Does the Level of NNK in Tobacco and Tobacco Products Depend on the Level of Pseudooxynicotine or of Nicotine-1′-N-Oxide?

The method for nicotine analysis started with sample extraction. For this purpose, 250 mg tobacco (“as is”) or tobacco products were weighed in a 20-mL scintillation vial. To this vial, 10 mL extraction solution was added. The extraction solution consisted of 100 mM ammonium acetate in water. The extraction was performed for 1 h on a wrist action shaker. The resulting solution was filtered through 0.45-µm pore size polyvinylidene fluoride (PVDF) filters. From the filtered solution 200 µL were placed in 4-mL screw-cap vials. To the vials were added 3.8 mL water. The analysis was performed on a 1100 Series High performance liquid chromatography (HPLC) system from Agilent (Wilmington, DE, USA) consisting of a solvent delivery system, a quaternary pump, autosampler, column temperature controller, and a variable wavelength detector. For the separation a Xterra^® RP 18,5 µm, 4.6 × 150 mm column from Waters (Milford MA, USA) was used. The detection was performed by ultraviolet (UV) absorption at 254 nm. The mobile phase was made from solvent A: 0.025 M NH₄HCO₃ in water brought to pH 9.5 with NH₄OH, and solvent B: acetonitrile. The separation was performed in gradient starting with 100% solvent A to 60% A at 10 min, hold for 2 min and then reverse to initial conditions. The total run time was 15 min. The injection volume was 5 µL. The quantitation was performed using calibration with six nicotine standard solutions obtained by sequential dilution having concentrations between 836.5 µg/mL and 26.1 µg/mL. The resulting calibration curve was linear with a correlation coefficient R² = 1.0000.

For the analysis of NNK, the same extraction procedure as for nicotine was utilized. The solution obtained from the extraction of 250 mg tobacco or tobacco products was analyzed without dilution. The separation was achieved on an Agilent 1290 HPLC binary system with a binary pump, an autosampler with cooling capability, and a column thermostatted compartment. The instrument was equipped with a Luna^® 3 µm C18 150 × 3 mm column from Phenomenex (Torrance, CA, USA). The tandem mass spectrometry (MS/MS) system was an API-5000 triple quadrupole mass spectrometer (AB Sciex, Framingham, MA, USA). The liquid chromatography with tandem mass spectrometry (LC-MS/MS) system was controlled using Analyst 1.6.2 software, and the peak integration was performed with MultiQuant 3.0.1 software. For the HPLC separation a gradient program with the mobile phase A: 100 mM ammonium formate pH 4.9 aqueous solution with 5% acetonitrile and mobile phase B: 0.2% formic acid in acetonitrile. The timetable for the gradient of the HPLC separation is reported in the literature (7). The detection was performed using scheduled multiple reaction monitoring (MRM) in positive mode, the mass spectrometric settings being previously reported (7).

For the quantitation of NNK, seven standard solutions were prepared by sequential dilution having concentrations between 141.10 ng/mL and 2.52 ng/mL. The solutions contained 40.28 ng/mL NNK-d₄ as internal standard (I.S.). Calibration curves were constructed by plotting the concentrations of the standard solutions as a function of the peak areas normalized to the area of the corresponding deuterated internal standard. The calibration was linear with correlation coefficients R² = 0.9998. More details about the method for NNK analysis are given in reference (7). One aspect requiring some comments is related to the presence in tobacco of free and bound NNK (8). Exhaustive extraction of tobacco in rather harsh conditions seems to increase the level of detected NNK.

In situ generation of NNK due to the extraction conditions is not excluded to be in part the cause of additional NNK formation during the harsh extraction procedure. In the present work, only the free NNK obtained by extraction using a buffer solution was measured.

