EXPERIMENTAL Derivatization

1. Oligo IRABA * Y3β 34 Y2 (-Glc) (-Gal) (-Fuc,-Gal,-GlcNAc) (-Fuc) IRMPD Y1 1265 446 1119 608 (-Hex,-HexNAc) (-Fuc) (-Hex,-Hex,-H2O) [LNFPI+IRABA-2H2O+H]+ Y2 * Y0 B Y1 284 15 412 B (-Gal,-GlcNAc) 608 (-Fuc) (-Glc) 446 (-Gal) 777 Y4 Y0 1119 923 IRMPD 11ms 284 (-Gal) Y3 0 (-HexNAc) (-Hex,-Hex,-H2O) [LNDFHIb+IRABA -2H2O+H]+ [LNFPI+IRABA-2H2O-H-E]- 1089 [LNFPI+IRABA-2H2O-H]- * B 42 Y3β Y1 Y2 B 100 (-Glc) (-Gal) (-Fuc,-Gal,-GlcNAc) (-Fuc) 1265 IRMPD 973 428 1117 Y0 446 1119 608 284 0 [LNFPII+IRABA-2H2O+H]+ * Relative Abundance (-Hex,-HexNAc) (-Fuc) (-Hex,-Hex,-H2O) [LNFPI-H]- Y2 Y1 852 B Y3β 100 1119 Y0 446 412 608 1140 B 973 777 IRMPD 11ms (-Glc) (-Gal) (-Gal,-GlcNAc) (-Fuc) 923 0 0 284 [LNDFHII+IRABA-2H2O+H]+ * (-Gal,-GlcNAc) (-Fuc) (-Gal) Y1β/Y2α B (-Glc) (-Fuc) Y1α/Y1β Y2α Y3α’’ 63 100 Relative Abundance (-HexNAc) (-Fuc) (-Hex,-Hex,-H2O) B 1119 [LNFPI+IRABA-2H2O+H]+ 1265 1141 1119 428 C4 (-Glc) IRMPD 446 876 [LNFPI+Na]+ a) CAD Y0 100 (-Fuc,-Gal,-GlcNAc) 777 [LNFPI-H]- 690 (-Hex,-HexNAc) (-Fuc,-Hex,-Hex,-H2O) 0 754 608 * 852 [LNFPIII+IRABA-2H2O+H]+ * B Y2 Relative Abundance 284 412 0 Y3β Y1 79 ] 1119 Y0 608 C3 (-Glc) (-Glc) (-Gal) (-Gal,-GlcNAc) (-Fuc) 777 b) CAD 0 52 446 973 [LNFPII-H]- IRMPD 10ms 690 0 284 * 852 850 900 950 1000 1050 1100 1150 1200 200 400 600 800 1000 1200 1400 (-HexNAc) (-Fuc) (-Hex,-Hex,-H2O) B B 0 m/z (-Glc) C3 428 c) CAD 73 [LNFPIII-H]- 690 * 852 777 LNFP-I Fucα1-2Galβ1-3GlcNAcβ1-3Galβ1-4Glc 0 0 (-Fuc,-Glc) [LNFPV+IRABA-2H2O+H]+ * C3α Y2α/Y1β Y1β d) CAD (-Glc) 446 Y1α/Y1β (-Gal) (-Gal,-GlcNAc) (-Fuc) (-Fuc,-H2O) 29 79 544 [LNFPV-H]- 608 973 Y0 1119 Relative Abundance Z1β * 852 688 (-Fuc) Y2 IRMPD 11ms 284 (-Gal,-GlcNAc) 0 C4 (-Glc) (-Hex,-Hex,-H2O) * 852 e) IRMPD 2ms 24 (-HexNAc) 690 B [LNFPI-H]- LNFP-II Galβ1-3(Fucα1-4)GlcNAcβ1-3Galβ1-4Glc B3 [LNFPI+IRABA-2H2O+H]+ 428 100 Y4α/Y3β * (-Hex,-Hex,-H2O) 973 B CAD 100 Y2 (-Gal) (-Gal,-GlcNAc) 608 1264 (-Sia) (-Sia) 0 (-Gal,-H2O) 754 Y1 Y2 (-Hex,-HexNAc) 1555 Z3α/C2 * 852 446 0 IRMPD f) IRMPD 2ms [LNFPI-H]- C2 (-Fuc,-H2O) C3 (-Glc) 608 32 (-Gal) 690 348 Y4 * 200 300 400 500 600 700 800 900 1000 1100 1200 777 Z3β/C2 528 922 364 973 m/z 0 (-Hex,-HexNAc) (-Sia,-Hex,-Hex,-H2O) 1119 981 Relative Abundance 0 g) IRMPD 3ms LNFP-III Galβ1-3(Fucα1-3)GlcNAcβ1-3Galβ1-4Glc Z3β/C2 C3 * 852 21 (-Fuc,-H2O) (-Gal) (-Glc) [LNFPII+IRABA-2H2O+H]+ (-Hex,-Hex,-H2O) (-Hex,-Hex,-H2O) 690 C2 364 85 557 [LNFPIII-H]- B2 CAD Y2 528 (-Hex,-HexNAc) (-Fuc) 631 Relative Abundance 0 (-Fuc,-Glc) 608 0 C3α (-Fuc,-H2O) * 852 h) IRMPD 3ms 16 Y3β 544 [LNFPV-H]- 200 400 600 800 1000 1200 1400 1600 1800 Z1β m/z * 777 973 688 1119 981 0 0 [LNFPIII+IRABA-2H2O+H]+ (-Hex,-Hex,-H2O) 200 300 400 500 600 700 800 900 91 B2 CAD (-Hex,-HexNAc) Y2 (-Fuc) 608 Y3β * 777 973 1119 981 0 [LNFPV+IRABA-2H2O+H]+ (-Hex,-Hex,-H2O) (-Fuc) DSLNT Neu5Acα2-3Galβ1-3(Neu5Acα2-6)GlcNAcβ1-3Galβ1-4Glc 72 CAD (-Hex,-HexNAc) B2 Y1β (-Fuc) 631 Y2α/Y1β 973 Y2α * 754 608 1119 981 0 400 500 600 700 800 900 1000 1100 1200 m/z Characterization of Oligosaccharides Using Infrared Multi-Photon Dissociation and Chemical Derivatization in a Quadrupole Ion Trap Michael Pikulski, Lisa Vasicek, Amanda Hargrove, Shagufta H. Shabbir, Eric Anslyn, Jennifer S. Brodbelt* Department of Chemistry and Biochemistry The University of Texas at Austin LNDFH Oligosaccharides This characterization strategy was expanded to the second oligosaccharide series and the same two pathways of dissociation that were observed for the LNFP series were seen again (Figure 7). The primary fragmentation pathway provides detailed sequence information