A New Look at Radicals in Cigarette Smoke - Analytical Chemistry

May 26, 2007 - Philip Morris USA Postgraduate Research Program, and Philip Morris USA Research Center, P.O. Box 26583, Richmond, Virginia 23261. Anal...
1 downloads 12 Views 94KB Size
Anal. Chem. 2007, 79, 5103-5106

Correspondence

A New Look at Radicals in Cigarette Smoke Judit Bartalis,† W. Geoffrey Chan,‡ and Jan B. Wooten*,‡

Philip Morris USA Postgraduate Research Program, and Philip Morris USA Research Center, P.O. Box 26583, Richmond, Virginia 23261

Radicals in cigarette smoke have been proposed to contribute to the harm caused by cigarette smoking. For the first time, using HPLC and high-resolution mass spectrometry analysis of stable radical adducts, we have identified specific radical species in cigarette smoke: 7 acyl and 11 alkylaminocarbonyl radicals. Their combined abundance measured in fresh whole smoke from a single 2R4F cigarette is ∼225 nmol (1.4 × 1017 radicals). The fiberglass Cambridge filter pad conventionally employed to separate the gas phase from mainstream smoke was found to reduce the apparent yield of these radicals, introducing artifacts of measurement. The long-accepted steady-state mechanism for the formation of carboncentered radicals in cigarette smoke involving NO2 chemistry cannot account for these newly identified radicals, and it does not in general appear to be a major source of carbon-centered radicals in fresh mainstream cigarette smoke. Consequently, we suggest that the precise nature of radicals in cigarette smoke warrants reexamination. Radicals in cigarette smoke were proposed to contribute to the harm caused by cigarette smoking by Pryor et al. in 1983, and numerous reports have associated radicals with smokingrelated oxidative stress and damage.1,2 Nevertheless, specific radical species in cigarette smoke have never been positively identified. Analysis of most smoke constituents, including radicals measurement by electron paramagnetic resonance (EPR), typically involves the separation of gas and particulate-phase constituents by passing the smoke through a 0.1 µm Cambridge filter pad, which can potentially introduce artifacts.3,4 We have adapted a * To whom correspondence should be addressed. E-mail: jan.b.wooten@ pmusa.com. † Philip Morris USA Postgraduate Research Program. ‡ Philip Morris USA Research Center. (1) (a) Pryor, W. A.; Hales, B. J.; Premovic, P. I.; Church, D. F. Science 1983, 220, 425-427. (b) Pryor, W. A.; Dooley, M. M.; Church, D. F. Biochem. Biophys. Res. Commun. 1984, 122, 676-681. (c) Pryor, W. A.; Dooley, M. M.; Church, D. F. Chem-Biol. Interact. 1985, 54, 171-183. (d) Church, D. F.; Pryor, W. A. Environ. Health Perspect. 1985, 64, 111-126. (2) Wooten, J. B.; Chouchane, S.; McGrath, T. E. In Smoking and Oxidative Stress; Halliwell, B. B., Poulsen, H. E., Eds.; Springer: Berlin, 2006; Chapter 2. (3) Baker, R. R. In Tobacco. Production Chemistry, Chemistry and Technology; Davis, E. L., Nielsen, M. T., Eds.; Blackwell Science: Oxford, 1999; pp 398439. (4) Shorter, J. H.; Nelson, D. D.; Zahniser, M. S.; Parrish, M. E.; Crawford, D. R.; Gee, D. L. Spectrochim. Acta, Part A 2006, 63, 994-1001. 10.1021/ac070561+ CCC: $37.00 Published on Web 05/26/2007

