Anal. Chem. 2004, 76, 4118-4122
Analytical Performance of a Venturi Device Integrated into an Electrospray Ionization Fourier Transform Ion Cyclotron Resonance Mass Spectrometer for Analysis of Nucleic Acids Adam M. Hawkridge,† Li Zhou,§ Milton L. Lee,§ and David C. Muddiman*,†,‡
W.M. Keck FT-ICR Mass Spectrometry Laboratory, Mayo Proteomics Research Center, and Department of Biochemistry and Molecular Biology, Mayo Clinic College of Medicine, Rochester, Minnesota 55905, and Department of Chemistry and Biochemistry, Brigham Young University, Provo, Utah 84602-5700
A voltage-assisted venturi device modeled after an industrial air amplifier was used to improve the ion transmission efficiency of a 16.2 kDa oligonucleotide and a 53mer PCR product in the high-pressure region between an electrospray ionization (ESI) emitter and the sampling orifices of two Fourier transform ion cyclotron resonance mass spectrometers (FT-ICR-MS). The venturi device increased the total ion abundance of the oligonucleotide and the PCR product by more than 6-fold relative to the best achievable signal without the device. Furthermore, the average charge states of the oligonucleotide and PCR product shifted from 12.5- to 14.5- and 10.9- to 12.6-, respectively, with the addition of the venturi device. Specific to FT-ICR mass spectrometry, this increase in the charge state directly translates to an increase in theoretical mass resolving power (>10 000 full width half-maximum for the results presented here at 7 T). Adduction was still observed while using the device, suggesting that it is “soft” relative to other high-pressure ion focusing methods. The development of electrospray ionization (ESI) for the mass spectral analysis of biomolecules by Fenn et al.1 has proven to be an invaluable technique for studying biological systems. One of the key attributes of ESI is its exceptionally high ionization efficiency (approaching unity). Under optimum ESI conditions, modern mass spectrometers (MS) can realistically achieve femtoto attomole limits of detection.2-6 However, there is significant potential for lowering the detection limits even further considering * To whom correspondence should be addressed: Medical Sciences Building 3-115, Mayo Clinic College of Medicine, 200 First Street SW, Rochester, MN 55905. Phone: 507-284-1997. Fax: 507-284-9261. E-mail: muddiman.david@ mayo.edu. † Mayo Proteomics Research Center, Mayo Clinic College of Medicine. ‡ Department of Biochemistry and Molecular Biology, Mayo Clinic College of Medicine. § Brigham Young University. (1) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64-70. (2) Valaskovic, G. A.; Kelleher, N. L.; McLafferty, F. W. Science 1996, 273, 1199-1202. (3) Valaskovic, G. A.; Kelleher, N. L.; Little, D. P.; Aaserud, D. J.; McLafferty, F. W. Anal. Chem. 1995, 67, 3802-3805.
4118 Analytical Chemistry, Vol. 76, No. 14, July 15, 2004
that the ion transmission efficiency between the ESI emitter and the inlet of the mass spectrometer is only 0.01-0.1%.7,8 Loss of more than 99% of the potential signal underscores the need for methods to focus more ions into the MS. Several approaches have been taken to improve ion transmission efficiencies in ESI-MS instruments with varying degrees of success.9-21 Focusing electrosprayed ions, especially multiply charged biomolecules, without inducing fragmentation has proven to be much more difficult at high pressures (e.g., atmospheric pressure) than at reduced pressures (e.g., 99.9%), 2-propanol (>99.9%), and water (HPLC grade) were purchased from Burdick & Jackson (Muskegon, MI) and used as received. Imidazole (99.0%), piperidine (>99.5%), and ammonium acetate (>98.0%) were purchased from Sigma-Aldrich (St. Louis, MO) and used as received. Venturi Device. Figure 1 shows a schematic of the venturi device used in this study, which is a modified version of the air amplifier used in the initial study.20 The body of the venturi device was fabricated from 304 stainless steel. A dielectric spacer (Ultem, GE Plastics, Berea, CA) was placed on the entrance of the device to create a voltage gradient along the z-axis (defined as the magnetic field axis of the FT-ICR-MS). To reduce charging, an antistatic insert (Semitron, Esd 520HR, Quadrant Engineering Plastic Products, Reading, PA) was incorporated in the venturi inlet. A Viton O-ring was used to seal the threaded joint between (22) Liu, C. L.; Wu, Q. Y.; Harms, A. C.; Smith, R. D. Anal. Chem. 1996, 68, 3295-3299. (23) Null, A. P.; George, L. T.; Muddiman, D. C. J. Am. Soc. Mass Spectrom. 2002, 13, 338-344. (24) Null, A. P.; Nepomuceno, A. I.; Muddiman, D. C. Anal. Chem. 2003, 75, 1331-1339. (25) Mangrum, J. B.; Flora, J. W.; Muddiman, D. C. J. Am. Soc. Mass Spectrom. 2002, 13, 232-240.
