Research Article Received: 23 December 2013
Revised: 5 March 2014
Accepted: 6 March 2014
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214 (wileyonlinelibrary.com) DOI: 10.1002/rcm.6892
Electrospray ionization tandem mass spectrometry study of six isomeric cationic amphiphiles with ester/amide linker D. Vijay Darshan1, B. G. N. Chandar1, M. Srujan2†, A. Chaudhuri2 and S. Prabhakar1* 1 2
National Centre for Mass Spectrometry, CSIR-Indian Institute of Chemical Technology, Hyderabad 500007, India Biomaterials Group, CSIR – Indian Institute of Chemical Technology, Hyderabad 500007, India
RATIONALE: Isomeric cationic amphiphiles differing only in the orientation of the linker group have been demonstrated
to possess dramatically changed gene transfer efficacies. Studies aimed at understanding structure-stability correlations of such isomeric cationic amphiphiles at the molecular level are yet to be undertaken. Such studies may throw significant new insights into the mechanistic origin on their contrasting bioactivities. METHODS: Electrospray ionization mass spectrometry (ESI-MS) and multi-stage tandem mass spectrometric (MSn) experiments were performed on a LCQ ion trap mass spectrometer. The decomposition pathway was confirmed by high-resolution mass spectrometry data from a quadrupole time-of-flight (Q-TOF) mass spectrometer. Dissociation curves were drawn based on the intensities of precursor and product ions. RESULTS: The collision-induced dissociation (CID) spectra of the M+ ion of each isomeric pair showed distinct product ions (3 pairs). Normal esters (1 and 3) showed abundant product ions with a neighboring group participation (NGP) reaction and reverse esters (lipid 2 and 4) showed McLafferty rearrangement product ions. The spectra of a normal amide (5) and a reverse amide (6) are similar to that found in the corresponding ester, except for the absence of the McLafferty rearrangement in 6. Dissociation curves revealed that normal esters/amide decompose at lower energy than those of corresponding reverse esters/amide. CONCLUSIONS: The lipids which easily decompose (flexible) show dramatically enhanced gene delivery capabilities and the lipids which decompose at higher collision energy (CE) values (rigid) are transfection incompetent. Copyright © 2014 John Wiley & Sons, Ltd.
The clinical success of gene therapy critically depends on the availability of safe and efficacious gene delivery reagents (popularly known as transfection vectors that are broadly categorized as viral and non-viral transfection vectors).[1,2] As the use of viral vectors is associated with bio-safety issues, researchers are developing a number of non-viral alternatives including cationic amphiphiles (also known as cationic transfection lipids),[3] cationic polymers[4,5] and dendrimers,[6–8] etc. Cationic transfection lipids are finding widespread exploitations as efficient liposomal gene delivery reagents in non-viral gene therapy. The gene transfer efficiencies of cationic amphiphiles critically depend on the nature of the head-group, alkyl chain-length of the tail group and the functionalities used in covalently linking the polar head-group and non-polar tail-groups. The effects of head and tail functions in gene delivery have been investigated previously. Recently, Chauduri and co-workers[9] reported on the dramatic influence of the linker group in liposomal gene delivery.
* Correspondence to: S. Prabhakar, National Centre for Mass Spectrometry, Indian Institute of Chemical Technology, Hyderabad 500007, India. E-mail: [email protected] †
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
Copyright © 2014 John Wiley & Sons, Ltd.
1209
Present address: Institute for Stem Cell Biology and Regenerative Medicine, Bangalore 560065, India.
Chauduri and co-workers[9] designed and synthesized a set of structurally isomeric cationic amphiphiles bearing the same hydrophobic tails and polar head groups. The only structural difference between each isomeric pair of cationic amphiphiles is the orientation of their linker, ester/amide functionality (Scheme 1). In one isomer (termed here as a ’normal’ ester/amide), the ester oxygen or amide NH is closer to the positively charged nitrogen atom (1, 3, and 5; Scheme 1), and in the second isomer (termed here as a ’reverse’ ester/amide), the ester or the amide carbonyl group is closer to the quaternary nitrogen atom (2, 4, and 6; Scheme 1). Despite the remarkable structural similarities between 1 and 2, only lipid 1 could efficiently deliver plasmid DNA encoding β-galactosidase enzyme into a number of cultured mammalian cells, whereas lipid 2 was found to be essentially incompetent.[10] Similarly, lipid 5 with an amide linker was found to be a highly serum-compatible cationic amphiphile and showed remarkable selectivity in transfecting mouse lung; however, lipid 6 showed compromised serum compatibility and gene transfer properties. It was found that the cationic liposomes of lipids 1 and 3 showed higher biomembrane fusibility while the cationic liposomes of lipids 2 and 4 showed rigid nature.[10] Studies aimed at understanding structure-stability correlations of such isomeric cationic amphiphiles at the molecular level are yet to be explored. Such studies may throw significant new insights into the mechanistic origin on their
D. Vijay Darshan et al.
Scheme 1. Chemical structures of the studied lipids (1–6).
contrasting bioactivities. Mass spectrometry has been used previously for the identification of lipids for their structural elucidation[11,12] and characterization.[13,14] However, mass spectral studies aimed at understanding structural stabilities vs gene transfer efficiency profiles of isomeric cationic amphiphiles have not yet been undertaken. To this end, we attempted to study the effect of linker group and its orientation (normal vs reverse ester/amide) on the decomposition of three pairs of isomeric cationic amphiphiles in the gas phase using electrospray ionization (ESI) mass spectrometry (MS). To the best of our knowledge, this is the first mass spectral study aimed at understanding correlation between the stability of molecules with their biological activity.
