CD22 Ligands on a Natural N-Glycan Scaffold Efficiently Deliver Toxins to B-Lymphoma Cells

Article abstract of DOI:10.1021/jacs.7b03208

CD22 is a sialic acid-binding immunoglobulin-like lectin (Siglec) that is highly expressed on B-cells and B cell lymphomas, and is a validated target for antibody and nanoparticle based therapeutics. However, cell targeted therapeutics are limited by their complexity, heterogeneity, and difficulties in production. We describe here a chemically defined natural N-linked glycan scaffold that displays high affinity CD22 glycan ligands and outcompetes the natural ligand for the receptor, resulting in single molecule binding to CD22 and endocytosis into cells. Binding affinity is increased by up to 1500-fold compared to the monovalent ligand, while maintaining the selectivity for hCD22 over other Siglecs. Conjugates of these multivalent ligands with auristatin and saporin toxins are efficiently internalized via hCD22 resulting in killing of B-cell lymphoma cells. This single molecule ligand targeting strategy represents an alternative to antibody- and nanoparticle-mediated approaches for delivery of agents to cells expressing CD22 and other Siglecs.

Full text of DOI:10.1021/jacs.7b03208

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
CD22 Ligands on a Natural N-Glycan Scaffold
Efficiently Delivers Tox-ins to B-Lymphoma Cells
Wenjie Peng, and James C Paulson
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b03208 • Publication Date (Web): 22 Aug 2017
Downloaded from http://pubs.acs.org on August 22, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 10
Journal of the American Chemical Society
1
2
3
4
5
CD22 Ligands on a Natural N-Glycan Scaffold Efficiently Delivers Toxins
to B-Lymphoma Cells
6
7
Wenjie Peng and James C. Paulson*
8
9
Departments of Molecular Medicine and Immunology & Microbiology, The Scripps Research Institute, 10550 N.
Torrey Pines Road, La Jolla, CA 92037, USA.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
KEYWORDS Siglec, N-linked glycan, biodegradable scaffold, drug delivery, enzymatic synthesis.
Supporting Information
S
ABSTRACT: CD22 is a sialic acidꢀbinding immunoglobulinꢀlike lectin (Siglec) that is highly expressed on Bꢀcells and B cell lymꢀ
phomas, and is a validated target for antibody and nanoparticle based therapeutics. However, cell targeted therapeutics are limited
by their complexity, heterogeneity and difficulties in producing. We describe here a chemicallyꢀdefined natural Nꢀlinked glycan
scaffold that displays high affinity CD22 glycan ligands and outcompetes the natural ligand for the receptor, resulting in single
molecule binding to CD22 and endocytosis into cells. Binding affinity is increased by up to 1500ꢀfold compared to the monovalent
ligand, while maintaining the selectivity for hCD22 over other Siglecs. Conjugates of these multivalent ligands with auristatin and
saporin toxins are efficiently internalized via hCD22 resulting in killing of Bꢀcell lymphoma cells. This single molecule ligand tarꢀ
geting strategy represents an alternative to antibody and nanoparticle mediated approaches for delivery of agents to cells expressing
CD22 and other Siglecs.
mond represents high affinity ligands for CD22. NꢀGlycan structure is
INTRODUCTION
displayed using the Symbol Nomenclature for Glycomics.16
Human CD22 (hCD22), or Siglecꢀ2, is a member of the sialic
acidꢀbinding immunoglobulinꢀlike lectin (Siglec) family and is
selectively expressed on the surface of B cells and B cell lymꢀ
phomas.1 CD22 is an inhibitory receptor of the B cell receptor
(BCR) whose activity is modulated by binding to its ligands,
comprising the sequence Neu5Acα2ꢀ6Galβ1ꢀ4GlcNAc that
terminate Nꢀlinked glycans on cell surface glycoproteins of B
cells and many other cells.2ꢀ5 In the right context, ligand bindꢀ
ing can recruit CD22 to the BCR to suppress autoimmune B
cell responses to antigens on ‘self’ cells, or sequester CD22
from the BCR to allow appropriate Bꢀcell activation against
pathogens that do not contain CD22 ligands.6ꢀ8
Targeting CD22 using sialic acid containing ligands provides
an alternative to antibodies (Figure 1).1, 17ꢀ18 To overcome the
low avidity of natural sialosides, several groups have emꢀ
ployed sialic acid analogues with substituents at the Cꢀ9
and/or Cꢀ5 positions to increase avidity and selectivity for
17ꢀ19
CD22.1,
Exemplary analogs for human CD22 are 9ꢀNꢀ
biphenyl carboxamideꢀNeu5Ac (BPCNeu5Ac; 1a)20 and 9ꢀNꢀmꢀ
phenoxybenzamideꢀNeu5FAc (MPBNeu5FAc; 1c)21 that have
affinities in the low micromolar range. Yet they do not have
enough avidity to compete with natural glycoprotein ligands
on B cells, so called cisꢀligands,22 requiring multivalent
presentation on polymers or nanoparticles for stable binding to
and uptake by cells.15, 17, 19, 23ꢀ26 Since CD22 is an endocytic
receptor that constitutively recycles between the cell surface
and endosomes, multivalent ligands bound to CD22 are rapidꢀ
ly endocytosed, released in the acidic endosome, and then
accumulate inside the cell over time.27 In contrast, antibodies
are not released in the endosome, and recycle with CD22 to
the cell surface, representing an advantage of ligandꢀmediated
vs antibody mediated delivery of cargos.15, 19, 27ꢀ29
Because CD22 is selectively expressed on Bꢀcell lymphomas
and leukemias it has become an attractive target for cell diꢀ
rected cancer therapy (Figure 1).9ꢀ10 Humanized antiꢀhCD22
antibodies (conjugated with drugs/toxins) have been used in
clinical trials to treat Bꢀcell NonꢀHodgkin’s Lymphoma and
acute lymphocytic leukemia.9, 11ꢀ13 However, these antibodyꢀ
based therapies often suffer from low efficacy and serious side
effects due to the recycling of the hCD22–antibody complex
back to the cell surface, difficulties in producing antibody–
drug conjugates and heterogeneity of conjugates.12, 14
Attempts to create paucivalent (diꢀ and triꢀvalent) ligands that
would compete with cisꢀligands, bind to CD22, and carry carꢀ
go into B cells has met with limited success.24, 30ꢀ31 Several
groups successfully produced paucivalent ligands using
BPCNeu5Ac as the high affinity sialic acid analog. While they
bound to CD22 with significantly higher avidity, the ligands
did not bind to B lymphoma cells without first treating cells
with sialidase to destroy endogenous cisꢀligands, or alternaꢀ
tively by further multimerization. We previously investigated
the impact of biological spacing on valency required for mulꢀ
tivalent binding using a biꢀfunctional CD22 ligand comprising
BPCNeu5Ac coupled to a hapten, nitrophenol (NP), recognized
by antiꢀNP antibodies. This biꢀfunctional ligand could mediꢀ
Figure 1. Schematic illustration of liposome15, Antibody Drug Conju-
gate (ADC)9-12 and N-glycan scaffold for targeting CD22-positive B-
lymphomas cells. Red circle with letter T represents toxin. Magenta diaꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
ate assembly of a multivalent complex between an antiꢀNP
antibody (IgM, IgA, IgG) and CD22 on cells that was then
endocytosed by CD22.24 This showed that a paucivalent ligand
complex with appropriate biological spacing, in this case deꢀ
fined by the valency and geometry of the antibody (n=2ꢀ10),
could bind to and be endocytosed by CD22.
