Development of non-aqueous single stage derivatisation method for the determination of putrescine and cadaverine using GC-MS

Transcription

1 Cent. Eur. J. Chem. 6(2) DOI: /s Central European Journal of Chemistry Development of non-aqueous single stage derivatisation method for the determination of putrescine and cadaverine using GC-MS M. Ali Awan Centre for Instrumentation and Analytical Science, School of Chemical Engineering and Analytical Science, The University of Manchester, Manchester M60 1QD, United Kingdom Research Article Received 21 October 2007; Accepted 14 January 2008 Abstract: A single step derivatisation for the determination of putrescine and cadaverine by gas chromatography using trifluoroacetylacetone (TFAA) in methanol or ethanol was studied and optimised. The derivatives were analysed by an iontrap gas chromatograph-mass spectrometer (GC-MS) operating with electron impact ionisation with selective ion storage (EI-SIS) mode. The optimised mole ratios for TFAA/putrescine and TFAA/cadaverine reactions were 5/1 and 5.8/1 respectively with a reaction time of 15 minutes at 95 o C. The retention times for the derivatised putrescine and cadaverine were 11.3 and 12.2 minutes respectively using the capillary column, CP-Sil 8CB; 30 m length x 0.25 mm i.d. x 0.25 mm film. The correlation coefficients (R 2 ) of calibration curves for putrescine and cadaverine were and respectively over a concentration range of 100 ng cm -3 to 1500 ng cm -3. The method developed was found to be simple (single-stage derivatisation), rapid (15 minutes derivatisation & 14 minutes GC/MS run) and accurate (putrescine and cadaverine recoveries 94.8% %). Keywords: Biogenic amines Trifluoroacetylacetone Putrescine Cadaverine Non-aqueous derivatisation Versita Warsaw and Springer-Verlag Berlin Heidelberg. 1. Introduction Putrescine and cadaverine are environmentally and biologically significant biogenic amines. They occur in all organisms where they are associated with cell growth [1,2]. They play an important role in the formation of nucleic acids and cell membranes [2]. The requirement for putrescine increases rapidly in growing tissues [3], therefore comparatively higher levels of putrescine are observed in rapidly dividing cells such as tumour cells [4]. The presence of biogenic amines in food, especially in fish, cheese and meat products is a result of microbial decarboxylation of amino acids [1]. So, biogenic amines in food and meat products are related both to food spoilage and food safety. Normally, in fresh meat putrescine levels are low, however, considerably higher levels will be observed in meat produced where hygiene is poorly controlled [5]. In vegetables biogenic amines are associated with spoilage due to poor storage [6]. In general the concentrations of biogenic amines give an indication of the levels of microbiological contamination in food products [7] and hence act as reliable quality indicators. Generally, high performance liquid chromatography (HPLC) and gas chromatography-mass spectrometry (GC-MS) are used to quantify biogenic amines in real samples [1]. Both of these techniques involve extensive procedural steps of amines extraction from complex sample matrices and then their derivatisation to improve chromatographic properties [2,8]. Usually, biogenic amines are first derivatised in the aqueous phase (sample matrix) and then the derivatives are extracted using organic solvents prior to GC and HPLC analyses [8]. The main features of the aqueous phase derivatisation procedures are long derivatisation/extraction times, low reproducibility owing to the stability of the derivatisation products and the difficulty in producing reproducible extractions [1]. * ali.awan@mail.com 229