for the LNDFH oligosaccharides with the exception of the two isomers that only differ in their reversed Gal and Fuc linkages to the central GlcNAc (Ia and Ib). Following the loss of terminal or branching Fuc moieties, fragmentation of the backbone for each of the isomers from the non-reducing end starts with the combined Gal-GlcNAc losses followed by two sequential Hex losses (Gal and Glc). Other possible starting sequences of these oligosaccharides are discounted because in each of the IRMPD spectra throughout this study there has been a key intersaccharide cleavage at the reducing end of GlcNAc. These fragments ions are not observed in the IRMPD mass spectra, thus discrediting these alternative sequences. Therefore the two Fuc must reside on the non-reducing end Gal and GlcNAc. The difference in the IRMPD spectrum for the LNDFH-II isomer is a result of a Fuc having a different position. Since the initial Fuc loss precedes the Gal-GlcNAc combined loss, it is evident that this Fuc resides on either the Gal or GlcNAc • OVERVIEW • A derivatization agent with high IR absorptivity, IRABA, was designed and synthesized to enhance the reactivity of the boronic acid functionality with oligosaccharides and improve the ability to sequence oligosaccharides without the need for MSn. • Two series of oligosaccharides were studied: the lacto-N-fucopentaoses (LNFPs), and the lacto-N-difucohexaoses (LNDFHs). • The IRABA-oligosaccharide products dissociated with high efficiency upon IR irradiation. • The resulting IRMPD mass spectra display a characteristic primary fragmentation pathway that simplifies the determination of oligosaccharide sequence. EXPERIMENTAL Derivatization The IRABA group was synthesized by a modification to a synthesis described by the Ansyln group (College of Chemistry and Biochemistry, University of Texas at Austin)3. First, 3.9 mmol 2-formylphenylboronic acid (mM) was dissolved in anhydrous CH3OH under argon protection. Then 16 mmol Hunig’s base was added followed by 3.9 mmol diethyl(aminomethyl)phosphonate oxalate. The solution was stirred for 16 hours before 2.9 mmol NaBH4 was added slowly, stirred an additional hour, followed by another addition of NaBH4. One hr later, the solvent was removed under vacuum and the residue was diluted with CH2Cl2. The white ppt was removed with vacuum filtration, with the filtrate subsequently concentrated. The residue was purified by flash chromatography on neutral alumina (2-5% NH3-saturated CH3OH in CH2Cl2) and at a final concentration of10 µM mixed with stock solutions of the glycans (500 µM) to make a 1:10 molar ratio along with a 5 µL aliquot of 0.5% TEA was added to a pH ~9. The solutions were then sonicated for 1 min and diluted with 0.1% TEA to pH ~8 and spiked with ammonium acetate (to 0.01%). Mass Spectrometry A 10 µM oligosaccharide solution was injected into a Finnigan LCQ Deca XP mass spectrometer equipped with an electrospray ionization source and interfaced with a Synrad CO2 50 W laser (10.6 µm). Typical IRMPD parameters included an irradiation time of 5 - 25 msec at a power of 50 W and a helium pressure of 2.8 x 10-5 Torr. For the underivatized oligosaccharides, the helium pressure was lowered to nominally 2.7 x 10-5 Torr to promote the IRMPD process. INTRODUCTION Oligosaccharides are in many ways more structurally complex than other biologically important molecules such as proteins and nucleic acids in that they are often highly branched and have several different linkage types between their fundamental monosaccharide units. The complex structures of oligosaccharides