MethodsExperimental The experimentaldata is obtained using collision induced dissociation (CID)in a LCQ Deca plusXP ion trap mass spectrometer.
Dilute samples (10-4 M) of the methyladeninespecies are prepared in a 1:1 distilled water:methanol solution (HPLC grade H2O,Sigma Aldrich, HPLC grade methanol, Fischer Scientific). 1, 3, and 7- methyladenine (98% Sigma Aldrich 1-methyladenine,Sigma Aldrich 3-methyladenine autophagy inhibitor, 97% Fischer Scientific7-methyladenine) are deprotonated by electrospray ionization (ESI). Theexperimental layout of the ESI process is shown in Figure 2 and the expectedstructures are displayed in Figure 1. A potential difference of 4.
5 kV isapplied between the spray nozzle and the heated ion transfer tube (250 °C)where the molecules are de-solvated, and a sample flow rate is maintained at 15?L min–1. The deprotonated methyladenine ions travel from the ESIsource and are focused by an octapole and quadrupole into the ion trap (Figure4) where the ions then collide with helium at a pressure of 0.177 Torr.This pressure is determined through a calibration process described in ourprevious work.11 Tandem mass spectrometry (MSn) is usedin order to isolate a desired ion and conduct additional fragmentation.
Thenormalized collision energy (NCE) is varied up to 50% in order to fullyfragment a specific ion within the trap. In addition, the activation quotientand activation times are set to .250 and 30 ms, respectively. Figure 2. Theschematic shown above describes the experimental design of the electrosprayionization process. 12 The schematic shows a Taylor cone containingthe anions being drawn into the heated ion transfer tube. From the ion transfertube molecules are de-solvated and focused through a quadrupole and octapoleinto the ion trap.
Computational Gaussian 09 software is utilized to further analyze the fragments of themethyladenine species produced by CID. Density functional theory is also employedin conjunction with Gaussian 09 at the B3LYP/6-311++G(d,p) level oftheory. Potential energy surfaces (PES) and acidity calculations are obtainedby optimization of molecular geometries alongside frequency calculations of thelocal minima and maxima of the PES, respectively. Gas-phase acidities arecalculated and reported as the change in the Gibbs free energy for reaction at298K.
Associated zero point energycorrections are applied. Figure 3. The staggeredspectra above displays the unique experimental spectrum for each of themethyladenine compounds. Shown in red, is the deprotonated anionic form (parention) of the functionalized purine molecule occurring at m/z 148 for all themethyladenine species. In all of the above spectra, a peak at m/z 133 wasdetected. This peak is characteristic of a commonly seen loss of methylradical. In addition to the peaks at m/z 133, 1-methyladenine, shown in black,displays a unique peak at m/z 107 likely resulting from the loss of CH3CN.
Seen in blue, the 7-methyladenine also displays a unique peak, this timeappearing at m/z 121. This peak is indicative of a loss of HCN. For all of the above spectra with theexception of the parent ion peak a normalized collision energy of 45% wasapplied, with an activation quotient and time of .
250 and 30 ms, respectively. Results And Discussion Herein, wereport the gas phase acidities, spectral analysis and density functional theory(DFT) computations regarding collision induced dissociation of 1, 3, and7-methyladenine within the LCQ Deca plus XP Quadrupole ion trap massspectrometer. Determined computationally, the gas phase acidities of the purinespecies led to the proposed anionic structure (Figure 1). These geometry optimizationsdescribed the most stable structures to be 1, and 7-methyladenine deprotonatedat their N9 position while 3-methyladenine experiences effective deprotonationat it N13 site. It is important to note that acidity calculations wereperformed for all potential sites including C-H bonds. Although deprotonationof the C-H bond is possible upon ESI,in the differing species, the occurrence of C-H deprotonation led to relativelyhigh energy conjugate bases.
Upon collision induced dissociation within the iontrap (Figure 4) each of the methyladenine compounds displayed unique spectraldifferences for each of the anionic molecules produced (Figure 3). In red isthe distinct peak corresponding to the “parent ion”, or deprotonated methyladeninecompound, occurring at m/z 148 in accordance with the mass of each ion. While eachspectrum is unique, similarities are identifiable in each. Primarily, eachspecies of ion displays a peak at m/z 133. This peak corresponds to a loss of15 AMUs and can most reasonably be explained by the loss of the methylsubstituent in its radical form. In addition to the peaks at m/z 133,1-methyladenine, shown in black, displays a unique peak at m/z 107. This peakrepresents a loss of 41 AMUs and likely results from the fragmentation of CH3CN.Seen in blue, the 7-methyladenine also displays a unique peak appearing at m/z121.