Page 2 of 10
1
2
3
4
5
6
7
8
9
In the search for a more chemically defined paucivalent conꢀ
struct, we were inspired by the work of Lee et al., who showed
that a branched Nꢀlinked glycan by itself provided biological
spacing of terminal galactose residues for the asialoglycoproꢀ
tein receptor on hepatocytes, with each branch adding 10ꢀ100
fold increased avidity.32 To explore the possibility that
branched Nꢀlinked glycans functionalized with sialic acids
might similarly increase affinity for CD22 Siglec ligands (Figꢀ
ure 1), we chemoenzymatically synthesized a library of Nꢀ
glycan scaffolds bearing hCD22 ligands at their nonꢀreducing
end and screened their relative inhibitory potency. Remarkaꢀ
bly, we find that diꢀ and triꢀbranched ligands based on natural
Nꢀlinked glycan scaffolds provide biological spacing that
dramatically increases affinity for CD22 from micromolar to
low nanomolar/high picomolar avidity, allowing competition
with cisꢀligands on B cells, and binding to and endocytosis by
CD22. Conjugates of these ligands with two toxins, saporin
and auristatin exhibit high cytotoxicity against hCD22ꢀpositive
Bꢀcell lymphoma cells.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
Synthesis of di- and trivalent substituted sialosides on
branched natural glycan scaffolds. A panel of branched
CD22 ligands were synthesized chemoenzymatically by chemꢀ
ical synthesis of 9ꢀC and 5ꢀC substituted sialic acids followed
by enzymatic transfer to glycan scaffolds (Figure 2B and Taꢀ
ble S1). The 9ꢀC and 5ꢀC sialic acids were prepared similarly
to our published methods21 with some modifications as deꢀ
scribed in Experimental Procedures and Supplementary Inꢀ
formation. These sialic acid analogs, 1a-c are readily prepared
to gram scale.
Figure 2. A library of glycan scaffolds bearing hCD22 synthetic lig-
ands via chemoenzymatic synthesis. A) Chemical structures of sialic
acid analog ligands of CD22 (1a-c). BPC: biphenyl carbonyl; MPB: mꢀ
phenoxybenzamide. B) Synthetic ligands of hCD22 based on linear, or
branched Oꢀ or Nꢀglycan scaffolds. Glycan structures are displayed using
the Symbol Nomenclature for Glycomics.16
To assess the ability of enzymatic transfer of the modified
sialic acids to linear and branched scaffolds in a oneꢀpot reacꢀ
tion we mixed LacNAc (2a), a sialic acid analog (1a-c) and
CTP with the CMPꢀsialic acid synthetase (NmCSS) and a siꢀ
alyltransferase (hST6GalꢀI or Pd2ꢀ6ST). Strikingly we found
ACS Paragon Plus Environment

Page 3 of 10
Journal of the American Chemical Society
that NmCSS and both sialyltransferases tolerated modification
of polymer ligand binding to beads coated with hCD22ꢀFc
chimera (Figure S1).36 In both assays, ligand 3b shows over
100ꢀfold higher binding affinity for hCD22 than 3a. For the
assay employing hCD22ꢀCHO cells, we reason that the two
BPCNeu5Ac residues of 3b simultaneously bind two hCD22
monomers resulting in a dramatic increase in binding affinity.
1
2
3
4
5
6
7
8
of Neu5Ac at Cꢀ9 with BPC or MPB resulting in synthesis of
the linear analog 3a or 3h, respectively, in excellent yield (Taꢀ
ble S1). We then applied the same strategy to other galactose
terminated Nꢀglycan scaffolds on 5ꢀ40 mg scale and obtained
a library of high affinity ligands for hCD22 (Figure 2B; Table
S1). After purification by gel filtration chromatography or
reverse phase HPLC, the final products were lyophilized and
the target ligands were obtained in excellent yield (88–96%,
Table S1). Among the targets, analogues 3a and 3h are monoꢀ
valent ligands, while branched glycans include an Oꢀlinked
glycan (3d), biantennary (3b-c, 3g, 3i) and triantennary (3e-f,
3j) Nꢀlinked glycans. Analogues 3k and 3l are multiꢀsialylated
diꢀLacNAc products, which were prepared with bacterial siꢀ
alyltransferase.33ꢀ34 Notably, the Nꢀglycans are derived from a
natural glycan isolated from eggs, (SGP, Figure 3A) which is
attached to asparagine, facilitating subsequent conjugation
strategies.35
The IC50 values for all the sialosides (3a-l) are also summaꢀ
rized in Figure 3B. Also shown is the relative inhibitory poꢀ
tency (rIP) as a fold increase in potency over the reference
compound 3a. All branched analogue ligands (3b-g, 3iꢀ3l)
demonstrated much higher binding affinity for hCD22 than
monovalent analogue ligand 3a with rIP values between 66
and 1500. Triantennary analogues 3e, 3f and 3j (IC50 = 0.22,
0.23 and 0.5 nM, respectively) exhibit better inhibitory potenꢀ
cy than biantennary analogues. The length of the branches
from the mannose core does not appear to influence signifiꢀ
cantly the binding affinity in 3b vs 3c and 3e vs 3f. Moreover,
further substitution of the internal LacNAc repeat units with
additional BPCNeu5Ac moieties actually reduces binding affiniꢀ
ty (e.g. 3c vs 3l).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For preparation of high affinity ligands for mCD22 and mSn,
the oneꢀpot reactions using the sialic acid analogs BPANeu5Gc
or TCCNeu5Ac were not successful. We found that the NmCSS
was unable to convert the BPANeu5Gc or TCCNeu5Ac to the
corresponding CMPꢀsialic acid derivative. Thus, we instead
transferred 9ꢀNH2ꢀNeu5Gc/Neu5Ac to galactose acceptors by
combination of NmCSS and sialyltransferase (hST6GalꢀI for
mCD22 ligands and rST3GalꢀIII for mSn ligands) as previousꢀ
ly reported,26, 36 followed by acylation of amine with BPAꢀ
NHS or TCCꢀNHS to form BPA or TCC ligands, respectively,
in excellent yield (Scheme S3).
Avidity and specificity for CD22. To assess the relative afꢀ
finities of the panel of CD22 ligands in Figure 2B we emꢀ
ployed a competition assay using Chinese hamster ovary
(CHO) cells expressing hCD22 (hCD22ꢀCHO), and a multivaꢀ
lent polymeric ligand, Neu5GcꢀPAA, comprising a low affiniꢀ
ty ligand (Neu5Gcα2ꢀ6Galβ1ꢀ4GlcNAc) coupled to polyꢀ
acrylamide (PAA) with a biotin tag. Sialoside ligands quantiꢀ
fied using the periodateꢀresorcinol assay,37 were serially dilutꢀ
ed and added with Neu5GcꢀPAA to the cells for 40 min at 4°
C, a temperature that prevents endocytosis. After washing the
cells, bound Neu5GcꢀPAA is detected using fluorescent laꢀ
beled streptavidin. Affinities of the ligands (3a-h) are comꢀ
pared by their ability to prevent binding of the Neu5GcꢀPAA
to cells (Figure 3).
Figure 3. Determination of the binding affinity and selectivity of syn-
thetic ligands for hCD22. A) Assessment of the ability of the hCD22
ligands to block binding of a multivalent ligand to hCD22ꢀCHO cells
using a flow cytometryꢀbased competitive binding assay. B) Summary of
IC50 and relative inhibitory potency (rIP) values of ligands in Figure 2B
assessed in the same assay described in panel A. Triantennary analogue
ligands are highlighted in red. C) Selective binding of monovalent ligand
3a and triantennary Nꢀglycan 3j for hCD22 assessed on a panel of Siglecꢀ
expressing CHO cells. Binding affinities were determined by flow
cytometry.
