Tobacco Documents Online

Accession number 632 Lorillard Research Center . Greensboro THE DETERMINATION OF THE EFFECTS OF VARIOUS ADDITIVES ON THE FORMATION OF PYRAZINE AND ALKYL-PYRAZINES ON ROASTED TOBACCO Submitted by: Tom M. Larson : Report number: Date: iof2/7s Summary or Abstract: ; A series of ten ex.~eriments was run, according to the ,Plackett-Burman Design using six variables, to determine .which of the tested variables contribute to the formation of pyrazine, methylpyrazine, and dimethylpyrazine. Results indicate that under these experimental conditions, the formation of pyrazine is the least specific reaction and the formation of dimethylpyrazine is the most specific. These data may be extrapolated to the conclusion that the formation of multisubstituted pyrazine compounds is more dependent upon. specific dicarbonyl compounds on tobaccos than upon any other naturally occurring compounds. TML:njw Xc: Dr. F. J. Schultz Dr. H. J. Minnemeyer Dt. C. I. Lewis

THE DETERMINATION OF THE EFFECTS OF VARIOUS ADDITIVES ON THE FORMATION OF PYRAZINE AND ALKYL-PYRAZINES ON ROASTED TOBACCO • In dealing with the formation of alkyl-pyrazines during the roasting of tobacco, it became evident that there were many possible pathways for their formations. To run enough single experiments to cover all of the possibilities would be time consuming and probably inconclusive. The Plackett- Burman Design for statistical analyses of multivariable systems gives a convenient method for obtaining an overall picture of some of the possible reactions taking place when tobacco is roasted. It has been shown in many previous papers that pyrazines can be formed from sugars and amino acids, dicarbonyls and amino acids, amino acids reacting with amino acids, ammonia reacting with sugars, etc. It was the aim of this experiment to narrow this field as it applies to roasting tobacco. EXPERIMENTAL A single lot of mixed, uncased burley tobaccos was used throughout the experiment. The six variables (additives) used were glyoxal, 2,3-butanedione, glucose, fructos5~ an amino acid mixture, and a divalent metal catalyst (Cu ), Glyoxal and 2,3-butanedione were added to the tobacco by spraying at a level of approximately 3%. Glucose and fructose (aqueous) were added at the 5% level. The catalyst was added at the 0.5% level. The amino acid mixture was a blend of amino acids having the same•profile-as amino acids found on burley tobaccos. This amino acid mixture was added to the tobacco dry, at a 5% level. Table 1 lists the composition of the amino acid mixture. The experimental design is given in Figure 1. The plus (+) indicates that a variable was added, and a minus (-) indicates that a variable was not added. The treated tobacco samples were stored in plastic bags. :e Samples were roasted in a random order. Previous roasting experiments, carried out in the laboratory roaster (pressure cooler), wep done at 120°C, with an optimum roasting time of 4 hours . The tobacco samples in this experiment had high moisture content as a result of applying additives. The roasting temperature had to be reduced ten degrees to avoid gross leakage and to keep the pressure within safe limits (25 psi). One hundred fifty grams of tobacco sample was roasted in each of two roasters ~ for 4 1/2 hours. a!'' N O ' .~ ' ?' . CD

After roasting, the cookers were allowed to cool overnight. The tobacco was extracted the following day using a slightly modified extraction procedure which increases pyrazine recovery by 10%. The procedure is given in Appendix 1. The concen- tration procedure and GC analysis are the same as described in Report # 543. Table 2 is a tabulation of the data obtained from the ten experiments by weighing the peak areas of pyrazine, methyl- pyrazine, and dimethylpyrazine. Relative amounts of each com- pound are listed. The quantitative determinations of the pyrazine compounds were not made as it is not a requirement for this type of experimental design. The relative amounts of each compbund were entered into the computer as required by the programl. Figures 2, 3, and 4 are copies of the print-out results. INTERPRETATION Note that while samples 2 and 4 are different under the dummy columns (Figure 1), they should be identically treated tobacco samples. Also samples 3 and 5 should be identical tobacco samples, and accordingly should give similar results. Obviously the results are not the same (Table 2). Therefore, one must consider that while the same amounts of compounds were applied to each, the applications were not necessarily uniform. The results of the experiment can only be interpreted with this in mind. Fortunately, it is*not necessarily a requirement of the design that experimental results be quanti- tative. The results in Figure 2 show that glyoxal, glucose, free amino acids, and Cu2+ enhance pyrazine formation (Effect (- to +) is positive), while fructose has a negative effect on pyrazine formation. It may be that fructose competes with glyoxal and glucose for available nitrogen otherwise incorporated into the pyrazine ring. The results in Figure 3 show that glyoxal, glucose and free amino acids enhance the formation of methylpyrazine. As in the formation of pyrazine, fructose (and 2,3-butanedione) may be in comp etition for available nitrogen. The negative effect of the CuZ+ catalyst may be due to a Cu2+ complex forming because of the2}sequence in which the application of amino acids, glyoxal, and Cu was made.- The results in Figure 4 show that only 2,3-butanedione has a significant positive effect on the formation of dimethylpyrazine. Fructose and the Cu2+ catalyst also have positive effects but are less significant. Glyoxal and glucose may compete with fructose and 2,3-butanedione for the available nitrogen, as they show negative effects on dimethylpyrazine formation. It may be that the formation of dimethylpyrazine is a very specific reaction in the case of roasted tobacco, 2,3- butanedione (and fructose) combining with a very small percentage of amino acids or available nitrogen to form dimethylpyrazine. •

