Anal. Chem. 1997, 69, 2525-2529
A MALDI Probe for Mass Spectrometers John D. Lennon, III, and Gary L. Glish
Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290
Matrix-assisted laser desorption/ionization (MALDI)1 mass spectrometry has become a powerful technique for the analysis of a wide variety of compounds. This ionization technique has primarily been coupled to time-of-flight mass spectrometers, but over the last few years a considerable amount of research has been devoted to coupling MALDI to the various other mass spectrometers.2-4 This effort has been aimed at obtaining the more favorable characteristics possessed by some of the other mass spectrometers such as higher mass resolution, higher mass accuracy, and the ability to more easily perform tandem mass spectrometry, (MS/MS). The vast majority of experiments have performed MALDI in top illumination geometry, i.e. the irradiating laser and mass analyzer are on the same side of the sample. In top illumination geometry the analyte/matrix sample is typically
deposited on a metallic substrate. However, it has been demonstrated5-7 that a metallic substrate is not necessary and as a result, transmission geometry can be used. In transmission geometry, the analyte/matrix sample is deposited on a transparent substrate and is irradiated from the direction opposite of the analyzer. One of the difficulties encountered when coupling MALDI to some of the various mass spectrometers is the inaccessibility of the ion source regions. Compact lens or obstructed views can make implementing MALDI in top illumination geometry arduous. Often the laser has to be focused through very shallow angles or through narrow passageways. This difficulty is exhibited in quadrupole ion trap mass spectrometers in which it is often desirable to generate ions directly in the trapping volume, which reduces transmission loss that can be encountered from external ion generation and injection. Desorbing MALDI ions directly into the trapping volume using top illumination geometry has been accomplished by two methods. The first method2 requires a small hole to be drilled radially through the ring electrode. The laser is then focused through this hole and impinges onto the sample that has been introduced through a hole on the other side of the ring electrode to the inner boundary. The other method8 focuses the laser through the narrow gap between the ring and entrance endcap electrodes onto a sample target brought to the inner boundary of the entrance endcap electrode. Both of these methods require somewhat stringent focusing and alignment. The work presented here describes a new MALDI probe that utilizes transmission geometry to overcome the aforementioned difficulties. The design allows MALDI ions to be generated in very confined locations. This new probe is fashioned after typical solids probes used for electron (EI) and chemical ionization (CI) in that it is introduced into the vacuum in the same manner and accommodates a sample holder on the end. The sample holder in this case is a solid quartz cylinder. A fiber optic is automatically aligned by the probe body to irradiate the sample from behind, obviating the need for stringent focusing and alignment. This MALDI probe was designed to desorb ions directly into the ion volume of a quadrupole ion trap. However, it should be adaptable for use with other mass spectrometers. While this report describes results obtained with a quadrupole ion trap we have
(1) Karas, M.; Bachmann, D.; Bahr, U.; Hillenkamp, F. Int. J. Mass Spectrom. Ion Processes 1987, 78, 53-68. (2) Chambers, D. M.; Goeringer, D. E.; McLuckey, S. A.; Glish, G. L. Anal. Chem. 1993, 65, 4-20. (3) Castoro, J. A.; Ko¨ster, C.; Wilkins, C. L. Rapid Commun. Mass Spectrom. 1992, 6, 239-241. (4) Hill, J. A.; Annan, R. S.; Biemann, K. Rapid Commun. Mass Spectrom. 1991, 5, 395-399.
(5) Vertes, A.; Balazs, L.; Gijbels, R. Rapid Commun. Mass Spectrom. 1990, 4, 228-231. (6) Ehring, H.; Costa, C.; Demirev, P; Sundqvist, B. Rapid Commun. Mass Spectrom. 1996, 10, 821-824. (7) Schu ¨renberg, M.; Schulz, T.; Dreisewerd, K.; Hillenkamp, F. Rapid Commun. Mass Spectrom. 1996, 10, 1873-1880. (8) Doroshenko, V. M.; Cornish, T. J.; Cotter, R. J. Rapid Commun. Mass Spectrom. 1992, 6, 753-757.
A new MALDI probe has been designed that uses transmission geometry. This geometry allows the probe to be fashioned after typical EI/CI solid probes which enables it to be introduced into spatially constrained ion source regions such as encountered in quadrupole ion trap mass spectrometers. In the probe design demonstrated here, light from a fiber optic irradiates the backside of a sample through a small piece of quartz on which the sample has been directly deposited. The performance characteristics exhibited by utilizing this probe for MALDI on a quadrupole ion trap mass spectrometer are similar to those which can be obtained through the traditional methods of implementing MALDI. Spectra have been obtained from 50 fmol of total loading of bombesin, MS/MS has been performed on 5 pmol of des-Arg9-bradykinin, and the analyte ion signal is shown to last for over 2500 laser shots for 2 pmol of bombesin. Optical micrographs showing the crystal distribution of a sample containing 2 pmol of bombesin have been obtained as a function of the number of laser shots for a single sample loading. Although this probe was designed for use with the quadrupole ion trap, it can be adapted for use with all types of mass spectrometers. Thus, with only one laser, one fiber optic, and this probe, MALDI can be performed on multiple instruments in a lab.