Shown as examples are titrations for 6’SLN (6’ꢀsialylꢀNꢀ
acetylꢀlactosamine; Neu5Acα2ꢀ6Galβ1ꢀ4GlcNAc), a fragment
of the natural ligand for hCD22, the biantennary Nꢀglycan
with the same terminal sequence SGP, and the corresponding
analogs of these glycans terminated with BPCNeu5Ac instead
of Neu5Ac, 3a and 3b are shown in Figure 3A. The IC50 valꢀ
ues for these compounds are summarized in Figure 3B. As a
reference, 6’SLN gave an IC50 of 62 µM, which is close to the
Kd value of 32 µM determined for binding to 6’SLN hCD22 at
4° C by equilibrium dialysis.5 Thus we can presume that the
IC50 values represent a close approximation of their affinities.
The corresponding biantennary Nꢀglycan SGP shows a threeꢀ
fold higher binding affinity for hCD22 compared to the monoꢀ
valent ligand. In contrast, the scaffolds with BPCNeu5Ac were
330 nM and 1.8 nM for the monovalent 3a and divalent 3b Nꢀ
glycan, respectively. The results were further validated by
using an independent analogous assay involving competition
Previously we reported that the MBPNeu5FAc ligand exhibited
high specificity for hCD22.21 To evaluate the selectivity of the
multivalent ligand for hCD22, we assessed binding to a panel
of recombinant Siglecꢀexpressing CHO cells as described preꢀ
viously.21, 36 Accordingly, we prepared probes by conjugating
3a and 3j to fluorescein (FITC) (See Scheme S2A; 3a-FITC,
3j-FITC). These were evaluated for binding to CHO cells
expressing no Siglec (WTꢀCHO) or CHO cells expressing
various human and murine Siglecs by flow cytometry (Figure
3C). Relative to WTꢀCHO cells, 3a and 3j bound only to
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 10
hCD22ꢀCHO cells and not cells expressing other siglecs,
showing that selectivity for hCD22 is maintained even for the
high affinity multivalent ligand 3j.
cargo to endosomal compartments.25, 27 To further establish
1
2
3
4
5
6
7
8
that the 3j-FITC was undergoing endocytosis we conducted
confocal imaging studies (Figure 4C). hCD22ꢀCHO and WTꢀ
CHO cells grown on glass cover slips were overlaid with meꢀ
dia containing the triꢀantennary ligand 3j-FITC and incubated
for 3h at 37 °C. To determine the endocytosis mechanism of
ligand the endosomal marker transferrinꢀAlexaFluor555 was
also added to the media concurrently with ligand. Following
incubation cells were fixed and permeabilized. Cells were then
stained with antiꢀhCD22ꢀAPC, with antiꢀFITCꢀAlexaFluor488
for enhancement of FITC signal, with antiꢀLAMPꢀbiotin
(CD107a) for localization of lysosomes, and finally with
Hoechst dye to stain nuclei.
Endocytosis of N-glycan ligands. We next investigated the
ability of the ligands to be endocytosed by hCD22. Four ligꢀ
ands (SGP, 3a, 3b and 3j) conjugated with FITC were added
to the media of WTꢀCHO or hCD22ꢀCHO cells at various
concentrations for 60ꢀminutes at 37 ºC. Cells were washed and
assessed for the amount of ligand taken up by flow cytometry.
As shown in Figure 4A, hCD22ꢀCHO cells showed no signifiꢀ
cant uptake of the monovalent ligand 3a-FITC and the natural
(low affinity) biantennary ligand SGP-FITC at concentrations
of up to 400 nM. In contrast, hCD22ꢀCHO cells exhibited
dramatic uptake of 3j-FITC, the triantennary Nꢀglycan capped
with MPBNeu5FAcs, at concentrations as low as 1.6 nM, and
approaching near maximal at 10ꢀ20 nM. This is consistent
with a receptorꢀmediated process by CD22 saturated at low
nM concentrations. The corresponding biantennary analog 3b-
FITC was also taken up by hCD22ꢀCHO cells but saturated at
higher concentrations than 3j-FITC, in keeping with its lower
affinity. The time course for uptake was rapid when measurꢀ
ing uptake of 50 nM 3bꢀ and 3j-FITC over time (Figure 4B).
The fact that hCD22 is indeed mediating uptake is confirmed
by complete lack of uptake of 3bꢀ and 3j-FITC by wild type
WTꢀCHO cells (Figure 4A and 4B).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We found that 3j-FITC was indeed endocytosed by hCD22ꢀ
CHO cells and coꢀlocalized in cells with hCD22 and with
transferrinꢀAlexaFluor555 (Figure 4C, top) in a periꢀnuclear
location consistent with the clathrin mediated endocytosis of
CD22 and transferrin to transꢀGolgi endosomal compartments
as documented previously.25, 38ꢀ39 Conversely, there was no coꢀ
localization with lysosomes (LAMP staining; Figure 4C, midꢀ
dle). As expected, there was no 3j-FITC staining of wild type
WTꢀCHO cells (Figure 4C, bottom). Monovalent ligand 3a-
FITC was not internalized by hCD22ꢀCHO cells (Figure S2).
We had previously demonstrated that CD22 mediates endocyꢀ
tosis through a clathrin dependent mechanism, and delivers
Figure 4. Bi- and tri-antennary ligands of hCD22 are endocytosed by CD22-CHO cells. A) Serial dilutions (400ꢀ0.4 nM) of ligand–FITC (SGPꢀ, 3a-,
3b-, and 3j-FITC) conjugates were incubated with hCD22ꢀCHO or WTꢀCHO cells at 37 °C for 60 min. Cells were washed and endocytosed ligand was
assessed by flow cytometry. B) Ligand–FITC conjugates (50 nM) were incubated with hCD22ꢀCHO or WTꢀCHO cells at 37 °C for various times. Conjuꢀ
gates 3b- and 3j-FITC were efficiently endocytosed by hCD22ꢀCHO cells, but not WTꢀCHO cells. C) Endocytosis of the triantennary hCD22 Nꢀglycan
ligand by hCD22ꢀCHO cells. hCD22ꢀCHO cells (top and middle) or WTꢀCHO (bottom) cells on coverslips were overlayed with the hCD22 ligand 3j-FITC
in culture media for 3 h at 37°C. Cells were then washed, fixed, and counterstained with nuclei (blue), ligand (green), hCD22 (yellow) and Tf receptor or
lysosome (red) followed by confocal fluorescence microscopy. DIC: differential interference contrast. Scale bar is 10 ꢀm.
ACS Paragon Plus Environment

Page 5 of 10
Journal of the American Chemical Society
We also assessed the uptake of the FITC conjugates of SPGꢀ, a 50 fold enhancement, and consistent with endocytosis of this
1
2
3
4
5
6
7
8
3aꢀ, 3bꢀ and 3j by Daudi, cells which highly express natural
α2ꢀ6ꢀsialosides as cis ligands that can compete with exogeꢀ
nous ligands for binding to CD22 (Figure 5A). The ligands
were incubated with Daudi cells in pure mouse serum at variꢀ
ous concentrations for 60ꢀminutes at 37 ºC. After washing, the
amount of ligand taken up was assessed by flow cytometry. As
illustrated in Figure 5A, the triantennary Nꢀglycan ligand 3j-
FITC bound to and was internalized by Daudi cells. As exꢀ
pected the lower affinity biꢀantennary ligand 3b was less effiꢀ
ciently taken up, and the monovalent analogue (3a) and the
natural ligand (SGP) had weak or no binding/uptake.
single molecule Nꢀglycan ligand conjugate by hCD22.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cytotoxicity of N-glycan ligand conjugates
With a goal of targeting hCD22ꢀexpressing cells for immunoꢀ
therapy, we designed conjugates of ligands with the saporin
and auristatin toxins and evaluated their ability to kill Daudi
lymphoma B cells. Saporin is a ribosome inactivating protein
and commercially available as a conjugate to streptavidin.