result in their characterization a challenging but vital task because they have important roles in numerous biological processes. Mass spectrometry (MS) has become an invaluable tool for the structural determination of glycans mainly due to the minimal sample consumption and specificity of the information obtained. To provide more extensive structural information and alleviate the need for elaborate MSn strategies, IRMPD1 has been implemented in quadrupole ion traps and affords a promising alternative to CAD for the characterization of oligosaccharides. In the present study, we report a simplified method for the sequencing of oligosaccharides in a single stage of activation by exploiting the supplementary information obtained from sequential fragmentation that is promoted by non-resonance IRMPD and derivatization with a boronic acid that has an incorporated phosphonate group, IRABA, to increase IR absorption efficiency. The first general type of pathway is defined by sequential losses from the non-reducing end of the oligosaccharides (losses highlighted in black type) and correspond to a loss of Fuc and a combined loss of Hex and HexNAc for the LNFPs. The persistent combined loss of Hex and HexNAc suggests that cleavage at the non-reducing end of the GlcNAc is not a facile process. The second general pathway observed in Figure 5 involves the combined loss of two Hex and W arising from cleavage at the reducing end of the oligosaccharide (losses highlighted in blue type). However, in this CAD spectra the sequence coverage of the oligosaccharides is still limited, therefore IRMPD was used to further the study. [LNDFHIa+IRABA-2H2O+H]+ IRABA Derivatized Oligosaccharides Since neither the CAD nor the IRMPD spectra of the underivatized LNFPs yielded diagnostic fragment ions or sufficient sequencing information for the isomers, the boronic acid derivatization strategy was explored next. -ESI +ESI RESULTS AND DISCUSSION LFNP Oligosaccharides CAD and IRMPD spectra of both the underivatized and derivatized species were taken to develop a method for sequencing the backbone of the LNFPs and to pinpoint the sites of attachment of the fucose moieties. The CAD mass spectra of the deprotonated oligosaccharides, ([L - H]-), are shown in Figures 2a-d. Figure 1: A scheme showing the derivatization strategy for oligosaccharides. [LNFPI+IRABA-2H2O+Na]+ The IR-active boronic acid (IRABA) was designed so that a nitrogen atom was adjacent to the boron atom to enhance the reactivity of the boronic acid functionality to the oligosaccharides2. When the IRABA is added to a solution containing an oligosaccharide of interest, the chemical reaction between the IRABA and diol functionalities of the oligosaccharide is responsible for the covalent addition of the IR-active phosphonate group. Two series of oligosaccharides were studied, the lacto-N-fucopentaoses (LNFPs), all have the same backbone sequence but differ either in the connectivity of fucose (Fuc) or the linkage of their terminal Gal and Fuc moieties, and the lacto-N-difucohexaoses (LNDFHs), also have the same backbone structure but have two Fuc residues and differ either in the connectivity or linkage of one of the Fuc residues. The IRABA-modified analytes proved to be ideal for characterization and differentiation using IRMPD. Structures of Oligosaccharides m/z Figure 4: ESI mass spectra for IRABA derivatized LNFP-I in both the negative and positive ion modes. The IRABA-derivatization efficiency is very high. Figure 7: IRMPD spectra of the IRABA derivatized LNDFH oligosaccharides LNFP LNDFH In negative