S0003-2700(96)01291-7 CCC: $14.00
© 1997 American Chemical Society
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997 2525
Figure 1. Transmission MALDI probe (not to scale).
obtained some preliminary results using this same probe with a sector instrument. With one laser and one fiber optic, MALDI can be performed on multiple instruments in a lab. This paper will describe the design of the probe, show results to date, and discuss some of the factors involved with performing MALDI in this manner on a quadrupole ion trap.
EXPERIMENTAL SECTION The mass spectrometer used for these experiments was a modified Finnigan MAT quadrupole ion trap. A 0.125 in. hole has been added radially to the ring electrode to allow the insertion of the MALDI probe. The laser, a Continuum Surelite II (Santa Clara, CA) Nd:YAG laser with the frequency tripled wavelength of 355 nm, was triggered by a TTL pulse from the quadrupole ion trap electronics, which was controlled by ICMS Ion Trap Software version 2.20.9 The transmission MALDI probe design is shown in Figure 1. The design was fashioned after typical EI/CI solids probes in that it is introduced into the vacuum in the same manner and accommodates a sample holder on the end. The main body of the probe consists of a 1/2 in. diameter o.d. stainless steel tube. A silica fiber optic (3M Specialty Optical Fibers, West Haven, CT) with a core diameter of 600 µm, a full acceptance angle of 19°, and currently about 8 m long is vacuum sealed through the center of this stainless steel tube via a 1/16 in. Swagelok fitting using a Teflon ferrule. The Teflon ferrule allows the fiber optic to be adjusted during analysis. Due to the divergent nature of the laser beam emitted from the fiber optic, the power density can be varied by adjusting the distance between the fiber and sample. The end of the probe is composed of a hollow stainless steel tip, which is electrically isolated from the probe body by a Lexan connector. A Lexan tip was initially used instead of this stainless steel tip and produced similar results; however, stainless steel was chosen for its greater structural integrity. The stainless steel tip grips a small cylinder of quartz that is 2 mm diameter and approximately 3 mm long. The tip was machined so that only a portion has an outer diameter (0.110 in.) small enough to fit into the 0.125 in. hole in the ring electrode. The remainder has a larger outer diameter which acts as a stop, only allowing enough of the tip to enter to bring the quartz sample holder just to the inner boundary of the ring electrode. (9) ICMS Ion Trap Software ver. 2.20, courtesy of Nathan Yates and Richard Yost at the University of Florida Department of Chemistry, Gainesville, FL, 1992.
2526
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
The quartz sample holder is cleaved from a longer rod by scoring and snapping. It is possible to obtain visibly flat surfaces on the quartz tip; however, rough spots often occur. These rough spots can act as nucleation sites for crystal growth. It has not yet been investigated as to whether this helps, hinders, or has no effect on ion production. The analyte/matrix mixture is deposited directly onto one end of the quartz cylinder. The quartz tip can be reused multiple times, though a fresh section was used for low analyte loading during detection limit measurements. The scan function utilized to obtain the spectra shown here involves trapping of ions at an rf level above the matrix ions’ mass as well as accumulating ions from multiple laser shots. Generating ions directly in the trapping volume of the quadrupole ion trap allows ions to be more easily trapped at rf levels above the matrix ions’ masses, whereas ions generated externally often possess an optimum rf level for trapping that is below the masses of the matrix ions. Accumulation of MALDI-generated ions from multiple laser shots has been demonstrated previously10 to improve ion intensity while matrix ions were excluded with the application of SWIFT excitation. Though not utilizing the SWIFT technique, it was observed that accumulating ions from as many as 100 or more laser shots improved the ion intensity. However, accumulating this many shots per scan reduces the number of scans that can be averaged before the sample is consumed. The matrix used for all spectra was R-cyano-4-hydroxycinnamic acid (4-HCCA) prepared in 7:3 acetonitrile/water to a concentration of 10 mM with 1% trifluoroacetic acid (TFA). The analytes that were used for the spectra shown were bombesin (MW ) 1619.9) and des-Arg9-bradykinin (MW ) 904.0). Both