Infrared Multiphoton Dissociation with a Hollow Fiber Waveguide

Ibis Therapeutics, a Division of Isis Pharmaceuticals, 2292 Faraday Avenue, Carlsbad, California 92008. Anal. Chem. , 2003, 75 (15), pp 3669–3674...
0 downloads 0 Views 94KB Size
Anal. Chem. 2003, 75, 3669-3674

Accelerated Articles

Infrared Multiphoton Dissociation with a Hollow Fiber Waveguide Jared J. Drader, James C. Hannis, and Steven A. Hofstadler*

Ibis Therapeutics, a Division of Isis Pharmaceuticals, 2292 Faraday Avenue, Carlsbad, California 92008

A novel scheme for performing infrared multiphoton dissociation (IRMPD) is presented in which a hollow fiber waveguide (HFWG) is used to transmit IR radiation into the ion storage region of a mass spectrometer. Efficient dissociation of oligonucleotide and protein ions is demonstrated on an ESI-FTICR instrument in which IRMPD is performed in the external ion reservoir and on a quadrupole ion trap. Using a simple optical scheme consisting of a single focusing lens and an x, y translator, the 10.6-µm IR laser beam, initially 3.5 mm in diameter, is focused into the vacuum-sealed HFWG. The small internal diameter and the high transfer efficiency of the waveguide allow IR radiation of high power density to be employed for IRMPD. In studies performed on a quadrupole ion trap, a 500-µm-i.d. waveguide was used as a medium to transmit IR radiation directly through a 700µm orifice in the ring electrode. Efficient IRMPD of both a 12-mer oligonucleotide and the protein melittin were performed at laser powers of 0.5 and 3.2 W, respectively. Following early fundamental work by Beauchamp and coworkers1 and later work by Eyler and co-workers,2,3 infrared multiphoton dissociation (IRMPD) has slowly evolved into a robust and widely employed method to dissociate biomolecular ions prior to interrogation by mass spectrometry. McLafferty and co-workers demonstrated the applicability of IRMPD for the efficient fragmentation of proteins4 and oligonucleotides5 in the trapped ion * To whom correspondence should be addressed. Phone: (760) 603-2599. E-mail: [email protected]. (1) Bomse, D. S.; Berman, D. W.; Beauchamp, J. L. J. Am. Chem. Soc. 1981, 103, 3967-3971. (2) Baykut, G.; Watson, C. H.; Weller, R. R.; Eyler, J. R. J. Am. Chem. Soc. 1985, 107, 8036-8039. (3) Watson, C. H.; Baykut, G.; Battiste, M. A.; Eyler, J. R. Anal. Chim. Acta 1985, 178, 125-136. (4) Little, D. P.; Speir, J. P.; Senko, M. W.; O’Connor, P. B.; McLafferty, F. W. Anal. Chem. 1994, 66, 2809-2815. 10.1021/ac030157k CCC: $25.00 Published on Web 06/25/2003

