Rewriting the bacterial glycocalyx via Suzuki-Miyaura cross-coupling

Article abstract of DOI:10.1039/c3cc38824g

Suzuki-Miyaura cross-coupling has been used to couple novel carbohydrate-based boronic acids, site-selectively, to the surface of E. coli at an unnatural amino acid. In this way, benign metal-catalyzed cellular switching allowed modulation of interactions with biomolecular partners via prokaryotic O-glycosylation mimics. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3cc38824g

Chemical Communications
www.rsc.org/chemcomm
Volume 49 | Number 27 | 7 April 2013 | Pages 2721–2836
ISSN 1359-7345
COMMUNICATION
Christopher D. Spicer and Benjamin G. Davis
Rewriting the bacterial glycocalyx via Suzuki–Miyaura cross-coupling
1359-7345(2013)49:27;1-3

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Rewriting the bacterial glycocalyx via Suzuki–Miyaura
cross-coupling†
Cite this: Chem. Commun., 2013,
49, 2747
Christopher D. Spicer and Benjamin G. Davis*
Received 9th December 2012,
Accepted 8th January 2013
DOI: 10.1039/c3cc38824g
www.rsc.org/chemcomm
Suzuki–Miyaura cross-coupling has been used to couple novel we demonstrated biological compatibility, as well as low asso-
carbohydrate-based boronic acids, site-selectively, to the surface ciated toxicity. In addition, a critical Pd threshold effect was
of E. coli at an unnatural amino acid. In this way, benign metal- observed, leading to a ‘switch-like’ dose response.
catalyzed cellular switching allowed modulation of interactions
Herein, we describe the synthesis of novel carbohydrate
with biomolecular partners via prokaryotic O-glycosylation mimics. boronic acids and their cross-coupling to the surface of
E. coli. This demonstrates the covalent conjugation of biomolecules
Until recently protein glycosylation was thought to be a primarily to cell surfaces and the potential applicability of this system in
eukaryotic phenomena.1 However, since the discovery of the first the elucidation of prokaryotic glycobiology, through the selective
prokaryotic glycoprotein in the archaebacteria Halobacterium modulation of cell surface binding partner interactions.
salinarium,2 and the first description of prokaryotic N-glycosyla-
We envisaged the synthesis of both aryl 2 (as mimics of Tyr
tion in Campylobacter jejuni,3 several cases have been described.4 O-glycosylation)19 and vinyl 10 (as mimics of Thr/Ser O-glyco-
Indeed, whilst the full extent can only now be estimated, it is sylation)20 boronic acids. D-Mannose (Man) and D-galactose
clear that prokaryotic protein glycosylation may be common, (Gal) were chosen as suitable glycans as the most common
particularly on bacterial cell surfaces.5
hexose motifs found in natural glycoproteins,21 whilst D-glucose
Over the past decade much research has been undertaken to (Glc) was also chosen due to its high natural abundance and
uncover the mechanisms and effects of glycosylation, and the widespread occurrence in biopolymers.22
selectivities towards interacting partners. Yet new tools are still
To synthesise the D-galactoside aryl boronic acid, 2Gal, the
required to allow further study and analysis.6 Of particular appeal corresponding b-benzyl bromide, 3, was first synthesized via
are new techniques for installing natural analogues or mimics of BF3-promoted glycosylation (Scheme 1a). Subsequent, Miyaura
such glycosyl modifications in a site-specific manner, into relevant borylation23 with bis(pinacolato)diboron gave the corresponding
biomolecules and complex biological contexts.7–9
boronic pinacol ester, 4. Removal of the ester was achieved through
We have recently developed a water soluble, phosphine-free treatment with diethanolamine followed by acidic hydrolysis.
palladium catalyst (Pd(OAc)2(ADHP)2, 1, see Fig. 1)10 for protein Global Zemplen deprotection with NaOMe yielded free aryl
´
modification via Suzuki–Miyaura cross-couplings11 at genetically sugar boronic acid 2Gal.