2 Development of non-aqueous single stage derivatisation method for the determination of putrescine and cadaverine using GC-MS This study sought to remove the need for an aqueous phase and to remove the solvent extraction procedures from the analysis of biogenic amines. Success here would enable the future development of automated gasphase derivatisation methodology that in turn would minimise the effect of the sample matrix. Additionally, we also aimed to significantly increase the sensitivity of the analysis by eliminating the losses of derivatised amines associated with solvent extraction procedures. Trifluoroacetylacetone (TFAA) was selected to derivatise putrescine and cadaverine in methanol and ethanol (solvents) as the introduction of the trifluoromethyl group enhances the volatility of derivatised molecules [9] making it suitable for use in conjunction with an iontrap GC-MS and GC-FID. 2. Experimental Procedures Six studies were undertaken to characterise amine derivatives (Studies 1-6) Chemicals All six amine test compounds; 1,4-butanediamine (putrescine) 99%, 1,5-pentanediamine (cadaverine) 95%, 1,2-diaminoethane 99%, 1,3-diaminipropane 99%, 1-butylamine 99% and 1-nonylamine 99% were purchased from Aldrich (Germany). Analytical grade solvents methanol, ethanol and dichloromethane were obtained from BDH (England). 1,1,1- trifluoroacetylacetone 98% and nonadecane (C 19 H 20 ) > 99% were purchased from Fluka (Switzerland) Preliminary characterization of derivatives Aqueous phase derivatisation using buffer and solvent extraction (Study 1) Preliminary mass spectrometric characterisation of the putrescine and cadaverine derivatives was based on an aqueous phase derivatisation followed by GC-MS ion-trap (Varian CP-3800 GC coupled to Saturn 2000 mass spectrometer) analysis. The procedure adopted in Study 1 was similar to the method reported previously [9]. TFAA, putrescine and cadaverine stock solutions were prepared in 10 cm 3 volumetric flasks with ethanol at concentrations of 38.2 mg cm -3, 8.77 mg cm -3 and 8.73 mg cm -3, respectively. Putrescine and cadaverine were derivatised (in triplicate) by pipeting 1 cm 3 of phosphate buffer (ph 7) into a 4 cm 3 reaction vial, followed by the addition of 0.8 cm 3 of the TFAA solution and 0.5 cm 3 of either putrescine or cadaverine solution. This protocol maintained a TFAA:amine mole ratio of 4:1, this being higher than the theoretical ratio of 2:1, in order to ensure 100% derivatisation. The reaction vial was sealed before heating it to 95 o C for 15 min, after which it was allowed to cool to room temperature. Once cool the reaction mixture was extracted with 1.5 cm 3 of dichloromethane (DCM) at room temperature. The extract was recovered and transferred into a 4 cm 3 vial. The extract was then evaporated to dryness under a stream of nitrogen and the residue was dissolved into 2 cm 3 of ethanol to form a stock solution. This stock solution was diluted further, twenty fold, in ethanol before analysis on GC-MS. A GC column 30 m 0.25 mm i.d. with a 0.25 µm thick 5% phenyl-95% methyl stationary phase (Varian CP Sil-8) was used during analysis. The instrumental parameters are summarized in Table Single stage derivatisation method (Study 2) In this study a modification in the derivatisation procedure (Study 1) was made by carrying out derivatisation reaction in non-aqueous phase (methanol/ethanol) in order to eliminate the solvent extraction step. TFAA, putrescine and cadaverine stock solutions were prepared in 10 cm 3 volumetric flasks with ethanol at concentrations of 38.2 mg cm -3, 8.77 mg cm -3 and 8.73 mg cm -3, respectively. Derivatives of putrescine and cadaverine were prepared in triplicate by reacting 0.5 cm 3 of their respective stock solutions with 0.8 cm 3 of TFAA in a 4 cm 3 reaction vial. The vial was sealed and heated to 95 o C on a hot plate for 15 min. The vial was then left to cool down to room temperature. The vial content was then evaporated to dryness under a stream of nitrogen and the residue was dissolved into 2 cm 3 of ethanol to form a stock solution. These stock solutions were diluted twenty fold, using ethanol, and analysed by GC-MS ion trap. A GC column 30 m 0.25 mm