For the analysis of PON and nicotine-1′-N-oxide, the same extraction procedure has been used for nicotine. Also the same dilution from 200 µL extract to 4 mL final volume by adding extracting solution was applied to the samples. The extraction of PON and the nicotine-1′-N-oxide with a buffer solution was reported in the literature as efficient (9). To each 1 mL of sample solution (or of standards used for quantitation) were added 20 µL of a solution of nicotine-d₃ containing 133 µg/mL. The concentration of the I.S. in the final solution was 2.66 µg/mL (not adjusted to 1.02 mL volume). This solution was further analyzed. The instrument used for the analysis was an Agilent 1290 HPLC system with a binary pump, an autosampler with cooling capability, and a column thermostatted compartment. The MS/MS system was an API-6500 triple quadrupole mass spectrometer. The LC-MS/MS system was controlled using Analyst 1.6.2 software, and the peak integration was performed with MultiQuant 3.0.1 software. The HPLC separation was performed on an Acquity UPLC™ BEH Phenyl 1.7 m column, 3 × 150 mm dimensions from Waters (10). The mobile phase was made from solvent A: 0.02 M NH₄COOH in water brought to pH 9.5 with NH₄OH, and solvent B: acetonitrile. The gradient program started with 4% solvent B kept constant for 2.0 min followed by an increase to 60% at 6 min. The 60% B solvent was kept constant for 1.0 min followed by decrease to the initial condition at 8 min and kept at 4% B for one more min. The electrospray ionization (ESI) detection was performed by MS/MS in MRM positive mode using the parameters indicated in Table 1.

Other parameters include curtain gas: 20 mL/min, N₂ collision gas: 4 mL/min, ion spray voltage: 4500 V, temperature: 500 °C, ion source gas (1) 40 mL/min, ion source gas (2): 50 mL/min.

A typical chromatogram obtained in the previously described conditions for a Burley tobacco sample showing the peaks of nicotine-1′-N-oxide, PON and of the internal standard nicotine-d₃ is given in Figure 1. Several additional peaks corresponding to the same transition as for PON molecule m/z 179.1 → 148.1 but at different retention times than for PON standard are presented in the chromatogram. The nature of these peaks was not determined.

Figure 1

For the quantitation of PON and nicotine-1′-N-oxide, calibration curves were constructed by plotting the concentrations of the standard solutions as a function of the peak areas. For this purpose, seven standards were used. They were obtained by sequential dilutions with water of an initial solution generating standards with concentrations between 1.419 µg/mL and 0.022 µg/mL for PON and between 2.000 µg/mL and 0.031 µg/mL for nicotine-1′-N-oxide for each sample (not adjusted from 1.0 mL to 1.02 mL volume, the difference caused by the addition of the internal standard).

The calibration curves were obtained for both (concentration)/(peak area) as well as for (concentration)/(normalized peak areas by the area of I.S.). Since the use of normalized peak areas in the back-calculation of concentrations did not provide better results, only the correlation standard concentration versus peak area was further used for quantitation. The peak area for nicotine-d₃ was used only as a chromatographic standard to verify the integrity of the separation. The best fit for the calibration curve for PON quantitation was quadratic and is expressed using the formula: [1] Y(conc)=a(Std. Area)2+b(St. Area)+c Y\left( {conc} \right) = a{\left( {Std.\,Area} \right)^2} + b\left( {St.\,Area} \right) + c

For nicotine-1′-N-oxide the calibration curve was linear and is expressed using the formula: [2] Y(conc)=b(Std. Area)+c Y\left( {conc} \right) = b\left( {Std.\,Area} \right) + c

Parameters a, b, and c as well as correlation coefficient R² for the calibration lines are given in Table 2.

The results from the calculations were corrected to “dry weight basis” values.

A summary method validation was further performed for the PON and nicotine-1′-N-oxide analysis. The selectivity of the method is based on using three parameters to differentiate each compound, m/z for the molecular ion (for Q1), m/z for the fragment ion (for Q3), and the retention time. No additional transition was utilized for verifying selectivity. The method accuracy was verified by obtaining the same back-calculation results from the areas in the chromatogram for all the standards. The precision of the method was verified to be very good.

For the lowest standard of PON the back calculation of the concentration for 3 replicates, RSD% was 7.06% with a deviation of the average by + 0.002 µg/mL from the standard value. For the lowest standard of nicotine-1′-N-oxide the back calculation of the concentration for 3 replicates, RSD% was 3.50% with a deviation of the average by + 0.001 µg/mL from the standard value.

The limit of quantitation (LOQ) for both compounds can be taken as equal with the value of the lowest standard, although based on signal to noise (S/N) values, a much lower LOQ would be obtained. The LOQ values resulting by this procedure are LOQ (PON) = 0.022 µag/mL and LOQ (nicotine-1′-N-oxide) = 0.031 µg/mL.

