NMR spectroscopic studies on the haemolymph of the tobacco hornworm,< i> Manduca sexta: assignment of< sup> 1 H and< sup> 13 C NMR

Insect Biochemistry and Molecular Biology 29 (1999) 795–805www.elsevier.com/locate/ibmb

NMR spectroscopic studies on the haemolymph of the tobaccohornworm,Manduca sexta: assignment of1H and 13C NMR

spectra

C. Phalaraksha, E.M. Lenza, J.C. Lindona, J.K. Nicholsona,*, R.D. Farrantb,S.E. Reynoldsc, I.D. Wilson d,e, D. Osbornf, J.M. Weeksf

a Biological Chemistry, Division of Biomedical Sciences, Imperial College of Science Technology and Medicine, Sir Alexander FlemingBuilding, South Kensington, London, SW7 2AZ, UK

b Physical Sciences, GlaxoWellcome Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts SG1 2NY, UKc Department of Biology and Biochemistry, University of Bath, Claverton Down, Bath, BA2 7AY, UK

d Department of Chemistry, University of Keele, Keele, Staffordshire ST5 5BG, UKe Current address. Department of Safety of Medicines, Zeneca Pharmaceuticals, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG,

UKf Institute of Terrestrial Ecology, Monks Wood, Abbots Ripton, Huntingdon, Cambridgeshire PE17 2LS, UK

Received 1 February 1999; received in revised form 26 April 1999; accepted 30 April 1999

Abstract

High resolution1H nuclear magnetic resonance (NMR) spectroscopy has been used to investigate the low molecular weightmetabolite composition of the whole haemolymph of 5th instar larvae of the Tobacco Hornworm,Manduca sexta. This techniqueprovides a rapid, multicomponent profile of a wide range of intermediary metabolites. Assignment of the1H NMR resonances inthe 1-dimensional spectrum has been facilitated by the use of high frequency 2-dimensional NMR techniques, including J-resolvedspectroscopy,1H–1H total correlation spectroscopy and1H–13C heteronuclear correlation spectroscopy. Amongst the biochemicallyimportant metabolic intermediates that were detected in whole haemolymph were trehalose, glucose, alanine, lactate, choline andbetaine. In total, it was possible to simultaneously detect and potentially quantify 19 endogenous metabolites. These studies indicatethat NMR spectroscopy of the whole larval haemolymph may offer the potential for evaluation of whole organism biochemicalstatus in relation to physiological or environmental stress. 1999 Elsevier Science Ltd. All rights reserved.

Keywords:NMR; Haemolymph; Manduca; metabolites

1. Introduction

Insect haemolymph is complex aqueous solution con-taining a wide range of inorganic ions, proteins andorganic acids and with particularly high concentrationsof sugar (trehalose and glucose) (Kerkut and Gilbert,1985), selected amino acids and organic phosphates. Thehaemolymph also contains blood cells, or haemocytesof several types (mainly plasmatocytes, cystocytes andgranular cells, Gullan and Cranston, 1994). Variations

* Corresponding author. Tel+44(171)-594-3195; Fax+44(171)-594-3221E-mail address:[email protected]. (J. Nicholson)

0965-1748/99/$ - see front matter. 1999 Elsevier Science Ltd. All rights reserved.PII: S0965-1748 (99)00053-3

in composition of the haemolymph might reasonably beexpected to reflect changes in insect physiology broughtabout by development and environmental factors. Therehave been several ecotoxicological studies monitoringvariations in lepidopteran larvae haemolymph compo-sition. Turunen (1992), using high performance liquidchromatography (HPLC) and thin-layer chromatography(TLC), showed trehalose to be the major sugar inPierisbrassicae haemolymph. Three dimensional HPLCshowed that tyrosine was present in the fourth and fifthinstar larvae of the silkwormBombyx mori(Takeda etal., 1991), and concentrations increased after exposureto low concentrations of chlordimeform (an insecticide)and clonidine (Shimizu and Takeda, 1991). Paper chro-matography has been used to study the effect of the pes-

