NMR IN BIOMEDICINE

NMR Biomed. 2002;15:37–44

DOI:10.1002/nbm.740

A comparison of cell and tissue extraction techniques using

high-resolution 1H-NMR spectroscopy

J. E. Le Belle,* N. G. Harris, S. R. Williams and K. K. Bhakoo

Unit of Biophysics, Institute of Child Health, UCL Medical School, London WC1N 1EH, UK

Received 11 November 2000; Revised 9 September 2001; Accepted 11 September 2001

ABSTRACT: Analysis of brain metabolites by a wide range of analytical techniques is typically achieved using
biochemical extraction methodologies that require either two separate samples or two separate extraction steps to
prepare both aqueous and organic metabolite fractions. However there are a number of brain pathologies in which
both aqueous metabolite and lipid changes occur so that a simultaneous extraction of both fractions would be
valuable. The methanol–chloroform (M/C) technique enables extraction of both aqueous metabolites and lipids
simultaneously. It is already well established for lipid extraction of cells and tissue but its efficiency and
reproducibility for extraction of aqueous metabolites is unknown. Therefore, we compared the aqueous metabolite
yield and the reproducibility of the M/C method to the commonly used perchloric acid (PCA) method, using 1H-NMR
spectroscopy of adult rat brain and purified rat astrocyte culture extracts. The results indicate that M/C is a superior
technique for aqueous metabolite extraction from both brain tissue and cells when compared to the PCA method. The
M/C extraction technique enables the simultaneous extraction of both lipids and aqueous metabolites from a single
sample using small solvent-volumes, making it well suited for NMR investigations of both tissues and cells.
Copyright  2002 John Wiley & Sons, Ltd.
KEYWORDS: extract; PCA; methanol; chloroform; astrocyte; rat

INTRODUCTION used to achieve this goal. It facilitates the simultaneous

Cell and tissue extraction is a widely used technique for
biochemical analysis and quantification of metabolite
profiles in NMR spectroscopic (NMRS) studies of
disease. While a number of different extraction tech-
niques exist for this purpose, the most ubiquitous is the
perchloric acid (PCA) method, which extracts water-
soluble metabolites. However, a number of in vivo 1H-
NMRS studies have demonstrated significant changes in
both lipids and water-soluble metabolites in many
different CNS diseases, for example multiple sclerosis1
and high-grade brain tumours.2,3 Therefore, a more
detailed in vitro analysis of both lipid and aqueous
neurochemicals may provide additional information to
aid the interpretation of in vivo spectra and enhance
clinical diagnosis. The methanol–chloroform–water ex-
traction (M/C) is a less widely used technique that can be
*Correspondence to: J. E. L. Belle, Centre for Brain Repair, University
of Cambridge, Forvie Site, Robinson Way, Cambridge CB2 2PY, UK.
Email: jl289@cam.ac.uk
Contract/grant sponsor: Harold Bridges Trust.
Contract/grant sponsor: Institute of Mental Health, USA; contract
grant number: 5-F31-MH11531-02.
Abbreviations used: Chols, total choline-containing compounds;
M/C, methanol–chloroform; PCA, perchloric acid; PtdCho, phospha-
tidylcholine.
Copyright  2002 John Wiley & Sons, Ltd.
extraction of both the water-soluble metabolites and the
organic-soluble lipid components from the same tissue
sample.
The important requirements of an extraction technique
are that it is efficient and that it produces a high total
tissue metabolite yield with low variability. It is well
established that PCA extraction fulfills these criteria for
the water-soluble metabolites4–10 and the M/C extraction
fulfills these criteria for lipids.11–18 However, M/C
extraction has not yet been established as a reliable
technique for proton NMR spectroscopy of water-soluble
metabolites. We have therefore compared the aqueous
metabolite yield and protein pellet content from the PCA
and the M/C extraction techniques for both rat brain
tissue and cellular preparations of purified astrocytes.
EXPERIMENTAL
Brain funnel freezing
Adult male Sprague–Dawley rats (200  20 g, n = 21)
were anaesthetized with 4% halothane vapourized in
N2O and O2 flowing at 0.7 and 0.3 l/min. Subsequent
anaesthesia during surgery and brain freezing was
maintained with 1.5 and 1% halothane, respectively.
NMR Biomed. 2002;15:37–44
38 J. E. LE BELLE ET AL.

