LC with ES/MS/MS for detection of trace amounts of native MA. (underivatized and ... In LC/MS, the composition of the LC mobile phase is an important ...
Anal. Chem. 1997, 69, 1956-1960
Trace Detection of Underivatized Muramic Acid in Environmental Dust Samples by Microcolumn Liquid Chromatography-Electrospray Tandem Mass Spectrometry Mona Shahgholi,* Susan Ohorodnik, and John H. Callahan
Naval Research Laboratory, Chemistry Division, Code 6113, Washington, D.C. 20375 Alvin Fox
Department of Microbiology & Immunology, School of Medicine, University of South Carolina, Columbia, South Carolina 29208
Muramic acid (MA) is a universal chemical marker for bacterial cell wall polymers (peptidoglycan) present in complex environmental, industrial, and clinical matrices. Trace analysis of MA has been limited to a handful of laboratories because of the lengthy nature of derivatization procedures necessary for GC/MS and GC/MS/MS. The focus of the current report concerns environmental analysis (organic dust) using LC/MS/MS of native MA. Dust was heated in sulfuric acid to release MA from cell wall polymers. Acid was removed by extraction with an organic base. The samples were passed through a hydrophobic C-18 extraction column to remove nonpolar contaminants. This was followed by binding of MA to a strong cation-exchange extraction column to eliminate neutral and anionic components of the sample. MA was then eluted with hydrochloric acid, which was removed by evaporation. Analysis was done by microscale LC ES/ MS/MS using a custom-made Lichrosphere-diol capillary column. 13C-Labeled MA was used as an internal standard. Total sample preparation takes approximately 12 h. The simplicity of the sample preparation may eventually allow MA analysis to be used in a host of applications, including air monitoring for biocontamination. This is the first report, to our knowledge, of microscale LC ES/MS/ MS analysis for MA (or, indeed, any other sugar monomer) in a complex matrix. Bacteria are conventionally detected in environmental and clinical matrices by microbiological culture (i.e., growth on liquid or solid media). Lack of growth of certain bacterial species can make the culture difficult to interpret. Furthermore, the culture is quite time consuming, often taking several days for successful isolation of colonies. Alternatively, bacteria can be monitored in complex matrices by determining the levels of certain compounds derived from the bacterial cell envelope.1,2 For example, muramic acid (MA), a simple amino sugar, has been used as a chemical (1) Fox, A.; Morgan, S.; Gilbart, J. In Analysis of carbohydrates by gas liquid chromatography (GC) and mass spectrometry (MS); Bierman, C. J., McGinnis, G., Eds.; CRC Press: Boca Raton, FL, 1989. (2) Black, G.; Fox, A.; Fox, K.; Smith, P.; Snyder, P. Anal. Chem. 1994, 66, 4171-4176.
1956 Analytical Chemistry, Vol. 69, No. 10, May 15, 1997
marker to detect the ubiquitous bacterial cell wall polymer peptidoglycan (PG). PG is a highly inflammatory substance and may be responsible for the respiratory problems associated with breathing air contaminated with microorganisms (e.g., the “sick building syndrome”).3 Muramic acid (3-O-lactylglucosamine) is a compound exclusively found in the PG of bacteria and not elsewhere in nature; thus, MA levels serve as an indicator of total PG levels. Classical methods for assaying MA involving spectrophotometry or chromatography (without MS detection) are still frequently employed;4-6 however, these methods are nonselective, providing equivocal results for complex biological samples. More commonly, trace analysis of MA in complex matrices is performed using GC/MS7-9 and GC/MS/MS10-12 methods, taking advantage of the high specificity of the mass spectrometric techniques. The main undesirable aspect of the GC/MS approach is the lengthy derivatization procedure used for converting MA into volatile acetylated alditols, a procedure which results in long analysis times, a minimum of 52 h.10 Trace analysis of MA using LC/MS methods would have substantial advantages because it offers the possibility for analysis without derivatization and the ability to handle complex mixtures (multicomponent) and complex matrices. This would simplify sample handling while also substantially shortening analysis time. Previous studies have shown that MA could be detected in bacterial cell hydrolysates by ES/MS/MS2 and by LC plasma spray13 using analytical-scale LC columns (4 mm i.d.). The samples examined in the aforementioned studies2,13 were cell (3) Fox, A. In Field Guide for the Determination of Biological Contaminants in Environmental Samples; Dillon, H., Heinsohn, P., Miller, D., Eds.; American Industrial Hygeine Association: 1996. (4) Hadzˇija, O. Anal. Biochem. 