796 C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

ticide, carbaryl, on the haemolymph of the mothSpodop-tera litura. In the study of fifth and sixth instar larvaeexposed to carbaryl, cysteine and glutamic acid wereeliminated from the haemolymph, and in the sixth instarand pupae, phenylalanine and aspartic acid werereduced. Concentrations of asparagine, glycine and pro-line were also reduced whereas lysine, threonine, ala-nine, valine and methionine were unaffected (Verma,1991).

Problems with conventional bio-analytical approachessuch as HPLC and other chromatographic techniquesinclude both the length of time and complexity necessaryfor the development of analytical procedures, and thelack of metabolite structural information. In addition,traditional bioanalytical procedures may cause both bio-logical and physiochemical property changes to the sam-ple, and hence the measured biochemical compositionmay differ from that actually occurring in the intactbiomatrix. These disadvantages can be partially over-come by using1H NMR spectroscopy to study complexbiomixtures, as measurements can often be made withminimal sample preparation (usually with only theaddition of 5–10% D2O), and a detailed multicomponentanalytical profile can be obtained on the whole biologicalsample simultaneously and potentially non-destructively(Nicholson and Wilson, 1989).

High resolution1H NMR spectroscopy of biologicalfluids has found widespread application in mammaliantoxicology and drug metabolism studies (e.g. Nicholsonand Wilson, 1989; Wilson and Nicholson, 1995; Holmeset al., 1996; Holmes et al., 1995a,b; Anthony et al., 1994;Bollard et al., 1996; Nicholls et al., 1995; Foxall et al.,1995, 1996; Fan, 1996). In mammals, there are manydifferent functional types of biofluid and these fluids(e.g. blood, urine, bile and cerebrospinal fluid) havediverse physico–chemical and biochemical featuresaccording to their mode of secretion or production. Alsoin mammals, blood serum contains a wide range of lip-ids, lipoproteins, proteins, and a range of low molecularweight (MW) organic and inorganic compounds(Nicholson et al., 1995). Preliminary studies (Wilson etal., 1989; Thompson, 1990) suggested that1H NMRspectroscopy of insect haemolymph was technicallyfeasible.

There have been few NMR spectroscopic studies onthe body fluids of invertebrate species, although thepossibilities for studying physiological and pathologicalprocesses are manifold (Higham et al., 1986).1H, 13Cand 31P NMR spectroscopy has been applied to per-chloric acid extracts of insect haemolymph (Thompson,1990). In addition,31P NMR spectra of the whole insecthas been reported (Thompson, 1990). We present herea detailed1H NMR spectroscopic assignment study onthe haemolymph of the tobacco hornworm (Manducasexta) utilising 500 MHz one- (1D) and two-dimensional(2D) experiments including 2D J-resolved, and 2D

homonuclear (1H–1H) correlation methods and a 750MHz heteronuclear (1H–13C) correlation technique forobtaining13C NMR data as previously applied to assignsignals in mammalian biofluids (Nicholson et al., 1995).

The overall aim of this work is to use NMR spec-troscopy to characterise whole invertebrate biofluids andsubsequently to develop biomarkers (changes in a rela-tively low levels of biochemicals that indicate exposureto, or the effect of, contaminants) as indicators of theenvironmental biochemical status of populations of fieldorganisms (Stamp and Casey, 1993). The biofluid usedhere was not deproteinised as the intention was to studythe intact sample. The necessary first step in this processis to establish baseline conditions for the unperturbedbiochemistry of suitable groups of organisms, and hereresults from the larval stages of a model lepidopteranspecies are described.