Figure 1. Schematic of the voltage-assisted venturi device.
the stainless steel body and venturi exit cone insert. The heated metal capillaries from both FT-ICR-MS instruments were inserted 28 mm into the exit cone of the venturi device. The voltage settings for the venturi entrance and venturi exit were held at -1000 and -500 V, respectively. The nitrogen pressure was maintained between 35 and 45 psi for all experiments. Figure 2 shows the schematic representation of the modified Analytica of Branford ESI source on the Ionspec FT-ICR (vide infra). Three separate sensors were used to measure the pressure at each stage within the ESI source with and without the venturi device. Stage 1 (purple region) is the nozzle-skimmer region, stage 2 (green region) is the intermediate pressure region (i.e., the skimmer cone region), and stage 3 (blue region) is the reduced pressure region. The pressure in stage 1 was measured with a MKS Baratron pressure transducer (Type 627 B). The pressure sensor was sampled from inside the ESI source to provide the highest measurement accuracy. A thermocouple pressure transducer (Duniway Stockroom, DST-531) was used in stage 2 (green) and a nude ion gauge filament (Duniway Stockroom, 1-100-K) was used in stage 3 to measure the respective pressures. The table inset in Figure 2 shows the pressures in each stage with and without the venturi device. All three stages showed slight decreases in pressure as a result of the reduced pressure region near the center of the venturi device entrance from which the MS inlet samples. Mass Spectrometry. Mass spectra were obtained with two ESI-FT-ICR mass spectrometers. Both the oligonucleotide and 53mer PCR product were electrosprayed in the negative-ion mode by direct infusion (Harvard Apparatus PHD 2000 syringe pump) at a flow rate of 0.5 µL/min. The distance-dependent data were obtained with a modified IonSpec (Irvine, CA) ESI-FT-ICR-MS equipped with an actively shielded 7 T superconducting magnet (Cryomagnetics, Oak Ridge, TN). The ESI source (Analytica of Branford, CT) was modified to accept a heated metal capillary (0.062 in. o.d. × 0.030 in. i.d. stainless steel) and a dual ESI source.26 The emitter tips used were 360 µm o.d. and 50 µm i.d. and tapered to 15 ( 1.0 µm i.d. (New Objective, Woburn, MA). The potentials applied to the emitter and heated metal capillary were -2000 and -125 V, respectively. All spectra were acquired with 1024 k points at an ADC rate of 1 MHz, a Blackman window function applied, and zero-filled three times prior to fast Fourier transformation to the frequency domain. The PCR product was analyzed with a modified Bruker Daltonics (Billerica, MA) FT-ICR-MS equipped with an actively (26) Nepomuceno, A. I.; Muddiman, D. C.; Bergen, H. R.; Craighead, J. R.; Burke, M. J.; Caskey, P. E.; Allan, J. A. Anal. Chem. 2003, 75, 3411-3418.
Analytical Chemistry, Vol. 76, No. 14, July 15, 2004
4119
Figure 2. Schematic of the modified Analytica of Branford ESI source with each general pressure stage indicated by purple (stage 1), green (stage 2), and blue (stage 3).