3 All of the interpretations made so far are compatable with current literature on the subject. It appears that the formation of pyrazine may be facilitated by a wide variety of carbonyl compounds and nitrogen sources, whereas, dimethyl- pyrazine and the pyrazine compounds with a high degree of substitution may not be so easily formed. Addition of aldehydes or dicarbonyls and nitrogen sources specific for the formation of multi-substituted alkyl pyrazine may be necessary to signi- ficantly increase the concentration of these compounds. SUGGESTIONS FOR FURTHER STUDIES Thus far, the pyrazine compounds studied have been in relationship to tobacco and roasted•tobacco. We have had to focus on only three pyrazine compounds when there are many,. most of which are more pleasant smelling and more potent. The next step in this pyrazine study should be a study of pyrazine compounds in tobacco smoke, aiming our study at answering the following questions: 1. What pyrazine compounds are found in tobacco smoke? 2. At what levels are these compounds found? 3. How do pyrazine compounds effect-the taste of tobacco smoke? 4." Would a change (either increase or decrease) in pyrazine compound concentration in tobacco smoke make a contribution to,tas te? - 5. Does high or low alkyl-pyrazine content correlate with high or low concentrations of other compounds important to cigarette smoke taste? WORK IN PROGRESS To try to determine the taste effect of various pyrazine compounds on tobacco smoke, I sprayed ethanol solutions of pyrazine, 2-methylpyrazine, 2,5-dimethylpyrazine, and 2,6- dimethylpyrazine on Kent tobacco at a level of 1000 ppm. Each sample-of treated Kent tobacco, as well as a control, was made into straight cigarettes and smoked by myself, Dr. Lewis, and Mr. Jim Bell. The results of the taste test can be summarized by saying that the taste modifications were very slight. The only difference I could tell between a control cigarette and any of the four test cigarettes was a tendency for the test cigarettes to be more mellow. Since these pyrazine compounds have odor thresholds from 10 to 100 ppm, I expected a significant change in taste. This significant change in taste was not evident to me, Dr. Lewis, or Mr. Bell.

Table 1- Composition of Amino Acid Mixture Amino Acid % Amino Acid of Total Amino Acid on Tobacco Grams Amino Acid in 100 g of Mixture Alanine 4.8 5 Valine .4.9 5 Glycine 5.2 5 Isoleucine 3.8 4 Leucine 5.7 6 Proline 4.7 5 Threonine 3.3 3 Serine 8.7 9 Methionine 3.3 3 Phenylalanine 4.1 4 Aspartic Acid 24.0 24 Glutamic Acid 12.4 12 Tyrosine 2.4 2 Lysine 4.8 5 Histidine 4.9 5 Arginine 2.0 2 Totals 99.0 $• 99 grams

Table 2 - Relative Amounts of Pyrazine and Alkylpyrazines Sample in Roasted Samples Pyrazine Methylpyrazine Dimethylpyrazine 1 400 110 250 2 460 150 90 3 10 85 480 4 420 500 20 5 30 25 110 6 1300 110 -40 7 10 100 90 8 60 10 60 9 950 140 250 10 80 270 100

Figure 1 - Experimental Design Sample Glyoxal 2,3-Butanedione Glucose Fructose Amino Acid Catalyst *D. *D. * D. 1 + + - + - + - + - 2 + - + - + - + - + 3 4 5 - + - + - + + - + 6 + - + - + + - + - 7 - + - + + - + - + 8 + - + + - + - + 9 - + + - + - + - + 10 * Dummy tPz,z4oz9to

