Compound-Specific Stable Isotope Analysis of Soil Mesofauna Using

Oct 18, 2003 - Cumbria LA11 6JU, U.K.. Philip Ineson. Department of Biology, University of York, P.O. Box 373, York YO10 5WY, U.K.. Stable isotope mas...
0 downloads 0 Views 101KB Size
Download PDF
Anal. Chem. 2003, 75, 6056-6062

Compound-Specific Stable Isotope Analysis of Soil Mesofauna Using Thermally Assisted Hydrolysis and Methylation for Ecological Investigations Claire J. Evans and Richard P. Evershed* Organic Geochemistry Unit, Biogeochemistry Research Centre, School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 1TS, U.K. Helaina I. J. Black Merlewood Research Station, Centre for Ecology and Hydrology, Windermere Road, Grange-over-Sands, Cumbria LA11 6JU, U.K. Philip Ineson Department of Biology, University of York, P.O. Box 373, York YO10 5WY, U.K.

Stable isotope mass spectrometric approaches are proving to be valuable tools in unravelling biotic interactions in complex ecosystems, yielding information on trophic preferences and functional roles of individual species. Gas chromatography/combustion/isotope ratio mass spectrometry (GC/C/IRMS) provides considerable opportunities to assist in studies concerned with ecosystem processes mediated by soil invertebrates and microorganisms by determination of δ13C values of individual compounds, for example, lipids, amino acids etc. However, techniques conventionally adopted for “wet” chemical extractions and derivatizations necessary for compound-specific stable isotope determinations restrict the size of soil organism that can be studied and can limit investigations of individuals or even parts of individuals. We demonstrate here that individual soil mesofauna can be probed directly for their fatty acid stable isotope signatures by pyrolysis-GC/ C/IRMS. A thermally assisted hydrolysis and methylation (THM) reaction is described for the determination of δ13C fatty acid values using trimethylsulfonium hydroxide (TMSH). Authentic fatty acids, acyl lipids, and individual Collembola (Folsomia candida) raised on C3 and C4 isotopically labeled yeast were analyzed initially by py-GC/ MS with TMSH and then by py-GC/C/IRMS. A kinetic isotope effect (KIE) observed with the THM reaction prevents direct calculation of the fatty acid δ13C values by simple mass balance equations. However, the KIE is shown to be both reproducible and robust and can therefore be accounted for by the use of correction factors. The fatty acid methyl ester compositions of individual F. candida and their respective δ13C values were determined and shown to agree with those obtained by conventional “wet” chemical procedures applied to much larger numbers of Collembola, thus enhancing the scope to which stable isotopes can be applied to the study of invertebrates in complex food webs in any environment. * Corresponding author. Fax: +44 117 9251295. E-mail: R.P.Evershed@ Bristol.ac.uk.

6056 Analytical Chemistry, Vol. 75, No. 22, November 15, 2003

Recent work has shown that stable isotope approaches can help to improve our understanding of trophic interactions in complex ecosystems, such as soil.1,2 The main advantage of this approach is that the determination of the natural variation in the stable isotope composition of an organism’s tissue can be utilized to give indications of trophic level and dietary behavior without the need for extensive visual observations or potentially unreliable gut content analyses. The naturally occurring differences in plant δ13C values produced by the C3 and C4 photosynthetic pathways also provide opportunities to utilize naturally labeled substrates for investigations of feeding behavior.3 The most widely used approach for stable isotope determinations of soil invertebrates involves bulk δ13C and δ15N analysis by isotope ratio mass spectrometry (IRMS).4-6 A major shortcoming of this approach is the limit set by the amount of C and N necessary for determining the 13C and 15N content of the soil fauna, such that tens to hundreds of organisms are often needed for a single determination. This factor dictates the size of invertebrate that can be studied and imposes practical limits on the investigation of very small, or even individual organisms. Furthermore, bulk isotopic determinations provide average values of a wide range of individual biochemical components, each of which will have characteristic isotopic compositions. Isotopic determinations of these biochemical components would have the advantage of providing additional isotopic information regarding the assimilation and biosynthesis of specific compounds from an organism’s diet, which would also assist in tracing the origins of nutrients necessary for biomass production. Recent technological developments have improved the specificity of stable isotope studies by use of a gas chromatograph (GC) interfaced to an IRMS via a combustion interface (and also a reduction interface for studies (1) Neilson, R.; Hamilton, D.; Wishart, J.; Marriot, C. A.; Boag, B.; Handley, L. L.; Scrimgeour, C. M.; McNicol, J. W.; Robinson, D. Soil Biol. Biochem. 1998, 13, 1773. (2) Boutton, T. W.; Archer, S. R. Rapid Commun. Mass Spectrom. 1999, 13, 1263. (3) Briones, M. J. I.; Ineson, P.; Sleep, D. Soil Biol. Biochem. 1999, 31, 937. (4) Boutton, T. W. In Carbon Isotope Techniques; Coleman, D. C., Ed.; Academic Press: San Diego, 1991; p 155. (5) Ponsard, S.; Arditi, R. Ecology 2000, 81, 852. (6) Scheu, S.; Falca, M. Oecologia 2000, 123, 285. 10.1021/ac034336d CCC: $25.00