EXPERIMENTAL Syntheses of all the studied compounds 1–6 were as per the literature reports.[9,10] The isomers were isolated by column chromatography and characterized by nuclear magnetic resonance (NMR) and MS. All the solvents used for MS analysis were of HPLC-grade purchased from Merck (Mumbai, India). Stock solutions (1000 ppm) of all the compounds were made in methanol and stored in a
refrigerator. The working solutions (50 ppm) were made by diluting the stock solutions with appropriate volumes of methanol before subjecting them to MS analysis. The lipids solutions were infused into the ESI source at a flow rate of 5 μL/min using a built-in syringe pump. The ESI-MS and multi-stage mass spectrometric (MSn) experiments were performed on a LCQ ion trap mass spectrometer (ThermoFinnigan, San Jose, CA, USA). The typical source conditions were: spray voltage, 4 kV; capillary voltage, 15 to 20 V; heated capillary temperature, 200 °C; tube lens offset voltage, 20 V. Nitrogen was used as the sheath gas and helium was used as the damping gas. In the MSn experiments, the precursor ion of interest was first isolated by applying an appropriate waveform across the end cap electrodes of the ion trap to resonantly eject all the trapped ions, expect those ions of the m/z ratio of interest. The isolated ions were then subjected to a supplementary ac signal to resonantly excite them and so cause collision-induced dissociation (CID). The collision energy used was 28 to 60 eV. The excitation time used was 30 ms. The high-resolution mass spectrometry (HRMS) data were recorded using a quadrupole time-of-flight (Q-TOF) mass spectrometer (QSTAR XL, Applied Biosystems/MDS Sciex, Foster City, CA, USA) equipped with an ESI source. The typical source conditions were: capillary voltage, 5 kV; declustering potential, 60 V; focusing potential, 250 V; 2nd declustering potential, 10 V; mass resolution 10000 FWHM (full-width at half-maximum). Nitrogen was used as the curtain gas and the collision gas. For the CID experiments, the precursor ion was selected by using the quadrupole analyzer and the product ion were analyzed using the TOF analyzer. The collision energies were between 5 to 70 eV, unless otherwise stated. To record the CID spectra of fragment ions, the fragment ions of interest were obtained in the source by applying higher DP values (in-source fragmentation). All the spectra of isomers were recorded under identical experimental conditions. The lipids were dissolved in HPLC-grade methanol, and infused in to the ESI source at a flow rate of 5 μL/min using a built-in syringe pump.
Table 1. MSn data of compounds 1–6
Lipid No. 1 2
3 4
5 6
1210
Ion (m/z)
MSn (n=)
CE (eV)
Product ion, m/z (Relative abundance, %)
M+;584 330 M+;584 402 298 220 M+;582 328 M+;582 400 298 218 M+; 582 329 M+; 582 329
2 3 2 3 4 4 2 3 2 3 4 4 2 3 2 3
32 35 48 39 39 39 32 37 48 39 39 39 32 35 45 35
330 (5), 255 (100) 255 (100), 120 (4),102 (86) 540 (4.5), 402 (62.4), 298 (37.6), 220 (22.4), 116 (17.6) 358 (3), 298 (12), 220 (100) 116 (100) 116 (100) 328 (6),255 (100) 255 (100),118 (5),100 (48) 540 (0.8), 400 (75.4), 298 (45.6), 218 (42.1), 116 (33.3) 358 (8), 298 (8),218 (83),116 (8) 116 (100) 116 (100) 329 (100), 254 (64.42) 254(100) 538 (0.5), 329 (34.7), 297 (6.4%), 254 (100) 254 (12)
wileyonlinelibrary.com/journal/rcm
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
ESI-MS/MS study of cationic amphiphiles with ester/amide linker
RESULTS AND DISCUSSION The chemical structures of the isomeric lipids (1–6) are given in Scheme 1. The studied lipids include quaternary normal esters (1 and 3), quaternary reverse esters (2 and 4), quaternary amide (5) and quaternary reverse amide (6). All the lipids were preformed ions, and hence they can be easily analyzed under positive ESI conditions. They exist in the salt form, M+Cl–, where M+ is the quaternary ammonium ion and Cl– is the counter ion. As expected the ESI analysis of these compounds showed the quaternary
ammonium ion (M+). We moved to CID experiments on M+ ions to understand the effect of linker orientation in dissociation of M+ ions. CID of M+ ions The CID spectra of the M+ ion of compounds 1–6 were recorded at different collision energies (between 5 and 70 eV). The spectra of each isomeric pair showed distinct fragment ions. The overall fragmentation pattern of each compound remained the same at different collision energies;
Table 2. High-resolution mass spectral data for compounds 1–6
Comp. No. 1 2
3 4
5 6
Ion (m/z) 584, M+ 255 330 584,M+ 402 298 220 116 582,M+ 255 582,M+ 400 298 218 116 582,M+ 329 254 582,M+ 329 297 254
Elemental composition C35H70NO5 C16H31O2 C19H40NO3 C35H70NO5 C22H44NO5 C18H36NO2 C9H18NO5 C5H10NO2 C36H72NO4 C16H31O2 C36H72NO4 C23H46NO4 C18H36NO2 C10H20NO4 C5H10NO2 C35H72N3O3 C19H41N2O2 C16H32NO C35H72N3O3 C19H41N2O2 C18H37N2O C16H32NO
Calculated mass (Da)
Measured mass (Da)
Error (ppm)
584.5254 255.2324 330.3008 584.5254 402.3219 298.2746 220.1185 116.0712 582.5461 255.2324 582.5461 400. 3427 298.2746 218.1392 116.0712 582.5574 329.3168 254.2484 582.5574 329.3168 297.2906 254.2484
584.5260 255.2340 330.3017 584.5255 402.3221 298.2744 220.1207 116.0715 582.5456 255.2347 582.5465 400.3425 298.2755 218.1401 116.0715 582.5578 329.3171 254.2488 582.5578 329.3164 297.2908 254.2486
+1.03 +1.15 –2.72 +0.17 +0.38 –0.69 +10.00 +2.99 –0.92 –1.54 +0.63 –0.46 +3.00 +3.97 +2.98 +0.74 +0.90 +1.61 +0.74 –1.23 +0.71 +0.83
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
1211
Scheme 2. Fragmentation pattern of the M+ ion from normal esters (1 and 3) and normal amide (5).