2. Materials and methods

2.1. Biological experiments

Manduca sextalarvae were obtained from the Depart-ment of Biology and Biochemistry, University of Bath,Bath, UK. The weight of the two day-old fifth instarlarvae was approximately 5–6 g. The haemolymphsamples (500–700µl) were obtained by exsanguination.Approximately 30 mg of phenylthiourea, a preservativepowder, was added to all haemolymph samples to inhibittyrosinase activity which leads to blackening of thehaemolymph.

2.2. NMR experiments

1H NMR spectra were measured on whole haemo-lymph diluted with 10% deuterium oxide (D2O) to pro-vide a field-frequency lock for the NMR spectrometer.The samples also contained the chemical shift andquantification standard sodium 3-trimethylsilyl-2H4-pro-pionic acid (TSP) at 100µM. One- and two dimensional1H NMR spectra were obtained using a Bruker DRX500NMR spectrometer (Bruker Spectrospin Ltd., Coventry,UK), operating at 500.13 MHz1H observation fre-quency. Typically, one-dimensional NMR spectra wereobtained either in the single pulse mode or using a spin-echo experiment (Carr–Purcell–Meiboom–Gill; CPMG)to remove resonances from macromolecules and otherspecies with short T2 relaxation times (Nicholson andWilson, 1989; Nicholson et al., 1995; Bollard et al.,1996) using 64 transients acquired into 32 768 datapoints covering a spectral width of 8012 Hz at a tem-perature of 300 K. For CPMG spectra, the total spin-spin relaxation delay was 90 ms. Presaturation of theH2O resonance was achieved by replacing the conven-tional 90° pulse in the single pulse experiment or the

797C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

initial 90° pulse in the CPMG experiment with the firstincrement of the Nuclear Overhauser Effect Spec-troscopy (NOESY) pulse sequence irradiating at thewater frequency for three seconds before the first 90°pulse and for 80 ms during the mixing period. A line-broadening of 0.5 Hz was applied before Fourier trans-

Fig. 1. 500 MHz1H NMR spin-echo (CPMG) spectrum of control haemolymph fromManduca sexta, spectrum regiond5.00–8.10 (a),d0.80–5.00 (b) and 500 MHz1H NMR single pulse spectrum of control haemolymph fromManduca sexta, spectrum regiond5.00–8.10 (c),d0.80–5.00(d). Assignments are as marked and as given in Table 1.

formation (FT). Two-dimensional J-resolved (JRES)NMR spectra were acquired using 64 transients collectedfor 64 increments into 8192 data points using a spectralwidth of 8012 Hz. The spectral width in the J-couplingdomain was 64 Hz. A relaxation delay of three secondswas included. These raw data were multiplied by sine-

798 C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

bell squared apodisation functions prior to zero-filling to512 points on the spin-coupling axis.1H–1H total corre-lation spectroscopy (TOCSY) spectra were acquiredusing 16 transients collected for 256 increments into4096 data points per increment over a spectral width of8012 Hz. The spin-lock was achieved using the MLEV-17 sequence for a period of 60 ms. A relaxation delayof two seconds was included and the water peak wassuppressed as described above. The data were multipliedby sine-bell squared functions in both domains beforeFT. 1H–13C heteronuclear correlation spectra were

Fig. 1. (continued).

acquired at 750 MHz using the Heteronuclear MultipleQuantum Coherence (HMQC) technique on a VarianInova spectrometer (Varian Instruments, Palo Alto,USA). Data were acquired using 512 transients collectedfor 204 increments into 1472 data points over a spectralwidth of 7273.4 Hz for1H and 34 144.3 Hz for13C.Broad band13C decoupling was carried out by the stan-dard GARP method (Shaka et al., 1985). These acquireddata were multiplied by an exponential factor equivalentto a line broadening of 1 Hz and extended by linear pre-diction to 4096 points using 512 coefficients. After the

799C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

Fig. 2. 500 MHz1H JRES NMR spectrum of control haemolymph fromManduca sexta, spectrum regiond4.60–5.25 (a),d3.00–4.10 (b),d2.10–2.80 (c) andd0.90–1.80 (d). Assignments are as marked and as given in Table 1.

