Communications
the biosynthesis of glycolipids and glycoproteins.[17] b-GalT
has the interesting characteristic that it recognizes N-acetyl
glucosamine as its substrate, but in the presence of lactalbu-
min it has an altered specificity and prefers glucose as the
substrate.[18] Our array presents three different terminal
carbohydrates (Gal, Glc, and GlcNAc) in the context of a
panel of disaccharides and is therefore well-suited for
assessing the extent to which the identity of the distal sugar
and the regiochemical linkage influences the activity of the
substrate. We treated disaccharides in the array with a
reaction cocktail (5 mL) that included b-GalT (20 UmLÀ1),
MnCl2 (10 mm), HEPES buffer (20 mm, pH 7.6), UDP-Gal
(0.5 mm) and in certain experiments, lactalbumin
(0.2 mgmLÀ1). The reactions were run for two hours in a
humidified chamber at 378C, stopped by rinsing the mono-
layers with water, and analyzed by SAMDI-TOF MS
techniques to give a spectrum for each spot in the array.
Prior to treatment of the array with the enzyme, the spot
that represents b-d-GlcNAc-1,4-b-d-Glc showed the expected
peaks at m/z 1353.6 and 1369.8 that correspond to the sodium
and potassium adducts (Figure 3a), respectively, of the
disulfide containing the disaccharide-terminated alkanethio-
late. After treatment with b-GalT, SAMDI-TOF MS revealed
newpeaks at m/z of 1517.1 and 1533.1, which correspond to
the expected trisaccharide product. The absence of the
original peaks at m/z of 1353.6 and 1369.8 demonstrated
that the enzymatic reaction was essentially complete. We
could estimate the approximate reaction yield by taking the
ratio of the sum of the intensities of product peaks (the
trisaccharide) and the sum of the intensities of the product
peaks and the reactant peaks (the disaccharide). We per-
formed this reaction on the full array and determined reaction
yields for each of the 24 disaccharides in the presence and
absence of lactalbumin (Figure 3b). We recognize that these
yields represent relative rather than absolute measures since
the products may have different ionization efficiencies
relative to the substrates.
Figure 3. b-GalTactivity assays on the 24 disaccharide array.
a) SAMDI-TOF MS spectrum for starting material b-d-GlcNAc-1,4-b-d-
Glc and product b-d-Gal-1,4-b-d-GlcNAc-1,4-b-d-Glc. b) Activities of
b-GalTfor each disaccharide. For each circle, the two halves represent
results from two different reaction conditions for the same disacchar-
ide. The left half circle represents activity in the presence of added
lactalbumin to the enzymatic solution and the right half circle
represents the activity in the absence of added lactalbumin.
an array of 24 disaccharides by using either Gal or Glc as the
first residue and Gal, Glc, or GlcNAc as the second residue.
By using four distinct building blocks each for Gal and Glc in
the first step—wherein the building blocks differed in the
position of the Lev group—and one building block each for
Gal, Glc, and GlcNAc in the second step, the array presented
the six disaccharides Gal–Gal, Gal–Glc, Glc–Gal, Glc–Glc,
GlcNAc–Gal, and GlcNAc–Glc having each of the four
linkages (b 1,2, b 1,3, b 1,4, and b 1,6). The reaction time for
the coupling of the carbohydrate building blocks is approx-
imately 30 minutes and the time required to remove the Lev
group is one minute. The final treatment to remove the
protecting groups and reveal the free disaccharides required
one hour; therefore the total time required to prepare the
array was about four hours. The mass-spectrometry results
showed that each disaccharide was formed as expected and no
peaks for the starting material remained, implying high yields
for the conversions.
We found that each of the eight GlcNAc-terminated
dissacharides was converted into the respective trisaccharide
in high yield. The addition of lactalbumin inhibited the
galactosylation of GlcNAc-terminated disaccharides, and
instead promoted addition of UDP-Gal to Glc-terminated
disaccharides. In the absence of lactalbumin, these same Glc-
terminated substrates were mostly inactive. We found that the
yields for galactosylation of the Glc-terminated disaccharides
(in the presence of lactalbumin) were substantially lower than
the yields for GlcNAc-terminated substrates (without lactal-
bumin), in agreement with previous reports.[17,18] A compar-
ison of the relative activities of all disaccharides revealed that
those having b-1,6 and b-1,4 linkages were more active, likely
owing to a better fit of the substrate within the enzyme active
site.[19,20] Finally, the Gal-terminated disaccharides showed no
activity, either with or without lactalbumin, which confirms
previous results that the enzyme does not use terminal
galactosides as substrates.[18]
We used the array to profile the substrate specificity of
bovine b-1,4-galactosyltransferase I (b-GalT). This enzyme
transfers galactoside from uridine diphosphogalactose (UDP-
Gal) to its substrate; it is involved in lactose synthesis and in
This work addresses the significant need for tools that can
facilitate the discovery and understanding of the many roles
that carbohydrates play in biology. At the molecular level,
these needs include the elucidation of binding preferences of
3398
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3396 –3399