Multiplexed Four-Channel Rectilinear Ion Trap Mass Spectrometer

Jan 27, 2009 - To utilize the advantages offered .... control electronics, hardware, and software used in the devel- ... The custom-built vacuum manif...
0 downloads 0 Views 446KB Size
Download PDF
Anal. Chem. 2009, 81, 1570–1579

Multiplexed Four-Channel Rectilinear Ion Trap Mass Spectrometer Sameer Kothari,† Qingyu Song,† Yu Xia,† Miriam Fico,† Dennis Taylor,‡ Jonathan W. Amy,†,‡ George Stafford,‡ and R. Graham Cooks*,† Department of Chemistry, Purdue University, West Lafayette, Indiana 47907-2084, and Thermo Fisher Scientific, 355 River Oaks Parkway, San Jose, California 95134 A four-channel multiplexed mass spectrometer with rectilinear ion trap (RIT) mass analyzers was designed, constructed, and characterized. The system consists of four parallel atmospheric pressure ion (API) sources, four RIT mass analyzers, four sets of ion optical elements, and four conversion dynode detectors. The complete instrument is housed in a single vacuum manifold with a common vacuum system. It has a relatively small footprint, and costs and complexity were minimized and controls simplified by sharing the electronics and control modules among different channels. Each channel of the instrument can be operated in either positive or negative ion mode with a choice of ionization methods to improve the information content from an experiment. Also, the instrument is equipped with simultaneous data acquisition capabilities from all four channels, but the use of a common RF electronics system limits the degree to which the analyzer channels can be scanned independently. The instrument was characterized over the mass/charge range of 150 to 1300 Th. Mass misassignments in different ion traps because of machining and assembly tolerances were avoided by the application of supplementary direct current signals to each mass analyzer to correct mass offsets. A multiplexed automatic gain control (AGC) scheme was developed to control the ion population in each of the traps independently. These two features allow tandem mass spectrometry to be performed with an isolation window of 1 Th so trapping identical ions in all four channels. There are two principal modes of operation. In one, the same sample is analyzed in all four channels using different ionization methods to increase the information content of the analysis. In the other mode of operation, different samples are analyzed in all four channels with the same ionization method, so providing higher throughput. These capabilities were demonstrated by examining lipids produced by Escherichia coli and complex mixtures containing drugs of abuse. Mass spectrometry plays an important role in many areas of science including in the pharmaceutical and biological sciences in such areas as library screening,1 proteomics,2 and metabolo* To whom correspondence should be addressed. E-mail: [email protected]. Phone: (765) 494-5262. Fax: (765) 494-0239. † Purdue University. ‡ Thermo Fisher Scientific.

1570

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

mics.3 The application of mass spectrometry to high throughput screening in drug discovery4 has altered the serial process of lead identification and optimization which is a traditional part of drug development. Instead, large numbers of new compounds are synthesized in parallel using combinatorial chemistry and tested rapidly using high throughput screening. In principle, if not always practice, the synthesis of new chemical entities is no longer the rate-limiting step during drug discovery. Increasingly, the new bottleneck has shifted to the analytical verification of the structure and purity of each library compound synthesized using combinatorial chemistry or of each lead compound obtained using high throughput screening. Consequently, more rapid and robust analytical approaches are needed to keep pace with these new paradigms.5 In these and other research areas, large numbers of highly similar samples are still analyzed via introduction of one sample into one mass spectrometer at a time. The process involves repeating the same operation many times on essentially identical samples. This is a circumstance in which differential measurements using multiplex instruments have much to offer. Although it is widely accepted that mass spectrometry provides a universal means to characterize and distinguish drug molecules based on both molecular weight and structural features, the cost of MS instrumentation is still relatively high. To utilize the MS system more effectively, several innovative liquid chromatography (LC) systems have been developed for use in combination with mass spectrometry. The use of a single mass analysis channel with provisions for time sharing of the mass spectrometer between several LC columns6 was introduced in the form of the MUX system (Micromass, Waters, Milford, MA). In one application7 the flow from a single HPLC pump was split four ways to feed four columns simultaneously, and the flows from each of the four columns were examined via a four-way multiplexed electrospray interface using an orthogonal time-of-flight (TOF) mass spectrometer. Each liquid stream was sampled in rapid succession for 0.1 s, and individual mass spectra for each column were acquired in four separate (1) Enjalbal, C.; Martinez, J.; Aubagnac, J. Mass Spectrom. Rev. 2000, 19, 139– 161. (2) Aebersold, R.; Goodlett, D. R. Chem. Rev. 2001, 101, 269–296. (3) Yanagida, M. J. Chromatogr. B 2002, 771, 89–106. (4) Thompson, L. A.; Ellman, J. A. Chem. Rev. 1996, 96, 555–600. (5) Kyranos, J. N.; Cai, H.; Wei, D.; Goetzinger, W. K. Curr. Opin. Biotechnol. 2001, 1, 105–111. (6) Deng, Y.; Wu, J.-T.; Lloyd, T. L.; Chi, C. L.; Olah, T. V.; Unger, S. E. Rapid Commun. Mass Spectrom. 2002, 16, 1116–1123. (7) de Biasi, V.; Haskins, N.; Organ, A.; Bateman, R.; Giles, K.; Jarvis, S. Rapid Commun. Mass Spectrom. 1999, 13, 1165–1168. 10.1021/ac8023284 CCC: $40.75  2009 American Chemical Society Published on Web 01/27/2009

