- Email: [email protected]
Novel approach for selective reduction of NNN in cigarette tobacco filler and mainstream smoke
Regulatory Toxicology and Pharmacology 89 (2017) 101e111
Contents lists available at ScienceDirect
Regulatory Toxicology and Pharmacology journal homepage: www.elsevier.com/locate/yrtph
Novel approach for selective reduction of NNN in cigarette tobacco filler and mainstream smoke M. Lusso a, **, I. Gunduz b, *, A. Kondylis c, G. Jaccard c, L. Ruffieux b, F. Gadani d, K. Lion a, A. Adams a, W. Morris a, T. Danielson a, U. Warek a, J. Strickland a a
Altria Client Services LLC, Research Development & Regulatory Affairs, 601 E. Jackson St., Richmond, VA 23219, USA Philip Morris International Management SA, Leaf Agricultural Programs, Avenue de Rhodanie 50, 1001 Lausanne, Switzerland Philip Morris International R&D, Philip Morris Products S.A. (Part of Philip Morris International Group of Companies), Quai Jeanrenaud 5, 2000 Neuchatel, Switzerland d Philip Morris Products SA, Product Development, Rue des Usines 56, 2000 Neuchatel, Switzerland b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 May 2017 Received in revised form 17 July 2017 Accepted 19 July 2017 Available online 20 July 2017
Research conducted during past decades to reduce the level of the tobacco specific nitrosamine Nnitrosonornicotine (NNN) and its precursor nornicotine in tobacco yielded identification of three tobacco genes encoding for cytochrome P450 nicotine demethylases converting nicotine to nornicotine. We carried out trials to investigate the effect of using tobaccos containing three non-functional nicotine demethylase genes on the selective reduction of NNN in cigarette tobacco filler and mainstream smoke. Our results indicate that the presence of non-functional alleles of the three genes reduces the level of nornicotine and NNN in Burley tobacco by 70% compared to the level observed in currently available low converter (LC) Burley tobacco varieties. The new technology, named ZYVERT™, does not require a regular screening process, while a yearly selection process is needed to produce LC Burley tobacco seeds for NNN reduction. The reduction of NNN observed in smoke of blended prototype cigarettes is proportional to the inclusion level of tobacco having ZYVERT™ technology. Inclusion of Burley tobacco possessing the new trait into a typical American blend resulted in a selective reduction of NNN in cigarette smoke, while the levels of other Harmful and Potentially Harmful Constituents (HPHC) currently in the abbreviated list provided by the US Food and Drug Administration are statistically equivalent in comparison with the levels obtained in reference prototype cigarettes containing LC Burley. © 2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Keywords: NNN N-nitrosonornicotine Nornicotine Nicotine demethylase Cytochrome P450 TSNA Tobacco regulation Harmful and potentially harmful constituents HPHC Selective smoke constituents' reduction
1. Introduction Some tobacco chemicals and smoke constituents have been classified as Harmful and Potentially Harmful Constituents (HPHC) (FDA, 2012). Therefore, approaches should be used to reduce or eliminate human exposure. Tobacco-specific N-nitrosamines (TSNA) are a class of chemical compounds found in cured, unburned tobacco and tobacco smoke. Due to their potential toxicological relevance, public health researchers have focused on these compounds for many years. For example, the International Agency for Research on Cancer, Working Group on the Evaluation of
* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (M. Lusso), [email protected] (I. Gunduz).
Carcinogenic Risks to Humans (IARC) evaluated the available scientific evidence for four TSNAs e 4-(methylnitrosamino)-1-(3pyridyl)-1-butanone (NNK), N0 -nitrosonornicotine (NNN), N0 nitrosoanabasine (NAB) and N-nitrosoanatabine (NAT) in an effort to determine the carcinogenic potential of these compounds. IARC (IARC, 2007) classified NNK and NNN as Group 1 carcinogens (carcinogenic to humans) and NAB and NAT as Group 3 compounds (not classifiable as to their carcinogenicity to humans). The United States National Toxicology Program (USNTP) classified NNN and NNK “to be reasonably anticipated to be a human carcinogen” based on sufficient evidence of carcinogenicity from studies in experimental animals as described in the 12th Report on Carcinogens (NTP, 2011). In addition, the FDA included NNN and NNK on its HPHC list (FDA, 2012) to be reported in tobacco products and tobacco smoke. Nicotine, the predominant alkaloid found in commercial tobacco varieties, represents typically 90e95% of the total
http://dx.doi.org/10.1016/j.yrtph.2017.07.019 0273-2300/© 2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
102
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
alkaloids in cured leaf. The 5e10% remaining alkaloids are called minor alkaloids and are comprised primarily of nornicotine, anabasine and anatabine. These minor alkaloids are the direct precursors of TSNA in tobaccos through their nitrosation during the plant life and especially during curing process, i.e., the leaf drying after harvest (Sophia et al., 1989). NNN is formed essentially via nitrosation of nornicotine, while NNK is formed from nitrosation of et al., nicotine or its oxidized products (Sophia et al., 1989; Piade 2013). The presence of TSNAs in mainstream smoke is essentially due to a direct transfer from tobacco filler to mainstream smoke. There is also a certain proportion (5e25% of the total amount) of NNN released by pyrosynthesis, strongly dependent on nornicotine level (Moldoveanu and Borgerding, 2008). For NNK, both pyrosynthesis and pyrorelease occur, due to a large proportion of NNK present in matrix-bound form in tobaccos (Lang and Vuarnoz, 2015). Nornicotine, the main direct precursor of NNN, is generated directly from nicotine by oxidative demethylation through the activity of N-demethylase enzymes (Siminszky et al., 2005a; Gavilano et al., 2006; Xu et al., 2007). Although nornicotine typically represents less than 5% of the total alkaloids, tobacco plants that initially produce very low amounts of nornicotine give rise to progeny that metabolically “convert” large percentages of leaf nicotine to nornicotine due to the genetic expression of nicotine demethylase enzymes. This process is called “conversion” which predominantly occurs during senescence and curing of the mature leaf (Chakrabarti et al., 2008). Among predominantly used tobacco types Burley tobaccos are particularly prone to nicotine to nornicotine conversion, with rates up to 90% observed in the progeny of some cultivars grown over multiple generations. It is understood that regardless of signals involved in “turning on” the genes, nicotine demethylase enzymes appear to be predominantly involved in nornicotine biosynthesis (Gavilano et al., 2006). Flue-cured tobacco is generally considered more stable in terms of nicotine conversion, although some nornicotine production does occur. Miller et al. (2004) discovered that levels of nornicotine present in a given Burley tobacco variety can be reduced up to 70e80% by identifying and eliminating individual converter plants prior to seed collection. This “roguing” process is accomplished by chemically analyzing seed plants as they are grown in the field and destroying all plants that have levels of nornicotine greater than 3% relative to nicotine content prior to seed harvest. This screening process results in the production of seeds giving a low converter (LC) phenotype, commonly referred to as a “LC variety.” While LC varieties substantially decrease the nicotine to nornicotine conversion, they remain the same as the original variety with respect to yield, quality, agronomic character and disease resistance. The use of LC varieties for the production of Burley tobacco has become the standard in the United States of America and many other parts of the world and has contributed together with other good agricultural and curing practices to downward trends in terms of NNN levels in tobaccos and commercial cigarettes (Gunduz et al., 2016; Appleton et al., 2013). More recent research on the mechanism of nornicotine formation led to the identification of three nicotine demethylase genes, CYP82E4 (Siminszky et al., 2005b), CYP82E5 (Dewey and Lewis, 2011) and CYP82E10 (Lewis et al., 2010) that facilitate conversion of nicotine to nornicotine in leaf. The genes designated CYP82E4 (E4) and CYP82E5 (E5) are expressed in leaf tissue of the plant and the gene designated CYP82E10 (E10) is expressed in the roots. Formation of NNN in tobacco is provided in Fig. 1. Tobacco varieties with non-functional alleles of these three genes show decrease in conversion to stable levels of about 0.5% nornicotine relative to nicotine (Lewis et al., 2010) and a
Fig. 1. Formation of NNN in tobacco Nicotine demetylation is catalyzed by a group of cytochrome P450 enzymes of the 82E family. The demethylation leads to the formation of nornicotine. Nornicotine itself is nitrosated in N0 -nitrosonornicotine (NNN) through a non enzymatic reaction occurring at low rates during the plant life and intensely during the leaf drying period.