D. Vijay Darshan et al. however, there were variations in the relative abundances of precursor and product ions. The fragmentation patterns of M+ ions from normal and reverse ester/amide, arrived at based on MSn (Table 1) and HRMS data (Table 2), are summarized in Schemes 2 and 3, respectively. Lipids 1 and 2 The CID spectrum of the M+ ion (m/z 584) from lipid 1 is completely different from that of lipid 2 (Fig. 1). Lipid 1 displays a prominent product ion at m/z 255 in addition to a low-abundance product ion at m/z 330. The ion m/z 255 is formed by the loss of C19H39NO3 as a neutral amine from M+ by neighboring group participation (NGP),[15,16] as described in Scheme 2 (path A). A similar kind of NGP-assisted fragmentation was reported in acetylcholine.[17] The NGP product ion (m/z 255) attains stability with a 1,3-dioxonium cation structure. The low-abundance product ion m/z 330 is formed by the loss of C16H30O2 and the plausible mechanism is shown in Scheme 2 (path B). Further fragmentation of the ion m/z 330 resulted in the ion at m/z 255 in addition to other product ions at m/z 120 and 102. The CID spectrum of M+ from lipid 2 showed a different set of product ions entirely, i.e., ions at m/z 540, 402, 298 and 220 (Fig. 1). The possible fragmentation pathway for formation of these products is summarized in Scheme 3. The ion at m/z 540 is formed by the loss of 44 u (C2H4O) and the more abundant product ion m/z 402 is the McLafferty rearrangement product ion.[18,19] Further fragmentation of m/z 402 showed the loss of 44 u (C2H4O) at m/z 358[20] and a second McLafferty rearrangement product ion at m/z 220. The ion m/z 298 further undergoes a McLafferty rearrangement, resulting in the ion at m/z 116. The ion m/z 116 can also be formed from m/z 402 or
Figure 1. CID spectra of M+ (m/z 584) from (A) 1 and (B) 2 at CE 41 eV. 220 from a different pathway. The NGP-type reaction that was prominent in lipid 1 is absent in lipid 2, probably because such a reaction leads to the formation of an unstable four-membered cyclic oxonium ion. Lipids 3 and 4 The lipids 3 and 4 are structurally similar to 1 and 2, respectively, except the hydroxyl ethyl group in 1 and 2 is replaced with a propyl group in 3 and 4. The fragmentation
1212
Scheme 3. Fragmentation pattern of M+ ion from reverse esters (2 and 4) and reverse amide (6).
wileyonlinelibrary.com/journal/rcm
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
ESI-MS/MS study of cationic amphiphiles with ester/amide linker
Figure 2. CID curves for the M+ ion from lipids 1–6.
pattern of 3 and 4 is almost similar to that observed in 1 and 2, respectively. When compared to 1 and 2, the product ions are shifted two mass units less in 3 and 4 when the propyl group is retained in the product ion, and some product ions remain unchanged when the propyl group is carried in the neutral part. The behavior of 3 and 4 reveals that there is no significant role of the hydroxyl group in the fragmentation of these compounds. Lipids 5 and 6
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
Apart from showing distinct CID spectra for isomeric lipids, it is apparent in the spectra that the precursor ion abundance of normal esters/amide (1, 3 and 5) is less than the corresponding reverse esters/amide (2, 4 and 6), which implies that M+ ions of reverse esters/amide are more stable than those of normal esters/amide. The same is clearly reflected when the percentage of the total ion current (% of TIC) of the M+ ion is plotted against collision energy (CE). The % of TIC is obtained by a/a + b, where a = relative abundance of M+ and b = sum of the relative abundances of all the product ions (Fig. 2). The CE values required for 50% dissociation of the M+ ion are marked with dotted lines in the plots. The plots reveal that the lipids 1, 3 and 5 underwent fragmentation at a lower CE value than lipids 2, 4 and 6, respectively. Lipids 5 and 6 decompose at a lower CE than 1–4. It is well known in CID experiments that the molecule with the higher collision cross-sectional area fragments easier (i.e., at lower CE) than the one with lower collision cross-sectional area. Based on this analogy, the dissociation of M+ ions of 1, 3, and 5 at lower CE values could be explained by presuming higher collision cross-sectional areas for 1, 3 and 5 than their corresponding isomers (2, 4 and 6, respectively). It also suggests that lipids 1, 3 and 5 must have flexible conformational structures by which their collision cross-sectional area increases. Interestingly, the lipids which decompose easily (flexible) show dramatically enhanced gene delivery capabilities and the lipids which decompose at higher CE values (rigid) are transfection incompetent. The flexible structures of lipids 1, 3 and 5 may be responsible for the higher biomembrane fusibility of their cationic liposomes when compared to those of lipids 2 and 4 and 6 that showed rigid nature.[10]
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
1213
The molecular weight (M+ = 582) of the lipids 5 and 6 is the same as that of 3 and 4, and they are structurally similar to 1 and 2, except that there is an amide functionality in 5 and 6 in the place of an ester functionality in 3 and 4. The fragmentation of lipid 5 is similar to that of lipid 1 but there is quite a difference in the abundance of the product ions (Table 1). [M–C16H31NO]+ (m/z 329) is the dominant ion in lipid 5, whereas the similar ion (m/z 330) is low in abundance in 1. The NGP product ion m/z 255 was the dominant ion in 1, while the analogous ion at m/z 254 is in low abundance in 5. The CID spectrum of m/z 582 from lipid 6 showed product ions at m/z 538, 329, 297 and 254. Like in lipids 2 and 4, lipid 6 shows loss of C2H4O (44 u) at m/z 538 and 297 (Scheme 3). The expected McLafferty rearrangement product ion (m/z 400) is absent in lipid 6. In the absence of the McLafferty rearrangement process, lipid 6 showed some of the product ions found in lipid 5, i.e., m/z 254 and 329, but their relative abundances were different between the two isomers. The ion at m/z 254 in lipid 6 may be formed by a different pathway when compared to lipid 5.