© 2003 American Chemical Society

cell of an FTICR. Subsequent work by that group demonstrated that gentle IR heating in the trapped ion cell can serve as an efficient postionization “cleanup” step which “boils off” noncovalently associated adducts, resulting in improved spectral quality for heavily adducted analytes.6 Recently, it has been shown that electron capture dissociation (ECD)7 of intact proteins can induce cleavage of the amide backbone in which the fragments remain noncovalently bound in the gas phase; gentle IR heating was shown to effectively dissociate such fragment aggregates. Although now widely employed 4,5,8-10 and commercially available on FTICR instruments, IRMPD on quadrupole ion trap (QIT) mass spectrometers has been slower to catch on owing to the higher operating pressure inside the trap and the concomitantly greater rate of collisional deactivation. As described in detail by McLuckey and Goeringer,11 the relatively high pressure of helium buffer gas inside the QIT can result in equivalent rates of ion activation and deactivation, resulting in a steady-state parent ion internal energy distribution which is below the dissociation threshold. Yost and co-workers have developed alternative QIT IRMPD schemes based on manipulating the rates of ion activation and ion cooling. One scheme involves an alternative trap design based on previous work by Eyler and co-workers with a modified FTICR cell.12 In this case, the ring electrode was modified with spherical concave mirrors mounted on the inner surface of the ring providing eight laser passes across the radial plane of the (5) Little, D. P.; Aaserud, D. J.; Valaskovic, G. A.; McLafferty, F. W. J. Am. Chem. Soc. 1996, 118, 9352-9359. (6) Little, D. P.; McLafferty, F. W. J. Am. Soc. Mass Spectrom. 1996, 7, 209210. (7) Kruger, N. A.; Zubarev, R. A.; Horn, D. M.; McLafferty, F. W. Int. J. Mass Spectrom. 1999, 185/186/187, 787-793. (8) Tonner, D. S.; McMahon, T. B. Anal. Chem. 1997, 69, 4735-4740. (9) Hofstadler, S. A.; Sannes-Lowery, K. A.; Griffey, R. H. Anal. Chem. 1999, 71, 2067-2070. (10) Li, W.; Hendrickson, C. L.; Emmett, M. R.; Marshall, A. G. Anal. Chem. 1999, 71, 4397-4402. (11) McLuckey, S. A.; Goeringer, D. E. J. Mass Spectrom. 1997, 32, 461-474. (12) Watson, C. H.; Zimmerman, J. A.; Bruce, J. E.; Eyler, J. R. J. Phys. Chem. 1991, 95, 6081-6086.

Analytical Chemistry, Vol. 75, No. 15, August 1, 2003 3669

ion cloud and was shown to increase the dissociation efficiency.13 In subsequent work, the incorporation of a dual-pulsed delivery system of He buffer gas into the experiment sequence allowed a reduced He pressure during IR irradiation while the second pulse of He prior to ion detection provided collisional focusing of the fragment ions for improved detection efficiency.14 Most recently, Glish and co-workers noted that IRMPD efficiency was significantly enhanced when the trap was operated at an elevated temperature.15 For example, at a trap temperature of 25 °C, no dissociation was observed following the irradiation of protonated pentapeptide for 700 ms at 50 W. The same molecular ion yielded significant fragment ions when irradiated for 100 ms at only 25 W when the trap temperature was raised to 160 °C. We recently demonstrated that performing IRMPD in the external ion reservoir of an ESI-FTICR instrument at relatively high pressure (>10-6 Torr) yielded improved fragment ion abundance and sequence coverage for oligonucleotide ions relative to in-cell IRMPD at low pressure (1000 W at 10.6 µm. Optical Configuration. A 25-W CW CO2 laser (Synrad Inc., Mukilteo, WA) operating at 10.6 µm was used for all IRMPD experiments. Laser power was measured using a Synrad PowerWizard. Optical rails (6 mm × 6 in., Thor labs, Newton, NJ) were mounted directly on the faceplate of the laser collinear with the axis of the laser beam. A single 5-in. focal length, 1.1-in. diameter, positive meniscus ZnSe IR-compatible focusing lens (Laser Power Corp., Murrieta, CA) was mounted on the optical rails with a standard 1-in. lens mount. The waveguide vacuum interface assembly (Figure 1a), which was mounted within an x, y translator stage, consisted of an aluminum beam aperture with an orifice diameter equal to the inside diameter of the HFWG (500 or 1000 µm), and a 13-mm diameter by 2-mm-thick BaF2 window (SaintGobain Crystals & Detectors, Solon, OH). A vacuum chamber was created between the BaF2 window and the HFWG by compressing a Viton O-ring against the BaF2 window and compressing a second Viton O-ring around the exterior of the HFWG. The exit end of the HFWG entered the mass spectrometer through another vacuum interface created by compressing a Viton O-ring against the exterior of the HFWG. Alignment of the focused laser beam and the entrance to the HFWG was accomplished by adjusting the focusing lens along the optical rails to the focal point (f 5 in.) with respect to the entrance of the HFWG and using the x, y manipulator to adjust the position of the waveguide entrance with respect to the laser beam. Instrumentation. FTICR. The external ion reservoir IRMPD experiments were performed on a Bruker Daltonics (Billerica, MA) Apex II 70e Fourier transform ion cyclotron resonance mass spectrometer. The spectrometer was equipped with an Analytica (Branford, CT) electrospray source utilizing a grounded ESI emitter, a glass desolvation capillary with countercurrent drying gas, a single skimmer cone, an rf-only hexapole ion guide, and a