800 C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

Fig. 2. (continued).

first Fourier transform the interferograms were extendedby linear prediction to 2048 points using 124 coef-ficients.

3. Results

The 1D single pulse and CPMG spin-echo 500 MHz1H NMR spectra of control haemolymph from a twoday-old larva of 5th instarManduca sextaare shown in

801C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

Fig. 3. 500 MHz1H–1H TOCSY NMR spectrum of control haemolymph fromManduca sexta. Assignments are as marked and as given in Table 1.

Fig. 1. There is little difference between the conventionaland the CPMG spectra consistent with the low concen-tration of macromolecules in the sample. Smalldecreases in intensity occurred to the citrate, lysine andtrehalose resonances in the CPMG spectrum when com-pared with the single pulse spectrum. The sharp lines inthe spectra are mainly from low MW substances manyof which could be assigned by reference to publisheddata (Nicholson et al., 1995). The assignments, parti-cularly in the very crowded region of the spectrumbetweend3–4, were aided by the measurement of JRESand TOCSY spectra. JRES spectroscopy results in thespin-coupled multiplets being displayed orthogonally tothe chemical shift axis, thereby reducing spectral overlapand this is viewed as a contour plot.1H–1H TOCSYspectroscopy displays the connectivities of protons thatare coupled in an unbroken chain. This is also presentedas a contour plot with the normal spectrum on the diag-onal and protons which belong to an unknown chain ofspin-spin coupling showing off-diagonal cross peaks.The JRES and TOCSY spectra of the same sample areshown in Figs. 2 and 3 with assignments as marked. Theassignments have been augmented by the measurementof two-dimensional1H–13C HMQC spectra, also viewedas a contour plot, which provide correlations between

the 1H and 13C chemical shifts of directly attached1Hand 13C nuclei. The1H–13C HMQC spectrum is shownin Fig. 4.

Identification of the resonances resulting from thepresence of a number of amino acids can be ambiguouslyachieved only by comparison of 1D and 2D data. Theseinclude the two methyl doublets of valine atd0.99 andd1.05 which are visible in the 1-dimensional, JRES,TOCSY and HMQC spectra coupled to theβ-CH atd2.29 in the TOCSY spectrum. Theγ-CH3 of isoleucinewas assigned atd1.02 visible in the 1D, JRES andTOCSY spectra and this was coupled to the signal atd1.99 (TOCSY). Theδ-CH3 of leucine visible in theJRES spectrum was also coupled to theβ-γ CH2CH atd1.73 (TOCSY). The methyl doublet of alanine atd1.48in the 1D, JRES, TOCSY and HMQC spectra wasassigned through coupling to theα-CH at d3.78 visiblein the TOCSY spectrum. The amino acid lysine gives aseries of distinctive resonances atd3.03 (e-CH2), d1.76(δ-CH2), d1.51 (γ-CH2), d1.91 (β-CH2), all easilymapped using the JRES spectra and the TOCSY connec-tivities. Glycine gives rise to a singlet atd3.60 visiblein the 1D and JRES spectra. Prominent resonances canbe observed from glutamate with theγ-CH2 at d2.46 andthe β-CH2 at d2.14 in 1D, JRES and TOCSY spectra

802 C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

Fig. 4. 750 MHz1H–13C HMQC NMR spectrum of control haemolymph fromManduca sexta,spectrum region1H d0.7–3.8,13C d15–52 (a);1Hd3.0–4.4,13C d50–85 (b),1H d4.55–5.3,13C d92–104 (c) and1H d6.8–8.1,13C d115–140 (d). Assignments are as marked and as given in Table 1.

