Glycodendriproteins: A Synthetic Glycoprotein Mimic Enzyme with Branched Sugar-Display Potently Inhibits Bacterial Aggregation

Article abstract of DOI:10.1021/ja031698u

The continuing ability of bacteria to resist current antibiotic treatments highlights the need for alternative strategies for inhibiting their pathogenicity. Bacterial attachment is a major factor in infectivity and virulence. This key binding phase of ba

Full text of DOI:10.1021/ja031698u

Published on Web 03/25/2004
Glycodendriproteins: A Synthetic Glycoprotein Mimic Enzyme with Branched
Sugar-Display Potently Inhibits Bacterial Aggregation
Phillip M. Rendle,†, Andreas Seger,†,# Joao Rodrigues,†,3 Neil J. Oldham,† Richard R. Bott,‡,
J. Bryan Jones,§ Marjorie M. Cowan,*,| and Benjamin G. Davis*,†
Department of Chemistry, UniVersity of Oxford, Mansfield Road, Oxford OX1 3TA, UK, Genencor International,
925 Page Mill Road, Palo Alto, California 94304, Department of Chemistry, UniVersity of Toronto,
Ontario M5S 5H6, Canada, and Department of Microbiology, Miami UniVersity, Oxford, Ohio 45056
Received December 13, 2003; E-mail: Ben.Davis@chem.ox.ac.uk; cowanmm@muohio.edu
The continuing ability of bacteria to resist current antibiotic
treatments highlights the need for alternative strategies for inhibiting
their pathogenicity. Branched, complex carbohydrate structures on
glycoproteins on host cell surfaces provide a binding point for many
pathogens, including bacteria.1-3 Monovalent carbohydrate ligands
are only poorly recognized4 by carbohydrate-binding proteins
(lectins), and therefore small molecules are only effective in
blocking bacterial adhesive events in high, therapeutically unrealistic
concentrations (IC50 ∼mM (vide infra)). When more than one
saccharide of the right type and orientation are clustered, there is
an increase in affinity and specificity.5,6 Elegant examples3,7,8 have
shown that exploitation of lectin binding does not require natural,
multi-antennary structures as long as energetically efficient methods
for carbohydrate presentation may be found.5,9 We confirm here
that glycodendrimers7,10 can mimic the branched carbohydrates on
glycoproteins. Moreover, when attached to a protein at a prede-
termined site, a new class of glycoconjugates, “glycodendripro-
teins”, designed to mimic glycoproteins is created. Such synthetic
glycoproteins created from protein-degrading enzymes potently
reduce the binding ability of pathogenic bacteria (Figure 1).
The chemical construction of glycoproteins offers certain ad-
vantages over their isolation.11 Isolation or recombinant expression
typically produces mixtures of differently glycosylated proteins,
termed glycoforms,12 that display crucially different properties.13
Figure 1. (A) Glycodendriprotein construction. (B) Putative anti-infective
Several techniques offer sources of controllable, well-defined,
mechanism. and visual comparison of A. naeslundii-S. oralis aggregation
assays: control (C) and treated with 15 µg/mL S156C-2b (D).
homogeneous glycoproteins (pure glycoforms),14-19 yet, to date,
few20,21 have allowed the incorporation of the branched, multivalent
carbohydrates needed for high affinity.22 Glycoprotein enzymes with
one to four carbohydrate-tipped antennae were constructed using
reagents 1-4 (Figure 1A and Scheme 1) activated with sulfhydryl-
specific methanethiosulfonate (MTS). Conjugation with a single
cysteine group in the protein-degrading proteinase, subtilisin, from
Bacillus lentus (SBL, EC 3.4.21.62) created hybrid glycoproteins
capable not only of binding but also of degradation (Figure 1B).
SBL displays broad selectivity and functional similarity to regula-
tory proteinases23 and contains no natural cysteines; therefore,
reagents 1-4 react only with cysteine introduced by mutagenesis.
Figure 2. ESMS of SBL, S156C, and glycodendriproteins S156C-2a,2b,3,4.
Protease localization strategies allow selective degradation of protein
targets provided that a single24 suitable homing ligand is used.25
However, until now, this strategy was limited to monovalent ligands
and was poorly effective for lectins. One mono-antennary 115 and
four multi-antennary reagents 2-4 were constructed (Scheme 1)22
containing either conformationally flexible TREN core 6 to create
2b, 3, 4, or more rigid mesitylene core 5 to create 2a. Since our
target pathogen Actinomyces naeslundii binds to â-D-galactose
(Gal),26,27 Gal units were attached using 1-thio-Gal without the need
for carbohydrate protection. Mutagenesis of Ser156 to create SBL-
Ser156Cys (S156C) was based on previous surveys of optimal
targeting ligand positioning.25 Quantitative modification of the
introduced cysteine (Figure 1) with 1-4 was established by ESMS
(Figure 2), electrophoresis, and thiol titration. All confirmed the
identity and high purity of the resulting synthetic glycoproteins.22
These Gal-presenting synthetic glycoproteins were tested against
surface-immobilized model Gal-binding lectin, peanut agglutinin
(PNA), in enzyme-linked lectin assay (ELLA)28,22 to mimic the
† University of Oxford.