Table 1. Summary of metabolite yields from the ®rst set of extractions of rat brain samples using either the
methanol±cholorform (M/C) or perchloric acid (PCA) extraction techniques

M/C CEV PCA CEV p-Values

Total Cr 8.75  0.15 8 7.74  0.24 14 0.001
Chols 1.36  0.03 9 1.28  0.04 14 0.091
Glutamine 5.37  0.07 6 5.18  0.20 18 0.396
Glutamate 9.40  0.11 5 8.64  0.32 17 0.033
NAA 7.09  0.13 8 6.51  0.22 16 0.030
Alanine 0.65  0.01 9 0.57  0.02 19 0.0003
Lactate 1.98  0.07 15 1.46  0.05 17 3  10 6
Protein 9.73  1.30 33 15.10  1.29 21 0.012
Metabolites expressed as mmol/g wet weight and mean  SEM. Protein is expressed as a percentage of the wet weight of the extracted tissue.
Abbreviations are as follows: Total Cr, total creatine (creatine ‡ phosphocreatine); Chols, choline-containing compounds (glycerophos-
phocholine ‡ phosphocholine ‡ choline); NAA, N-acetylaspartate; CEV, coefficient of variation (SD  100/mean). The critical value a, adjusted
for the number of comparisons, is 0.007 (0.05/7) for p-values to achieve significance.

Brain tissue was frozen in situ by pouring liquid nitrogen
through a funnel directly onto the exposed skull of the
rat.19 The brain tissue was removed intact by chiseling
away skin and bone while the brain was submerged in a
reservoir of liquid nitrogen to prevent thawing.
Extraction protocols
Brain tissue. The frozen brain tissue (n = 21) was
divided into two halves for extraction by either PCA or
M/C and the aqueous metabolites were analysed by high-
resolution 1H-NMR spectroscopy (n = 21/group). Six
tissue pellets from each extraction group were then
chosen at random for protein content determination.20
Since the aqueous metabolite yields were greater using
the M/C technique (Table 1), the remaining tissue pellets
from the original M/C and PCA extracts (n = 15/group)
were then re-extracted using the M/C technique. This re-
extraction was analysed separately from the first extrac-
tion using 1H-NMR spectroscopy to determine the
metabolite yield remaining in the pellets after the first
extraction.
Cells. Primary rat astrocyte cultures were grown to
sufficient numbers21 for 12 separate NMR samples
(approximately 107 cells/ NMR sample) for extraction
by either PCA or M/C (n = 6/group). Cell purity (>98%)
was determined by glial fibrillary acidic protein im-
munostaining. Each group of cells was extracted twice
using a single extraction technique and the supernatants
were pooled before drying and analysed by high-resolu-
tion 1H-NMR spectroscopy.
Tissue and cell extraction
Brain tissues were kept under liquid nitrogen and ground
to a fine powder with a mortar and pestle before the
extract solvents were added and allowed to thaw while in
Copyright  2002 John Wiley & Sons, Ltd.
contact with the ground tissue. Ice-cold solvents were
added to the frozen cell samples from both groups and the
cell-solvent mixtures were sonicated.
Methanol±chloroform±water (M/C) extraction. Re-
agent-grade methanol and chloroform (4 °C) in a ratio of
2:1 (v/v; 3 ml/g tissue or 250 ml/cell pellet) were added to
the frozen, ground tissue and to the frozen cell pellets.
The tissue–solvent mixture was allowed to thaw before
being transferred to Teflon centrifuge tubes (Oakridge,
UK). The cell pellet–solvent mixture was sonicated.
After approximately 15 min in contact with the first
solvents, chloroform and distilled water were added to
the samples in a ratio of 1:1 (1 ml/g tissue or 250 ml/cell
pellet) to form an emulsion. The samples were then
centrifuged at 13,000 rpm for 20 min. The upper phase
(methanol and water) was separated from the lower
(organic) phase using a glass syringe and both fractions
were dried at room temperature under a stream of
nitrogen gas. The protein pellets from the first tissue
extraction were re-extracted and dried separately from
the original tissue extract. The protein pellets from the
cells were also re-extracted but the separated fractions
were pooled with the original extracted fractions before
drying.
Perchloric acid (PCA) extraction. Cold 12% perchloric
acid (4 °C) was added (3 ml/g tissue or 250 ml/pellet) to
the ground tissue under liquid nitrogen or to the frozen
cells that were then sonicated. The tissue–solvent mixture
was allowed to thaw before being transferred into
centrifuge tubes. All samples were then centrifuged at
13,000 rpm for 20 min. The supernatant was removed and
the pellet was re-extracted using either M/C (tissue) or
PCA (cells). The PCA supernatants were neutralized with
1 M NaOH and the protein pellet kept for further analysis.
The precipitated salt was removed by centrifugation and
the supernatant was freeze-dried overnight.
NMR Biomed. 2002;15:37–44
 NMR EXTRACT COMPARISON 39

