Journal of Analytical Toxicology 2015;39:96 –105 doi:10.1093/jat/bku135 Advance Access publication December 16, 2014
Article
Fragmentation Pathways and Structural Characterization of 14 Nerve Agent Compounds by Electrospray Ionization Tandem Mass Spectrometry Kathleen J. Housman, Austin T. Swift and Jonathan M. Oyler* US Army Medical Research Institute of Chemical Defense, 3100 Ricketts Point Road, Aberdeen Proving Ground, MD 21010, USA *Author to whom correspondence should be addressed. Email: [email protected]
Organophosphate nerve agents (OPNAs) are some of the most widely used and proliferated chemical warfare agents. As evidenced by recent events in Syria, these compounds remain a serious military and terrorist threat to human health because of their toxicity and the ease with which they can be used, produced and stored. There are over 2,000 known, scheduled compounds derived from common parent structures with many more possible. To address medical, forensic, attribution, remediation and other requirements, laboratory systems have been established to provide the capability to analyze ‘unknown’ samples for the presence of these compounds. Liquid chromatography/mass spectrometric methods have been validated and are routinely used in the analysis of samples for a very limited number of these compounds, but limited data exist characterizing the electrospray ionization (ESI) and mass spectrometric fragmentation pathways of the compound families. This report describes results from direct infusion ESI/MS, ESI/MS2 and ESI/MS3 analysis of 14 G and V agents, the major OPNA families, using an AB Sciex 4000 QTrap. Using a range of conditions, spectra were acquired and characteristic fragments identified. The results demonstrated that the reproducible and predictable fragmentation of these compounds by ESI/MS, ESI/MS2 and ESI/MS3 can be used to describe systematic fragmentation pathways specific to compound structural class. These fragmentation pathways, in turn, may be useful as a predictive tool in the analysis of samples by screening and confirmatory laboratories to identify related compounds for which authentic standards are not readily available.
Introduction Organophosphate nerve agents (OPNAs) are highly toxic compounds which act as cholinesterase inhibitors. According to the Nuclear Threat Initiative (NTI), the most widely used and proliferated weapons of mass destruction have been chemical weapons in large part because of the ease with which they can be produced, stored and weaponized, and the reduced production and storage costs as compared with biological or nuclear weapons (NTI: Understanding Chemical Threats Home Page. http://www.nti.org/threats/chemical/, accessed 5 November 2013). Since 1997, the 188 member countries of the Organization for the Prohibition of Chemical Weapons (OPCW) have agreed to ban the production, stockpiling and use of chemical warfare agents (CWA) in accordance with the Chemical Weapons Convention (CWC) (1). The list of known scheduled target compounds is comprised of over 2,000 compounds many of which are derived from common parent structures (2). The theoretical possibilities for nerve agents derived from the Schedule 1.A.1 and 1.A.3 core structures exceed 270000 (3). While member countries had, as of 31 December 2011, eliminated 51.5 metric tons of their declared
71.2 metric tons of Category 1 CWA stockpiles, the remaining stockpiles, undeclared stockpiles possessed by nonparticipating states and non-state actors and frequently discovered new agent caches continue to represent a potential threat to human health through both military and terrorist use. To verify treaty compliance and investigate potential events, the OPCW has established a proficiency testing program which maintains an analytical system comprised of 22 expert-designated laboratories and a mobile laboratory (4). In addition to the OPCW laboratory system, several other laboratory systems have been established to address this potential threat. The Laboratory Response Network (LRN) was established by the US Department of Health and Human Services, Centers for Disease Control and Prevention (Centers for Disease Control and Prevention (CDC) Emergency Preparedness and Response: The LRN Partners in Preparedness Home Page. http://www.bt.cdc. gov/lrn/, accessed 6 November 2013) in collaboration with the US Department of Homeland Security (DHS), the US Federal Bureau of Investigation (FBI) and the Association of Public Health Laboratories (APHL). It is responsible for providing analytical capabilities to address potential chemical and biological terrorist events occurring within the USA. The LRN strategy is to maintain a national and international system of laboratories capable of rapidly identifying agents that threaten public health. The system links state and local public laboratories with veterinary, agriculture, military, environmental and food-testing laboratories. Other federal organizations including the FBI, the US Environmental Protection Agency (EPA) and branches of the US Department of Defense (DoD) have also put in place both fixed-site and field laboratory systems to provide diagnostic, forensic and verification analyses of clinical and environmental samples for these target analytes. For this reason, it is imperative that highly specific and robust analytical methods for identifying intact OPNAs be developed and validated to verify the presence of CWAs in unknown samples. Environmental and biological samples are routinely analyzed by gas chromatography – mass spectrometry (GC – MS) for CWAs using both electron impact ionization (EI) (5 – 7) and chemical ionization (8, 9). The in-source fragmentation of an extensive list of CWAs has been rigorously characterized using GC-EI – MS and extensive spectral libraries have been acquired using standardized ionization conditions (NIST Chemistry WebBook, NIST Standard Reference Database Number 69, eds. P.J. Linstrom and W.G. Mallard, http://webbook.nist.gov, accessed 6 November 2013) (10). However, spectral differences across structural analogs may be subtle when the spectral profile is solely dependent on in-source fragmentation. This can make compound identification difficult when authentic reference standards are not available for comparison. While GC – MS has proven to be a robust instrumental approach for identifying and quantifying OPNAs, additional sample extraction steps are
Published by Oxford University Press 2014. This work is written by (a) US Government employee(s) and is in the public domain in the US.
typically required to partition compounds from relatively polar solvents like water into GC-friendly volatile organic solvents. These additional sample preparation steps are time consuming, costly and can be the source of additional analytical error. Furthermore, the identification of the more polar OPNA methylphosphonic hydrolysis products via GC requires an additional derivatization step. These compounds are not discussed in this report, but the benefit of liquid chromatography (LC) is that samples could be analyzed for the presence of both CWA(s) and their respective hydrolysis breakdown products. While providing highly specific spectra, the single quadrupole (GC– MS and LC–MS) method specificity is largely dependent on a single process, in-source fragmentation. Relatively controlled tandem mass spectrometry (MSn) processes can be used to systematically acquire compound-specific spectra as well as to characterize the fragmentation patterns of families of compounds. These fragmentation patterns can, in turn, be used to predict compound identity with greater confidence when comparison against authentic reference standards is not possible. Though GC/MS2 methods used to identify and measure OPNAs provide increased specificity through the use of MS2 fragmentation, they still require the additional sample extraction steps characteristic of GC methods, and derivatization is still required for target analytes of higher relative polarity and those containing multiple reactive substituent groups. LC separation coupled with MS and MS2 detection provides highly specific methods with sensitivities and analysis times comparable to those obtained with GC/MS and GC/MS2 analysis, respectively. These methods do not require that target analytes be isolated in GC-friendly, volatile organic solvents, and derivatization is not required for more polar analytes, resulting in more efficient and cost-effective analyses. Despite the fact that LC/MS is currently being used routinely to detect and identify CWAs and their degradation products in both environmental and biological samples employing both electrospray ionization (ESI) (5, 11, 12) and atmospheric pressure chemical ionization (APCI) (13, 14), little work has been published documenting the complete MS2 fragmentation pathways of OPNAs using these ionization methods. In 2001, Bell et al. characterized the ESI products and MS2 fragmentation pathways of VX and RVX using a three-dimensional ion trap (Thermo Finnigan, San Jose, CA) (15). In 2006 using ESI, EllisSteinborner et al. postulated the MS2 fragmentation pathway of the V analog, ethyl S-2-diethylaminoethyl methylphosphonothiolate (Amiton, VG) (16). Weissberg et al. used an AB Sciex 5500 QTrap (Foster City, CA) to perform MS2 and MS3 to characterize VX, RVX, VM and two additional V analogs (17). Results from these studies indicated that the MS fragmentation of these dialkylphosphonothiolates followed the same pathways. This paper further verifies these previously reported data. This study focused on the two main OPNA classes, the G and V analogs. The compounds analyzed are listed in Table I. Structurally, these classes can be easily distinguished as V analogs have a common aminoethylphosphonothiolate backbone, and G analogs are alkoxyphosphonofluorides with the exception of GA, which is a phosphoroamidocyanidate. Using ESI on an AB Sciex 4000 QTrap (Foster City, CA), the MS2 fragmentation pathways have been characterized for 14 OPNA compounds over a range of collision energies.