Dissociation curves
D. Vijay Darshan et al.
CONCLUSIONS Cationic amphiphiles have been exploited as efficient liposomal gene delivery reagents in non-viral gene therapy and each subunit of cationic amphiphiles, i.e., head, tail and linker, plays a critical role in their gene transfer efficiencies. For instance, isomeric cationic amphiphiles with a varied linker group (normal vs reverse esters/amide), in which normal esters/amide showed remarkable gene transfer, whereas its isomeric reverse esters/amide have been reported to be essentially transfection incompetent. To this end, we attempted to study the effect of linker group and its orientation (normal vs reverse ester/amide) on the decomposition of three pairs of isomeric cationic amphiphiles in the gas phase using ESI-MS. The CID spectra of normal esters/amide are very different from the corresponding reverse esters/amide. The major product ions were formed by a neighboring group participation process or a McLafferty rearrangement process. Dissociation curves were drawn based on the relative abundances of precursor ion and product ions in the CID spectra at each CE value. The dissociation curves clearly revealed that the normal esters/amide dissociate at lower CE values than compared to those for the corresponding reverse esters/amide. Such findings are consistent with the supposition that lipids possessing better gene delivery capabilities are more easily decomposed than transfection-incompetent lipids. Stated differently, the normal esters/amides (1, 3, and 5) possess flexible conformational structures by which their collision cross-sectional area increases resulting in their decomposition at lower CE value, whereas the reverse esters/amide are likely to be structurally rigid with less flexibility in their structures causing lower collision cross-sectional area. Such flexible conformational structures possibly impart multiple hydrogen bonding with the bio-environment leading to enhanced bioactivities.
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
Acknowledgements The authors thank the Director, CSIR-IICT, Hyderabad, for facilities and encouragement. Senior research fellowships for VD and MS and funding (CSC-0406) from CSIR, New Delhi are greatly acknowledged.
[15]
[16] [17]
REFERENCES [1] I. M. Verma, M. Somina. Gene therapy – promises and prospects. Nature 1997, 389, 239. [2] W. F. Anderson. Human gene therapy. Nature 1998, 392, 25. [3] P. P. Karmani, A. Chaudhuri. Cationic liposomes as nonviral carriers of gene medicines: resolved issues, open questions, and future promises. Med. Res. Rev. 2007, 27, 696. [4] S. Mishra, J. D. Heidel, P. Webster, M. E. Davis. Imidazole end groups on a linear, cyclodextrin-containing polycation produce enhanced gene delivery via multiple processes. J. Control. Rel. 2006, 116, 179. [5] D. M. Lynn, D. G. Anderson, D. Putman, R. Langer. Accelerated discovery of synthetic transfection vectors:
[18]
[19]
[20]
parallel synthesis and screening of a degradable polymer library. J. Am. Chem. Soc. 2001, 123, 8155. M. X. Tang, C. T. Redmann, F. C. Szoka. In vitro gene delivery by degraded polyamidoamine dendrimers. Bioconjugate Chem. 1996, 7, 703. J. F. Kukowska-Latallo, A. U. Bienlinska, J. Jonhson, R. Spinder, D. A. Tomalia, J. R. Baker. Efficient transfer of genetic material into mammalian cells using Starburst polyamidoamine dendrimers. Proc. Natl. Acad. Sci. USA 1996, 93, 4897. D. Luo, K. Haverstick, N. Belecheva, E. Han, W. M. Saltzman. Poly(ethylene glycol)-conjugated PAMAM dendrimer for biocompatible, high-efficiency DNA delivery. Macromolecules 2002, 35, 3456. M. Rajesh, J. Sen, M. Srujan, K. Mukherjee, B. Sreedhar, A. Chaudhuri. Dramatic influence of the orientation of linker between hydrophilic and hydrophobic lipid moiety in liposomal gene delivery. J. Am. Chem. Soc. 2007, 129, 11408. M. Srjun, V. Chandrashekhar, R. C. Reddy, R. Prabhakar, B. Sreedhar, A. Chaudhuri. The influence of the structural orientation of amide linkers on the serum compatibility and lung transfection properties of cationic amphiphiles. Biomaterials 2011, 32, 5231. J. W. Russell, C. M. Robert. Mass spectrometric analysis of four regioisomers of F2-isoprostanes formed by free radical oxidation of arachidonic acid. J. Am. Soc. Mass Spectrom. 1996, 7, 490. P. Wheelan, J. A. Zirrolli, R. C. Murphy. Electrospray ionization and low energy tandem mass spectrometry of polyhydroxy unsaturated fatty acids. J. Am. Soc. Mass Spectrom. 1996, 7, 140. F.-F. Hsu, J. Turk. Differentiation of 1-O-alk-1′-enyl-2-acyl and 1-O-alkyl-2-acyl glycerolphospholipids by multiplestage linear ion-trap mass spectrometry with electrospray Ionization. J. Am. Soc. Mass Spectrom. 2007, 18, 2065. F.-F. Hsu, J. Turk, Y. Shi, E. A. Grosiman. Characterization of acylphosphatidylglycerols from Salmonella typhimurium by tandem mass spectrometry with electrospray ionization. J. Am. Soc. Mass Spectrom. 2004, 15, 1. S. H. Wilen, L. Delguzzo, R. Safersstein. Experimental evidence for aco-7 neighboring group participation. Tetrahedron 1987, 43, 5089. Y. He, J. P. Reilly. Does a charge tag really provide a fixed charge? Angew. Chem. 2008, 120, 2497. H. Lioe, C. K. Barlow, R. A. J. O′Hair. How does acetylcholine lose trimethylamine? A density functional theory study of four competing mechanisms. J. Am. Soc. Mass Spectrom. 2009, 20, 238. K. Bulleigh, A. Howard, T. Do, Q. Wu, V. Anbalagan, M. Van Stipdonk. Investigation of intramolecular proton migration in a series of model, metal-cationized tripeptides using in situ generation of an isotope label. Rapid Commun. Mass Spectrom. 2006, 20, 227. G. Wood, A. M. Falick, A. L. Burlingame. Apparent charge localization effects in the McLafferty rearrangement of acetate ester; field ionization kinetic measurements. Org. Mass Spectrom. 1974, 8, 279. T. J. Reddy, S. P. Mirza, U. V. R. V. Saradhi, V. J. Rao, M. Vairamani. Mass spectral studies of N,N-dialkylaminoethanols. Rapid Commun. Mass Spectrom. 2003, 17, 746.