Table 2. Summary of metabolite yields from the tissue M/C re-extraction

M/C pellets CEV PCA pellets CEV P-values

Total Cr 0.74  0.05 24 1.04  0.08 30 0.005
Chols 0.10  0.01 31 0.91  0.05 22 1.2  10 10

Glutamine 0.49  0.02 12 0.81  0.05 25 5.9  10 6

Glutamate 0.80  0.07 33 1.55  0.08 19 3.1  10 7

NAA 0.55  0.04 26 1.06  0.04 15 7.9  10 7

Alanine 0.11  0.01 28 0.18  0.01 30 0.005

Lactate 0.35  0.01 15 0.26  0.02 35 0.011
Metabolites are expressed as mmol/g wet weight and mean  SEM. See Table 1 for abbreviations. The critical value a, adjusted for the number of
comparisons, is 0.007 (0.05/7) for p-values to achieve significance.

Protein determination Data analysis and statistics

Cell and tissue pellets were solublised in 2 ml/g (tissue)
or 200 ml/pellet (cells) of 1 M NaOH and analysed for
protein content against a set of bovine serum albumin
protein standards using the Bio-Rad (UK) reagent kit and
an LKB Ultrospec II spectrophotometer (wavelength,
595 nm).
Phosphatidylcholine extraction
M/C re-extraction of the tissue pellets that were initially
extracted with PCA resulted in the release of an unusually
large amount of choline (Table 2). Therefore, in order to
determine the source of the high choline we investigated
the effects of PCA extraction followed by M/C extraction
on the membrane phospholipid, phosphatidylcholine
(PtdCho). PtdCho standard, 1 mmol (from bovine brain
Sigma, UK) was exposed to 12% PCA for 10 min, 24 h,
48 h and 2 weeks (n = 3 per group) followed by M/C
extraction. 1H-NMR spectroscopy was used to determine
the concentration of free choline, a breakdown-product of
PtdCho in the aqueous fraction of the extract.
Tissue and cell metabolite concentrations are expressed
as per gram wet weight of brain tissue and per mg protein,
respectively. All values are expressed as means  SEM.
From an analysis of the inter-animal variability of the
metabolite yields, it was judged appropriate to analyse
seven different metabolite peaks in order to maintain
sufficient statistical power to resolve any group differ-
ences. These metabolites represent a range of chemically
varied compounds present in brain tissue including
alanine, glutamine, glutamate, lactate, N-acetyl-aspar-
tate, choline-containing compounds, and creatine. In the
astrocyte extracts, hypotaurine was analysed in place of
NAA, which is not present in astrocytes.21 Paired t-tests
were used to compare the metabolite concentrations and
the protein content from the two extraction techniques. P-
values are quoted without correction for multiple com-
parisons. Therefore, the critical value for assessing
significance was adjusted as a function of the number
of metabolites being compared (a = 0.007 at the 5%
significance level).
RESULTS

1
H-NMR spectroscopy
All dried extract samples were re-dissolved in D2O and
adjusted to pH 7.0. A known amount of an internal
standard (trimethylsilylproprionate) was added to each
sample. Spectroscopy was performed at 25 °C with a
Varian Unity plus spectrometer operating at 500 MHz.
Fully relaxed, one-dimensional spectra were acquired
using a pulse-acquire sequence with the following
parameters: repetition time = 22 s, number of transi-
ents = 256, and a 90° flip angle. Data was Fourier-
transformed with 0.1 Hz line broadening. Resonance
assignments were made based on published chemical
shifts and coupling patterns of known compounds. Peak
areas were integrated using the manufacturer’s (Varian)
standard software, after local baseline flattening around
each integration region.
Copyright  2002 John Wiley & Sons, Ltd.
Tissue extractions
The first tissue extraction compared the effect of
extracting each brain half once with either the PCA or
M/C technique. The results show that the metabolite
yields from M/C extraction were greater than those with
PCA extraction with significant differences in total Cr,
Ala and Lac yields (Fig. 1 and Table 1, p < 0.007). In
addition, with the exception of lactate, the coefficient of
variation was consistently lower using the M/C technique
(5–9% compared with 14–19%). The variation in protein
content was similar for both techniques (Table 1). This is
an important consideration for studies of cell extracts
where metabolite concentrations are expressed relative to
protein. The protein content (as a percentage of the wet
weight of the extracted tissue) left in the pellet after the
first extraction was significantly higher in the PCA
extraction (p = 0.012, a = 0.05).
NMR Biomed. 2002;15:37–44
40 J. E. LE BELLE ET AL.