data files that were synchronized with the spray being sampled. The total cycle time from channel to channel of this four-way multiplexed electrospray interface was 0.8 s, allowing full mass spectra to be recorded using a quadrupole mass spectrometer. In another approach, an array of electrospray tips was designed as a disposable counterpart of the standard 96-microtiter well plate, enabling throughput of 12 samples per second.8 The NanoMate system (Advion BioSciences, Ithaca, NY) provides yet another solution to the problem of high-throughput analysis. In this system, rapid robotic sample changing at the inlet of a single mass spectrometer couples it in turn to each of an array of micromachined silicon nanospray tips.9,10 All these systems provide an increase in sample throughput over the traditional one sample per mass spectrometer approach, but there remains an opportunity for multiplexed analytical systems that extend scalability beyond currently available technology. Analytical studies aim at a comprehensive and unbiased view of the system under study. All ionization techniques discriminate against analytes with certain physical-chemical properties.11 The use of multiple ionization methods in tandem will increase the number of species detected, increasing the information acquired from the analysis.11,12 Even more information can be obtained by switching polarity (positive and negative) during ionization, as first demonstrated by Hunt13 using pulsed positive and negative ion chemical ionization. It is estimated that more than 80% of compounds can be ionized readily by ESI.14 To significantly increase this number, it is desirable to choose a different ionization method, such as atmospheric pressure chemical ionization (APCI), atmospheric pressure photo ionization (APPI), or atmospheric pressure matrix-assisted laser desorption (AP MALDI). This typically requires a different ion source and reoptimization of the ion source parameters, which increases analytical cycle time and makes automation impractical. To utilize the advantages offered by various ionization methods in an efficient manner, ion sources combining multiple ionization methods in series and in parallel have been introduced. The so-called multimode11,14-17 ion sources combine multiple ionization modes in series, and the resulting spectrum is a combination of the data from the combined action of the ionization processes. These sources have the potential to increase throughput and information content to a considerable extent although they may increase difficulties in spectral interpretation. A multiplexed instrument would provide independent control of each source to optimize ionization conditions in each source. (8) Liu, H.; Felten, C.; Xue, Q.; Zhang, B.; Jedrzejewski, P.; Karger, B. L.; Foret, F. Anal. Chem. 2000, 72, 3303–3310. (9) Schultz, G. A.; Corso, T. N.; Prosser, S. J.; Zhang, S. Anal. Chem. 2000, 72, 4058–4063. (10) Van Pelt, C. K.; Zhang, S.; Henion, J. D. J. Biomol. Tech. 2002, 13, 72–84. (11) Nordstro ¨m, A.; Want, E.; Northen, T.; Lehtio ¨, J.; Siuzdak, G. Anal. Chem. 2008, 80, 421–429. (12) Yu, K.; Di, L.; Kerns, E.; Li, S.; Alden, P.; Plumb, R. S. Rapid Commun. Mass Spectrom. 2007, 21, 893–902. (13) Hunt, D. F.; Stafford, G. C.; Crow, F. W.; Russel, J. W. Anal. Chem. 1976, 48, 2098–2104. (14) Gallagher, R. T.; Balogh, M. P.; Davey, P.; Jackson, M. R.; Sinclair, I.; Southern, L. J. Anal. Chem. 2003, 75, 973–977. (15) Syagea, J. A.; Hanolda, K. A.; Lynna, T. C.; Horner, J. A.; Thakur, R. A. J. Chromatogr., A 2004, 1050, 137–149. (16) Fisher, S. M.; Perkins, P. D. Agilent Technical Note 2005. (17) Laiko, V. V.; Moyer, S. C.; Cotter, R. J. Anal. Chem. 2000, 72, 5239–5243.