1214 wileyonlinelibrary.com/journal/rcm
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
Revised: 5 March 2014
Accepted: 6 March 2014
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214 (wileyonlinelibrary.com) DOI: 10.1002/rcm.6892
Electrospray ionization tandem mass spectrometry study of six isomeric cationic amphiphiles with ester/amide linker D. Vijay Darshan1, B. G. N. Chandar1, M. Srujan2†, A. Chaudhuri2 and S. Prabhakar1* 1 2
National Centre for Mass Spectrometry, CSIR-Indian Institute of Chemical Technology, Hyderabad 500007, India Biomaterials Group, CSIR – Indian Institute of Chemical Technology, Hyderabad 500007, India
RATIONALE: Isomeric cationic amphiphiles differing only in the orientation of the linker group have been demonstrated
to possess dramatically changed gene transfer efficacies. Studies aimed at understanding structure-stability correlations of such isomeric cationic amphiphiles at the molecular level are yet to be undertaken. Such studies may throw significant new insights into the mechanistic origin on their contrasting bioactivities. METHODS: Electrospray ionization mass spectrometry (ESI-MS) and multi-stage tandem mass spectrometric (MSn) experiments were performed on a LCQ ion trap mass spectrometer. The decomposition pathway was confirmed by high-resolution mass spectrometry data from a quadrupole time-of-flight (Q-TOF) mass spectrometer. Dissociation curves were drawn based on the intensities of precursor and product ions. RESULTS: The collision-induced dissociation (CID) spectra of the M+ ion of each isomeric pair showed distinct product ions (3 pairs). Normal esters (1 and 3) showed abundant product ions with a neighboring group participation (NGP) reaction and reverse esters (lipid 2 and 4) showed McLafferty rearrangement product ions. The spectra of a normal amide (5) and a reverse amide (6) are similar to that found in the corresponding ester, except for the absence of the McLafferty rearrangement in 6. Dissociation curves revealed that normal esters/amide decompose at lower energy than those of corresponding reverse esters/amide. CONCLUSIONS: The lipids which easily decompose (flexible) show dramatically enhanced gene delivery capabilities and the lipids which decompose at higher collision energy (CE) values (rigid) are transfection incompetent. Copyright © 2014 John Wiley & Sons, Ltd.
The clinical success of gene therapy critically depends on the availability of safe and efficacious gene delivery reagents (popularly known as transfection vectors that are broadly categorized as viral and non-viral transfection vectors).[1,2] As the use of viral vectors is associated with bio-safety issues, researchers are developing a number of non-viral alternatives including cationic amphiphiles (also known as cationic transfection lipids),[3] cationic polymers[4,5] and dendrimers,[6–8] etc. Cationic transfection lipids are finding widespread exploitations as efficient liposomal gene delivery reagents in non-viral gene therapy. The gene transfer efficiencies of cationic amphiphiles critically depend on the nature of the head-group, alkyl chain-length of the tail group and the functionalities used in covalently linking the polar head-group and non-polar tail-groups. The effects of head and tail functions in gene delivery have been investigated previously. Recently, Chauduri and co-workers[9] reported on the dramatic influence of the linker group in liposomal gene delivery.
* Correspondence to: S. Prabhakar, National Centre for Mass Spectrometry, Indian Institute of Chemical Technology, Hyderabad 500007, India. E-mail: [email protected] †
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
Copyright © 2014 John Wiley & Sons, Ltd.
1209
Present address: Institute for Stem Cell Biology and Regenerative Medicine, Bangalore 560065, India.
Chauduri and co-workers[9] designed and synthesized a set of structurally isomeric cationic amphiphiles bearing the same hydrophobic tails and polar head groups. The only structural difference between each isomeric pair of cationic amphiphiles is the orientation of their linker, ester/amide functionality (Scheme 1). In one isomer (termed here as a ’normal’ ester/amide), the ester oxygen or amide NH is closer to the positively charged nitrogen atom (1, 3, and 5; Scheme 1), and in the second isomer (termed here as a ’reverse’ ester/amide), the ester or the amide carbonyl group is closer to the quaternary nitrogen atom (2, 4, and 6; Scheme 1). Despite the remarkable structural similarities between 1 and 2, only lipid 1 could efficiently deliver plasmid DNA encoding β-galactosidase enzyme into a number of cultured mammalian cells, whereas lipid 2 was found to be essentially incompetent.[10] Similarly, lipid 5 with an amide linker was found to be a highly serum-compatible cationic amphiphile and showed remarkable selectivity in transfecting mouse lung; however, lipid 6 showed compromised serum compatibility and gene transfer properties. It was found that the cationic liposomes of lipids 1 and 3 showed higher biomembrane fusibility while the cationic liposomes of lipids 2 and 4 showed rigid nature.[10] Studies aimed at understanding structure-stability correlations of such isomeric cationic amphiphiles at the molecular level are yet to be explored. Such studies may throw significant new insights into the mechanistic origin on their
D. Vijay Darshan et al.
Scheme 1. Chemical structures of the studied lipids (1–6).
contrasting bioactivities. Mass spectrometry has been used previously for the identification of lipids for their structural elucidation[11,12] and characterization.[13,14] However, mass spectral studies aimed at understanding structural stabilities vs gene transfer efficiency profiles of isomeric cationic amphiphiles have not yet been undertaken. To this end, we attempted to study the effect of linker group and its orientation (normal vs reverse ester/amide) on the decomposition of three pairs of isomeric cationic amphiphiles in the gas phase using electrospray ionization (ESI) mass spectrometry (MS). To the best of our knowledge, this is the first mass spectral study aimed at understanding correlation between the stability of molecules with their biological activity.