The mass of the fragment represented by this peak is likely indicative ofthe parent ion having a lost an HCN molecule. It is considered highly likelythat the unique losses may be isomers of CH3CN and HCN seen inprevious literature.12,14,15 All of the spectra were obtained at anunusually low normalized collision energy value (45%). Cole et al2 describescomplete dissociation to occur well over an NCE value of 50 percent.
Thisanomaly is likely to be caused by the high coefficient of pressure (.177 torr) withinthe ion trap, thus causing more frequent and higher energy collisions. Figure 4. Theschematic above describes of the focusing of ions by a quadrupole and octapoleinto the ion trap. Once inside the trap, the ions undergo collision induceddissociation with helium gas. This process results in anionic fragments thatcan then be detected by an electron multiplier. Computations revealed the potential energy surfaces for the loss of methylradical for each of the methyladenine molecules (Figure 5). The dissociationevents are considered to take place without a transition state and occur atunusually high energys.
For 1-methyladenine the energy of the resultantfragments lies at approximately 92.79 kcal·mol-1 whereas3-methyladenine exhibited a slighly lower energy of 91.2 kcal·mol-1. The lowest energy severance is shown in 7-methyladeninearound 87.67 kcal ·mol-1. Excited state intermolecular protontranfers may be key in the detatchment of methyl radical upon CID as Protontransfers have been previously noted as the only method for certain reactions tooccur (Freeman et al. 1978).
It has been postulated that the formation oftetrahedral carbons within the ion trap drastically change the energy in whichsubstituents undergo any dissociation event. In this case, it is likely that anintermolecular proton transfer occurs from C2 to N5, C2 to N8, and C6 to N1 in1, 3, and 7-methyladeinine respectively, in order to weaken the CH3-purinebond. Figure 5.
Exhibited above are the energetics involved in the loss of methylradical from each methyladenine derivative. Appearing from left to right are 7-3- and 1-methyladenine with dissociation energies of 87.67, 91.22, 92.79 kcalmol-1, respectively.
1-Methyladenine 1-Methyladenine displaysthe unique decay of the parent ion peak (m/z 148) into peaks at m/z 133 and m/z107 upon NCE values increasing from 25-45 percent (Figure 6). The large peakappearing at m/z 107 indicates a mass of 41 AMUs and is likely the result of aseverance event including CH3CN, and resulting in the detectedmolecule C3H3N3-. Although eachspectrum contains a peak at m/z 133, a large aberration in the intensity isclear in the spectrum for 1-methyladenine juxtaposed to that of 3- and 7- methyladenine.Elucidation of the mechanism behind the unusual peak intensity is achieved bycomputationally modeling the lowest energy path for the loss of CH3CNcompared to that of the methyl radical (Figures 5 & 7.) In spite of a .
4isolation width, a shoulder is detected at m/z 108 and m/z 149 due to thepossible formation of a radical 1-methyladenine species during the ESI process. Figure 6. The fragmentation pattern of1-methyladenine with respect to increasing normalized collision energy (NCE %)is defined. Loss of the CH3CN fragment is emphasized. A peakcorresponding to the loss of methyl radical is seen developing in the center. Ashoulder is detected at 108 and 149 m/z due to the possible formation of aradical 1-methyladenine species. A potential energy surfacedescribing the energetics involved in the complete dissociation of the acetonitrilegroup from 1-methyladenine is obtained computationally (Figure 7). Thefragmentation event is first facilitated by the migration of the methylsubstituent to the C2 position forming a tetrahedral carbon and destabilizingthe bond between C2 and N5.
This transition state of approximately 65 kcal·mol-1yields a relatively low energy geometry optimization of 35 kcal·mol-1relative to the methyladenine parent. The now weakened C-N bond then undergoesa ring opening transition state occurring at 75.51 kcal·mol-1. Theresultant open ring compound rests at an energy of 55.59 kcal·mol-1.At this juncture in the computing process several routes for the transfer ofH13 were explored. Transition states of reasonable energy were found for the transferof H13 to carbons 1, 3, and 5. In addition, proton transfers were conducted tonitrogens 5 and 9.