shielded 12 T superconducting magnet (Magnex Scientific Limited, Oxford, UK). The Apollo API ESI source (Bruker Daltonics) was modified to accept a heated metal capillary (0.062 in. o.d. × 0.020 in. i.d. stainless steel) and dual ESI source adapted from Nepomuceno et al.26 The ESI emitters used were the same as the distance study (vide supra) and the emitter and heated metal capillary voltages were -1700 and -150 V, respectively. All spectra were acquired with 1024 k points at an ADC rate of 1 MHz, and a sin-bell window function was applied and zero-filled twice prior to fast Fourier transformation to the frequency domain. RESULTS AND DISCUSSION Increase in Ion Abundance Observed. Figure 3 shows ESIFT-ICR mass spectra of a 53-mer oligonucleotide obtained using the venturi device as a function of the ESI emitter distance from the sampling orifice. Two general observations follow regarding the ion abundances and the charge-state distributions for these data. However, it is first important to note that as the ESI emitter distance is increased (venturi mode), adduction to the oligonucleotide is also increased, suggesting that the venturi device is a “soft” ion collection device, which broadens the applicability of the source (e.g., noncovalent complexes). Figure 4a shows the ion abundance data as a function of distance of the ESI emitter with respect to the orifice of the mass spectrometer; five replicate measurements were made at each distance examined, and the trend was found to be reproducible. These data were generated using only the un-adducted molecular ion abundances and were normalized for charge state as we have previously reported.27 The mean ion abundance obtained using our standard operating configuration (i.e., without the use of the venturi device) is denoted in Figure 4a by the arrow (vide infra). A one-way analysis of variance (ANOVA) was carried out to test the null hypothesis that the mean amplitude is equal at all distances (i.e., that differences observed among these means are (27) Gordon, E. F.; Mansoori, B. A.; Carroll, C. F.; Muddiman, D. C. J. Mass Spectrom. 1999, 34, 1055-1062.
4120
Analytical Chemistry, Vol. 76, No. 14, July 15, 2004
due to random error). The null hypothesis was rejected (p < 0.05). To determine where significant differences occurred, two-sample t-tests, assuming unequal variances, were performed between data from adjacent distances (i.e., 1 vs 2 mm, 2 vs 3 mm, 3 vs 4 mm, etc.). Table 1 lists the results of these analyses; if the absolute value of tcalc > tcrit, the null hypothesis is rejected, indicating that the means are statistically different at the 5% level. It should be noted that the degrees of freedom for the tcrit values were computed using the Satterthwaite approximation due to the fact that the variances were unequal. These data clearly indicate that the changes in ion abundance are real and not attributable to random error. Combined with the mean ion abundance data, we can conclude that the optimum range of the venturi device is from 8 to 10 mm. The data obtained without the assistance of the venturi device (N ) 5) resulted in a mean ) 0.122 and a standard deviation of 0.031 (% RSD ) 25%). In comparison, the data obtained with the venturi device operating at 10 mm resulted in a mean ) 0.761 and a standard deviation of 0.055 (% RSD ) 7%), a significant improvement. To be complete, a two-sample t-test was carried out to determine if the ion abundance without the venturi device was equal to that obtained with the device at 8 and 10 mm; the data from 8 and 10 mm were used to represent the abundances achievable with the venturi device since the difference in the means at these two distances could be accounted for by random error alone (vide supra). The tcalc and tcrit values obtained from this analysis were -25.752 and 2.160, respectively, clearly indicating that the >6-fold observed improvement in ion abundance is real. Increase in the Average Charge State Observed. Figure 4b shows the average charge state observed as a function of distance of the ESI emitter with respect to the orifice of the mass spectrometer; five replicate measurements were made at each distance examined and the trend was found to be reproducible. Figure 3 qualitatively illustrates this effect.
Figure 3. Single-acquisition negative-ion ESI-FT-ICR mass spectra of a 53-mer oligonucleotide as a function of distance of the ESI emitter from the orifice of the heated metal capillary. Note the shift in the charge-state distribution to higher charge states as the distance is increased from 1 to 10 mm.
