Integrative Biology
Dynamic Article Links
Cite this: Integr. Biol., 2011, 3, 1135–1142
www.rsc.org/ibiology
TECHNICAL INNOVATION
Comparative bioenergetic assessment of transformed cells using a cell
energy budget platformw
A. V. Zhdanov,a C. Favre,a L. O’Flaherty,b J. Adam,b R. O’Connor,a
P. J. Pollardb and D. B. Papkovsky*a
Received 30th May 2011, Accepted 2nd October 2011
DOI: 10.1039/c1ib00050k
The aberrant expression and functional activity of proteins involved in ATP production pathways
may cause a crisis in energy generation for cells and compromise their survival under stressful
conditions such as excitation, starvation, pharmacological treatment or disease states. Under
resting conditions such defects are often compensated for, and therefore masked by, alternative
pathways which have significant spare capacity. Here we present a multiplexed ‘cell energy
budget’ platform which facilitates metabolic assessment and cross-comparison of different cells
and the identification of genes directly or indirectly involved in ATP production. Long-decay
emitting O2 and pH sensitive probes and time-resolved fluorometry are used to measure changes
in cellular O2 consumption, glycolytic and total extracellular acidification (ECA), along with the
measurement of total ATP and protein content in multiple samples. To assess the extent of spare
capacity in the main energy pathways, the cells are also analysed following double-treatment with
carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone and oligomycin. The four-parametric
platform operating in a high throughput format has been validated with two panels of
transformed cells: mouse embryonic fibroblasts (MEFs) lacking the Krebs cycle enzyme fumarate
hydratase (Fh1) and HeLa cells with reduced expression of pyrimidine nucleotide carrier 1. In
both cases, a marked reduction in both respiration and spare respiratory capacity was observed,
accompanied by a compensatory activation of glycolysis and consequent maintenance of total
ATP levels. At the same time, in Fh1-deficient MEFs the contribution of non-glycolytic pathways
to the ECA did not change.
a
Biochemistry Department, University College Cork,
Cavanagh Pharmacy Building, College Road, Cork, Ireland.
E-mail: [email protected]; Fax: +353-21-490-1698;
Tel: +353-21-490-1698
b
Henry Wellcome Building for Molecular Physiology,
University of Oxford, Oxford OX3 7BN, UK
w Electronic supplementary information (ESI) available. See DOI:
10.1039/c1ib00050k
Introduction
Energy stress, perturbed metabolism and mitochondrial
dysfunction are all hallmarks of abnormal cell development
and the progression of various diseases.1–3 Glycolysis and
oxidative phosphorylation (OxPhos) interlinked via the Krebs
cycle are the main energy generating pathways in eukaryotic
Insight, innovation, integration
We present a multi-parametric cell energy budget (CEB)
approach for assessing the roles of different factors in cell
bioenergetics. The aim was to standardize and minimize the
number of experiments required to identify metabolic
abnormalities associated with a particular protein, gene
mutation, drug treatment or disease state. This was achieved
by measuring and comparing the O2 consumption rate
(OCR), total and glycolytic extracellular acidification
This journal is
c
The Royal Society of Chemistry 2011
(ECA) and ATP levels in wild type and transformed cells
at rest and upon mitochondrial uncoupling. Kinetic timeresolved fluorescence detection of the OCR and ECA using
the long-decay emitting O2 and pH sensitive probes and high
throughput microplate format provides information-rich
data and leaves large scope for further multiplexing with
other fluorescent probes and assays, e.g. cellular Ca2+ or
mitochondrial membrane potential.
Integr. Biol., 2011, 3, 1135–1142
1135
Fig. 1 Schematic representation of cell bioenergetics and Cell Energy Budget concept. (A) Glycolysis, OxPhos and the Krebs cycle (provides the
substrates for OxPhos) usually work in a synchronised manner maintaining constant ATP levels within the cell. Each pathway has feedback
regulation (reciprocal scales for glycolysis and OxPhos) and significant spare capacity (dark grey areas outside the dashed box), thus allowing the
cell to withstand fluctuations in energy demand. If cell transformation reduces the efficiency and/or spare capacity of one pathway, the reciprocal
pathways are activated in order to maintain cellular ATP levels, as indicated by filled arrows. (B) The four-parametric assessment under energy
stress conditions helps uncover metabolic status of different cell types. FCCP/oligomycin treatment blocks ATP production by OxPhos, brings
respiration to its maximal level, activates glycolysis and the Krebs cycle. Measurement of the OCR and T-ECA in sealed samples and the L-ECA
and ATP in unsealed samples allows for interrogation of the cells probed for these primary pathways. Thus, the effects of mutations/deletions on
cell efficiency (at rest) and spare capacity (upon uncoupling) can be ascertained. For more robust analysis and cross-comparison, raw data are
normalized for total protein content in each sample well containing cells.
organisms (Fig. 1A). Working together they maintain an
optimal energy status for the cell, while their relative contributions
vary broadly for different cell types and physiological conditions.
