Rapid Peptide Mapping by Reversed-Phase Liquid Chromatography

A peptide map of horse heart myoglobin, completed in 3.5 min, is shown as an example of the results which can be obtained from combining this ... Cove...
4 downloads 0 Views 267KB Size
Download PDF
Anal. Chem. 1997, 69, 3973-3978

Rapid Peptide Mapping by Reversed-Phase Liquid Chromatography on Nonporous Silica with On-Line Electrospray Time-of-Flight Mass Spectrometry J. Fred Banks* and Erol E. Gulcicek

Analytica of Branford, Inc., 29 Business Park Drive, Branford, Connecticut 06405

Increased demands for faster analysis times have fostered the development of a number of new liquid chromatography (LC) techniques which can accelerate the chromatographic process and reduce the resources required for analytical investigations. In both academic and industrial venues, time has quickly become the critical element in research work. This trend is concurrent with some of the basic goals of separation science, i.e., providing the greatest allowable performance (efficiency) in the least amount of time possible. As an example, an excellent review on the topic of fast separations for proteins and peptides, in particular, has recently been published.1 One approach to achieving the goal of faster separations has been the development of “perfusion” or gigaporous supports, manufactured from polymeric bidisperse poly(styrene-divinylbenzene) particles.2-7 These materials have very large flowthrough pores of 6000-8000 Å as well as diffusion pores of 500-

1500 Å and have been demonstrated previously for use with fast separations and electrospray mass spectrometry (ES/MS) detection.8,9 The large flow-through pores allow convective flow to occur inside the particles and permit the use of greatly elevated mobile phase linear velocities. This result is particularly useful for the separation of larger biomacromolecules which suffer from slow diffusion in conventional porous silica matrices. Because of the greatly elevated volumetric flow rates used, 10-25 times higher than normal, this technique has been especially popular in the capillary or microcolumn LC format,8,9 where flow rates are naturally lower. The use of nonporous supports for LC stationary phases has also been under exploration for some time,10 but relatively recent improvements in particle production have allowed for the preparation of micropellicular spheres of 2 µm or less in diameter11,12 which are useful for the fast separation of proteins and peptides. Both polymeric and silica materials have been successfully developed for this application. While polymeric materials have the advantage of increased stability in alkaline pH, they have a more hydrophobic surface, may experience swelling from organic solvents, and often cannot endure pressures in excess of 1500 psi as can silica material. The reasoning for working with nonporous supports is the potential of increased column efficiency due to decreased mass transfer resistance in the stationary phase and a reduction of diffusional broadening owing to the lack of a pore structure. These advantages are especially pronounced with the use of smaller particles. Along with the reduced mass transfer resistance, elevated flow rates may be used with very little loss of column efficiency. However, the back-pressure can be substantial with these smaller nonporous particles, so often elevated temperatures are used to reduce solvent viscosity and take advantage of the “flattened” Van Deemter response at higher mobile phase velocities.13,14 Also, the use of very high operating pressures (>60 000 psi) has recently been demonstrated with nonporous silica (NPS) LC.15 Both perfusion and NPS LC are capable of producing complete separations in a matter of seconds with peak widths on the order of 1 s or less. This type of rapid separation places severe demands on detector response time. Because of the depth of information

(1) Chen, H.; Horva´th, C. J. Chromatogr. A 1995, 705, 3. (2) Lloyd, L. L.; Warner, F. P. J. Chromatogr. 1990, 512, 365. (3) Afayan, N. B.; Fulton, S. P.; Regnier, F. E. LC-GC 1992, 9, 824. (4) Afeyan, N. B.; Gordon, N. F.; Mazsaroff, I.; Varady, L.; Fulton, S. P.; Yang, Y. B.; Regnier, F. E. J. Chromatogr. 1990, 519, 1. (5) Afeyan, N. B.; Fulton, S. P.; Regnier, F. E. LC-GC 1991, 9, 824. (6) Afeyan, N. B.; Fulton, S. P.; Regnier, F. E. J. Chromatogr. 1991, 544, 267. (7) Fulton, S. P.; Afeyan, N. B.; Gordon, N. F.; Regnier, F. E. J. Chromatogr. 1991, 547, 452.