Each of these approaches discussed so far allows users to analyze more samples in less time and provides superior information. But in each of the above approaches, mass analysis is performed with a single mass analyzer. To scale up the throughput from mass spectrometric analysis even further, a plurality of mass analyzers is required. Such a system apparently was first implemented by Byrdwell18 who split a liquid chromatographic effluent and used two mass spectrometers in parallel to provide ESI and APCI data simultaneously. This particular system used a triple quadrupole instrument and a single quadrupole instrument. The data were obtained under different ionization conditions on components separated in a single chromatographic run but with selection of virtually identical retention times by the different systems. Also the amount of time, as well as the amount of solvent and sample used, was cut by half, whereas the information obtained per run was increased. This study was followed by the description of another dual mass spectrometry system using a triple quadrupole and an ion-trap instrument.19 In 2005, Cornish’s group presented a two-channel arrayed MALDI-TOF20 mass spectrometer with two linear analyzers employing a single laser and vacuum system. This instrument was aimed at use in time-critical detection of hazardous agents. These parallel mass spectrometers achieve higher throughput and chemical specificity by using multiplexed mass spectrometry. Several other multiplexed mass spectrometers have been reported. A four-channel multiplexed instrument with electron ionizationchemical ionization (EI-CI) based on cylindrical ion trap (CIT) mass analyzers was developed.21 To improve the sensitivity of this instrument, the CIT mass analyzers were later replaced by higher capacity RIT mass analyzers.22 At the same time, the first multiplexed instrument equipped with atmospheric pressure ionization (API) sources was reported.23 This four channel ESICIT instrument had limited pumping capacity allowing only two channels to be used simultaneously. The aim of the current work was to build a mass spectrometer that is multiplexed in a manner that allows each channel to be operated in as fully parallel and independent a fashion as possible. This means multiplexing the ionization source, ion-transfer optics, mass analyzer and detector in such a way that each channel operates as a complete mass spectrometer. In the multiplexed ion source, two ionization methods (ESI and APCI) operating in two possible modes (positive and negative) were implemented to allow for the detection of more analytes in each sample. Through the use of a common vacuum system, shared control electronics and a common data acquisition system, multiplexing of the largest and most expensive portions on the instrument were avoided, and an instrument with relatively small size was developed. At the same time, analytical capabilities of the instrument were greatly enhanced by allowing universal analysis using different ionization methods and polarities on the different channels. While the system (18) Byrdwell, W. C. Rapid Commun. Mass Spectrom. 1998, 12, 256–272. (19) Byrdwell, W. C.; Neff, W. E. Rapid Commun. Mass Spectrom. 2002, 16, 300–319. (20) Cornish, T. J.; Antoine, M. D.; Ecelberger, S. A.; Demirev, P. A. Anal. Chem. 2005, 77, 3954–3959. (21) Amy, M.; Tabert, J. G.-R.; Andrew, J.; Guymon, R.; Cooks, R. G. Anal. Chem. 2003, 75, 5656–5664. (22) Tabert, A. M.; Goodwin, M. P.; Duncan, J. S.; Fico, C. D.; Cooks, R. G. Anal. Chem. 2006, 78, 4830–4838. (23) Misharin, A. S.; Laughlin, B. C.; Vilkov, A.; Takats, Z.; Ouyang, Z.; Cooks, R. G. Anal. Chem. 2005, 77, 459–470.