incorporated unnatural amino acids.12 This system and related
The vinyl boronic acids were synthesised via two convergent
variants have subsequently shown utility in a variety of protein methods (Scheme 1b). In our initial approach (Route 1), hydro-
systems.13,14 One such amino acid, the aryl halide p-iodophenyl- boration of protected propargyl alcohol 5 and subsequent
alanine (pIPhe), can be incorporated into both bacterial15 and acidic deprotection gave the corresponding vinyl boronic ester
eukaryotic16 systems site-selectively via amber-stop codon sup- alcohol, 6, in good yields. Glycosylation of this acceptor, 6,
pression.17 We recently reported the use of this method to using the appropriate per-acetylated sugars as glycosyl donors
incorporate a pIPhe ‘tag’ onto cell surfaces and demonstrated was subsequently undertaken in the presence of a Lewis acidic
the applicability of our catalyst for the cell surface Suzuki–Miyaura promoter.10 Whilst the use of BF3Áetherate proved sufficient for
fluorescent labelling of ‘tagged’ cells.18 In the course of this work, glycosylation using D-glucose and D-galactose donors in mod-
erate yields (9Glc and 9Gal), the use of TMSOTf was required to
Chemistry Research Laboratory, University of Oxford, 12 Mansfield Road, Oxford,
access the corresponding D-mannoside, 9Man. Whilst complete
consumption of the glycosyl donors was observed in all cases,
overall reduced yield was attributed to hydrodeboronation of
OX1 3TA, UK. E-mail: Ben.Davis@chem.ox.ac.uk; Fax: +44 (0)1865 275674;
Tel: +44 (0)1865 275652
† Electronic supplementary information (ESI) available: Full experimental details,
the product under the acidic reaction conditions, giving corres-
ponding deborylated vinyl sugars as unwanted side products.24
compound characterisation and fluorescence microscopy images. See DOI:
10.1039/c3cc38824g
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 2747--2749 2747

View Article Online
Communication
ChemComm
Fig. 1 Proposed procedure for cell surface carbohydrate modification of E. coli.
Scheme 1 Synthesis of sugar functionalised boronic acids 2Gal, 10Glc, 10Gal and 10Man.
As a result, we investigated alternative access (Route 2) to 9 via were then labelled at an OD600 = 0.2 using a palladium
hydroboration of the corresponding propargyl derivatives, 8, concentration of 0.5 mM and a boronic acid concentration of
themselves synthesised via previously reported boron trifluoride 3.5 mM. After labelling at 37 1C for 1 h, cells were collected by
promoted glycosylations.25 A number of hydroboration conditions centrifugation and washed extensively to ensure the complete
were investigated with little success: use of an excess of removal of unreacted boronic acid. Disappointingly, the Gal-tyrosinyl
catecholborane26 under solvent-less conditions failed to yield mimic reagent 2Gal proved to be poorly soluble in water under
desired product, nor use of dibromoborane,27 dicyclohexylborane28 conditions amenable to biological systems and so could not be used
or diisopinocampheylborane.29 However, the use of pinacolborane for cellular labelling. The glycosyl-serinyl mimic reagents 10Glc,
in the presence of catalytic amounts of Schwartz’s reagent 10Gal and 10Man, however, were all highly soluble under relevant
(Cp2ZrHCl) and triethylamine30 provided the protected boronic conditions and proved to be highly effective coupling partners.
esters, 9, albeit in modest yields. Deprotection of the boronic esters
Having thus synthetically altered the composition of the cell
was again achieved with diethanolamine, followed by deprotection surface glycocalyx through C–C coupling, we investigated the
with sodium methoxide to yield the desired sugars, 10, as crystal- modulation of its interactions with glycan-selective binding
line solids (Scheme 1).