(8) Banks, J. F. J. Chromatogr. A 1995, 691, 325. (9) Kassel, D. B.; Shushan, B.; Sakumua, T.; Salzmann, J.-P. Anal. Chem. 1994, 66, 236. (10) Horvath, C.; Preiss, B. A.; Lipsky, S. R. Anal. Chem. 1967, 39, 1422. (11) Unger, K. K.; Jilge, G.; Kinkel, J. N.; Hearn, M. T. W. J. Chromatogr. 1986, 359, 61. (12) Kalghati, K.; Horvath, C. J. Chromatogr. 1987, 398, 335. (13) Chen, H.; Horvath, C. Anal. Methods Instrum. 1993, 1, 213. (14) Kalghati, K.; Horvath, C. J. Chromatogr. 1988, 443, 343. (15) McNair, J. E.; Lewis, K. C.; Jorgenson, J. W. Anal. Chem. 1997, 69, 983.

Reversed-phase liquid chromatography (LC) using a nonporous silica support has been combined with electrospray (ES) time-of-flight (TOF) mass spectrometry (MS) for the fast separation and mass detection of peptides. Using this LC method, the resolution of a peptide mixture can be completed is less than 35 s. The resulting chromatographic peak widths are less than 1 s wide. Because of the unique nature of a TOF mass analyzer, complete mass spectra can be acquired at a rate which is sufficient to sample these narrow peaks. When compared with conventional LC, the same separation takes nearly 20 min to complete, and the signal-to-noise ratio observed in the total ion chromatogram is dramatically lower due to the influence of increased background noise in the mass spectra. The limit of detection for a low molecular weight peptide, Val-Pro-Leu, was found to be 6 pmol with the total ion chromatogram and 500 fmol with the reconstructed ion chromatogram. A peptide map of horse heart myoglobin, completed in 3.5 min, is shown as an example of the results which can be obtained from combining this fast LC method with fast ES/TOF/MS detection capability.