© 2003 American Chemical Society Published on Web 10/18/2003

involving 15N) that is, GC/C/IRMS.7,8 This advance provides opportunities to access information regarding the isotopic composition of individual biochemical components. Fatty acids are ubiquitous constituents of all organisms, arising from the diet directly (essential) or biosynthesized de novo (nonessential) from dietary carbon sources. They occur in the diet of soil fauna in many different forms, typically triacylglycerols, phospholipids, and steryl esters, performing vital roles as energy stores, membrane components, and mediators. The specific fatty acid composition is dependent on a number of individual and environmental factors, including life stage, species, and diet, and often characteristic fatty acids are associated with a particular organism, for example, iso and anteiso odd carbon number fatty acids in bacteria.9,10 In addition, fatty acids have relatively rapid turnover rates, and their δ13C values are readily determined by GC/C/IRMS; thus, they constitute ideal targets for compoundspecific stable isotope analysis.11 The determination of soil mesofauna δ13C fatty acid values would be extremely beneficial for the interpretation of trophic interactions while also aiding in the determination of their functional role in nutrient cycles and improving our understanding of their lipid biochemistry. Compound-specific isotopic analysis requires the compound of interest to be isolated from the sample matrix and to be of suitable volatility for GC analysis; hence, derivatization of fatty acids is necessary prior to isotopic determination. Methyl esters remain the most popular derivative and can be prepared by a variety of methods. Conventional “wet” chemical procedures typically involve a solvent extraction followed by saponification and acid-catalyzed methylation. However, the small size of many soil fauna limits the use of conventional methods for fatty acid methyl ester (FAME) preparation. Although the amount of C required for GC/C/IRMS is significantly lower than for bulk isotopic analysis, the small size of soil mesofauna can still present problems, especially when information is required for single organisms, such as, for the study of biological variation within a species, or when subsamples of an organism are required for complementary investigations, for example, genetic identification.12 The steps involved in sample preparation also require a considerable amount of sample handling, which as well as being timeconsuming introduces the risk of contamination, especially when sample sizes are small. Thermally assisted hydrolysis and methylation (THM) reactions using tetraalkylammonium or trialkysulfonium salts are wellestablished derivatization methods for the chemical characterization of a wide range of synthetic and natural products, including waxes, soil, sediments, carbohydrates, and lipids.13-16 Fatty acids esterified in biomolecules, such as triacylglycerols and phospho(7) Matthews, D. E.; Hayes, J. M. Anal. Chem. 1978, 50, 1465. (8) Merrit, D. A.; Hayes, J. M. J. Am. Soc. Mass Spectrom. 1994, 5, 387. (9) Tunlid, A.; White, D. C. In Soil Biochemistry; Stotzky, G., Bollag, J.-M., Eds.; Marcel Dekker: New York, 1992, p 229. (10) Zelles, L. Chemosphere 1997, 35, 275. (11) Meier-Augenstein, W. Anal. Chim. Acta 2002, 465, 63. (12) Black, H. I. J.; Piertney, S. B.; Macdonald, C.; Standen, V.; Bull, I. D.; Evershed, R. P. Eur. J. Soil. Sci., in press. (13) El- Hamdy, A. H.; Christie, W. W. J. Chromatogr. 1993, 630, 438. (14) Asperger, A.; Engewald, W.; Fabian, G. J. Anal. Appl. Pyrolysis 1999, 52, 51. (15) Fabbri, D.; Helleur, R. J. Anal. Appl. Pyrolysis 1999, 49, 277. (16) Knicker, H.; del Rio, J. C.; Hatcher P. G.; Minard, R. D. Org. Geochem. 2001, 32, 397.