Figure 1. Representative 1H-NMR spectra from the initial M/C (top) and
PCA (bottom) extracts of whole rat brain frozen in situ. The spectra were
acquired from extracts prepared from the two halves of the same brain.
Abbreviations are as follows: Lac, lactate; Ala, alanine; NAA, N-
acetylaspartate; Glu, glutamate; Gln, glutamine; PCr, phosphocreatine;
Cr, creatine; Chols, choline-containing compounds

Tissue re-extractions with M/C
Subsequent M/C re-extraction of the tissue pellets from
the original PCA or M/C extractions resulted in a
substantially higher yield of all metabolites (except for
lactate) from the PCA pellet (Table 2). Some metabolites
were also re-extracted from the M/C pellets, indicating
the need for at least one re-extraction of the tissue pellet
in order to maximize the metabolite yield for analysis.
Comparison of total metabolite yield (original extraction
plus M/C re-extraction) revealed no significant difference
between the groups except for choline-containing
compounds and lactate, which were significantly lower
(p = 8  10 8) and higher (p = 1  10 6; a = 0.05/7)
from the M/C pellets, respectively (Table 3). When the
original M/C or PCA extraction is considered as a
percentage of this total yield, the M/C technique resulted
in consistently higher extraction efficiency (85–93% of
total) compared with PCA (58–88% of total; Table 3).
However, it should be noted when considering the first
extraction relative to the total cell extract that the data on
the PCA extraction of the Chols (58%) was skewed by the
release of free choline from membrane lipids in the re-
extraction of the tissue pellets.
Cell extractions
The tissue extractions demonstrated a need for re-
extracting the protein pellet with either M/C or PCA in
order to maximize the metabolite yield. Therefore, the
supernatants from the initial extraction and re-extraction
of cells using the same technique were pooled prior to
analysis.
The cellular metabolite yields from the M/C extraction
were once again greater than the PCA extraction (Fig. 2

Table 3. Total metabolite yields (original plus re-extraction) for tissue extraction

M/C pellets Percentage first extraction PCA pellets Percentage first extraction p-Values
Total Cr 9.49  0.20 92 8.78  0.34 88 0.027
Chols 1.46  0.04 93 2.19  0.11 58 8.0  10 8

Glutamine 5.86  0.12 92 6.00  0.27 86 0.605

Glutamate 10.20  0.19 92 10.19  0.40 85 0.897
NAA 7.64  0.17 93 7.56  0.27 86 0.872
Alanine 0.76  0.03 86 0.72  0.04 79 0.223
Lactate 2.33  0.10 85 1.73  0.09 84 1.1  10 6

The percentage first extraction represents the percentage of the total metabolite yield from the original tissue extraction by M/C or PCA. Metabolite
yields are expressed as mmol/g wet weight and a = 0.007.

Copyright  2002 John Wiley & Sons, Ltd. NMR Biomed. 2002;15:37–44
 NMR EXTRACT COMPARISON 41

Figure 2. Representative 1H-NMR spectra from the M/C (top) and PCA (bottom) extracts of
rat astrocytes grown under identical culture conditions. See Fig. 1 for abbreviations; H-tau,
hypotaurine

and Table 4), with highly significant differences detected the M/C extract when the PtdCho standard had not
between extraction techniques for all of the metabolites previously been exposed to PCA.
measured. In addition, the coefficient of variation was
consistently lower using the M/C technique (5–12%
compared with 12–45% for PCA). DISCUSSION

The results show that methanol-chloroform extraction of

Phosphatidylcholine extraction
Exposing a standard amount of phosphatidylcholine to
PCA for varying times before subsequent extraction with
M/C resulted in the recovery of small amounts of choline
in the aqueous fraction, which increased with the time of
PCA exposure, ranging from 0.025 to 0.059 mmol.
Choline was not detected in the aqueous fraction of
both brain tissue and cultured astrocytes produce a
significantly greater metabolite yield with lower varia-
bility when compared to extraction by perchloric acid.
The M/C procedure
Although several different solvent-systems exist for the

