Synthesis of fluorous tags for incorporation of reducing sugars into a quantitative microarray platform

Article abstract of DOI:10.1021/ol702915e

Carbohydrate microarrays can map out key interactions of carbohydrates and proteins in a high-throughput manner, but require the inclusion of a range of sugars for their optimal use. Here we present the synthesis and use of a new hydroxylamine-modified fluorous tag that allows the facile incorporation of reducing sugars into a noncovalent fluorous-based microarray after simple purification by fluorous solid-phase extraction (FSPE). The microarray supports quantitative screening against carbohydrate-binding proteins.

Full text of DOI:10.1021/ol702915e

ORGANIC
LETTERS
2
008
Vol. 10, No. 5
85-788
Synthesis of Fluorous Tags for
Incorporation of Reducing Sugars into a
Quantitative Microarray Platform
7
Guo-Song Chen and Nicola L. Pohl*
Department of Chemistry and the Plant Sciences Institute, Gilman Hall, Iowa State
UniVersity, Ames, Iowa 50011-3111
npohl@iastate.edu
Received December 4, 2007
ABSTRACT
Carbohydrate microarrays can map out key interactions of carbohydrates and proteins in a high-throughput manner, but require the inclusion
of a range of sugars for their optimal use. Here we present the synthesis and use of a new hydroxylamine-modified fluorous tag that allows
the facile incorporation of reducing sugars into a noncovalent fluorous-based microarray after simple purification by fluorous solid-phase
extraction (FSPE). The microarray supports quantitative screening against carbohydrate-binding proteins.
Carbohydrate-protein interactions are critical in many
important biological events, including cell-cell communica-
tion, pathogen invasion, and inflammatory responses. Plot-
for attachment of sugars to the solid support and are not
readily amenable to the quick fabrication of small arrays to
ask targeted questions of a small range of isolated carbohy-
drates by nonexperts. Recently we demonstrated a simple
method to synthesize oligosaccharides using fluorocarbon
tags that also allows the direct formation of microarrays by
1
ting these carbohydrate-protein interactions using a minimal
amount of material has been facilitated to a large extent by
the development of carbohydrate microarrays. Multiple
sugars can be screened in parallel in these chip-based
approaches. The successful fabrication of carbohydrate
microarrays depends on efficient and facile immobilization
methods for carbohydrate probes on a proper surface. The
orientation and glycosidic stereochemistry of the incorporated
sugars must be consistent for reproducible results. A range
of methods to immobilize sugars for preparing carbohydrate
3
noncovalent fluorous-fluorous interactions. More recently,
this microarray system has been applied by other groups to
the fabrication of small molecule arrays for the discovery
4
of histone deacetylase inhibitors and protein-ligand binding
5
interactions. Our carbohydrate microarray method does not
require unique reactive functional handles and multiple
2
microarrays have been described to date. Unfortunately,
(3) (a) Ko, K.-S.; Jaipuri, F. A.; Pohl, N. L. J. Am. Chem. Soc. 2005,
27, 13162. (b) Mamidyala, S. K.; Ko, K.-S.; Jaipuri, F. A.; Park, G.; Pohl,
1
these methods require optimization of the reaction chemistry
N. L. J. Fluorine Chem. 2006, 127, 571. (c) Pohl, N. L. Carbohydrate
Microarrays and Fluorous-Phase Synthesis: Interfacing Fluorous-Phase Tags
with the Direct Formation of Glycoarrays. In Current Fluoroorganic
Chemistry: New Synthetic Directions, Technologies, Materials, and Biologi-
cal Applications; Soloshonok, V. A., Mikami, K., Yamazaki, T., Welch, J.
T., Honek, J. F., Eds.; ACS Symp. Ser. No. 949; American Chemical
Society: Washington, DC, 2007; p 261.
