Analysis of α-Tocopherol in Tobacco and Cigarette Smoke

α-Tocopherol, α-tocopheryl acetate, 2,5-bis(5-tert-butyl-benzoxazol-2-yl)thiophene (Uvitex-OB), and 2,6-di-tert-butyl-4-methylphenol (BHT) as well as acetonitrile, formic acid, isopropanol, and methanol were purchased from Sigma-Aldrich (St. Louis, MO, USA).

The GC vials used were 2-mL with screw top caps with septa. Also 4-mL vials with screw top caps with septa were used. Polyvinylidene difluoride (PVDF) filters of 0.45-μm (Whatman Autovial, GE Healthcare, Little Chalfort, UK) were used for filtration.

Two HPLC systems were used for the analysis. One system with UV detection was a 1260–1290 HPLC system from Agilent (Wilmington, DE, USA). This HPLC system consisted of a binary pump, autosampler, column thermostatted compartment, and UV detector (diode array).

The peak integration was performed with OpenLab ChemStation. The other system was an Agilent 1200 HPLC binary system that consisted of a binary pump, an autosampler with cooling capability, and a column thermostatted compartment coupled with an API-5000 triple quadrupole mass spectrometer (AB Sciex, Framing-ham, MA, USA) controlled using Analyst 1.6.2 software. The peak integration was performed with MultiQuant 3.0.1 software. Both HPLC systems were equipped with a Zorbax Eclipse XDB-C18 column, 4.6 × 250 mm, 5 μm, also from Agilent. An ultrasonic bath 5510 from Branson (Danbury, CT, USA) was used. For smoking, a Cerulean SM 450 smoking machine (Cerulean, Linford Wood East, UK) has been used. Tobacco moisture was measured with halogen moisture analyzer HE53 (Mettler Toledo, Greifensee, Switzerland).

Two initial standard stock solutions - one containing 98.4 μg/mL α-tocopheryl acetate and the other 100.6 μg/mL α-tocopherol - were prepared in methanol. Each stock solution also contained 5604.0 μg/mL BHT. BHT is used as an antioxidant to prevent tocopherol and tocopheryl acetate oxidation. The stock solutions were also used as the highest standard.

The Level 2 of dilution was prepared by mixing equal volumes of the two stock solutions and the other standards were obtained by successive dilutions with methanol. 20 μL of Uvitex-OB internal standard (I.S.) were added to 1 mL of each standard and the resulting solution concentrations for the calibration is given in Table 1. The concentration of the I.S. in each solution was 525.6 μg/mL.

Before the analysis, the moisture of each tobacco sample was measured. The analysis of tobacco was performed on tobacco “as is”, and 200 mg (± 5 mg) tobacco was precisely weighed in a 4-mL vial. 2 mL methanol containing 560.4 μg/mL BHT, and 40 μL internal standard were added to the vial. The capped vials were heated at 78 °C for 30 min with occasional agitation. After cooling, at room temperature the solutions were filtered through a 0.45-μm pore size PVDF filter and submitted for analysis.

For the analysis of cigarette smoke, smoking was performed using 35 mL puff volume, 2 s puff duration, and 60 s puff interval, with the cigarette filters not having the ventilation blocked (indicated as ISO (International Organization for Standardization)). The machine airflows were tuned for ISO conditions (40, 41). The smoking was performed at Eurofins/Lancaster Labs. Smoke from five cigarettes was collected in each run on a 44-mm diameter Cambridge pad. The pads were further extracted on a mechanical shaker for 30 min with 5 mL methanol containing 560.4 μg/mL BHT, and 100 μL solution of internal standard (Uvitex-OB) were added. An aliquot from the extract was filtered through a 0.45-μm pore size PVDF filter and submitted for HPLC analysis.

The HPLC separation was performed using the gradient with solvent A being acetonitrile and solvent B isopropanol. For the MS/MS analysis the acetonitrile also contained 0.1% formic acid. The gradient timetable is described in Table 2.

The detection by UV absorption was performed at 210 nm and the injection volume was 20 μL. The flow rate was 1.5 mL/min. For the MS/MS, because the flow rate of 1.5 mL/min was too high, the effluent from the chromatographic column was split, and only 0.5 mL/min were introduced into the mass spectrometer. Also, the injection volume was 10 μL.

The detection was kept on only for 16 min, the last part of the gradient being used to elute highly hydrophobic compounds from tobacco or smoke that have a long retention time in the separation. The chromatogram of standard Level 5 for the HPLC separation with UV detection is indicated in Figure 1.

Figure 1

The MS/MS detection using electrospray ionisation (ESI) was performed in multiple reaction monitoring (MRM) positive mode. The transitions utilized for detection are indicated in Table 3. The table also indicates the collision energy that was different for each analyte.

The other parameters for the MS/MS detection included:

The chromatogram for the extracted ions of the analyzed compounds in Level 5 standard is shown in Figure 2.

Figure 2

For UV detection, considering the stability of the UV response, the quantitation for the two analytes was performed using calibrations of amount of the analyte versus the peak area of the analyte (no normalization by the internal standard). The internal standard was used only to verify the repeatability of the chromatographic process. The analytes concentrations in standards were those indicated in Table 1.

The calibration curves were linear and are shown in Figure 3 for α-tocopherol and in Figure 4 for α-tocopheryl acetate.

Figure 3

Figure 4

The coefficients a and b for the calibration curves of the form Y = a X + b, where Y is the concentration of the analyte and X is the peak area are given in Table 4.