processing, and plotting, were performed using Bruker XMASS version 6.0 software running on a dual processor Pentium PC. Quadrupole Ion Trap. A modified Bruker Daltonics (Billerica, MA) Esquire 3000 running software version 5.0 was used for all QIT studies. To allow direct observation of the waveguide and trap, the aluminum top plate of the vacuum manifold was replaced with a clear polycarbonate cover into which the compressed O-ring vacuum feedthrough was mounted. The ring electrode was modified to accommodate a HFWG mounting fixture that aligned the waveguide orthogonal to the ion injection axis and provided alignment with the single 700-µm through-hole drilled in the ring electrode (Figure 1c). The IR laser was triggered from an Agilent (Palo Alto, CA) 33120A arbitrary waveform generator that was initialized after ion accumulation by the “ExCon1” signal of the Esquire 3000. To compensate for the additional capacitance in the high voltage rf circuit brought about from the close proximity of the waveguide to the ring electrode, 11 turns were removed from the secondary rf coil. To reduce the pressure in the trap during IRMPD, a needle valve was installed in-line with the He buffer gas to control the flow of buffer gas. Although it was not possible to directly measure the pressure of the He buffer gas inside the trap, relative changes were tracked by measuring the indicated pressure in the primary vacuum manifold.

Figure 1. (a) Schematic representation of the HFWG-IRMPD coupling approach used in this work. A single focusing lens and x, y manipulator are mounted on a set of small optical rails that mount directly to the faceplate of the CO2 laser. (b) Schematic showing the orientation of the 1-mm waveguide as interfaced to the external ion reservoir of the ESI-FTICR instrument. (c) Schematic showing the orientation of the 500-µm-i.d. waveguide as interfaced to the quadrupole ion trap through a 700-µm hole drilled in the ring electrode.

gate electrode. The external ion reservoir consisted of a biased skimmer cone, an rf-only hexapole ion guide, and a gate electrode at the low-pressure end of the hexapole ion guide. An ∼40-cm length of the 1-mm-i.d. waveguide was coupled to the hexapole accumulation region, as described above. The exit of the waveguide was aligned such that the beam passed directly through the hexapole, orthogonal to the ion-axis, so as not to directly impinge on the hexapole elements (Figure 1b). Ions were accumulated in the external ion reservoir (concurrent with IR irradiation for the IRMPD experiments) for 0.5-1 s and pulsed into the Infinity trapped ion cell for analysis by FTICR. The oligonucleotides were analyzed in the negative mode with the potential applied to the capillary exit maintained at 10 W. Dissociation of protein ions was less straightforward, even at full laser power 3672

Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

(∼33 W), and was accomplished only at very high ion numbers that were just below the threshold of the onset of multipole storage-assisted dissociation (MSAD)24,25 or at reduced pressure in the hexapole. These studies suggested that higher power densities would induce protein fragmentation. Furthermore, although aligning the laser to traverse the hexapole was rather straightforward, readjustment of the alignment was required following vacuum chamber maintenance (in which the vacuum chamber is rolled out of the magnet). Thus, there were several motivating factors behind the decision to attempt to couple the HFWG to the hexapole ion reservoir. In this work, a 1-mm-i.d. HFWG was brought into the region of the vacuum chamber containing the external ion reservoir such that it was orthogonal to the axis of the ion optics (Figure 1b), resulting in the beam making a single pass across the longitudinal axis of the ion volume. HFWG-IRMPD studies with multiply charged oligonucleotide anions resulted in extensive fragmentation, comparable to that observed with the on-axis configuration, except that 5-fold lower irradiation power could be utilized. Additionally, initial experiments aimed at mapping the shape and location of the ion cloud in the external ion reservoir were unsuccessful because efficient dissociation was observed at all positions employed (data not shown). These data suggest that the ion cloud is at a dynamic equilibrium that traverses the length of the external ion reservoir irradiated by the HFWG. The IRMPD spectrum in Figure 2b was acquired following the irradiation of (24) Sannes-Lowery, K.; Griffey, R. H.; Kruppa, G. H.; Speir, J. P.; Hofstadler, S. A. Rapid Comm. Mass Spectrom. 1998, 12, 1957-1961. (25) Sannes-Lowery, K. A.; Hofstadler, S. A. J. Am. Soc. Mass Spectrom. 2000, 11, 1-9.