© 2007 American Chemical Society

radical-trapping method5 that is specific for carbon-centered radicals that does not rely on either the separation of the mainstream smoke phases or the use of EPR spin-traps. Moreover, the method utilizes a solvent-free trapping approach suitable for analyzing radicals in freshly emitted whole cigarette smoke. Using this approach, we have identified carbon-centered radicals with chemical structures that are different from those previously postulated and measured in significantly higher yields than the reported generic alkyl radicals. Flicker and Green introduced a method to trap and quantify carbon-centered radicals in whole cigarette smoke where smoke emitted from the filter end of a burning cigarette is passed through a column packed with glass beads coated with a nitroxide radical scavenger, 3-amino-proxyl radical (3AP).5 Reaction of 3AP with carbon-centered radicals is highly selective, kinetically rapid (k ) 2 × 1010 M-1 s-1), and forms stable diamagnetic adduct.6 The radical adducts are washed from the glass beads and derivatized with a fluorescent agent, naphthalene-2,3-dicarboxaldehyde (NDA). The fluorescent products were subsequently separated by HPLC and quantified with very high sensitivity by fluorescence detection (HPLC/FLD). The HPLC/FLD chromatograms reported by Flicker and Green for smoke from different tobacco products exhibited distinctive patterns, but the chemical constituents corresponding to individual chromatographic peaks were not characterized. We have made several improvements to the HPLC/FLD method including using standard FTC smoking conditions with a smoking machine,7 optimizing the radical trapping parameters, and eliminating interferences and false assignments in the LC chromatograms. In addition, we extended the method by measuring radicals in both gas phase and aged whole smoke and by employing mass spectrometry and NMR spectroscopy for chemical identification. By omitting the fluorophore derivatization step, we excluded numerous false radical candidates that appear as extraneous NDA adducts of miscellaneous primary amines in smoke. Detection of the 3AP radical adducts without NDA derivatization was by triple-quadrupole mass spectrometry (HPLCMS/MS). In separate experiments, we achieved further discrimination by employing an alternative nitroxide radical scavenger, (5) (a) Flicker, T. M.; Green, S. A. Anal. Chem. 1998, 70, 2008-2012. (b) Flicker, T. M.; Green, S. A. Environ. Health Perspect. 2001, 109, 765771. (6) Blough, N. V. Environ. Sci. Technol. 1988, 22, 77-82. (7) Pillsbury, H. C., Jr. NIH Publication 96-4028, 1996.

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007 5103

3-cyanoproxyl radical (3CP), which lacks the amine group of 3AP that introduces artifacts into the mass spectral analysis by reaction with aldehydic and carboxylic smoke constituents. Exchangeable protons on the radicals were identified by deuterium labeling experiments. Individual molecular formulas of the 3AP adducts from isolated HPLC fractions were determined from high-precision mass measurements by linear ion trap Fourier transform ion cyclotron resonance mass spectrometry (LIT-FTICR). Molecular structures of the most abundant adducts were confirmed by chemical synthesis8 and by 500-MHz NMR spectroscopy. EXPERIMENTAL SECTION Materials. HPLC grade organic solvents and water, NaCN, and solid borosilicate glass beads of 2- and 3-mm diameter were purchased from Fisher Scientific; NDA and 3CP were from Aldrich; and Na2B4O7, 3AP, cyclopentylamine, acetohydroxamic acid, and acetaldoxime were from Acros Organics. All chemicals were purchased at their highest quality. The glass columns for trapping radicals and quartz column for pyrolysis were ordered from Research Glass (Richmond, VA). The smoke filter assembly consists of a Cambridge fiberglass sheet inserted into a disposable holder (Gelman Scientific, Ann Arbor, MI). Kentucky 2R4F research grade cigarettes9 were obtained from the University of Kentucky (Lexington, KY). The blend composition and cigarette design10 of 2R4F cigarettes is representative of domestic cigarettes manufactured by major U.S. cigarette companies, and its smoke chemistry is very similar.11 Inert Kynar tubing of 1/4-in. i.d. (Cole Parmer Instrument Co.) was purchased through Fisher Scientific. Cellulose cigarettes were machine-made and contained 40% cellulose and 60% Ca2CO3. Cigarettes were conditioned at 24 °C and at a humidity of 60% in a controlled laboratory for at least 24 h prior to smoking. Smoking and Pyrolysis. The smoke collection method was adapted from Green and Flicker.5 The surface area of 3AP- or 3CP-coated glass beads loaded in the trap of 35-mL void volume was 860 cm2 (3-mm-diameter beads) for standard measurements. Additionally, a larger surface area (1290 cm2, 2-mm-diameter beads) was considered for the evaluation of trapping efficiency. Three cigarettes were smoked in series on a five-port smoking machine (KC Automation, Inc., Richmond, VA) in a conditioned room. For gas-phase smoke collection, a Cambridge filter pad in disposable holder was inserted between the cigarette and the trap. For the collection of aged smoke, Kynar tubing was inserted between the cigarette and the trap in order to age smoke for 0.5, 1, 2, or 3 min. All smoking experiments involved smoke controls with no 3AP or 3CP deposited on glass beads. For pyrolysis, tobacco filter and 3 mm of the tobacco rod was cut off the 2R4F research cigarette and the rest placed into the pyrolyzing chamber. More details about pyrolysis are presented by McGrath et al.12 (8) Bartalis, J.; Flora, J. W.; Paine, J. B. III; Zhao, Y.-L.; Wooten, J. B.; Chan, W. G. In preparation. (9) Chen, P. X.; Moldoveanu, S. C. Beitrage Tabakforsch. 2003, 20, 448-458. (10) Davies, M. H.; Vaught, A. The reference cigarette. Kentucky Tobacco Research and Development Center, University of Kentucky, Lexington, KY, 1990. (11) Counts, M. E.; Hsu, F. S.; Tewes, F. J. Regul. Toxicol. Pharmacol. 2006, 46, 225-242. (12) McGrath, T. E.; Chan, W. G.; Hajaligol, M. R. J. Anal. Appl. Pyrolysis 2003, 66, 51-70.