lipids, are rapidly converted to their fatty acid methyl esters in a single-step THM reaction in situ, avoiding the more complicated sample preparation protocols normally involved with wet chemical derivatization methods. Consequently, sample size requirements are lower, and the speed of analysis is increased. In addition, unlike conventional chemical extraction and derivatization procedures that generate large volumes of solvent waste, the in situ THM methodology requires only small volumes of reagent (low microliters), generating no solvent waste and also reducing the possibility of sample contamination. The principal derivatizing agent employed for such reactions is tetramethylammonium hydroxide (TMAH), and thorough evaluations of this approach and its applications have been presented by Kossa17 and Challinor.18 However, TMAH has been shown to produce isomerization or degradation of polyunsaturated fatty acid (PUFA) components when used in a THM reaction.19,20 Trimethylsulfonium hydroxide (TMSH) is a milder reagent21 that has been shown to be a very efficient methylating agent, and has been applied to the analysis of animal fats and vegetable oils,22 phenols,23 and the fatty acid profiling of microbial cells.24 Furthermore, it has proven to be a more favorable derivatizing agent for the analysis of PUFAs because of no significant isomerization or degradation being observed.22,24,25 The TMSH reagent can be used directly in combination with pyrolysis-gas chromatography, whereby THM is conducted in situ in the pyrolysis unit and derivatives are introduced directly into the GC for analysis. Thus, on-line THM using TMSH in the pyrolysis-GC inlet has considerable potential for the lipid analysis of individual micro- and mesofauna, particularly in conjunction with GC/C/IRMS for the determination of δ13C fatty acid values. We present here the results of a systematic study aimed at developing a protocol using a THM reaction with py-GC/C/IRMS for the isotopic determination of fatty acids from soil mesofauna. Our investigations proceeded with a preliminary assessment of the THM reaction using TMSH as the derivatizing agent; the method was applied to authentic acyl lipids and to the Collembolan species Folsomia candida using py-GC/MS. Subsequently, the possibility of a kinetic isotope effect (KIE) occurring during the derivatization step was examined, followed by determinations of the δ13C values of the major fatty acid components from individual F. candida using py-GC/C/IRMS. We demonstrate that this approach can provide compositional and stable isotope information identical to that obtained using conventional wet chemical procedures, but with the advantage of allowing determinations of single organisms. EXPERIMENTAL SECTION Reagents and Organisms. TMSH was purchased as a 0.25 M methanolic solution from Fluka (Sigma-Aldrich Co. Ltd., (17) Kossa, W. C.; Machee, J.; Ramachandran, S.; Webber, A. J. J. Chromatgr. Sci. 1979, 17, 177. (18) Challinor, J. M. J. Anal. Appl. Pyrolysis 2001, 61, 3. (19) Challinor, J. M. J. Anal. Appl. Pyrolysis 1996, 37, 185. (20) Jun-Kai, D.; Wei, J.; Tian-Zhi, Z.; Ming, S.; Xiao-Guang, Y.; Chui-Chang, F. J. Anal. Appl. Pyrolysis 1997, 42, 1. (21) Yamauchi, K.; Tanabe, T.; Kinoshita, M. J. Org. Chem. 1979, 44, 638. (22) Butte, W. J. Chromatogr. 1983, 261, 142. (23) Zapf, A.; Stan, H.-J. J. High Resolut. Chromatogr. 1999, 22, 83. (24) Blokker, P.; Pel, R.; Akoto, L.; Brinkman, U. A. T.; Vreuls, R. J. J. J. Chromatogr., A 2002, 959, 191. (25) Ishida, Y.; Wakamatsu, S.; Yokoi, H.; Ohtani, H.; Tsuge, S. J. Anal. Appl. Pyrolysis 1999, 49, 267.

Analytical Chemistry, Vol. 75, No. 22, November 15, 2003

6057

Table 1. Fatty Acid Compositions of F. candida Expressed as Percentage of Total Fatty Acids individual F. candidaa fatty acid

1

2

3

4

5

meanb

F. candidac

C14:0 C16:0 C16:1 C18:0 C18:1(n-9) C18:1(n-7) C18:2 C20:4 C20:5

1.6 21.9 10.7 12.6 41.3 5.6 0.9 4.0 1.6

1.5 19.9 12.6 10.7 38.2 5.8 1.4 7.2 2.8

2.8 25.2 4.0 20.1 21.0 2.9 5.9 13.6 4.6

1.5 16.9 7.3 15.9 24.9 4.6 1.5 18.1 9.3

2.6 23.9 14.5 8.7 39.3 2.9 1.2 4.0 3.0

2.0 (0.4) 21.6 (3.3) 9.8 (4.2) 13.6 (4.5) 27.5 (10.9) 4.4 (1.4) 2.2 (2.1) 9.4 (6.3) 4.2 (3.0)

1.7 20.9 6.9 20.8 30.9 3.5 1.6 9.0 4.6

a FAMEs obtained by THM. b Mean values of five individuals with associated standard deviations. c FAMEs obtained by conventional wet chemical procedures applied to solvent extract of 15 F. candida.

