Automated Sampling and Imaging of Analytes Separated on Thin

Olga S. Ovchinnikova , Vilmos Kertesz , and Gary J. Van Berkel ... Griffin K. Barbula , Matthew D. Robbins , Oh Kyu Yoon , Ignacio Zuleta and Richard ...
0 downloads 0 Views 333KB Size
Anal. Chem. 2006, 78, 4938-4944

Automated Sampling and Imaging of Analytes Separated on Thin-Layer Chromatography Plates Using Desorption Electrospray Ionization Mass Spectrometry Gary J. Van Berkel* and Vilmos Kertesz

Organic and Biological Mass Spectrometry Group, Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6131

Modest modifications to the atmospheric sampling capillary of a commercial electrospray mass spectrometer and upgrades to an in-house-developed surface positioning control software package (HandsFree TLC/MS) were used to enable the automated sampling and imaging of analytes on and within large area surface substrates using desorption electrospray ionization mass spectrometry. Sampling and imaging of rhodamine dyes separated on TLC plates were used to illustrate some of the practical applications of this system. Examples are shown for user-defined spot sampling from separated bands on a TLC plate (one or multiple spots), scanning of a complete development lane (one or multiple lanes), or imaging of analyte bands in a development lane (i.e., multiple lane scans with close spacing). The post data acquisition processing and data display aspects of the software system are also discussed. Desorption electrospray ionization mass spectrometry (DESIMS) is rapidly developing as a surface sampling/ionization source for the interrogation of a wide variety of analytes on a broad range of surfaces under ambient conditions.1-3 Most analytes amenable to analysis by ES-MS appear to be amenable to analysis by DESIMS, and in many cases, little if any preparation of the surfaces needs to be performed to obtain at least a qualitative identification of the analytes present. Some of the demonstrated application areas for DESI-MS involve the analysis of pharmaceutical and illicit drug tablets,1,4-9 plant and animal tissues,1,10,11 and thin-layer chromatography * Corresponding author. Phone: 865-574-1922. Fax: 865-576-8559. E-mail: [email protected]. (1) Taka´ts, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471-473. (2) Taka´ts, Z.; Wiseman, J. M.; Cooks, R. G. J. Mass Spectrom. 2005, 40, 12611275. (3) Cooks, R. G.; Ouyang, Z.; Taka´ts, Z.; Wiseman, J. M. Science 2006, 311, 1566-1570. (4) Chen, H.; Talaty, N. N.; Taka´ts, Z.; Gooks, R. G. Anal. Chem. 2005, 77, 6915-6927. (5) Williams, J. P.; Scrivens, J. H. Rapid Commun. Mass Spectrom. 2005, 19, 3643-3650. (6) Rodriguez-Cruz, S. E. Rapid Commun. Mass Spectrom. 2006, 20, 53-60. (7) Leuthold, L. A.; Mandscheff, J.-F.; Fathi, M.; Giroud, C.; Augsburger, M.; Varesio, E.; Hopfgartner, G. Rapid Commun. Mass Spectrom. 2006, 20, 103110.

4938 Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

(TLC) plates.12 In most of these and other DESI-MS reports, the sample surfaces in question were interrogated with manual sample manipulation. In a few cases, results were presented that pointed to the possibility of using DESI-MS for automated high-throughput analysis or spatial chemical imaging of analytes on and within the surfaces. For example, Cooks and co-workers showed that with tablet samples mounted on a variable-speed moving belt, sample analysis throughput as high as 2-3 tablets/s could be achieved with selected ion monitoring.4 In the very first paper on DESIMS,1 the Cooks group showed an intensity versus position plot indicating the variable distribution of γ-coniceine across a 6-mm portion of a Conium maculatum stem. This plot was derived from the manually acquired spot samples across the stem. In another case, Cooks’ group in collaboration with Caprioli’s group demonstrated the manual analysis of different areas of tissue sections.10 In one example, 20 consecutive spots at 1-mm intervals were sampled across a section of metastatic human liver from the healthy region to a cancerous region. They found that the intensity distribution of particular sphingomyelin species distinguished the nontumor and tumor regions of the tissue. Our group demonstrated the first automated DESI-MS experiments to distinguish the spatial distribution of analytes on a surface.12 In that work, DESI-MS with computer-controlled plate movement was used to accomplish the scanning of a single development lane on a TLC plate to detect various dyes and pharmaceuticals separated on normal-phase and hydrophobic and wettable reversed-phase plates. In the present report, we expand on that first automated DESI-MS work. Modifications were made to the sampling capillary of a commercial ES-MS instrument and to the software controlling the x/y/z stage holding the sample to be interrogated as well as the image-generating data processing package, both of which were developed in-house (HandsFree (8) Kauppila, T. J.; Wiseman, J. M.; Ketola, R. A.; Kotiaho, T.; Cooks, R. G.; Kostiainen, R. Rapid Commun. Mass Spectrom. 2006, 20, 387-392. (9) Weston, D. J.; Bateman, R.; Wilson, I. D.; Wood, T. R.; Creaser, C. S. Anal. Chem. 2005, 77, 7572-7580. (10) Wiseman, J. M.; Puolitaival, S. M.; Taka´ts, Z.; Cooks, R. G.; Caprioli, R. M. Angew. Chem., Int. Ed. 2005, 44, 7094-7097. (11) Talaty, N.; Taka´ts, Z.; Cooks, R. G. Analyst 2005, 130, 1624-1633. (12) Van Berkel, G. J.; Ford, M. J.; Deibel, M. A. Anal. Chem. 2005, 77, 12071215. 10.1021/ac060690a CCC: $33.50