S0003-2700(97)00226-6 CCC: $14.00

© 1997 American Chemical Society

Analytical Chemistry, Vol. 69, No. 19, October 1, 1997 3973

possible from mass spectrometry (MS) via electrospray (ES) ionization, MS is arguably the most desirable and powerful type of LC detection available. Traditional types of mass spectrometers based on quadrupole, magnetic sector, or ion-trap technology cannot acquire data in the time frame required to sample these very fast separations. This failure is due to fundamental behavior and is not resolvable with current technology. Because TOF/ MS operates in a substantially different mode compared to these other technologies, it is uniquely well-suited for the adequate sampling of rapid analyses. Indeed, TOF/MS technology has been previously developed for use in MS detection for fast GC separations.16 Ultimately, the TOF machine described in this study is capable of acquiring complete mass spectra at a rate of 8.2 kHz. These continuous spectra may then be summed for an arbitrary length of time, so that up to 100 complete mass spectra/s may be recorded. In the present study, we have explored the use of reversedphase LC using a NPS support for the fast separation and subsequent ES/TOF/MS detection of peptides. A peptide map was generated for equine myoglobin in 3.5 min to demonstrate the viability of this technique. EXPERIMENTAL SECTION Electrospray. The electrospray ionization source was from Analytica of Branford, Inc. (Branford, CT) and was similar to its original form.17 Large windows on three sides of the atmospheric region of the source allowed for the visualization of the electrospray process itself and greatly aided in the optimization of instrumental conditions necessary to obtain good results. The voltages in the source were set as follows: the cylinder electrode (Vcyl) was at -3800 V, the endplate electrode (Vend) at -4100 V, and the capillary entrance (Vcap) at -4480 V. The use of a dielectric capillary having a metallic coating on both ends to transfer the ions generated from atmosphere into vacuum is important in this case, as it allows for the LC column exit and ES spray needle to be maintained at ground potential while the internal source electrode voltages are maintained at high negative potential. ES source designs which depend on the application of a positive voltage to the spray needle itself and ground potential to the source electrodes have the disadvantage of the possibility of electrical shock to the user. Time-of-Flight Mass Spectrometry. From the exit of the ES source and subsequent ion guiding lenses, the ion beam is focused through a 2 mm aperture into the TOF analyzer chamber of the Wiley/McLaren type.18 The distance from the center of the extraction volume to the detector surface was 400 mm, and the drift energy in the field-free flight tube was 2700 eV for a singly charged ion. The continuous beam of ions from the ES source enters the first stage of the Wiley/McLaren accelerator orthogonally to the axis of the TOF spectrometer.19-23 In each scan cycle, a portion (16) Ji, Q.; Davenport, M. R.; Enke, C. G.; Holland, J. F. J. Am. Soc. Mass Spectrom. 1996, 7, 1009. (17) Banks, J. F.; Dresch, T. Anal. Chem. 1996, 68, 1480. (18) Wiley, W. C.; McLaren, I. H. Rev. Sci. Instrum. 1955, 26, 1150. (19) Mirgorodskaya, O. A.; Shevchenko, A. A.; Chernushevich, I. V.; Dodonov, A. F.; Miroshnikov, A. I. Anal. Chem. 1994, 66, 99. (20) Dodonov, A. F.; Chernushevich, I. V.; Laiko, V. V. Proceedings from the 12th International Mass Spectrometry Conference, Amsterdam, 1991. (21) Dawson, J. H. J.; Guilhaus, M. Rapid Commun. Mass Spectrom. 1994, 3, 155. (22) Verentchikov, A. N.; Ens, W.; Standing, K. G. Anal. Chem. 1994, 66, 126.

3974

Analytical Chemistry, Vol. 69, No. 19, October 1, 1997

of this beam is sent toward the detector by applying a voltage pulse to the repeller electrode. At high cycle repetition rates, this method renders a highly efficient way of adapting a continuous external ion source to the inherently pulsed TOF analyzer. Electrostatic steering deflectors were used to compensate for the effect of the initial orthogonal velocity component. The injection potential that corresponds to the initial perpendicular motion is constant and equal for all ions. Hence, the steering potential is independent of m/z values. The distribution of ion flight times at the end of the flight tube was measured with an ETP Model AF820 secondary electron multiplier (Auburn, MA). A mass resolution of R(fwhm) ) 1200 was determined for leucine enkephalin at m/z ) 556 using a Tektronix Model TDS 540 digital oscilloscope (Wilsonville, OR), operated at 500 MHz. For rapid and continuous analog-to-digital conversion and signal averaging, the output signal from the electrically floating ion detector was capacitively coupled into a Model 9825 integrating transient recorder (ITR) data acquisition device (200 MHz) from Precision Instruments (Knoxville, TN). During the experiments described here, the scan repetition rate of the TOF spectrometer was fixed at 8192 Hz. Thus, summing of 8192, 4096, 2024, ... scans/spectrum results in the acquisition of 1, 2, 4, ... spectra/s. At this scan rate, the dead time of the data acquisition system due to data transfer was 15%, and thus, to allow for precise reconstruction of ion chromatograms, a real-time label was attached to the data set of each spectrum. Liquid Chromatography. The LC system used was a 1090 Series II L from Hewlett-Packard (Palo Alto, CA), The mobile phase was taken directly from the high-pressure output pump using a 5 ft section of 0.005 in. i.d. PEEK tubing and connected to a Valco CI4W (Houston, TX) internal loop injection valve with a 1 µL internal volume rotor. This arrangement created a gradient delay time of 11.4 s, which was compensated for by delaying the sample injection time by the same value. The NPS reversed-phase column, 4.6 mm i.d. × 30 mm, was provided by Micra Scientific (Northbrook, IL). The conventional LC column was Microsorb MV C18 reversed-phase, 4.6 mm i.d. × 50 mm, 300 Å pore size, from Rainin Instruments (Woburn, MA). Chemicals. Buffers were prepared from distilled water obtained from a Barnstead NANOpure II (Boston, MA) system. Acetic acid, ULTREX II ultrapure reagent grade, was purchased from J. T. Baker, Inc. (Phillipsburg, NJ). Methanol, acetonitrile (ACN), and trifluoroacetic acid (TFA) were obtained from Mallinckrodt (Paris, KY). All solvents were filtered through nylon 66 membranes from Anspec Co. (Ann Arbor, MI). Peptide and protein samples were purchased from Sigma (St. Louis, MO). The trypsin digest of myoglobin was purchased from Michrome Bioresources (Auburn, CA). Procedures. The protein digest was subjected to a thorough cleaning procedure in order to remove surfactants and other contaminants which were found to contribute greatly to the background chemical noise. Myoglobin digest (500 pmol) was dissolved in 25 µL of 2% acetic acid. The sample was then centrifuged through a Microcon ion-exchange membrane system from Amicon (Beverly, MA). After the filtrate was discarded, the (23) O’Halloran, R. A.; Fluegge, R. A.; Betts, W. L.; Everette, W. L. Technical Report ASD-TDR 62-644, Prepared under Contract AF 33(616)-8374, The Bendix Corp. Research Laboratories Division, Southfield, MI, 1964.