For the MS/MS detection, the quantitation for the two analytes was performed using calibrations of the amount of the analyte versus the normalized peak area of the analyte by the area of the internal standard. The calibration curves were linear and are shown in Figure 5 for α-tocopherol and in Figure 6 for α-tocopheryl acetate.

Figure 5

Figure 6

The coefficients a and b for the calibration curves of the form Y = a X + b, where Y is the concentration of the analyte and X is the peak area of the standard normalized by the area of the internal standard are given in Table 5.

Both LC-UV and LC-MS/MS methods for α-tocopherol and α-tocopheryl acetate analysis can be considered as having good specificity. For the UV detection, retention time is basically the only criterion for selectivity, but for the MS/MS detection, besides the retention time are the masses of the ion for Q1 (for the parent ion) as the ion for Q3 (for the daughter ion).

The validation was further performed regarding limits of detection (LOD) and limits of quantitation (LOQ), linearity range, precision and intermediate precision, spike/recovery, accuracy, as well as solutions stability (see e.g. (42)).

The application of the procedure of considering LOD = 3 × SD (SD = standard deviation of the lowest standard) (42) would indicate from the SD values for α-tocopherol (SD = 0.008 μg/mL, three replicates) and for α-tocopheryl acetate (SD = 0.001 μg/mL, three replicates) that LOD = 0.024 μg/mL for tocopherol and LOD = 0.003 μg/mL for tocopheryl acetate (the LOQ values are three times higher). However, these values are not realistic since the true measurements are not feasible at these levels. For this reason, the limit of quantitation (LOQ) can be considered as equal with the lowest standard for both UV detection and MS/MS detection, and the values are indicated in Table 6.

Considering that the samples for tobacco analysis are prepared using a 1/10 dilution in methanol, the LOQ per g of sample is 10 times lower. The same LOQ values were considered for smoke analysis.

The precision of the method can be considered very good, based on the R² values of all the calibrations. However, samples were not injected a sufficient number of times to generate specific precision levels.

The recovery of α-tocopherol from the tobacco matrix was evaluated by performing a second extraction of the tobacco sample. For this reason, after the first extraction and the filtration of the liquid, the remaining solid tobacco was placed on a filter paper and, as much as possible, the remaining liquid was adsorbed on the paper. After that the extracted tobacco was allowed to dry. From this material, 100 mg were weighed in a 2-mL GC vial and 1 mL methanol containing BHT, plus 20 μL solution of the internal standard were added. The sample was heated at 78 °C for 30 min with occasional agitation and after cooling the solutions were filtered through a 0.45-μm pore size PVDF filter and submitted for the HPLC analysis. The procedure was performed for four tobacco samples. The results are indicated in Table 7.

As indicated in Table 7, only a very small proportion of α-tocopherol was recovered in a second extraction, indicating a very good recovery of the analytical technique.

Based on literature reports (35) in the presence of BHT, the solutions are stable up to two weeks when stored at −5 °C.

The list of tobaccos evaluated for the level of α-tocopherol and the potential presence of α-tocopheryl acetate is given in Table 8. The table also indicated wether cigarettes were made or not made from the corresponding tobacco.

The cigarettes were manufactured in the R.J. Reynolds Tobacco Co. Fabrication Laboratory to achieve a draft with perforated filter holes closed of about 155 mm H₂O and a draft with holes open of about 125 mm H₂O and different tobacco weights, with cigarette length: 83 mm, rod length: 56 mm, filter length: 27 mm, and circumference: 24.48 mm. The cigarettes had cellulose acetate filters with segment draft of about 100 mm H₂O and laser perforations.

The analysis was performed for both α-tocopherol and for α-tocopheryl acetate. However, α-tocopheryl acetate was not detected in tobacco. The results of α-tocopherol as measured in the tobacco samples are given in Table 9. The results from Table 9 indicate that flue-cured tobaccos have typically higher levels of α-tocopherol than Burley and Oriental tobaccos with the upper stalk leaves slightly higher in α-tocopherol compared to the lower stalk leaves. Oriental and Burley tobacco had about the same range of levels of the compound, with Samsun Oriental having the lowest level.

The results for the same analysis performed using the MS/MS detection generated very similar results. A graph showing the correlation between the UV data and the MS/MS data is given in Figure 7.

Figure 7

The total particulate matter (TPM) results in g for five cigarettes smoked under ISO type conditions are indicated in Table 10.

The same table gives the average results for α-tocopherol in the collected TPM from the UV analysis.

Very similar results with those obtained using UV detection were obtained using MS/MS detection for α-tocopherol. A graph showing the correlation between the UV data and the MS/MS data for the smoke levels of α-tocopherol is shown in Figure 8.

Figure 8

As expected, the correlation between the UV and the MS/MS results is very good.

The results for α-tocopherol in smoke show a relatively good correlation with the level of the compound in the tobacco. In Figure 9 the correlation between the level of α-tocopherol in smoke as a function of the level of the compound in tobacco is shown. In the LC/MS smoke analysis, the trace of the transition m/z = 473.4 → 165.1 corresponding to α-tocopheryl acetate shows a very small peak at the retention time of α-tocopheryl acetate. The peak area was below the quantitation limit, and it could not be determined if it belongs indeed to α-tocopheryl acetate, or it is an interference. The formation of even very low traces of α-tocopheryl acetate in cigarette smoke is not likely.

Figure 9