Figure 3. IRMPD spectrum of the isolated (M - 4H+)4- charge state of a 12-mer DNA oligonucleotide acquired using the 500-µm-i.d. waveguide coupled to a quadrupole ion trap, as shown in Figure 1b. An 83-ms laser pulse at 0.5 W was shown to effectively dissociate the molecular ion when the He pressure was reduced in the trap (see text). Note that since the 0.5-W beam which initially had a 3.5-mm diameter (0.05 W/mm2) was focused and transmitted through the 500-µm waveguide, the effective power density in the trap was increased to 2.6 W/mm2 (∼49-fold increase relative to the unfocused beam).

the external ion reservoir at 3.5 W for 1 s concurrent with the 1-s ion accumulation event. Note that significant fragmentation is observed consistent with previous hexapole IRMPD studies using the coaxial scheme.16 Although the terminus of the HFWG was in close proximity to the hexapole elements, no change in hexapole performance was observed. HFWG IRMPD in a Quadrupole Ion Trap. The motivation for coupling the IR laser to the QIT via the HFWG is perhaps even more compelling than that of the external ion reservoir. While several researchers have demonstrated IRMPD on a QIT, the technique has not gained widespread acceptance (or commercial support) owing to the intrinsic difficulty in overcoming the collisional inactivation brought about by the very high operating pressure inside the QIT, typically 1 × 10-3 Torr He buffer gas. Thus, in addition to operating at reduced trap pressures, the approaches employed to date use high laser power (up to 50 W). A key attribute of the HFWG is that it can deliver very high irradiance at relatively low total power. For example, previous work by Glish and co-workers employed a 50-W CO2 laser in which the unfocused 3.5-mm beam traversed the QIT through two 5-mm holes in the ring electrode, resulting in a net power density of about 5.2 w/mm2. By prefocusing the beam and transmitting it through a 500-µm-i.d. HFWG, the power density of a 1-W input is 5.1 W/mm2. Note that a if a 50-W beam were focused through the 500-µm or 300-µm waveguides, the resulting power density of the beams would be ∼255 and 707 W/mm2, respectively. The physical coupling of the HFWG to the QIT, illustrated in Figure 1c, is very straightforward. The ring electrode was modified with a 5-mm hole drilled 1.5 cm radially into the ring electrode. A 700-µm hole was drilled through one side of the ring electrode with a drill bit mounted on a 5-mm drill bushing that kept the hole centered through the remaining 2 mm of the ring electrode. Since the HFWG is conductive, a Teflon guide was used to center the HFWG in the 5-mm hole. The distance between the throughhole and the exit end of the waveguide was adjusted while under vacuum by loosening the compression O-ring in the polycarbonate lid and moving the waveguide to a position that prevented a