EXPERIMENTAL Syntheses of all the studied compounds 1–6 were as per the literature reports.[9,10] The isomers were isolated by column chromatography and characterized by nuclear magnetic resonance (NMR) and MS. All the solvents used for MS analysis were of HPLC-grade purchased from Merck (Mumbai, India). Stock solutions (1000 ppm) of all the compounds were made in methanol and stored in a
refrigerator. The working solutions (50 ppm) were made by diluting the stock solutions with appropriate volumes of methanol before subjecting them to MS analysis. The lipids solutions were infused into the ESI source at a flow rate of 5 μL/min using a built-in syringe pump. The ESI-MS and multi-stage mass spectrometric (MSn) experiments were performed on a LCQ ion trap mass spectrometer (ThermoFinnigan, San Jose, CA, USA). The typical source conditions were: spray voltage, 4 kV; capillary voltage, 15 to 20 V; heated capillary temperature, 200 °C; tube lens offset voltage, 20 V. Nitrogen was used as the sheath gas and helium was used as the damping gas. In the MSn experiments, the precursor ion of interest was first isolated by applying an appropriate waveform across the end cap electrodes of the ion trap to resonantly eject all the trapped ions, expect those ions of the m/z ratio of interest. The isolated ions were then subjected to a supplementary ac signal to resonantly excite them and so cause collision-induced dissociation (CID). The collision energy used was 28 to 60 eV. The excitation time used was 30 ms. The high-resolution mass spectrometry (HRMS) data were recorded using a quadrupole time-of-flight (Q-TOF) mass spectrometer (QSTAR XL, Applied Biosystems/MDS Sciex, Foster City, CA, USA) equipped with an ESI source. The typical source conditions were: capillary voltage, 5 kV; declustering potential, 60 V; focusing potential, 250 V; 2nd declustering potential, 10 V; mass resolution 10000 FWHM (full-width at half-maximum). Nitrogen was used as the curtain gas and the collision gas. For the CID experiments, the precursor ion was selected by using the quadrupole analyzer and the product ion were analyzed using the TOF analyzer. The collision energies were between 5 to 70 eV, unless otherwise stated. To record the CID spectra of fragment ions, the fragment ions of interest were obtained in the source by applying higher DP values (in-source fragmentation). All the spectra of isomers were recorded under identical experimental conditions. The lipids were dissolved in HPLC-grade methanol, and infused in to the ESI source at a flow rate of 5 μL/min using a built-in syringe pump.
Table 1. MSn data of compounds 1–6
Lipid No. 1 2
3 4
5 6
1210
Ion (m/z)
MSn (n=)
CE (eV)
Product ion, m/z (Relative abundance, %)
M+;584 330 M+;584 402 298 220 M+;582 328 M+;582 400 298 218 M+; 582 329 M+; 582 329
2 3 2 3 4 4 2 3 2 3 4 4 2 3 2 3
32 35 48 39 39 39 32 37 48 39 39 39 32 35 45 35
330 (5), 255 (100) 255 (100), 120 (4),102 (86) 540 (4.5), 402 (62.4), 298 (37.6), 220 (22.4), 116 (17.6) 358 (3), 298 (12), 220 (100) 116 (100) 116 (100) 328 (6),255 (100) 255 (100),118 (5),100 (48) 540 (0.8), 400 (75.4), 298 (45.6), 218 (42.1), 116 (33.3) 358 (8), 298 (8),218 (83),116 (8) 116 (100) 116 (100) 329 (100), 254 (64.42) 254(100) 538 (0.5), 329 (34.7), 297 (6.4%), 254 (100) 254 (12)
wileyonlinelibrary.com/journal/rcm
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
ESI-MS/MS study of cationic amphiphiles with ester/amide linker
RESULTS AND DISCUSSION The chemical structures of the isomeric lipids (1–6) are given in Scheme 1. The studied lipids include quaternary normal esters (1 and 3), quaternary reverse esters (2 and 4), quaternary amide (5) and quaternary reverse amide (6). All the lipids were preformed ions, and hence they can be easily analyzed under positive ESI conditions. They exist in the salt form, M+Cl–, where M+ is the quaternary ammonium ion and Cl– is the counter ion. As expected the ESI analysis of these compounds showed the quaternary
ammonium ion (M+). We moved to CID experiments on M+ ions to understand the effect of linker orientation in dissociation of M+ ions. CID of M+ ions The CID spectra of the M+ ion of compounds 1–6 were recorded at different collision energies (between 5 and 70 eV). The spectra of each isomeric pair showed distinct fragment ions. The overall fragmentation pattern of each compound remained the same at different collision energies;
Table 2. High-resolution mass spectral data for compounds 1–6
Comp. No. 1 2
3 4
5 6
Ion (m/z) 584, M+ 255 330 584,M+ 402 298 220 116 582,M+ 255 582,M+ 400 298 218 116 582,M+ 329 254 582,M+ 329 297 254
Elemental composition C35H70NO5 C16H31O2 C19H40NO3 C35H70NO5 C22H44NO5 C18H36NO2 C9H18NO5 C5H10NO2 C36H72NO4 C16H31O2 C36H72NO4 C23H46NO4 C18H36NO2 C10H20NO4 C5H10NO2 C35H72N3O3 C19H41N2O2 C16H32NO C35H72N3O3 C19H41N2O2 C18H37N2O C16H32NO
Calculated mass (Da)
Measured mass (Da)
Error (ppm)
584.5254 255.2324 330.3008 584.5254 402.3219 298.2746 220.1185 116.0712 582.5461 255.2324 582.5461 400. 3427 298.2746 218.1392 116.0712 582.5574 329.3168 254.2484 582.5574 329.3168 297.2906 254.2484
584.5260 255.2340 330.3017 584.5255 402.3221 298.2744 220.1207 116.0715 582.5456 255.2347 582.5465 400.3425 298.2755 218.1401 116.0715 582.5578 329.3171 254.2488 582.5578 329.3164 297.2908 254.2486
+1.03 +1.15 –2.72 +0.17 +0.38 –0.69 +10.00 +2.99 –0.92 –1.54 +0.63 –0.46 +3.00 +3.97 +2.98 +0.74 +0.90 +1.61 +0.74 –1.23 +0.71 +0.83
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
1211
Scheme 2. Fragmentation pattern of the M+ ion from normal esters (1 and 3) and normal amide (5).