compounds were obtained from Sigma Chemical Co. (St. Louis, MO) and used with no additional purification. Samples were prepared by first depositing 0.5 µL of matrix onto the quartz tip and allowing it to completely dry. Then 0.5 µL of the analyte solution was deposited followed by an immediate deposition of another 0.5 µL of the matrix solution for cocrystallization. The typical procedure for collecting spectra after the probe has been introduced into the vacuum is as follows. Before any laser shots, the fiber optic is first retracted a sufficient distance from the back of the sample holder so that the irradiating laser intensity is below threshold for desorption. At this point the data acquisition is started and the fiber optic is moved closer to the probe tip until ions are observed. The distance from the end of the fiber optic to the probe tip is typically 4-6 cm, depending upon the laser power at the input of the fiber optic. After some number of laser shots, when the ion intensity decreases, the fiber is moved closer to the tip, increasing the ion intensity again. This can be repeated until the sample is exhausted. When the sample is completely gone the fiber is typically 2-3 cm away from the quartz. During the entire process the quartz tip is fully illuminated by direct as well as an unknown amount of reflected light. Considering a 5 cm starting distance and the acceptance angle, the diameter of the laser beam irradiating the sample, if there were no obstructions, would be around 1.7 cm. Therefore, the sample holder (2 mm diameter) is directly irradiated by only 1.4% of the area of the laser beam. To perform MS/MS, CID was effected with argon as the collision gas along with the helium bath gas. Argon was pulsed (10) Doroshenko, V. M.; Cotter, R. J. Rapid Commun. Mass Spectrom. 1996, 10, 65-73.
Figure 2. Transmission MALDI spectrum (one scan, 25 laser shots) of 50 fmol of total loading of bombesin.
into the vacuum system for 1500 µs. After a 15 ms delay, an excitation frequency near the calculated resonant frequency corresponding to the q value of the desired parent ion was applied to the endcap electrodes for 60 ms at 700 mVp-p for the experiment presented here. A Nikon Microphot-fx optical microscope (Melville, NY) with a Sony CCD Iris digital video camera Model SSC-C374 was used to obtain micrographs of a sample of bombesin at various times during a single analysis. The analyte/matrix was deposited on the quartz as usual. The mass analysis was stopped periodically, and the quartz was removed from the probe tip to obtain the micrographs. The probe tip was then returned and mass analysis was continued with the same sample and with the fiber optic in the same position as it was left prior to microscopic imaging.
RESULTS AND DISCUSSION To demonstrate the utility of this transmission probe, characteristics such as detection limit, ability to perform MS/MS, and sample lifetimes were investigated. Two common peptides, bombesin and des-Arg9-bradykinin, were chosen as model compounds. The matrix, 4-HCCA, was chosen because it has been shown to give good results for a variety of peptides. The smallest quantity of analyte that has been reproducibly observed using the transmission MALDI probe is 50 fmol of total loading of bombesin. A mass spectrum from a single scan accumulating ions from 25 laser shots per scan is shown in Figure 2. Analyte ions were present for at least 10 scans (250 laser shots) each time the experiment was performed. Lower detection limits for similar molecular weight peptides have been demonstrated using top illumination geometry MALDI on another quadrupole ion trap.2 However, it is difficult to make a direct comparison between the ultimate detection limit of transmission geometry versus top illumination geometry from different instruments. Different scan functions, pressures, etc. may lead to differences in trapping efficiencies, detection efficiencies, etc. In the future, as top illumination geometry MALDI is implemented on our instrument, a direct comparison can be made. One of the advantages of using a quadrupole ion trap for MALDI versus the traditional TOF instruments is the former’s much more powerful MS/MS capabilities. The most common method of performing MS/MS on the quadrupole ion trap is via CID using resonant excitation. The helium bath gas (at a pressure of ∼1 × 10-3 Torr) is typically used as the CID target gas. However, performing CID in this manner on MALDI-generated
Figure 3. MS/MS spectrum of 5 pmol of des-Arg9-bradykinin using Transmission MALDI (average of two scans, 150 laser shots).