with a connectivity to theα-CH atd3.76 in the TOCSYspectrum. Some minor resonances, which were detectedat d2.01 (β-CH2), d2.35 (β9-CH2) could be identified dueto connectivities in the TOCSY spectrum, and were con-sistent with the presence of proline. Resonances from theamino acid tyrosine could be seen in the aromatic region

of the 1D NMR spectrum atd6.90 andd7.20 (H3, H5

and H2, H6, respectively).A number of organic acids were also assigned. The

lactate methyl doublet was detected atd1.33 and iscoupled to theα-CH at d4.13 in the 1D and TOCSYspectra. A small doublet resonance consistent with 3-

803C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

hydroxybutyrate was observed atd1.20 in the 1D spec-trum. Succinate gave a singlet atd2.41 visible in the1D and JRES spectra. Citrate gives a characteristic ABdoublet pattern which was visible atd2.54 andd2.69 inthe 1D spectra.

High concentrations of carbohydrates were alsoobserved. Thus many resonances forα- and β-glucosecan be assigned using the JRES, TOCSY and HMQCspectra. In addition, the H1 signal of trehalose appearsas a doublet atd5.14 and this shows a TOCSY con-nectivity to H2 at d3.61. A further carbohydrate with aspin coupling typical of theα-configuration is seen atd5.24 and this remains unidentified at present.

Several nitrogen containing molecules were visible inthe NMR spectra, with dimethylamine providing acharacteristic singlet atd2.74 visible in the 1D and JRESspectra, and betaine giving rise to two singlets in the 1Dspectrum (atd3.27 andd3.85 corresponding to the N-Me and CH2 resonances, respectively). Choline was alsopresent and gave a prominent singlet atd3.22 from the2NMe+

3 group. The two coupled CH2 groups were vis-ible in the TOCSY spectrum atd3.55 (N-CH2) andd4.06(O-CH2), respectively. A multiplet was observed in the1H NMR spectrum atd4.19 that was coupled to anotherpeak atd3.59 as shown by the TOCSY spectrum. Theseresonances could arise from glycerophosphorylcholine.

Having assigned the1H NMR chemical shifts of manyof the endogenous low MW metabolites (as given inTable 1) a 750 MHz1H–13C two-dimensional correlationspectrum was also obtained on the intact haemolymph.This is shown in Fig. 4 and enabled the assignment ofthe 13C NMR chemical shifts of many species, whichserved to confirm the identification of the individualmetabolites. The13C chemical shifts are also given inTable 1.

4. Discussion

High resolution1H NMR spectroscopy has shown thatthe whole haemolymph of the tobacco hornwormMand-uca sextacontains a highly complex mixture of metab-olites (Figs. 1–4). Despite the high degree of signal over-lap observed in most of the spectra, many resonancescould be assigned (particularly with the aid of 2-dimen-sional methods).

The range of low molecular weight substancesdetected inManduca sextalarvae by NMR was similarto that found in the larvae of other species that had notbeen exposed to contaminants (Verma, 1991; Takeda etal., 1991; Shimizu and Takeda, 1991) and that had beenexamined by traditional methods of bioanalysis. How-ever, in the present study, it was found that the glucoseconcentration in the haemolymph was higher than thatof trehalose. This is in contrast to previous studies(Thompson 1990, 1998; Thompson and Borchardt,

1996). As trehalose is chemically stable, it could be sur-mised therefore that enzymes in the haemolymph causehydrolysis of the trehalose.

Thus, a non-selective, non-destructive analytical tech-nique requiring little sample pretreatment, NMR, hasproduced data equivalent to that obtained with a wholesuite of more labour-intensive techniques (e.g. HPLC,TLC, paper chromatography) that involve extensivemethodological development, sometimes for singledeterminant assays. Typical data collection times persample for 1D1H NMR spectra are in of the order offive minutes whilst1H–1H TOCSY spectra can requireas little as one to two hours.