Figure 1. Comparison of TICs from the separation of nine peptides (20 pmol each) using the NPS LC column at 1 mL/min flow rate with varying TOF/MS spectral acquisition rates. Gradient: 15-90% ACN, 0.1% TFA, in 1 min. Peaks: (1) Arg-Gly-Asp, (2) kyotorphin, (3) Val-ProLeu, (4) morphiceptin, (5) impurity, (6) methionine enkephalin, (7) bradykinin, (8) leucine enkephalin, (9) angiotensin II, and (10) angiotensin I. M/z range: 325-2000.

retained peptides were eluted from the filter disk by centrifuging again after addition of 50 µL of 1.0 M ammonium hydroxide in 1:1 methanol/water to the top of the filter disk. The resulting filtrate solution containing the peptides was then injected onto the LC column at the amount required. RESULTS AND DISCUSSION Effect of Integration Time. In order to test the performance of this system, a mixture of peptides was prepared and injected (20 pmol/component) onto a 4.6 mm i.d. × 3.0 cm column filled with the NPS LC adsorbent while a series of total ion chromatograms (TICs) were acquired from the TOF/MS at various integration times. “Integration time”, as used here, refers to the arbitrary grouping of scans recorded, since these scans are being acquired continually at the nominal rate of 8 kHz and can be grouped, or, more accurately, summed, in data registers for a length of time determined by the user. With the current data system, up to 100 individual summed mass spectra/s can be stored in this way. The data obtained for this experiment are shown in Figure 1. The TIC recorded at 0.5 spectra/s shows that the separation has been severely distorted by the integration time used. This spectral acquisition rate also represents the maximum speed that a quadrupole MS could be scanned over the same m/z region (3252000). As the integration time was decreased, the sampling reached 16 spectra/s, which was the minimum acquisition rate that could adequately represent the separation. The last peak visible (angiotensin I) is completely separated from the others and has a width of only 500 ms at the base. Note also that, by using the nonporous adsorbent and a fast gradient (15-90% ACN, 0.1% TFA in 1 min), the separation can be executed in less than 35 s with nearly complete resolution of all the components. Sensitivity. The dynamic range and limit of detection for this system were next determined by injecting increasing amounts of

Figure 2. S/N vs sample injection amount for Val-Pro-Leu. Conditions are the same as in Figure 1, with spectral acquisition at 16 spectra/s.