Table I Compound List Compound
CAS #
Name
Formula
GA (Tabun) GB (Sarin) GF (Cyclosarin) GD (Soman) G1 (GB Analog) G2 (GF Analog) G3 VM
77-81-6 107-44-8 329-99-7
Ethyl N, N-dimethylphosphoramidocyanidate Isopropyl methylphosphonofluoridate Cyclohexyl methylphosphonofluoridate
C5H11N2O2P C4H10FO2P C7H14FO2P
96-64-0 648-59-9
Pinacolyl methylphosphonofluoridate Methyl isopropylphosphonofluoridate
C7H16FO2P C4H10FO2P
NA
Methyl cyclohexylphosphonofluoridate
C7H14FO2P
1426-08-0 21770-86-5
Ethyl isopropylphosphonofluoridate Ethyl S-2-diethylaminoethyl methylphosphonothiolate Isobutyl S-2-diethylaminoethyl methylphosphonothiolate Ethyl S-2-dimethylaminoethyl methylphosphonothiolate Ethyl S-2-diisopropylaminoethyl methylphosphonothiolate Propyl S-2-diisopropylaminoethyl methylphosphonothiolate Isopropyl S-2-diisopropylaminoethyl methylphosphonothiolate Isobutyl S-2-diisopropylaminoethyl methylphosphonothiolate
C5H12FO2P C9H22NO2PS
RVX
159939-87-4
Vx
20820-80-8
VX
50782-69-9
V1
52364-45-1
V2
51446-23-2
V3
855307-83-4
C11H26NO2PS C7H18NO2PS C11H26NO2PS C12H28NO2PS C12H28NO2PS C13H30NO2PS
Table II Characteristic In-source Fragment Ions Compound
In-source fragment ion
GA GB GD GF G1 G2 G3 VM RVX Vx VX V1 V2 V3
135 (20) 99 (27), 79 (7), 97 (52), 81 (8), 117 (13) 99 (10), 97 (100), 85 (20) 99 (30), 79 (3), 97 (70), 81 (10), 117 (7) 83 (12) 83 (6) 127 (12), 67 (5) 167 (25), 139 (129), 79 (24), 97 (27), 100 (125), 72 (26), 44 (40) 212 (50), 100 (100), 139 (25), 61 (10) 139 (7), 72 (20) 240 (40), 128 (115), 86 (22), 167 (10), 139 (10), 79 (8) 240 (12), 128 (100), 86 (20), 181 (20), 121 (20), 97 (17), 139 (13) 240 (90), 128 (95), 86 (15), 181 (10), 121 (5), 79 (5), 97 (15), 139 (13) 240 (20), 128 (105), 86 (12), 195 (9)
Note. Value in parentheses indicates the percent abundance of each ion relative to the [MþH]þ ion.
Materials and methods Materials and sample preparation Reference standards for GA, GB, GD, GF and VX were acquired from the CASARM Program of the US Army Edgewood Chemical Biological Center (ECBC), Aberdeen Proving Ground (APG), MD, with purities verified at 95% by NIST-traceable 1 H NMR, 13C NMR, 31P NMR, GC/TCD, GC/MSD and/or acid – base titration methods. All other reference standards were acquired from the Chemical Sciences Division at ECBC with purities verified at 92% by 1H NMR, 13C NMR or 31P NMR. Dilute authentic standards were formulated in chloroform at ,2 mg/ mL for G agents and ,1 mg/mL for V agents. Samples were then diluted to 500 ng/mL in 3:1 methanol:water (v:v) with 0.1% formic acid prior to analysis. The methanol, water and formic acid used were LC/MS-grade and were purchased from Fisher Scientific (Waltham, MA).
Fragmentation Pathways of 14 Nerve Agent Compounds 97
Figure 1. Summed product ion scan spectra of the [MþH]þ for all G analogs over a CE range of 0 –120 V.
98 Housman et al.
Figure 2. Proposed CAD pathway for [GAþH]þ.
MS and MS2 experiments Samples were analyzed by direct infusion at 10 mL/min on an AB Sciex 4000 QTrap (Foster City, CA) hybrid triple quadrupole/linear ion trap (LIT) mass spectrometer using a standard Turbo V, ESI source operated in positive mode. Data acquisition and analysis were performed using the AB Sciex Analyst 1.5 software platform. The 4000 QTrap mass spectrometer is comprised four quadrupoles, Q0 – Q3. Q0 is a high pressure, RF ion-focusing region which can also function as a LIT to enhance the collection of ions for subsequent MS, MS2 and MS3 analysis. In this region, the declustering potential (DP), a voltage differential between the orifice plate and Q0, is used to minimize solvent ion pairing, thereby optimizing ion transfer into the mass spectrometer. At relatively higher voltages, it can be used to generate collisionally induced dissociation (CID) with the aid of atmospheric gas particles and the nitrogen curtain gas. Q1 functions as a mass filter which can be operated as an ion guide to transfer ions up to m/z 2800 (full scan) or select ions for further analysis in Q2. Q2 can function as a focusing region in which ions are cooled by a small amount of relatively inert gas or as a collision cell in which ions are accelerated to collide with the gas to produce controlled fragmentation, a process labeled collisionally activated dissociation (CAD). Q3 can operate as an ion guide for full scan analysis, as a mass filter to select ions or as a LIT to enhance ion collection for CAD MS3 analysis. Initially, optimal source parameters for the low flow, direct infusion experiments were identified for each compound and included a probe height of 0 mm, a capillary voltage of 5,500 V, a sheath gas pressure of 20 psi and a curtain gas pressure of 20
psi. At the flow rates used, the use of turbo heaters and desolvation gas impaired the ionization efficiency of all compounds, so neither was employed in subsequent experiments. A Q1 full scan experiment (40–1,000 m/z) was performed to characterize the ESI fragmentation of each compound and identify possible precursor ions to be used in MS2 analyses. Ion transfer efficiency into the mass filter is dependent on various voltage differentials applied to the ions. These differentials desolvate and guide ions while the ions transition from the relatively high pressures in the source and Q0 to the high vacuum in the mass filter. Q1 full scan data were summed over 10 cycles while holding the DP at 20, 60 or 100 V. Once the protonated molecule ([MþH]þ) for each compound was identified, the DP was stepped over a 0 –120 V range in 1 V increments while scanning for the ion to identify the optimal DP at which maximal [MþH]þ ion transfer into the mass spectrometer occurred. Using optimal DP values, the [MþH]þ was selected in Q1 for CAD fragmentation in Q2. Q3 full scan data (m/z 40 – 1,000) were collected while varying the collision energy (CE) and the CAD gas pressure to identify as many important CAD-generated product ions as possible. The rate at which ions selected in Q1 are transferred through the collision cell (Q2) is directly proportional to the CE, the voltage differential between Q0 and Q2. Precursor ions are fragmented as they collide with a collision gas; in this case nitrogen was used. Each precursor-to-product ion transition occurred optimally at a specific CE, and for this reason product ion spectra for each [MþH]þ were summed while stepping the CE over a 0 – 120 V range in 5 V increments to identify important CAD-generated product ions. Since CAD fragmentation is also directly proportional to the CAD gas pressure, the same CAD experiments were performed with the collision cell nitrogen set at three different gas pressures, 1.7 – 2.9, 3.7 –3.9 and 4.7 –4.9 psi (low, medium and high settings, respectively, in the Analyst 1.5 software).
Linear ion trap experiments The Q3 LIT capability of the 4000 QTrap was employed to enhance both sensitivity and resolution in the structural analysis of these compounds. Ions were trapped both radially and axially through voltages applied by the quadrupoles, and the entrance and exit barrier lenses, respectively. After enough ions were trapped, voltages were ramped, and ions scanned out of the LIT. Enhanced MS (EMS) scanning, in which Q1 and Q2 operate in RF-only mode to focus and transfer the ion beam directly into the LIT, was used to collect full scan data. Data were summed for 25 cycles, while holding CE at 10 V, and employing the dynamic fill time (DFT) function in which the software optimized the fill time based on relative ion abundance. Enhanced product ion (EPI) scanning (a MS2 full scan using the increased sensitivity and resolution of the LIT) was used to collect product ion data on a precursor ion selected in Q1, fragmented in Q2 and trapped and scanned out of the LIT. Using optimal DP values for each precursor, EPI scans for each precursor ion were summed using a fill time of either 100 or 250 ms while stepping the CE over a 0 –120 V range in 5 V increments. MS3 experiments were performed in Q3 on single product ions generated from Q2 CAD fragmentation of a precursor selected in Q1. In Q3, the characteristic product ion was isolated, and through the application of auxiliary voltages, excited and Fragmentation Pathways of 14 Nerve Agent Compounds 99
Figure 3. Proposed CAD pathways for [GBþH]þ, [GDþH]þ and [GFþH]þ.
fragmented through collision with residual nitrogen from Q2. To enhance nitrogen bleed into Q3 sufficient to produce MS3 fragmentation, the CAD gas pressure in Q2 was set at either medium or high. For compounds from which stable CAD-generated and structurally significant product ions were produced, MS3 scans were collected. MS3 experiments were performed on [MþH]þ and other precursor ions using their optimal DP values and by employing the following parameters: CE 10 – 20 V, fill time of 250 ms or the DFT function, excitation energies of 10 – 50 V and MS3 fragmentation excitation time of 100 or 250 ms. EMS and EPI experiments were performed at low, medium and high CAD gas pressures, and MS3 experiments were only performed using medium or high CAD gas. All LIT experiments were performed at a scan speed of 1,000 Da/s over a scan range of 50–1,000 m/z. Results and discussion ESI and in-source fragmentation Over a range of source voltages, ESI of each compound produced characteristic ions. For all 14 compounds, prominent [MþH]þ 100 Housman et al.
peaks were identified. ESI of the G analogs produced prominent sodium adduct ([MþNa]þ) and sodiated dimer ([2MþNa]þ) and trimer ([3MþNa]þ) peaks, and also produced protonated dimers ([2MþH]þ) for all G analogs with the exception of G1. Other than [MþH]þ ions and characteristic in-source fragment ions, no sodiated adducts or protonated or sodiated dimers or trimers were identified for any V analogs. Characteristic in-source fragments along with percent abundances, relative to the [MþH]þ ion, are listed in Table II. No multiply charged species were identified following direct infusion ESI of these compounds.