A technique as flexible and powerful as NMR shouldfind wide application in ecotoxicological studies, wherethe need for a rapid system of non-selective detection isessential if the effects of environmental contaminants areto be identified and their mechanisms of action deter-mined. NMR should prove a useful tool in the develop-ment of biomarkers, or combinations of biomarkers, thatare specific to certain groups of contaminant or toxicant,because it provides a set of fingerprints of toxicologicalresponses that can be related to ecologically relevantphysiological or biochemical end-points. This shouldhelp free ecotoxicological studies from the current over-dependence on the use of lethal end-points. Some otherpreliminary studies, in other invertebrate species, of theuse of NMR spectroscopy in this way have already beenreported (Gibb et al., 1997a,b).

NMR spectroscopy is a powerful method for providinga multivariate fingerprint of a complex mixture such ashaemolymph. Especially in situations where the identity ofthe components is not known beforehand, NMR spectracan be used in an exploratory fashion. This contrastswith chromatographic and mass spectrometricapproaches where it is usual to develop an assay for aspecific substance. As a consequence, these latter tech-niques can prove to be more sensitive than NMR. How-ever, using appropriate internal standards all techniquescan be considered quantitative. Two-dimensional NMRmethods are powerful approaches for the assignment ofNMR resonances and hence identification of metabolites.On the other hand two-dimensional NMR spectra are dif-ficult to use for quantitative purposes.

In this case,1H NMR spectroscopy has provided aunique and characteristic fingerprint of endogenousmetabolites for a species in its basal metabolic state andat a particular stage of development. As a wide range ofbiochemically important intermediary pathways can bemeasured simultaneously by1H NMR spectroscopy, thepotential for the detection of biochemical perturbationsis high. The establishment of biomarkers may be aidedby the application of computational pattern recognitiontechniques which have been extensively developed forbiomarker exploration in vertebrate species (Gartland etal., 1991; Anthony et al., 1994). Comprehensive assign-

804 C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

Table 1Assignments in the1H NMR spectra of haemolymph fromManduca sextaa

Molecule 1H shift (d) Multiplicity Assignment 13C shift(d)*

leucine 0.96 d δ-CH3 19.5valine 0.99 d γ-CH3 19.6isoleucine 1.02 d γ-CH3

valine 1.05 d γ9-CH3

3-hydroxybutyrate 1.20 d CH3lactate 1.33 d β-CH3 21.9alanine 1.48 d β-CH3 18.6lysine 1.51 m γ-CH2 23.9leucine 1.73 m β-CH2 42.8lysine 1.76 m δ-CH2 26.2lysine 1.91 m β-CH2

isoleucine 1.99 m β-CH2

proline 2.01 m β-CH2

glutamate 2.14 m β-CH2 28.7valine 2.29 m β-CHproline 2.35 m β9-CH2

succinate 2.41 s CH2glutamate 2.46 m γ-CH2 33.5citrate 2.54 d 1/2-CHcitrate 2.72 d 1/2-CHdimethylamine 2.74 s CH3lysine 3.03 m e-CH2 41.2choline 3.22 s CH3 56.5β-glucose 3.24 t H2 77.0betaine 3.27 s CH3 56.1β-glucose 3.40 m H4 72.1α-glucose 3.45 dd H4 72.1β-glucose 3.47 ddd H5 78.6β-glucose 3.49 m H3 78.6α-glucose 3.53 dd H2 74.1choline 3.55 t NCH2glycine 3.60 s CH2trehalose 3.61 m H2 64.8β-glucose 3.72 dd H6 63.1glutamate 3.76 t α-CHα-glucose 3.76 m H6 63.1alanine 3.78 q α-CH 16.8betaine 3.85 s CH2α-glucose 3.88 m H69 63.1β-glucose 3.95 dd H69 63.1choline 4.06 t OCH2lactate 4.13 q α-CH 71.2GPC 4.19 t NCH2GPC 4.59 t OCH2β-glucose 4.65 d H1 98.5trehalose 5.14 t H1 102.8α-glucose 5.20 d H1 94.6tyrosine 6.90 d H3,H5 118.2tyrosine 7.20 d H2,H6 133.1phenylthiourea 7.33 d H2phenylthiourea 7.41 t H3phenylthiourea 7.51 t H4

a s=singlet; d=doublet; t=triplet; q=quartet; m=multiplet; dd=doublet of doublets; ddd=doublet of doublets of doublets; *blank absent in HMQC.