i.d. with a 0.25 µm thick 5% phenyl- 95% methyl stationary phase (Varian CP Sil-8) was used during analysis. The instrumental parameters are summarized in Table Examination of single stage derivatisation method (Study 3) In order to explore the reaction mechanisms associated with the single stage derivatisation method two other diamines, 1,2-diaminoethane, 1,3-diaminopropane, and two monoamines, 1-butylamine and 1-nonylamine were also investigated using the procedure described in Study 2. TFAA, stock solutions were prepared in 10 cm 3 volumetric flasks with ethanol at concentrations of 38.2 mg cm -3, 8.99 mg cm -3, 8.88 mg cm -3, 7.40 mg cm -3 and 7.82 mg cm -3, respectively. The derivatives were obtained by reacting a mixture of diamines containing 0.5 cm 3 of each 1,2-diaminoethane and 1,3-diaminopropane stock solutions with 1.6 cm 3 of the TFAA in a 4 cm 3 reaction vial. Similarly, a mixture of monoamines containing 230

3 M.A. Awan Table 1. Summary of instrumental parameters used in this study. The values in brackets indicate GC column oven temperature hold times (min). Study number 1, 2 3 4, 5 6 Injector (Varian 1177) Split ratio 50:1 50:1 10:1 Splitless Temperature / C Injection volume / µl GC-MS or GC-FID Column flow / cm 3 min Start temp/ C (time/min) 40(2) 90(2) 150(2) 150(2) Temp ramp rate/ C min Final temp/ C (time/min) 280(9) 250(12) 280(3.3) 280(3.3) Transfer line temp/ C n/a 300 Iontrap temperature n/a 200 Ionisation / ev EI EI n/a EI Scan range / m/z n/a Ion storage: EI-SIS m/z n/a n/a n/a Putrescine: 69, 96, 110,138 Cadaverine: 69, 96, 152, 305, 334 Nonadecane: 57, 71, 85, 97 Chemical ionisation CI (methanol): m/z n/a n/a n/a FID detector temp/ C n/a n/a 300 n/a 0.5 cm 3 of each 1-butylamine and 1-nonylamine stock solutions were reacted with 1.6 cm 3 of the TFAA in a 4 cm 3 reaction vial. The remaining procedural steps have been described in Study 2 above. Each derivatisation was undertaken in triplicate. The GC-MS instrumental conditions are reported in Table Optimisation of derivatisation methods The derivatisation yield was optimised with respect to the stoichiometric ratios of the reagents and the derivatisation reaction time. The derivative yields were obtained using a GC-FID (Varian CP-3800 GC coupled to flame ionization detector). The sequence of studies that were run to optimise the proposed methods are summarised in Table 2. In this and subsequent studies we started the GC method with an oven temperature of 150 o C (see Table 1) and thus decreased elution times of the derivatised putrescine and cadaverine to 11.3 and 12.2 min, respectively Effect of TFAA:amine stoichometric ratio (Study 4) The effect of varying TFAA levels on derivative yield was studied by monitoring the peak area (yield) of each respective derivative at constant concentrations of putrescine and cadaverine. The volumes of TFAA stock solution for both methods described in Study 1 & 2 above were 0.35, 0.50, 0.75, 1.00 and 1.25 cm 3 (no. of moles , , , and ), respectively. Each solution was added separately to 0.50 cm 3 of either putrescine ( mol) or cadaverine ( mol) stock solution contained in a 4 cm 3 reaction vials. The remaining procedural steps were the same as described in the Studies 1 & 2 above. Each derivatisation was undertaken in triplicate. The GC-FID instrumental conditions are reported in Table Effect of derivatisation time at 95 o C (Study 5) Data from TFAA:amine stochiometric optimisation indicated that the single stage modification (Study 2) was significantly more sensitive than the aqueous phase method (Study 1). Therefore, in this work only the single stage modification was studied. The effect of reaction time at 95 C was investigated using the optimised stoichometric mole ratios of 5.0:1.0 and 5.8:1.0 for TFAA/putrescine or cadaverine, respectively, which were determined in the previous study. The optimum TFAA level 1 cm 3 ( mol) was added to 0.5 cm 3 of putrescine or cadaverine ( or mol) in each vial. Six reaction times were studied: 1, 5, 10, 15, 20 and 30 min using the single stage procedure described Study 2 above. Each derivatisation was undertaken in triplicate. The GC-FID instrumental conditions are reported in Table Calibration and validation (Study 6) Calibration curves were generated over the range 100 ng cm -3 to 1500 ng cm -3 using five putrescine or cadaverine standards. Since these standards were approximately one thousand times less concentrated, compared to 231