Thus, we prepared biotinylated ligands by treatment 2a, 3h
and 3j with NHS activated biotin reagent (Scheme S2B). All
the products were subjected to semiꢀpreparative RPꢀHPLC
separation and the purities were further determined by analytiꢀ
cal HPLC (See supplementary HPLC Spectra). By mixing the
biotinylated ligands (2a-, 3h- and 3j-Biotin) with the streptavꢀ
idinꢀsaporin by 1:1 molar ratio, ligandꢀsaporin conjugates (2a-
, 3h- and 3j-saporin, Scheme S2B) were obtained. Without
further purification, the mixture was serially diluted to the
indicated concentrations and used in the cytotoxicity assay.
Figure 5. The triantennary glycan scaffold can be endocytosed and
deliver conjugated toxins to Daudi B lymphoma cells. A) Daudi cells
were incubated with serial dilutions of ligand–FITC conjugates in pure
mouse serum at 37 °C for 60 mins. Endocytosed ligands were assessed by
flow cytometry. B) Serial dilutions of saporinꢀstreptavidin complex with
biotinylatedꢀglycan ligands were mixed with Daudi cells for 72h, followed
by assessment of viability using the chromogenic MTT assay. The absorbꢀ
ance was measured at 570 nm. Ligands were LacNAcꢀbiotin (2a-biotin)
with no sialic acid, and MBPNeu5FAc bearing monovalent (3h-biotin) and
triꢀvalent Nꢀglycan (3j-biotin). C) Serial dilutions of ligandꢀMMAF conꢀ
jugates were mixed with Daudi cells for 72h, followed by assessment of
viability using the chromogenic MTT assay. The absorbance was measꢀ
ured at 570 nm. Ligands were MBPNeu5FAc bearing monovalent (3h) and
triꢀvalent Nꢀglycan (3j). MMAF alone was used as a control.
As shown in Figure 5B, the CD22 ligandꢀtoxin conjugate with
the triantennary Nꢀglycan 3j (3j-saporin) had extremely high
cytotoxicity with an EC50 of 0.21 nM, over 1300ꢀfold higher
than the monovalent ligand conjugate 3h-saporin (EC50 = 287
nM), and >5000 fold higher potency than the asialo conjugate
2a-saporin, used as a negative control, which showed undeꢀ
tectable cytotoxicity at 1 µM, in keeping with the differences
in the avidities of the ligands (Figure 3 B).
Since streptavidin is tetravalent, the saporinꢀstreptavidinꢀ
CD22 ligand complex is in principle a mixture of conjugates
with 1 to 4 CD22 ligands. Thus, it is possible that more than
one ligand is required for binding and endocytosis of the 3j-
saporin conjugate.
Our results show that triantennary glycan scaffold not only
delivers large toxin complexes such as the saporin streptavidin
conjugates, but also small chemically defined peptide like
toxins like auristatin to B lymphoma cells. These show that the
triantennary Nꢀglycan scaffold has sufficient avidity that it can
outcompete the natural cisꢀligands on B cells to effectively
deliver toxins into the cell by hCD22 mediated endocytosis.
This single molecule targeting strategy therefore represents an
alternative to nanoparticle and antibody targeting of hCD22 to
deliver chemotoxins to B cell lymphomas.
To investigate the potential of 3j to mediate single molecule
targeting we turned to the chemically defined toxin monomeꢀ
thyl auristatin F (MMAF), which has previously been studied
for targeting B lymphoma cells when conjugated to antiꢀ
hCD22 antibodies.11, 40 To introduce a sulfhydryl (–SH) funcꢀ
tional group we treated hCD22 ligands (3h and 3j) with NHS
activated SPDP reagent, followed by reduction with TCEP.
The ligands, 3h-SH and 3j-SH, were then coupled to MMAF
by thiolꢀmaleimide coupling (Scheme S2C).
Targeting other Siglecs with N-glycan ligands. Given our
success in targeting hCD22 with conjugates of 3j, we explored
if this Nꢀglycan scaffold could also be applied to target other
Siglecs. Thus, we prepared conjugates of ligands for other
Siglecs, including BPANeu5Gc26 terminated analogs analogues
(4a-d) for targeting mCD22 and TCCNeu5Ac36 analogues (5a,
b) for targeting mSn (murine sialoadhesin). Because the CMPꢀ
sialic acid synthetase (NmCSS) and sialyltransferases do not
efficiently transfer BPANeu5Gc and TCCNeu5Ac to acceptor
substrates, we first transfered 9ꢀNH2ꢀNeu5Gc using hST6GalꢀI
or 9ꢀNH2ꢀNeu5Ac using rST3GalꢀIII to the desired galactose
terminated scaffolds, followed by acylation of the 9ꢀNH2 with
These toxin conjugates were then assessed in the cytotoxicity
assay as described above (Figure 5C). While free MMAF exꢀ
hibited toxicity with an EC50 of 200 nM, the monovalent ligꢀ
and exhibited reduced potency with EC50 of 1000 nM, preꢀ
sumably a result of reduced cell permeability and little or no
endocytosis by hCD22. In contrast, the triantennary Nꢀglycan
conjugate 3j-MMAF exhibited an EC50 = 20 nM representing
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 10
the corresponding NHS activated substituent to yield the corꢀ
Here we show that Siglec ligands based on diꢀ and triꢀvalent
responding BPANeu5Gc (for mCD22) and TCCNeu5Ac (for
mSn) products in excellent yield (Figure 6C, Scheme S3B,C).
Nꢀglycan scaffold results in dramatically increased affinity
over the corresponding monovalent ligand. Indeed, Nꢀglycan
based ligands of hCD22 exhibited high up to 1,500ꢀfold inꢀ
creased affinity with low nM/high pM avidity capable of bindꢀ
ing to and being endocytosed by CD22 on B lymphoma cells.
Since CD22 is a monomer, we suggest that the avidity gain
observed in a cell based binding assay must come from the
individual branches of the glycans interacting simultaneously
with multiple CD22 receptors on surface of the cell, which
would require that the ligand binding sites of two CD22 moleꢀ
cules come within 30ꢀ50 Å of each other.41 Of particular note
was that these ligands, as single molecules and without further
multimerization, are capable of being bound and endocytosed
by hCD22 on Daudi B lymphoma cells, which are known to
have abundant natural ligands of CD22 that bind in cis and
prevent binding of weak exogenous ligands in trans.22ꢀ23, 30
These ligands provides an alternative to nanoparticleꢀ and Abꢀ
mediated approaches for delivery of toxins and other cargo to
CD22 bearing B lymphoma cells. The fact that avidity gains
are also seen with Nꢀglycan based ligands of mCD22 and mSn
suggests that avidity gains could be achieved using the Nꢀ
1
2
3
4
5
6
7
8
Analysis of the BPANeu5Gc ligands using mCD22ꢀCHO cells
in the cellꢀbased competitive binding assay, showed that the
triantennary BPANeu5Gc ligand 4d (IC50 = 2.9 nM) had 360
fold higher binding affinity to mCD22 than monovalent ligand
4a and 13 fold higher than biantennary ligand 4b (Figure 6A).
The diꢀLacNAc extended ligand 4c has similar binding affinity
to mCD22 as monoꢀLacNAc extended ligand 4b, which is
consistent with the results for hCD22 (3b vs 3c and 3e vs 3f).
Analysis of the TCCNeu5Ac ligands using mSn Fc chimera in
the beadꢀbased competitive assay showed that the biantennary
ligand 5b had a 21ꢀfold higher binding affinity monovalent
ligand 5a (Figure 6B). The results show for several Siglecs
that the branched Nꢀglycan scaffold provides dramatically
enhanced avidities over monovalent ligands.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
linked glycan scaffold for other members of the Siglec family
.