commensurate reduction in NNN content of the leaf. Mutation of these three P450 functional genes resulting in non-functional versions is commercially referred to as tobacco having ZYVERT™ technology, and the tobacco varieties using the technology are called ZYVERT™ tobaccos, hereafter. Major Burley commercial varieties containing three non-functional versions of these genes met minimum standards and were approved in 2013, 2014 and 2015 by the Regional Burley Variety Evaluation program conducted cooperatively by Universities in Kentucky, North Carolina, Tennessee, and Virginia. While ZYVERT™ technology is being integrated to commonly grown tobacco varieties, the impact of tobaccos containing this technology on cigarette smoke chemistry is not known. Therefore the objective of this study is to evaluate the impact of tobaccos containing ZYVERT™ in American blended cigarettes on the chemical composition of tobacco cut-filler and mainstream smoke. Data presented in this study include the percentage of nicotine conversion in low converter and ZYVERT™ plant populations, and the comparative level of chemical constituents identified in the list of HPHC (FDA, 2012) in tobacco products and tobacco smoke by the US FDA (abbreviated list). 2. Methods 2.1. Tobacco production Tobaccos used in the experiments described in this paper were produced and cured according to the recommended practices for Burley tobacco (Cooperative effort of University of Kentucky). Seeds from the experimental ZYVERT™ Burley tobacco varieties were received from the tobacco breeding program at North Carolina State University, Raleigh, NC except for the KT 204 ZYVERT™ experimental variety which was produced by Altria Client Services (ALCS), Richmond, VA. To evaluate the genetic stability of the ZYVERT™ technology, the experimental Burley tobacco variety KT 204 having ZYVERT™ technology and its low converter version (KT 204 LC) were grown in Virginia in 2013. Eighty-one plants from each variety were individually sampled and processed according to the low-converter protocol (Jack, 2007) and submitted for chemical analysis. To assess the impact of the ZYVERT™ technology on the chemical composition of cigarette filler and smoke, experimental Burley tobacco varieties containing the ZYVERT™ technology and commercially available tobacco varieties were grown in Virginia and Kentucky, respectively, during the 2012 and 2013 crop years. Cured leaves were de-stemmed, dried, cut and used for making machine made cigarettes prototypes. 2.2. Prototype cigarette construction Experimental cigarettes were prepared with components (cellulose acetate filters, papers, and adhesives) and designs consistent
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
103
in filler and smoke, which were performed by a liquid chromatography-tandem mass spectrometry method, according to Labstat internal methods T-309 and T-311 respectively (Canada, 2000). For HPHC filler analyses of group 2 prototypes, it was requested that nicotine, NNN and NNK analysis be conducted with three replicates. For cigarette smoke of group 2 prototypes, it was requested seven replicates, except for nicotine and carbon monoxide, in which eight replicates were requested. Two different prototypes were manufactured in group 2, all Burley cigarette models, which were analyzed for HPHCs in cigarette smoke with three replicates and HCI smoking conditions only. All the conducted studies quantified NNN, NNK, and nicotine both in tobacco filler and in mainstream smoke. ISO and HCI results are available for mainstream smoke in blended cigarettes, while only HCI results are available for mainstream smoke in Burley single component cigarette prototypes. NAB, NAT and minor alkaloids were measured in the prototypes cut fillers of group 2 prototypes only.
with commercial US cigarette manufacturing. Cigarettes measured 84 mm in length (57 mm tobacco rod, 27 mm filter) and 25 mm in circumference. The prototype cigarettes were produced at the same time and the same constant weight at Altria Client Services (ALCS) (Group 1), or Philip Morris International (PMI) (Group 2); see Table 1. One sub-group of cigarette prototypes was made as a single component containing only Burley tobacco, while the three other sub-groups were designed as typical American blended cigarettes. Single component Burley cigarettes were comprised of 100% or 0% of Burley tobacco having ZYVERT™ technology. Blended cigarette prototypes constructed with 2012 tobacco having ZYVERT™ technology comprised of Bright (35%), Burley (23%), Oriental (15%), and reconstituted tobacco sheet (27%). Blended cigarette prototypes constructed with 2013 tobacco having ZYVERT™ technology comprised of Bright (35%), Burley (23%), Oriental (12%), and reconstituted tobacco sheet (30%). All prototypes used the same ZYVERT™ tobacco produced in the corresponding year.
2.4. Data treatment and analysis 2.3. Tobacco filler and mainstream smoke chemical analyses Data analysis include tobacco and smoke chemistry results, and are presented separately for Burley single component and blended cigarette prototypes in the following section. Genetic stability results of nicotine conversion are also presented therein. Smoke HPHC are reported on a per cigarette basis for both ISO and HCI smoking conditions. Cut filler results are reported per gram of tobacco. The principal statistical results are reported and illustrated in the main body of the paper. Further results, tables, and data summaries are provided in the Appendix. For single component cigarette prototypes our main point of investigation was testing whether there was a significant difference in TSNA levels in cigarette smoke and tobacco filler as well as on some TSNA precursors in tobacco. To do so we used the Welch twosample t-test (Rupert G. Miller, 1998) for testing the zero mean difference between the two groups, with and without ZYVERT™ tobacco. Note that differences were tested for single component cigarette prototypes on a limited number of targeted constituents. Significance levels for testing group differences have been adjusted
Three ISO 17025 accredited laboratories generated the HPHC data in cut fillers and in cigarette mainstream smoke with ISO (ISO, 2000) and Health Canada Intense (HCI) (Canada, 1999) machine smoking regimes. Specifically, group 1 cigarette prototypes samples were submitted to Arista Laboratories (Richmond, VA), Enthalpy Analytical, Inc. (Durham, NC), and Labstat International ULC (Ontario, Canada). All analytical methods for the determination of the HPHCs were on the laboratories’ ISO scope of accreditation at the time of testing. To minimize analytical variability, the prototype products were submitted to a single laboratory, for a given method. Each laboratory conducted seven replicates for 2013 and five replicates for 2012 for all smoke and cigarette filler analyses except for nicotine and carbon monoxide, for which the laboratory conducted 20 replicates. Group 2 cigarette protypes were submitted to Labstat International ULC (Ontario, Canada) for HPHC analyses, according to the Health Canada methods, with the exception of the analyses of TSNA
Table 1 Cigarette construction and related information for the experimental prototypes. Cigarette Prototype Groups
Cigarette Prototype ZYVERT™ Sub-Groups Crop Year
Prototype Control LC Burley % ZYVERT™ Burley % Bright Tobacco Leaf Reconstituted Leaf Name of Inclusion of Inclusion % of Inclusion % of Inclusion
Oriental Tobacco Leaf % of Inclusion
Group 1
ALCS 2012a
12 ALCSB0 12 ALCSB8 12 ALCSB15 12 ALCSB23 13 ALCSB0 13 ALCSB8 13 ALCSB15 13 ALCSB23 13 PMIB0 13 PMIB8 13 PMIB15 13 PMIB23 13 PMIS0 13 PMIS100
2012 2012 2012 2012
Group 1
ALCS 2013b
2013 2013 2013 2013
Group 2
PMI 2013c
2013 2013 2013 2013 2013 2013
a b c
23 15 8
0 8 15
35 35 35
27 27 27
15 15 15
0
23
35
27
15
23 15 8
0 8 15
35 35 35
30 30 30
12 12 12
0
23
35
30
12
23 15 8
0 8 15
35 35 35
30 30 30
12 12 12
0
23
35
30
12
100 0
0 100
0 0
0 0
0 0
Blended cigarettes produced by ALCS in 2012. Blended cigarettes produced by ALCS in 2013. Blended or single component cigarettes produced by PMI in 2013.
104
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
Fig. 2. Percentage of nicotine conversion in individual plants of a population of 81 plants of KT 204 LC and 81 plants of KT 204 having ZYVERT™ technology. For illustrative purposes the bars are plotted on the logarithmic scale.
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
105
Table 2 Smoke and tobacco chemistry comparative results for single Burley component cigarette prototypes with and without ZYVERT™ tobacco. Smoke chemistry results are available only under HCI smoking conditions.
Smoke HCI
Tobacco filler
Constituent
unit
Estimated difference (%)
p-value
NNN NNK Carbon monoxide Nicotine NNN NNK NAB NAT Nicotine Nornicotine Anabasine Anatabine
[ng/cig] [ng/cig] [mg/cig] [mg/cig] [ng/g] [ng/g] [ng/g] [ng/g] [mg/g] [mg/g] [mg/g] [mg/g]
75.50% 24.90% 0.90% 18.20% 73.90% 30.10% 30.60% 33.20% 7.10% 73.20% 12.80% 14.90%
0.002 0.023 0.673 0.000 0.000 0.009 0.001 0.000 0.016 0.001 0.002 0.000
Note: The values for tobacco filler constituents are on dry weight basis.
Fig. 3. Tobacco filler constituent yields for blended cigarette prototypes with ZYVERT™ technology relative to the yields for blended cigarette prototypes without ZYVERT™ technology. Intervals correspond to 90% confidence interval on the estimated ratio. Results are reported by study (squares for ALCS, 2012; triangles for ALCS, 2013; and diamonds for PMI, 2013).
for simultaneously testing using a Bonferroni-type correction (Rupert G. Miller, 1981). For blended cigarette prototypes our goal was to control whether the two cigarettes (with or without ZYVERT™ tobacco for the Burley portion) provide similar smoke and tobacco chemistry profiles targeting a wider class of smoke constituents. This was better achieved by using an equivalence-type approach rather than formally testing for statistical differences. To do so, we plotted the 90% confidence intervals for the ratio of yields for the cigarette prototypes with and without ZYVERT™ tobacco in the Burley portion and checked whether it was included within the 0.8 and 1.25 lower and upper limits. The region within the 0.8 and 1.25 limits is commonly used as an equivalence zone; while the two limits are often considered as lower and upper limits for equivalence between a new and a reference product (FDA, 2001). The statistical analysis and all the provided figures were produced using TIBCO Spotfire Sþ® 8.2 for Windows.