5104

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

After smoke or pyrolysate collection, the beads were transferred into an Erlenmeyer flask containing 25 mL of 5% aqueous acetonitrile. The flask was closed and shaken gently for 30 min. Part of the sample was filtered through a 0.2-µm syringe and analyzed immediately or stored at +4 °C in a glass vial for later analysis. The 3AP-R and 3CP-R adducts were stable for at least 1 week. HPLC/FLD. Samples containing 150 µL of 3AP-R were mixed with 50 µL of 10 mM cyanide and excess NDA (50 µL/10 mM), and the volume was adjusted to 1 mL with 25 mM, pH 9.2 borate buffer. To the control samples (smoke or pyrolysate) 18 µL of 1 mg/mL 3AP was added. The HP-1100 HPLC system (Agilent Technology) parameters were as follows: autosampler at ambient temperature; injection volume 10 µL; column temperature 30 °C; and flow rate 0.5 mL/min. The LC column was Symmetry C18 (150 mm × 3.9 mm, 5 µm, Waters). The LC conditions were as follows: solvent A, water; solvent B, methanol; column equilibration time with 75% B 10 min before injection; elution gradient, 0-3 min 75% B, 3-20 min 75-90% B, 20-30 min 90% B. The FLD was operated at 420/480 nm excitation/emission wavelengths. The NDA adduct of cyclopentylamine was used as the external standard for quantitation. A standard curve was created from six points for a concentration range of 8.6-841.5 nM (r2 >0.9999), with a LOD of 0.2 nM (57.5 pg/mL). ESI(+)-MS/MS Infusion. Mass spectrometric analyses were conducted with a Micromass Quattro Ultima triple-quadrupole tandem mass spectrometer (Waters Corp., Milford, MA). Smoke samples were infused in normal or deuterated 5% aqueous methanol (5% CH3OD in D2O) and 0.1% formic acid; the deuterated solvent afforded the number of active hydrogens for each radical. Precursor ion monitoring technique was employed for both 3AP-R and 3CP-R samples, considering major fragments m/z 98 and 151, respectively. HPLC-ESI(+)-MS/MS Analysis. Instruments presented above were employed, and both eluents A and B contained 0.1% formic acid. The LC column C18 Symmetry (100 mm × 2.1 mm, 3.5 µm, Waters) was used at flow rate 0.3 mL/min. Equilibration time was 10 min with the initial eluent mixture. Gradient elution for 3AP-R analysis is as follows: 0-5 min 5% B; 5-20 min 5-45% B; 20.1 min 75% B; 20.1-25 min 75% B. The gradient elution for 3CP-R analysis is as follows: 0-4 min 35% B; 4-20 min 35-65% B; 20-25 min 65-75% B. Precursor ion monitoring by focusing on the base peak (m/z 98 for 3AP-R and m/z 151 for 3CP-R) permitted the detection of less abundant radical adducts. Multiple reaction monitoring by focusing on the base peak permitted monitoring radical adducts in various samples. ESI(+)-LTQ-FTICR-MS. The Finnigan LTQ-FT mass spectrometer comprised an ion-trap coupled to a 7-T Fourier transform ion cyclotron resonance MS (Thermo Electron). Pure HPLC fractions for 3AP-R and 3CP-R smoke samples were infused for exact mass measurements. RESULTS AND DISCUSSION In our analysis, 10 radicals with two distinct formulas, CnH2n-1O and CnH2nNO, were identified in whole smoke from 2R4F research cigarettes using the 3AP trapping agent. Due to interferences, absolute quantification of only three major radicals was possible by HPLC/FLD (Figure 1). Based upon the HPLC-MS/MS screening, these radicals contributed 87% to the total radical yield