Table 4. Summary of metabolite yields from the rat astrocyte cultures using either the methanol±cholorform (M/C)
or perchloric acid (PCA) extraction techniques

M/C CEV PCA CEV p-Values

Total Cr 21.72  0.63 7 11.16  2.53 45 0.006
Chols 40.72  1.34 8 26.80  1.55 12 4  10 5
Glutamine 29.09  0.89 7 15.19  3.15 41 0.001
Glutamate 12.01  0.24 5 7.25  0.50 14 2  10 5
Hypotaurine 39.99  2.02 12 19.25  3.08 32 5  10 5
Alanine 3.90  0.18 11 1.98  0.26 26 0.0004
Lactate 14.82  0.49 8 8.68  1.05 24 5  10 5
Metabolites are expressed as nmol/mg protein and mean  SEM. n = 6 samples/extraction group. Abbreviations are as follows: Total Cr, total
creatine (creatine ‡ phosphocreatine); Chols, choline-containing compounds (glycerophosphocholine ‡ phosphocholine ‡ choline); CEV, coeffi-
cient of variation (SD  100/mean). The critical value a, adjusted for the number of comparisons, is 0.007 (0.05/7) to achieve significance.

Copyright  2002 John Wiley & Sons, Ltd. NMR Biomed. 2002;15:37–44
42
Table 5. Summary of the individual choline-containing compounds extracted from whole rat brain using the M/C
and PCA techniques for the ®rst extraction and M/C only for the re-extraction of the protein pellets
GPC
First M/C extract 0.87  0.02
First PCA extract 0.82  0.03
M/C pellet re-extraction 0.07  0.01
PCA pellet re-extraction 0.16  0.01
Abbreviations are as follows: GPC, glycerophosphocholine; PC, phosphocholine; Cho, choline. Metabolites are expressed as mmol/g wet weight and
as mean  SEM. GPC > PC > Cho, except for the PCA pellet re-extraction by M/C which yielded Cho > GPC > Cho.
extraction of particular lipid species, Folch et al.22 first
introduced the use of a methanol and chloroform solvent
system for the efficient extraction of total lipids from
animal tissues. This method was later improved by Bligh
and Dyer23 by carrying out the extraction and separation
of solvent layers almost simultaneously, retaining the
precipitated protein between phases. Although the Bligh–
Dyer method was designed to extract lipid species, the
technique also eliminated the addition of salts to the
aqueous layer, which had previously rendered this
fraction unsuitable for analysis using the Folch method.
The Bligh–Dyer method also established the proportions
of solvents (methanol:chloroform:water) based on the
water content of the specific tissues being extracted. The
use of smaller solvent volumes makes this method
applicable to both dilute cell suspensions and tissue
samples. The M/C extraction method employed in the
present study is based on the Bligh–Dyer technique.
Both the Folch and Bligh–Dyer extraction techniques
have been widely used, although often in modified form,
in both NMR and non-NMR studies of lipid composition
in normal and pathological tissue, for example in human
liver,24 rat brain,17 multiple sclerosis25 and mouse
myeloma cells.11 NMR analyses of lipids extracted in
this way have also been cross-validated by chromato-
graphic techniques and shown to be in good agreement.26
Thus, the use of M/C extraction for the analysis of cell
and tissue lipids is well established. However, the
information from 1H-NMRS of the aqueous phase of
the M/C extraction has not been utilized until now.
Although a combination of PCA followed by M/C
extraction can be used on a sample to obtain both
aqueous and lipid metabolites,12 this requires an extra
extraction step compared to the methodology used in the
present study. Furthermore, there is some evidence that
the PCA extraction may result in a loss of total
phospholipids and cholesterol from the sample.14
Tissue extraction
With the exception of the Chols and lactate, the
concentrations of the aqueous metabolites extracted by
either M/C or PCA in the present study are in good
agreement with published values on the rat brain.4–10,27–
30
However, the finding of higher lactate concentrations
Copyright  2002 John Wiley & Sons, Ltd.
J. E. LE BELLE ET AL.
PC CHO
0.39  0.01 0.10  0.002
0.39  0.01 0.07  0.002
0.02  0.002 0.01  0.001
0.07  0.003 0.65  0.04
in all M/C extractions could indicate that the technique