G analog CAD fragmentation Figure 1 contains spectra resulting from CAD analysis of the [MþH]þ for all G analogs and provides evidence for the majority of G analog fragmentation steps observed. Postulated fragmentation pathways based on results from CAD analysis of the [MþH]þ of each compound are illustrated in Figures 2 –4. CAD of all protonated G analogs initially produced characteristic fragments because of the loss of the alkoxy alkyl group, and CAD of
Minimal CAD fragmentation was observed for the protonated phosphonofluoridates G1, G2 and G3 (Figure 4). The only characteristic product ions resulting from CAD of protonated G1 and G2 were m/z 99 and 83. It is likely that the formation of the m/z 99 ion resulted from a cleavage of the isopropyl and cyclohexyl groups from G1 and G2, respectively, and that the m/z 83 ion common to both formed as a result of subsequent cleavage of the methyl from the methoxy group. In addition, CAD of the CID-generated ion, m/z 99, from G1 and G2 produced characteristic m/z 83 fragment ions. Two prominent characteristic fragments at m/z 127 and 67 were identified from CAD of [G3þH]þ, an isopropyl phosphonofluoridate, originating from loss of its alkoxy ethyl group and its primary isopropyl and hydroxyl groups, respectively. Similar to GD, CAD of [G1þH]þ and [G3þH]þ also produced prominent m/z 43 peaks resulting most likely from protonation of the cleaved primary propylene group, and a prominent m/z 55 peak was produced from CAD of [G2þH]þ through a mechanism similar to that proposed for CAD of its cyclohexyl phosphonofluoridate analog, GF.
Figure 4. Proposed CAD pathways for [G1þH]þ, [G2þH]þ and [G3þH]þ.
proton-bound dimers produced fragmentation patterns similar to those observed following ESI-CAD of the corresponding [MþH]þ. Because of its unique phosphoroamidocyanidate structure, the Q3 full scan spectrum resulting from CAD of [GAþH]þ shared no common product ions with spectra collected from CAD of the other protonated G analogs (Figure 2). Initially, GA fragmentation produced an ion at m/z 135 due to the loss of the alkoxy ethyl group. This ion further fragmented producing an ion at m/z 117 resulting from loss of water and m/z 108 resulting from loss of the cyano group. The observed m/z 135, 117 and 108 GA fragment ions were consistent with those reported by D’Agostino and Chenier in 2010 (18). Relative abundance increases for the m/z 126 ion with corresponding increases in CAD energy during MS2 of [GAþH]þ coupled with an absence of the m/z 126 ion in product ion scans from negative blank solvent samples supported the formation of a water adduct. Subsequent CAD of in-source produced m/z 108 ions produced a prominent m/z 126 peak resulting from a gain of water which confirmed adduct formation; we also observed the formation of doublet water adducts at m/z 144. CAD of protonated GB, GD and GF (Figure 3), induced the loss of the alkoxy alkyl group and produced a characteristic ion at m/z 99, owing to their core methyl phosphonofluoridate structure. This was also previously reported by D’Agostino et al. in 2001 and 2010 (18, 19). Similar prominent fragment ions at m/z 81, 117, 79 and 97 were also observed for all three compounds. Since GD had the largest alkyl side chain, CAD of [GDþH]þ (Figure 1) produced three additional characteristic protonated alkene fragments at m/z 85, 57 and 43 as reported by D’Agostino and Chenier (18). A prominent m/z 55 peak was also produced by CAD of [GFþH]þ most likely originating from protonation of a fragment of the cleaved primary cyclohexene group.
V analog CAD fragmentation V series analog CAD occurred through two main mechanisms (Figure 5), and because V analogs have common aminoethylphosphonothiolate backbones, summed spectra from CAD of the [MþH]þ of each of the seven V analogs resulted in many similar product ions (Table III and Figure 6). CAD fragmentation patterns for V analogs through the two pathways we propose are consistent with those previously proposed by Bell et al. in their study of VX and RVX (17), Ellis-Steinborner et al. in work with Amiton (VG) (18) and Weissberg et al. in their mass spectrometric characterization of several V analogs (19). The fragmentation pathway for VX proposed in this study also agrees with the fragmentation of VX extracted from soil and concrete particles by ion trap secondary ion MS2 observed by Groenewold et al. in 1999 (20, 21). Cleavage appeared to occur at two locations along the ethyl bridge corresponding to neutral losses of the methylphosphono group (A) and dialkylamine (B). The first fragmentation pathway was initiated at low relative CAD energies by the neutral loss of the alkoxy alkyl (Figure 5, A1) as confirmed by the presence of a m/z 212 peak in the CAD full scan spectrum of [RVXþH]þ and m/z 240 peaks in the CAD full scan spectra of protonated VX, V1, V2 and V3. With increasing CAD energy, the methylphosphono group (Figure 5, A2) was subsequently cleaved as evidenced by the presence of a m/z 134 peak in the CAD scan spectrum of [RVXþH]þ and m/z 162 peaks in the CAD scan spectra of protonated V1, V2 and V3. There was no evidence that CAD of [VMþH]þ or [VxþH]þ led to the loss of either the alkoxy alkyl- or methylphosphono groups prior to the formation of aziridine or imminium ions, and no peak could be identified following CAD of [VXþH]þ that corresponded to the loss of the methylphosphono group. The final step in Pathway A was the removal of sulfur and formation of either an alkylated aziridine ion or an imminium ion (Figure 5, A3). Alkylated aziridine or imminium ions and subsequent fragment ions appeared as m/z 100, 72 and 44 peaks in CAD-generated full scan spectra of protonated VM, RVX and Vx and as m/z 128, 86 and 44 peaks following CAD of protonated VX, V1, V2 and V3. Fragmentation Pathways of 14 Nerve Agent Compounds 101
Figure 5. Proposed CAD pathways for protonated V analogs. Pathway A: neutral loss of alkoxy R-group (A1) followed by cleavage of methylphosphono group (A2) and formation of the aziridine ion (A3). Pathway B: neutral loss of dialkylamine (B1) followed by loss of the alkylated thiol group (B2) and/or cleavage of the alkoxy R-group (B3) and subsequent fragmentation. Table III Prominent Characteristic Ions Produced from CAD MS2 of Protonated V Analogs VM
RVX
Vx
VX
V1
V2
V3
[MþH]þ
240
268
212
268
282
282
296
Pathway A A1 A2 A3 A4 A5
– – 100 72 44
212 134 100 72 44
– – 72 – 44
240 – 128 86 44
240 162 128 86 44
240 162 128 86 44
240 162 128 86 44
Pathway B B1 B2 B3 B4 B5 B6
167 107 139 79 97 61
– – 139 – – 61
167 107 139 79 97 61
167 107 139 79 97 61
181 121 139 79 97 61
181 121 139 79 97 61
195 135 139 79 97 61
Structures
102 Housman et al.
Figure 6. Summed product ion scan spectra of the [MþH]þ for all V analogs over a CE range of 0– 120 V.
Fragmentation Pathways of 14 Nerve Agent Compounds 103
The initial step in the second major pathway (Figure 5, B1) resulted from the neutral loss of dialkylamine. It is proposed that this loss would likely result in the formation of either a thiiranium ion or an ion resulting from formation of an ethylene bridge between the sulfur and phosphono oxygen. This was supported by the identification of m/z 167 (VM, Vx and VX), 181 (V1 and V3) and 195 (V3) peaks. No stable product ion was observed following CAD of [RVXþH]þ indicative of this cleavage. Subsequent fragmentation followed two additional paths initiated by elimination of the alkylated thio group (Figure 5, B2) or cleavage of the alkoxy alkyl group (B3). Ions resulting from the loss of the alkylated thio group and corresponding with the number of carbons in the alkoxy alkyl group were identified for VM, Vx and VX as peaks at m/z 107, for V1 and V2 as peaks at m/z 121 and for RVX and V3 as peaks at m/z 135. CAD of all protonated V analogs produced successive m/z 139, 79, 97 and 61 fragment ions with the exception of [RVXþH]þ which did not produce m/z 79 or 97 peaks. Water adduction of all the V analogs to form methyl phosphonic acid was also observed.
LIT experiments Summed spectra resulting from EMS and EPI scanning were similar to those described above for all G and V analogs, further substantiating the proposed fragmentation pathways. Somewhat unexpectedly, product ion absolute abundances resulting from EPI scanning were reduced compared with those acquired using ‘pure quadrupole’ MS2 methods. This suggested that these ions were not efficiently retained in Q3 during the trapping process or that further fragmentation of resonant ions occurred in Q3 as a result of collision with CAD gas bleeding from Q2 during trapping. Table IV lists ions identified from EPI scanning of all protonated G and V analogs. While prominent precursor and product ions were identified in LIT experiments, many of the lower mass CAD ions did not appear in these spectra. No characteristic product ions were identified following MS3 fragmentation of the G analogs. This could be due to these compounds having reduced bond energy distributions compared with the V analogs. Each of the seven V analogs produced MS3 transitions correlating with at least one of the two aforementioned pathways which also supported the proposed two-step fragmentation pathways. As expected, product ion absolute abundances were reduced using these techniques compared with those acquired using MS2 techniques, and confirmation of some multi-step fragmentation was limited by the ‘one-third rule’ commonly associated with ion trapping (Table V). The m/z 162 ion was observed in MS3 spectra of the m/z 240 product of protonated V1, V2 and V3. In addition, m/z 86 was seen in MS3 spectra of the m/z 128 product from protonated VX, V2 and V3 which further confirmed the postulated fragmentation in Figure 5A. MS3 fragmentation of the m/z 212 product of protonated RVX produced m/z 134 and 100; this was previously reported by Weissberg et al. (2013) using a 5500 QTrap (17). In the case of VM, Vx and VX, an m/z 139 ion was observed in MS3 fragmentation of the product ion m/z 167 from m/z 212. MS3 fragmentation of the m/z 181 product from protonated V1 and V2 and the m/z 196 product from protonated V3 also resulted in the formation of an m/z 139 ion. This further confirms the proposed fragmentation pathway described in Figure 5B. In 2013, Weissberg et al. provided more in-depth MS3 characterization 104 Housman et al.