(1) (a) Bertozzi, C. R.; Kiessling, L. L. Science 2001, 291, 2357. (b)
Roth, J. Chem. ReV. 2002, 102, 285. (c) Ritchie, G. E.; Moffatt, B. E.;
Sim, R. B.; Morgan, B. P.; Dwek, R. A.; Rudd, P. M. Chem. ReV. 2002,
1
02, 305. (d) Zachara, N. E.; Hart, G. W. Chem. ReV. 2002, 102, 431.
(
2) For recent reviews on carbohydrate microarrays see: (a) Horlacher,
T.; Seeberger, P. H. OMICS 2006, 10, 490-498. (b) Culf, A. S.; Cuperlovic-
Culf, M.; Oullette, R. J. OMICS 2006, 10, 289-310. (c) Larsen, K.;
Thygesen, M. B.; Guillaumie, F.; Willats, W. G.; Jensen, K. J. Carbohydr.
Res. 2006, 341, 1209-1234. (d) Paulson, J. C.; Blixt, O.; Collins, B. E.
Nat. Chem. Biol. 2006, 2, 238-248.
(4) Vegas, A. J.; Bradner, J. E.; Tang, W.; McPherson, O. M.; Greenberg,
E. F.; Koehler, A. N.; Schreiber, S. L. Angew. Chem., Int. Ed. 2007, 46,
1-6.
(5) Nicholson, R. L.; Ladlow, M. L.; Spring, D. R. Chem. Commun.
2007, 3906-3908.
1
0.1021/ol702915e CCC: $40.75
© 2008 American Chemical Society
Published on Web 02/05/2008

chemical steps required by covalent immobilization methods.
The introduction of fluorous tags also can substantially
simplify the purification of synthetic carbohydrate intermedi-
ates. Although isolation of homogeneous sugars from natural
sources is challenging, some reducing sugars are available
and ideally also could be interfaced easily with this carbo-
hydrate array platform. Herein we report the design and
synthesis of a fluorous tag for reaction with unprotected
reducing sugars and incorporation of the tagged sugars in a
fluorous microarray platform as the first step in producing a
carbohydrate microarray kit for incorporation of natural
sugars. We also demonstrate that the fluorous array platform
with this new tag can support quantitative binding measure-
ments.
Scheme 1. Synthesis of Fluorous Tag
reaction conditions to produce a small library of sugars
including polysaccharides and N-acetylamino sugars com-
monly found as ligands for plant lectins (Figure 1). Of
A fluorous tag for incorporation of reducing sugars into
our microarray platform would require addition of a func-
tional group that can easily react with unmodified sugars at
the reducing end and produce primarily the chain-closed
version of the sugar. The conjugation method must be under
mild reaction conditions and be amenable to parallel
synthesis. Reactions of aminooxy or hydrazide groups with
free carbohydrates are known to be chemoselective and have
been widely used for the synthesis of various glycoconju-
6
gates. Because formation of the hydrazide resulted in a
cyclized product (see the Supporting Information), we next
tried to build a fluorous tag with an aminooxy reactive group.
Previously, a peptide bearing an aminooxy function was
reacted with the anomeric center of a free aldose in a
7
chemoselective way without using protecting groups and
neoglycolipid probes have been made by oxime ligations for
8
microarrays. After we began our studies, the functional
9
group has also been reacted with a range of reducing sugars.
2
Although reaction of a primary-O-NH group produces
mostly open-chain sugar oximes, the cyclic form of the sugar
is restored when a peptide with a secondary-hydroxylamino
7
group (R-O-NH-R′) is used. Preparation of the desired
aminooxy-functionalized fluorous tag commenced with
protection of commercially available N-methylhydroxylamine
1
0
hydrochloride to form the known N-tert-butylcarbamate 2
followed by reaction of this protected hydroxylamine with
-(perfluorooctyl)propanyl methyl sulfonate3b to provide 3
Scheme 1). Removal of the Boc protecting group with
3
(
trifluoroacetic acid gave the desired fluorous tag 4.
With the N,O-substituted hydroxylamine modified fluorous
tag 4 in hand, we next experimented with the glycosylation
Figure 1. Fluorous-tagged sugars synthesized as targets for
carbohydrate microarray production.