does not inactivate tissue enzymes as efficiently as PCA.
On the other hand, the relatively high PCr/Cr ratio (PCA,
0.59  0.015; M/C, 0.56  0.017; mean  SEM) would
suggest that enzyme inactivation has been relatively
efficient. It is more likely, however, that the source of this
discrepancy is the relative inaccuracy of lactate peak
integration due to an overlapping broad signal from
contaminating protein or lipid in the M/C extracts.
Overall, the M/C technique proved to be more efficient
and less variable than PCA in the initial extraction of the
normal brain tissues.
Tissue re-extraction
With the exception of lactate, re-extraction of the original
M/C or PCA protein pellet with subsequent M/C re-
extraction resulted in a greater metabolite yield from the
PCA pellet, indicating a less efficient initial PCA
extraction. As mentioned previously, the high yield of
lactate from the M/C samples was most likely due to the
overlap of signal from small amounts of contaminating
protein and/or lipid in the samples. Thus, this further
demonstrates the higher efficiency of the original M/C
extraction and the ability of the M/C re-extraction to
maximize the total metabolite yield extractable from the
brain tissues.
The very high yield of choline-containing compounds
resulting from M/C re-extraction of the PCA-pellet was
an unexpected finding. Similar results have been reported
in human tumour biopsy tissue extracted by PCA
followed by ethanol–chloroform,31 where it was hy-
pothesized that extra choline was produced by separation
of the membrane-contained hydrophilic choline head-
groups by the chlorform–alcohol extraction. In support of
this hypothesis, analysis of the individual choline-
containing compounds in the PCA tissue-pellet re-
extracts in this study revealed that the main source of
the increased total choline-containing compound signal
was the free-choline component rather than either
glycerophosophocholine, or phosophocholine (Table 5).
Furthermore, exposure of a phosphatidylcholine (Ptdcho)
standard to PCA for varying times before extraction by
M/C resulted in the extraction of increasing concentra-
tions of choline. Since M/C does not normally extract
NMR Biomed. 2002;15:37–44
 NMR EXTRACT COMPARISON
choline from PtdCho, which is partitioned in the lipid
fraction of the extract, it is likely that a small amount of
PCA remaining with the tissue pellets after the original
tissue extraction may have hydrolysed the C-O-P
phosphoester linkages in membrane Ptdcho molecules,
releasing choline into the aqueous fraction of the
subsequent M/C re-extractions. This may explain the
significant difference for total yield of Chols between the
two extraction techniques. Comparison of the yields for
the choline-containing compounds in the original extrac-
tion demonstrates that, as a separate technique, the M/C
extraction is superior to PCA.
Cell extraction
Comparison of M/C and PCA extraction of purified rat
astrocytes yielded similar results to those observed in
tissue extractions: the M/C extraction proved to be a less
variable and more efficient technique. The M/C tech-
nique may be ideally suited to cellular extract studies
where very small amounts of biological material are used
because the entire aqueous fraction can be easily re-
dissolved in D2O for NMR analysis. In contrast, the total
dry weight of the cellular residue after PCA extraction is
often two to three times greater than the maximum
amount that will dissolve fully in 0.4–0.7 ml D2O, the
volume of solvent typically used in high-resolution NMR
sample tubes. This is due to the precipitation of salts,
which occurs despite neutralization and centrifugation
prior to drying, so that a maximum concentration of
30 mg/ml can be reconstitution in D2O without satura-
tion. However, reconstituted of the whole sample is
required to achieve adequate sensitivity for NMR signal
acquisition from dilute cellular extracts. Therefore, when
the whole PCA-extract sample is reconstituted in D2O,
the remaining un-dissolved salt must be centrifuged and
removed again prior to NMR analysis. Although the salt
pellet can be washed, these extra steps undoubtedly add