Table IV LIT EPI Scan-Generated Ions from Protonated Molecular Ions Compound
EPI ions
Compound
EPI Ions
GA GB GD GF G1 G2 G3
135, 117, 108, 126 99, 79, 97, 81 99, 79, 81 99, 79, 81 99 99, 83 127
VM RVX Vx VX V1 V2 V3
167, 139, 97, 100, 72 212, 134, 100, 72, 139 72, 167, 139, 79, 97 128, 86, 139, 79, 97 240, 128, 86, 181, 79, 97, 139 240, 162, 128, 121, 139 240, 162, 128, 86
Table V MS3 Transitions Observed in the Analysis of V Analogs Compound
MS3 confirmation Pathway A
VM RVX Vx VX V1 V2 V3
– 268/212/134 – 268/128/86 282/240/162 282/240/162 296/240/162
Pathway B – 268/212/100 – – – 268/128/86 296/128/86
212/167/139 – 212/167/139 212/167/139 282/181/139 282/181/139 296/195/139
of VX, RVX, VM and several additional V analogs using a 5500 QTrap, a more sensitive instrument with a faster LIT scan rate (17).
Conclusions The results presented in this study demonstrate that ESI/MS, ESI/ MS2 and ESI/MS3 can be used to structurally identify the individual compounds evaluated and substantiate previously published data for a very limited number of these compounds using similar methods. The results also demonstrate that the reproducible and predictable fragmentation of these compounds by ESI/MS, ESI/ MS2 and ESI/MS3 can be used to describe systematic fragmentation pathways specific to compound structural class. We observed the unique formation of water adducts during CAD fragmentation of GA and all V analogs with the exception of RVX. Although water adduct formation during CAD fragmentation of VX and RVX (15) and other small molecules (22, 23) has previously been reported by researchers using a threedimensional ion trap and an ion cyclotron, respectively, water adduct formation during CAD in a hybrid quadrupole LIT instrument has not been reported previously with these compounds. Due to security regulations agreed to by OPCW signatories that limit access to these compounds, few synthetic labs exist from which authentic standards can be acquired for a wide variety of the structurally related analogs. As predictive reference tools, the pathways described here could be used in laboratories charged with ‘unknown’ sample analysis to initially screen for CWA analogs for which authentic standards may not be readily available; since these compounds generally have very short clinical half-lives, use of these pathways as predictive tools would likely be limited to the analysis of environmental samples. The data presented will also expand the method-specific CWA spectral library, and this work, in addition to previously published
work, should serve as a reference for researchers and analysts in the treaty compliance, public health, forensic and defense communities. Acknowledgments The views expressed in this article are those of the author(s) and do not reflect the official policy of the Department of Army, Department of Defense or the U.S. Government. This research was supported by the Defense Threat Reduction Agency—Joint Science and Technology Office, Medical S&T Division. References 1. Convention on the Prohibition of the Development, Production, Stockpiling, and Use of Chemical Weapons and on Their Destruction. (2005). Technical Secretariat of the Organisation of Prohibition of Chemical Weapons (OPCW). The Hague, the Netherlands. http://www.opcw.org/index.php?eID=dam_frontend _push&docID=6357.pdf (6 November 2013, date last accessed). 2. Nyanyira, C. (2006) Sampling and analysis in the Chemical Weapons Convention and the OPCW Mobile Laboratory. In: Mesilaakso, M. (ed). Chemical Weapons Convention Chemicals Analysis: Sample Collection, Preparation and Analytical Methods, 1st edition, Chapter 2. Chichester, UK: John Wiley & Sons, Ltd, 133 pp. 3. Mogl, S. (2006) Sampling and analysis in the Chemical Weapons Convention and the OPCW Mobile Laboratory. In: Mesilaakso, M. (ed). Chemical Weapons Convention chemicals analysis: Sample collection, preparation and analytical methods, 1st edition, Chapter 2. Chichester, UK: John Wiley & Sons, Ltd, 7 pp. 4. Report of the Scientific Advisory Board on developments in science and technology for the Third Special Session of the Conference of the States Parties to Review the Operation of the Chemical Weapons Convention (2012) OPCW RC-3/DG.1. Organisation for the Prohibition of Chemical Weapons (OPCW) Third Review Conference, The Hague, The Netherlands. http://www.opcw.org/index.php?eID= dam_frontend_push&docID=15865.pdf (6 November 2013, date last accessed). 5. Koller, M., Becker, C., Thiermann, H., Worek, F. (2010) GC – MS and LC–MS analysis of nerve agents in body fluids: Intra-laboratory verification test using spiked plasma and urine samples. Journal of Chromatography B, 878, 1226– 1233. 6. D’Agostino, P.A., Porter, C.J. (1992) Capillary column gas chromatography/tandem mass spectrometry verification of chemical warfare agents. Rapid Communications in Mass Spectrometry, 6, 717–718. 7. D’Agostino, P.A., Provost, L.R. (1992) Determination of chemical warfare agents, their hydrolysis products and related compounds in soil. Journal of Chromatography, 589, 287– 294. 8. Barr, J.R., Driskell, W.J., Aston, L.S., Martinez, R.A. (2004) Quantitation of metabolites of the nerve agents sarin, soman, cyclohexylsarin, VX, and Russian VX in human urine using isotope-dilution gas chromatography-tandem mass spectrometry. Journal of Analytical Toxicology, 28, 372–378. 9. Degenhardt, C.E., Plieijster, K., van der Schans, M.J., Langenberg, J.P., Preston, K.E., Solano, M.I. et al. (2004) Improvements of the fluoride reactivation method for the verification of nerve agent exposure. Journal of Analytical Toxicology, 28, 364– 371.
10. Enqvist, J., Rautio, M. (eds) (1977) A. General Reports. A.1 Chemical and Instrumental Verification of Organophosphorus Warfare Agents, The Ministry for Foreign Affairs of Finland, Helsinki. 11. D’Agostino, P.A., Hancock, J.R., Provost, L.R. (1999) Analysis of O-ethyl S-[2-diisopropylamino)ethyl] methylphosphonothiolate (VX) and its degradation products by packed capillary liquid chromatographyelectrospray mass spectrometry. Journal of Chromatography A, 837, 93 –105. 12. D’Agostino, P.A., Hancock, J.R., Provost, L.R. (1999) Packed capillary chromatography-electrospray mass spectrometry analysis of organophosphorus chemical warfare agents. Journal of Chromatography A, 840, 289–294. 13. Smith, J.R. (2004) Analysis of the enantiomers of VX using normalphase chiral liquid chromatography with atmospheric pressure chemical ionization-mass spectrometry. Journal of Analytical Toxicology, 28, 390– 392. 14. Smith, J.R., Shih, M. (2001) Analysis of the degradation compounds of chemical warfare agents using liquid chromatography/mass spectrometry. Journal of Applied Toxicology, 21, S27– S34. 15. Bell, A.J., Murrell, J., Timperley, C.M., Watts, P. (2001) Fragmentation and reactions of two isomeric o-alkyl s-(2-dialkylamino)ethyl methylphosphonothiolates studied by electrospray ionization/ion trap mass spectrometry. Journal of the American Society for Mass Spectrometry, 12, 902–910. 16. Ellis-Steinborner, S., Ramachandran, A., Blanksby, S.J. (2006) The fragmentation pathways of protonated Amiton in the gas phase: towards the structural characterisation of organophosphorus chemical warfare agents by electrospray ionisation tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 20, 1939– 1948. 17. Weissberg, A., Tzanani, N., Dagan, S. (2013) Specificity enhancement by electrospray ionization multistage mass spectrometry—a valuable tool for differentiation and identification of ‘V’-type chemical warfare agents. Journal of Mass Spectrometry, 48, 1340–1348. 18. D’Agostino, P.A., Chenier, C.L. (2010) Desorption electrospray ionization mass spectrometric analysis of organophosphorus chemical warfare agents using ion mobility and tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 24, 1617–1624. 19. D’Agostino, P.A., Hancock, J.R., Provost, L.R. (2001) Determination of sarin, soman and their hydrolysis products in soil by packed capillary liquid chromatography–electrospray mass spectrometry. Journal of Chromatography A, 912, 291– 299. 20. Groenewold, G.S., Appelhans, A.D., Gresham, G.L., Olson, J.E., Jeffery, M., Wright, J.B. (1999) Analysis of VX on soil particles using ion trap secondary ion mass spectrometry. Analytical Chemistry, 71, 2318–2323. 21. Groenewold, G.S., Appelhans, A.D., Gresham, G.L., Olson, J.E., Jeffery, M., Weibel, M. (2000) Characterization of VX on concrete using ion trap secondary ionization mass spectrometry. Journal of the American Society for Mass Spectrometry, 11, 69 –77. 22. Beuck, S., Schwabe, T., Grimme, S., Schlorer, N., Kamber, M., Schanzer, W. et al. (2009) Unusual mass spectrometric dissociation pathway of protonated isoquinoline-3-carboxamides due to multiple reversible water adduct formation in the gas phase. Journal of the American Society for Mass Spectrometry, 20, 2034– 2048. 23. Guan, Z., Liesch, J.M. (2001) Solvation of acylium fragment ions in electrospray ionization quadrupole ion trap and Fourier transform ion cyclotron resonance mass spectrometry. Journal of Mass Spectrometry, 36, 264– 276.