mode, there is no indication of whether the Hex residues are cleaved from the reducing or non-reducing ends.Therefore, the analytical value is similar to that obtained from the CAD and IRMPD spectra of the deprotonated underivatized oligosaccharides. However in positive mode, the data suggests that there are two general dissociation pathways: the most prominent one from sequential losses from the non-reducing end of the oligosaccharides (a result of Y-type cleavages) and the other entailing losses from the reducing end (a result of B-type cleavages). DSLNT Oligosaccharide The IRABA derivatization strategy was further extended to a hexasaccharide dicarboxylic acid, disialyllacto-N-tetraose(DSLNT). As observed with the previous series, the principal fragments are a series that sequence the backbone from the non-reducing to reducing end of the oligosaccharide (Figure 8). Following the loss of the two Sia moieties, a combined loss of Gal-GlcNAc resulting from cleavage at the reducing end of GlcNAc indicates the sequence Sia-Gal-(Sia)-GlcNAc. Subsequent losses reveal the completion of the sequence, Gal-Glc. The successful sequencing of this oligosaccharide suggests that the method may be further applied to larger and different types of oligosaccharides. LNDFH-I(a) Fucα1-2Galβ1-3(Fucα1-4)GlcNAcβ1-3Galβ1-4Glc [DSLNT+IRABA-2W+H]+ Y4α LNDFH-I(b) Fucα1-2Galβ1-4(Fucα1-3)GlcNAcβ1-3Galβ1-4Glc (-Fuc) Figure 6:IRMPD spectra for the IRABA derivatized LNFPs In addition to the same losses observed in the CAD spectra, the IRMPD spectra also reveal ions that have undergone subsequent dissociation. These secondary IRMPD events are responsible for the two additional Hex losses observed for all of the LNFPs in Figures 6. As a result, there is complete sequence coverage for each of the LNFPs. In addition, the diagnostic pathway for the LNFP-V is present (initial loss of Gal + GlcNAc) along with a diagnostic ion for the LNFP-I isomer that was not observed in the CAD spectra (Y2). It is surmised that this ion is due to sequential fragmentation of the fragment ion at 973 Da due to the initial loss of Fuc. We believe the differences in location of the charge site after the initial loss of Fuc for LNFP-I with a non-reducing terminal Fuc, LNFP-II and III  with a GlcNAc bound Fuc, and LNFP-V with a Glc bound Fuc are what make this loss for the LNFP-I isomer possible. The primary fragmentation pathway is the key to reliable sequencing of the oligosaccharides since the fragments that occur in this pathway only occur from one end of the oligosaccharide. This primary fragmentation pathway is observed only when the oligosaccharides are derivatized. Therfore, derivatization coupled with IRMPD has proven to be an excellent technique for sequencing the LNFPs in one activation event. Site of Derivatization The losses observed in the primary fragmentation pathway in the IRMPD spectra suggest that the IRABA is attached to the reducing sugar. Since the reducing sugar is the most reactive and may have both the alpha and beta configuration, it may bear the cis-diol functionality that readily reacts with boronic acids. The existence of the secondary pathway suggests that there is at least one other site of derivatization. Since the first losses in the secondary pathway include two Hex and Fuc, it is surmised that the secondary site of attachment is at one of the Gal moieties that also possess the cis-diofunctionality. To further elucidate the site of attachment, the reducing end of the LNFP-II isomer was reduced to an alditol with sodium borohydride. The alditol form of