3. Results 3.1. Genetic stability of nicotine conversion in experimental KT 204 having ZYVERT™ technology Based on the alkaloid data measured in a single leaf of 81 individual field grown plants, the average percent of nicotine conversion for the tobacco varieties KT 204 LC and KT 204 having ZYVERT™ technology plant populations was 5.32% and 0.43%, respectively (see Fig. 2). The percent of nicotine conversion for the commercial variety KT 204 LC ranged from 1.31% in the lower end to 86.72% in the upper end; see the lower panel in Fig. 2. For 72% of the KT 204 LC plants the percentage of nicotine conversion was below 3%, 26% had the nicotine conversion between 3% and 22%, and 2% of the plants had the nicotine conversion above 80%. For the experimental KT 204 having ZYVERT™ technology the range of nicotine conversion was from 0.32% to 0.63% with all 81 plants tested below 1% (Fig. 2).
106
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
Fig. 4. NNN levels in tobacco filler for increasing ZYVERT™ tobacco inclusion rates. From left to right: ALCS study 2012, ALCS study 2013, PMI study 2013.
3.2. Burley single component cigarette analyses 3.2.1. Cut filler results The quantity of all four TSNAs were significantly lower in the cut filler of ZYVERT™ cigarette prototypes made with 100% Burley tobacco as compared to non ZYVERT™ cigarette prototypes (see Table 2 and Tabe 3 in Supplementary material). Data indicate that NNN and its precursor were reduced by 73% while NNK, NAT and NAB were reduced around 30%. A 7% increase observed for nicotine was not statistically significant on the adjusted 5% family wise error rate. 3.2.2. Mainstream smoke HPHCs in cigarettes The yield of NNN in the mainstream smoke of Burley single component was significantly lower in the cigarettes made with tobacco having ZYVERT™ technology than in the cigarettes with LC tobacco, while the nicotine yield was significantly higher (18.2%). The content of NNK, while it has a p-value less than 5% (see Table 2) is not statistically significant after using a Bonferroni correction for multiple testing. Finally, for Carbon monoxide (CO), the observed increase (0.9%) was far from being statistically significant (pvalue ¼ 0.673).
substantial reduction of NNN and nornicotine by 31% and 36%, respectively. For the remaining tobacco filler constituents the results showed non consistent (across studies) and smaller in magnitude differences between the ZYVERT™ and the non ZYVERT™ blended cigarette prototypes. Only for NNK in the ALCS 2012 study the 90% interval upper limit was slightly above the 1.25 upper limits, most likely due to the high variability of NNK. Furthermore we computed and illustrated in Fig. 4 the NNN content in filler for increasing ZYVERT™ tobacco inclusion rate. Four levels of replacement were chosen to cover a range from 0% to 100% Burley tobacco replacement. These are: 0%, 33%, 66%, and 100% replacement of LC-Burley tobacco with ZYVERT™-Burley that correspond to 0, 0.33, 0.66, 1.0 ZYVERT™ fractional inclusion rates. Fig. 4 highlights the link between the NNN reduction and the increasing ZYVERT™ tobacco inclusion rate. Our results confirmed a significant decreasing NNN trend for increasing ZYVERT™ tobacco inclusion rate. When 100% of the Burley component in the blend (23% in the total blend) was replaced with tobacco having ZYVERT™ technology, around 30% reduction of NNN was observed in tobacco filler. All cut filler results for blended cigarette prototypes are reported in Table 3 in the Appendix, including averages and standard errors values for cigarette prototypes with and without tobacco having ZYVERT™ technology.
3.3. Blended cigarette analyses 3.3.1. Cut filler constituents’ results Cut filler constituents for blended cigarette prototypes with and without ZYVERT™ technology have been analyzed. Relative yields for blended cigarette prototypes with ZYVERT™ technology relative to prototypes without ZYVERT™ technology were quantified and are depicted in Fig. 3. The resulting ratios are plotted together with their 90% confidence intervals. Inspecting Fig. 3 highlights the
3.3.2. Smoke HPHCs results Mainstream smoke HPHC for blended cigarette prototypes were quantified under both ISO and HCI smoking conditions. For both smoking conditions the resulting data confirmed that NNN yields were strongly reduced for blended prototypes containing ZYVERT™ tobacco. This is illustrated in Fig. 5 and in Fig. 6 where the relative smoke constituents (abbreviated list of FDA) yields of the
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
107
Fig. 5. ISO smoke constituent yields for cigarette prototypes with ZYVERT™ technology relative to the yields for cigarette prototypes without ZYVERT™ technology. Intervals correspond to 90% confidence interval on the estimated ratio. Results are reported by study (squares for ALCS, 2012; triangles for ALCS, 2013; and diamonds for PMI, 2013) and concern the comparison of prototypes B23 relative to B0 for all three studies: ALCS 2012, ALCS 2013 and PMI 2013.
blended cigarette prototypes with and without tobacco having ZYVERT™ technology are plotted together with their 90% confidence intervals. The observed NNN reduction is more than 30% and is consistent across all three studies. Evaluation of the remaining HPHC showed relatively small and inconsistent (across studies) reductions or increases for blended prototypes containing ZYVERT™ tobacco relative to prototypes containing LC Burley. With the exception of crotonaldehyde, formaldehyde and ammonia, the 90% confidence intervals for all HPHC were fully included within the 0.8 and 1.25 limits. In contrast only the NNN confidence intervals were fully included in the region below the 0.8 lower limit, confirming the selective reduction of NNN. The comparative smoke
chemistry results including the ratio yields and their 90% confidence limits are also reported in the Appendix. The average smoke constituent yields together with their associated standard errors are given in Tables 4 and 5 in the Appendix for both ISO and HCI smoking conditions. The smoke chemistry results are reported by study, and for both cigarettes with and without ZYVERT™ technology. The smoke chemistry results presented so far focus on the statistical assessment of blended cigarette constructs with and without ZYVERT™ technology on the whole Burley portion of the blend. Additional blended cigarette prototypes have been included in the studies in order to assess the impact of ZYVERT™ tobacco
108
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
Fig. 6. HCI smoke constituent yields for cigarette prototypes with ZYVERT™ technology relative to the yields for cigarette prototypes without ZYVERT™ technology. Intervals correspond to 90% confidence interval on the estimated ratio. Results are reported by study (squares for ALCS, 2012; triangles for ALCS, 2013; and diamonds for PMI, 2013) and concern the comparison of prototypes B23 relative to B0 for all three studies: ALCS 2012, ALCS 2013 and PMI 2013.
inclusion for each portion of Burley tobacco being replaced by ZYVERT™ tobacco. Four levels of replacement have been chosen to cover a range from 0% to 100% burley tobacco replacement. These are 0%, 33%, 66%, and 100% replacement of Burley tobacco with ZYVERT™ technology that correspond to 0, 0.33, 0.66, and 1.0 ZYVERT™ fractional inclusion rates. Note that 100% of Burley tobacco replacement equals in terms of total blend to 23% of replacement, or equally to a 0.23 fraction of ZYVERT™ technology inclusion rate. The NNN yields with increasing ZYVERT™ tobacco inclusion rate are illustrated in Fig. 7. In all studies and for both smoking conditions the decreasing trend in NNN was confirmed.
4. Discussion Our experimental data indicated that the introgression of cyp82e4, cyp82e5 and cyp82e10 nonfunctional version of nicotine demethylase genes in tobacco varieties reduces the conversion of nicotine to nornicotine, the precursor of the NNN, to less than 1% compared to 3e5% conversion rates observed with low converter varieties. These results are in line with a previous publication showing that tobacco varieties with a mutant versions of these three genes reach stable levels of about 0.5% nornicotine relative to nicotine (Lewis et al., 2010). The 3e5% conversion rate observed for LC varieties is not uniform across the populations and it is not
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
109
Fig. 7. Smoke NNN yields for increasing ZYVERT™ tobacco inclusion rates. Upper panel: ISO smoking condition. Lower panel: HCI smoking conditions. From left to right: ALCS study 2012, ALCS study 2013, PMI study 2013.