D. Vijay Darshan et al. however, there were variations in the relative abundances of precursor and product ions. The fragmentation patterns of M+ ions from normal and reverse ester/amide, arrived at based on MSn (Table 1) and HRMS data (Table 2), are summarized in Schemes 2 and 3, respectively. Lipids 1 and 2 The CID spectrum of the M+ ion (m/z 584) from lipid 1 is completely different from that of lipid 2 (Fig. 1). Lipid 1 displays a prominent product ion at m/z 255 in addition to a low-abundance product ion at m/z 330. The ion m/z 255 is formed by the loss of C19H39NO3 as a neutral amine from M+ by neighboring group participation (NGP),[15,16] as described in Scheme 2 (path A). A similar kind of NGP-assisted fragmentation was reported in acetylcholine.[17] The NGP product ion (m/z 255) attains stability with a 1,3-dioxonium cation structure. The low-abundance product ion m/z 330 is formed by the loss of C16H30O2 and the plausible mechanism is shown in Scheme 2 (path B). Further fragmentation of the ion m/z 330 resulted in the ion at m/z 255 in addition to other product ions at m/z 120 and 102. The CID spectrum of M+ from lipid 2 showed a different set of product ions entirely, i.e., ions at m/z 540, 402, 298 and 220 (Fig. 1). The possible fragmentation pathway for formation of these products is summarized in Scheme 3. The ion at m/z 540 is formed by the loss of 44 u (C2H4O) and the more abundant product ion m/z 402 is the McLafferty rearrangement product ion.[18,19] Further fragmentation of m/z 402 showed the loss of 44 u (C2H4O) at m/z 358[20] and a second McLafferty rearrangement product ion at m/z 220. The ion m/z 298 further undergoes a McLafferty rearrangement, resulting in the ion at m/z 116. The ion m/z 116 can also be formed from m/z 402 or
Figure 1. CID spectra of M+ (m/z 584) from (A) 1 and (B) 2 at CE 41 eV. 220 from a different pathway. The NGP-type reaction that was prominent in lipid 1 is absent in lipid 2, probably because such a reaction leads to the formation of an unstable four-membered cyclic oxonium ion. Lipids 3 and 4 The lipids 3 and 4 are structurally similar to 1 and 2, respectively, except the hydroxyl ethyl group in 1 and 2 is replaced with a propyl group in 3 and 4. The fragmentation
1212
Scheme 3. Fragmentation pattern of M+ ion from reverse esters (2 and 4) and reverse amide (6).
wileyonlinelibrary.com/journal/rcm
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
ESI-MS/MS study of cationic amphiphiles with ester/amide linker
Figure 2. CID curves for the M+ ion from lipids 1–6.
pattern of 3 and 4 is almost similar to that observed in 1 and 2, respectively. When compared to 1 and 2, the product ions are shifted two mass units less in 3 and 4 when the propyl group is retained in the product ion, and some product ions remain unchanged when the propyl group is carried in the neutral part. The behavior of 3 and 4 reveals that there is no significant role of the hydroxyl group in the fragmentation of these compounds. Lipids 5 and 6
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
Apart from showing distinct CID spectra for isomeric lipids, it is apparent in the spectra that the precursor ion abundance of normal esters/amide (1, 3 and 5) is less than the corresponding reverse esters/amide (2, 4 and 6), which implies that M+ ions of reverse esters/amide are more stable than those of normal esters/amide. The same is clearly reflected when the percentage of the total ion current (% of TIC) of the M+ ion is plotted against collision energy (CE). The % of TIC is obtained by a/a + b, where a = relative abundance of M+ and b = sum of the relative abundances of all the product ions (Fig. 2). The CE values required for 50% dissociation of the M+ ion are marked with dotted lines in the plots. The plots reveal that the lipids 1, 3 and 5 underwent fragmentation at a lower CE value than lipids 2, 4 and 6, respectively. Lipids 5 and 6 decompose at a lower CE than 1–4. It is well known in CID experiments that the molecule with the higher collision cross-sectional area fragments easier (i.e., at lower CE) than the one with lower collision cross-sectional area. Based on this analogy, the dissociation of M+ ions of 1, 3, and 5 at lower CE values could be explained by presuming higher collision cross-sectional areas for 1, 3 and 5 than their corresponding isomers (2, 4 and 6, respectively). It also suggests that lipids 1, 3 and 5 must have flexible conformational structures by which their collision cross-sectional area increases. Interestingly, the lipids which decompose easily (flexible) show dramatically enhanced gene delivery capabilities and the lipids which decompose at higher CE values (rigid) are transfection incompetent. The flexible structures of lipids 1, 3 and 5 may be responsible for the higher biomembrane fusibility of their cationic liposomes when compared to those of lipids 2 and 4 and 6 that showed rigid nature.[10]
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
1213
The molecular weight (M+ = 582) of the lipids 5 and 6 is the same as that of 3 and 4, and they are structurally similar to 1 and 2, except that there is an amide functionality in 5 and 6 in the place of an ester functionality in 3 and 4. The fragmentation of lipid 5 is similar to that of lipid 1 but there is quite a difference in the abundance of the product ions (Table 1). [M–C16H31NO]+ (m/z 329) is the dominant ion in lipid 5, whereas the similar ion (m/z 330) is low in abundance in 1. The NGP product ion m/z 255 was the dominant ion in 1, while the analogous ion at m/z 254 is in low abundance in 5. The CID spectrum of m/z 582 from lipid 6 showed product ions at m/z 538, 329, 297 and 254. Like in lipids 2 and 4, lipid 6 shows loss of C2H4O (44 u) at m/z 538 and 297 (Scheme 3). The expected McLafferty rearrangement product ion (m/z 400) is absent in lipid 6. In the absence of the McLafferty rearrangement process, lipid 6 showed some of the product ions found in lipid 5, i.e., m/z 254 and 329, but their relative abundances were different between the two isomers. The ion at m/z 254 in lipid 6 may be formed by a different pathway when compared to lipid 5.