1974, 60, 512-517. (5) Hoijer, M.; Melief, M.; van Helden-Meeuwsen, C.; Eulderink, F.; Hazenberg, M. Infect. Immun. 1995, 63, 1652-1657. (6) Tipper, D. J. Biochemistry 1968, 7, 1441-1449. (7) Findlay, R.; Moriarty, D. J. W.; White, D. C. J. Geomicrobiol. 1983, 3, 135150. (8) Fox, A.; Rosario, R.; Larsson, L. Appl. Environ. Microbiol. 1993, 59, 43544360. (9) Mielniczuk, Z.; Mielniczuk, E.; Larsson, L. J. Chromatogr. Biomed. Appl. 1995, 670, 167-172. (10) Fox, A.; Wright, L.; Fox, K. J. Microbiol. Methods 1995, 22, 11-26. (11) Saraf, A.; Larsson, L. J. Mass Spectrom. 1996, 31, 389-396. (12) Fox, A.; Kramer, M.; Harrelson, D. J. Microbiol. Methods. In press. S0003-2700(96)00914-6 CCC: $14.00
© 1997 American Chemical Society
hydrolysates of pure bacterial cultures containing high levels of MA (parts per thousand or higher) in relatively homogeneous matrices. The current work, on the other hand, uses microscale LC with ES/MS/MS for detection of trace amounts of native MA (underivatized and unlabeled MA) in environmental dust with matrices significantly more complex than those of pure bacterial hydrolysates. The analysis of mixtures of sugars and complex matrices has generally required the use of both chromatographic and MS/MS14 (tandem MS) methods to ensure high selectivity, specificity, and sensitivity. MS/MS is an effective tool, relaxing the need for highresolution chromatography. However, it cannot entirely eliminate the need for chromatographic separations because of ion suppression phenomena that could influence the observed abundances of detected species in complex mixtures,15,16 thereby limiting the sensitivity of ES/MS detection. Ion suppression effects on quantitation can be reduced even further by using an internal standard consisting of the isotope-labeled analog of the compound of interest. A recent development in LC/MS technology is the trend toward microscale LC instead of the more traditional analyticalscale LC. Microscale LC is superior in separation and higher in sensitivity compared with analytical-scale LC.17 In addition, the lower chromatographic flow rates are compatible with flow rates for optimal electrospray ionization. Microscale LC/MS is now routinely used in many laboratories for analysis of peptides and proteins;18,19 however, its use for analysis of underivatized sugars is still limited. For example, Jorgenson et al. demonstrated the use of packed microcolumns with CF-FAB for analysis of a variety of compounds including maltotetraose, a tetrasaccharide.20 The use of reversed-phase capillary LC for analysis of derivatized complex carbohydrates released from glycoproteins was demonstrated by Reinhold et al.21 We recently reported the detection of subpicomolar amounts of underivatized chitobiose, a disaccharide, using capillary columns packed with porous graphitic carbon.22 The achievement of subpicomolar detection limits for chitobiose was ascribed to the use of microscale LC and selection of chromatographic packing materials and solvents that were compatible with electrospray ionization. In LC/MS, the composition of the LC mobile phase is an important consideration for achieving good chromatographic separation, but it is even more important for MS detection, as (13) Elmroth, I.; Larsson, L.; Westerdahl, G.; Odham, G. J. Chromatogr. 1992, 598, 43-50. (14) Busch, K. L.; Glish, G. L.; McLuckey, S. A. Mass Spectrometry/Mass Spectrometry; Techniques and Applications of Tandem Mass Spectrometry; VCH Publishers: New York, 1988. (15) Ligon, W. V.; Dorn, S. B. Int. J. Mass Spectrom. Ion Processes 1985, 63, 315-324. (16) Ligon, W. V.; Dorn, S. B. Int. J. Mass Spectrom. Ion Processes 1986, 68, 337-340. (17) Kennedy, R. T.; Jorgenson, J. W. Anal. Chem. 1989, 61, 1128-1135. (18) Emmett, M. R.; Caprioli, R. M. J. Am. Soc. Mass Spectrom. 1994, 5, 605613. (19) Hunt, D. F.; Alexander, J. E.; McCormack, A. L.; Martineo, P. A.; Michel, H.; Shabanowitz, J.; Sherman, N. In Techniques in Protein Chemistry II; Villafranca, J. J., Ed.; Academic Press: New York, 1991; pp 441-454. (20) Moseley, M. A.; Deterding, L. J.; de Wit, J. S. M.; Tomer, K. B.; Kennedy, R. T.; Bragg, N.; Jorgenson, J. W. Anal. Chem. 1989, 61, 1577-1584. (21) Mora, R.; Drazen, J. M.; Reinhold, V. N. Proceedings of the 40th ASMS Conference on Mass Spectrometry and Allied Topics, Washington, D.C., May 31-June 5, 1992; 100a-b. Chan, S.; Reinhold, B.; Reinhold, V. Proceedings of the 40th ASMS Conference on Mass Spectrometry and Allied Topics, Washington, D.C., May 31-June 5, 1992; 102a-b. (22) Shahgholi, M.; Ross, M. M.; Callahan, J. H.; Smucker, R. A. Anal. Chem. 1996, 68, 1335-1341.