uncommon to observe individual plants with conversion rates of 10% or more. These individual plants are called “converters” as compared to “non-converters” plants that have conversion rates of 3% or below. The sporadic transition of “non-converters” to “converters” can occur in up to 20% of the population per generation
(Gavilano et al., 2006). On the other hand data provided in this paper show that conversion of nicotine to nornicotine is stable in the population of varieties containing the ZYVERT™ technology, and in average is less than 1%. This lower conversion rate contributed to about a 70% reduction of NNN in the Burley tobaccos and
110
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
cigarette smoke of prototypes made from 100% Burley tobacco. Data derived from this study indicate that NNK, NAB and NAT are also decreased in Burley ZYVERT™ tobaccos. However, these decreases may not result from the presence of non-functional three mutant P450 genes and are most likely derived from the variability observed during tobacco growing, harvesting and curing. Variability of TSNA content in tobacco material is well described (Gunduz et al., 2016). Moreover in blend inclusion studies, NNK levels both in tobacco filler and in cigarette smoke of prototypes produced by PMI and ALCS with ZYVERT™ were substantially equivalent in comparison to the levels observed in cigarette prototypes containing LC Burley tobacco. An increase in nicotine levels in Zyvert Burley tobacco was not statistically significant in tobacco however significant in the yields of machine smoked prototypes. This difference potentially could arise due to the variability of cigarette making and machine smoking. When 100% of LC Burley component (23% of the blend) was replaced with ZYVERT™ Burley in blend inclusion studies, NNN was reduced by 30% both in tobacco filler and in mainstream smoke of the prototypes. This reduction is proportional to the replacement portion of Burley tobacco containing nonfunctional versions of three P450 demethylase genes. With the exception of NNN, the levels of the other FDA abbreviated list HPHCs were substantially equivalent between prototypes containing LC Burley tobacco and prototypes containing Burley tobacco containing the three nonfunctional demethylase genes. We therefore demonstrated that the use of the Burley tobacco having ZYVERT™ technology selectively reduced NNN in the tobacco filler and smoke. The selective reduction of smoke constituents in conventional cigarettes is very challenging since a decrease of one smoke constituent may be accompanied by an increase of other smoke con et al., 2013). For example the adoption of stituents (Piade recommended nitrogen usage for the fertilization that aims to reduce TSNA, has the potential to decrease the ammonia and nitrogen oxide in mainstream smoke, therefore causing an increase of formaldehyde (Baker, 2006; Paine et al., 2004; Glarborg et al., 2003). Conversely, the use of ZYVERT™ Burley tobaccos minimized the nicotine to nornicotine conversion without changing the nitrogen balance that could lead to an increase of formaldehyde and potentially other smoke constituents. Although selective reduction of smoke constituents in conventional cigarettes that burn tobacco is unlikely to reduce the risk of tobacco related diseases, the ZYVERT™ technology provides an opportunity to substantially reduce NNN and its precursor nornicotine in tobacco and cigarette smoke. In conclusion, the use of a non-functional version of three P450 genes reduces nornicotine, the precursor of NNN and consequently NNN in Burley tobacco by 70% compared to levels observed in currently available LC Burley varieties. The technology does not require a further screening process and is genetically stable over multiple generations. The selective reduction of NNN in cigarette smoke is proportional to the inclusion level of Burley tobacco having ZYVERT™ technology in the blend, with a 23% inclusion level resulting into a 30% selective reduction of NNN for a typical American blended cigarette. This reduction could even be higher if tobacco having ZYVERT™ technology would be used for replacement of the Burley component in reconstituted leaves and stem inclusions.
Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.yrtph.2017.07.019.
Transparency document Transparency document related to this article can be found online at http://dx.doi.org/10.1016/j.yrtph.2017.07.019. References Appleton, S., Olegario, R.M., Lipowicz, P.J., 2013. TSNA levels in machine-generated mainstream cigarette smoke: 35 years of data. Regul. Toxicol. Pharmacol. 66 (2), 197e207. http://dx.doi.org/10.1016/j.yrtph.2013.03.013. Baker, R.R., 2006. The generation of formaldehyde in cigarettes-Overview and recent experiments. Food Chem. Toxicol. 44 (11), 1799e1822. http://dx.doi.org/ 10.1016/j.fct.2006.05.017. Canada, Health, 1999. Method-T115, determination of tar, nicotine and carbon monoxide in mainstream tobacco smoke. In: SOR/200e273. Registration 200006-26. Canadian Ministry of Justice: Tobacco Reporting Regulations, Part 3: Emissions for Designated Tobacco Products. Canada, Health, 2000 (Accessed 29 August 2016). In: Health Canada (Ed.), Tobacco Reporting Regulations. Chakrabarti, M., Bowen, S.W., Coleman, N.P., Meekins, K.M., Dewey, R.E., Siminszky, B., 2008. CYP82E4-mediated nicotine to nornicotine conversion in tobacco is regulated by a senescence-specific signaling pathway. Plant Mol. Biol. 66 (4), 415e427. http://dx.doi.org/10.1007/s11103-007-9280-6. Cooperative effort of University of Kentucky, University of Tennessee, Virginia Tech and NC State University. Burley and Dark Tobacco Production Guide [cited June 8th, 2016]. Available from: https://tobacco.ces.ncsu.edu/wp-content/uploads/ 2015/04/2015-2016-Burley-Production-Guide.pdf?fwd¼no. Dewey, R.E., and R.S. Lewis., 2011. Compositions and methods for minimizing nornicotine synthesis in tobacco. Google Patents. FDA, 2001. Statistical Approaches to Establishing Bioequivalence. In: Guidelines for Industry: Food and Drug Administration. Center for Drug Evaluation and Research. FDA, 2012. In: Food and Drug Administration (Ed.), Harmful and Potentially Harmful Constituents in Tobacco Products and Tobacco Smoke. Gavilano, L.B., Coleman, N.P., Burnley, L.E., Bowman, M.L., Kalengamaliro, N.E., Hayes, A., Bush, L., Siminszky, B., 2006. Genetic engineering of Nicotiana tabacum for reduced nornicotine content. J. Agric. Food Chem. 54 (24), 9071e9078. http://dx.doi.org/10.1021/jf0610458. Glarborg, P., Alzueta, U., Kjaergaard, K., Dam-Johanse, K., 2003. Oxidation of formaldehyde and its interaction with nitric oxide in a flow reacto. Combust. Flame 132, 629e638. Gunduz, I., Kondylis, A., Jaccard, G., Renaud, J.M., Hofer, R., Ruffieux, L., Gadani, F., 2016. Tobacco-specific N-nitrosamines NNN and NNK levels in cigarette brands between 2000 and 2014. Regul. Toxicol. Pharmacol. 76, 113e120. http:// dx.doi.org/10.1016/j.yrtph.2016.01.012. IARC, 2007. Smokeless tobacco and some tobacco-specific N-Nitrosamines. In: International agency for research on cancer (Ed.), IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. World Health Organisation, Lyon, France. ISO, 2000. ISO-3308. Routine Analytical Cigarette-smoking Machine-definitions and Standard Conditions. Ineternational Organization for Standardization. Jack, Anne, 2007. The“LC” protocol. Available from: https://www.uky.edu/Ag/ Tobacco/Pdf/LC-Protocol.pdf. . Lang, G., Vuarnoz, A., 2015. Matrix-bound 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone in tobacco: quantification and evidence for an origin from ligninincorporated alkaloids. J. Nat. Prod. 78 (1), 85e92. http://dx.doi.org/10.1021/ np500725a. Lewis, R.S., Bowen, S.W., Keog, M.R., Dewey, R.E., 2010. Three nicotine demethylase genes mediate nornicotine biosynthesis in Nicotiana tabacum L.: functional characterization of the CYP82E10 gene. Phytochemistry 71, 1988e1998. Miller, B., Jack, A., Bush, L., 2004. Screening commercial burley tobacco cultivars for nornicotine conversion, Paper read at 41st Tobacco Workers Conference, at Nashville, TN. Moldoveanu, S.C., Borgerding, M., 2008. Formation of tobacco specific nitrosamines in mainstream cigarette smoke; Part 1, FTC smoking. Beitrage zur Tabakforschung Int. Contrib. Tob. Res. 23 (1), 19e31. NTP, 2011. NTP 12th Report on carcinogens. Rep. Carcinog. 12 iii-499. Paine, J.B., Pithawalla, Y.B., Hollins, S.R., Naworal, J., 2004. Influence of ammonia on carbohydrate pyrolysis chemistry: Co-pyrolysis of 13C-labeled D-glucose with ammonium bicarbonate. In: 6th PM-USA Symposium on Fundamental Science. Richmond, VA, USA. , J.J., Wajrock, S., Jaccard, G., Janeke, G., 2013. Formation of mainstream cigaPiade rette smoke constituents prioritized by the World Health Organization - yield patterns observed in market surveys, clustering and inverse correlations. Food Chem. Toxicol. 55, 329e347. http://dx.doi.org/10.1016/j.fct.2013.01.016. Miller Jr., Rupert G., 1981. Simultaneous Statistical Inference. Edited by 2nd Edition. Springer-Verlag, New York. Miller Jr., Rupert G., 1998. Beyond ANOVA. Chapman & Hall, New York. Siminszky, B., Gavilano, L., Bowen, S.W., Dewey, R.E., 2005a. Conversion of nicotine to nornicotine in Nicotiana tabacum is mediated by CYP82E4, a cytochrome P450 monooxygenase. Proced. Natl. Acad. Sci. U. S. A. 102 (41), 14919e14924. http://dx.doi.org/10.1073/pnas.0506581102. Siminszky, B., Gavilano, L., Bowen, S.W., Dewey, R.E., 2005b. Conversion of nicotine
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111 to nornicotine in Nicotiana tabacum is mediated by CYP82E4, a cytochrome P450 monooxygenase. Proc. Natl. Acad. Sci. U. S. A. 102 (41), 14919e14924. http://dx.doi.org/10.1073/pnas.0506581102. Sophia, F., Spiegelhalder, B., Preussmann, R., 1989. Preformed tobacco-specific nitrosamines in tobacco-role of nitrate and influence of tobacco type.
111
Carcinogenesis 10 (8), 1511e1517. http://dx.doi.org/10.1093/carcin/10.8.1511. Xu, D., Shen, Y., Chappell, J., Mingwu, C., Nielsen, M., 2007. Biochemical and molecular characterizations of nicotine demethylase in tobacco. Physiol. Plant. 129, 307e319.