805C. Phalaraksh et al. / Insect Biochemistry and Molecular Biology 29 (1999) 795–805

ment of resonances is a prerequisite for any such1HNMR spectroscopic investigations.

Acknowledgements

We thank the Royal Thai Government for provisionof a postgraduate studentship (to C. Phalaraksh) andNERC for financial support.

References

Anthony, M.L., Lindon, J.C., Beddell, C.R., Nicholson, J.K., 1994.Pattern recognition classification of the site of nephrotoxicity basedon metabolic data derived from high resolution proton nuclear mag-netic resonance spectra of urine. Molecular Pharmacology 46,199–211.

Bollard, M.E., Holmes, E., Blackledge, C.A., Wilson, I.D., Lindon,J.C., Nicholson, J.K., 1996. High resolution1H and19F NMR stud-ies on the metabolism of substituted phenols in the rat. Xenobiotica26, 255–274.

Fan, T.W.-M., 1996. Metabolite profiling by one- and two-dimensionalNMR analysis of complex mixtures. Progress in NMR spec-troscopy 28, 161–219.

Foxall, P.J.D., Bewley, S., Neild, G., Rodeck, C.H., Nicholson, J.K.,1995. Analysis of fetal and neonatal urine using proton NMR spec-troscopy. Archives of Disease in Childhood 73, 153–157.

Foxall, P.J.D., Lenz, E.M., Neild, G.H., Lindon, J.C., Wilson, I.D.,Nicholson, J.K., 1996. Nuclear magnetic resonance and high per-formance liquid chromatography-nuclear magnetic resonance spec-troscopic studies on the toxicity and metabolism of ifosfamide.Therap. Drug Mon. 18, 498–505.

Gartland, K.P.R., Beddell, C.R., Lindon, J.C., Nicholson, J.K., 1991.Application of pattern-recognition methods to the analysis andclassification of toxicological data derived from proton nuclearmagnetic resonance spectroscopy of urine. Molecular Pharma-cology 39, 629–642.

Gibb, J.O.T., Holmes, E., Nicholson, J.K., Weeks, J.M., 1997a. ProtonNMR spectroscopic studies on tissue extracts of invertebrate spec-ies with pollution indicator potential. Comparative Biochemistryand Physiology B 118, 587–598.

Gibb, J.O.T., Swendsen, C., Weeks, J.M., Nicholson, J.K., 1997b.1HNMR spectroscopic investigations of tissue metabolite biomarkersresponse to Cu(II) exposure in terrestrial invertebrates: Identifi-cation of free histidine as a novel biomarker of exposure to Cu inearthworms. Biomarkers 2, 295–302.

Gullan, P.J., Cranston, P.S., 1994. The Insect: An Outline of Ento-mology. Chapman and Hall, London.

Higham, D.P., Nicholson, J.K., Overnell, J., Sadler, P.J., 1986. NMRstudies of crab and plaice metallothioneins. Environmental HealthPerspectives 65, 157–165.

Holmes, E., Bonner, F., Nicholson, J.K., 1995a. Comparative studieson the nephrotoxicity of 2-bromoethanamine hydrobromide in theFisher 344 rat and the multimammate desert mouse (Mastomysnatalensis). Archives of Toxicology 70, 89–95.

Holmes, E., Caddick, S., Lindon, J.C., Wilson, I.D., Kryvawych, S.,Nicholson, J.K., 1995b.1H and2H NMR spectroscopic studies onthe metabolism and biochemical effects of 2-bromoethanamine inthe rat. Biochemical Pharmacology 49, 1349–1359.