Figure 1. Abundance of major carbon-centered radicals trapped with 3AP in fresh and 1-min aged, whole and gas-phase smoke of 2R4F cigarettes and fresh whole smoke of cellulose cigarettes, measured by HPLC/FLD after derivatization with NDA. External standard cyclopentylamine-NDA was employed for quantification. The cellulose cigarette contained a small amount of nitrogen impurity (∼0.38%), which accounts for the presence of •C(O)NHCH3. Table 1. Molecular Structures and Corresponding Precise Masses for the Three Most Abundant Carbon-Centered Radicals Detected in Smoke from 2R4F Cigarettes Ha

∆ 3APR (ppm)

∆ 3CP-R (ppm)

mass R• (amu)b

1

1

0.370

0.398

58.029

2

1

0.348

0.375

72.044

3

0

0.398

0.284

43.018

no.

formula R•

a Number of exchangeable protons for each radical. b Exact mass of radicals calculated from the theoretical mass of radical adducts.

and the concentration of all 10 radicals amounted to ∼225 nmol (1.4 × 1017 radicals) per cigarette. The chemical formulas of the nitrogen-containing radicals are consistent with three possible structures: two oxime-type isomers, •CH(R)sCHdNsOH and •C(R)dNsOH, and an alkylaminocarbonyl type, •C(O)sNHsR. 1H NMR analysis of the isolated HPLC fraction for 3CP-C2H4NO showed that the alkylaminocarbonyl radical is the correct structure,13 supported by additional spectra of standard compounds with structures similar to the oxime radicals, e.g., CH3sCHdNsOH and CH3sC(OH)dNsOH. Thus, two classes of carbon-centered radicals were discovered: acyl, •C(O)-R, and •C(O)-NH-R (Table 1). Structures for the most abundant representatives of both classes, •C(O)-CH3 and •C(O)-NH-CH3, were further confirmed by chemical synthesis of their 3CP adducts. Eight additional radicals were detected at low intensity in smoke using 3CP, including two unsaturated alkylaminocarbonyl radicals (Figure 2). Both the acyl and the saturated alkylaminocarbonyl radicals are present as a homologous series of masses corresponding to R ) C1-C4 and C1-C5, respectively. The widely accepted mechanism for the formation of both alkyl and alkoxy radicals in cigarette smoke was proposed by Pryor

Figure 2. Representative chromatograms of radicals in fresh mainstream smoke of 2R4F cigarettes trapped with nitroxides 3AP (A) or 3CP (B) and detected by MS/MS by precursor ion monitoring. Radicals formula: 1, C2H4NO; 2, C3H6NO; 3, C2H3O; 4,11, C4H8NO; 5, C3H5O; 6,7, C5H10NO; 8,13, C4H7O; 9,10,14, C6H12NO; 12, C5H8NO; 15, C7H12NO; 16-18, C5H9O.

and co-workers1,14 and supported by persuasive, but primarily indirect, evidence based on comparisons of gas-phase cigarette smoke and model gas mixtures.15 Carbon-centered radicals are postulated to evolve from gas-phase reactions within the smoke itself. NO2, which forms by oxidation of NO, reacts with various dienes in smoke, e.g., isoprene, to yield nitroalkyl radicals that undergo rapid air oxidation to form peroxy and alkoxy radicals. The apparent persistence and high concentration of these gasphase radicals was attributed to a steady-state process whereby NO2 is continually regenerated. Recently, Shorter et al.4 used tunable infrared laser differential absorption spectroscopy to measure NO2 with high sensitivity in whole and gas-phase cigarette smoke on a puff-by-puff basis. Surprisingly, NO2 was detected in whole cigarette smoke only in the lighting puff. The appearance of NO2 in gas-phase smoke is apparently an artifact introduced by the use of the fiberglass Cambridge filter pad. With our experimental setup, we did not detect any radicals containing NO2 groups of the type postulated by Pryor et al.1 and by Green and Flicker5 involving the addition of NO2 to dienes in either cigarette smoke, cigarette pyrolysate, or model gas mixtures of NO, air, and isoprene. Finally, when we aged the smoke 1 min before the radical trapping step inside highly inert tubing with an internal volume of 35 mL, the sum of all carbon-centered radicals decreased, in sharp contrast to the results of Pryor et al.1 If one considers the low yield of NO2 in whole smoke, the rapid velocity of the air passing through a cigarette during puffing (30-40 cm/s)3, and the slow rate of oxidation of NO to NO2 in air (k ) 7 × 103 M-2 s-1),17 then the steady-state model not only fails to account for the acyl and alkylaminocarbonyl radicals reported here but it cannot in general (13) 1H NMR results for 3CP-C(O)NHCH3 (smoke fraction 1): 2.95 δ (t) CH (J ) 9.2 Hz); 2.14 δ (d) CH2 (J ) 9.2 Hz, ∆ν/J ) 44.4) 2.76 δ CH3 (s). 1H NMR spectra were obtained in methanol solution on a Varian 500-MHz NMR system. (14) (a) Pryor, W. A.; Prier, D. G.; Church, D. F. Environ. Health Perspect. 1983, 47, 345-355. (b) Pryor, W. A.; Tamura, M.; Church, D. F. J. Am. Chem. Soc. 1984, 106, 5073-5079. (15) (a) Cueto, R.; Church, D. F.; Pryor, W. A. Anal. Lett. 1989, 22, 751-763. (b) Cueto, R.; Pryor, W. A. Vib. Spectrosc. 1994, 7, 97-111. (16) Burton, H. R.; Childs, G., Jr. Beitrage Tabakforsch. 1977, 9, 45-52. (17) Kinetics database at National Institute of Standards and Technology (http:// kinetics.nist.gov).

