Anal. Chem. 2007, 79, 1626-1633
Gold Nanoparticles as Assisted Matrix for Determining Neutral Small Carbohydrates through Laser Desorption/Ionization Time-of-Flight Mass Spectrometry Chih-Lin Su and Wei-Lung Tseng*
Department of Chemistry, National Sun Yat-sen University and National Sun Yat-Sen University-Kaohsiung Medical University Joint Center, Kaohsiung, Taiwan
A method for the analysis of small neutral carbohydrates by surface-assisted laser desorption/ionization mass spectrometry (SALDI-MS) has been developed by using bare AuNPs as matrices. In comparison with citrate-capped and DDAB-capped AuNPs, bare AuNPs can capture the analytes on their surface; therefore, small neutral carbohydrates, which are difficult to ionize by matrix-assisted desorption/ionization mass spectrometry, could be cationized very efficiently by SALDI-MS with AuNPs as matrices. By using SALDI-MS in the positive ion mode, many molecular ions are obtained from monosaccharides and disaccharides. Without derivatization, the limits of detection at a S/N ratio of 3 are 82, 41, 144, and 151 nM for ribose, glucose, cellobiose, and maltose. In comparison with conventional organic matrixes (e.g., 2,5dihydroxybenzoic acid), bare AuNPs as SALDI matrices offer many advantages, such as easy sample preparation, high ionization efficiency, and high shot-to-shot reproducibility. To validate the applicability of our method, a calibration curve is created from urine spiked with standard glucose (0.5-10 mM). We strongly believe that this approach can potentially be applied to diagnosis and glycomics. Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) developed by Karas and Hillemkamp1 has become a powerful tool for the analysis of large biomolecules, such as proteins, peptides, and nucleic acids.2-4 The UV-absorbing matrices that function as energy mediators effectively transfer the absorbed photoenergy from an irradiation source to the surrounding sample molecules, resulting in minimum fragmentation.5 However, because the matrix, which is also ionized by the application of the laser beam, produces a high background signal in the low-mass region ( Na > Li > H; therefore, no MH+ signal could be detected by SALDI-MS. Additionally, the MS spectra of the carbohydrate mixture (100 µM), including ribose, glucose, and maltose, are shown in Figure 4. We assign the peak at m/z ) 172.71, 202.91, and 364.35 to [ribose + Na]+, [glucose + Na]+, and [maltose + Na]+, respectively. In comparison with the MALDI-MS spectra obtained with DHB as a matrix (no molecular ions or alkali metal adduct ions were observed), the SALDI mass spectra produce significantly lower background levels and showed minimum molecular fragmentation. We suggest that bare AuNPs as SALDI matrices not only provide excellent sensitivity toward glucose but also remove the matrix interference. 1632
Analytical Chemistry, Vol. 79, No. 4, February 15, 2007
Quantitative Analyses of Four Neutral Carbohydrates. The homogeneity achieved using nanomaterials for assisted-matrix desorption/ionization MS is higher than obtained by the formation of crystals of the analyte molecule with the conventional organic matrix; this results in an improvement in quantitative analysis. Recently, Chang’s group has demonstrated that excellent reproducibility of the spectrum signals is obtained using modified AuNPs as the matrix for determining aminothiols in SALDI-MS.12b A similar phenomenon is observed in the case of carbon nanotubes.15a In this study, we also investigate whether neutral carbohydrates could be quantified by SALDI-MS according to the intensities of the sodium adduct ions of carbohydrate. When bare AuNPs are used as the SALDI matrix, the intensity of the glucose signal varies within less than 12% over 20 sample spots (Supporting information Figure S3). The calibration curve is constructed by a linear regression of the signal intensity (m/z ) 202.91) against the glucose concentrations. The plot presented in Figure 5D exhibits good linearity (R2 ) 0.9872) over the range of 1-100 µM. The limit of detection (LOD) for glucose was