Fragmentation Pathways of 14 Nerve Agent Compounds 105
Article
Fragmentation Pathways and Structural Characterization of 14 Nerve Agent Compounds by Electrospray Ionization Tandem Mass Spectrometry Kathleen J. Housman, Austin T. Swift and Jonathan M. Oyler* US Army Medical Research Institute of Chemical Defense, 3100 Ricketts Point Road, Aberdeen Proving Ground, MD 21010, USA *Author to whom correspondence should be addressed. Email: [email protected]
Organophosphate nerve agents (OPNAs) are some of the most widely used and proliferated chemical warfare agents. As evidenced by recent events in Syria, these compounds remain a serious military and terrorist threat to human health because of their toxicity and the ease with which they can be used, produced and stored. There are over 2,000 known, scheduled compounds derived from common parent structures with many more possible. To address medical, forensic, attribution, remediation and other requirements, laboratory systems have been established to provide the capability to analyze ‘unknown’ samples for the presence of these compounds. Liquid chromatography/mass spectrometric methods have been validated and are routinely used in the analysis of samples for a very limited number of these compounds, but limited data exist characterizing the electrospray ionization (ESI) and mass spectrometric fragmentation pathways of the compound families. This report describes results from direct infusion ESI/MS, ESI/MS2 and ESI/MS3 analysis of 14 G and V agents, the major OPNA families, using an AB Sciex 4000 QTrap. Using a range of conditions, spectra were acquired and characteristic fragments identified. The results demonstrated that the reproducible and predictable fragmentation of these compounds by ESI/MS, ESI/MS2 and ESI/MS3 can be used to describe systematic fragmentation pathways specific to compound structural class. These fragmentation pathways, in turn, may be useful as a predictive tool in the analysis of samples by screening and confirmatory laboratories to identify related compounds for which authentic standards are not readily available.
Introduction Organophosphate nerve agents (OPNAs) are highly toxic compounds which act as cholinesterase inhibitors. According to the Nuclear Threat Initiative (NTI), the most widely used and proliferated weapons of mass destruction have been chemical weapons in large part because of the ease with which they can be produced, stored and weaponized, and the reduced production and storage costs as compared with biological or nuclear weapons (NTI: Understanding Chemical Threats Home Page. http://www.nti.org/threats/chemical/, accessed 5 November 2013). Since 1997, the 188 member countries of the Organization for the Prohibition of Chemical Weapons (OPCW) have agreed to ban the production, stockpiling and use of chemical warfare agents (CWA) in accordance with the Chemical Weapons Convention (CWC) (1). The list of known scheduled target compounds is comprised of over 2,000 compounds many of which are derived from common parent structures (2). The theoretical possibilities for nerve agents derived from the Schedule 1.A.1 and 1.A.3 core structures exceed 270000 (3). While member countries had, as of 31 December 2011, eliminated 51.5 metric tons of their declared
71.2 metric tons of Category 1 CWA stockpiles, the remaining stockpiles, undeclared stockpiles possessed by nonparticipating states and non-state actors and frequently discovered new agent caches continue to represent a potential threat to human health through both military and terrorist use. To verify treaty compliance and investigate potential events, the OPCW has established a proficiency testing program which maintains an analytical system comprised of 22 expert-designated laboratories and a mobile laboratory (4). In addition to the OPCW laboratory system, several other laboratory systems have been established to address this potential threat. The Laboratory Response Network (LRN) was established by the US Department of Health and Human Services, Centers for Disease Control and Prevention (Centers for Disease Control and Prevention (CDC) Emergency Preparedness and Response: The LRN Partners in Preparedness Home Page. http://www.bt.cdc. gov/lrn/, accessed 6 November 2013) in collaboration with the US Department of Homeland Security (DHS), the US Federal Bureau of Investigation (FBI) and the Association of Public Health Laboratories (APHL). It is responsible for providing analytical capabilities to address potential chemical and biological terrorist events occurring within the USA. The LRN strategy is to maintain a national and international system of laboratories capable of rapidly identifying agents that threaten public health. The system links state and local public laboratories with veterinary, agriculture, military, environmental and food-testing laboratories. Other federal organizations including the FBI, the US Environmental Protection Agency (EPA) and branches of the US Department of Defense (DoD) have also put in place both fixed-site and field laboratory systems to provide diagnostic, forensic and verification analyses of clinical and environmental samples for these target analytes. For this reason, it is imperative that highly specific and robust analytical methods for identifying intact OPNAs be developed and validated to verify the presence of CWAs in unknown samples. Environmental and biological samples are routinely analyzed by gas chromatography – mass spectrometry (GC – MS) for CWAs using both electron impact ionization (EI) (5 – 7) and chemical ionization (8, 9). The in-source fragmentation of an extensive list of CWAs has been rigorously characterized using GC-EI – MS and extensive spectral libraries have been acquired using standardized ionization conditions (NIST Chemistry WebBook, NIST Standard Reference Database Number 69, eds. P.J. Linstrom and W.G. Mallard, http://webbook.nist.gov, accessed 6 November 2013) (10). However, spectral differences across structural analogs may be subtle when the spectral profile is solely dependent on in-source fragmentation. This can make compound identification difficult when authentic reference standards are not available for comparison. While GC – MS has proven to be a robust instrumental approach for identifying and quantifying OPNAs, additional sample extraction steps are
Published by Oxford University Press 2014. This work is written by (a) US Government employee(s) and is in the public domain in the US.
typically required to partition compounds from relatively polar solvents like water into GC-friendly volatile organic solvents. These additional sample preparation steps are time consuming, costly and can be the source of additional analytical error. Furthermore, the identification of the more polar OPNA methylphosphonic hydrolysis products via GC requires an additional derivatization step. These compounds are not discussed in this report, but the benefit of liquid chromatography (LC) is that samples could be analyzed for the presence of both CWA(s) and their respective hydrolysis breakdown products. While providing highly specific spectra, the single quadrupole (GC– MS and LC–MS) method specificity is largely dependent on a single process, in-source fragmentation. Relatively controlled tandem mass spectrometry (MSn) processes can be used to systematically acquire compound-specific spectra as well as to characterize the fragmentation patterns of families of compounds. These fragmentation patterns can, in turn, be used to predict compound identity with greater confidence when comparison against authentic reference standards is not possible. Though GC/MS2 methods used to identify and measure OPNAs provide increased specificity through the use of MS2 fragmentation, they still require the additional sample extraction steps characteristic of GC methods, and derivatization is still required for target analytes of higher relative polarity and those containing multiple reactive substituent groups. LC separation coupled with MS and MS2 detection provides highly specific methods with sensitivities and analysis times comparable to those obtained with GC/MS and GC/MS2 analysis, respectively. These methods do not require that target analytes be isolated in GC-friendly, volatile organic solvents, and derivatization is not required for more polar analytes, resulting in more efficient and cost-effective analyses. Despite the fact that LC/MS is currently being used routinely to detect and identify CWAs and their degradation products in both environmental and biological samples employing both electrospray ionization (ESI) (5, 11, 12) and atmospheric pressure chemical ionization (APCI) (13, 14), little work has been published documenting the complete MS2 fragmentation pathways of OPNAs using these ionization methods. In 2001, Bell et al. characterized the ESI products and MS2 fragmentation pathways of VX and RVX using a three-dimensional ion trap (Thermo Finnigan, San Jose, CA) (15). In 2006 using ESI, EllisSteinborner et al. postulated the MS2 fragmentation pathway of the V analog, ethyl S-2-diethylaminoethyl methylphosphonothiolate (Amiton, VG) (16). Weissberg et al. used an AB Sciex 5500 QTrap (Foster City, CA) to perform MS2 and MS3 to characterize VX, RVX, VM and two additional V analogs (17). Results from these studies indicated that the MS fragmentation of these dialkylphosphonothiolates followed the same pathways. This paper further verifies these previously reported data. This study focused on the two main OPNA classes, the G and V analogs. The compounds analyzed are listed in Table I. Structurally, these classes can be easily distinguished as V analogs have a common aminoethylphosphonothiolate backbone, and G analogs are alkoxyphosphonofluorides with the exception of GA, which is a phosphoroamidocyanidate. Using ESI on an AB Sciex 4000 QTrap (Foster City, CA), the MS2 fragmentation pathways have been characterized for 14 OPNA compounds over a range of collision energies.