Dissociation curves
D. Vijay Darshan et al.
CONCLUSIONS Cationic amphiphiles have been exploited as efficient liposomal gene delivery reagents in non-viral gene therapy and each subunit of cationic amphiphiles, i.e., head, tail and linker, plays a critical role in their gene transfer efficiencies. For instance, isomeric cationic amphiphiles with a varied linker group (normal vs reverse esters/amide), in which normal esters/amide showed remarkable gene transfer, whereas its isomeric reverse esters/amide have been reported to be essentially transfection incompetent. To this end, we attempted to study the effect of linker group and its orientation (normal vs reverse ester/amide) on the decomposition of three pairs of isomeric cationic amphiphiles in the gas phase using ESI-MS. The CID spectra of normal esters/amide are very different from the corresponding reverse esters/amide. The major product ions were formed by a neighboring group participation process or a McLafferty rearrangement process. Dissociation curves were drawn based on the relative abundances of precursor ion and product ions in the CID spectra at each CE value. The dissociation curves clearly revealed that the normal esters/amide dissociate at lower CE values than compared to those for the corresponding reverse esters/amide. Such findings are consistent with the supposition that lipids possessing better gene delivery capabilities are more easily decomposed than transfection-incompetent lipids. Stated differently, the normal esters/amides (1, 3, and 5) possess flexible conformational structures by which their collision cross-sectional area increases resulting in their decomposition at lower CE value, whereas the reverse esters/amide are likely to be structurally rigid with less flexibility in their structures causing lower collision cross-sectional area. Such flexible conformational structures possibly impart multiple hydrogen bonding with the bio-environment leading to enhanced bioactivities.
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
Acknowledgements The authors thank the Director, CSIR-IICT, Hyderabad, for facilities and encouragement. Senior research fellowships for VD and MS and funding (CSC-0406) from CSIR, New Delhi are greatly acknowledged.
[15]
[16] [17]
REFERENCES [1] I. M. Verma, M. Somina. Gene therapy – promises and prospects. Nature 1997, 389, 239. [2] W. F. Anderson. Human gene therapy. Nature 1998, 392, 25. [3] P. P. Karmani, A. Chaudhuri. Cationic liposomes as nonviral carriers of gene medicines: resolved issues, open questions, and future promises. Med. Res. Rev. 2007, 27, 696. [4] S. Mishra, J. D. Heidel, P. Webster, M. E. Davis. Imidazole end groups on a linear, cyclodextrin-containing polycation produce enhanced gene delivery via multiple processes. J. Control. Rel. 2006, 116, 179. [5] D. M. Lynn, D. G. Anderson, D. Putman, R. Langer. Accelerated discovery of synthetic transfection vectors:
[18]
[19]
[20]
parallel synthesis and screening of a degradable polymer library. J. Am. Chem. Soc. 2001, 123, 8155. M. X. Tang, C. T. Redmann, F. C. Szoka. In vitro gene delivery by degraded polyamidoamine dendrimers. Bioconjugate Chem. 1996, 7, 703. J. F. Kukowska-Latallo, A. U. Bienlinska, J. Jonhson, R. Spinder, D. A. Tomalia, J. R. Baker. Efficient transfer of genetic material into mammalian cells using Starburst polyamidoamine dendrimers. Proc. Natl. Acad. Sci. USA 1996, 93, 4897. D. Luo, K. Haverstick, N. Belecheva, E. Han, W. M. Saltzman. Poly(ethylene glycol)-conjugated PAMAM dendrimer for biocompatible, high-efficiency DNA delivery. Macromolecules 2002, 35, 3456. M. Rajesh, J. Sen, M. Srujan, K. Mukherjee, B. Sreedhar, A. Chaudhuri. Dramatic influence of the orientation of linker between hydrophilic and hydrophobic lipid moiety in liposomal gene delivery. J. Am. Chem. Soc. 2007, 129, 11408. M. Srjun, V. Chandrashekhar, R. C. Reddy, R. Prabhakar, B. Sreedhar, A. Chaudhuri. The influence of the structural orientation of amide linkers on the serum compatibility and lung transfection properties of cationic amphiphiles. Biomaterials 2011, 32, 5231. J. W. Russell, C. M. Robert. Mass spectrometric analysis of four regioisomers of F2-isoprostanes formed by free radical oxidation of arachidonic acid. J. Am. Soc. Mass Spectrom. 1996, 7, 490. P. Wheelan, J. A. Zirrolli, R. C. Murphy. Electrospray ionization and low energy tandem mass spectrometry of polyhydroxy unsaturated fatty acids. J. Am. Soc. Mass Spectrom. 1996, 7, 140. F.-F. Hsu, J. Turk. Differentiation of 1-O-alk-1′-enyl-2-acyl and 1-O-alkyl-2-acyl glycerolphospholipids by multiplestage linear ion-trap mass spectrometry with electrospray Ionization. J. Am. Soc. Mass Spectrom. 2007, 18, 2065. F.-F. Hsu, J. Turk, Y. Shi, E. A. Grosiman. Characterization of acylphosphatidylglycerols from Salmonella typhimurium by tandem mass spectrometry with electrospray ionization. J. Am. Soc. Mass Spectrom. 2004, 15, 1. S. H. Wilen, L. Delguzzo, R. Safersstein. Experimental evidence for aco-7 neighboring group participation. Tetrahedron 1987, 43, 5089. Y. He, J. P. Reilly. Does a charge tag really provide a fixed charge? Angew. Chem. 2008, 120, 2497. H. Lioe, C. K. Barlow, R. A. J. O′Hair. How does acetylcholine lose trimethylamine? A density functional theory study of four competing mechanisms. J. Am. Soc. Mass Spectrom. 2009, 20, 238. K. Bulleigh, A. Howard, T. Do, Q. Wu, V. Anbalagan, M. Van Stipdonk. Investigation of intramolecular proton migration in a series of model, metal-cationized tripeptides using in situ generation of an isotope label. Rapid Commun. Mass Spectrom. 2006, 20, 227. G. Wood, A. M. Falick, A. L. Burlingame. Apparent charge localization effects in the McLafferty rearrangement of acetate ester; field ionization kinetic measurements. Org. Mass Spectrom. 1974, 8, 279. T. J. Reddy, S. P. Mirza, U. V. R. V. Saradhi, V. J. Rao, M. Vairamani. Mass spectral studies of N,N-dialkylaminoethanols. Rapid Commun. Mass Spectrom. 2003, 17, 746.
1214 wileyonlinelibrary.com/journal/rcm
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1209–1214