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

1571

described here does not achieve all the objectives fully, it represents significant steps toward these goals. MATERIALS AND METHODS All the chemicals used were obtained from Sigma Aldrich (Milwaukee, WI) and used without further purification. The total lipid extract from Escherichia coli was obtained from Avanti (Alabaster, AL) and was used without further purification. Bacteria samples (SAR A19 and E. coli K12) were provided by FDA laboratory. The Bligh and Dyer method24 was used for lipid extraction from these samples prior to MS analysis. Solutions of the compounds used were prepared using HPLC grade solvents and deionized water. Throughout this paper, mass/charge ratios for ions are expressed in the Thomson25 unit, where 1 Thomson (Th) ) 1 Da/atomic charge. RESULTS AND DISCUSSION Instrumentation. The four-channel multiplexed mass spectrometer used rectilinear ion trap (RIT) mass analyzers, a choice based on previous work26 in single-stage LTQ instrument (Thermo Fisher Scientific, San Jose, CA) in which an RIT analyzer replaced a hyperbolic linear ion trap analyzer. The resulting single-channel RIT instrument exhibited performance characteristics not greatly different from those of a standard hyperbolic linear ion trap in terms of resolution, mass range, mass accuracy, and mass linearity, as well as tandem mass spectrometric capabilities. The ease with which the simplified geometry RIT analyzer could be machined compared to a conventional hyperbolic trap made it the choice for use in the multiplexed mass spectrometer. All the traps and the ion optics in the new four-channel instrument are housed in a single, four-sectioned aluminum manifold and share the same pumping system. The mass spectrometer is equipped with four atmospheric pressure ionization sources and four ion optical systems to allow simultaneous data acquisition from the four channels. All traps are powered by the same monopolar RF system but different direct current (DC) offsets are applied to the y-electrode pairs individual ion traps for mass correction. The instrument has four detectors, one for each trap. The ion sources, ion optics, and ion traps are controlled using LTQ/LXQ controls. The control electronics, hardware, and software used in the development of this instrument represent items obtained from standard commercial LTQ platform instruments unless mentioned otherwise. The existing LTQ control software was modified to control the newly added devices. Vacuum System. The custom-built vacuum manifold is made of aluminum and accommodates four stages of differential pumping. Various transfer capillary diameters (0.250 mm, 0.375 mm, and 0.500 mm) were tested to achieve a satisfactory balance between the sensitivity of the instrument and the pumping requirements needed to accommodate the gas load. Four stainless steel (SS 316) capillaries, which measure 0.375 (24) Bligh, E. G.; Dyer, W. J. Can. J. Biochem. Physiol. 1959, 37, 911–917. (25) Cooks, R. G.; Rockwood, A. L. Rapid Commun. Mass Spectrom. 1991, 5, 83. (26) Song, Q.; Kothari, S.; Senko, M. A.; Schwartz, J. C.; Amy, J. W.; Stafford, G. C.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2006, 78, 718–725.

1572

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

mm inner diameter, 1/16 in. outer diameter, and 100 mm length, are used for ion transfer from atmospheric pressure to the first vacuum region. Following the capillaries, an array of skimmers of 1 mm aperture diameter serves to limit gas transfer into the second vacuum stage. Thin lenses with apertures of 1 mm are used as conductance limiting orifices for the third and the fourth stages of pumping. The first pumping stage uses two 40 m3/hr rough pumps (Leybold SV40BI), the second stage uses two 70 L/sec hybrid turbomolecular pumps (Edwards EXT 70H) backed by a rough pump (18 m3/hr), the third stage uses a 350 L/sec turbomolecular pump from Edwards, and the fourth stage uses a Pfeiffer 521 L/sec turbomolecular pump. The third and fourth stage turbo pumps are backed by a common rough pump (18 m3/hr). This system was successfully tested under full gas load with all four capillaries exposed to atmospheric pressure and helium buffer gas added at the desired pressure into the analysis chambers. A schematic showing the vacuum chamber and the pumping system is represented in Figure 1a. Three pressure gauges were used to monitor the pressure in the vacuum chambers continuously. Multiplexed Ionization Source. The custom-built multiplexed ion source is mounted on the system manifold. The source is designed to allow ionization by any combination of APCI and ESI in the desired polarity to the four analysis channels. Two independently controlled power supplies, capable of producing up to 8 KV potential in the positive or negative ion modes power the ion sources, make it possible to generate positive, as well as negative, ions at the same time in different channels. A 4-channel syringe pump (Pump-22, Harvard Apparatus, MA) is used for simultaneous sample infusion into all four sources. The transfer capillary tubes are heated by means of heater blocks, one for each channel. Omega (series CN900) controllers, with a resistance feedback sensor, are used for controlling the temperature of the heaters. In all the experiments described here, a heater temperature of 200 °C was used. Metered sheath and auxiliary gas supplies for individual sources are used. An LC pump is used to provide the excess flow requirement of the APCI sources. Ion Optics. The instrument uses tube lens/skimmer interfaces followed by three sets of square quadrupoles as shown in Figure 1b. Each quadrupole set has these four ion optical elements connected in parallel. Adjustable DC potentials can be added to the sets of tube lenses, skimmers, interquadrupole lenses, and square quadrupoles. All the square quadrupoles share the same RF coil and control electronics. Their operating frequency is 1.589 KHz, and the maximum available voltage is 400 V. In the current set up, the ion optics are configured in such a way that two sets of ion optical channels are dedicated to positive ion transport and the other two sets dedicated to negative ion transport by sharing lens voltages. However, this choice can easily be changed. The tube lens in each channel is provided with a separate DC potential, as it is also used for controlling the number of ions entering each trap.