Contents lists available at ScienceDirect
Regulatory Toxicology and Pharmacology journal homepage: www.elsevier.com/locate/yrtph
Novel approach for selective reduction of NNN in cigarette tobacco filler and mainstream smoke M. Lusso a, **, I. Gunduz b, *, A. Kondylis c, G. Jaccard c, L. Ruffieux b, F. Gadani d, K. Lion a, A. Adams a, W. Morris a, T. Danielson a, U. Warek a, J. Strickland a a
Altria Client Services LLC, Research Development & Regulatory Affairs, 601 E. Jackson St., Richmond, VA 23219, USA Philip Morris International Management SA, Leaf Agricultural Programs, Avenue de Rhodanie 50, 1001 Lausanne, Switzerland Philip Morris International R&D, Philip Morris Products S.A. (Part of Philip Morris International Group of Companies), Quai Jeanrenaud 5, 2000 Neuchatel, Switzerland d Philip Morris Products SA, Product Development, Rue des Usines 56, 2000 Neuchatel, Switzerland b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 May 2017 Received in revised form 17 July 2017 Accepted 19 July 2017 Available online 20 July 2017
Research conducted during past decades to reduce the level of the tobacco specific nitrosamine Nnitrosonornicotine (NNN) and its precursor nornicotine in tobacco yielded identification of three tobacco genes encoding for cytochrome P450 nicotine demethylases converting nicotine to nornicotine. We carried out trials to investigate the effect of using tobaccos containing three non-functional nicotine demethylase genes on the selective reduction of NNN in cigarette tobacco filler and mainstream smoke. Our results indicate that the presence of non-functional alleles of the three genes reduces the level of nornicotine and NNN in Burley tobacco by 70% compared to the level observed in currently available low converter (LC) Burley tobacco varieties. The new technology, named ZYVERT™, does not require a regular screening process, while a yearly selection process is needed to produce LC Burley tobacco seeds for NNN reduction. The reduction of NNN observed in smoke of blended prototype cigarettes is proportional to the inclusion level of tobacco having ZYVERT™ technology. Inclusion of Burley tobacco possessing the new trait into a typical American blend resulted in a selective reduction of NNN in cigarette smoke, while the levels of other Harmful and Potentially Harmful Constituents (HPHC) currently in the abbreviated list provided by the US Food and Drug Administration are statistically equivalent in comparison with the levels obtained in reference prototype cigarettes containing LC Burley. © 2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Keywords: NNN N-nitrosonornicotine Nornicotine Nicotine demethylase Cytochrome P450 TSNA Tobacco regulation Harmful and potentially harmful constituents HPHC Selective smoke constituents' reduction
1. Introduction Some tobacco chemicals and smoke constituents have been classified as Harmful and Potentially Harmful Constituents (HPHC) (FDA, 2012). Therefore, approaches should be used to reduce or eliminate human exposure. Tobacco-specific N-nitrosamines (TSNA) are a class of chemical compounds found in cured, unburned tobacco and tobacco smoke. Due to their potential toxicological relevance, public health researchers have focused on these compounds for many years. For example, the International Agency for Research on Cancer, Working Group on the Evaluation of
* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (M. Lusso), [email protected] (I. Gunduz).
Carcinogenic Risks to Humans (IARC) evaluated the available scientific evidence for four TSNAs e 4-(methylnitrosamino)-1-(3pyridyl)-1-butanone (NNK), N0 -nitrosonornicotine (NNN), N0 nitrosoanabasine (NAB) and N-nitrosoanatabine (NAT) in an effort to determine the carcinogenic potential of these compounds. IARC (IARC, 2007) classified NNK and NNN as Group 1 carcinogens (carcinogenic to humans) and NAB and NAT as Group 3 compounds (not classifiable as to their carcinogenicity to humans). The United States National Toxicology Program (USNTP) classified NNN and NNK “to be reasonably anticipated to be a human carcinogen” based on sufficient evidence of carcinogenicity from studies in experimental animals as described in the 12th Report on Carcinogens (NTP, 2011). In addition, the FDA included NNN and NNK on its HPHC list (FDA, 2012) to be reported in tobacco products and tobacco smoke. Nicotine, the predominant alkaloid found in commercial tobacco varieties, represents typically 90e95% of the total
http://dx.doi.org/10.1016/j.yrtph.2017.07.019 0273-2300/© 2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
102
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
alkaloids in cured leaf. The 5e10% remaining alkaloids are called minor alkaloids and are comprised primarily of nornicotine, anabasine and anatabine. These minor alkaloids are the direct precursors of TSNA in tobaccos through their nitrosation during the plant life and especially during curing process, i.e., the leaf drying after harvest (Sophia et al., 1989). NNN is formed essentially via nitrosation of nornicotine, while NNK is formed from nitrosation of et al., nicotine or its oxidized products (Sophia et al., 1989; Piade 2013). The presence of TSNAs in mainstream smoke is essentially due to a direct transfer from tobacco filler to mainstream smoke. There is also a certain proportion (5e25% of the total amount) of NNN released by pyrosynthesis, strongly dependent on nornicotine level (Moldoveanu and Borgerding, 2008). For NNK, both pyrosynthesis and pyrorelease occur, due to a large proportion of NNK present in matrix-bound form in tobaccos (Lang and Vuarnoz, 2015). Nornicotine, the main direct precursor of NNN, is generated directly from nicotine by oxidative demethylation through the activity of N-demethylase enzymes (Siminszky et al., 2005a; Gavilano et al., 2006; Xu et al., 2007). Although nornicotine typically represents less than 5% of the total alkaloids, tobacco plants that initially produce very low amounts of nornicotine give rise to progeny that metabolically “convert” large percentages of leaf nicotine to nornicotine due to the genetic expression of nicotine demethylase enzymes. This process is called “conversion” which predominantly occurs during senescence and curing of the mature leaf (Chakrabarti et al., 2008). Among predominantly used tobacco types Burley tobaccos are particularly prone to nicotine to nornicotine conversion, with rates up to 90% observed in the progeny of some cultivars grown over multiple generations. It is understood that regardless of signals involved in “turning on” the genes, nicotine demethylase enzymes appear to be predominantly involved in nornicotine biosynthesis (Gavilano et al., 2006). Flue-cured tobacco is generally considered more stable in terms of nicotine conversion, although some nornicotine production does occur. Miller et al. (2004) discovered that levels of nornicotine present in a given Burley tobacco variety can be reduced up to 70e80% by identifying and eliminating individual converter plants prior to seed collection. This “roguing” process is accomplished by chemically analyzing seed plants as they are grown in the field and destroying all plants that have levels of nornicotine greater than 3% relative to nicotine content prior to seed harvest. This screening process results in the production of seeds giving a low converter (LC) phenotype, commonly referred to as a “LC variety.” While LC varieties substantially decrease the nicotine to nornicotine conversion, they remain the same as the original variety with respect to yield, quality, agronomic character and disease resistance. The use of LC varieties for the production of Burley tobacco has become the standard in the United States of America and many other parts of the world and has contributed together with other good agricultural and curing practices to downward trends in terms of NNN levels in tobaccos and commercial cigarettes (Gunduz et al., 2016; Appleton et al., 2013). More recent research on the mechanism of nornicotine formation led to the identification of three nicotine demethylase genes, CYP82E4 (Siminszky et al., 2005b), CYP82E5 (Dewey and Lewis, 2011) and CYP82E10 (Lewis et al., 2010) that facilitate conversion of nicotine to nornicotine in leaf. The genes designated CYP82E4 (E4) and CYP82E5 (E5) are expressed in leaf tissue of the plant and the gene designated CYP82E10 (E10) is expressed in the roots. Formation of NNN in tobacco is provided in Fig. 1. Tobacco varieties with non-functional alleles of these three genes show decrease in conversion to stable levels of about 0.5% nornicotine relative to nicotine (Lewis et al., 2010) and a
Fig. 1. Formation of NNN in tobacco Nicotine demetylation is catalyzed by a group of cytochrome P450 enzymes of the 82E family. The demethylation leads to the formation of nornicotine. Nornicotine itself is nitrosated in N0 -nitrosonornicotine (NNN) through a non enzymatic reaction occurring at low rates during the plant life and intensely during the leaf drying period.