the peptide Val-Pro-Leu. The same chromatography conditions as used above were employed. The integration rate for TOF/ MS was 16 spectra/s, and signal-to-noise ratio (S/N) measurements were calculated from the peak height from triplicate injections and then averaged. Figure 2 shows a graph relating the observed S/N in both the TIC and reconstructed ion chromatogram (RIC) formats. The smallest quantity which could be detected in the TIC data was 6 pmol, while the smallest quantity which could be detected in the RIC data was 500 fmol, both at S/N ) 3. While not as good as the sensitivity available from CEES/TOF/MS in previous work17 or as capillary LC could, in theory, provide, these results are, nonetheless, impressive, considering the use of a 4.6 mm i.d. column at a liquid flow rate of 1 mL/min. The linear dynamic range was found to be 3 and 3.5 orders of magnitude in TIC and RIC modes, respectively. Analytical Chemistry, Vol. 69, No. 19, October 1, 1997

3975

The loss of linearity at higher concentrations is a function of the both the ES process and the loadability of the LC column. At a flow rate of 1 mL/min and a peak width of 1 s, the 1000 pmol injection produces a zone concentration of 60 pmol/µL. Previous studies24,25 have shown that electrospray signal can lose linearity above this level. Although, at first thought, concentrations such as these would not seem to be excessive, the actual solute concentration in solution at the time of ion desorption is substantially greater, since the ES droplets are evaporated during the ionization process. Thus, the next point on the graph, showing a 4000 pmol injection, has a reduced peak height but equal peak width. Finally, the last point, at 20 000 pmol, not only is reduced in height but also suffers from poor peak shape due to the column loadability. Nonetheless, the overall sensitivity for this system is attractive due to the narrowness of the chromatographic zone, and a substantial dynamic range may still be demonstrated. The signal observed in these experiemnts also suffers from at least a 20-fold loss in magnitude due to the well-known phenomena of ion signal suppression from the presence of TFA. Although the use of a postcolumn addition of 2-propanol and propionic acid has been demonstrated to restore the signal lost from TFA suppression,26 this technique was not used here. Comparison of Nonporous and Conventional LC. In the next series of experiments, the performance of both an NPS column and a conventional column was compared using the same nine-component peptide mixture. The conventional column had dimensions of 4.6 mm i.d. × 5.0 cm and was filled with a C18, 5 µm particle size, 120 Å pore size adsorbent. It was operated at the same flow rate as the NPS column, although a gradient with a more gradual slope was used. The NPS column was the same as that employed previously. Figure 3A shows the TIC collected from a separation of the nine peptides on the NPS column. At the amount of 20 pmol, each component is fully resolved and visible from the TIC data. The peak width of the final component, angiotensin I, is 500 ms. At a spectral recording rate of 16 spectra/s, this allows seven mass spectra to be saved while the peak elutes. Figure 3B shows the individual mass spectra recorded at the angiotensin I peak maximum. In contrast, Figure 4A shows the TIC produced from the separation of the same sample quantity but using the conventional LC column. The first and most obvious result seen here is the loss in ability to discriminate the eluting peaks in the TIC from the baseline noise. Also, the analysis time has increased to approximately 1150 s, or about 19 min, a reasonable result for conventional LC. The angiotensin peak, which could be located only by observing the RIC data, is now about 9 s wide, a typical, if not better than typical, peak width from a conventional 4.6 mm i.d. column. Because the peak width is wider, the TOF integration time was adjusted so that only 1 mass spectrum/s was recorded. This rate produced about the same data point density across the peak, eight complete mass spectra being collected, that had been used with the NPS experiment. Figure 4B shows the individual mass spectra recorded at the angiotensin I peak maximum. It is interesting to note that the abscissa or signal scales in Figures 3A and 4A are approximately the same. Examination of (24) Banks, J. F.; Whitehouse, C. W. In Methods in Enzymology; Karger, B. L., Hancock, W. S., Eds.; Academic Press: San Diego, CA, 1996; Vol. 270A. (25) Banks, J. F.; Whitehouse, C. W. Int. J. Mass Spectrom. Ion Processes 1997, 162, 163. (26) Kuhlmann, F. E.; Apffel, A.; Fischer, S. M.; Goldberg, G.; Goodley, P. C. J. Am. Soc. Mass Spectrom. 1995, 6, 1221.