© 2006 American Chemical Society Published on Web 06/13/2006

Figure 1. Structure and mass-to-charge ratio for the rhodamine dyes.

TLC/MS).12,13 With these modifications in place, we demonstrate the use of DESI-MS in computer-controlled spot sampling of bands on a TLC plate, scanning of multiple development lanes on a TLC plate, and imaging of analyte bands in a development lane. More broadly, these results illustrate the ability to automate a DESIMS system to sample and determine the spatial distribution of analytes on and within any of a number of analytically important surfaces. EXPERIMENTAL SECTION Materials and Reagents. HPLC grade methanol was obtained from Burdick & Jackson (Muskegon, MI). Ammonium acetate (99.999%) was obtained from Sigma Aldrich (Milwaukee, WI). Rhodamine B (1) and rhodamine 6G (2) were purchased from Eastman Kodak Company (Rochester, NY), and rhodamine 123 (3) was obtained from Sigma Aldrich (Figure 1). Standard stock solutions were prepared for TLC and ES-MS detection optimization by dissolving the dyes in methanol at a concentration of 0.2 ng/ µL (450 fmol/µL for rhodamines B and 6G and 580 fmol/µL for rhodamine 123). TLC. TLC was carried out using hydrophobic Merck RP C8 plates (P/N 13725/5, EM Science, Gibbstown, NJ). Rhodamine standards were manually spotted on the plates in 1.0-µL aliquots using a 1.0-µL disposable pipet (Alltech, Deerfield, IL). Plates were developed in ascending mode in a saturated chamber with 75/25 (v/v) methanol/water containing ∼500 mM ammonium acetate. Developed plates were air-dried for at least 30 min prior to mass spectrometric analysis. Fluorescence images of the plates were taken using a Versadoc fluorescence imaging system (BioRad Laboratories, Hercules, CA) operating in UV transillumination mode with 5-s exposure time, 1.0× gain, and 1 × 1 binning. DESI-MS. The DESI-MS setup with manual and computercontrolled x/y/z sample stage is shown in Figure 2. The mass spectrometer was a ThermoFinnigan LCQ Deca ion trap (Thermo(13) Kertesz, V.; Ford, M. J.; Van Berkel, G. J. Anal. Chem. 2005, 77, 71837189.