commensurate reduction in NNN content of the leaf. Mutation of these three P450 functional genes resulting in non-functional versions is commercially referred to as tobacco having ZYVERT™ technology, and the tobacco varieties using the technology are called ZYVERT™ tobaccos, hereafter. Major Burley commercial varieties containing three non-functional versions of these genes met minimum standards and were approved in 2013, 2014 and 2015 by the Regional Burley Variety Evaluation program conducted cooperatively by Universities in Kentucky, North Carolina, Tennessee, and Virginia. While ZYVERT™ technology is being integrated to commonly grown tobacco varieties, the impact of tobaccos containing this technology on cigarette smoke chemistry is not known. Therefore the objective of this study is to evaluate the impact of tobaccos containing ZYVERT™ in American blended cigarettes on the chemical composition of tobacco cut-filler and mainstream smoke. Data presented in this study include the percentage of nicotine conversion in low converter and ZYVERT™ plant populations, and the comparative level of chemical constituents identified in the list of HPHC (FDA, 2012) in tobacco products and tobacco smoke by the US FDA (abbreviated list). 2. Methods 2.1. Tobacco production Tobaccos used in the experiments described in this paper were produced and cured according to the recommended practices for Burley tobacco (Cooperative effort of University of Kentucky). Seeds from the experimental ZYVERT™ Burley tobacco varieties were received from the tobacco breeding program at North Carolina State University, Raleigh, NC except for the KT 204 ZYVERT™ experimental variety which was produced by Altria Client Services (ALCS), Richmond, VA. To evaluate the genetic stability of the ZYVERT™ technology, the experimental Burley tobacco variety KT 204 having ZYVERT™ technology and its low converter version (KT 204 LC) were grown in Virginia in 2013. Eighty-one plants from each variety were individually sampled and processed according to the low-converter protocol (Jack, 2007) and submitted for chemical analysis. To assess the impact of the ZYVERT™ technology on the chemical composition of cigarette filler and smoke, experimental Burley tobacco varieties containing the ZYVERT™ technology and commercially available tobacco varieties were grown in Virginia and Kentucky, respectively, during the 2012 and 2013 crop years. Cured leaves were de-stemmed, dried, cut and used for making machine made cigarettes prototypes. 2.2. Prototype cigarette construction Experimental cigarettes were prepared with components (cellulose acetate filters, papers, and adhesives) and designs consistent
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
103
in filler and smoke, which were performed by a liquid chromatography-tandem mass spectrometry method, according to Labstat internal methods T-309 and T-311 respectively (Canada, 2000). For HPHC filler analyses of group 2 prototypes, it was requested that nicotine, NNN and NNK analysis be conducted with three replicates. For cigarette smoke of group 2 prototypes, it was requested seven replicates, except for nicotine and carbon monoxide, in which eight replicates were requested. Two different prototypes were manufactured in group 2, all Burley cigarette models, which were analyzed for HPHCs in cigarette smoke with three replicates and HCI smoking conditions only. All the conducted studies quantified NNN, NNK, and nicotine both in tobacco filler and in mainstream smoke. ISO and HCI results are available for mainstream smoke in blended cigarettes, while only HCI results are available for mainstream smoke in Burley single component cigarette prototypes. NAB, NAT and minor alkaloids were measured in the prototypes cut fillers of group 2 prototypes only.
with commercial US cigarette manufacturing. Cigarettes measured 84 mm in length (57 mm tobacco rod, 27 mm filter) and 25 mm in circumference. The prototype cigarettes were produced at the same time and the same constant weight at Altria Client Services (ALCS) (Group 1), or Philip Morris International (PMI) (Group 2); see Table 1. One sub-group of cigarette prototypes was made as a single component containing only Burley tobacco, while the three other sub-groups were designed as typical American blended cigarettes. Single component Burley cigarettes were comprised of 100% or 0% of Burley tobacco having ZYVERT™ technology. Blended cigarette prototypes constructed with 2012 tobacco having ZYVERT™ technology comprised of Bright (35%), Burley (23%), Oriental (15%), and reconstituted tobacco sheet (27%). Blended cigarette prototypes constructed with 2013 tobacco having ZYVERT™ technology comprised of Bright (35%), Burley (23%), Oriental (12%), and reconstituted tobacco sheet (30%). All prototypes used the same ZYVERT™ tobacco produced in the corresponding year.
2.4. Data treatment and analysis 2.3. Tobacco filler and mainstream smoke chemical analyses Data analysis include tobacco and smoke chemistry results, and are presented separately for Burley single component and blended cigarette prototypes in the following section. Genetic stability results of nicotine conversion are also presented therein. Smoke HPHC are reported on a per cigarette basis for both ISO and HCI smoking conditions. Cut filler results are reported per gram of tobacco. The principal statistical results are reported and illustrated in the main body of the paper. Further results, tables, and data summaries are provided in the Appendix. For single component cigarette prototypes our main point of investigation was testing whether there was a significant difference in TSNA levels in cigarette smoke and tobacco filler as well as on some TSNA precursors in tobacco. To do so we used the Welch twosample t-test (Rupert G. Miller, 1998) for testing the zero mean difference between the two groups, with and without ZYVERT™ tobacco. Note that differences were tested for single component cigarette prototypes on a limited number of targeted constituents. Significance levels for testing group differences have been adjusted
Three ISO 17025 accredited laboratories generated the HPHC data in cut fillers and in cigarette mainstream smoke with ISO (ISO, 2000) and Health Canada Intense (HCI) (Canada, 1999) machine smoking regimes. Specifically, group 1 cigarette prototypes samples were submitted to Arista Laboratories (Richmond, VA), Enthalpy Analytical, Inc. (Durham, NC), and Labstat International ULC (Ontario, Canada). All analytical methods for the determination of the HPHCs were on the laboratories’ ISO scope of accreditation at the time of testing. To minimize analytical variability, the prototype products were submitted to a single laboratory, for a given method. Each laboratory conducted seven replicates for 2013 and five replicates for 2012 for all smoke and cigarette filler analyses except for nicotine and carbon monoxide, for which the laboratory conducted 20 replicates. Group 2 cigarette protypes were submitted to Labstat International ULC (Ontario, Canada) for HPHC analyses, according to the Health Canada methods, with the exception of the analyses of TSNA
Table 1 Cigarette construction and related information for the experimental prototypes. Cigarette Prototype Groups
Cigarette Prototype ZYVERT™ Sub-Groups Crop Year
Prototype Control LC Burley % ZYVERT™ Burley % Bright Tobacco Leaf Reconstituted Leaf Name of Inclusion of Inclusion % of Inclusion % of Inclusion
Oriental Tobacco Leaf % of Inclusion
Group 1
ALCS 2012a
12 ALCSB0 12 ALCSB8 12 ALCSB15 12 ALCSB23 13 ALCSB0 13 ALCSB8 13 ALCSB15 13 ALCSB23 13 PMIB0 13 PMIB8 13 PMIB15 13 PMIB23 13 PMIS0 13 PMIS100
2012 2012 2012 2012
Group 1
ALCS 2013b
2013 2013 2013 2013
Group 2
PMI 2013c
2013 2013 2013 2013 2013 2013
a b c
23 15 8
0 8 15
35 35 35
27 27 27
15 15 15
0
23
35
27
15
23 15 8
0 8 15
35 35 35
30 30 30
12 12 12
0
23
35
30
12
23 15 8
0 8 15
35 35 35
30 30 30
12 12 12
0
23
35
30
12
100 0
0 100
0 0
0 0
0 0
Blended cigarettes produced by ALCS in 2012. Blended cigarettes produced by ALCS in 2013. Blended or single component cigarettes produced by PMI in 2013.
104
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
Fig. 2. Percentage of nicotine conversion in individual plants of a population of 81 plants of KT 204 LC and 81 plants of KT 204 having ZYVERT™ technology. For illustrative purposes the bars are plotted on the logarithmic scale.
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
105
Table 2 Smoke and tobacco chemistry comparative results for single Burley component cigarette prototypes with and without ZYVERT™ tobacco. Smoke chemistry results are available only under HCI smoking conditions.
Smoke HCI
Tobacco filler
Constituent
unit
Estimated difference (%)
p-value
NNN NNK Carbon monoxide Nicotine NNN NNK NAB NAT Nicotine Nornicotine Anabasine Anatabine
[ng/cig] [ng/cig] [mg/cig] [mg/cig] [ng/g] [ng/g] [ng/g] [ng/g] [mg/g] [mg/g] [mg/g] [mg/g]
75.50% 24.90% 0.90% 18.20% 73.90% 30.10% 30.60% 33.20% 7.10% 73.20% 12.80% 14.90%
0.002 0.023 0.673 0.000 0.000 0.009 0.001 0.000 0.016 0.001 0.002 0.000
Note: The values for tobacco filler constituents are on dry weight basis.
Fig. 3. Tobacco filler constituent yields for blended cigarette prototypes with ZYVERT™ technology relative to the yields for blended cigarette prototypes without ZYVERT™ technology. Intervals correspond to 90% confidence interval on the estimated ratio. Results are reported by study (squares for ALCS, 2012; triangles for ALCS, 2013; and diamonds for PMI, 2013).
for simultaneously testing using a Bonferroni-type correction (Rupert G. Miller, 1981). For blended cigarette prototypes our goal was to control whether the two cigarettes (with or without ZYVERT™ tobacco for the Burley portion) provide similar smoke and tobacco chemistry profiles targeting a wider class of smoke constituents. This was better achieved by using an equivalence-type approach rather than formally testing for statistical differences. To do so, we plotted the 90% confidence intervals for the ratio of yields for the cigarette prototypes with and without ZYVERT™ tobacco in the Burley portion and checked whether it was included within the 0.8 and 1.25 lower and upper limits. The region within the 0.8 and 1.25 limits is commonly used as an equivalence zone; while the two limits are often considered as lower and upper limits for equivalence between a new and a reference product (FDA, 2001). The statistical analysis and all the provided figures were produced using TIBCO Spotfire Sþ® 8.2 for Windows.