Table 2. Effect of the KIE on the Predicted FAME δ13C Values for Different Fatty Acids fatty acid

δ13C (‰)

meas FAME δ13C (‰)a

predicted FAME δ13C (‰)b

∆ (‰)c

KIEd

C14:0 C15:0 C16:0 C18:0 C19:0

-26.6 -26.9 -29.1 -28.6 -31.0

-30.5 (0.2) -30.5 (0.1) -32.2 (0.2) -31.6 (0.2) -33.6 (0.3)

-26.6 -26.9 -29.0 -28.5 -30.8

3.9 3.6 3.2 3.1 2.8

1.059 1.058 1.055 1.059 1.056

a Values are means with associated standard deviations and n ) 10. b Calculated from rearranged eq 1, where δ13C of TMSH ) -27.1‰. c (Measured FAME δ13C) - (predicted FAME δ13C). d KIE ) 1 + (∆ncd)/1000x.

Dorset, U.K.). The authentic compounds myristic acid (C14:0), pentadecanoic acid (C15:0), palmitic acid (C16:0), stearic acid (C18:0), oleic acid (C18:1, (Z)-9), linoleic acid (C18:2, (Z,Z)-9,12), nonadecanoic acid (C19:0), arachidonic acid (C20:4, (Z,Z,Z,Z)-5,8,11,14), tristearin (C18:0), triolein (C18:1, (Z)-9), L-R-phosphatidylcholine, distearoyl (C18:0), and dioleoyl (C18:1, (Z)-9) were supplied by Sigma-Aldrich Co. Ltd. (Dorset, U.K.); cholesteryl stearate (C18:0) and oleate (C18:1, (Z)-9) were supplied by Lancaster Synthesis Ltd. (Lancashire, U.K.). All compounds were dissolved in HPLC grade chloroform to produce stock solutions of 1 mg mL-1. The δ13C values of the saturated fatty acyl compounds had previously been determined by IRMS (Table 2) and were used to determine any KIE associated with the THM reaction. The Collembola F. candida was cultured in plastic containers on an absorbent base comprising of plaster of Paris and activated charcoal.26 The cultures were maintained in the dark at 20 °C and fed continually on isotopically labeled yeast (Saccharomyces cerevisiae), cultured using either beet sugar (“C3” yeast; δ13Csugar, -24.5‰) or cane sugar (“C4” yeast; δ13Csugar, -10.7‰). Preparation of FAMEs. Approximately 1 µL of a chloroform solution containing 1 µg of authentic lipid was added to a glass capillary pyrolysis tube containing glass wool and dried at room temperature to allow the solvent to evaporate. After complete drying, 1 µL of 0.25 M TMSH in methanol was applied to the glass wool, and the solvent was evaporated before insertion into the pyroprobe inlet. The THM reaction was performed at 350 °C (26) Snider, R. Rev. Ecol. Biol. Soil 1972, 10, 103.