various buffers and additives can suppress ionization and compromise sensitivity.23 For example, high-performance anionexchange chromatography (HPAEC) is the most frequently used procedure for chromatographic analysis of simple and complex carbohydrates.24 However, the strongly alkaline solutions used in HPAEC of monosaccharides are not compatible with MS detection and require the use of ion suppressors for removal of sodium ions. This has been demonstrated by Fenselau et al.25 for LC thermospray and Conboy and Henion26 for ion spray. Ion suppressors are effective in lowering the ionic strength of the eluents; however, amino sugars are removed from the effluent stream by these devices, rendering them unsuitable for the analysis of MA. In such instances, alternative columns using EScompatible mobile phases are necessary. Another difficulty commonly encountered in analysis of complex biological and environmental samples by LC concerns column fouling. This problem is less severe when the analysis involves sample derivatization because many impurities are removed during this process. For untreated samples, a standard method for dealing with the column fouling problem is the use of off-line cleanup, focusing the analysis on one type of analyte. For example, in the determination of MA in dust hydrolysates, strong cation-exchange (SCX) extraction columns were used to isolate the positively charged species selectively; these included MA, other positively charged sugars, and amino acids from the hydrolysates. This work describes the use of microscale LC with ES/MS/ MS for trace analysis of native MA in environmental dust samples. The complexity of the dust matrix necessitated the use of off-line sample cleanup in addition to on-line LC for reliable determination of low levels of MA. 13C-Labeled MA was used as internal standard for quantitation and to account for potential matrix interferences with MA determination. MS/MS was used for positive identification of trace levels of MA in the dust hydrolysates with a complex matrix. EXPERIMENTAL SECTION Materials. Muramic acid (MA), methionylarginylphenylalanylalanine (MRFA), and myoglobin were from Sigma Chemical Co. (St. Louis, MO). HPLC-grade acetonitrile, methanol, and water were purchased from Fisher Scientific (Fair Lawn, NJ). Acetic acid was from Aldrich Chemical Co., Inc. (Milwaukee, WI). Dioctylmethylamine was from Fluka (Ronkonkoma, NY). 13CO2Labeled blue-green algae, used as a source of [13C]MA (universally labeled [13C]muramic acid, the internal standard), was obtained from Isotec (Miamisburg, OH). Surface dust was collected from a barn in Blackville, SC. To ensure homogeneity, the sample was frozen in liquid nitrogen and ground before lyophilization. A fungus, Torulopsis glabrata (used as a negative control), was grown in Saborauds media and water washed before freeze drying. Hydrolysis and Sample Preparation Procedures. Lyophilized dust and fungal samples were hydrolyzed in 2 N H2SO4 for 3 h at 100 °C. 13CO2-Labeled blue-green algae were hydrolyzed in a similar manner, and 500 µg aliquots were added to the dust and fungal hydrolysates as internal standard. The algae contained (23) Roboz, J.; Yy, Q.; Meng, A.; van Soest, R. Rapid Commun. Mass Spectrom. 1994, 8, 621-626. (24) Lee, Y. C. Anal. Biochem. 1990, 189, 151. (25) Simpson, R. C.; Fenselau, C. C.; Hardy, M. R.; Townsend, R. R.; Lee, Y. C.; Cotter, R. J. Anal. Chem. 1990, 62, 248-252. (26) Conboy, J. J.; Henion, J. Biol. Mass Spectrom. 1992, 21, 397-407.
