cells by copper-free and fast huisgen cycloadditions, Angew.

Chem., Int. Ed., 2008, 47(12), 2253–2255.

Org. Biomol. Chem.

Article Doi

One-pot: N -glycosylation remodeling of IgG with non-natural sialylglycopeptides enables glycosite-specific and dual-payload antibody-drug conjugates

DOI: 10.1039/c6ob01751g

Source and publish data:

Authors:

Article abstract of DOI:10.1039/c6ob01751g

Full text of DOI:10.1039/c6ob01751g

Products guided by the article

R&D Labs maybe for 134272-63-2

Relevant to this article

Hot Product