LNFP-II was then reacted with the IRABA ligand, followed by analysis by ESI-MS. No product of derivatization was observed in the spectra. Given these results, it is surmised that the IRABA reacts primarily at the reducing sugar and that any other sites of derivatization are minor products. LNDFH-II Galβ1-3(Fucα1-4)GlcNAcβ1-3Galβ1-4(Fucα1-3)Glc CONCLUSIONS The sequencing of oligosaccharides has been simplified with the use of IRMPD in a QIT. This was accomplished with the use of a boronic acid derivatizing reagent that was also functionalized with an IR-active phosphonate group to facilitate the photon absorption process. The oligosaccahrides underwent modification by simple addition of the IRABA with reaction times of ~1 min and did not require sample cleanup prior to analysis by ESI-MS. The resulting IRABA-oligosaccharide products dissociated with high efficiency upon IR irradiation, and the degree of secondary fragmentation may be controlled by adjusting the irradiation time. The resulting IRMPD mass spectra display a characteristic primary fragmentation pathway that consisted of one uniform type of ion, thereby simplifying the determination of oligosaccharide sequence. As a result, the method should be generally applicable to the sequencing of even larger unknown oligosaccharides. Figure 8: IRMPD of IRABA derivatized negatively-charged hexasacharide, disiayllacto-N-tetraose (DSLNT) LNFP-V Galβ1-3GlcNAcβ1-3Galβ1-4(Fucα1-3)Glc m/z Figure 3: CAD and IRMPD spectra of the deprotonated oligosaccharides The spectra for the first three isomers (Figure 3a-c) all result in losses of a hexose residue (Hex, -162 Da) as a Y- or B-type cleavage, therefore not providing any valuable structural information. In the CAD spectrum for LNFP-V (Figure 3d), a loss of 164 Da indicates the elimination of a fucose residue (Fuc) with water (H2O). This additional water loss is indicative of a Z- or C-type cleavage where the intersaccharide oxygen atom is associated with the non-reducing end of the oligosaccharide upon cleavage. In addition, a loss of two saccharide groups, both hexose and fucose residues, is observed. The IRMPD spectra with an increase in the laser irradiation time resulted in an increase in intensity of lower mass fragments coupled with the decrease in intensity of higher mass fragments. The spectrum for deprotonated LNFP-I through LNFP-V are displayed above (Figure 3e-h). Unfortunately, for each isomer there is a lack of a significant amount of information for sequencing the backbone and Fuc position. In addition, the sequential losses are limited to one to three monosaccharide units, thus restricting the coverage of information obtained. Due to the unsatisfactory structural characterization of the deprotonated oligosaccharides, the CAD and IRMPD spectra of the sodium-cationized compounds were obtained for comparison. In the positive ion mode, sodium-cationized complexeswere formed almost exclusively over the protonated species, and thus they were therefore used for the analyses. ACKNOWLEDGEMENTS Support from the Robert A. Welch Foundation (F1155) and the National Science Foundation (CHE-0315337) is gratefully acknowledged. REFERENCES (1) Wilson, J. J. Brodbelt, J.S. A. Chem.2006, 78, 6855-6862. (2) Zhu, L.; Shabbir, S. H.; Gray, M.; Lynch, V.; Sorey, S.; Anslyn, E. V. J. Am. Chem. Soc.2006, 128, 1222-1232. (3) Zhu, L.; Anslyn, E. V. J. Am. Chem. Soc.2004, 126, 3676-3677. =Glc =Gal =GlcNAc =Fuc =Sia Figure 5: CAD spectra for the derivatized LNFPs of the type [LNFP+IRABA-2H2O+H]+ Figure 2: Structures of LNFP and LNDFH oligosaccharides