Figure 1. (a) Vacuum manifold and pumping system configuration for four channel instrument. (b) Ion optical configuration. (c) Arrangement of four ion traps and detectors.

Ion Traps and Detectors. Four RITs are mounted in parallel, each equipped with a detector, as shown in Figure 1c. All the traps share the same RF and control electronics. The traps operate at the main RF frequency of 804 kHz. An auxiliary RF is provided for resonance excitation during scanning. A custom-built RF coil having four filars in monopolar configuration is used to provide

DC bias voltages for the individual traps making it possible to trap either positive or negative ions. Each filar has an independent DC voltage that can also be used to change the secular frequency of ions during scanning. Each detector contains a dynode and electron multiplier. Each trap is provided with a single detector because of space conAnalytical Chemistry, Vol. 81, No. 4, February 15, 2009

1573

Figure 2. (a) Spectrum for Cal-mix calibration solution before applying the mass correction. (b) Principle of the mass correction method. (c) Cal-Mix spectra after applying mass correction to channels 2, 3, and 4.

straints. Two pairs of dynodes are connected together and positive or negative potentials (± 15 KV) can be independently applied to 1574

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

each pair. High voltage (up to -1.5 KV) can be applied to all four of the electron multipliers and can be controlled independently.

Figure 3. Simultaneous four channel positive ESI MS mixture analysis of drug and metabolite standards.

Data Acquisition System. The Thermo Xcalibur data acquisition and processing system can handle only one stream of data. To achieve simultaneous data acquisition from four channels, a customized data handling system was used. This system uses National Instrument DAQ card (NI PCI-6071E, National Instruments Corp., Austin, TX) and a custom program written using LabVIEW software. Data acquired by the LabVIEW system is converted into the Xcalibur format for further processing. Analytical Characterization. Figure 2a shows full scan mass spectra recorded separately in each of the four channels for a standard calibration solution containing caffeine, the tetrapeptide MRFA, and the mass calibration standard Ultramark. Data were acquired in the positive ion electrospray ionization mode at a scan speed of 2778 Th/sec. To obtain unit resolution at 525 Th, the scan speed was reduced to 1389 Th/sec. The mass range available is 50-1330 Th. Resonance excitation was carried out at q ) 0.88 to maximize performance. Mass calibration was performed only on channel 1. During simultaneous mass analysis using all four channels, it was observed that the same ion species is assigned slightly different mass (errors range from 1 Th to 6 Th for the mass range used and are proportional to mass/charge of the ion) in the different channels. It was concluded that geometric differences in machining and assembly of the traps causes the ion to have slightly different apparent masses in the different traps. For example, it is evident from Figure 2a that for channel 3 the mass is shifted by about 1 Th for caffeine (m/z 195) and by about 6 Th for ultramark (m/z 1222). The same phenomenon was also observed previously.22