Table I Compound List Compound
CAS #
Name
Formula
GA (Tabun) GB (Sarin) GF (Cyclosarin) GD (Soman) G1 (GB Analog) G2 (GF Analog) G3 VM
77-81-6 107-44-8 329-99-7
Ethyl N, N-dimethylphosphoramidocyanidate Isopropyl methylphosphonofluoridate Cyclohexyl methylphosphonofluoridate
C5H11N2O2P C4H10FO2P C7H14FO2P
96-64-0 648-59-9
Pinacolyl methylphosphonofluoridate Methyl isopropylphosphonofluoridate
C7H16FO2P C4H10FO2P
NA
Methyl cyclohexylphosphonofluoridate
C7H14FO2P
1426-08-0 21770-86-5
Ethyl isopropylphosphonofluoridate Ethyl S-2-diethylaminoethyl methylphosphonothiolate Isobutyl S-2-diethylaminoethyl methylphosphonothiolate Ethyl S-2-dimethylaminoethyl methylphosphonothiolate Ethyl S-2-diisopropylaminoethyl methylphosphonothiolate Propyl S-2-diisopropylaminoethyl methylphosphonothiolate Isopropyl S-2-diisopropylaminoethyl methylphosphonothiolate Isobutyl S-2-diisopropylaminoethyl methylphosphonothiolate
C5H12FO2P C9H22NO2PS
RVX
159939-87-4
Vx
20820-80-8
VX
50782-69-9
V1
52364-45-1
V2
51446-23-2
V3
855307-83-4
C11H26NO2PS C7H18NO2PS C11H26NO2PS C12H28NO2PS C12H28NO2PS C13H30NO2PS
Table II Characteristic In-source Fragment Ions Compound
In-source fragment ion
GA GB GD GF G1 G2 G3 VM RVX Vx VX V1 V2 V3
135 (20) 99 (27), 79 (7), 97 (52), 81 (8), 117 (13) 99 (10), 97 (100), 85 (20) 99 (30), 79 (3), 97 (70), 81 (10), 117 (7) 83 (12) 83 (6) 127 (12), 67 (5) 167 (25), 139 (129), 79 (24), 97 (27), 100 (125), 72 (26), 44 (40) 212 (50), 100 (100), 139 (25), 61 (10) 139 (7), 72 (20) 240 (40), 128 (115), 86 (22), 167 (10), 139 (10), 79 (8) 240 (12), 128 (100), 86 (20), 181 (20), 121 (20), 97 (17), 139 (13) 240 (90), 128 (95), 86 (15), 181 (10), 121 (5), 79 (5), 97 (15), 139 (13) 240 (20), 128 (105), 86 (12), 195 (9)
Note. Value in parentheses indicates the percent abundance of each ion relative to the [MþH]þ ion.
Materials and methods Materials and sample preparation Reference standards for GA, GB, GD, GF and VX were acquired from the CASARM Program of the US Army Edgewood Chemical Biological Center (ECBC), Aberdeen Proving Ground (APG), MD, with purities verified at 95% by NIST-traceable 1 H NMR, 13C NMR, 31P NMR, GC/TCD, GC/MSD and/or acid – base titration methods. All other reference standards were acquired from the Chemical Sciences Division at ECBC with purities verified at 92% by 1H NMR, 13C NMR or 31P NMR. Dilute authentic standards were formulated in chloroform at ,2 mg/ mL for G agents and ,1 mg/mL for V agents. Samples were then diluted to 500 ng/mL in 3:1 methanol:water (v:v) with 0.1% formic acid prior to analysis. The methanol, water and formic acid used were LC/MS-grade and were purchased from Fisher Scientific (Waltham, MA).
Fragmentation Pathways of 14 Nerve Agent Compounds 97
Figure 1. Summed product ion scan spectra of the [MþH]þ for all G analogs over a CE range of 0 –120 V.
98 Housman et al.
Figure 2. Proposed CAD pathway for [GAþH]þ.
MS and MS2 experiments Samples were analyzed by direct infusion at 10 mL/min on an AB Sciex 4000 QTrap (Foster City, CA) hybrid triple quadrupole/linear ion trap (LIT) mass spectrometer using a standard Turbo V, ESI source operated in positive mode. Data acquisition and analysis were performed using the AB Sciex Analyst 1.5 software platform. The 4000 QTrap mass spectrometer is comprised four quadrupoles, Q0 – Q3. Q0 is a high pressure, RF ion-focusing region which can also function as a LIT to enhance the collection of ions for subsequent MS, MS2 and MS3 analysis. In this region, the declustering potential (DP), a voltage differential between the orifice plate and Q0, is used to minimize solvent ion pairing, thereby optimizing ion transfer into the mass spectrometer. At relatively higher voltages, it can be used to generate collisionally induced dissociation (CID) with the aid of atmospheric gas particles and the nitrogen curtain gas. Q1 functions as a mass filter which can be operated as an ion guide to transfer ions up to m/z 2800 (full scan) or select ions for further analysis in Q2. Q2 can function as a focusing region in which ions are cooled by a small amount of relatively inert gas or as a collision cell in which ions are accelerated to collide with the gas to produce controlled fragmentation, a process labeled collisionally activated dissociation (CAD). Q3 can operate as an ion guide for full scan analysis, as a mass filter to select ions or as a LIT to enhance ion collection for CAD MS3 analysis. Initially, optimal source parameters for the low flow, direct infusion experiments were identified for each compound and included a probe height of 0 mm, a capillary voltage of 5,500 V, a sheath gas pressure of 20 psi and a curtain gas pressure of 20
psi. At the flow rates used, the use of turbo heaters and desolvation gas impaired the ionization efficiency of all compounds, so neither was employed in subsequent experiments. A Q1 full scan experiment (40–1,000 m/z) was performed to characterize the ESI fragmentation of each compound and identify possible precursor ions to be used in MS2 analyses. Ion transfer efficiency into the mass filter is dependent on various voltage differentials applied to the ions. These differentials desolvate and guide ions while the ions transition from the relatively high pressures in the source and Q0 to the high vacuum in the mass filter. Q1 full scan data were summed over 10 cycles while holding the DP at 20, 60 or 100 V. Once the protonated molecule ([MþH]þ) for each compound was identified, the DP was stepped over a 0 –120 V range in 1 V increments while scanning for the ion to identify the optimal DP at which maximal [MþH]þ ion transfer into the mass spectrometer occurred. Using optimal DP values, the [MþH]þ was selected in Q1 for CAD fragmentation in Q2. Q3 full scan data (m/z 40 – 1,000) were collected while varying the collision energy (CE) and the CAD gas pressure to identify as many important CAD-generated product ions as possible. The rate at which ions selected in Q1 are transferred through the collision cell (Q2) is directly proportional to the CE, the voltage differential between Q0 and Q2. Precursor ions are fragmented as they collide with a collision gas; in this case nitrogen was used. Each precursor-to-product ion transition occurred optimally at a specific CE, and for this reason product ion spectra for each [MþH]þ were summed while stepping the CE over a 0 – 120 V range in 5 V increments to identify important CAD-generated product ions. Since CAD fragmentation is also directly proportional to the CAD gas pressure, the same CAD experiments were performed with the collision cell nitrogen set at three different gas pressures, 1.7 – 2.9, 3.7 –3.9 and 4.7 –4.9 psi (low, medium and high settings, respectively, in the Analyst 1.5 software).
Linear ion trap experiments The Q3 LIT capability of the 4000 QTrap was employed to enhance both sensitivity and resolution in the structural analysis of these compounds. Ions were trapped both radially and axially through voltages applied by the quadrupoles, and the entrance and exit barrier lenses, respectively. After enough ions were trapped, voltages were ramped, and ions scanned out of the LIT. Enhanced MS (EMS) scanning, in which Q1 and Q2 operate in RF-only mode to focus and transfer the ion beam directly into the LIT, was used to collect full scan data. Data were summed for 25 cycles, while holding CE at 10 V, and employing the dynamic fill time (DFT) function in which the software optimized the fill time based on relative ion abundance. Enhanced product ion (EPI) scanning (a MS2 full scan using the increased sensitivity and resolution of the LIT) was used to collect product ion data on a precursor ion selected in Q1, fragmented in Q2 and trapped and scanned out of the LIT. Using optimal DP values for each precursor, EPI scans for each precursor ion were summed using a fill time of either 100 or 250 ms while stepping the CE over a 0 –120 V range in 5 V increments. MS3 experiments were performed in Q3 on single product ions generated from Q2 CAD fragmentation of a precursor selected in Q1. In Q3, the characteristic product ion was isolated, and through the application of auxiliary voltages, excited and Fragmentation Pathways of 14 Nerve Agent Compounds 99
Figure 3. Proposed CAD pathways for [GBþH]þ, [GDþH]þ and [GFþH]þ.
fragmented through collision with residual nitrogen from Q2. To enhance nitrogen bleed into Q3 sufficient to produce MS3 fragmentation, the CAD gas pressure in Q2 was set at either medium or high. For compounds from which stable CAD-generated and structurally significant product ions were produced, MS3 scans were collected. MS3 experiments were performed on [MþH]þ and other precursor ions using their optimal DP values and by employing the following parameters: CE 10 – 20 V, fill time of 250 ms or the DFT function, excitation energies of 10 – 50 V and MS3 fragmentation excitation time of 100 or 250 ms. EMS and EPI experiments were performed at low, medium and high CAD gas pressures, and MS3 experiments were only performed using medium or high CAD gas. All LIT experiments were performed at a scan speed of 1,000 Da/s over a scan range of 50–1,000 m/z. Results and discussion ESI and in-source fragmentation Over a range of source voltages, ESI of each compound produced characteristic ions. For all 14 compounds, prominent [MþH]þ 100 Housman et al.
peaks were identified. ESI of the G analogs produced prominent sodium adduct ([MþNa]þ) and sodiated dimer ([2MþNa]þ) and trimer ([3MþNa]þ) peaks, and also produced protonated dimers ([2MþH]þ) for all G analogs with the exception of G1. Other than [MþH]þ ions and characteristic in-source fragment ions, no sodiated adducts or protonated or sodiated dimers or trimers were identified for any V analogs. Characteristic in-source fragments along with percent abundances, relative to the [MþH]þ ion, are listed in Table II. No multiply charged species were identified following direct infusion ESI of these compounds.