3. Results 3.1. Genetic stability of nicotine conversion in experimental KT 204 having ZYVERT™ technology Based on the alkaloid data measured in a single leaf of 81 individual field grown plants, the average percent of nicotine conversion for the tobacco varieties KT 204 LC and KT 204 having ZYVERT™ technology plant populations was 5.32% and 0.43%, respectively (see Fig. 2). The percent of nicotine conversion for the commercial variety KT 204 LC ranged from 1.31% in the lower end to 86.72% in the upper end; see the lower panel in Fig. 2. For 72% of the KT 204 LC plants the percentage of nicotine conversion was below 3%, 26% had the nicotine conversion between 3% and 22%, and 2% of the plants had the nicotine conversion above 80%. For the experimental KT 204 having ZYVERT™ technology the range of nicotine conversion was from 0.32% to 0.63% with all 81 plants tested below 1% (Fig. 2).
106
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
Fig. 4. NNN levels in tobacco filler for increasing ZYVERT™ tobacco inclusion rates. From left to right: ALCS study 2012, ALCS study 2013, PMI study 2013.
3.2. Burley single component cigarette analyses 3.2.1. Cut filler results The quantity of all four TSNAs were significantly lower in the cut filler of ZYVERT™ cigarette prototypes made with 100% Burley tobacco as compared to non ZYVERT™ cigarette prototypes (see Table 2 and Tabe 3 in Supplementary material). Data indicate that NNN and its precursor were reduced by 73% while NNK, NAT and NAB were reduced around 30%. A 7% increase observed for nicotine was not statistically significant on the adjusted 5% family wise error rate. 3.2.2. Mainstream smoke HPHCs in cigarettes The yield of NNN in the mainstream smoke of Burley single component was significantly lower in the cigarettes made with tobacco having ZYVERT™ technology than in the cigarettes with LC tobacco, while the nicotine yield was significantly higher (18.2%). The content of NNK, while it has a p-value less than 5% (see Table 2) is not statistically significant after using a Bonferroni correction for multiple testing. Finally, for Carbon monoxide (CO), the observed increase (0.9%) was far from being statistically significant (pvalue ¼ 0.673).
substantial reduction of NNN and nornicotine by 31% and 36%, respectively. For the remaining tobacco filler constituents the results showed non consistent (across studies) and smaller in magnitude differences between the ZYVERT™ and the non ZYVERT™ blended cigarette prototypes. Only for NNK in the ALCS 2012 study the 90% interval upper limit was slightly above the 1.25 upper limits, most likely due to the high variability of NNK. Furthermore we computed and illustrated in Fig. 4 the NNN content in filler for increasing ZYVERT™ tobacco inclusion rate. Four levels of replacement were chosen to cover a range from 0% to 100% Burley tobacco replacement. These are: 0%, 33%, 66%, and 100% replacement of LC-Burley tobacco with ZYVERT™-Burley that correspond to 0, 0.33, 0.66, 1.0 ZYVERT™ fractional inclusion rates. Fig. 4 highlights the link between the NNN reduction and the increasing ZYVERT™ tobacco inclusion rate. Our results confirmed a significant decreasing NNN trend for increasing ZYVERT™ tobacco inclusion rate. When 100% of the Burley component in the blend (23% in the total blend) was replaced with tobacco having ZYVERT™ technology, around 30% reduction of NNN was observed in tobacco filler. All cut filler results for blended cigarette prototypes are reported in Table 3 in the Appendix, including averages and standard errors values for cigarette prototypes with and without tobacco having ZYVERT™ technology.
3.3. Blended cigarette analyses 3.3.1. Cut filler constituents’ results Cut filler constituents for blended cigarette prototypes with and without ZYVERT™ technology have been analyzed. Relative yields for blended cigarette prototypes with ZYVERT™ technology relative to prototypes without ZYVERT™ technology were quantified and are depicted in Fig. 3. The resulting ratios are plotted together with their 90% confidence intervals. Inspecting Fig. 3 highlights the
3.3.2. Smoke HPHCs results Mainstream smoke HPHC for blended cigarette prototypes were quantified under both ISO and HCI smoking conditions. For both smoking conditions the resulting data confirmed that NNN yields were strongly reduced for blended prototypes containing ZYVERT™ tobacco. This is illustrated in Fig. 5 and in Fig. 6 where the relative smoke constituents (abbreviated list of FDA) yields of the
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
107
Fig. 5. ISO smoke constituent yields for cigarette prototypes with ZYVERT™ technology relative to the yields for cigarette prototypes without ZYVERT™ technology. Intervals correspond to 90% confidence interval on the estimated ratio. Results are reported by study (squares for ALCS, 2012; triangles for ALCS, 2013; and diamonds for PMI, 2013) and concern the comparison of prototypes B23 relative to B0 for all three studies: ALCS 2012, ALCS 2013 and PMI 2013.
blended cigarette prototypes with and without tobacco having ZYVERT™ technology are plotted together with their 90% confidence intervals. The observed NNN reduction is more than 30% and is consistent across all three studies. Evaluation of the remaining HPHC showed relatively small and inconsistent (across studies) reductions or increases for blended prototypes containing ZYVERT™ tobacco relative to prototypes containing LC Burley. With the exception of crotonaldehyde, formaldehyde and ammonia, the 90% confidence intervals for all HPHC were fully included within the 0.8 and 1.25 limits. In contrast only the NNN confidence intervals were fully included in the region below the 0.8 lower limit, confirming the selective reduction of NNN. The comparative smoke
chemistry results including the ratio yields and their 90% confidence limits are also reported in the Appendix. The average smoke constituent yields together with their associated standard errors are given in Tables 4 and 5 in the Appendix for both ISO and HCI smoking conditions. The smoke chemistry results are reported by study, and for both cigarettes with and without ZYVERT™ technology. The smoke chemistry results presented so far focus on the statistical assessment of blended cigarette constructs with and without ZYVERT™ technology on the whole Burley portion of the blend. Additional blended cigarette prototypes have been included in the studies in order to assess the impact of ZYVERT™ tobacco
108
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
Fig. 6. HCI smoke constituent yields for cigarette prototypes with ZYVERT™ technology relative to the yields for cigarette prototypes without ZYVERT™ technology. Intervals correspond to 90% confidence interval on the estimated ratio. Results are reported by study (squares for ALCS, 2012; triangles for ALCS, 2013; and diamonds for PMI, 2013) and concern the comparison of prototypes B23 relative to B0 for all three studies: ALCS 2012, ALCS 2013 and PMI 2013.
inclusion for each portion of Burley tobacco being replaced by ZYVERT™ tobacco. Four levels of replacement have been chosen to cover a range from 0% to 100% burley tobacco replacement. These are 0%, 33%, 66%, and 100% replacement of Burley tobacco with ZYVERT™ technology that correspond to 0, 0.33, 0.66, and 1.0 ZYVERT™ fractional inclusion rates. Note that 100% of Burley tobacco replacement equals in terms of total blend to 23% of replacement, or equally to a 0.23 fraction of ZYVERT™ technology inclusion rate. The NNN yields with increasing ZYVERT™ tobacco inclusion rate are illustrated in Fig. 7. In all studies and for both smoking conditions the decreasing trend in NNN was confirmed.
4. Discussion Our experimental data indicated that the introgression of cyp82e4, cyp82e5 and cyp82e10 nonfunctional version of nicotine demethylase genes in tobacco varieties reduces the conversion of nicotine to nornicotine, the precursor of the NNN, to less than 1% compared to 3e5% conversion rates observed with low converter varieties. These results are in line with a previous publication showing that tobacco varieties with a mutant versions of these three genes reach stable levels of about 0.5% nornicotine relative to nicotine (Lewis et al., 2010). The 3e5% conversion rate observed for LC varieties is not uniform across the populations and it is not
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
109
Fig. 7. Smoke NNN yields for increasing ZYVERT™ tobacco inclusion rates. Upper panel: ISO smoking condition. Lower panel: HCI smoking conditions. From left to right: ALCS study 2012, ALCS study 2013, PMI study 2013.