To solve this problem, a supplementary DC voltage was applied to the individual traps, and its magnitude was scanned with that of the RF. With careful calibration of supplementary DC potential, the secular frequency of ions can be changed in such a way that any given ion is placed at the same Mathieu “q” value in all the traps, resulting in the same mass assignment for a given species. The ability to add this supplementary DC potential to the electrodes of the individual traps was achieved by designing an RF coil having four filars in a monopolar configuration to provide two pairs of trap biases (for positive and negative ion storage, respectively). Different DC voltages can be applied to each filar, and each DC voltage was ramped with the RF voltage. Application of these small DC voltage changes the q value of the ions, as can be seen by the operating line in Figure 2b, which is at an angle to the q-axis. Figure 2c shows that the correct masses are also measured for channels 2, 3, and 4 after implementation of the ramped DC potential. A series of experiments was then performed to investigate the capabilities of the multiplexed instrument. Mixture Analysis. Four channels of simultaneous mass analysis allow analysis of more samples in unit time. To explore this feature a mixture containing drugs of abuse was examined in the positive ion electrospray ionization mode. The ionization conditions (spray voltage, sheath gas pressure, source-inlet distance, and tube lens voltages) for each channel were individually optimized prior to acquisition of data on all four channels. The analyte solution flow rate for each analyte was held constant. In this experiment, morphine, cannabinol, and oxazepam were each prepared as 50 µM in 50/50 methanol/water and were Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

1575

Figure 4. (a) Simultaneous positive mode ESI-MS of 3 individual barbiturates and their mixture. (b) Comparison of MS analysis of a barbiturate mixture (top) and post-analysis addition of the spectra of the individual analytes (bottom).

sprayed into channels 2, 3, and 4. A mixture of three drugs in 50/50 methanol/water in concentrations methamphetamine (116 µM), cocaine (56 µM), and heroin (46 µM) was sprayed in channel 1. Figure 3 shows the mass spectra obtained from each of the four channels. In each of the channels the [M+H]+ for each drug was a prominent peak. Note, however, that a peak corresponding to protonated heroin, m/z 370, also appears in channels 2 and 4. In a second mixture analysis experiment, a set of three barbiturates (amobarbital, hexobarbital, and phenobarbital) was analyzed as a mixture in channel 1 and separately in channels 1576

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

2-4. The individual analytes were dissolved in a 50/50 MeOH/ H2O, 0.1% formic solution to a concentration of 100 µM. An equimolar (100 µM) mixture of each analyte was examined in the same solvent mixture. The solutions were electrosprayed simultaneously into the four separate channels and were analyzed in the positive ion mode. The resulting mass spectra are depicted in Figure 4a. Prominent peaks at m/z of 316, 322, and 326 of amobarbital, phenobarbital, and hexobarbital are present in the channels 2, 3, and 4, respectively. All these peaks can be observed in the data from channel 1 into which a mixture of these barbiturates was sprayed. Note that the difference in ion

Figure 5. Mass spectra of lipid extracts from E. coli with simultaneous acquisition of both positive and negative ESI and APCI data.

Figure 6. MS spectra of lipid extracts from SAR A19 (top) and E. coli K12 using negative ion ESI in channels 1 and 2.

abundance of ions in different channels is due to variation among ionization sources. A very useful feature of the four channel instrument is the capability of manipulating spectra from separate channels to look for differences in the samples. Alternatively, in another method of data processing, the spectrum of a complex mixture can be

compared to the spectra of individual components. This comparison may allow for rapid identification of unexpected peaks, for example, from a contaminant in the mixture. Currently, these capabilities do not exist in real time; therefore, these steps were taken post-analysis. The m/z and intensity data for the spectra from channels 2-4 were normalized to each other and then Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

1577

Figure 7. CID spectrum of a PE lipid [M-H]- in the negative ion mode from channel 3 using negative ESI (top) and CID spectrum of [M+H]+ in positive mode from channel 2 using positive ESI (bottom).