Each pathway has a substantial spare capacity outside the
physiological range, thus allowing the cells to withstand stress
conditions and to maintain ATP levels by multiple feedback
mechanisms. In general terms, any spare capacity within either
of these energy producing pathways can be probed by limiting the
substrates required for ATP production, or by pharmacological
treatments. Thus, in normal cells with active OxPhos and
glycolytic machineries, when ATP production is inhibited by
mitochondrial uncoupling (OxPhos) or substitution of glucose
with galactose (glycolysis),4 total ATP levels remain unchanged
for many hours.5
On the other hand, metabolic abnormalities and mutations
in genes related to OxPhos, glycolysis or the Krebs cycle may
lead to redistribution of ATP fluxes5–7 and reduced spare
capacities. For such cells, sustained excitation or an imbalance
in energy-generating pathways may lead to an energy crisis,
increased susceptibility to stress factors, and ultimately death.
This highlights the importance of assessment and crosscomparison of the key metabolic parameters of normal and
transformed cells, their relative contribution and impact on
cell bioenergetics and pathophysiology. Measurement of cellular
NAD(P)H, ATP, Ca2+, mitochondrial pH, membrane potential
and redox state can highlight changes in cell metabolism, however
1136
Integr. Biol., 2011, 3, 1135–1142
ATP fluxes through OxPhos and glycolysis, and the activity of
the Krebs cycle are more informative and direct bioenergetic
parameters. Other metabolic pathways, including b-oxidation,
the pentose phosphate pathway and glutaminolysis, also
contribute to cell bioenergetics,8,9 but in the context of this
study they are set aside.
Various platforms have been described previously to
determine key metabolic parameters individually in a cell line
under one set of conditions, but this strategy cannot provide
the required details. The use of an automated system for
simultaneous measurement of cellular oxygen consumption
rate (OCR) and extracellular acidification (ECA) in dedicated
microchamber plates, Extracellular Flux Analyzer developed
by Seahorse Biosciences10 allows for an accurate bioenergetic
assessment of different cell types and metabolic effectors. This
system has been successfully used in a number of metabolic
studies (e.g. ref. 11 and 12), but currently it is not able
to differentiate the contribution of glycolysis (lactate) and
non-glycolytic activities (mostly CO2) to the ECA. Another
platform uses long-decay emitting O2 and pH sensitive probes
to analyse cells on a standard time-resolved fluorescence
(TR-F) reader in microtiter plates.13,14 This platform can devise
the glycolytic and non-glycolytic ECA components by measuring
unsealed (lactate, L-ECA, while CO2 is allowed to escape) and
sealed samples (total, T-ECA, lactate and CO2 combined).
Together with the measurement of the OCR (sealed) and total
This journal is
c
The Royal Society of Chemistry 2011
ATP levels (unsealed), it provides a detailed assessment of ‘cell
energy budget’ (CEB), i.e. the contribution of each of the three main
energy-generating pathways to cell bioenergetics. High flexibility,
sample throughput and potential for further multiplexing are the
other attractive features of this measurement approach.
Nonetheless, to elaborate the effects of mutations and metabolic
deficiencies on CEB, the measurement methodology needs to be
developed further. Many non-lethal metabolic mutations change
the cell phenotype (size, metabolic activity, proliferation rate and
viability), making the comparison of different cells difficult,
particularly when prolonged growth, differentiation and treatments
are involved. In order to achieve reliable and reproducible results,
researchers often have to adjust various parameters for example,
optimize culturing conditions, synchronise growth, and adjust cell
numbers for each cell type.