(6) (a) Leteux, C.; Childs, R. A.; Chai, W.; Stoll, M. S.; Kogelberg, H.;
Feizi, T. Glycobiology 1998, 8, 227. (b) Shinohara, Y.; Sota, H.; Gotoh,
M.; Hasebe, M.; Tosu, M.; Nakao, J.; Hasegawa, Y. Anal. Chem. 1996,
particular interest was testing the amount of cyclized versus
open-chain form of monosaccharides formed after tagging
as only the closed-chain pyranose forms will bind to a range
of lectins. We also wished to find conditions to separate
anomers if mixtures were formed in order to test pure
compounds in our arrays and delineate if this separation was
actually always necessary. The reducing sugars glucose,
maltose, lactose, galactose, and mannose were mixed with
the fluorous tag 4 in a polar organic solvent mixture (DMF/
6
8, 2573. (c) Cervigni, S. E.; Dumy, P.; Mutter, M. Angew. Chem., Int.
Ed. 1996, 35, 1230.
7) (a) Peri, F.; Jim e´ nez-Barbero, J.; Garc ´ı a-Aparicio, V.; Tvaro sˇ ka, I.;
(
Nicotra, F. Chem. Eur. J. 2004, 10, 1433. (b) Peri, F.; Dumy, P.; Mutter,
M. Tetrahedron 1998, 54, 12269.
(8) (a) Liu, Y.; Feizi, T.; Campanero-Rhodes, M. A.; Childs, R. A.;
Zhang, Y.; Mulloy, B.; Evans, P. G.; Osborn, H. M. I.; Otto, D.; Crocker,
P. R.; Chai, W. Chem. Biol. 2007, 14, 847. (b) Liu, Y.; Chai, W.; Childs,
R. A.; Feizi, T.; Methods Enzymol. 2006, 415, 326.
(
9) Bohorov, O.; Andersson-Sand, H.; Hoffmann, J.; Blixt, O. Glycobi-
ology 2006, 16, 21C-27C.
10) Beshara, C. S.; Hall, A.; Jenkins, R. L.; Jones, K. L.; Jones, T. C.;
(
AcOH, v:v ) 1:1). As expected, compounds 5-7 were
Killeen, N. M.; Taylor, P. H.; Thomas, S. P.; Tomkinson, N. C. O. Org.
Lett. 2005, 7, 5729.
4
obtained exclusively in the pyranose form with a typical C
1
786
Org. Lett., Vol. 10, No. 5, 2008

Scheme 2. Addition of Fluorous Tag to Reducing Sugars with Fluorous Solid-Phase Extraction (FSPE) and Separation of Mixtures
1
conformation as ascertained by H NMR in deuterated
methanol. It was noteworthy that, based on analysis of the
J
This aqueous/organic method was successfully applied to
tag a range of sugars (Figure 1) and could be easily run on
a larger micromolar scale. These reaction conditions worked
well to quantitatively tag N-acetylglucosamine (GlcNAc) and
N-acetylgalactosamine (GalNAc) in 24 h. Only the â-pyra-
nose form of the glycosylation product of GlcNAc (13) was
detected. In contrast, a 1:1 ratio of GalNAc derivative
H1-H2 coupling constant values, the reactions proceed with
complete diastereoselectivity; only the â-pyranose forms
were observed for glucose, maltose, and lactose, which all
contain D-glucose at the reducing end.
