Anal. Chem. 1999, 71, 1914-1917
Automated Derivatization Instrument: Preparation of Alditol Acetates for Analysis of Bacterial Carbohydrates Using Gas Chromatography/Mass Spectrometry Paul Steinberg and Alvin Fox*
Department of Microbiology and Immunology, University of South Carolina, School of Medicine, Columbia, South Carolina 29208
An automated derivatization instrument was developed for preparation of alditol acetates for GC/MS profiling of bacterial carbohydrates. The multistage alditol acetate method is now performed sequentially by a computercontrolled instrument. A series of electrically driven solenoid valves are in-line with a 21-sample manifold. A set of solvent valves controls the input of solvent and/or nitrogen gas to each sample chamber. A set of gas valves controls output to atmosphere or vacuum. Additionally, closure of all valves allows the sample to be sealed in a closed chamber. Temperature is also determined automatically. The availability of this instrument could help popularize the alditol acetate procedure and serve as a prototype for automation of other complex derivatization procedures. Chemotaxonomy-based methods for microbial differentiation, using chromatography and/or mass spectrometry, are primarily used in research or reference laboratories.1 Microorganisms are classified on the basis of their chemical constituents. Commonly, low-molecular-weight fatty acid monomers are released by methanolysis and converted to fatty acid methyl esters (FAMES) for analysis by GC using a flame ionization detector (FID).2 This allows universal bacterial speciation. Although data processing is automated, sample preparation is not. Profiling of whole cell hydrolysates for their carbohydrate composition also allows bacterial speciation. To prepare samples for gas chromatography/mass spectrometry (GC/MS) analysis, the sugars are first released from whole cells by acid hydrolysis and then converted into alditol acetates. Due to the timeconsuming nature of the multistep derivatization, carbohydrate profiling of bacteria has not been routinely performed outside a few specialized laboratories.3,4 Carbohydrate profiling of bacteria * Corresponding Author: (phone) 803 733 3288; (fax) 803 733 3192; (e-mail)
[email protected]. (1) Fox A., Larsson L., Morgan S. L., Odham, G., Eds. Analytical Microbiology Methods: Chromatography and Mass Spectrometry; Plenum Press: New York, 1990. (2) Moss, C. W. In Analytical Microbiology Methods: Chromatography and Mass Spectrometry; Fox, A., Larsson L., Morgan S. L., Odham, G., Eds; Plenum Press: New York, 1990; pp 59-70.
1914 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999
can be performed by GC with an FID,5 but GC/MS is preferred.6-10 In the selected ion monitoring (SIM) MS mode, simple chromatograms free of background interferences from other components of the bacterial cell (e.g., amino acids, fatty acids, and nucleotides) are generated. Total ion spectra can also be used to identify each sugar component. In contrast, derivatization is unnecessary prior to liquid chromatography/electrospray mass spectrometry analysis of bacterial carbohydrates but requires further development for routine analysis.11-16 Automation could help enormously in popularizing the alditol acetate procedure and, indeed, other multistep derivatization techniques. In the current report, the development and application of a computer-controlled derivatization instrument for automated preparation of alditol acetates is described. After hydrolysis and prederivatization cleanup, the