Analytical Chemistry, Vol. 69, No. 10, May 15, 1997
1957
approximately 0.4% [13C]MA on a dry weight basis. Standards consisted of varying amounts of MA (0-3240 ng) and 500 µg of 13CO -labeled blue-green algae (approximately 2000 ng of [13C]MA). 2 Samples were neutralized by mixing with 1.5 mL of N,Ndioctylmethylamine/chloroform (50:50 v/v). The aqueous phase was passed through a C-18 extraction column (Analytichem, Harbor City, CA) and frozen until ready for MS analysis. Prior to MS analysis, the aqueous samples were treated with SCX extraction columns (Varian, Harbor City, CA) as described in ref 27. Briefly, the SCX extraction columns were conditioned with three column volumes of methanol, three column volumes of water, and six column volumes of 4 N HCl and washed with six column volumes of deionized water. The hydrolysates were applied to the SCX columns, followed by 0.5 mL of 0.1 N HCl. The SCX resin was washed with two column volumes of water to remove the neutral and anionic species. Cations, including MA, were eluted with two column volumes of 4 N HCl. The samples were dried in a vacuum centrifuge to remove the hydrochloric acid and reconstituted in 250 µL of acetonitrile for mass spectrometric analysis. Chromatography. Microcolumns were made according to the method of Kennedy and Jorgenson28 using a 50 cm segment of fused-silica capillary (SGE Inc., Austin, TX) with 75 µm inner diameter. The columns were slurry-packed to a length of 5.5 cm with 5 µm diameter particles of Lichrosphere-diol (Phenomenex, Torrance, CA) and washed for 30 min with methanol, the packing solvent. This was followed by a 30 min wash with 90:10 (v/v) acetonitrile/water, the elution solvents, after which the column was inserted into the ES needle. A Model 140B microgradient dual-syringe pump (Applied Biosystems, Foster City, CA) was used for delivery of the eluents. The flow rate through the column was adjusted to 0.5 µL/min with the aid of a splitting tee and a length of restriction tubing (50 µm i.d. fused-silica capillary). The column flow was supplemented with an auxiliary flow of a sheath liquid, 70:30 (v/v) methanol/0.02 M acetic acid at 2 µL/min. The microcolumn was attached to a high-pressure injector for sample loading, and then 0.5-1 µL of sample was injected on the column by displacement,22 after which the microcolumn was reconnected to the microsyringe pump and the chromatographic run was started. The chromatographic run consisted of an acetonitrile/water gradient of 95% to 5% over 10 min, holding at 5% for 5 min, back to 95% over 3 min, and holding at 95% for 5 min. The microcolumn was conditioned at the beginning and end of each day with a blank chromatographic run. Mass Spectrometry. The experiments were performed on a TSQ-70 triple-quadrupole mass spectrometer (Finnigan-MAT Corp., San Jose, CA) fitted with an electrospray interface. The electrospray needle was held at a potential of +4.5 kV, and the capillary inlet was heated to 200 °C. The mass spectrometer was initially tuned by infusing a standard solution of 20 pm/µL MRFA and 5 pm/µL myoglobin in 50:50 MeOH/0.5% acetic acid at 2 µL/ min. Following tuning and mass calibration, optimal collisionactivated dissociation (CAD) conditions for fragmentation of MA were determined by direct infusion of a 10 µM standard solution of MA in acetonitrile. These experiments were performed with argon as collision gas at a pressure of 2.0-2.2 mTorr, with a collision offset of -25 eV. The direct infusion line was then (27) Ueda, K.; Morgan, S. L.; Fox, A.; Gilbart, J.; Sonesson, A.; Larsson, L.; Odham, G. Anal. Chem. 1989, 61, 265-270. (28) Kennedy, R. T.; Jorgenson, J. W. Anal. Chem. 1989, 56,1128-1135.