Figure 2 CI1l _ 09:2br-JT 0:3/07/75 PLACi:LTT-:j U'Ri4i Ai1 1);j:iIGid FOR -10 EXPcRI,.SE,dTS SYSl'::: ;-1iUA:i T ",:U TUiiACCU R E:iP7i1; E-PYRAZ I.dE 1.lIi:IJ,iU;:: SI( . I:IFICA+~CE IINCICATED ABOVE 80 PERCENT VAR. I1J1,ii's~.R DU1,i.;tY3' 1 • GLYJXAL..- .2 . 3J"i A: I i:v 10 'r:L 3 - . GLUCUSc 4 FRU::TU SE . 5 'A;il:d" AC.IDS 6 ' CAi ALYST i.._ .7 DU:i;.tY 1 , . 8 DUi;aY2 •, AVE:2AGE VALUE OF RESPUPISE VAR. ' NA,:i E VAaI A:dCc = 14519.7 ., : .STA:lDA!:i) i:RROR =.. 120.498 ji EG:3 ELS OF .}=R EL-'DG; ~.= 3 EFFECT(-TU +) 315 ' -183.6 533.8 -543.4. ~ 509.8 214.5 3.4 129.8 -163.4 T-VALUE 2.6141•6 1.56517 4.42996 . 4. 5U 963 4.23078 1.78012 2.82163C-2 1.0772 1.35604 . SIGN. PERCE;JT 90 -PcRCENT 0 P~~tCL;1T 95 PERCE;dT 95 PERCEIVT . 95 PERCE-idT 80 PERCElIdT . 0 PERCENT 0 PERCENT 0_ PERDEIdT

Figure 3 09:2f3f:OT 0ii/07/75 PLACi ETf-3URi,',Ahl UES IGI1 FUR - . 10 SYSTL;:(-f;UAJTEli TU;3ACCU RESP~Ji1jE-; ;c'TFIYPY:?.1ZI :iE llli~dl~iJ;4 SIG:.IFICAi~iCE Ii•1CICATED AI30VL 80 PE:~CEPI"f `. VAR. N U:aB, _Z R VAR. I tk i a - GLYOXAL BUTAI!EUIUI:L GLUi.USc_ Fc2Jc,Y~~SE . AMl;+u ACIDS CATALYST DU:.;,.iY 1 ixJ::::.iY2 UU i k':iY3 AVERAGE VALUE UF ?ESPi);ySE IIAR I.1NCE = 4946: 51 , STANDARD EF?R0R. = 70.3314 DEGR EES OF FR EEDu;a = 3 e 'EFFCCT(-T0 -~) T-VALU E SI J1'. PcI1CE:JT . 51.2 0.727982 0 PE'2C~l-'i ~11::-).6 .i . 64365 80 !'=.ZCr-Wf 63.6 0.90429 0 PE:~Cra1 -171.6 2:43985 90 P=,QCE;aT 102 ' 1.45028 0 PCR Cr~Ii -175 . 9.17 2.50125 90 PE:~Cr::if 68.4 .0.972539 J PIF i?Cc- ::T .23.6 0.335554 _ 0 P'.'^_'r•Z(~i`::•il -98 1.3934 0 PFRC%~:1f 149.2

Figure 4 !iU;•I-9 CII' U~~:3UL~11 O8/07/75 I'LAC,:;.:1'1•-i;UI7i.~AiJ l~i:SIGIJ 1=UII 10 i=XP?RIi.iiri:l'S JYJ j Li•S-RuASI'EU TUAt;CU :~ la3i'LJ;l:i L-•-11ji I:.11: f I{YLP YR AZ I;JE ~.'IIiI;:iJ+.i 6ICi,IFIl:A:iCE Ii:"ICA'I`ED A`30VE 80 NFRCEidT VAR' . :iU.:ii;~~ VAIs.' ' EFI=EC"l'(-TU +) T-=VALUE SIGN. PERCE:,Ff • . . • . . • 1 GLYc3XAL- -113.8 •1 . 42 519 0 Pi=RCEIJT 2 EiU 1•A.1I UIt ;L 173:4 2.17161 80 PcRCE,dT s- GLti ~U:i~ -1 17.•8 1.47529 0 Pi=i?CEIJT 4. i=C105'_:.' 101.4 1'. 26 99 0 Pt=?CE;J"!' 5- A; T43 :1t;IDS -105.4 •.1 .32 0 PERCENT ci CATALYST ' 12U: 333 1.50702 0 PERCE;Jl 7 OU ris,iY 1 . , -75. 4 0.944285 0 PcRC_?dT 8. DUt`:;aY2 40.6 0.508461 0 PERCENT 9 DU:.':.,Y3 103.6 1.36007 0 PERCENT AVERAGE VALUE OF RJ=S"r'W::;iE = 148.7• VAR I A•iJCL = 6375,83 SiA:JDA~~ ~~;RuR = '19.8488 DEG;3LES U'r FREEi)U:t = 3 '