estimated from Figure 5C as 41 nM (41 fmol). Our results are significantly better than other reported data; the LODs of glucose obtained using MALDI and electrospray ionization MS are 5 and 1.2 pmol, respectively.31 We state that a potentially precise and time-saving procedure for the quantitative assays of glucose can be developed using bare AuNPs as the SALDI matrix provides; it obviates the need to perform calibrations by an internal standard. Moreover, Table 2 shows that when the four carbohydrates, namely, ribose, glucose, cellobiose, and maltose, are determined by SALDI-MS, the intensity changes linearly with their concentrations. Therefore, the plot of each carbohydrate exhibits good linearity (R2 > 0.98). The LODs for ribose, cellobiose, and maltose at a S/N ratio of 3 are 82, 144, and 151 nM, respectively. It is interesting to note that the LODs for the disaccharide are relatively higher than those for the monosaccharides. The lower ionization efficiency may be attributed to larger steric hindrance, thereby resulting in less adsorption between disaccharide molecules and nanoparticles. To further support our hypothesis, we have performed SALDI-MS analysis of 10 µM β-cyclodextrin (m/z 1157.95) with a matrix of bare AuNPs, and its mass spectrum is shown in Figure 6. Evidently, the relatively poor sensitivity for β-cyclodextrin is relatively less than that observed in Figure 3. Urine Glucose Analysis. We tested the application of our methods for the practical analyses of glucose in urine samples from a healthy female. The urine glucose levels for healthy subject and diabetic patient are 5-120 mg/L and 1600-2000 mg/L, respectively.32 By simple dilution without further pretreatment, an apparent peak was obtained at m/z ) 202.91 by SALDI-MS after 10 mM glucose was spiked into the urine samples (Figure 7A). In contrast, there was no visible glucose signal in the spectrum obtained by MALDI-MS using DHB as a matrix (data not shown). A calibration curve derived from solution containing glucose standards is shown in Figure 7C; it exhibited linearity in the working range of 0.5-10 mM, which includes the upper limit of the healthy glucose concentration. These results suggest that (31) (a)Wan, E. C. H.; Yu, J. Z. J. Chromatogr. A 2006, 1107, 175-181. (b) Bungert, D.; Heinzle, E.; Tholey, A. Anal. Biochem. 2004, 326, 167-175. (32) Jin, L. J.; Li, S. F. Y. Electrophoresis 1999, 20, 3450-3454.
the proposed method will be suitable for routine urine assay in clinical studies. CONCLUSIONS Bare AuNPs, prepared from the chemical reduction of ionic gold using sodium borohydride, were investigated as the matrix for the analysis of small neutral carbohydrates by SALDI-MS. In comparison with AuNPs with different capping systems (citrate, DDAB, glucose), bare AuNPs offer the following advantages: (a) high ionization efficiency for neutral carbohydrates, (b) simple sample preparation, (c) no requirement of derivatization procedures, and (d) quantitative improvement in the analysis of small molecules. When we performed SALDI-MS analysis by using the bare AuNPs (1×), the LODs for neutral carbohydrate, including pentoses, hexoses, and disaccharides, were in the range of 41151 nM. However, the sensitivity for β-cyclodextrin was relatively low. Additionally, a calibration curve was constructed by a linear regression of the signal intensity (m/z ) 202.91) against glucose
concentrations (0.5-10 mM) in urine samples. On the basis of these results, we strongly believe that this approach can potentially be applied to the analysis of glycoproteins and oligosaccharides. ACKNOWLEDGMENT We thank the National Science Council (NSC 95-2113-M-110020-) of Taiwan and Aim for the Top University Plan, Ministry of Education, Taiwan for financial support of this work. We thank Professor J. Shiea for allowing us to use the MALDI-TOF instrument. SUPPORTING INFORMATION AVAILABLE Four additional figures. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review September 16, 2006. Accepted December 7, 2006. AC061747W
Analytical Chemistry, Vol. 79, No. 4, February 15, 2007
1633