instrument performs sequentially all steps of the alditol acetate procedure including borodeuteride reduction, elimination of borate, drying, and acetylation. A simple postderivatization cleanup is then performed. (3) Fox A.; Morgan S. L.; Gilbart J. In Analysis of Carbohydrates by GLC and MS; Biermann C., McGinnis, G., Eds; CRC Press: Boca Raton, 1989; pp 87-117. (4) Fox A.; Black G. In Mass Spectrometry for the Characterization of Microorganisms; Fenselau, C., Ed.; ACS Symposium Series 541; American Chemical Society: Washington, DC, 1994; pp 107-131. (5) Fox, A.; Lau, P.; Brown, A.; Morgan, S. L.; Zhu, Z.-T.; Lema, M.; Walla, M. J. Clin. Microbiol. 1984, 19, 326-332. (6) Walla, M.; Lau, P.; Morgan, S. L.; Fox, A.; Brown, A. J. Chromatogr. 1984, 288, 399-413. (7) Fox, A.; Rogers, J.; Fox K.; Schnitzer, G.; Morgan, S. L.; Brown, A.; Aono, A. J. Clin. Microbiol. 1990, 28, 546-552. (8) Fox, A.; Black G.; Fox, K.; Rostovtseva, R. J. Clin. Microbiol. 1993, 31, 887-994. (9) Fox, K.; Fox, A.; Nagpal, M.; Steinberg, P.; Heroux, K. J. Clin. Microbiol. 1998, 36, 3217-3222. (10) Wunschel, D.; Fox, K.; Black, G.; Fox, A. Syst. Appl. Microbiol. 1994, 17, 625-635. (11) Wunschel, D.; Fox, K.; Fox, A.; Nagpal, M.; Kim, K.; Stewart, G.;Shahgholi, M. J. Chromatogr., A 1997, 776, 205-219. (12) Fox, K.; Wunschel, D.; Fox, A.; Stewart, G. J. Microbiol. Methods 1998, 33, 1-11. (13) Fox K., Stewart G.; Fox A. Infect. Immunol. 1998, 66, 4004-4007. (14) Simpson, R. C.; Fenselau, C. C.; Hardy, M. R.; Townsend; R. R.; Lee, Y. C.; Cotter, R. J. Anal. Chem. 1990, 62, 248-252. (15) Conboy, J. J.; Henion, J. Biol. Mass Spectrom. 1992, 21, 397-407. (16) Shahgholi, M.; Ohorodnik, S.; Callahan, S. J.; Fox, A. Anal. Chem. 1997, 69, 1956-1960. 10.1021/ac981155g CCC: $18.00
© 1999 American Chemical Society Published on Web 03/20/1999
Figure 2. Individual reaction chamber of the automated derivatization instrument: 1, ferrule (Teflon); 2, bushing (nylon); 3, adaptor (Teflon); 4, nut with ferrule (Teflon); 5, input line (FEP); 6, output line (stainless steel); 7, tube insert (glass); 8, sample tube (glass); 9, adaptor (glass).
Figure 1. Outline of the steps in the alditol acetate procedure using an automated derivatization instrument.
EXPERIMENTAL SECTION Automated Derivatization Instrument. The multistage alditol acetate method is now performed sequentially by a computercontrolled instrument. The procedure, which previously took 2.5 working days, is performed in an automated fashion, requiring a total of 90 min of manual work. The samples are processed in four stages: (1) evacuation/hydrolysis (3 h and 15 min, automated); (2) prederivatization cleanup (1 h, manual); (3) alditol acetate derivatization (23 h, automated); and (4) postderivatization cleanup (30 min, manual). This procedure is overviewed in Figure 1. The core of the machine, where chemical manipulations and reactions are performed, consists of a custom-built manifold with 21 glass chambers, to each of which a custom-made test tube (13 mm o.d. × 140 mm, Ace Glass, Vineland, NJ) can be attached. Input, at the top of each chamber, is through a tapered hemostat tube (Fisher, Swanee, GA) connected by a Teflon ferrule (Ace Glass). The output from each chamber is connected to the other chambers via stainless steel sidearms with one central exit. At the bottom of each chamber, the removable test tubes are also attached by a Teflon ferrule (Ace Glass). The manifold is seated in a movable heating block (Barnstead, Thermolyne, Dubuque, IA). Figure 2 shows an individual reaction chamber.