To make this technique more amenable for a simple
microarray kit, we wanted to eliminate the need for dim-
ethylformamide, a solvent that is known to decompose upon
prolonged storage, as well as simplify the procedure for use
only with equipment found in most molecular biology
laboratories. As expected, aminoxy-functionalized fluorous
tag 4 is not very water soluble and requires an organic
cosolvent. Acetonitrile was chosen based on its ease of
removal, miscibility with water, and stability in room
temperature storage. Our glycan derivatives then were
1
anomers (14) was evident from H NMR data. For maltotet-
raose and maltohexaose, a molar ratio of sugar to fluorous
tag of 2:1 was used to affect quantitative conversion of the
fluorous tag after 36 h. In these cases also, only the
â-pyranose forms 15 and 16 were detected. Unlike the
glucose and glucosamine conjugations, the reactions with
1
mannose and galactose were more complex. The H NMR
spectrum of the D-galactose derivative revealed the presence
of a mixture of pyranose forms (R to â: 2 to 3, as calculated
from the integration ratio of the methyl group on the
hydroxylamine linker). Three forms of D-mannose derivatives
prepared in milligram scale using an aqueous/organic mixture
Scheme 2) and identified by H NMR and ESI mass
spectrometry. Briefly, representative disaccharide maltose
1
(
1
were evident in the H NMR data (ratio of 5:2:3 from less
and the fluorous tag 4 (molar ratio 1.5:1) were each
polar to more polar) and could not be readily separated by
silica gel chromatography. However, acylation of the man-
nose and galactose reaction mixtures with acetic anhydride
in pyridine resulted in easily separable acylated derivatives
(Scheme 2 and Supporting Information). Because the ano-
meric position of mannose cannot be characterized simply
by JH1-H2 coupling constants, NOESY NMR experiments
were performed for compounds 17-19 to establish the
anomeric stereochemistry. In the NOE spectrum of 19, cross-
separately dissolved in aqueous buffer at pH 2.5 in NaH
PO /H PO (0.7 mL) and acetonitrile (0.2 mL) (pH 3.0 after
addition of CH CN). The two solutions were then mixed in
2
-
4
3
4
3
a centrifuge tube and shaken at 37 °C. After 24 h, the reaction
mixture was directly loaded on a cartridge packed with silica
1
2
gel containing a perfluorooctylethylsilyl (Si(CH
bonded phase for fluorous solid-phase extraction (FSPE).
2
)
2
C
8
F
17
1
)
1
Fluorous-tagged sugars were eluted with methanol whereas
other components, including the unreacted sugar and buffer
salts, were eluted first with a fluorophobic phase. Two other
buffers were also assayed: pH 4 buffer (NaOAc/HOAc, 0.2
M) and pH 1.5 buffer (HCl/KCl, 0.2 M). After mixing with
acetonitrile, the pH values were found to be 4.1 and 1.9.
1 3 1 5
peaks between H and H as well as H and H were seen,
but no comparable cross-peaks were found in compounds
17 and 18. Proton identities were assigned by COSY NMR.
13
Standard Zempl e´ n deacylation conditions readily converted
the acylated compounds to their final forms for incorporation
into microarrays.
2 4 3 4
The NaH PO /H PO buffer provided the most reliable
results.
With the fluorous-tagged sugars in hand, we probed the
ability of the new fluorous tag to bind to a fluorous-coated
(11) (a) Curran, D. P. Separations with Fluorous Silica Gel and Related
Materials. In The Handbook of Fluorous Chemistry; Gladysz, J. A.; Horv a´ th,
I.; Curran, D. P. , Eds.; Wiley-VCH: Weinheim, Germany, 2004; pp 101.
(12) Duus, J. Ø.; Gotfredsen, C. H.; Bock, K. Chem. ReV. 2000, 100,
4589.
(b) Zhang, W.; W.; Curran, D. P. Tetrahedron 2006, 62, 11837.
(13) Zempl e´ n, G.; Pacsu, E. Ber. Dtsch. Chem. Ges. 1929, 62, 1613.
Org. Lett., Vol. 10, No. 5, 2008
787

glass slide surface and support plant lectin binding (Figure
). Differences in binding between the pure R-/â-isomers
intensities of replicate spots to give a curve (Figure 2). The
curves were analyzed as Langmuir isotherms, assuming that
the system reached equilibrium during incubation (see the
Supporting Information). KD,surf is found to be 1.2 ( 0.2 µM
for the mannose mixture and 0.8 ( 0.1 µM for 9. We also
calculated the binding constants for the sugar mixture and 9
at different concentrations, KD,surf ) 1.2 ( 0.2 µM for the
mixture at 6.25 µM, and 0.8 ( 0.1 µM for 9 at 100 µM.