MATERIALS AND METHODS
Materials: The recombinant enzymes, a CMPꢀsialic acid synꢀ
thetase (NmCSS), two mammalian sialyltransferases rat
ST3GalꢀIII and human ST6GalꢀI, and a ST6Gal sialyltransferꢀ
ase from Photobacterium damsella (Pd2,6ST) were prepared
as previously described.41ꢀ42 The Chinese hamster ovary
(CHO) parental (wild type) cell line (WTꢀCHO) and cells exꢀ
pressing human or murine Siglecs (SiglecꢀCHO) were deꢀ
scribed previously.21, 36 The recombinant siglecꢀFc chimeras,
hCD22 Fc23 and mSn Fc36, were prepared as previously deꢀ
scribed. A polymeric CD22 ligand (1 M Da), Neu5Gcα2ꢀ
6LacNAcꢀPAAꢀBiotin (Neu5GcꢀPAA), was obtained from the
Consortium for Functional Glycomics (CFG #: PA365). All
other chemicals and buffers were obtained in the highest puriꢀ
ty available from commercial suppliers.
Preparation of substituted sialic acids
BPCNeu5Ac (1a, Scheme S1A): Crude 9ꢀNH2ꢀNeu5Ac (12)21
and DIEA (5.0 equiv.) were dissolved in H2O, followed with
addition of 3.0 equivalents of biphenyl carboxylic acidꢀ(BPC)ꢀ
Nꢀhydrozysuccinimide ester (NHS). The reaction was stirred
º
at 0 C until the starting material was consumed. Then the reꢀ
action was purified by SepꢀPak C18 column (2 g, Waters
Corp.) and eluted with H2OꢀMeOH. The product was obtained
as inseparable α/β isomers in 95% yield.
MPBNeu5Ac (1b, Scheme S1A): was prepared similarly as
described above by stirring 12 with metaꢀphenoxy benzoic
acidꢀNHS (MPBꢀNHS) in water. After purification, α/β mixꢀ
ture of 1b was afforded in 93% yield.
Figure 6. Determination of the binding affinity of synthetic ligands for
mCD22 and mSn. A) Assessment of the ability of the mCD22 ligands to
block binding of a multivalent ligand to mCD22ꢀCHO cells using a flow
cytometryꢀbased competitive binding assay. B) Assessment of the ability
of the mSn ligands to block binding of mSn to the magnetic beads coated
with multivalent natural ligand using a flow cytometryꢀbased competitive
binding assay. C) Synthetic ligands of mCD22 and mSn based on linear,
or branched Nꢀglycan scaffolds. Glycan structures are displayed using the
Symbol Nomenclature for Glycomics.16 IC50 values were listed in parenꢀ
thesis behind the compound numbers. BPA: Biphenyl acetyl. TCC: 4Hꢀ
thieno[3,2ꢀc]chromeneꢀ2ꢀcarbamoyl.
MPBNeu5FAc (1c, Scheme S1B): In brief, the Nꢀacetyl group
of sialic acid derivative 13 was removed by heating to 60 °C
with MsOH in CH3OH. After neutralization with Et3N the
amine group of neuraminic acid 14 was reacted with fluoroaꢀ
cetyl chloride and DIEA at 0 °C to provide fluorinated sialic
acid 15. 9ꢀN3ꢀNeu5FAc 17 was obtained by treatment comꢀ
pound 15 with LiOH in water to hydrolyze methyl ester, and
Summary
ACS Paragon Plus Environment

Page 7 of 10
Journal of the American Chemical Society
with additional of I2 to hydrolyzed the thiol group. Azido linked 9ꢀNH2ꢀNeu5Gc sialosides 21a-d. By mixing 21a-d with
1
2
3
4
5
6
7
8
group of 17 was further reduced to amine by treatment with
PMe3. All of the reactions were monitored by thin layer chroꢀ
matography (TLC). All the intermediates were purified by
silica gel chromatography. The final product 1c was obtained
by stirring 9ꢀNH2ꢀNeu5FAc 18 with MPBꢀNHS, followed by
purification with SepꢀPak C18 column.
3.0 equivalents/amine BPAꢀNHS (biphenyl acetyl NHS ester)
and DIEA in 0.4 mL H2O at 0 °C for 4h, α2ꢀ6 linked
BPANeu5Gc sialosides 4a-d were obtained. The final products
were purified by semiꢀpreparative RPꢀHPLC.
TCCNeu5Ac (9-N-(4H-thieno[3,2-c]chromene-2-carbonyl)-
Neu5Ac) sialosides for mSn (Scheme S3C): the procedure
was applied similarly as described previously.36 Briefly, a
galactose terminated glycan (2a, b), 9ꢀNH2ꢀNeu5Ac (12, 1.5
equivalents/galactose) and CTP (2.0 equivalents/galactose)
were dissolved in TrisꢀHCl (100 mM, pH9.0) containing
MgCl2 (20 mM final concentration). The NmCSS (~100ꢀ500
mU) and α2ꢀ3ꢀsialyltransferase (~100ꢀ300 mU of rST3Galꢀ
III41) were added and incubated at 37 °C for overnight. Reacꢀ
tions were monitored by mass spectrometry and thin layer
chromatography (TLC). After completion, the reactions were
centrifuged and the supernatants were loaded to size exclusion
chromatography Sephadex Gꢀ25 (1 x100 cm, eluted with waꢀ
ter) to provide α2ꢀ3 linked 9ꢀNH2ꢀNeu5Ac sialosides 22a and
22b. By stirring 22a,b with 5 equivalents/amine TCCꢀNHS
ester and DIEA in 0.4 mL H2O at 0 °C for 4h, α2ꢀ3 linked
TCCNeu5Gc sialosides 5a and 5d were obtained. The final
products were purified by semiꢀpreparative RPꢀHPLC.
One-pot chemo-enzymatic synthesis of high affinity ligands
for hCD22
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Linear and biantennary scaffolds (2a-d, and 2h) were prepared
as described previously.33, 41ꢀ43 The synthetic strategy for trianꢀ
tennary scaffolds were illustrated in Scheme S4. In brief, trianꢀ
tennary glyopeptide 2341 was digested to the asparaginyl oliꢀ
gosaccharide 2e by treatment with the enzyme Actinase E
(Kaken Pharmaceutical, Japan).35 Next, the diꢀLacNAc (LN,
Nꢀacetylꢀlactosamine) extended triantennary glycan 2e was
further extended to triꢀLacNAc glycan 2f by using bacterial
β1ꢀ3GlcNAcꢀtransferase (from H. pylori) and β1ꢀ4Galꢀ
transferase (LgtB).43 Because the NH2 group of asparagine is
poorly reactive, it was acylated by NHS activated (2ꢀ
azido)glycine to form glycan 2g to facilitate subsequent conꢀ
jugation to biotin, fluorescent groups and toxins.
With the acceptors 2a-h in hand, we performed a oneꢀpot twoꢀ
enzyme strategy to generate the corresponding sialoside liꢀ
brary (Table S1). In brief, a galactose terminated acceptor (2a-
h, 5 ꢀ 40 mg) was mixed with 1.5 molar equivalents/galactose
of a sialic acid analog monosaccharide (1a-c) and 2.0 equivaꢀ
lents/galactose of CTP in Tris buffer (100 mM, pH8.5) with
MgCl2 (15 mM) to give a final concentration of 0.5 mM for
the galactose acceptor. NmCSS (~400ꢀ1000 mU) and an α2ꢀ
6ꢀsialyltransferase (~100ꢀ500 mU of either hST6GalꢀI41 or
Pd2ꢀ6ST44) were added to the reaction and incubated at 37 °C .