G analog CAD fragmentation Figure 1 contains spectra resulting from CAD analysis of the [MþH]þ for all G analogs and provides evidence for the majority of G analog fragmentation steps observed. Postulated fragmentation pathways based on results from CAD analysis of the [MþH]þ of each compound are illustrated in Figures 2 –4. CAD of all protonated G analogs initially produced characteristic fragments because of the loss of the alkoxy alkyl group, and CAD of
Minimal CAD fragmentation was observed for the protonated phosphonofluoridates G1, G2 and G3 (Figure 4). The only characteristic product ions resulting from CAD of protonated G1 and G2 were m/z 99 and 83. It is likely that the formation of the m/z 99 ion resulted from a cleavage of the isopropyl and cyclohexyl groups from G1 and G2, respectively, and that the m/z 83 ion common to both formed as a result of subsequent cleavage of the methyl from the methoxy group. In addition, CAD of the CID-generated ion, m/z 99, from G1 and G2 produced characteristic m/z 83 fragment ions. Two prominent characteristic fragments at m/z 127 and 67 were identified from CAD of [G3þH]þ, an isopropyl phosphonofluoridate, originating from loss of its alkoxy ethyl group and its primary isopropyl and hydroxyl groups, respectively. Similar to GD, CAD of [G1þH]þ and [G3þH]þ also produced prominent m/z 43 peaks resulting most likely from protonation of the cleaved primary propylene group, and a prominent m/z 55 peak was produced from CAD of [G2þH]þ through a mechanism similar to that proposed for CAD of its cyclohexyl phosphonofluoridate analog, GF.
Figure 4. Proposed CAD pathways for [G1þH]þ, [G2þH]þ and [G3þH]þ.
proton-bound dimers produced fragmentation patterns similar to those observed following ESI-CAD of the corresponding [MþH]þ. Because of its unique phosphoroamidocyanidate structure, the Q3 full scan spectrum resulting from CAD of [GAþH]þ shared no common product ions with spectra collected from CAD of the other protonated G analogs (Figure 2). Initially, GA fragmentation produced an ion at m/z 135 due to the loss of the alkoxy ethyl group. This ion further fragmented producing an ion at m/z 117 resulting from loss of water and m/z 108 resulting from loss of the cyano group. The observed m/z 135, 117 and 108 GA fragment ions were consistent with those reported by D’Agostino and Chenier in 2010 (18). Relative abundance increases for the m/z 126 ion with corresponding increases in CAD energy during MS2 of [GAþH]þ coupled with an absence of the m/z 126 ion in product ion scans from negative blank solvent samples supported the formation of a water adduct. Subsequent CAD of in-source produced m/z 108 ions produced a prominent m/z 126 peak resulting from a gain of water which confirmed adduct formation; we also observed the formation of doublet water adducts at m/z 144. CAD of protonated GB, GD and GF (Figure 3), induced the loss of the alkoxy alkyl group and produced a characteristic ion at m/z 99, owing to their core methyl phosphonofluoridate structure. This was also previously reported by D’Agostino et al. in 2001 and 2010 (18, 19). Similar prominent fragment ions at m/z 81, 117, 79 and 97 were also observed for all three compounds. Since GD had the largest alkyl side chain, CAD of [GDþH]þ (Figure 1) produced three additional characteristic protonated alkene fragments at m/z 85, 57 and 43 as reported by D’Agostino and Chenier (18). A prominent m/z 55 peak was also produced by CAD of [GFþH]þ most likely originating from protonation of a fragment of the cleaved primary cyclohexene group.
V analog CAD fragmentation V series analog CAD occurred through two main mechanisms (Figure 5), and because V analogs have common aminoethylphosphonothiolate backbones, summed spectra from CAD of the [MþH]þ of each of the seven V analogs resulted in many similar product ions (Table III and Figure 6). CAD fragmentation patterns for V analogs through the two pathways we propose are consistent with those previously proposed by Bell et al. in their study of VX and RVX (17), Ellis-Steinborner et al. in work with Amiton (VG) (18) and Weissberg et al. in their mass spectrometric characterization of several V analogs (19). The fragmentation pathway for VX proposed in this study also agrees with the fragmentation of VX extracted from soil and concrete particles by ion trap secondary ion MS2 observed by Groenewold et al. in 1999 (20, 21). Cleavage appeared to occur at two locations along the ethyl bridge corresponding to neutral losses of the methylphosphono group (A) and dialkylamine (B). The first fragmentation pathway was initiated at low relative CAD energies by the neutral loss of the alkoxy alkyl (Figure 5, A1) as confirmed by the presence of a m/z 212 peak in the CAD full scan spectrum of [RVXþH]þ and m/z 240 peaks in the CAD full scan spectra of protonated VX, V1, V2 and V3. With increasing CAD energy, the methylphosphono group (Figure 5, A2) was subsequently cleaved as evidenced by the presence of a m/z 134 peak in the CAD scan spectrum of [RVXþH]þ and m/z 162 peaks in the CAD scan spectra of protonated V1, V2 and V3. There was no evidence that CAD of [VMþH]þ or [VxþH]þ led to the loss of either the alkoxy alkyl- or methylphosphono groups prior to the formation of aziridine or imminium ions, and no peak could be identified following CAD of [VXþH]þ that corresponded to the loss of the methylphosphono group. The final step in Pathway A was the removal of sulfur and formation of either an alkylated aziridine ion or an imminium ion (Figure 5, A3). Alkylated aziridine or imminium ions and subsequent fragment ions appeared as m/z 100, 72 and 44 peaks in CAD-generated full scan spectra of protonated VM, RVX and Vx and as m/z 128, 86 and 44 peaks following CAD of protonated VX, V1, V2 and V3. Fragmentation Pathways of 14 Nerve Agent Compounds 101
Figure 5. Proposed CAD pathways for protonated V analogs. Pathway A: neutral loss of alkoxy R-group (A1) followed by cleavage of methylphosphono group (A2) and formation of the aziridine ion (A3). Pathway B: neutral loss of dialkylamine (B1) followed by loss of the alkylated thiol group (B2) and/or cleavage of the alkoxy R-group (B3) and subsequent fragmentation. Table III Prominent Characteristic Ions Produced from CAD MS2 of Protonated V Analogs VM
RVX
Vx
VX
V1
V2
V3
[MþH]þ
240
268
212
268
282
282
296
Pathway A A1 A2 A3 A4 A5
– – 100 72 44
212 134 100 72 44
– – 72 – 44
240 – 128 86 44
240 162 128 86 44
240 162 128 86 44
240 162 128 86 44
Pathway B B1 B2 B3 B4 B5 B6
167 107 139 79 97 61
– – 139 – – 61
167 107 139 79 97 61
167 107 139 79 97 61
181 121 139 79 97 61
181 121 139 79 97 61
195 135 139 79 97 61
Structures
102 Housman et al.
Figure 6. Summed product ion scan spectra of the [MþH]þ for all V analogs over a CE range of 0– 120 V.
Fragmentation Pathways of 14 Nerve Agent Compounds 103
The initial step in the second major pathway (Figure 5, B1) resulted from the neutral loss of dialkylamine. It is proposed that this loss would likely result in the formation of either a thiiranium ion or an ion resulting from formation of an ethylene bridge between the sulfur and phosphono oxygen. This was supported by the identification of m/z 167 (VM, Vx and VX), 181 (V1 and V3) and 195 (V3) peaks. No stable product ion was observed following CAD of [RVXþH]þ indicative of this cleavage. Subsequent fragmentation followed two additional paths initiated by elimination of the alkylated thio group (Figure 5, B2) or cleavage of the alkoxy alkyl group (B3). Ions resulting from the loss of the alkylated thio group and corresponding with the number of carbons in the alkoxy alkyl group were identified for VM, Vx and VX as peaks at m/z 107, for V1 and V2 as peaks at m/z 121 and for RVX and V3 as peaks at m/z 135. CAD of all protonated V analogs produced successive m/z 139, 79, 97 and 61 fragment ions with the exception of [RVXþH]þ which did not produce m/z 79 or 97 peaks. Water adduction of all the V analogs to form methyl phosphonic acid was also observed.
LIT experiments Summed spectra resulting from EMS and EPI scanning were similar to those described above for all G and V analogs, further substantiating the proposed fragmentation pathways. Somewhat unexpectedly, product ion absolute abundances resulting from EPI scanning were reduced compared with those acquired using ‘pure quadrupole’ MS2 methods. This suggested that these ions were not efficiently retained in Q3 during the trapping process or that further fragmentation of resonant ions occurred in Q3 as a result of collision with CAD gas bleeding from Q2 during trapping. Table IV lists ions identified from EPI scanning of all protonated G and V analogs. While prominent precursor and product ions were identified in LIT experiments, many of the lower mass CAD ions did not appear in these spectra. No characteristic product ions were identified following MS3 fragmentation of the G analogs. This could be due to these compounds having reduced bond energy distributions compared with the V analogs. Each of the seven V analogs produced MS3 transitions correlating with at least one of the two aforementioned pathways which also supported the proposed two-step fragmentation pathways. As expected, product ion absolute abundances were reduced using these techniques compared with those acquired using MS2 techniques, and confirmation of some multi-step fragmentation was limited by the ‘one-third rule’ commonly associated with ion trapping (Table V). The m/z 162 ion was observed in MS3 spectra of the m/z 240 product of protonated V1, V2 and V3. In addition, m/z 86 was seen in MS3 spectra of the m/z 128 product from protonated VX, V2 and V3 which further confirmed the postulated fragmentation in Figure 5A. MS3 fragmentation of the m/z 212 product of protonated RVX produced m/z 134 and 100; this was previously reported by Weissberg et al. (2013) using a 5500 QTrap (17). In the case of VM, Vx and VX, an m/z 139 ion was observed in MS3 fragmentation of the product ion m/z 167 from m/z 212. MS3 fragmentation of the m/z 181 product from protonated V1 and V2 and the m/z 196 product from protonated V3 also resulted in the formation of an m/z 139 ion. This further confirms the proposed fragmentation pathway described in Figure 5B. In 2013, Weissberg et al. provided more in-depth MS3 characterization 104 Housman et al.