Finnigan, San Jose, CA) operated using Xcalibur version 1.3 software. Analysis of large-area surfaces required that the atmospheric sampling heated capillary be extended out from the main body of the instrument. To accomplish this, a shallow counter bore (∼1 mm) was drilled into the heated sampling capillary, and into this was press-fit a 7.1-cm-long, 21-RW-gauge stainless steel tube extension (508 µm i.d., 813 µm o.d., Scientific Instrument Services, Inc. Ringoes, NJ). In addition to the press fit, the thermal expansion of the capillaries under normal operation of the source sealed the extension in place. DESI mass spectra were acquired in positive ion mode with a heated capillary temperature of 250 °C, 0- or 4.0-kV spray voltage applied to the DESI emitter, 12-V capillary voltage, and 15-V tube lens offset. These and other optimum detection parameters were determined by ES-MS analysis of the analytes in question using the DESI emitter probe with the appropriate positioning as the ES emitter. Note that optimum signal levels for the rhodamine dyes often could be achieved with the DESI emitter voltage set to 0. The ionization process in this case appears to be related to that observed in sonic spray ionization.14,15 This phenomenon in DESI-MS has been noted by others using different analytes2 and requires further fundamental explanation that is beyond the scope of the present study. The DESI emitter used for these experiments was identical to that used in our prior report.12 The spray emitter was a 5.2-cmlong, taper-tip, fused-silica capillary (50 µm i.d., 360 µm o.d., New Objective, Woburn, MA). The inner diameter of the nebulizing gas tube was 500 µm, providing a nebulizing gas (nitrogen) jet annulus area of ∼1.5 × 10-7 m2. The ES emitter was mounted from 1 to 4 mm from the surface to be analyzed at an ∼50° angle to the surface. The nebulizer gas flow rate was set between 2.8 and 3.3 L/s for the various experiments (285-380 m/s nebulizing gas jet linear velocity). The ES high voltage was applied to the stainless steel body of the microionspray head. The ES solvent was delivered to the emitter by a syringe pump using a 2.5-mL glass syringe. A grounded union was placed in the transfer line ( ∼40 cm of 100 µm i.d. (1/16 in. o.d.) Teflon tubing) between the syringe pump and the DESI emitter. For initial positioning and mass spectral detection optimization, the surface position was controlled with the joystick (x/y) and jog wheel (z surface to DESI emitter axis) of the manual control unit of the x/y/z sample platform (see below). Signal levels were optimized when the surface was positioned so that the bottom of the sampling capillary was 100 µm or less above the surface and ∼2 mm back from the position where the DESI plume jet impacted the surface. To accomplish the positioning of the sampling capillary above the surface, a black and white CCD camera and associated monitor (Protana A/S, Odense, Denmark) was used to observe the DESI emitter and sampling capillary from an angle just above the surface. A color webcam (Creative Technology, Ltd., Milpitas, CA) was mounted directly over the surface to be analyzed and centered on the DESI emitter and sampling capillary. The webcam image was output to the screen of the HandsFree TLC/MS software running on the mass spectrometer PC. This bird’s eye view of the surface was used for manual and computercontrolled positioning of the surface in the x/y plane. Samples to (14) Hirabayashi, A.; Sakairi, M.; Koizumi, H. Anal. Chem. 1994, 66, 45574559. (15) Hirabayashi, A.; Sakairi, M.; Koizumi, H. Anal. Chem. 1995, 67, 28782882.

Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

4939

Figure 2. Schematic illustration of the DESI-MS setup showing side and top views. Inset shows rhodamine B band on TLC plate sampled by multiple spot sampling and one lane scan and a similar band sampled by multiple lane scans (raster-scanning with a 400-µm separation between lanes) as done for imaging.

be analyzed were fixed with double-sided tape to the top of a rectangular block made of an insulating material that in turn was attached to the top of the positioning stage. The surface could be illuminated with white light or light from a long-wavelength UV lamp. Automated Sampling and Data Processing. The MS2000 x/y/z robotic platform (Applied Scientific Instrumentation Inc., Eugene, OR) and the basic control software used to manipulate the TLC plate relative to the stationary DESI emitter have been described previously.12,13 The platform could be operated manually or by computer control in the x, y, and z planes for spot sampling. With computer control, it was possible to perform spot sampling (one or multiple spots), array sampling (spot sampling from prespecified array coordinates, not illustrated here), lane scanning (one or multiple lanes), or imaging (i.e., multiple lane scans with close spacing). The data collection module of the HandsFree TLC/ MS software package was upgraded and now can access and store the real-time mass spectral data being acquired by the mass spectrometer with the Xcalibur software. This made possible the display on the user interface of real-time ion maps and images of the surface being examined. 4940

Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

To enable detailed post-data acquisition processing of a data set, our software created a file with an “.IMA” extension (IMA to reflect the word “image”). This file stored all mass spectral data, the corresponding surface location data, and the time elapsed from the beginning of the analysis. When an IMA file was opened, the interrogation route (surface locations sampled) within a predefined area of the surface analyzed was displayed. The webcam image of the surface examined could be loaded with the interrogation path superimposed. By selecting an area of interest on the interrogation route, the respective mass spectra within that area were extracted from the IMA file and displayed for further processing (e.g., averaging). For (selected) ion map creation and imaging purposes, the user could define multiple m/z ratios or m/z ranges to plot in different colors. The color intensity displayed was calculated on the basis of the specified minimum and maximum intensities of the respective ion signal. More details on the data postprocessing module can be found in the Supporting Information section. Safety Considerations. The DESI emitter floats at the high ES voltage, and appropriate shields and interlocks should be used to avoid accidental contact with this component.