4 Development of non-aqueous single stage derivatisation method for the determination of putrescine and cadaverine using GC-MS Table 2. Summary of the optimisation studies following triplicate analysis at 95 o C. Study Derivatisation Mole ratio Reaction time / min Method TFAA:Put TFAA:Cad TFAA:amine mole ratio 1 Aqueous Phase TFAA:amine mole ratio 2 Aqueous Phase TFAA:amine mole ratio 3 Aqueous Phase TFAA:amine mole ratio 4 Aqueous Phase TFAA:amine mole ratio 5 Aqueous Phase TFAA:amine mole ratio 1 Single Stage TFAA:amine mole ratio 2 Single Stage TFAA:amine mole ratio 3 Single Stage TFAA:amine mole ratio 4 Single Stage TFAA:amine mole ratio 5 Single Stage Effect of reaction time 1 Single Stage Effect of reaction time 2 Single Stage Effect of reaction time 3 Single Stage Effect of reaction time 4 Single Stage Effect of reaction time 5 Single Stage Effect of reaction time 6 Single Stage those in the optimization study, a corresponding diluted TFAA solution ( g cm -3 ) of low volume (0.1 cm 3 ) was used to maintain constant mole fractions. 2 cm 3 of each standard was pipetted into 4 cm 3 vials to which were added 0.1 cm 3 of TFAA and 0.1 cm 3 of nonadecane (38 µg cm -3 ) solutions. The total volume of the vial was 2.2 cm 3. The vials were sealed and heated to 95 o C on a hot plate for 15 min. The derivatised calibration standards were significantly less concentrated than those used in the optimisation studies and were analysed by GC-MS without further dilution in triplicate under electron impact selective ion storage (EI-SIS) conditions, Table 1. Accuracy of the developed method was calculated as recovery percentage (%R= [C o /C s ] x 100) between found (C o ) and known (C s ) concentrations of putrescine and cadaverine by running their standards ranging ng cm -3 in triplicate. 3. Results and Discussion 3.1. Evaluation of the aqueous phase (Studies 1) & single stage (Study 2) derivatisation procedures Study 1 provided reference putrescine and cadaverine derivatives for comparison against the products obtained from the single stage method (Study 2). Putrescine and cadaverine derivatives were obtained by treating putrescine or cadaverine standards with TFAA in the presence of buffer (Study 1). The chromatogram showed peaks appearing at retention times of 24.5 and 25.4 min for the putrescine and cadaverine derivatives, respectively. The mass spectra were acquired at 70eV in the positive-ion electron ionization (+EI) mode using an ion-trap instrument, scanning over the range m/z The main ion fragments of the derivatives along with intensities relative to their base peaks (in brackets) are as follows: Putrescine derivative m/z: 69(22%), 96(22%), 110(45%), 111(14%), 138(100%), 166(13%), 207(53%), 208(18%), 320(14%), 361(4%); Cadaverine derivative m/z: 69(46%), 96(68%), 110(19%), 124(44%), 152(97%), 166(28%), 220(65%), 305(31%), 334(100%), 375 (11%). The reproducibility of the derivatisation process was assessed by repeating each derivatisation three times. On each occasion the chromatography and associated mass spectra were the same. Study 2 tested the feasibility of eliminating water and using non-aqueous solvents in the derivatisation process in order to remove the time consuming steps of solvent extraction. Fig. 1 shows the chromatogram of the putrescine and cadaverine derivatives obtained by treating mixtures of putrescine and cadaverine standards with TFAA in ethanol. The chromatogram shows principal peaks appearing at retention times of 24.5 and 25.4 min for the putrescine and cadaverine derivatives, respectively. The mass spectra have been examined under same conditions as in Study 1 above. The retention times and mass fragments of the principal peaks confirmed that the same derivatives were obtained both in Study 1 & 2. The derivative peak intensities in Study 2 were much higher than the derivative peak intensities in 232