Table IV LIT EPI Scan-Generated Ions from Protonated Molecular Ions Compound
EPI ions
Compound
EPI Ions
GA GB GD GF G1 G2 G3
135, 117, 108, 126 99, 79, 97, 81 99, 79, 81 99, 79, 81 99 99, 83 127
VM RVX Vx VX V1 V2 V3
167, 139, 97, 100, 72 212, 134, 100, 72, 139 72, 167, 139, 79, 97 128, 86, 139, 79, 97 240, 128, 86, 181, 79, 97, 139 240, 162, 128, 121, 139 240, 162, 128, 86
Table V MS3 Transitions Observed in the Analysis of V Analogs Compound
MS3 confirmation Pathway A
VM RVX Vx VX V1 V2 V3
– 268/212/134 – 268/128/86 282/240/162 282/240/162 296/240/162
Pathway B – 268/212/100 – – – 268/128/86 296/128/86
212/167/139 – 212/167/139 212/167/139 282/181/139 282/181/139 296/195/139
of VX, RVX, VM and several additional V analogs using a 5500 QTrap, a more sensitive instrument with a faster LIT scan rate (17).
Conclusions The results presented in this study demonstrate that ESI/MS, ESI/ MS2 and ESI/MS3 can be used to structurally identify the individual compounds evaluated and substantiate previously published data for a very limited number of these compounds using similar methods. The results also demonstrate that the reproducible and predictable fragmentation of these compounds by ESI/MS, ESI/ MS2 and ESI/MS3 can be used to describe systematic fragmentation pathways specific to compound structural class. We observed the unique formation of water adducts during CAD fragmentation of GA and all V analogs with the exception of RVX. Although water adduct formation during CAD fragmentation of VX and RVX (15) and other small molecules (22, 23) has previously been reported by researchers using a threedimensional ion trap and an ion cyclotron, respectively, water adduct formation during CAD in a hybrid quadrupole LIT instrument has not been reported previously with these compounds. Due to security regulations agreed to by OPCW signatories that limit access to these compounds, few synthetic labs exist from which authentic standards can be acquired for a wide variety of the structurally related analogs. As predictive reference tools, the pathways described here could be used in laboratories charged with ‘unknown’ sample analysis to initially screen for CWA analogs for which authentic standards may not be readily available; since these compounds generally have very short clinical half-lives, use of these pathways as predictive tools would likely be limited to the analysis of environmental samples. The data presented will also expand the method-specific CWA spectral library, and this work, in addition to previously published
work, should serve as a reference for researchers and analysts in the treaty compliance, public health, forensic and defense communities. Acknowledgments The views expressed in this article are those of the author(s) and do not reflect the official policy of the Department of Army, Department of Defense or the U.S. Government. This research was supported by the Defense Threat Reduction Agency—Joint Science and Technology Office, Medical S&T Division. References 1. Convention on the Prohibition of the Development, Production, Stockpiling, and Use of Chemical Weapons and on Their Destruction. (2005). Technical Secretariat of the Organisation of Prohibition of Chemical Weapons (OPCW). The Hague, the Netherlands. http://www.opcw.org/index.php?eID=dam_frontend _push&docID=6357.pdf (6 November 2013, date last accessed). 2. Nyanyira, C. (2006) Sampling and analysis in the Chemical Weapons Convention and the OPCW Mobile Laboratory. In: Mesilaakso, M. (ed). Chemical Weapons Convention Chemicals Analysis: Sample Collection, Preparation and Analytical Methods, 1st edition, Chapter 2. Chichester, UK: John Wiley & Sons, Ltd, 133 pp. 3. Mogl, S. (2006) Sampling and analysis in the Chemical Weapons Convention and the OPCW Mobile Laboratory. In: Mesilaakso, M. (ed). Chemical Weapons Convention chemicals analysis: Sample collection, preparation and analytical methods, 1st edition, Chapter 2. Chichester, UK: John Wiley & Sons, Ltd, 7 pp. 4. Report of the Scientific Advisory Board on developments in science and technology for the Third Special Session of the Conference of the States Parties to Review the Operation of the Chemical Weapons Convention (2012) OPCW RC-3/DG.1. Organisation for the Prohibition of Chemical Weapons (OPCW) Third Review Conference, The Hague, The Netherlands. http://www.opcw.org/index.php?eID= dam_frontend_push&docID=15865.pdf (6 November 2013, date last accessed). 5. Koller, M., Becker, C., Thiermann, H., Worek, F. (2010) GC – MS and LC–MS analysis of nerve agents in body fluids: Intra-laboratory verification test using spiked plasma and urine samples. Journal of Chromatography B, 878, 1226– 1233. 6. D’Agostino, P.A., Porter, C.J. (1992) Capillary column gas chromatography/tandem mass spectrometry verification of chemical warfare agents. Rapid Communications in Mass Spectrometry, 6, 717–718. 7. D’Agostino, P.A., Provost, L.R. (1992) Determination of chemical warfare agents, their hydrolysis products and related compounds in soil. Journal of Chromatography, 589, 287– 294. 8. Barr, J.R., Driskell, W.J., Aston, L.S., Martinez, R.A. (2004) Quantitation of metabolites of the nerve agents sarin, soman, cyclohexylsarin, VX, and Russian VX in human urine using isotope-dilution gas chromatography-tandem mass spectrometry. Journal of Analytical Toxicology, 28, 372–378. 9. Degenhardt, C.E., Plieijster, K., van der Schans, M.J., Langenberg, J.P., Preston, K.E., Solano, M.I. et al. (2004) Improvements of the fluoride reactivation method for the verification of nerve agent exposure. Journal of Analytical Toxicology, 28, 364– 371.
10. Enqvist, J., Rautio, M. (eds) (1977) A. General Reports. A.1 Chemical and Instrumental Verification of Organophosphorus Warfare Agents, The Ministry for Foreign Affairs of Finland, Helsinki. 11. D’Agostino, P.A., Hancock, J.R., Provost, L.R. (1999) Analysis of O-ethyl S-[2-diisopropylamino)ethyl] methylphosphonothiolate (VX) and its degradation products by packed capillary liquid chromatographyelectrospray mass spectrometry. Journal of Chromatography A, 837, 93 –105. 12. D’Agostino, P.A., Hancock, J.R., Provost, L.R. (1999) Packed capillary chromatography-electrospray mass spectrometry analysis of organophosphorus chemical warfare agents. Journal of Chromatography A, 840, 289–294. 13. Smith, J.R. (2004) Analysis of the enantiomers of VX using normalphase chiral liquid chromatography with atmospheric pressure chemical ionization-mass spectrometry. Journal of Analytical Toxicology, 28, 390– 392. 14. Smith, J.R., Shih, M. (2001) Analysis of the degradation compounds of chemical warfare agents using liquid chromatography/mass spectrometry. Journal of Applied Toxicology, 21, S27– S34. 15. Bell, A.J., Murrell, J., Timperley, C.M., Watts, P. (2001) Fragmentation and reactions of two isomeric o-alkyl s-(2-dialkylamino)ethyl methylphosphonothiolates studied by electrospray ionization/ion trap mass spectrometry. Journal of the American Society for Mass Spectrometry, 12, 902–910. 16. Ellis-Steinborner, S., Ramachandran, A., Blanksby, S.J. (2006) The fragmentation pathways of protonated Amiton in the gas phase: towards the structural characterisation of organophosphorus chemical warfare agents by electrospray ionisation tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 20, 1939– 1948. 17. Weissberg, A., Tzanani, N., Dagan, S. (2013) Specificity enhancement by electrospray ionization multistage mass spectrometry—a valuable tool for differentiation and identification of ‘V’-type chemical warfare agents. Journal of Mass Spectrometry, 48, 1340–1348. 18. D’Agostino, P.A., Chenier, C.L. (2010) Desorption electrospray ionization mass spectrometric analysis of organophosphorus chemical warfare agents using ion mobility and tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 24, 1617–1624. 19. D’Agostino, P.A., Hancock, J.R., Provost, L.R. (2001) Determination of sarin, soman and their hydrolysis products in soil by packed capillary liquid chromatography–electrospray mass spectrometry. Journal of Chromatography A, 912, 291– 299. 20. Groenewold, G.S., Appelhans, A.D., Gresham, G.L., Olson, J.E., Jeffery, M., Wright, J.B. (1999) Analysis of VX on soil particles using ion trap secondary ion mass spectrometry. Analytical Chemistry, 71, 2318–2323. 21. Groenewold, G.S., Appelhans, A.D., Gresham, G.L., Olson, J.E., Jeffery, M., Weibel, M. (2000) Characterization of VX on concrete using ion trap secondary ionization mass spectrometry. Journal of the American Society for Mass Spectrometry, 11, 69 –77. 22. Beuck, S., Schwabe, T., Grimme, S., Schlorer, N., Kamber, M., Schanzer, W. et al. (2009) Unusual mass spectrometric dissociation pathway of protonated isoquinoline-3-carboxamides due to multiple reversible water adduct formation in the gas phase. Journal of the American Society for Mass Spectrometry, 20, 2034– 2048. 23. Guan, Z., Liesch, J.M. (2001) Solvation of acylium fragment ions in electrospray ionization quadrupole ion trap and Fourier transform ion cyclotron resonance mass spectrometry. Journal of Mass Spectrometry, 36, 264– 276.
Fragmentation Pathways of 14 Nerve Agent Compounds 105