RESULTS AND DISCUSSION To illustrate the automated sampling and imaging capabilities of our DESI-MS setup, we chose to focus on the analysis of analytes separated on TLC plates. This application is a logical continuation of our prior work coupling TLC and MS using either our surface sampling probe approach (with ES ionization13,16-19 or atmospheric pressure chemical ionization (APCI)20) or DESIMS.12 Moreover, TLC remains an important analytical separation technology, and a simple and robust automated TLC/MS system does not exist. DESI-MS may offer one of the more useful TLC/ MS interfaces developed to date. Hydrophobic reversed-phase C8 plates were used in this work only because the rhodamine dyes used to illustrate the practical application of the system separated well on this phase. As previously shown,12 DESI-MS can also be used to analyze analytes separated on normal phase and other more commonly used TLC phases. Spot Sampling of Separated Analyte Bands on a TLC Plate. More often than not, the separated analyte bands on a TLC plate are visible in white or UV light. This allows the few points of interest on the plate to be readily discerned from the much larger area of the total development lane. In these cases, a selected spot-sampling mode of analysis may be applied, as illustrated by the data in Figure 3. Figure 3a is the webcam image of a developed TLC plate under UV illumination. Spotted in lanes 1, 2, and 3, respectively, were 200 pg of rhodamine B (1), rhodamine 6G (2), and rhodamine 123 (3). On the developed plate, one major component, as expected, was observed in each lane, each with a different color and different retention factor (RF) value. Each lane contained lesser amounts of one readily visible additional component at an RF value greater than that of the respective major component. For this experiment, the three major component bands and the three minor component bands were each chosen for interrogation (six bands total) with a sampling time of 20 s at each band. The bands chosen and the order in which they were analyzed are illustrated by the black “+” symbols and the red path, respectively, on the webcam image of the plate (Figure 3a). With our software interface, the areas of interest to interrogate were selected with a point and click at the webcam image of the surface displayed on the PC screen. Under the conditions used here, the spot size sampled from the surface with the DESI plume jet was roughly 400 µm in diameter (see Figure 2, top view and inset therein). When all sampling points were selected, the software mapped out an efficient route of travel between the selected points and offered the analyst the opportunity to alter the route. At the start of the automated experiment, the program moved the position of the plate so that the first sampling point was in the optimum sampling position along the line between the sampling capillary and the impact area of the DESI plume jet on (16) Van Berkel, G. J.; Sanchez, A. D.; Quirke, J. M. E. Anal. Chem. 2002, 74, 6216-6223. (17) Ford, M. J.; Van Berkel, G. J. Rapid. Commun. Mass Spectrom. 2004, 18, 1303-1309. (18) Ford, M. J.; Kertesz, V.; Van Berkel, G. J. J. Mass Spectrom. 2005, 40, 866875. (19) Ford, M. J.; Deibel, M. A.; Tomkins, B. A.; Van Berkel, G. J. Anal. Chem. 2005, 77, 4385-4389. (20) Asano, K. G.; Ford, M. J.; Tomkins, B. A.; Van Berkel, G. J. Rapid Comm. Mass Spectrom. 2005, 19, 2305-2312.

Figure 3. (a) Webcam image of the UV-illuminated RP C8 TLC plate showing the separated components in the respective development lanes in which were spotted 200 ng of rhodamine B (1, lane 1), rhodamine 6G (2, lane 2), and rhodamine 123 (3, lane 3). Black “+” symbols show the spots sampled, and the red line shows the route of the probe during the automated spot sampling. Summed, full-scan mass spectra collected for 20 s at each spot are shown for sampling spots (b) b, (c) c, (d) d, (e) e, (f) f, and (g) g. Spray solvent was methanol at a flow rate of 10 µL/min.

the surface. The surface was then raised from a distance of 3 mm below the sampling capillary to the ideal surface-to-sampling capillary distance (∼100 µm). At this point, full scan mass spectra were collected, summed, and displayed. Full-scan mass spectra were the preferred method to interrogate the bands for compound confirmation and determination of the mass-to-charge (an initial identification) ratio of the unexpected components. Following interrogation of a particular spot, the individual spectra and the summed spectrum were stored (