An additional method for the analysis of PON has been developed and tested. This method includes a double extraction in the sample preparation with the intention to eliminate any potential interferences in the PON measurement. In this procedure, 100 mg tobacco or tobacco product were placed in a scintillation vial and extracted with 10 mL of a buffer at pH = 3.0 containing 1:1 (v/v) 0.16 M citric acid and 0.08 M disodium phosphate containing 39.5 µg/mL PON-d₃ as the internal standard. After shaken for 10 min, the extract solution was cleaned twice based on the following steps:

1) 100 µL of 5M NaOH and 1 mL hexane were added to 1 mL of the above extract solution and the mixture was vortexed for 30 s, which transferred PON into hexane layer,

The instrument used for the analysis was a ACQUITY UPLC™ H-Class system connected to a Xevo TQD triple quadrupole mass spectrometry (Waters Corporation, Milford, MA, USA). The sample separation was achieved using a Waters ACQUITY UPLC™ XBridge BEH C18 2.1 × 50 mm column with 1.7 µm particles at a flow rate of 0.25 mL/min. The elution was made with mobile phases consisting of solvent A 10 mM ammonium acetate with 0.1% NH₄OH and solvent B 100% acetonitrile. The ultra performance liquid chromatography (UPLC) gradient program was an initial 2% solvent B, increased linearly to 90% in 3.5 min and further linearly to 100% B in 0.1 min and held for 1 min followed by a sharp decrease to 2% solvent B where it was held for 4 min for re-equilibration. The injection volume was 2 µL. The Waters Xevo TQD used ESI in positive mode with MRM detection. The quantitative fragmentation transitions for MRM were m/z 179 → 106 for PON, and m/z 182 → 106 for PON-d₃, → while the qualitative fragmentation transitions for MRM were m/z 179 → 148 for PON, and m/z 182 → 148 for PON-d₃. The source and desolvation temperatures were 150 and 500 °C, respectively. The desolvation gas flow was 800 L/hr. The capillary voltage was 0.75 kV. Using these conditions, the typical chromatograms for PON and PON-d₃ in a 3R4F sample and in a standard with a concentration of 46.2 ng/mL are shown in Figure 2.

Figure 2

For the quantitation of PON, a standard calibration curve was established by injecting a series of individual PON standard solutions of known concentrations in the range between 45 ng/mL and 2200 ng/mL. The PON standard calibration solutions were made by mixing a known amount of PON and an internal standard (PON-d₃) into pH = 3.0 buffer. Masslynx software was used to process the data. The ratio of the peak area of PON to that of PON-d₃ was calculated. The graph of the peak area ratio versus the concentration ratio of PON to PON-d₃ was plotted. The calibration type was linear with 1/x weighting. The regression line was not forced through the origin and generated R² = 0.999.

The validation of the second method for PON analysis was performed according to literature recommendations (11). The method was validated for the accuracy of PON at different concentrations with the recovery study. The reference cigarette 3R4F ground filler extract solution was used as a matrix for the recovery study. Since tobacco samples usually contain a certain amount of PON, PON-d₄ (quantitative transition 183 → 152) was spiked into 3R4F extract solution for the recovery test, and PON-d₃ was used again as an internal standard. The low, medium, and high levels of the PON-d₄ were spiked into 3R4F and followed the sample preparation procedure as previously described for PON. Three replicates of each concentration level were tested. The average recovery of PON-d₄ in the extraction procedure was 102% and it is indicated in Table 3. The very good recovery of PON-d₄ is a reasonable proof that PON recovery is also very good.

For the evaluation of precision of the UPLC-MS/MS method, the amount of PON in 3R4F filler was measured on three replicates each day for three different days. The coefficient of variations (RSD%) of intra-day and inter-day were found to be within 4% and 8%, respectively, indicating the repeatability of this method.

The limit of detection (LOD) and the limit of quantitation (LOQ) were based on the standard deviation of y-intercepts of the regression line and the slope. The PON calibration curve was made by injecting a series of known concentration PON standards into the UPLC-MS/MS. The LOD and LOQ were estimated by the standard deviation of the response (SD) and slope (S) from the standard curve using the formula LOD = 3.3 × SD/S and LOQ = 10 × SD/S. The LOD and LOQ of PON are 0.3 ng/mL and 1 ng/mL, respectively.