ions is problematic due to the shot-to-shot variation of the ion intensity. This variation changes the number density of trapped ions and therefore changes the secular frequency of the ions due to space charge. This can substantially reduce the MS/MS efficiency when using a fixed single resonant excitation frequency. Accumulating ions from a number of laser shots tends to average out some of this intensity variation, but does not satisfactorily solve the problem. A method of performing MS/MS on MALDIgenerated ions in the quadrupole ion trap that has been recently demonstrated11 involves the use of heavy gases as the CID target gas. Heavy gases deposit more internal energy into the parent ion per collision, and thus enough energy can be deposited to induce dissociation even when the secular frequency is slightly shifted from the applied frequency.12 Heavy gases were used to obtain MS/MS spectra using the transmission MALDI probe. An MS/MS spectrum of 5 pmol of total loading of des-Arg9-bradykinin is shown in Figure 3. This spectrum is an average of two scans accumulating ions from 75 shots per scan. More fragmentation than shown in Figure 3 can be induced by several means. A heavier gas can be used, the length of the Ar pulse can be extended, and the delay time between the Ar pulse and excitation frequency can be varied.12 This is, however, a typical quadrupole ion trap MS/MS spectrum of des-Arg9-bradykinin. To investigate sample consumption, optical micrographs were obtained at periodic intervals during analysis of 2 pmol of bombesin. Parts a-f of Figure 4 show the initial analyte/matrix distribution on the quartz surface and the distribution following 100, 600, 1100, 1600, and 2775 laser shots after the first observation of bombesin ions, respectively. The signal intensity remained fairly constant over the course of the analysis, with an occasional adjustment of the fiber optic, until a few hundred laser shots before the last image when the sample was exhausted. This is typical of sample lifetimes that are obtained with this MALDI probe. An attempt was made to obtain the image in the same location on the quartz cylinder. What is thought to be the same aggregate of crystals, though some rotation from image to image has occurred, is circled in each figure. Figure 4a shows that there is an even distribution of crystals formed across the surface. A large portion of the surface is uncovered which allows the laser beam to irradiate the sides as well as the bottom of the individual crystals. The remaining parts (11) Doroshenko, V. M.; Cotter, R. J. Anal. Chem. 1996, 68, 463-472. (12) Vachet, R. W.; Glish, G. L. J. Am. Soc. Mass Spectrom. 1996, 7, 11941202.
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
2527
Figure 4. Optical micrographs of a sample of 2 pmol of bombesin deposited on a quartz tip (a) before laser irradiation, (b) 100, (c) 600, (d) 1100, (e) 1600, and (f) 2775 laser shots after ions were first observed.
b-f of Figure 4 show a slow uniform progression of desorption of the crystals across the surface, which may indicate that the entire surface is experiencing relatively the same power density. Schu¨renberg et al. recently performed a preliminary study7 on desorption mechanisms of transmission geometry MALDI coupled to a delayed extraction time-of-flight mass spectrometer. Their findings show that ions are being generated at the sample holder/ sample interface and are extracted with thermalized velocities. In the absence of extraction potentials, as is encountered with the quadrupole ion trap, it is unclear whether the ions generated at the sample holder/sample interface would desorb off of the sample and be trapped. We believe that our results as well as Schu¨renberg, et al.s indicate that it is not necessary to invoke a different mechanism for analyte ion formation with this transmission geometry compared to the top illumination geometry.
CONCLUSIONS The work presented here as well as elsewhere has demonstrated that MALDI does not need to be performed on a metallic 2528
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
substrate and that a transmission geometry can be used. Therefore, a probe can be designed, such as the one described here, that allows MALDI to be implemented in spatially constrained ion source regions. One of the benefits of this probe is that the stringent laser focusing and alignment requirements that are often encountered in spatially constrained ion source regions are not necessary. The fiber optic, which is used to direct the light onto the sample, is automatically aligned with the sample and can be readily varied in distance from the sample to vary the laser intensity striking the sample. Another advantage of this approach is that with just one fiber optic, a laser, and this probe, MALDI can be performed on multiple instruments in the lab. The performance characteristics of transmission MALDI using this probe are favorable. The lowest level of detection, though not as low as has been reported for traditional MALDI on a quadrupole ion trap, is in the tens of femtomole range. A direct comparison between transmission MALDI and top illumination MALDI can not be made at this time on this instrument, so it is unknown whether the reported detection limit is affected by the quartz substrate or if other instrumental parameters are involved.
MS/MS has also been obtained using this probe with 5-pmol of des-Arg9-bradykinin. Shot-to-shot variation of analyte ion intensity, common to MALDI in general, makes performing MS/MS via resonant excitation using He as the collision gas difficult. Ar was utilized in these experiments as a collision gas to aid in the fragmentation. Typical sample lifetimes that can be obtained irradiating the sample in this manner is over 2500 laser shots. It is also believed that it is not necessary to invoke a different mechanism for analyte ion formation with this transmission geometry compared to top illumination geometry.
ACKNOWLEDGMENT We thank Theo Dingemans for his help obtaining the optical micrographs. This work was supported by NIH GM 49852. Received for review December 30, 1996. April 15, 1997.X
Accepted
AC961291Q X
Abstract published in Advance ACS Abstracts, June 1, 1997.
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
2529