6058

Analytical Chemistry, Vol. 75, No. 22, November 15, 2003

for 10 s after optimization. For the determinations of fatty acids in F. candida, single organisms were added to the capillary tube, and 1 µL of TMSH was applied directly onto the organism; identical pyrolysis conditions were applied. To obtain reference fatty acid compositions for the F. candida, conventional wet chemical procedures were employed to prepare FAMEs similar to those described by Hamilton and Hamilton.27 Batches of Collembola samples (∼15 individuals) were extracted with dichloromethane/methanol (2:1 v/v; 3 × 2 mL; ultrasonication). Solvent was removed from extracts under a gentle stream of nitrogen, and lipid extracts were stored at 4 °C until required. Total lipid extracts were saponified with 0.5 M sodium hydroxide in methanol (70 °C, 90 min), acidified to pH 2, extracted with dichloromethane (3 × 2 mL), and passed through an anhydrous sodium sulfate column to remove residual water. The “acidic” fraction was separated using an extraction cartridge packed with an aminopropyl bonded phase, eluting with a 2% acetic acid solution in diethyl ether. Fatty acids were methylated using a boron trifluoride/methanol complex (14% w/v, 30 µL, 70 °C, 60 min), and the reaction was quenched with water. The resultant FAMEs were extracted into hexane (3 × 2 mL), and the solvent was removed under a gentle stream of nitrogen. After evaporation of the solvent, the FAMEs were redissolved in hexane prior to analyses by GC, GC/MS, and GC/C/IRMS. Instrumental Analysis. Py-GC/MS analyses were performed using a CDS (CDS Analytical, Oxford, PA) 1000 Pyroprobe connected to a Carlo Erba (Milan, Italy) 4130 GC coupled to a Finnigan MAT (San Jose, CA) 4500. THM products were separated using a SGE BPX-70 capillary column of 50-m length and 0.32-mm i.d. coated with a 70% cyanopropyl polysilphenylenesiloxane stationary phase (0.25-µm film thickness). Following an initial isothermal hold for 5 min at 50 °C, the oven temperature was increased to 220 °C at 5 °C min-1. The MS was operated in full scan electron ionization mode (m/z 35-650 Da at 1 scan s-1, electron energy 70 eV, filament current 0.35 mA, source temperature 280 °C). A pyrolysis temperature of 350 °C was used, and the carrier gas was helium. Isotopic determinations were performed with the CDS 1000 Pyroprobe connected to a Varian 3400 GC fitted with the capillary column described above, coupled to a Finnigan MAT Delta S isotope ratio monitoring mass spectrometer (electron ionization 100 eV; 3 Faraday cup collectors m/z 44, 45, and 46) via a Finnigan MAT combustion interface (CuO/Pt combustion reactor set to 850 °C). Chromatographic and pyrolysis conditions were as described above. δ13C values were calibrated against reference CO2 of known isotopic composition, previously calibrated against PDB, introduced directly into the source three times at the beginning and end of every isotopic determination. GC analyses were performed on a Carlo Erba 5160 HRGC with FID detection. Samples prepared by conventional wet chemical procedures were dissolved in hexane, and 1 µL was injected on column. All other chromatographic conditions were as described above. Compound identification by GC/MS was performed on a Thermoquest Finnigan Trace GC/MS using the GC conditions described above. Ionization was by EI at 70 eV, and the quadrupole analyzer mass range was scanned from m/z 50 to 650 at 1 scan s-1. GC/C/IRMS analyses were performed on the instrumentation (27) Hamilton, R. J.; Hamilton, S. Lipid AnalysissA Practical Approach; Oxford University Press: Oxford, U.K., 1992; Chapter 2.

described above but excluding the pyroprobe, using 1-µL injections taken from a sample prepared by conventional wet chemical procedures; all other chromatographic and mass spectrometric conditions were as above. RESULTS AND DISCUSSION The overall aim of this investigation was to validate a py-GC/ C/IRMS method for the determination of the δ13C values of fatty acids in individual soil mesofauna for the purposes of ecological research. The method development proceeded in three steps: (i) assessment of the effectiveness of the THM reaction by py-GC/ MS for the generation of FAMEs from a range of key acyl lipids in both their pure form and from the complex biochemical matrix of an individual soil mesofauna, (ii) coupling of the THM reaction with GC/C/IRMS and testing its ability to generate reproducible δ13C values of fatty acids, and (iii) assessment of a KIE associated with the TMSH derivatization reaction and calculation of correction factors such that meaningful δ13C values could be reported. Py-GC/MS. Initially, the effectiveness of TMSH as a derivatizing agent for the THM reaction was evaluated by using solutions of commonly occurring saturated and unsaturated free fatty acid standards. Typical partial gas chromatograms for the methyl ester derivatives of the standards, obtained by TMSH treatment are shown in Figure 1a and b. Intense, fully resolved GC peaks were achieved, which are essential for isotopic determinations. Furthermore, when polyunsaturated fatty acids were derivatized, no additional peaks, potentially arising from isomerization or degradation products, were observed, and peak intensities were comparable to those obtained for saturated fatty acid standards present in the stock solution in similar concentrations (Figure 1b). Subsequently, standard triacylglycerols, phospholipids and steryl esters were also subjected to TMSH treatment, and their fatty acyl moieties were converted quantitatively to FAMEs in a singlestep THM reaction. These results confirm the capabilities of TMSH as an extremely effective reagent for the methylation of the major classes of acyl lipids that are present in soil mesofauna. The ability of the THM reaction to generate a fatty acid signature from individual soil mesofauna was investigated using reaction conditions identical to those employed above for the authentic compounds. In this part of the study, individual F. candida (average weights ∼40 µg), which had been raised on isotopically labeled S. cerevisiae, were treated with TMSH, and their fatty acid profile was determined by py-GC/MS. To provide reference fatty acid compositions and distributions, F. candida from the same colony were extracted, saponified, and derivatized using the conventional wet chemical procedures described previously, followed by GC and GC/MS analysis using chromatographic conditions comparable to those for py-GC/MS. Partial chromatograms from both methods are shown in Figure 2a and b. Eluting components were identified on the basis of their mass spectral characteristics, by comparison to standards, and with reference to data from a previous investigation performed in this laboratory on the lipid composition of F. candida raised on an identical diet.28 The chromatogram of the FAMEs obtained by THM of an individual F. candida shows all the characteristics of that obtained from a larger number of individuals by conventional wet chemical and GC procedures. However, Figure 2a and b also (28) Chamberlain, P. Ph.D. Thesis, University of Bristol, Bristol, U.K., 2001.