Holmes, E., Bonner, F., Nicholson, J.K., 1996. Comparative biochemi-cal effects of low doses of mercury II chloride in the F344 ratand the multimammate mouse (Mastomys natalensis). ComparativeBiochemistry and Physiology C 114 (1), 7–15.

Kerkut, G.A., Gilbert, L.T., 1985. Comprehensive Insect PhysiologyBiochemistry and Pharmacology vol. 3, Integument, Respirationand Circulation. Pergamon Press, Oxford.

Nicholls, A.W., Caddick, S., Wilson, I.D., Farrant, R.D., Lindon, J.C.,Nicholson, J.K., 1995. High resolution1H and 2H NMR studieson the metabolism and futile deacetylation of 4-hydroxyacetanilide(paracetamol-d3) in the rat. Biochemical Pharmacology 49 (8),1155–1164.

Nicholson, J.K., Foxall, P.J.D., Spraul, M., Farrant, R.D., Lindon, J.C.,1995. 750 MHz1H and1H-13C NMR spectroscopy of human bloodplasma. Analytical Chemistry 67, 793–811.

Nicholson, J.K., Wilson, I.D., 1989. High resolution proton NMR spec-troscopy of biological fluids. Progress in Nuclear Magnetic Reson-ance Spectroscopy 21, 449–501.

Shaka, A.J., Barker, P.B., Freeman, R., 1985. Computer-optimiseddecoupling scheme for wideband applications and low-level oper-ation. Journal of Magnetic Resonance 64, 547–552.

Shimizu, T., Takeda, N., 1991. Biogenic-amine levels in the haemo-lymph of the cabbage armyworm larvae (Mamestra brassicae) fol-lowing injection of octopaminergic insecticides. Zeitschrift fur Nat-urforschung C-A Journal Biosciences 46 (1-2), 127–132.

Stamp, N.E., Casey, T.M., 1993. Caterpillars: Ecological and Evol-utionary Constraints on Foraging. Chapman and Hall, London.

Takeda, N., Takaoka, H., Shimizu, T., Yazawa, M., Yagi, S., 1991.Biogenic-amine levels in the central nervous system and haemo-lymph of the silkwormBombyx mori. Comparative Biochemistryand Physiology C 100 (3), 677–682.

Thompson, S.N., 1990. NMR Spectroscopy: its basis, biological appli-cation and use in studies of insect metabolism. Insect Biochemistry20 (3), 223–237.

Thompson, S.N., 1998. Long-term regulation of glucogenesis by diet-ary carbohydrate and relevance to blood sugar level in an insectManduca sextaL. The International Journal of Biochemistry andCell Biology 30, 987–999.

Thompson, S.N., Borchardt, D.B., 1996. An in vivo13C NMR studyon the effect of dietary sugar on pyruvate cycling during glucogen-esis from (2-13C) pyruvate inManduca sextaL. Insect Biochemis-try and Molecular Biology 26 (10), 1047–1054.

Turunen, S., 1992. Efficient use of dietary galactose inPieris bras-sicae. Journal of Insect Physiology 38 (7), 503–509.

Verma, A., 1991. Effect of carbaryl toxication on hemolymph freeamino acid ofSpodoptera litura. Fab. Nat. acad. science letters-India 14 (3), 151.

Wilson, I.D., Nicholson, J.K., 1995. NMR spectroscopic analysis ofbiological fluids. In: The Encyclopaedia of Analytical Science.(Editor A. Townshend) Academic Press, New York, pp. 3518–3531.

Wilson, I.D., Wade, K.E., Nicholson, J.K., 1989. Analysis of biologicalfluids by high-field nuclear magnetic resonance spectroscopy.Trends in Analytical Chemistry 8 (10), 368–374.