discharge between the high-voltage rf and the grounded HFWG. The 700-µm through-hole in the ring electrode is expected to induce minimal perturbation to the electric field inside the trap. Unfortunately, because the HFWG is conductive and in close proximity to the ring electrode, the interface changes the capacitance beyond the range of the trim capacitors in the highvoltage rf drive circuit. To roughly balance the trap circuitry, 11 turns were removed from the secondary coil of the rf drive circuit, reducing the mismatch to a range which could be balanced by the tunable trim capacitors. By removing the turns from the secondary coil, the drive voltage was proportionally decreased; thus, the m/z range of the instrument was reduced from 3000 to 2200. The additional capacitance induced by the HFWG should not pose a significant limitation to broad deployment of this approach because rf drivers can readily be redesigned with the additional power required to extend the range back to 3000 m/z. Despite the relatively high power densities delivered by the HFWG, it was found that dissociation efficiency was significantly improved when the pressure of the He buffer gas was reduced by a metering valve (not shown) such that the indicated pressure in the main a vacuum manifold was on the order of 1.2 × 10-5 Torr. At the normal operating pressure in the trap (∼1 mTorr), limited fragmentation was observed, and operation at higher power resulted in a significant increase in chemical noise, likely from desorption of low molecular weight species adsorbed on the ring electrode. The spectrum in Figure 3 illustrates the efficiency of the approach with IRMPD of the isolated (M - 4H+)4- charge state a 12-mer DNA oligonucleotide. What is particularly striking about this spectrum is the fact it was generated with a laser power of only 0.5 W and an irradiation interval of 83 ms. While the laser output was only 0.5 W of power, the power density delivered to the trapped ion cell was ∼2.6 W/mm2 owing to the focal properties of the lens and the 500-µm i.d. of the HFWG. It has been previously shown that IRMPD of proteins requires more power for efficient dissociation compared to oligonucleotides 5,26 because the amide backbone is refractory to dissociation as compared to the relatively fragile phosphodiester oligonucleotide Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

3673

Figure 4. IRMPD spectrum of bee venom melittin acquired using the 500-µm-i.d. waveguide coupled to a quadrupole ion trap, as shown in Figure 1c. The spectrum in (a) was acquired under normal MS conditions, while the IRMPD spectrum in (b) was acquired following a 62.5-ms laser pulse at a power of 3.2 W (16.3 W/mm2).

backbone. Initial attempts at IRMPD of proteins with the HFWG have been successful only at reduced trap pressures. An example is shown in Figure 4, in which bee venom melittin (MWmono ) 2845.7 Da) is effectively dissociated with a laser power of only 3.2 W applied over a 63-ms irradiation interval. A series of b and y ions (Figure 4b) is generated, which provides 44% sequence coverage. As expected, the parent ions of higher charge dissociate to a greater extent, as observed by the disappearance of the 5+ charge state and the altered distribution of the 4+ and 3+ charge states. Charge-stripping is also indicated by the appearance of the 2+ charge state following IRMPD. Several of the peaks are readily defined as due to H2O loss, whereas others are consistent with internal fragments. In addition to serving as an effective means of inducing IRMPD, the favorable transmission characteristics and mechanical flexibility of the HFWG may prove to have additional utility in the field of analytical mass spectrometry. It may be possible to use the high power density transmitted through the HFWG to perform IR MALDI at relatively low laser power. Direct heating of components, such as trapped ion cells or gas inlets, might be easily accomplished by employing the HFWG, for example, heating and volatilizing nonvolatile samples for electron impact ionization. The HFWG might serve as a convenient means of bringing IR radiation into the trapped ion cell of an FTICR instrument in which optical access to the central axis is blocked by hardware required for electron capture dissociation. Preliminary attempts at using the HFWG to effect more efficient desolvation of nanospray generated droplets is presently underway in our lab. CONCLUSIONS In this work, we have demonstrated a novel approach for performing infrared multiphoton dissociation in which a hollow (26) Hofstadler, S. A.; Griffey, R. H.; Pasa-Tolic, L.; Smith, R. D. Rapid Comm. Mass Spectrom. 1998, 12, 1400-1404.

3674 Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

fiber waveguide is used to transmit coherent infrared radiation into the vacuum chamber of a mass spectrometer. This approach was demonstrated on both the external ion reservoir of an external source FTICR instrument and on a quadrupole ion trap mass spectrometer. Both oligonucleotides and protein ions were effectively dissociated at relatively low laser powers as the laser beam focused into the waveguide has a significantly higher power density than the unfocused beam. Oligonucleotides are dissociated in the quadrupole ion trap at a laser power of