3976 Analytical Chemistry, Vol. 69, No. 19, October 1, 1997

Figure 3. (A) TIC from the separation of the nine-component peptide mixture using NPS LC. Conditions and peak assignments are the same as in Figure 1. (B) Single mass spectrum from apex of angiotensin I peak. Data were acquired in 1/16 s.

the RIC data for the angiotensin I peak (data not shown) reveals that the signals for the m/z ) 649, (M + 2H)2+ ion are approximately the same in both cases. Therefore, the loss in S/N in the TIC is almost entirely attributable to an increase in the background noise. This effect can be best seen by comparing the individual mass spectra from the angiotensin I peak maxima obtained with the NPS and conventional columns, as shown in Figures 3B and 4B, respectively. In both cases, the m/z ) 649 ion signal height is 60-70 units in magnitude. In the conventional LC case, however, the background chemical noise dominates the spectrum. This result is reasonable upon considering the relationship between solute zone concentration, spectral integration time, and data point density. The 9 s peak of angiotensin I obtained using conventional LC has the same number of data points as the 500 ms peak from the NPS LC. Each point in the conventional data, however, represents an integration time (1 s) which is about 16fold longer than the integration time with the NPS data. Because the solute zone is also about 16-fold more dilute in the conventional case, one would anticipate that the observed magnitude of the ion signal of interest should be approximately the same in both cases, ignoring, of course, ion statistical limits and detection probabilities at the lower concentration. In fact, the m/z ) 649 ion signal is about the same in both cases. However, the critical difference lies in the background noise, which, as observed here, is the result of real ions always present in the mobile phase and almost impossible to completely remove.

Figure 5. Comparison of peptide ion signals obtained from NPS LC and FIA sample injections. LC conditions are the same as in Figure 3A. Numbers above the bar graphs coincide with peak identifications in Figure 1.

Figure 4. (A) TIC from separation of the nine-component peptide mixture using conventional porous LC. Gradient: 15-90% ACN, 0.1% TFA in 60 min. Flow rate: 1 mL/min. Peak assignments are the same as in Figure 1. (B) Single mass spectrum from apex of angiotensin 1 peak. Data were acquired in 1 s.

The fact that real chemical noise background, as opposed to simple electronic or statistical noise, ultimately determines the S/N in on-line separations and ES/MS detection has become an increasingly accepted proposition. Although the 16-fold greater integration time increased 16-fold the ion signal for the peak of interest, which was diluted 16-fold due to a broader liquid zone, it also integrated the background noise for a longer period of time and increased its value as well. These data suggest the quite rational conclusion that the greatest TIC sensitivity will be observed from the narrowest possible peak, which has the highest zone concentration against the constant background noise. The result implies that a fast data acquisition time with a narrow peak width should be used, even though the required decrease in integration time decreases the ion signal itself. Again, this conclusion is based on the observation that the S/N limit in most applications is invariably dictated by the background chemical noise consisting of real ions, and not the various electronic sources of noise. Comparison with FIA. To determine what kind of suppression effects from other sample components and contaminants might be present if no separation process were used, the same quantity of sample injected in Figure 3A was injected using an isocratic mobile phase of 40% methanol (0.1% TFA) but with no column present. A loop size of 1 µL was used in order to create FIA peak widths essentially identical to the ones found in the separation example, and the concentration of the sample compnments was adjusted accordingly. In this way, peak heights from