to the experimental error of the PCA technique.
The aqueous metabolite yields that we obtained for rat
astrocytes are broadly similar to other studies in the
literature using PCA extraction,21,32 although compari-
sons are difficult due to variations in experimental
techniques such as cell media compositions and harvest-
ing protocols. However, the cells in the present study
were grown and harvested under identical conditions and
contained very similar protein content (PCA, 15.4 
0.6 mg/sample; M/C, 15.2  1.2 mg/sample; mean 
SEM) so that any differences in metabolite yield can be
attributed to the extraction technique alone. Although
there have been a small number of studies using 31P-NMR
to analyse the aqueous phosphate-containing metabo-
lites from M/C extracted tissue and cells,11,28 these
are the first reported quantitative values of cellular
aqueous metabolites using M/C extraction and 1H-NMR
analysis.
Copyright  2002 John Wiley & Sons, Ltd.
43
CONCLUSIONS
In conclusion, we have demonstrated that the M/C
method is an efficient and reproducible technique for
extracting the aqueous metabolites from tissue and cell
culture. Using the M/C extraction technique, based on the
Bligh–Dyer method, both the lipid and aqueous com-
pounds are obtained from a single sample, making it
well-suited for NMR investigations. This technique is
superior to either PCA extraction alone, or PCA
extraction followed by lipid extraction.
Acknowledgements
This work was supported by the Harold Bridges Trust,
UK, and the Institute of Mental Health, USA, grant no. 5-
F31-MH11531-02. The authors would like to gratefully
acknowledge Dr Martin King for his assistance and
expertise with statistical analysis.
REFERENCES
1. Davie CA, Hawkins CP, Barker GJ, Brennan A, Tofts PS, Miller
DH, McDonald WI. Detection of myelin breakdown products by
proton magnetic resonance spectroscopy. [Letter.] Lancet 1993;
341: 630–631.
2. Remy C, Arus C, Ziegler A, Lai ES, Moreno A, Le Fur Y, Decorps
M. In vivo, ex vivo, and in vitro one- and two-dimensional nuclear
magnetic resonance spectroscopy of an intracerebral glioma in rat
brain: assignment of resonances. J. Neurochem. 1994; 62: 166–
179.
3. Sijens PE, Levendag PC, Vecht CJ, van Dijk P, Oudkerk M. 1H
MR spectroscopy detection of lipids and lactate in metastatic brain
tumors. NMR Biomed. 1996; 9: 65–71.
4. Evanochko WT, Sakai TT, Ng TC, Krishna NR, Kim HD, Zeidler
RB, Ghanta VK, Brockman RW, Schiffer LM, Braunschweiger
PG. NMR study of in vivo RIF-1 tumors. Analysis of perchloric
acid extracts and identification of 1H, 31P and 13C resonances.
Biochim. Biophys. Acta 1984; 805: 104–116.
5. Peeling J and Sutherland G. High-resolution 1H NMR spec-
troscopy studies of extracts of human cerebral neoplasms. Magn.
Reson. Med. 1992; 24: 123–136.
6. Glonek T, Kopp SJ, Kot E, Pettegrew JW, Harrison WH, Cohen
MM. P-31 nuclear magnetic resonance analysis of brain: the
perchloric acid extract spectrum. J. Neurochem. 1982; 39: 1210–
1219.
7. Burri R, Bigler P, Straehl P, Posse S, Colombo JP, Herschkowitz
N. Brain development: 1H magnetic resonance spectroscopy of rat
brain extracts compared with chromatographic methods. Neuro-
chem. Res. 1990; 15: 1009–1016.
8. Bates TE, Williams SR, Gadian DG, Bell JD, Small RK, Iles RA.
1
H NMR study of cerebral development in the rat. NMR Biomed.
1989; 2: 225–229.
9. Fan TW, Higashi RM, Lane AN, Jardetzky O. Combined use of
1
H-NMR and GC-MS for metabolite monitoring and in vivo 1H-
NMR assignments. Biochim. Biophys. Acta 1986; 882: 154–167.
10. Sze DY, Jardetzky O. Determination of metabolite and nucleotide
concentrations in proliferating lymphocytes by 1H-NMR of acid
extracts. Biochim. Biophys. Acta 1990; 1054: 181–197.
11. Tyagi RK, Azrad A, Degani H, Salomon Y. Simultaneous
extraction of cellular lipids and water-soluble metabolites:
evaluation by NMR spectroscopy. Magn. Reson. Med. 1996; 35:
194–200.
12. Henke J, Willker W, Engelmann J, Leibfritz D. Combined
extraction techniques of tumour cells and lipid/phospholipid
NMR Biomed. 2002;15:37–44
44 J. E. LE BELLE ET AL.