Figure 1. Partial py-GC/MS chromatograms of (a) authentic saturated fatty acid standards and (b) authentic unsaturated fatty acid standards, derivatized to FAMEs by THM with TMSH.

illustrates a degree of variation in the fatty acid distributions, which may be a result of individual biological variation within a species. Table 1 shows the fatty acid compositions for five individual F. candida obtained by the THM reaction with TMSH. A variation of the fatty acid distributions is observed, and the most significant differences are exhibited by F. candida 3 and 4. This degree of variation could not previously have been determined by conventional wet chemical methods because of the number of F. candida that would have to be combined for analysis. However, the mean values of the fatty acid compositions of the five individual F. candida proved to be comparable to the compositions obtained by conventional wet procedures values, and no significant difference at 95% confidence limits is observed when a Student’s t-test is performed. These results demonstrate that fatty acid compositions can be readily achieved via a THM reaction employing TMSH and further highlights the advantages of the methodology compared to conventional wet chemical procedures. Kinetic Isotope Effect. Fatty acids are polar compounds that must be converted to less polar derivatives, typically FAMEs, to improve their GC behavior. However, derivatization prior to compound-specific isotope analysis requires careful consideration, because it will significantly affect the result of the isotopic Analytical Chemistry, Vol. 75, No. 22, November 15, 2003

6059

Scheme 1. Proposed THM of Free Fatty Acids with TMSH

Scheme 2. Proposed THM of Triacylglycerols with TMSH

Figure 2. Partial chromatograms of the fatty acids of F. candida obtained by (a) THM with TMSH and (b) conventional wet chemical procedures.

δ13C

determinations. The most common effects are changes on the values of the target analyte due to carbon present in the reagent used for derivatization; such effects are readily corrected for by mass balance calculations.29 In addition, fractionation effects, such as KIEs, can cause alterations to the δ13C values of derivatized compounds. Such effects need to be identified and the δ13C values corrected for accordingly. Consequently, a thorough treatment of the precision of the derivatization procedure and the mass balance equations employed is essential in order to obtain meaningful δ13C values and to prevent generalized interpretation of results based only on the precision of the analyses. KIEs occur because of the differences in the rate of transport or rate of reaction of isotopic species. In addition, bonds involving heavier isotopes have higher potential energies; therefore, when different isotopic species are involved, different rates of reaction are observed.30 The primary isotope effect is the most significant and arises as a result of a change in a bond at a specific position during the rate-determining step of the derivatization reaction. A KIE normally occurs on one of the added derivative carbons when a bond involving that carbon is broken and the reaction involving that carbon atom is nonquantitative, causing a fractionation of carbon isotopes at a specific position in the derivatized compound and, subsequently, an alteration to the δ13C value. Rieley31 (29) Jones, D. M.; Carter, J. F.; Eglinton, G.; Jumea, E. J.; Fenwick, C. S. Biol. Mass Spectrom. 1991, 20, 641. (30) Melander, L. G. S.; Saunders, W. H. Reaction Rates of Isotopic Molecules; Wiley: New York, 1980. (31) Rieley, G. Analyst 1994, 119, 915.