A series of electrically driven solenoid valves are in-line with the manifold. A set of solvent valves controls the input of solvent and/or nitrogen gas to each sample chamber (Bio-Chem Valve, Booton, NJ). A set of gas valves (Cole-Parmer, Vernon Hills, IL) controls output to atmosphere or vacuum. Additionally, closure of all valves allows the sample to be sealed in a closed chamber. Computer control of the individual stages of the derivatization process is a major feature of the system. A program was written for this purpose using Lab Windows/CVI (National Instruments, Austin, TX). Timed events in the software activate a DIO-24 I/O board (National Instruments) which in turn controls a modified ERA-01 eight-way relay board (Keithley, Cleveland, OH). The signal from the latter board controls valves, cooling fans, vacuum pump, and heating block temperature. A diagram of the complete instrument is shown in Figure 3. Automated Alditol Acetate Method. Bacillus anthracis strain Sterne and Bacillus cereus ATCC 12826 were grown in nutrient broth and washed 3 times in water. Samples were autoclaved and freeze-dried prior to analysis. Carbohydrate profiles were determined using the automated derivatization instrument described above and adapted from a manual procedure reviewed elsewhere.3,4 In the automated alditol acetate procedure, 10 mg of bacteria, in 1 mL of 2 N sulfuric acid, is placed in the custom test tubes mentioned above. The samples are attached to the manifold and the program is started. Oxygen is evacuated by repeated alternate exposure to nitrogen and vacuum. After evacuation, the program sets the heating block at 100 °C for hydrolysis. Heating continues for 3 h at under nitrogen. Following hydrolysis, the samples are removed from the instrument, neutralized with 2 mL of 50% N,N-dioctylmethylamine (Fluka, Buchs, Switzerland) in chloroform, and centrifuged. The Analytical Chemistry, Vol. 71, No. 9, May 1, 1999
1915
Figure 3. Diagrammatic representation of the automated derivatization instrument: 1, acetic anhydride reservoir (10 mL); 2, chloroform reservoir (250 mL); 3, methanol-acetic acid (200:1) reservoir (250 mL); 4, two-way solvent solenoid valve*; 5, three-way solenoid valve*; 6, 1/8-in. FEP tubing; 7, eight-way adjustable distribution valve; 8, reaction chamber assembly; 9, sample output manifold (1/2-in. stainless steel); 10, two-way gas solenoid valve*; 11, three-way gas solenoid valve*; 12, vacuum pump*; 13, heating block*; 14, movable platform; 15, waste gases and vapor; 16, sample lines exiting from distribution valves to reaction chamber assembly units. Note for simplicity only 6 sample lines and 6 of the 21 reaction chambers are shown. *Controlled by ERA-01 board.
aqueous phase is removed and passed through C-18 ODS columns (J & W, Fulsom, CA) into 21 new sample tubes via the evacuated 21-chamber manifold described above. Aqueous sodium borodeuteride (200 µL of 25 mg/mL) is then added to each sample. The derivatization procedure is entirely under computer control. After a 2-h delay, in which sample reduction occurs, methanol-acetic acid (200:1 v/v) is added by activation of a solenoid valve connected to a reagent reservoir. The program then sets the heating block at 60 °C, and evaporation under nitrogen occurs for 30 min. This addition and evaporation is automatically repeated 6 times to remove borodeuteride as tetramethyl borate gas. After the last addition, the system is evacuated by activation of the attached vacuum pump and the samples are dried for 4 h at room temperature. Acetic anhydride is then added to the samples from another reservoir, and the samples are acetylated for 13 h at 100 °C. Finally, the samples are evaporated to dryness under nitrogen and chloroform is added from a third reservoir. The final postderivatization cleanup (taking 30 min) is performed manually but also uses the 21-sample manifold. Thus, ancillary equipment is not required. Samples are passed through a pair of connected Chem-Elut columns (Varian, Walnut Creek, CA), the first being pretreated with 2 M acetic acid and the second with 14.8 M ammonium hydroxide. The chloroform eluent is evaporated under nitrogen, and samples are reconstituted for analysis. GC/MS analyses was carried out with a mass-selective detector (model 5970; Hewlett-Packard Co., Palo Alto, CA) interfaced to a GC (model 5890; Hewlett-Packard) equipped with an automated sample injector. Chromatography was accomplished on a DB5MS fused silica capillary column (0.25-mm i.d., 0.25-µm film thickness, 30-m length; J & W Scientific, Folsom, CA). Electron 1916 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999