5 M.A. Awan were observed at retention times of 25.7 and 26.5 min for putrescine and cadaverine, respectively. Their corresponding main mass fragments were observed at m/z 69, 84, 98, 110, 138, 168, 188, 207, 228 (putrescine) and m/z 69, 84, 98, 110, 140, 188, 194, 242 (cadaverine). Elucidation of exact structures of these products will be the subject of future work. However, their presence did not interfere with the chromatography and reproducibility of the putrescine and cadaverine derivatives. Minutes Minutes Figure 1. Chromatograms observed in Study 2 produced by putrescine (Top) and cadaverine (Bottom). A and C are the principal peaks while B and D are the side products. Study 1 because the single stage method eliminated the possibility of derivatives loss during extraction. The extent of increase in yield is evaluated in Study 4. In addition to the principal derivative peaks observed, less intense side-product peaks (not seen in Study 1) were also noted, see Fig. 1. The additional products 3.2. Examination of single stage derivatisation method (Study 3) The validity of the single stage derivatisation method was tested by derivatising two monoamines (1-butylamine & 1-nonylamine) and a series of diamines (C2-C5) with TFAA in ethanol using the single stage method. This study was carried out under different chromatographic conditions thus putrescine and cadaverine derivatives showed different retention times from Studies 1 & 2, see Table 1 for GC-MS instrumental settings. Each derivative chromatogram showed a principal derivative peak and an additional product peak. The mass spectra have been examined under positive-ion electron ionization (+EI) and positive-ion chemical ionization (+CI) modes using the ion-trap instrument with a scan range m/z The retention times and main ion fragments of the derivatives along with relative intensities are listed in Table 3. In the case of the monoamines (1-butylamine & 1-nonylamine), the spectra of these derivatives show typical Schiff s base [10] characteristics, see Table 3. Table 3. Main mass fragments (+EI) and protonated molecular ions [M+H] + (+CI) of 6 amine derivatives obtained by single stage method (Study 3). The values in brackets are % relative intensities of the ions. Amine Retention time / min (+EI) Principal mass fragments(m/z) 1-Butylamine (17), 69(18), 70(16), 84(23), 96(16), 98(37), 122(43), 140(100), 166(20), 209(24), 210(18) 1-Nonylamine (16), 70(15), 84(21), 96(11), 98(37), 166(23), 168(34), 180(18), 182(34), 192(25), 208(33), 210(100), 264(24), 279(12) 1,4-Diaminobutane (putrescine) (16), 96(13), 110(34), 111(10), 138(100), 166(8), 207(65), 208(20), 291(11), 361(6), 414(10) 1,5-Diaminopentane (cadaverine) (36), 96(24), 110(14), 152(82), 166(21), 180(11), 208(12), 220(56), 221(24), 222(28), 305(35), 334(100), 375(10), 428(14) 1,2-Diaminoethane (27), 69(33), 82(59), 96(24), 98(100), 110(38), 166(98), 179(90), 263(25), 333(24), 386(8) 1,3-Diaminopropane (15), 69(19), 82(15), 96(33), 98(100), 124(34), 162(17), 166(17), 180(47), 192(31), 194(21), 277(23), 326(22), 347(36), 400(5) (+CI) [M+H] +. ion (m/z) (100), 415(3) 375(100), 429(2) 333(100), 387(2) 347(100), 401(2) 233