The optimized geometric structures of all the newlyprotonated carbon species yielded unreasonable energies when juxtaposed to theion trap. All computed energies must fall within a rangethat is accessible in the trap. When conducting a transfer to the N5 position ahigh energy compound formed.
Due to the relationship between HCN and the formationof adenine partially through addition to 4-amino-5-cyanoimidazole (4-aminoimidazole-5-carbonitrile)17,an attempt was made to lower the aforementioned energy by shuttling proton 12to N5 in order to create a potentially stable NH2 moiety. Thisapproach lent a reasonable transition state energy of 82 kcal·mol-1however, the newly formed NH2 containing species optimized to anenergy of 123.15 kcal·mol-1. In order to render these tribulations,the excited state intermolecular proton transfer from C2 to N9 was analyzed.Despite a long range atomic scan, the transition state was shown to occur at 95.56kcal·mol-1 with an optimized geometric structure at 93.
63 kcal·mol-1.Although this is a high energy transition state and geometry, it has not beenunusual in the explorations of CID fragmentation involving adenine.14The PES effectively describes the loss of CH3CN however, it ishypothesized that a lower energy pathway exists explaining the relative intensitiesof the m/z 107 and m/z 133 peak observed in the experimental result (Figure 6). The lower energy pathway may include theformation of the CH3NC isomer.
Bond scans were employed again inorder to study the possibility of this occurrence. The result of the bond scan between C3 and N5yielded no discernable transition state, however, it is highly probably thatthe separation occurs and possesses a low energy transition state. 3-Methyladenine The experimental CID evolution spectrum with respect to increasing NCE value obtainedfrom 1-methyladenine provided curious results (Figure 8). Unlike the relativelydynamic fragmentation patterns seen in 1- and 7- methyladenine this isomer displayedthe sole characteristic loss of methyl radical. It is evident that this arises due to the position of the methyl subtituentin the N3 position. DFT computations for 1- and 7-methyladenine show that excitedstate intermolecular proton transfer to form tetrahedral intermediates givesrise to the loss of alternative fragments.
With the electron donating propertiesof the CH3 subtituent it is likely that the aromaticity and electrondensity provides a stabilizing effect in which only the methyl moiety can dissociatebefore complete degradation of the molecule. Much like the experimental spectrumfor 1-methyladenine a small shoulder can be seen at the m/z 133 peak correspondingto an NCE % of 30. This shoulder is resultant of the formation of a radical3-methyladenine species during the ESI process. Figure 8. Thespectrum above displays the fragmentation pattern of 3-methyladeninewith respect to increasing normalized collision energy (NCE%). Loss of methylradical fragment is emphasized.
Figure 7: The potential energy surface for the loss of CH3CN from 1-methyladenine. 7-Methyladenine 7-Methyladenine displays an experimentalspectrum (Figure 9) disimilar from both 1- and 3-methyladenine. A low relativeintensity peak is observed at m/z 121 representing a loss of27 AMUs. This loss has been been observed in several anions previously studiedincluding adenine, guanine, oxazole, as well a many others13,14 and in nearly every case this mass hascorreponded to an HCN molecule or one of it isomers. The mechanism by which this severance occurs islikely very similar to that of CH3CN from 1-methyladenine. It ishighly likely that a ring opening event occurs beteen N7 and C3 thus exposingthe HCN fragment as H5 N2 C7.
In order for the ring opening event to take placean intermolecular proton transfer of H11is probable. The energy of this dissociation is thought to be relatively highdue to the vast difference in intensity compared to the loss of methyl radical.An alternative route to describe the fragment is the formation of the HNC isomer.This event could likely involve an intermolecular proton tranfer from C2 or N9to N7 in order to destabilize the N7-C3 bond. The destabilization of the bond willlikely result in a ring opening followed by dissociation of HNC as H11N9C3 orH5N7C2.
It Is once again significant to describe the formation of the7-methyladenine radical during the ESI process observed as a shoulder in theexperimental spectrum. Attempts were once made to solve this problem byreducing the isolation width of the spectrum to .4. Despite the small isolationwidth the shoulder persisted to appear on the experimental spectrum.