ionization was performed at 70 eV for both full-scan monitoring and SIM. RESULTS An automated derivatization instrument has been developed at the University of South Carolina. The machine is described here for the first time. The first use of the instrument in computerassisted preparation of alditol acetates is illustrated below. Standard curves were generated for each of 15 sugars (The numbers in brackets refer to elution order on chromatography.): deoxyribose [1], rhamnose [2], ribose [3], fucose [4], xylose [6], inositol [7], mannose [8], glucose [9], galactose [10], glucosamine [11], mannosamine [12], galactosamine [13], D-glycero-D-mannoheptose [14], muramic acid [15], and D-glycero-L-mannoheptose [16]. As in the manual procedure, arabinose [5] was used as an internal standard for neutral sugars and methylglucamine for amino sugars [17]. Duplicate samples (seven levels ranging from 500 ng to 128 µg of each sugar) and two blank samples were analyzed. To each, 25 µg of arabinose and methylglucamine was added. All samples, in sulfuric acid, were taken through the entire analytical procedure. The ratio of peak area for each sugar compared to internal standard was plotted against total amount of that sugar in the sample. R2 values ranged from 0.992 to 0.9998. The relative standard deviations (RSD) for each sugar, determined using four replicates containing 64 µg of each sugar and 32 µg of internal standard, were all less than 6.39%. It is possible that RSD might be greater when the amount of sugar analyzed is lowered. The automated derivatization instrument was also used to prepare alditol acetates to compare the carbohydrate profiles of B. anthracis (The numbers are the percent dry weight of bacteria.) (rhamnose, 0.14; ribose, 0.91; glucose, 4.0; galactose, 4.7; glucosamine, 1.6; mannosamine, 0.36; galactosamine, 0.05; muramic
replicates for each of the eight sugars present in B. anthracis were 5.08% or less. CONCLUSIONS
Figure 4. SIM chromatogram of carbohydrates present in bacilli after release by hydrolysis and automated alditol acetate derivatization: (A) B. anthracis Sterne; (B) B. cereus ATCC 12826. See Results section for peak identification.
acid, 0.32) and B. cereus (ribose, 1.3; mannose, 0.03; glucose, 2.5; glucosamine, 2.8; mannosamine, 0.76; galactosamine; 0.84; muramic acid, 0.29). Both organisms contain large amounts of ribose, glucose, glucosamine, and muramic acid. In agreement with results obtained with the manual alditol acetate procedure, B. anthracis is readily discriminated from B. cereus by the presence of galactose and presence of small amounts of galactosamine (See Figure 4). The presence of rhamnose indicates the presence of some spores in the B. anthracis preparation. The RSD for five (17) Fox, A.; Wright L.; Fox, K. J. Microbiol. Methods 1995, 22, 11-26. (18) Fox, A.; Krahmer, M.; Harrelson, D. J. Microbiol. Methods 1996, 27, 129138. (19) Fox, A.; Fox, K.; Christensson, B.; Harrelson, D.; Krahmer M. Infect. Immunol. 1996, 64, 3911-3915. (20) Saraf, A.; Larsson, L. J. Mass Spectrom. 1996, 31, 389-396. (21) Zenie, F. Laboratory Robotics, Handbook; Zymark, Hopkinton, MA, 1988.
Optimization of the preparation of alditol acetates and their analysis by GC/MS analysis has been reviewed elsewhere. The procedure allows profiling of the whole cell carbohydrate profiles of bacteria using GC/MS. Identification of bacterial cells allows chemotaxonomic discrimination or determination of physiological status.5-13 Analysis of bacterial sugars in complex environmental and clinical matrixes is optimally performed using gas chromatography/tandem mass spectrometry, after alditol acetate derivatization.17-20 Development of the automated derivatization instrument has greatly simplified the existing method. Computer control enables a process to take place unattended so that it can proceed during the day or night. For example, a robotic system, the Zymark robot (Zymark Inc., Hopkinton, MA), is widely used.21 Samples are moved by a robotic arm from module to module and multiple operations, at different x-y-z coordinates, can be programmed into the instrument. This instrument is extremely flexible, but considerable setup time may be required. This is suitable for large-scale industrial operations. The instrument described here leaves samples fixed in space, delivers solvents and gases, and controls vacuum and heat. While this does limit flexibility, setup time is minimized and routine analysis simplified. The current instrument is suitable for small laboratories. Automation of the alditol acetate procedure demonstrated here could encourage widespread application of the method and, indeed, other complex derivatization methods outside of specialist analytical laboratories. ACKNOWLEDGMENT This work was supported by the Army Research Office (Grant DAAH04-95-1-0359).
Received for review October 20, 1998. Accepted February 2, 1999. AC981155G
Analytical Chemistry, Vol. 71, No. 9, May 1, 1999
1917