Reactions were monitored by mass spectrometry analysis and
thin layer chromatography (TLC). After the acceptor was conꢀ
sumed, the reaction was centrifuged and the supernatant was
subjected to a centrifuge filter with molecular mass cutꢀoff 10
kDa (Amicon Ultra, Millipore) to remove proteins. The filtrate
was purified with a semiꢀpreparative RPꢀHPLC. The products
were eluted by water and acetonitrile solvents containing with
0.1% TFA (trifluoroacetic acid). The purity of the final prodꢀ
uct was determined by an analytical HPLC monitoring with
UV wavelength at 215 nm and 254 nm (See HPLC spectra in
Supporting Information).
Measurement of IC50 by competitive cell-based binding
assay
Serial dilutions of the sialoside ligands (3a-l) were made in a
96 well plastic plate in 100 ꢁL cold FACS buffer (Hank’s Balꢀ
anced Salt Solution containing 0.5% BSA) the polymeric
CD22 ligand Neu5GcꢀPAA at 1 ꢁg/mL. Approximately 106
WTꢀCHO or SiglecꢀCHO cells were added and the plate was
maintained in dark for 40 mins at 4 ºC. After washing the cells
twice by suspension in cold FACS buffer (200 ꢁL) and cenꢀ
trifugation (300 x g), cells were incubated with Streptavidinꢀ
PE (1:1000 dilution, Biolegend) in 100 ꢁL FACS buffer in
dark for 30 mins at 4 ºC. After washing with 200 ꢁL cold
FACS buffer, cells were reꢀsuspended in FACS buffer (200
ꢁL) and analyzed by a flow cytometry (LSR II, BD Biosciencꢀ
es). Flow cytometry data were analyzed using the Flowjo
software. All titrations were performed in triplicate. Inhibition
data were analyzed and graphed with Prism software
(GraphPad) for the curveꢀfitting and IC50 calculations. No
binding (0%) was defined as background without the PA365
ligand. Complete binding (100%) was determined as incubaꢀ
tion binding of the PA365 ligand without any competing sialoꢀ
side.
Synthesis of BPANeu5Gc and TCCNeu5Ac sialosides
Synthesis of BPANeu5Gc (9-N-biphenylacetamide-N-
glycollyl-neuraminic acid) sialosides (Scheme S3B): The
synthesis was performed similar to that described
previously.23, 26 Briefly, a galactose terminated glycan (2a-c,
2g), 9ꢀNH2ꢀNeu5Gc (20, 1.5 equivalents/galactose) and cytiꢀ
dine 5’ꢀtriphosphate (CTP, 2.0 equivalents/galactose) were
dissolved in TrisꢀHCl (100 mM, pH9.0) containing MgCl2 (20
mM final concentration). NmCSS (~100ꢀ400 mU) and an α2ꢀ
6ꢀsialyltransferase (~100ꢀ200 mU of hST6GalꢀI41) were added
and incubated at 37 °C for overnight. The reactions were monꢀ
itored by mass spectrometry and thin layer chromatography
(TLC). After completion, the reactions were centrifuged and
the supernatant was loaded to size exclusion chromatography
(Sephadex Gꢀ25, 1 x 100 cm, eluted with water) to afford α2ꢀ6
Synthesis of fluorescent ligand conjugates. Fluorescine
isothiocyanate (FITC) was conjugated to hCD22 ligands (3a,
3b, 3j and SGP) by mixing with FITCꢀNCS under basic conꢀ
ditions (Scheme S2A). The final products were purified with
semipreparative HPLC by monitoring with UV wavelength at
215 nm and 254 nm using waterꢀacetonitrile with 0.1% TFA
as an eluent.
Cell binding and uptake of FITC-ligands
To assess Siglec mediated binding and endocytosis of the ligꢀ
ands, ligandꢀFITC conjugates (3a-, 3b-, 3j- or SGP-FITC)
were diluted to the indicated concentrations and were incubatꢀ
ed with approximately 106 WTꢀCHO, SiglecꢀCHO, or Daudi B
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 10
cells in 100 uL HBSS buffer (with 0.5% BSA or pure mouse 0 ºC. After the starting material was consumed, the product
1
2
3
4
5
6
7
8
serum) at 37 ºC for certain time. Cells were washed twice by
suspension in HBSS buffer (200 uL) and analyzed by flow
cytometry. Binding data were processed and graphed with
Prism software (GraphPad) for the curveꢀfitting and calculaꢀ
tions.
was purified by reverse phase HPLC to afford 3h- or 3j-
PEG4-SPDP in excellent yield. To reduce the thiol bond, the
SPDP derivative (3h- or 3-SPDP) was treated with 2.0 equivaꢀ
lents Tris(2ꢀcarboxyethyl)phosphine (TCEP, Acros Orgaics) in
200 ꢀL PBS at room temperature. The pH value was adjusted
to 7.0. After completion, 3h-SH and 3j-SH were obtained in
excellent yield by separation with SepꢀPak C18 column. The
freshly prepared 3h-SH and 3j-SH were mixed with 1.1 equiv.
MCꢀvcꢀPABꢀMMAF in 100ꢀ200 ꢀL PBS to provide the ligꢀ
andꢀMMAF conjugates in excellent yield. The MMAFꢀ
conjugates were purified by semiꢀpreparative RPꢀHPLC and
the purity was further determined by analytical HPLC.
Confocal fluorescence microscopy study of ligand endocy-
tosis
9
WTꢀCHO and hCD22ꢀCHO cells were grown on 18 mm
round cover slips in a 12 well plate. The cells were washed
twice with 1 mL PBS and incubated with 400 nM ligandꢀFITC
conjugate (3a-FITC or 3j-FITC) in 100 ꢁL DMEM (Dulbecꢀ
co's Modified Eagle Medium) buffer containing 0.5% BSA at
37 ºC for 3h. To detect the transferrin receptor, transferrinꢀ
AlexaFluor555 (1:100 dilution, Invitrogen catalg #: T35352)
was added concurrently with the ligandꢀFITC conjugate. After
washed with PBS, cells were fixed with 4% PFA for 20 mins
at room temperature. Then, cells were washed with PBS and
permeabilized with 0.1% Triton Xꢀ100 in PBS for 15 mins at
room temperature, followed by block with 5% normal goat
serum (NGS) in PBS for 60 mins at room temperature. Antiꢀ
hCD22ꢀAPC (1:25 dilution, Biolegend catalg #: 302510) and
antiꢀhuman CD107a (1:100 dilution, Biolegend catalg #:
328604) in 1% NGS/PBS solution were used to detect hCD22
and LAMPꢀ1 (lysosome biomarker), respectively, at 4 ºC for
overnight. After washing three times with PBS (2 mL), Strepꢀ
tavidinꢀAlexaFluor555 (1:500 dilution) and antiꢀFITCꢀ
AlexaFluor488 (1:100 dilution, Jackson Immuno Research
catalg #: 200ꢀ542ꢀ037) in 1% NGS/PBS solution were used to
detect biotin and FITC group, respectively, in dark at room
temperature for 60 min. After washing, cells were incubated
with Hoechst to stain the nuclei (0.1 ꢁg/mL in PBS) for 10
mins at room temperature. Cells were mounted with Vecꢀ
tashield antifade reagent (Vectorlabs, Hꢀ1000) and analyzed
using a Zeiss LSM710 laser scanning confocal microscope.