protein partners. Lens culinaris agglutinin (LCA)32 and Griffonia
With these boronic acids (2Gal, 10Glc, 10Gal and 10Man) in simplicifolia lectin I (GSL)33 were chosen as a-Man- and b-Gal-
hand, Suzuki–Miyaura cross-couplings were undertaken on selective lectins, respectively, whilst Concanavalin A (ConA)
pIPhe ‘tagged’ E. coli.18 Briefly, E. coli JW2203-1 (an OmpC was used as a positive control due to its known affinity for
knockout strain from the Keio collection31) was co-transformed the existing glycan lipo-polysaccharide (LPS) of E. coli outer
with plasmids pEVOL(IPhe)17 and pOmpC-(Y232Á), to allow membranes.34 Gratifyingly, when cell surface glycans were
amber-suppressed incorporation of pIPhe into a mutant OmpC ‘switched’ to alternatives via Suzuki–Miyaura cell-surface
membrane protein at position Y232 (see ESI†). After overnight modification, a high specificity of recognition was observed.
induction of protein expression with 1 mM IPTG, 0.02% L-arabinose Modulation of the cell surface with Gal (from 10Gal) resulted in
and 2 mM pIPhe, cells were collected by centrifugation and strong interaction with GSL (Fig. 2 middle), whilst synthetic
washed extensively with pH 8.0 phosphate buffer. The cells ‘switching’ to Man (from 10Man) resulted in a high interaction
c
2748 Chem. Commun., 2013, 49, 2747--2749
This journal is The Royal Society of Chemistry 2013

View Article Online
ChemComm
Communication
We would like to thank UCB and the BBSRC for funding and
Drs R. Alexander and J. Porter for helpful discussions. BGD is a
Royal Society Wolfson Research Merit Award recipient.
Notes and references
1 R. K. Upreti, M. Kumar and V. Shankar, Proteomics, 2003, 3, 363–379.
2 M. F. Mescher, J. L. Strominger and S. W. Watson, J. Bacteriol., 1974,
120, 945–954.
3 M. Wacker, D. Linton, P. G. Hitchen, M. Nita-Lazar, S. M. Haslam,
S. J. North, M. Panico, H. R. Morris, A. Dell, B. W. Wren and M. Aebi,
Science, 2002, 298, 1790–1793.
4 N. M. Young, J.-R. Brisson, J. Kelly, D. C. Watson, L. Tessier,
P. H. Lanthier, H. C. Jarrell, N. Cadotte, F. St. Michael, E. Aberg
and C. M. Szymanski, J. Biol. Chem., 2002, 277, 42530–42539.
5 C. M. Szymanski and B. W. Wren, Nat. Rev. Microbiol., 2005, 3,
225–237.
6 E. Weerapana and B. Imperiali, Glycobiology, 2006, 16, 91R–101R.
7 B. G. Davis, Chem. Rev., 2002, 102, 579–602.
8 D. P. Gamblin, E. M. Scanlan and B. G. Davis, Chem. Rev., 2009, 109,
131–163.
9 J. M. Chalker, G. J. L. Bernardes and B. G. Davis, Acc. Chem. Res.,
2011, 44, 730–741.
10 J. M. Chalker, C. S. C. Wood and B. G. Davis, J. Am. Chem. Soc., 2009,
131, 16346–16347.
11 N. Miyaura and A. Suzuki, Chem. Rev., 1995, 95, 2457–2483.
12 C. D. Spicer and B. G. Davis, Chem. Commun., 2011, 47, 1698–1700.
13 N. Li, R. K. V. Lim, S. Edwardraja and Q. Lin, J. Am. Chem. Soc., 2011,
133, 15316–15319.
14 Y.-S. Wang, W. K. Russell, Z. Wang, W. Wan, L. E. Dodd, P.-J. Pai,
D. H. Russell and W. R. Liu, Mol. BioSyst., 2011, 7, 714–717.
15 J. Xie, L. Wang, N. Wu, A. Brock, G. Spraggon and P. G. Schultz, Nat.
Biotechnol., 2004, 22, 1297–1301.
Fig. 2 Interaction of fluorescein–lectin conjugates with E. coli labelled with
monosaccharide boronic acids via Suzuki–Miyaura coupling.
16 J. W. Chin, T. A. Cropp, J. C. Anderson, M. Mukherji, Z. Zhang and
P. G. Schultz, Science, 2003, 301, 964–967.
17 T. S. Young, I. Ahmad, J. A. Yin and P. G. Schultz, J. Mol. Biol., 2010,
395, 361–374.
specificity for LCA (Fig. 2 top). Glc modification of the cell
surfaces resulted in no interaction with either of these proteins.