A one-way analysis of variance (ANOVA) was carried out to test the null hypothesis that the mean charge state is equal at all distances (i.e., that differences observed among these means are due to random error). The null hypothesis was rejected (p < 0.05). To determine where significant differences occurred, two-sample t-tests assuming unequal variances were performed between data from adjacent distances (i.e., 1 vs 2 mm, 2 vs 3 mm, 3 vs 4 mm, etc.). Table 1 lists the results of these analyses; if the absolute value of tcalc > tcrit, the null hypothesis is rejected, indicating that the means are statistically different at the 5% level. These data indicate that, over the distance interval of 10-16 mm, the observed changes in the mean charge state are not statistically significant, while over the distance interval of 2-10 mm, the observed change is real. The increase in the average charge state observed through the implementation of the venturi device can make biological macromolecules more amenable to dissociation. Moreover, due to the fact that FT-ICR is uniquely suited for top-down sequencing of nucleic acids, proteins, and other biopolymers, mass resolving power greatly benefits these types of analyses for both the intact (i.e., ensuring the presence of only one species, affording more accurate mass measurements) and the dissociated species (direct
Figure 4. (a) Normalized27 ion abundance and (b) charge state and theoretical mass resolving power as a function of distance of the ESI emitter from the orifice of the mass spectrometer. The arrow in Figure 4a shows the mean ion abundance obtained without the venturi device. The shaded box indicates where the maximum ion abundance was achieved, which also coincides with the maximum average charge state observed. Table 1 ion abundance
charge state
distance interval (mm)
tcalc
tcrit
tcalc
tcrit
1-2 2-4 4-6 6-8 8-10 10-12 12-14 14-16
-2.648 12.500 -7.191 -5.339 -1.099 2.561 5.598 3.079
2.365 2.365 2.777 2.306 2.306 2.447 2.365 2.365
0.9488 -11.268 -12.085 -7.254 -4.377 -1.143 -1.470 -0.156
2.447 2.365 2.447 2.365 2.365 2.447 2.571 2.306
charge-state determination). The theoretical mass resolving power achievable in FT-ICR at the zero-pressure limit is given by
RPFWHM )
(1.274 × 107)zB0T m
(1)
where z is the number of charges, B0 equals the magnetic strength in Tesla, T is the transient length in seconds, and m is the molecular mass in Da. Inspection of this relationship demonstrates the inverse relationship between resolving power and the m/z ratio. Figure 4b also shows the theoretical mass resolving power as a function of distance of the ESI emitter with respect to the orifice of the mass spectrometer for a 1-s time domain at 7 T. These calculations used the average charge state observed for z Analytical Chemistry, Vol. 76, No. 14, July 15, 2004
4121
Adaptation of the Venturi Device to a 12 T ESI-FT-ICRMS. Figure 5 shows the effect of the venturi device on the analysis of a 53-mer PCR product on a different FT-ICR-MS platform. Figure 5a shows the mass spectrum of the PCR product without the assistance of the venturi device; this was the best signal that could be obtained for a wide variety of experimental conditions tested. However, with the venturi mode on (Figure 5b) and keeping the experimental sequence identical to that used to obtain the spectrum in Figure 5a, the PCR product signal increased by 6.5-fold and the average charge state shifted from 10.9- to 12.6-. These results provide further evidence for the general applicability of the venturi device with nearly identical improvements across MS platforms. Clearly, the incorporation of a venturi device can improve the observed ion abundances by nearly an order of magnitude. Specific to FT-ICR, higher charge states are observed in the venturi mode, which directly translates into increased mass resolving power. Future studies include modeling of the electrostatics/gas dynamics, defining its utility for top-down proteomics, noncovalent complexes, and liquid separations coupled with mass spectrometry.
Figure 5. Single-acquisition negative-ion ESI-FT-ICR mass spectrum of a PCR product (a) without and (b) with the assistance of the venturi device. The insets show the -10 charge state with the primer (*), coding strand (C), noncoding strand (NC), and the coding strand minus one nucleotide (this is not neutral base loss). The absolute abundance (arbitrary units) of selected peaks is given in parentheses. CSavg ) average charge state of the coding strand.
in eq 1. Thus, the increase in the average charge state directly translates into higher mass resolving power. Specifically, the increase in the theoretical mass resolving power afforded by the venturi device is over 10 000 (full-width half-maximum) under these conditions at 7 T. Note the large working distance available (>6 mm) to achieve this enhancement in mass resolving power. Importantly, the increase in the mass resolving power coincides with the improved ion abundance (see gray box in Figure 4).
4122
Analytical Chemistry, Vol. 76, No. 14, July 15, 2004
CONCLUSIONS We have integrated a venturi device into two ESI-FT-ICR mass spectrometers and have demonstrated an enhancement in the ion abundance by almost an order of magnitude and an increase in the theoretical mass resolving power due to detection of higher charge states; the latter may also prove useful for top-down sequencing of biological macromolecules. We are continuing to exploit these attributes as well as improve our understanding of the device. ACKNOWLEDGMENT We thank Dr. Ann Oberg, Division of Biostatistics, Mayo Clinic, for her assistance, Angelito Nepomuceno for preparing the oligonucleotide and PCR product, Dr. Ryan Danell for assistance with the 12 T FT-ICR data, and Dr. Jason W. Flora for his assistance with artwork. The authors are grateful for the financial support of the National Institutes of Health (R01HG02159), the W.M. Keck Foundation, and the Mayo Clinic College of Medicine.
Received for review February 28, 2004. Accepted April 27, 2004. AC049677L