‡ Genencor International.
§ University of Toronto.
| Miami University.
Current address: Carbohydrate Chemistry Team, Industrial Research, Ltd.,
PO Box 31310, Lower Hutt, NZ.
# Current address: Johnson Matthey Catalysts, Cambridge CB4 0FP, UK.
3 Current address: Department Chemistry, York YO10 5DD, UK.
9
4750
J. AM. CHEM. SOC. 2004, 126, 4750-4751
10.1021/ja031698u CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
Scheme 1. Synthesis of Glycodendrimer Glycosylating Reagents
1-4a
its protein degrading activity, and (iii) Gal presentation. Protein
lacking carbohydrate (SBL-WT) was only moderate (IC50 ) 2.4
µM) in its inhibition and glycoprotein lacking protein-degrading
activity (S156C-2b treated with irreversible inhibitor PMSF22) was
ineffective (IC50 > 5 µM). Replacement of Gal in S156C-2b by
glucose (Glc) gave a control glycodendriprotein S156C-2b(Glc)
that showed only similar levels of inhibition to SBL-WT. The
nanomolar level of co-aggregation inhibition displayed by S156C-
2b is, to the best of our knowledge, the most potent to date.
Furthermore, the glycodendriprotein strategy allows ready carbo-
hydrate retooling for an alternative lectin or pathogen.
Acknowledgment. We thank Genencor International for funding
(B.G.D., M.M.C., J.B.J.), Dr Karl Sanford for his continued interest
in this work, and Dr. W. Bruce Turnbull for his critical reading of
this manuscript.
Supporting Information Available: Glycodendrimer reagent and
glycodendriprotein syntheses and characterization, binding assays, dose
response curves, and further results, figures, and discussion (PDF). This
material is available free of charge via the Internet at http://pubs.acs.org.
References
(1) Beuth, J.; Stoffel, B.; Pulverer, G. AdV. Exp. Med. Biol. 1996, 408, 51.
(2) Lingwood, C. A. Curr. Biol. 1999, 2, 695.
(3) Mulvey, G.; Kitov, P. I.; Marcato, P.; Bundle, D. R.; Armstrong, G. D.
Biochimie 2001, 83, 841.
(4) Lee, Y. C.; Lee, R. T. Acc. Chem. Res. 1995, 28, 321.
(5) Lee, Y. C.; Towsend, R. R.; Hardy, M. R.; Lo¨nngren, J.; Arnarp, J.;
Haraldsson, M.; Lo¨nn, H. J. Biol. Chem. 1983, 258, 199.
(6) Lundquist, J. J.; Toone, E. J. Chem. ReV. 2002, 102, 555.
(7) Lindhorst, T. K. Top. Curr. Chem. 2002, 218, 217.
(8) Kitov, P. I.; Sadowska, J. M.; Mulvey, G.; Armstrong, G. D.; Ling, H.;
Pannu, N. S.; Read, R. J.; Bundle, D. R. Nature 2000, 403, 669.
(9) Kitov, P. I.; Shimizu, H.; Homans, S. W.; Bundle, D. R. J. Am. Chem.
Soc. 2003, 125, 3284.
a
Reagents and conditions: A: a) Gal(Ac)5, BF3‚Et2O, DCM, 76%; b)
MeONa, MeOH, 83%; c) NaSSO2CH3, DMF, 78%; see ref 15. B: d)
NaSSO2CH3, DMF, 56%; e) 2 equiv Gal-S-Na+, DMF, 0 °C 55%. C: f)
Boc2O, DCM, -78 °C, 68%; g) (ClCH2CO)2O, DCM, 97%; h) 2 equiv
Gal-S-Na+, DMF, 88%; i) CF3COOH, DCM, 91%; j) thiobutyrolactone,
dithiothreitol, NaHCO3, water, EtOH, 69%; k) NHS-butyl-MTS, DMF, 87%
over 2 steps from 8; l) 1 equiv Gal-S-Na+, DMF, 48%; m) 9, DMF, 78%;
n) CF3COOH, DCM, 94%; o) NHS-butyl-MTS, DMF, 77%; p) NHS-butyl-
MTS, DMF, 87%; q) 9, DMF, 81%; r) CF3COOH, DCM; s) NHS-butyl-
MTS, DMF, 67% over 2 steps from 12.
(10) Roy, R.; Zanini, D.; Meunier, S. J.; Romanowska, A. J. Chem. Soc., Chem.
Commun. 1993, 1869.
(11) Davis, B. G. Chem. ReV. 2002, 102, 579.
(12) Rademacher, T. W.; Parekh, R. B.; Dwek, R. A. Annu. ReV. Biochem.
1988, 57, 785.