the various ions could be directly compared between the LC and FIA experiments. These data are shown in Figure 5 as bar graphs for the individual ions and clearly demonstrate that signal from many of the ions is severely suppressed when there is no separation of species. These kinds of variable suppression effects in ES/MS are well-known and are presumed to be the result of competition between ions or neutrals for the limited amount of charge and, perhaps, even physical space present on the surface of an ES droplet just before ion desorption occurs. The samples used were synthetic and so contained substantial quantities of involatile salts, which are known to suppress ES/MS signal as well. The relative suppression of ion signal is a result of the complex relation between pKa, solubility, polarity, and surface activity and is an active area of study.27,28 Regardless of the cause, these results indicate that there are real benefits to the use of liquid phase separation schemes with ES/MS detection due to possible interference from the solute and matrix and that simple FIA analysis may lead to confusing or inaccurate data, especially where complex mixtures are involved and quantitation is desired. Peptide Mapping. As an example of the fast separation capability of the NPS material and the fast data acquisition capability of the TOF/MS, a peptide mapping application was explored. Figure 6 shows the TIC from a peptide map of trypsindigested myoglobin, taking approximately 3.5 min to complete. The amount of material injected here was equivalent to 300 pmol of the myoglobin protein sample. Peak widths generated here are on the order of 0.5-3 s wide, and thus a data acquisition rate of 16 spectra/s was used. The fragments identified in Figure 6 correspond to 98% of the myoglobin sequence or 151 of 153 amino acids, more than adequate for a positive identification. Of the 16 expected fragments above m/z ) 300, 14 were identified here. Additionally, five fragments from a single cleavage miss, one from a double cleavage miss, and one from a triple cleavage miss were identified. Based on the above comparison between conventional and NPS LC, it is expected that the performance from a conventional column would be considerably worse for this application, in terms of both a decreased TIC signal and a much longer analysis time. (27) Buhrman, D. L.; Price, P. I.; Rudewicz, P. J. J. Am. Soc. Mass Spectrom. 1996, 7, 1099. (28) Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1990, 62, 957.

Analytical Chemistry, Vol. 69, No. 19, October 1, 1997

3977

use of a 4.6 mm i.d. column with a 1 mL/min flow rate. Previous studies29 using a quadrupole MS have shown that orders of magnitude in sensitivity may be gained through the use of capillary LC with ES/MS detection of peptide maps. CONCLUSIONS The development of faster liquid phase separation techniques has progressed sufficiently in recent years so that resolution of peptide and protein mixtures can be easily completed in less than 1 min, thus generating chromatographic peaks which are less than 1 s in width. On-line detection of these solute zones by mass spectrometry using electrospray is, of course, a desirable capability due to the value and depth of information which can be obtained, i.e., molecular weight and chemical structure. Time-of-flight mass spectrometry is uniquely suited for this strategy because it has the capability to acquire data at the speed needed to adequately sample these narrow solute zones. Using the combination of reversed-phase nonporous LC with a 4.6 mm i.d. column and TOF/ES/MS detection, as little as 500 fmol of Arg-Gly-Asp can be observed. A peptide mixture can be completely resolved and the components identified by molecular weight in less than 35 s. Finally, a peptide map with MS detection can be completed in 3.5 min. Future work will include the development of packed-capillary NPS columns to improve overall sensitivity for the ES/TOF/MS system. Figure 6. TIC from 300 pmol of trypsin-digested myoglobin using NPS LC. Gradient: 5-90% ACN, 0.1% TFA in 5 min. Flow rate: 1 mL/min. Storage rate: 16 spectra/s. M/z range: 150-2000.

Although the absolute sensitivity demonstrated with this peptide map is not impressive in and of itself, it is good considering the (29) Banks, J. F. J. Chromatogr. A 1996, 743, 99.

3978 Analytical Chemistry, Vol. 69, No. 19, October 1, 1997

ACKNOWLEDGMENT This work was supported in part by NIH Grant 1R43GM5388601. Received for review Feburary 27, 1997. Accepted July 2, 1997.X AC970226T X

Abstract published in Advance ACS Abstracts, August 15, 1997.