Images were Zꢀstacked and analyzed with Adobe Photoshop.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cytotoxicity assay for ligand-toxin conjugates
To determine the potency of the ligandꢀtoxin conjugates Daudi
cells were subjected to the toxins and cell viability was perꢀ
formed using the colorimetric 3ꢀ(4,5ꢀdimethylthiazolꢀ2ꢀyl)ꢀ
2,5ꢀdiphenyl tetrazolium bromide (MTT) assay.45 Approxiꢀ
mately 5000 Daudi cells in 50 ꢁL of complete culture medium
(RPMI 1640 buffer with 10% FBS and 1% PenStrep) were
plated in 96ꢀwell flat bottom plate (Corning, #3595) and incuꢀ
bated overnight. Serial dilutions of ligandꢀtoxin conjugates in
50 ꢁL RPMI 1640 buffer were added to each well and incuꢀ
bated for 72h at 37 ºC. Then, MTT (10 ꢁL, 5 mg/mL in PBS)
was added to each well and incubated for 3h at 37 ºC. The
blue formazan salts were dissolved in 100 ꢁL isopropanol with
0.1% NP40. The absorbance was measured at 570 nm with a
reference wavelength of 630 nm using a microplate reader
(BioTek Synergy H1). Data were collected in triplicate. Cell
viability data were processed and graphed with Prism software
(GraphPad) for the curveꢀfitting and calculations. The maxiꢀ
mum cell viability was defined as untreated cells (100%). The
complete killing was defined as background without cells.
ASSOCIATED CONTENT
Supporting Information
Preparation of hCD22 ligand-toxin conjugates.
The Supporting Information is available free of charge on the
ACS Publications website at http://pubs.acs.org.
Saporin conjugates were prepared in situ and applied in the
assay without further purification (Scheme S2B). Briefly,
hCD22 high affinity ligands (3h and 3j) were mixed, respecꢀ
tively, with 1.5 equivalents of NHSꢀPEGꢀbiotin or NHSꢀ
LCLCꢀBiotin Thermo Fisher Scientific) in H2O under basic
condition at 0 ºC to form biotinylated ligands (3h- and 3j-
Biotins). Asialo glycan LacNAcꢀbiotin conjugate (2a-Biotin)
was prepared and used as a negative control. After purification
by HPLC, the biotinylated ligand solution (100 ꢀM) was preꢀ
pared in PBS and mixed with StreptavidinꢀSaporin (20.7 ꢀM,
Advanced Targeting Systems Biotech catalg #: ITꢀ27) by 1:1
molar ratio at room temperature. The mixture was used in the
cytotoxicity assay without further purification.
Table S1. Chemo-enzymatic synthesis of high affinity ligands
for hCD22 (pdf)
Figure S1. Determination of IC50 values by bead binding as-
say; Figure S2. Confocal imaging of endocytosis of monova-
lent analogue (pdf)
Scheme S1. Chemical synthesis of 1a, 1b and 1c; Scheme S2.
Chemical synthesis of ligand-FITC and -toxin conjugates;
Scheme S3. Chemical synthesis of ligands for mCD22 and
mSn; Scheme S4. Chemo-enzymatic synthesis of triantennary
N-glycans (pdf)
Materials; General methods; Experimental section; Refer-
ences;
Spectra (pdf)
Conjugation to monomethyl auristatin F (MMAF) was accomꢀ
plished by the thiolꢀmaleimide coupling of a thiol modified
hCD22 ligand derivative (3h-SH or 3j-SH) to commercially
AUTHOR INFORMATION
available
maleimidocaproylꢀvalineꢀcitrullineꢀpꢀ
Corresponding Author
aminobenzoyloxycarbonyl (MCꢀvcꢀPAB)ꢀMMAF (Levena
Biopharma; Scheme S2C). Briefly, the 4ꢀ5 mg ligand, 3h or 3j
was stirred with 1.5 equivalents NHSꢀPEG4ꢀSPDP (Thermo
Fisher Scientific) in 100ꢀ200 ꢀL H2O under basic condition at
* jpaulson@scripps.edu.
ORCID
James C. Paulson: 0000-0003-4589-5322
ACS Paragon Plus Environment

Page 9 of 10
Journal of the American Chemical Society
(15) Chen, W. C.; Completo, G. C.; Sigal, D. S.; Crocker, P.
Wenjie Peng: 0000-0001-5093-7115
1
2
R.; Saven, A.; Paulson, J. C., Blood 2010, 115 (23), 4778ꢀ
4786.
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version
of the manuscript.
3
4
5
6
7
8
9
(16) Varki, A.; Cummings, R. D.; Aebi, M.; Packer, N. H.;
Seeberger, P. H.; Esko, J. D.; Stanley, P.; Hart, G.; Darvill, A.;
Kinoshita, T.; Prestegard, J. J.; Schnaar, R. L.; Freeze, H. H.;
Marth, J. D.; Bertozzi, C. R.; Etzler, M. E.; Frank, M.;
Vliegenthart, J. F.; Lutteke, T.; Perez, S.; Bolton, E.; Rudd, P.;
Paulson, J.; Kanehisa, M.; Toukach, P.; AokiꢀKinoshita, K. F.;
Dell, A.; Narimatsu, H.; York, W.; Taniguchi, N.; Kornfeld, S.,
Glycobiology 2015, 25 (12), 1323ꢀ4.
(17) Bull, C.; Heise, T.; Adema, G. J.; Boltje, T. J., Trends
Biochem Sci 2016, 41 (6), 519ꢀ531.
(18) O'Reilly, M. K.; Paulson, J. C., Trends Pharmacol Sci
2009, 30 (5), 240ꢀ8.
Funding Sources
This work was supported by grants R01AI050143,
P01HL107151, R01AI099141 from the NIH (to JCP).
Notes
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The authors declare no conflict of interest.
ACKNOWLEDGMENT
We thank Jennifer Pranskevich for expressing enzymes used
to assemble the target molecules and Dr. Todd L Lowary for
his help and advice on the manuscript.
(19) Angata, T.; Nycholat, C. M.; Macauley, M. S., Trends
Pharmacol Sci 2015, 36 (10), 645ꢀ60.
(20) Kelm, S.; Gerlach, J.; Brossmer, R.; Danzer, C. P.;
Nitschke, L., J Exp Med 2002, 195 (9), 1207ꢀ1213.
(21) Rillahan, C. D.; Macauley, M. S.; Schwartz, E.; He, Y.;
McBride, R.; Arlian, B. M.; Rangarajan, J.; Fokin, V. V.;
Paulson, J. C., Chem Sci 2014, 5 (6), 2398ꢀ2406.
(22) Razi, N.; Varki, A., P Natl Acad Sci USA 1998, 95 (13),
7469ꢀ7474.
(23) Collins, B. E.; Blixt, O.; Han, S.; Duong, B.; Li, H.;
Nathan, J. K.; Bovin, N.; Paulson, J. C., J Immunol 2006, 177
(5), 2994ꢀ3003.
(24) O'Reilly, M. K.; Collins, B. E.; Han, S.; Liao, L.; Rillahan,
C.; Kitov, P. I.; Bundle, D. R.; Paulson, J. C., J Am Chem Soc
2008, 130 (24), 7736ꢀ7745.
(25) Tateno, H.; Li, H. Y.; Schur, M. J.; Bovin, N.; Crocker, P.
R.; Wakarchuk, W. W.; Paulson, J. C., Mol Cell Biol 2007, 27
(16), 5699ꢀ5710.
REFERENCES
(1) Macauley, M. S.; Crocker, P. R.; Paulson, J. C., Nat Rev
Immunol 2014, 14 (10), 653ꢀ666.
(2) Nitschke, L., Glycobiology 2014, 24 (9), 807ꢀ817.
(3) Powell, L. D.; Sgroi, D.; Sjoberg, E. R.; Stamenkovic, I.;
Varki, A., J Biol Chem 1993, 268 (10), 7019ꢀ7027.
(4) Powell, L. D.; Varki, A., J Biol Chem 1994, 269 (14),
10628ꢀ10636.