(13) Rudd, P. M.; Joao, H. C.; Coghill, E.; Fiten, P.; Saunders: M. R.;
Opdenakker, G.; Dwek, R. A. Biochemistry 1994, 33, 17.
(14) Witte, K.; Sears, P.; Martin, R.; Wong, C.-H. J. Am. Chem. Soc. 1997,
119, 2114.
(15) Davis, B. G.; Lloyd, R. C.; Jones, J. B. J. Org. Chem. 1998, 63, 9614.
(16) Shin, Y.; Winans, K. A.; Backes, B. J.; Kent, S. B. H.; Ellman, J. A.;
Bertozzi, C. R. J. Am. Chem. Soc. 1999, 121, 11684.
(17) Macmillan: D.; Bill, R. M.; Sage, K. A.; Fern, D.; Flitsch, S. L. Chem.
Biol. 2001, 8, 125.
display of surface lectins on bacterial surfaces. This revealed
increasing affinity with increasing Gal-antennae (KD for S156C-1,
2a, 2b, 3, 4 ≈ 1.8 × 10-3, 1.1 × 10-3, 1.5 × 10-3, 3.4 × 10-4, 1.4
× 10-7 M, respectively). Only a low-level (KD ≈ 10-2 M) binding
of S156C-4 to the control mannose-binding lectin from snowdrop29
was observed.
(18) Liu, H. T.; Wang, L.; Brock, A.; Wong, C.-H.; Schultz, P. G. J. Am.
Chem. Soc. 2003, 125, 1702.
(19) Watt, G. M.; Lund, J.; Levens, M.; Kolli, V. S. K.; Jefferis, R.; Boons,
G-J. Chem. Biol. 2003, 10, 807.
(20) Davis, B. G. Chem. Commun. 2001, 351.
(21) Nagahori, N.; Lee, R. T.; Nishimura, S.-I.; Page, D.; Roy, R.; Lee, Y. C.
ChemBioChem 2002, 3, 836.
(22) For further discussion, see SI.
With this model Gal-binding lectin interaction established, we
next evaluated the ability to inhibit the function of a Gal-binding
pathogen. Gram-positive A. naeslundii aggressively colonizes oral
cavities,27 surgical prostheses,30 and internal cavities.31 We tested
the ability of our glycoproteins to inhibit the co-aggregation of A.
naeslundii with co-pathogen Streptococcus oralis (Figure 1C,D).22
A. naeslundii uses the pili fimA adhesin, which binds Gal-tipped
structures,26 to bind structures on the surface of S. oralis during
this crucial phase in the colonization process.25,32 The inhibitory
potency of S156C-1-4, designed to degrade the A. naeslundii Gal-
binding adhesin using 1-4 as homing ligands, depended strongly
on carbohydrate structure.25 Optimal inhibitory potency was
observed for S156C-2b that has a bi-antennary carbohydrate
structure.33 The limited inhibitory potency of S156C-2a demon-
strated that bi-antennary display alone is not enough. Dose
response22 revealed that S156C-2b is a nanomolar inhibitor of A.
naeslundii co-aggregation: IC50 ) 20 nM, >106 more potent than
small-molecule adhesin inhibitor lactose (IC50 ) 33 mM). Three
key properties of hybrid glycoprotein S156C-2b are essential for
optimal inhibition: (i) its multi-antennary carbohydrate display, (ii)
(23) Rawlings, N. D.; O’Brien E.; Barrett, A. J. Nucleic Acids Res. 2002, 30,
343.
(24) Multiple, indiscriminate ligands (e.g. ref 21) fail.
(25) Davis, B. G.; Sala, R. F.; Hodgson, D. R. W.; Ullman, A.; Khumtaveeporn,
K.; Estell, D. A.; Sanford, K.; Bott, R. R.; Jones, J. B. ChemBioChem
2003, 4, 533.
(26) Hallberg, K.; Holm, C.; O¨ hman, U.; Stro¨mberg, N. Infect. Immun. 1998,
66, 4403.
(27) Ruhl, S.; Cisar, J. O.; Sandberg, A. L. Infect. Immun. 2000, 68, 6346.
(28) Page, D.; Roy, R. Bioconjugate Chem. 1997, 8, 714.
(29) Shibuya, N.; Goldstein, I. J.; Van Damme, E. J. M.; Peumans, W. J. J.
Biol. Chem. 1988, 263, 728.
(30) Wust, J.; Steiger, U.; Vuong, H.; Zbinden, R. J. Clin. Microbiol. 2000,
38, 929.
(31) Ormsby, A. H.; Bauer, T. W.; Hall, G. S. Pathology 1998, 30, 65.
(32) Ochiai, K.; Kurita-Ochiai, T.; Kamino, Y.; T., I. J. Med. Microbiol. 1993,
39, 183.
(33) A. naeslundii fimA may be functionally active as a dimer,26 PNA as a
tetramer. Bi-antennary complex glycoproteins are postulated to be natural
fimA ligands.27
JA031698U
9
J. AM. CHEM. SOC. VOL. 126, NO. 15, 2004 4751