assignment by two dimensional NMR spectroscopy. Anticancer
Res. 1996; 16: 1417–1427.
13. Jones RM, Richards T, Budinger TF. Proton and carbon spec-
troscopy of lipids and aqueous metabolites extracted from brain.
Proc. ISMRM, 1996; 395.
14. Katyal SL, Barilaro L, Hanin I. Lipid composition of different
areas of murine brain: effects of lipid extraction procedures. Lipids
1985; 20: 201–203.
15. Edzes HT, Teerlink T, van der Knaap MS, Valk J. Analysis of
phospholipids in brain tissue by 31P NMR at different compositions
of the solvent system chloroform–methanol–water. Magn. Reson.
Med. 1992; 26: 46–59.
16. Sappey-Marinier D. High-resolution NMR spectroscopy of
cerebral white matter in multiple sclerosis. Magn. Reson. Med.
1990; 15: 229–239.
17. Capuani G, Aureli T, Miccheli A, Di Cocco ME, Ramacci MT,
Delfini M. Improved resolution of 31P nuclear magnetic resonance
spectra of phospholipids from brain. Lipids 1992; 27: 389–391.
18. Sze DY and Jardetzky O. Characterization of lipid composition in
stimulated human lymphocytes by 1H-NMR. Biochim. Biophys.
Acta 1990; 1054: 198–206.
19. Lust WD, Ricci AJ, Selman WR, Ratcheson RA. Methods of
fixation of nervous tissue for use in the study of cerebral energy
metabolism, in Neuromethods: Carbohydrates and Energy Meta-
bolism, Eds: Boulton AA, Baker GB and Butterworth RF, Humana
Press, New Jersey pp. 1–36, 1989.
20. Smith PK, Krohn RI, Hermanson GT, Mallia AK, Gartner FH,
Provenzano MD, Fujimoto EK, Goeke NM, Olson BJ, Klenk DC.
Measurement of protein using bicinchoninic acid. [Published
erratum appears in Anal. Biochem. 1987; 163(1): 279.] Anal.
Biochem. 1985; 150: 76–85.
21. Urenjak J, Williams SR, Gadian DG, Noble M. Proton nuclear
magnetic resonance spectroscopy unambiguously identifies differ-
ent neural cell types. J. Neurosci. 1993; 13: 981–989.
22. Folch J, Lees M, Sloane-Stanley G. A simple method for the
isolation and purification of total lipides from animal tissues. Biol.
Chem. 1957; 226: 497–509.
Copyright  2002 John Wiley & Sons, Ltd.
23. Bligh EG and Dyer WJ. A rapid method of total lipid extraction
and purification. Can. J. Biochem. 1959; 37: 911–917.
24. Pollesello P, Eriksson O, Hockerstedt K. Analysis of total lipid
extracts from human liver by 13C and 1H nuclear magnetic
resonance spectroscopy. Anal. Biochem. 1996; 236: 41–48.
25. Sappey-Marinier DS, Letoublon R, Delmau J. Phosphorus NMR
analysis of human white matter in mixed non-ionic detergent
micelles. J. Lipid Res. 1988; 29: 1237–1243.
26. Meneses P and Glonek T. High resolution 31P NMR of extracted
phospholipids. J. Lipid Res. 1988; 29: 679–689.
27. Aureli T, Miccheli A, Ricciolini R, Di Cocco ME, Ramacci MT,
Angelucci L, Ghirardi O, Conti F. Aging brain: effect of acetyl-L-
carnitine treatment on rat brain energy and phospholipid metab-
olism. A study by 31P and 1H NMR spectroscopy. Brain Res.
1990; 526: 108–112.
28. Aureli T, Miccheli A, Di Cocco ME, Ghirardi O, Giuliani A,
Ramacci MT, Conti F. Effect of acetyl-L-carnitine on recovery of
brain phosphorus metabolites and lactic acid level during
reperfusion after cerebral ischemia in the rat – study by 13P- and
1
H-NMR spectroscopy. Brain Res. 1994; 643: 92–99.
29. Florian CL, Williams SR, Bhakoo KK, Noble MD. Regional and
developmental variations in metabolite concentration in the rat
brain and eye: a study using 1H NMR spectroscopy and high
performance liquid chromatography. Neurochem. Res. 1996; 21:
1065–1074.
30. Harris NG, Plant HD, Inglis BA, Briggs RW, Jones HC.
Neurochemical changes in the cerebral cortex of treated and
untreated hydrocephalic rat pups quantified with in vitro 1H-NMR
spectroscopy. J. Neurochem. 1997; 68: 305–312.
31. Bruhn H, Michaelis T, Merboldt KD, Hanicke W, Gyngell ML,
Hamburger C, Frahm J. On the interpretation of proton NMR
spectra from brain tumours in vivo and in vitro. NMR Biomed.
1992; 5: 253–258.
32. Serkova N, Brand A, Christians U, Leibfritz D. Evaluation of the
effects of immunosuppressants on neuronal and glial cells in vitro
by multinuclear magnetic resonance spectroscopy. Biochim.
Biophys. Acta 1996; 1314: 93–104.
NMR Biomed. 2002;15:37–44