6060

Analytical Chemistry, Vol. 75, No. 22, November 15, 2003

presented a thorough treatment of the kinetic isotope effects associated with the derivatization of a wide range of organic compounds in relation to their preparation for GC/C/IRMS analysis. If no KIE is associated with the derivatization reaction, the exogenous carbon atoms can be corrected for using a mass balance equation,

ncdδ13Ccd ) ncδ13Cc + ndδ13Cd

(1)

where n is the number of moles of carbon, subscript c represents the compound of interest, d the derivative group, and cd the derivatized compound. A KIE prevents the δ13C determination of the compound of interest by this method. Potential isotopic fractionations must be investigated prior to sample analysis. Close inspection of the mechanism for the derivatization reaction should highlight any potential KIEs. Schemes 1 and 2 show the proposed THM reaction using TMSH for the derivatization of free (Scheme 1) and esterified (Scheme 2) fatty acids. Since a bond to the methylating carbon is broken in the reaction intermediate and is remade in forming the final derivative, a KIE would be expected at this methyl carbon. To investigate the potential KIE associated with the TMSH methylation of fatty acids, the δ13C values of FAMEs prepared from standard compounds of established isotope values were determined by py-GC/C/IRMS and compared with the δ13C values predicted for those compounds from the mass balance eq 1. Table 2 shows that δ13C values of FAMEs determined by py-GC/C/

IRMS are between 2.0 and 4.0‰ lower than those predicted by mass balance, thus confirming that an isotopic fractionation has arisen during the addition of the derivative carbon as a result of a KIE. The magnitude of the KIE can be determined from the following equation,

KIE ) 1 +

∆ncd 1000x

(2)

where ∆ is the difference between the measured δ13C and the predicted δ13C, and x is the number of carboxyl groups available for methylation. Table 2 lists the KIE calculated for each of the fatty acid derivatizations with TMSH. The mean KIE is calculated as 1.057 (SD ) 0.002), demonstrating very little variation in the magnitude of the KIE between the different fatty acids. Rieley31 reported a KIE of 1.019 (SD ) 0.03) associated with the trifluoroacetyl carbon in the trifluoroacetate/isopropyl ester derivatives of amino acids. A wider range of individual isotope effects was observed than for the fatty acids in this study. It was hypothesized that the variation was due to substituent effects, different degrees of reaction, or possible systematic errors that are clearly not evident in the TMSH reaction because of the close similarities of the fatty acid structures compared to amino acids. Since the preparation of the methyl ester derivatives has the potential to cause variations in the measured isotope values, the derivatization procedure must be shown to be reproducible. In a study involving the acetylation of carbohydrates, Macko et al.32 demonstrated that by using the same batch of reagents, amounts of reagents, and identical reaction conditions, the fractionation associated with acetylation of carbohydrates was indeed reproducible. A similar approach was adopted herein, with 10 separate derivatizations being performed on identical aliquots of the same mixture of authentic fatty acids. Hence, each of the reported values in Table 2 is the mean of 10 individual measurements. The respective standard deviations were never greater than 0.3‰, thereby confirming the high degree of reproducibility for the THM derivatization reaction using TMSH. An isotope fractionation effect is generally expressed in the derivatized compound as a result of a nonquantitative reaction of the reagent. Varying the amount of analyte present may therefore affect the extent to which the KIE is expressed. This was also investigated by diluting the standard solutions of authentic fatty acid to give a range of concentrations, aliquots of which were then derivatized under identical reaction conditions using identical reagent volumes and heating conditions. The results of this experiment presented graphically in Figure 3 (mean values of triplicate measurements) show that there is no significant variation between the amount of derivatized fatty acid and the corresponding δ13C values. Correction Factors. Since the KIE associated with the THM derivatization using TMSH has been shown to be reproducible and robust for the preparation of FAMEs, the KIE can be corrected for by the use of correction factors. The correction factor can be defined as the “effective” stable isotope composition of the derivative carbon introduced during derivatization, taking into account the isotopic fractionation associated with the reaction.32 (32) Macko, S. A.; Ryan, M.; Engel, M. H. Chem. Geol. 1998, 152, 205.

Figure 3. Variation in the expression of the KIE during TMSH derivatization with varying amounts of authentic fatty acid standards. (Plotted values have not been corrected for the addition of extra carbon.) Table 3. Correction Factors Determined in Order to Calculate the Original Fatty Acid δ13C Value fatty acid

correction factor δ13Ccorr (‰)a

error ((‰)b

C14:0 C15:0 C16:0 C18:0 C19:0

-85.6 -85.1 -82.0 -86.0 -83.0

0.5 0.5 0.5 0.5 0.5

a Calculated from rearranged eq 1, where δ13C ) δ13C b d corr. Calculated from propagation of errors.

