conditions were similar to those previously reported.19 The
integrated areas under the extracted ion chromatogram (EIC)
were determined by Xcalibur software version 2.0.5 and used
to evaluate the electrospray response of different species.
Four 100 mL aliquots were taken from a stock standard
solution of maltoheptaose (1 mg mLÀ1) and pipetted into
4 Eppendorf tubes. Three aliquots were individually derivatized
with the two newly synthesized reagents (1 and 2) and
Girard’s T reagent (3), while one was left underivatized (native
species) and served as a control. Hydrazone formation was
completed as previously discussed by Naven and Harvey,17
with approximately a 20-fold excess of reagent. The samples
were diluted, combined in equal volumes for an equimolar
mixture and subsequently analyzed by HILIC nanoLC-MS.
Calculations for determining the non-polar surface area
(i.e., hydrophobicity) of each tag were estimated based on
bond lengths and van der Waals radii of each atom.20
The procedure for cleavage and purification of N-linked
glycans from plasma was performed as previously reported.19
For derivatization of N-linked glycans derived from plasma
glycoproteins, a 65-fold excess of reagent to internal standard
was utilized. Five mL of the glycan mixture were utilized after
lyophilization and reconstitution of the solid phase extraction
eluents.
Fig. 2 The relative responses for the analysis of an equal molar
mixture of the native and various maltoheptaose derivatives.
markers for disease are most likely at relatively low abundances
and any increase in electrospray response (i.e., lower limit of
detection) afforded by a fast and high yielding derivatization
procedure would be of considerable importance. Since
reagent 1 afforded the greatest ion abundance from the model
experiments, we chose to investigate the utility of this reagent
for derivatization of N-linked glycans derived from plasma
glycoproteins.
Fig. 2 displays an EIC of the different maltoheptaose species
from the nano HILIC-MS analysis of an equal molar mixture.
In addition, the derivatives formed as a result of various
tags are shown. The tripropyl derivatized species afforded
approximately a 5-fold and 12-fold increase in electrospray
response compared to the native species and the species
derivatized with the Girard’s T reagent, respectively. In
addition, this species was determined both theoretically and
experimentally to be the most hydrophobic, as it was calculated
to have the largest non-polar surface area and was the least
retained by HILIC (Table 1). Although the Girard’s T reagent
did impart additional non-polar surface area to the analyte, it
was experimentally discovered to be the most hydrophilic
(most retained) and yielded the lowest ion abundance. We
hypothesize that the limited amount of hydrophobic moieties
(3 methyl groups) was not sufficient to overcome the hydro-
philicity imparted by the permanent charge. To further
investigate this result, an equal molar direct infusion experiment
was performed with all four species and it was again found the
Girard’s T derivative was less abundant than the native. This
result, different than previously reported,17 may be attributed
to the addition of ammonium acetate to the solvent system
used for both LC-MS and direct infusion experiments. The
dominant ionization pathway observed was ammonium
adduction. Data summarizing the non-polar surface area
(i.e., degree of hydrophobicity) and response ratios of the
different species from both sets of experiments are summarized
in Table 1. Although significant differences existed amongst
the degrees of signal increase between the two sets of experiments,
all species followed the same general trend (i.e., the more
hydrophobic the greater the signal response).
Fig. 3 displays an EIC of a monosialylated glycan (native)
overlaid with its tripropyl derivatized counterpart. As shown,
approximately a 4.5-fold (ratio of integrated areas) increase in
electrospray response was observed for the derivatized species
when compared to the native. Similar to the model system, we
observed a decrease in retention time which indicated an
increase in hydrophobicity for the derivatized glycan.
It is possible that increasing the number of vibrational
modes and/or imparting a permanent charge on the molecule,
as others have previously theorized,15 would negatively effect
glycan fragmentation. Therefore, it was necessary to evaluate
the effect of derivatization on glycan fragmentation via collision
induced dissociation (CID). The inset in Fig. 3 compares the
fragmentation spectra of the derivatized species to that of
the native species. A high similarity exists between both
fragmentation spectra. The signal-to-noise ratio of the
tandem MS spectrum resulting from the derivatized species
is approximately 10-fold greater compared to the native
glycan. The majority of fragment peaks greater than m/z
1000 contain the reducing end of this particular glycan
(y-type) which explains the offset of approximately 198 Daltons,
the added mass of the tag, between the derivatized and native
species. Species less than m/z 1000 often correspond to b-type
ions or internal fragments and do not contain the reducing
terminus; thus, these peaks have the same m/z for both the
derivatized and native species.
This preliminary work indicated the potential of these tags
to increase glycan electrospray response; however, the ultimate
goal is the application of these reagents to glycan biomarker
discovery where samples are much more complex. Potential
In this communication we have synthesized two new
hydrazide reagents. Both reagents afford greater ESI response
of glycans as compared to Girard’s T reagent. It is clear that
ꢀc
This journal is The Royal Society of Chemistry 2010
238 | Chem. Commun., 2010, 46, 237–239
Bereman, Michael S.