exported into a Microsoft Excel spreadsheet, where they were added. The spectrum resulting from the addition is shown in Figure 4b. The dotted lines highlight the peak similarities between the two spectra, emphasizing the overlap in peak locations achieved by the addition method. The agreement between the two spectra is promising for future experiments since in this case it demonstrates the absence of significant matrix suppression effects. Comprehensive Lipid Mapping from E. coli Extract. The goal of this experiment was to simultaneously generate complementary data sets from four different ionization modes for comprehensive lipid mapping of a biological sample. Figure 5 demonstrates the data acquired simultaneously from each of the four channels for the same lipid extract from E. coli using four distinct ionization modes, positive/negative atmospheric pressure chemical ionization (APCI), and positive/negative electrospray (ESI). It is evident from the data that complementary information with regard to the lipid composition can be obtained using the different ionization modes. Therefore, by combining all the information from the four channels, comprehensive lipid mapping is more likely to be achieved in a high throughput fashion. High Throughput Comparison of Lipidome from Different Bacteria. MS spectra of lipid extracts from two different bacteria, E. coli K12 and SAR A19, were compared using ESI in the negative mode in channels 1 and 2 of the multiplex instrument. The data in Figure 6 show that distinctive profiles are obtained from different bacteria, which may enhance fast detection of bacteria using MS. High Throughput Structure Elucidation. The goal of this phase was to study high throughput MS/MS simultaneously on positively and negatively charged species to demonstrate access to complementary structural information. Phosphatidylethanola1578

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

mine (PE) is one of the major classes of lipids in bacterial lipidome, and collisional activation of the protonated molecule ([M+H]+) typically results in the loss of the phosphoethanolamine headgroup (141 Da). Therefore, constant neutral loss of 141 Da can be used as a diagnostic tool for the determination of PE species in complex lipid mixtures.27 However, the dominant headgroup loss depresses other fragmentation channels which are necessary for acyl chain identification. The acyl compositions of PE lipids are therefore obtained by doing CID on the deprotonated ions ([M-H]-), where diagnostic RCOOions are normally observed. Figure 7 shows an example of CID of protonated and deprotonated PE (Mw: 718) ions acquired simultaneously from channels 3 and 2. In the positive ion MS/ MS mode, the dominant peak at m/z 578, due to the loss of 141 Da, suggests the existence of PE lipid, while the major ions at m/z 258, 282, and 452 indicate the composition of the two acyl chains (18:1 and 16:1). Combining the information obtained from positive and negative mode, the structure of the PE lipid can be deduced rapidly with high confidence. CONCLUSIONS A new multiplexed mass spectrometer with four parallel massanalysis channels has been designed, constructed, and characterized. This instrument allows for simultaneous high-throughput mass analysis of analytes ionized using atmospheric pressure ionization techniques. Increased information is obtained by using different ionization methods and polarities (for example, simultaneous positive and negative mode APCI and ESI). Increased specificity is achieved using simultaneous tandem mass spectrometry analysis. Errors in mass assignments for the same ions in different channels because of geometric differences between (27) Pulfer, M.; Murphy, R. C. Mass Spectrom. Rev. 2003, 22, 332–364.

the traps were minimized by applying appropriate supplementary DC potentials to each channel. This improvement will be especially important in future versions of the instrument where more channels will be utilized. Applications of the instrument for complex mixture analysis, high throughput identification of bacterial strains, and lipid structure elucidation were demonstrated. In future multiplexed experiments, the large quantity of data generated will be analyzed with algorithms appropriate to analytical informatics. The current rf electronics will be significantly redesigned to apply unique auxiliary alternating current waveforms, at either the same time or in a rapid, sequential fashion, for the purpose of performing multichannel MS/MS on different and arbitrarily selected precursor m/z ions. The vacuum system for the instrument will be redesigned using the recently introduced discontinuous atmospheric pressure ionization (DAPI) ion introduction method.28 This will result in better management of the gas load allowing for increased number of mass spectrometer channels without additional vacuum pumps.

ACKNOWLEDGMENT The authors thank Zheng Ouyang, Jae Schwartz, and Michael Senko for important contributions in this work, Nigel Gore for his support in the development of this instrument, and John Syka for the suggested method of mass correction. The work at Purdue was supported by Thermo Fisher and by the Office of Naval Research.

(28) Gao, L.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2008, 80, 4026–4032.

AC8023284

SUPPORTING INFORMATION AVAILABLE Picture of the four channel multiplex instrument and further details about the effect of dimensional variation on mass assignment in array of ion traps. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review December 18, 2008.

November

4,

2008.

Accepted

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

1579