6 Development of non-aqueous single stage derivatisation method for the determination of putrescine and cadaverine using GC-MS The nitrogen of amines undergo a nucleophilic attack on the carbonyl carbon (bearing CF 3 group) of TFAA and a condensation reaction occurred with loss of a water molecule, see Scheme 1. TFAA has been shown to undergo such condensation reactions with amines [11]. CH 3 (CH 2 ) n NH 2 + O Scheme 1. Condensation reaction between monoamines and TFAA. The formation of butylamine and nonylamine derivatives was verified by +EI [M] +. and their +CI [M+H] + molecular ions, see Table 3. The mass to charge ratio of [M] +. & [M+H] + for butylamine derivatives are m/z 209 & 210 and for nonylamine derivative m/z 279 & 280. The +EI spectra showed characteristic loss of m/z 69 (-CF 3 ) from their respective [M] +. to form base peak ions at m/z 140 (butylamine derivative) and m/z 210 (nonylamine derivative). Unlike monoamine which forms a Schiff s base after derivatisation, diamine (C2-C5) derivatives exhibited complicated mass spectra. The reaction of TFAA with diamines should produce a derivative structure as suggested below, Scheme 2. For putrescine and cadaverine derivatives the theoretical molecular masses of m/z 360 and m/z 374 should be observed in their respective spectra: O O CH 3 C (CH 2 ) n C CF N N 3 Where n is a number of methylene groups in a diamine molecule. C C CH 3 CF 3 Monoamine TFAA Derivative Where n = 3 & 8 for 1-butylamine & 1-nonylamine, respectively. Scheme 2. Structure of diamine derivatives. H 3 C (CH 2 ) n N C + H 2 O The above mentioned molecular structures were reported in an aqueous phase multi-step derivatisation of putrescine and cadaverine [9]. The difference between the previously published spectra and Study 3 s mass spectra were the molecular ions. +EI mass spectra indicated that the molecular ions for the putrescine and cadaverine derivatives were m/z 414 and m/z 428, respectively, whereas previously [9] the molecular ions [M] +. were assigned to m/z 360 and m/z 374. The chemical ionisation (+CI) mass spectra obtained in Study 3 show that the principal species for the putrescine derivative was m/z 361, [M+H] +, along with a lower intensity ion at m/z 415, corresponding to [M+54] H+. Similarly, for the cadaverine derivative the principal species was observed at m/z 375, [M+H] +, with a lower intensity ion m/z 429, corresponding to [M+54] H+. The mass spectra of derivatives of two other CH 3 O O C CH 3 CF 3 CF 3 diamines; 1, 2-diaminoethane and 1, 3-diaminopropane formed by the single stage method also yielded [M+54] +. under +EI conditions and [M+H] + with a lower intensity ion corresponding to [M+54] H+ under +CI conditions, see Table 3. The mass difference of 54 amu between these observations and previous reports for this reaction merits further elucidation and will be the subject of future work. Nevertheless, the quantitation in this study was based upon total ion current (TIC) measurements. All diamines generated stable chromatographic and mass spectral responses throughout this study and the difference was judged not to materially affect the analytical outcomes of this work. The reproducibility of the derivatisation process was assessed by repeating each derivatisation. On each occasion the chromatographic data and associated mass spectra were identical The effect of TFAA:amine stoichometric ratio (Study 4) Varying the TFAA/amine ratio had a significant effect on the derivative yield and hence the sensitivity of any related analytical method, see Fig. 2. In the aqueous phase derivatisation method the yield increased 3.4 times for putrescine and 4.2 times for cadaverine as the mole-ratio [TFAA]/[Amine] increased from 1.75 to 6.23 and 2.0 to 7.25 for putrescine and cadaverine, respectively. Above a mole ratio of 4 the derivative yield shows stability (graph plateaus). With the single stage method the increase in yield over the above mentioned mole-ratio range was greater, with a maximum yield at a ratio of 5 for putrescine with a yield 2.1 times greater than that observed from the aqueous phase method. This effect was larger with cadaverine with an increase in yield of 4.1 observed for the optimum mole-ratio of 5.8. Increasing the mole ratios beyond these levels resulted in a decrease in the yield. The optimum mole ratios corresponding to the maximum