(5) Powell, L. D.; Jain, R. K.; Matta, K. L.; Sabesan, S.; Varki,
A., J Biol Chem 1995, 270 (13), 7523ꢀ7532.
(6) Crocker, P. R.; Paulson, J. C.; Varki, A., Nat Rev Immunol
2007, 7 (4), 255ꢀ266.
(7) Poe, J. C.; Tedder, T. F., Trends Immunol 2012, 33 (8),
413ꢀ420.
(8) Muller, J.; Nitschke, L., Nat Rev Rheumatol 2014, 10 (7),
422ꢀ428.
(9) Polson, A. G.; Williams, M.; Gray, A. M.; Fuji, R. N.;
Poon, K. A.; McBride, J.; Raab, H.; Januario, T.; Go, M.; Lau,
J.; Yu, S. F.; Du, C.; Fuh, F.; Tan, C.; Wu, Y.; Liang, W. C.;
Prabhu, S.; Stephan, J. P.; Hongo, J. A.; Dere, R. C.; Deng, R.;
Cullen, M.; de Tute, R.; Bennett, F.; Rawstron, A.; Jack, A.;
Ebens, A., Leukemia 2010, 24 (9), 1566ꢀ1573.
(10) Piccaluga, P. P.; Arpinati, M.; Candoni, A.; Laterza, C.;
Paolini, S.; Gazzola, A.; Sabattini, E.; Visani, G.; Pileri, S. A.,
Leukemia Lymphoma 2011, 52 (2), 325ꢀ327.
(11) Polson, A. G.; CalemineꢀFenaux, J.; Chan, P.; Chang, W.;
Christensen, E.; Clark, S.; de Sauvage, F. J.; Eaton, D.; Elkins,
K.; Elliott, J. M.; Frantz, G.; Fuji, R. N.; Gray, A.; Harden, K.;
Ingle, G. S.; Kljavin, N. M.; Koeppen, H.; Nelson, C.; Prabhu,
S.; Raab, H.; Ross, S.; Stephan, J. P.; Scales, S. J.; Spencer, S.
D.; Vandlen, R.; Wranik, B.; Yu, S. F.; Zheng, B.; Ebens, A.,
Cancer Res 2009, 69 (6), 2358ꢀ2364.
(12) Yu, S. F.; Zheng, B.; Go, M.; Lau, J.; Spencer, S.; Raab,
H.; Soriano, R.; Jhunjhunwala, S.; Cohen, R.; Caruso, M.;
Polakis, P.; Flygare, J.; Polson, A. G., Clin Cancer Res 2015,
21 (14), 3298ꢀ3306.
(13) Kantarjian, H. M.; DeAngelo, D. J.; Stelljes, M.;
Martinelli, G.; Liedtke, M.; Stock, W.; Gokbuget, N.; O'Brien,
S.; Wang, K. M.; Wang, T.; Paccagnella, M. L.; Sleight, B.;
Vandendries, E.; Advani, A. S., New Engl J Med 2016, 375 (8),
740ꢀ753.
(26) Macauley, M. S.; Pfrengle, F.; Rademacher, C.; Nycholat,
C. M.; Gale, A. J.; von Drygalski, A.; Paulson, J. C., J Clin
Invest 2013, 123 (7), 3074ꢀ3083.
(27) O'Reilly, M. K.; Tian, H.; Paulson, J. C., J Immunol 2011,
186 (3), 1554ꢀ1563.
(28) Shan, D. M.; Press, O. W., J Immunol 1995, 154 (9),
4466ꢀ4475.
(29) Ingle, G. S.; Scales, S. J., Traffic 2014, 15 (3), 255ꢀ272.
(30) Schweizer, A.; Wohner, M.; Prescher, H.; Brossmer, R.;
Nitschke, L., Eur J Immunol 2012, 42 (10), 2792ꢀ2802.
(31) StꢀPierre, G.; Pal, S.; Ostergaard, M. E.; Zhou, T. Y.; Yu,
J. H.; Tanowitz, M.; Seth, P. P.; Hanessian, S., Bioorgan Med
Chem 2016, 24 (11), 2397ꢀ2409.
(32) Lee, Y. C.; Townsend, R. R.; Hardy, M. R.; Lonngren, J.;
Arnarp, J.; Haraldsson, M.; Lonn, H., J Biol Chem 1983, 258
(1), 199ꢀ202.
(33) Nycholat, C. M.; Peng, W. J.; McBride, R.; Antonopoulos,
A.; de Vries, R. P.; Polonskaya, Z.; Finn, M. G.; Dell, A.;
Haslam, S. M.; Paulson, J. C., J Am Chem Soc 2013, 135 (49),
18280ꢀ18283.
(34) Chien, W. T.; Liang, C. F.; Yu, C. C.; Lin, C. H.; Li, S. P.;
Primadona, I.; Chen, Y. J.; Mong, K. K. T.; Lin, C. C., Chem
Commun 2014, 50 (43), 5786ꢀ5789.
(35) Yamamoto, N.; Ohmori, Y.; Sakakibara, T.; Sasaki, K.;
Juneja, L. R.; Kajihara, Y., Angew Chem Int Ed Engl 2003, 42
(22), 2537ꢀ40.
(36) Nycholat, C. M.; Rademacher, C.; Kawasaki, N.; Paulson,
J. C., J Am Chem Soc 2012, 134 (38), 15696ꢀ15699.
(37) Jourdian, G. W.; Dean, L.; Roseman, S., J Biol Chem
1971, 246 (2), 430ꢀ435.
(14) Chari, R. V. J.; Miller, M. L.; Widdison, W. C., Angew
Chem Int Edit 2014, 53 (15), 3796ꢀ3827.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 10
(38) Yamashiro, D. J.; Tycko, B.; Fluss, S. R.; Maxfield, F. R.,
Cell 1984, 37 (3), 789ꢀ800.
W.; Wilson, I. A.; Paulson, J. C., Angew Chem Int Edit 2012,
51 (20), 4860ꢀ4863.
1
2
3
4
5
6
7
8
9
(39) Mukherjee, S.; Ghosh, R. N.; Maxfield, F. R., Physiol
Rev 1997, 77 (3), 759ꢀ803.
(43) Peng, W. J.; Pranskevich, J.; Nycholat, C.; Gilbert, M.;
Wakarchuk, W.; Paulson, J. C.; Razi, N., Glycobiology 2012,
22 (11), 1453ꢀ1464.
(44) Yu, H.; Huang, S.; Chokhawala, H.; Sun, M.; Zheng, H.;
Chen, X., Angew Chem Int Ed Engl 2006, 45 (24), 3938ꢀ44.
(45) Li, X.; Fang, T.; Boons, G. J., Angew Chem Int Ed Engl
2014, 53 (28), 7179ꢀ82.
(40) Stephan, J. P.; Chan, P.; Lee, C.; Nelson, C.; Elliott, J. M.;
Bechtel, C.; Raab, H.; Xie, D.; Akutagawa, J.; Baudys, J.;
Saad, O.; Prabhu, S.; Wong, W. L. T.; Vandlen, R.; Jacobson,
F.; Ebens, A., Bioconjugate Chem 2008, 19 (8), 1673ꢀ1683.
(41) Peng, W.; de Vries, R. P.; Grant, O. C.; Thompson, A. J.;
McBride, R.; Tsogtbaatar, B.; Lee, P. S.; Razi, N.; Wilson, I.
A.; Woods, R. J.; Paulson, J. C., Cell Host Microbe 2017, 21
(1), 23ꢀ34.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(42) Nycholat, C. M.; McBride, R.; Ekiert, D. C.; Xu, R.;
Rangarajan, J.; Peng, W. J.; Razi, N.; Gilbert, M.; Wakarchuk,
TOC Graphic:
ACS Paragon Plus Environment