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

5105

be the major source of radicals in fresh whole cigarette smoke. Formation of NO2 by the steady-state mechanism in inhaled smoke, however, remains a possibility. Radicals of type •C(O)-NH-R have been detected as degradation products of proteins/peptides.20 Therefore, tobacco proteins/ peptides are potential precursors of similar radicals in smoke. Acyl radicals appear to be related to the abundance of aldehydes in smoke. The total measured amount of acyl radicals represents ∼0.6% of the total aldehyde content in smoke from 2R4F cigarettes.9 Furthermore, we observed that pyrolysis of tobacco in 5% O2 in He relative to inert environment at 300 °C increases the concentration of acyl radicals several fold, similar to the increase reported for acetaldehyde.16 Low molecular weight aldehydes in tobacco smoke are mainly thermal decomposition products of polysaccharides such as cellulose.18 Carbon-centered radicals from cigarettes made with cellulose rather than tobacco have been observed by Pryor et al. in gas-phase smoke19 and by us in whole smoke. In our experiments, the carbon-centered radicals from cellulose cigarettes were found to be almost entirely acyl radicals (Figure 1.). Using the Cambridge filter pad, the acyl radicals in 2R4F gas-phase smoke dropped by ∼96% compared to whole smoke. Based on the aging experiments, the lifetime of (18) Seeman, J. I.; Dixon, M.; Haussmann, H.-J. Chem. Res. Toxcicol. 2002, 11, 1331-1350. (19) Pryor, W. A.; Tamura, M.; Dooley, M. M.; Premovic, P.; Hales, B. J.; Church, D. F. In Oxy Radicals and Their Scavenger Systems. Volume II: Cellular and Medical Aspects; Greenwald, R. A., Cohen, G., Eds.; Elsevier: New York, 1983; pp 185-192. (20) Davies, M. J. Arch. Biochem. Biophys. 1996, 336, 163-172.

5106

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

the acyl radicals in smoke was found to be significantly shorter than the lifetime of alkylaminocarbonyl radicals (estimated t1/2 is ∼1 min), which are resonance stabilized and therefore more stable and longer lived. The alkylaminocarbonyl and acyl radicals reported here are the first individual radical species to be positively identified in cigarette smoke. In our measurements, the abundance of these radicals in cigarette smoke was ∼2 orders of magnitude higher than alkyl radicals previously measured by EPR spin-trapping methods. Their persistence in smoke suggests that they would be inhaled by the smoker. Because the radicals are reactive, they may oxidize quickly to peroxy and alkoxy radicals or they may quench by reaction with other radical or molecular species. Identification of specific organic radicals in cigarette smoke thus will enable a better understanding of the impact of cigarette smoking on health. ACKNOWLEDGMENT The authors benefited from discussions with Dr. Jia Wang, Dr. Salem Chouchane, Dr. Jeffrey Seeman, Dr. John B. Paine, Prof. Sarah Green, Prof. Barry Dellinger, and Prof. W. A. Pryor. We also thank Dr. Jason W. Flora for HRMS measurements and Dr. William D. Thweatt for his assistance in measuring the NOx level by TDL. Received for review March 20, 2007. Accepted May 9, 2007. AC070561+