peak areas were 5.0:1.0 and 5.8:1.0 for TFAA/putrescine and TFAA/ cadaverine, respectively. The mole ratios are significantly higher than the theoretical ratio of 2:1, increasing the TFAA concentration which shifts the equilibrium of the derivatisation processes further to completion Effect of derivatisation time at 95 o C (Study 5) Fig. 3 shows the effect of reaction time, which followed the same general trend for putrescine and cadaverine. The yield of the putrescine derivative product increased to a maximum level of 1.55 at approximately 15 min, after which the yield reduced significantly and for cadaverine the maximum level was 1.81 with similar losses observed after 15 min. The reduction in yield beyond an optimum reaction time was anticipated with 234

7 M.A. Awan Peak area Ratio (Derivatives/Int. Std) Mole Ratio (TFAA/Amines) Figure 2. A comparison of the effect of the TFAA/Amine mole ratio on the yield of the putrescine (triangles) and cadaverine (circles) derivatives. The data obtained from the aqueous phase method are denoted by the solid symbols while open symbols represent the findings from the single stage modified method. Note that the peak areas are expressed as a ratio of the internal standard used in the mole-ratio optimisation study. Error bars denote the 95% confidence level and are based on triplicate measurements. the semi-volatile derivative being driven into the headspace of the vial, and hence lost to the analysis; this effect was more pronounced for the putrescine, the more volatile of the two derivatives. A reaction time of 15 min was therefore chosen for all subsequent studies Calibration and Validation Electron impact selective ion storage (EI-SIS) (see Table 1) was used to generate calibration curves for putrescine and cadaverine. Here the aim was to examine linearity of the calibration only and not to quantify amines in real samples. EI-SIS mode helped in generating putrescine and cadaverine chromatograms with low background noise as compared to analyte peak signal. Calibration functions were constructed from triplicate independent derivatisations with five putrescine and cadaverine standards ranging from 100 ng cm -3 to 1500 ng cm -3 using nonadecane as the internal standard, see Fig. 4. The calibration function was non-linear, reflecting a more complex derivatisation reaction. A second-order polynomial approach (MS Excel) was used to model the relationship between concentration and response. The approach provided good calibration functions with correlation coefficients (R 2 ) of and for putrescine and cadaverine respectively. Such a secondorder polynomial approach has also been reported to accommodate non-linear calibration data obtained by a GC-MS Study of VOCs in water [12]. The ultimate limits of detection were not part of this study. It is likely that with decreasing concentrations of the analytes the optimum reaction conditions will change. What was established was that the sensitivity and signal to noise characteristics of the responses are adequate for analysis of putrescine Peak Area Ratio (Derivatives/Int. Std) Reaction Time/minutes Figure 3. Effect of reaction time on the yield of derivatives. Increasing the reaction time allowed the derivatisation to proceed towards completion. This process was counteracted by increasing analyte losses from the vial (into the headspace) over time. This effect was greater with the more volatile putrescine derivative. Putrescine is denoted by solid triangles and cadaverine by solid circles- three derivatisations were performed for each datum point. Error bars denote the 95% confidence level and are based on triplicate measurements. Peak Area Ratio (Derivatives/Int. Std) Concentration/ ng cm -3 Figure 4. Calibration curves for putrescine (triangles) and cadaverine (circles) over a concentration range 100 to 1500 g cm -3 (in triplicate) obtained by polynomial least-square regression analysis. Correlation coefficients (R 2 ) are 0.99 and 