uncommon to observe individual plants with conversion rates of 10% or more. These individual plants are called “converters” as compared to “non-converters” plants that have conversion rates of 3% or below. The sporadic transition of “non-converters” to “converters” can occur in up to 20% of the population per generation
(Gavilano et al., 2006). On the other hand data provided in this paper show that conversion of nicotine to nornicotine is stable in the population of varieties containing the ZYVERT™ technology, and in average is less than 1%. This lower conversion rate contributed to about a 70% reduction of NNN in the Burley tobaccos and
110
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111
cigarette smoke of prototypes made from 100% Burley tobacco. Data derived from this study indicate that NNK, NAB and NAT are also decreased in Burley ZYVERT™ tobaccos. However, these decreases may not result from the presence of non-functional three mutant P450 genes and are most likely derived from the variability observed during tobacco growing, harvesting and curing. Variability of TSNA content in tobacco material is well described (Gunduz et al., 2016). Moreover in blend inclusion studies, NNK levels both in tobacco filler and in cigarette smoke of prototypes produced by PMI and ALCS with ZYVERT™ were substantially equivalent in comparison to the levels observed in cigarette prototypes containing LC Burley tobacco. An increase in nicotine levels in Zyvert Burley tobacco was not statistically significant in tobacco however significant in the yields of machine smoked prototypes. This difference potentially could arise due to the variability of cigarette making and machine smoking. When 100% of LC Burley component (23% of the blend) was replaced with ZYVERT™ Burley in blend inclusion studies, NNN was reduced by 30% both in tobacco filler and in mainstream smoke of the prototypes. This reduction is proportional to the replacement portion of Burley tobacco containing nonfunctional versions of three P450 demethylase genes. With the exception of NNN, the levels of the other FDA abbreviated list HPHCs were substantially equivalent between prototypes containing LC Burley tobacco and prototypes containing Burley tobacco containing the three nonfunctional demethylase genes. We therefore demonstrated that the use of the Burley tobacco having ZYVERT™ technology selectively reduced NNN in the tobacco filler and smoke. The selective reduction of smoke constituents in conventional cigarettes is very challenging since a decrease of one smoke constituent may be accompanied by an increase of other smoke con et al., 2013). For example the adoption of stituents (Piade recommended nitrogen usage for the fertilization that aims to reduce TSNA, has the potential to decrease the ammonia and nitrogen oxide in mainstream smoke, therefore causing an increase of formaldehyde (Baker, 2006; Paine et al., 2004; Glarborg et al., 2003). Conversely, the use of ZYVERT™ Burley tobaccos minimized the nicotine to nornicotine conversion without changing the nitrogen balance that could lead to an increase of formaldehyde and potentially other smoke constituents. Although selective reduction of smoke constituents in conventional cigarettes that burn tobacco is unlikely to reduce the risk of tobacco related diseases, the ZYVERT™ technology provides an opportunity to substantially reduce NNN and its precursor nornicotine in tobacco and cigarette smoke. In conclusion, the use of a non-functional version of three P450 genes reduces nornicotine, the precursor of NNN and consequently NNN in Burley tobacco by 70% compared to levels observed in currently available LC Burley varieties. The technology does not require a further screening process and is genetically stable over multiple generations. The selective reduction of NNN in cigarette smoke is proportional to the inclusion level of Burley tobacco having ZYVERT™ technology in the blend, with a 23% inclusion level resulting into a 30% selective reduction of NNN for a typical American blended cigarette. This reduction could even be higher if tobacco having ZYVERT™ technology would be used for replacement of the Burley component in reconstituted leaves and stem inclusions.
Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.yrtph.2017.07.019.
Transparency document Transparency document related to this article can be found online at http://dx.doi.org/10.1016/j.yrtph.2017.07.019. References Appleton, S., Olegario, R.M., Lipowicz, P.J., 2013. TSNA levels in machine-generated mainstream cigarette smoke: 35 years of data. Regul. Toxicol. Pharmacol. 66 (2), 197e207. http://dx.doi.org/10.1016/j.yrtph.2013.03.013. Baker, R.R., 2006. The generation of formaldehyde in cigarettes-Overview and recent experiments. Food Chem. Toxicol. 44 (11), 1799e1822. http://dx.doi.org/ 10.1016/j.fct.2006.05.017. Canada, Health, 1999. Method-T115, determination of tar, nicotine and carbon monoxide in mainstream tobacco smoke. In: SOR/200e273. Registration 200006-26. Canadian Ministry of Justice: Tobacco Reporting Regulations, Part 3: Emissions for Designated Tobacco Products. Canada, Health, 2000 (Accessed 29 August 2016). In: Health Canada (Ed.), Tobacco Reporting Regulations. Chakrabarti, M., Bowen, S.W., Coleman, N.P., Meekins, K.M., Dewey, R.E., Siminszky, B., 2008. CYP82E4-mediated nicotine to nornicotine conversion in tobacco is regulated by a senescence-specific signaling pathway. Plant Mol. Biol. 66 (4), 415e427. http://dx.doi.org/10.1007/s11103-007-9280-6. Cooperative effort of University of Kentucky, University of Tennessee, Virginia Tech and NC State University. Burley and Dark Tobacco Production Guide [cited June 8th, 2016]. Available from: https://tobacco.ces.ncsu.edu/wp-content/uploads/ 2015/04/2015-2016-Burley-Production-Guide.pdf?fwd¼no. Dewey, R.E., and R.S. Lewis., 2011. Compositions and methods for minimizing nornicotine synthesis in tobacco. Google Patents. FDA, 2001. Statistical Approaches to Establishing Bioequivalence. In: Guidelines for Industry: Food and Drug Administration. Center for Drug Evaluation and Research. FDA, 2012. In: Food and Drug Administration (Ed.), Harmful and Potentially Harmful Constituents in Tobacco Products and Tobacco Smoke. Gavilano, L.B., Coleman, N.P., Burnley, L.E., Bowman, M.L., Kalengamaliro, N.E., Hayes, A., Bush, L., Siminszky, B., 2006. Genetic engineering of Nicotiana tabacum for reduced nornicotine content. J. Agric. Food Chem. 54 (24), 9071e9078. http://dx.doi.org/10.1021/jf0610458. Glarborg, P., Alzueta, U., Kjaergaard, K., Dam-Johanse, K., 2003. Oxidation of formaldehyde and its interaction with nitric oxide in a flow reacto. Combust. Flame 132, 629e638. Gunduz, I., Kondylis, A., Jaccard, G., Renaud, J.M., Hofer, R., Ruffieux, L., Gadani, F., 2016. Tobacco-specific N-nitrosamines NNN and NNK levels in cigarette brands between 2000 and 2014. Regul. Toxicol. Pharmacol. 76, 113e120. http:// dx.doi.org/10.1016/j.yrtph.2016.01.012. IARC, 2007. Smokeless tobacco and some tobacco-specific N-Nitrosamines. In: International agency for research on cancer (Ed.), IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. World Health Organisation, Lyon, France. ISO, 2000. ISO-3308. Routine Analytical Cigarette-smoking Machine-definitions and Standard Conditions. Ineternational Organization for Standardization. Jack, Anne, 2007. The“LC” protocol. Available from: https://www.uky.edu/Ag/ Tobacco/Pdf/LC-Protocol.pdf. . Lang, G., Vuarnoz, A., 2015. Matrix-bound 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone in tobacco: quantification and evidence for an origin from ligninincorporated alkaloids. J. Nat. Prod. 78 (1), 85e92. http://dx.doi.org/10.1021/ np500725a. Lewis, R.S., Bowen, S.W., Keog, M.R., Dewey, R.E., 2010. Three nicotine demethylase genes mediate nornicotine biosynthesis in Nicotiana tabacum L.: functional characterization of the CYP82E10 gene. Phytochemistry 71, 1988e1998. Miller, B., Jack, A., Bush, L., 2004. Screening commercial burley tobacco cultivars for nornicotine conversion, Paper read at 41st Tobacco Workers Conference, at Nashville, TN. Moldoveanu, S.C., Borgerding, M., 2008. Formation of tobacco specific nitrosamines in mainstream cigarette smoke; Part 1, FTC smoking. Beitrage zur Tabakforschung Int. Contrib. Tob. Res. 23 (1), 19e31. NTP, 2011. NTP 12th Report on carcinogens. Rep. Carcinog. 12 iii-499. Paine, J.B., Pithawalla, Y.B., Hollins, S.R., Naworal, J., 2004. Influence of ammonia on carbohydrate pyrolysis chemistry: Co-pyrolysis of 13C-labeled D-glucose with ammonium bicarbonate. In: 6th PM-USA Symposium on Fundamental Science. Richmond, VA, USA. , J.J., Wajrock, S., Jaccard, G., Janeke, G., 2013. Formation of mainstream cigaPiade rette smoke constituents prioritized by the World Health Organization - yield patterns observed in market surveys, clustering and inverse correlations. Food Chem. Toxicol. 55, 329e347. http://dx.doi.org/10.1016/j.fct.2013.01.016. Miller Jr., Rupert G., 1981. Simultaneous Statistical Inference. Edited by 2nd Edition. Springer-Verlag, New York. Miller Jr., Rupert G., 1998. Beyond ANOVA. Chapman & Hall, New York. Siminszky, B., Gavilano, L., Bowen, S.W., Dewey, R.E., 2005a. Conversion of nicotine to nornicotine in Nicotiana tabacum is mediated by CYP82E4, a cytochrome P450 monooxygenase. Proced. Natl. Acad. Sci. U. S. A. 102 (41), 14919e14924. http://dx.doi.org/10.1073/pnas.0506581102. Siminszky, B., Gavilano, L., Bowen, S.W., Dewey, R.E., 2005b. Conversion of nicotine
M. Lusso et al. / Regulatory Toxicology and Pharmacology 89 (2017) 101e111 to nornicotine in Nicotiana tabacum is mediated by CYP82E4, a cytochrome P450 monooxygenase. Proc. Natl. Acad. Sci. U. S. A. 102 (41), 14919e14924. http://dx.doi.org/10.1073/pnas.0506581102. Sophia, F., Spiegelhalder, B., Preussmann, R., 1989. Preformed tobacco-specific nitrosamines in tobacco-role of nitrate and influence of tobacco type.
111
Carcinogenesis 10 (8), 1511e1517. http://dx.doi.org/10.1093/carcin/10.8.1511. Xu, D., Shen, Y., Chappell, J., Mingwu, C., Nielsen, M., 2007. Biochemical and molecular characterizations of nicotine demethylase in tobacco. Physiol. Plant. 129, 307e319.