These results show that the binding constants measured are
not affected by sugar concentrations in a certain range. As
expected, the KD,surf value of 9 is smaller than that of the
mixture, which contains only 50% of 9. The nonbinding
components of the mixture did not significantly affect the
binding constant under these conditions and therefore even
mixtures of sugars obtained after conjugation can be used
directly in a microarray screen after analysis of the compo-
nent ratio. Compared to other mannose-ConA surface binding
results with multivalent displays, our value is in excellent
agreement with surface plasmon resonance (SPR) data (KADS
2
6
14
)
5.6 ( 1.7 × 10 M). Interestingly, a recently reported
study on a carbohydrate microarray capped with a proprietary
protein prior to ConA binding found K
) 80.4 nM.15
Figure 2. Two scans at 488 nm and the binding curve of glass
slides arrayed with a fluorous-tagged carbohydrate mixture (a) and
pure mannose 11 (b) after incubation with FITC-labeled ConA.
The curve was obtained with use of FITC-ConA. The KD,surf values
were obtained by fitting the curve to Langmuir isotherms.
D
Clearly, because the linkers to mannose and the surfaces
differ substantially between techniques, quantitative data
should be compared only with those of comparable systems.
In conclusion, the new fluorous glycosylation method can
directly link reducing saccharides from natural sources to a
fluorous tag with retention of a cyclic structure and without
the requirement of activation of the anomeric center. The
tag can be synthesized in three linear steps from commercial
material in 84% overall yield prior to reaction with reducing
sugars and arraying on commercially available fluorous-
coated glass slides. Because of the ease of the overall
procedure, we envision this new tag as the basis of a simple
kit aimed at glycobiologists for the fabrication of small
targeted carbohydrate microarrays.
and their mixture (containing the three forms of fluorous
tagged mannose with the same ratio from reaction) were also
of interest. The fluorous-tagged carbohydrates were dissolved
in methanol/DMSO/water (2:2:1) and were spotted onto
commercially available fluoroalkylsilane-derivatized glass
6
a
slides employing a standard robot used for DNA array
formation. The resulting slides were then incubated with
FITC (fluorescein isothiocyanate)-labeled concanavalin A
(jack bean, FITC-ConA). Clearly the new tag could support
qualitative microarray experiments.
We next wished to test the feasibility of getting quantitative
binding data of both pure R-mannose 9 and the mixture of
mannose isomers 8, 9, and 10 to ConA. To determine a
suitable sugar concentration for these quantitative measure-
ments, control experiments were first carried out for different
sugar concentrations (Figure 2). The lower sugar concentra-
tions (0.1 M), especially for the mixture of mannose isomers,
proved optimal. Higher concentrations of sugars that cannot
bind to ConA (namely 9 and 10 in the mixture) then would
only crowd the surface and provide a suboptimal binding
surface for the protein.
Acknowledgment. This material is based in part upon
work supported by the National Science Foundation under
CAREER Grant No. 0349139. N.L.P. is an Alfred P. Sloan
Research Fellow. We thank the Microarray Facility in the
Center for Plant Genomics at Iowa State University.
Supporting Information Available: Experimental pro-
1
13
1
1
cedures and copies of H, C, and H- H COSY NMR
spectra for all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
OL702915E
The arrays of compound 8 and mannose mixture at 250
µM were then probed with eight concentrations of protein-
FITC labeled ConA, ranging from 1.25 µM to 100 nM. ConA
concentrations were plotted against mean fluorescence
(14) Smith, E. A.; Thomas, W. D.; Kiessling, L. L.; Corn, R. M. J. Am.
Chem. Soc. 2003, 125, 6140.
(15) Liang, P.-H.; Wang, S.-K.; Wong, C.-H. J. Am. Chem. Soc. 2007,
129, 11177.
788
Org. Lett., Vol. 10, No. 5, 2008