Experiments run in the absence of the key coupling partners 18 C. D. Spicer, T. Triemer and B. G. Davis, J. Am. Chem. Soc., 2012, 134,
800–803.
(palladium, boronic acid or in cells grown in the absence of
pIPhe) also failed to result in any binding, strongly supporting a
19 I. R. Rodriguez and W. J. Whelan, Biochem. Biophys. Res. Commun.,
1985, 132, 829–836.
Suzuki–Miyaura-induced switching mechanism (see ESI†). Nota- 20 B. C. O’Connell and L. A. Tabak, J. Dent. Res., 1993, 72, 1554–1558.
21 A. Varki, R. D. Cummings, J. D. Esko, H. H. Freeze, P. Stanley,
bly, the existing, natural E. coli glycocalyx was not disrupted during
the course of the reaction, allowing the inherent interaction with
C. R. Bertozzi, G. W. Hart and M. E. Etzler, Essentials of Glycobiology,
Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2nd edn,
ConA to be maintained in all cases (Fig. 2 bottom). In addition to
confirming Suzuki–Miyaura glycoconjugations at bacterial cell
surfaces, these results also demonstrated the accessibility and
potential applicability of these biologically relevant ligands for
future applications.
2009.
22 W. S. Hu and A.-P. Zeng, Genomics and Systems Biology of Mammalian
Cell Culture, Springer, 2012.
23 T. Ishiyama, M. Murata and N. Miyaura, J. Org. Chem., 1995, 60,
7508–7510.
24 H. C. Brown, D. Basavaiah, S. U. Kulkarni, H. D. Lee, E. Negishi and
J. J. Katz, J. Org. Chem., 1986, 51, 5270–5276.
25 H. B. Mereyala and S. R. Gurrala, Carbohydr. Res., 1998, 307, 351–354.
In conclusion, we have synthesised a series of novel carbo-
hydrate based boronic acids, and applied these as cross- 26 H. C. Brown and S. K. Gupta, J. Am. Chem. Soc., 1972, 94, 4370–4371.
27 H. C. Brown and J. B. Campbell, J. Org. Chem., 1980, 45, 389–395.
28 H. C. Brown, A. K. Mandal and S. U. Kulkarni, J. Org. Chem., 1977,
coupling partners for the Suzuki–Miyaura labelling of pIPhe
‘tagged’ cells. The cognate interactions of these biologically
42, 1392–1398.
relevant ligands could be visualised on the cell surface via the 29 P. Martinez-Fresneda and M. Vaultier, Tetrahedron Lett., 1989, 30,
2929–2932.
selective binding of fluorescein–lectin conjugates, thus demon-
strating the functionality of such ligands in a complex and
30 Y. D. Wang, G. Kimball, A. S. Prashad and Y. Wang, Tetrahedron
Lett., 2005, 46, 8777–8780.
relevant biological context. Importantly, we have also demon- 31 T. Baba, T. Ara, M. Hasegawa, Y. Takai, Y. Okumura, M. Baba,
strated18 low associated toxicity. The use of such couplings
further demonstrates the power of Pd-mediated control of
K. A. Datsenko, M. Tomita, B. L. Wanner and H. Mori, Mol. Syst.
Biol., 2006, 2, 2006.0008.
32 A. Foriers, E. Lebrun, R. Van Rapenbusch, R. de Neve and
A. D. Strosberg, J. Biol. Chem., 1981, 256, 5550–5560.
33 L. A. Murphy and I. J. Goldstein, J. Biol. Chem., 1977, 252,
4739–4742.
34 T. G. Pistole, Annu. Rev. Microbiol., 1981, 35, 85–112.
35 J. Li and P. R. Chen, ChemBioChem, 2012, 13, 1728–1731.
Biology,35 here in cellular interactions. We are currently work-
ing towards utilising such methods in eukaryotic systems,
particularly in the potential blood typing (‘blood groups’) of
mammalian samples.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 2747--2749 2749