These are determined indirectly by measuring the δ13C value of an underivatized standard of the compound of interest (by IRMS), the δ13C value of the standard after derivatization using TMSH (by py-GC/C/IRMS), and using a rearranged eq 1 to determine δ13Cd. The δ13Cd value calculated represents the corresponding correction factor for the compound of interest. The correction factors determined for each of the fatty acids targeted in this study are listed in Table 3. These values can be used with a rearranged eq 1 to calculate the original δ13C value of a fatty acid within a sample. However, Docherty et al.33 showed how the use of such correction factors can result in substantial analytical imprecision arising from the analysis of small polyfunctional compounds, which places constraints on how the resultant calculated δ13C values can be interpreted. Errors associated with the correction factors applied must therefore be quantified to prevent generalized interpretations of results that are beyond what can be justified on the basis of the precision of the instrumental analyses. The instrumental precision for the δ13C determinations by py-GC/C/ IRMS is (0.3‰ (the intrinsic precision of the instrument), and the errors associated with the correction factors were calculated as described by Docherty et al.33 and shown to be (0.5‰ for each fatty acid standard. Hence, there is only a slight reduction (relative to the instrument precision) in the precision of the analysis when correction factors are utilized in the mass balance equation for determination of the δ13C value of the fatty acid. When the corrected δ13C FAME values obtained by TMSH derivatization and py-GC/C/IRMS were compared directly to the same underivatized fatty acids analyzed by IRMS, the values were accurate to well within their associated precisions. (33) Docherty, G.; Jones, V.; Evershed, R. P. Rapid Commun. Mass Spectrom. 2001, 15, 730.

Analytical Chemistry, Vol. 75, No. 22, November 15, 2003

6061

those of the diets on which they were raised and show excellent agreement within the associated error to the δ13C values determined by conventional wet chemical procedures. This was confirmed by using a Student’s t-test to compare the δ13C values determined by both methods, and no significant difference was observed at 95% confidence limits.

Figure 4. δ13C values of the major fatty acids of C3 F. candida (closed symbols) and C4 F. candida (open symbols). 4, corrected δ13C values of the fatty acids derivatized by THM with TMSH; O, δ13C values of fatty acids extracted and derivatized by conventional wet chemical procedures. All data are mean values with associated standard deviations.

δ13C Analysis of Fatty Acids in Collembola. The δ13C values of fatty acids of 10 individuals of the Collembolan species F. candida, raised on C3 and C4 isotopically labeled S. cerevisiae were determined by py-GC/C/IRMS using TMSH and compared with the values obtained by GC/C/IRMS of FAMEs prepared via wet chemical methods involving solvent extraction, saponification, and methylation using BF3/MeOH. Figure 4 shows a plot of the δ13C values obtained by both methods for the major fatty acids present. The δ13C values obtained by py-GC/C/IRMS have been corrected for the KIE shown to be associated with the TMSH derivatization reaction, using the correction factors calculated for the authentic fatty acids presented in Table 3. F. candida saponified and derivatized by wet chemical procedures were analyzed by GC/ C/IRMS. In this instance, each sample was run in duplicate, and the mean value was reported. Because there is no KIE associated with this methylation reaction, the contribution of the derivatizing carbon added by this method was corrected for by mass balance (eq 1). Corrected isotopic values of the TMSH derivatized fatty acids for both C3 and C4 isotopically labeled F. candida reflect

6062 Analytical Chemistry, Vol. 75, No. 22, November 15, 2003

CONCLUSIONS The application of a THM reaction in the presence of TMSH with Py-GC/C/IRMS has been shown to be a powerful new tool for the compound-specific δ13C determination of fatty acids in soil mesofauna. The fatty acid profiles and corrected δ13C values show excellent agreement with those obtained by conventional wet chemical procedures. A KIE is associated with the THM reaction; however, it has been proven to be reproducible and robust, thereby the exogenous carbon can be accurately corrected for by the use of correction factors. The magnitude of the error associated with the determined δ13C value of a fatty acid is fully quantified ((0.5‰); therefore, δ13C values of fatty acids isolated from soil mesofauna can be quoted to determined precisions and meaningful interpretations thus achieved. The results presented herein demonstrate for the first time the δ13C values and fatty acid composition of individual soil mesofauna and also the possibility for smaller organisms or parts of individuals to be analyzed, thereby expanding the scope to which isotopic studies can be applied to the investigation of nutrient cycling and trophic behavior of complex invertebrate ecosystems. ACKNOWLEDGMENT NERC is thanked for financial support for mass spectrometry facilities (Grant nos. GR3/2951, GR3/3758, and FG6/36101), and Dr. J. F. Carter and Dr. I. D. Bull are thanked for their help with instrumental analysis. The study was undertaken while C.J.E. was in receipt of a NERC CASE studentship. Received for review April 1, 2003. Accepted August 20, 2003. AC034336D