0.99 for putrescine and cadaverine, respectively. Error bars denote the 95% confidence level and are based on triplicate measurements. and cadaverine in food and biological samples. The accuracy of this approach was estimated in terms of percentage recovery obtained by analysing putrescine and cadaverine standards at the lower, middle and upper parts of the calibration ranges, Table 4. The recoveries observed were in the range 94.8% % indicating a satisfactory level of accuracy over the selected range of study. The student t test also confirmed the validity of the developed method as all the calculated t values are less than the reported [13] value t = for 2 degree of freedom (N-1) & 95% confidence level. 235

8 Development of non-aqueous single stage derivatisation method for the determination of putrescine and cadaverine using GC-MS Table 4. Summary of the accuracy assessment showing the estimates of the observed material (C O ) against the mass present in the standard (C S ). Each assessment was undertaken in triplicate and the standard deviations (s) on the observed values is given in brackets. Analyte C S / µg cm -3 C O / µg cm -3 %Recovery t-score* Comments Putrescine (0.007) no significant difference 4. Conclusions (0.026) no significant difference (0.097) no significant difference Cadaverine (0.01) no significant difference (0.037) no significant difference (0.056) no significant difference C S - mass of analyte in standard C O - observed mass of analyte (average of triplicate), the figures in brackets are the standard deviations from replicate standards CO %Recovery - % R = 100 C ( CO CS) N t - Test* - ± t = s S Where N is number of measurements and s is standard deviation, for 2 degree of freedom & 95% confidence level t = A single stage derivatisation method for putrescine and cadaverine has been developed using trifluoroacetylacetone (TFAA) in non-aqueous solvents (methanol/ethanol). The time consuming, and expensive steps of ph control and solvent extraction have been eliminated from the methodology. The resultant procedure is significantly faster (15 min derivatisation and 14 min GC/MS run), up to 5 times more sensitive and accurate (putrescine and cadaverine recoveries 94.8% %). The ultimate goal of the current research is to develop a derivatisation procedure for biogenic amines that may be undertaken within gas chromatographic inlet systems. This non-aqueous single stage derivatisation method is an important first stage in the realisation of this goal. The next studies used this approach to develop an automated vapour-phase derivatisation References method for biogenic amines and in doing so eliminated the laborious and expensive procedures associated with sample preparation and extraction. Acknowledgements The financial support provided by the National University of Sciences and Technology, Pakistan, for this research is highly acknowledged. Thanks are also due to Professor Paul Thomas for his supervision and support during this study and to Dr Ian Fleet for his valuable collaboration on mass spectra interpretation and technical guidance. [1] A. Önal, Food Chem. 103, 1475 (2007) [2] D. Teti, M. Visalli, H. McNair, J. Chromatogr. B. 781, 107 (2002) [3] A. E. Pegg, D. J. Feith, Biochem. Soc. T. 35 (2), 295 (2007) [4] P. Kalač, P. Krausová, Food Chem. 90, 219 (2005) [5] P. Kalač, Meat Sci. 73, 1 (2006) [6] S. Moret, D. Smela, T. Populin, L. S. Conte, Food Chem. 89, 355 (2005) [7] A. R. Shalaby, Food Res. Int. 29, 675 (1996) [8] M.Y. Khuhawar, G.A. Qureshi, J. Chromatogr. B. 764, 385 (2001) [9] M. Y. Khuhawar, A. A. Memon, P.D. Jaipal, M.I. Bhanger, J. Chromatogr. B. 723, 17 (1999) [10] E. H. Cordes, W. P. Jencks, J. Am. Chem. Soc. 84, 832 (1962) [11] K. Abbasi, M.I. Bhanger, M.Y. Khuhawar, J. Pharmaceut. Biomed. 41, 998 (2006) [12] I. Lavagnini, G. Favaro, F. Magno, Rapid Commun. Mass Sp. 18, 1383 (2004) [13] G. D. Christian, Analytical Chemistry. 6th Edition, John Wiley & Sons, Inc. (2004) 236