1958
Analytical Chemistry, Vol. 69, No. 10, May 15, 1997
removed from the ES needle and replaced with the packed microcolumn. Data were obtained in the full-scan and multiple-reactionmonitoring (MRM) modes for the microscale LC ES/MS/MS experiments. Full mass spectra were acquired over m/z 150400 at 0.5 s/scan. For MRM experiments, a simple instrument control language procedure was written to monitor the fragmentation pathways specific for native and 13C-labeled muramic acid. These were m/z 252f216, 252f144 for MA and m/z 261f225, 261f150 for 13C-labeled MA. The dwell time was 0.13 s per dissociation. The fragmentation pathways for MA are discussed by Fox et al.2 RESULTS AND DISCUSSION The purpose of this work was to develop an LC/MS/MS method for trace analysis of MA in environmental dust samples in order to eliminate the lengthy derivatization procedure necessary for analysis by GC/MS/MS. The most significant challenge of this assay was the complexity of the sample matrix, which was addressed in the following manner: off-line sample cleanup to simplify the sample matrix; chromatographic separation with a packing material that uses ES-compatible eluents; use of a suitable internal standard for reliable quantitative analysis; standard calibration to test the quantitative nature of the assay; and data acquisition in the MRM mode to improve specificity and limits of detection. For the LC/MS method, off-line sample cleanup was used to reduce matrix complexity, which lowers sensitivity and degrades column performance. Both C-18 and SCX extraction columns were used for the cleanup procedure. The C-18 columns were used to remove the nonpolar contaminants in the hydrolysates, while neutrals and anionic components were removed with SCX treatment. Cleanup with SCX was particularly effective in eliminating column fouling problems. The main considerations in selecting the LC packing material were the chromatographic performance of the column and the nature of the eluents used for separation. In this work, three different chromatographic packing materials were tested for separation of MA from other sample components, all using EScompatible mobile phases such as water and acetonitrile. These were PolyGLYCOPLEX (PolyLC Inc., Columbia, MD), Lichrosphere-amino, and Lichrosphere-diol (EM Separations Technology, Gibbstown, NJ, purchased from Phenomenex, Torrance, CA). Although all three packing materials performed well with standard solutions of MA and internal standard, Lichrosphere-diol was selected for the assay because it showed lower background levels and minimal fouling with the environmental samples. It should be noted that the PolyGLYCOPLEX material was used in a capacity different from its suggested use, which is the analysis of complex carbohydrates in simple matrices. The choice of an appropriate internal standard was important for reliable quantitation and critical in accounting for potential interferences from other matrix components. Preliminary experiments used N-methylglucamine (NMG), previously used in related work dealing with determination of MA.13 Although useful for quantitation of MA in standard solutions, NMG was not ideal for the complex environmental samples under study because it would not correct for potential suppression effects from coeluting peaks or background (NMG retention time differs from that of MA). The alternative internal standard selected was 13C-labeled MA, found in the peptidoglycan of blue-green algae that were grown
Figure 1. Microscale LC/MS extracted ion chromatogram of a 360 ng standard solution of MA spiked with approximately 2000 ng of [13C]MA from blue-green algae. MA is monitored at m/z 252 and [13C]MA at m/z 261. Absolute signal abundances are displayed in the upper right corner of each chromatogram.
Figure 3. MRM chromatogram of a 360 ng standard solution of MA containing approximately 2000 ng of [13C]MA. The top trace shows the sum of m/z 261f225 and 261f150, the specific dissociations of [13C]MA. The lower trace shows the sum of m/z 252f216 and 252f144, the specific dissociations of native MA. Absolute signal abundances are displayed in the upper right corner of each figure. Table 1. Microscale LC/MS/MS Results for Muramic Acid (MA) in Environmental Dust
Figure 2. CAD spectrum of a 10 µM standard solution of MA showing the fragmentation pathways of MA: m/z 252f234, 252f216, 252f144, and 252f126. The dissociations monitored for MRM analysis are marked with an asterisk.
in 13CO2-enriched media. (A commercial source of pure 13Clabeled MA is not currently available.) Samples of the blue-green algae were thus hydrolyzed in a manner similar to the dust samples and spiked into all samples as a source of 13C-labeled MA. The 13C-labeled MA was seen at m/z 261 in positive ion ES and the native MA at m/z 252. This is demonstrated in Figure 1, which shows a microscale LC-extracted ion chromatogram of a 360 ng standard solution of MA spiked with 500 µg of blue-green algae. Three different mass spectrometric experiments were performed in this study. Full-scan spectra were obtained for standards and the environmental samples to optimize the chromatographic separation of MA from other sample components. CAD spectra were then collected for MA, native and 13C-labeled, to identify their specific product ions. Having identified the specific dissociations for MA, all other mass spectrometric runs were done in the MRM mode, which showed superior signal-tonoise ratios because of chemical noise reduction. Figure 2 shows a CAD spectrum of a 10 µM standard solution of MA, indicating (with an asterisk) the dissociations monitored for MRM detection. As noted in Figure 2, only two of MA’s four specific fragmentation pathways (252f216 and 252f144) were monitored during MRM because the remaining two pathways, although abundant, were
samplea
MA (ng)
R252/261
standard 0 standard 40 standard 3240 dust 1 (21.4 mg) dust 2 (21.3 mg) dust 3 (21.7 mg) dust 4 (13.2 mg) dust 5 (12.2 mg) dust 6 (11.6 mg) fungus 1 (21.8 mg) fungus 2 (16.2 mg) fungus 3 (22.2 mg)
0 40 3240 2683 2567 3317 1900 1717 1466
0.0 2.35 × 10-2 1.94 × 10-2 1.61 1.54 1.99 1.14 1.03 0.88 c c c
ng of MA/mg of dust b 125 121 153 144 141 126