Here we describe the application of the CEB approach to the
assessment of multiple cell types, which overcomes these difficulties
and provides reliable, good quality data. It involves measurement
of ATP levels, activities of OxPhos, glycolysis and CO2 producing
pathways (e.g. the Krebs cycle) in resting cells and their spare
capacity upon energy stress induced by pharmacological treatment. The utility of this approach is demonstrated with two
representative biological models: mouse embryonic fibroblasts
(MEFs) deficient in fumarate hydratase (Fh1, mouse orthologue
of human FH)7 and HeLa cells with decreased expression of
mitochondrial pyrimidine nucleotide carrier 1 (PNC1).6
Results
Measurement strategy
To assess CEB with a minimal set of experiments, we
performed multiplexed measurement of the OCR, T-ECA,
L-ECA and ATP values for different cells under resting
conditions and upon double-treatment with FCCP/oligomycin
(Fig. 1B). The first two parameters are measured in sealed
samples (i.e. covered with oil to prevent back-diffusion of
ambient O2 and escape of CO214), while the last two parameters are measured in unsealed samples exposed to air. The
OCR and T-ECA values are obtained in kinetic assays which
can be multiplexed13 and expressed as changes of primary
fluorescent parameters over time, i.e. slopes. Cellular ATP is
measured following cell lysis at the end of the experiment,
and this can be multiplexed with L-ECA measurement. For
accurate analysis and cross-comparison of different cell types,
raw values of experimental parameters are normalized for the
cell mass in each sample, which is determined by measuring
total protein concentration. All these assays are performed in
standard 96-well plates with cultured cells. Typically, 30–50
sample wells (to include repeats, different treatments and
controls) are measured in parallel on a multi-label reader with
TR-F capabilities, such as Victor 2 (PerkinElmer) or FLUOstar Omega (BMG).
Genetic mutations, changes in gene expression and protein
activity may affect the bioenergetic status of the cell, directly
or indirectly. In order to compensate for disruption of energy
pathways the cell utilises alternative pathway(s) to maintain
physiological ATP levels, and this can mask the underlying
metabolic defects. For some mutations that have a marked
This journal is
c
The Royal Society of Chemistry 2011
influence on bioenergetics, changes in respiration, glycolysis or
the Krebs cycle can be seen already in resting cells. However,
further analysis of the cells treated with FCCP/oligomycin helps
to visualize minute or masked changes in cell metabolism.
Inhibition of OxPhos by oligomycin can be used to probe
the respiration uncoupled from mitochondrial ATP production.
However, oligomycin alone increases proton motive force
across the mitochondrial inner membrane and elevates the rate
of proton leak,15,16 therefore the results are difficult to interpret.
In contrast, oligomycin/FCCP double treatment strongly
activates the OCR due to mitochondrial uncoupling and prevents
the production and hydrolysis of ATP by mitochondrial
Complex V. To compensate for a decrease in mitochondrial flux,
the cells are forced to increase glycolytic ATP production. For
the purpose of this paper, we assign the terms maximal
respiratory capacity, Rmax, and maximal glycolytic capacity,
Gmax, to this uncoupled metabolic state. These parameters are
assessed by the OCR and L-ECA measurements, respectively,
and are compared with resting OCR and L-ECA values. This
enables the spare capacities within the main pathways to be
probed. Spare capacities are defined as the difference between
Rmax and OCR (or Gmax and L-ECA) at rest. Note that these
parameters relate to intact cells, i.e. not the same as for
permeabilized cells or isolated mitochondria. Increased NADH
and FADH2 consumption by uncoupled mitochondria also
leads to activation of the Krebs cycle and to elevated CO2
production, which can be quantified via L-ECA/T-ECA
measurement. It is worth noting that the Krebs cycle is the
largest, but not the only producer of CO2 in the cell. For
example, the pentose phosphate pathway also produces
CO2.17 In addition, lactate-independent pathways of H+
extrusion (e.g. through the Na+/H+ exchangers18) may contribute to ECA.
A simple experiment with one mutant and one control
cell line can be performed on one plate in approximately
2 hours, including plate preparation and measurement in quadruplicates (four wells). The proposed layout for simultaneous
T-ECA, OCR (with oil seal) and L-ECA followed by ATP
and protein measurement (unsealed samples) is shown in ESIw
(Fig. S1). Negative controls (wells containing probes in cell-free
medium) are also included. Preliminary experiments may be
required to determine optimal seeding cell numbers and FCCP
concentrations which provide reliably detectable OCR and ECA.
For larger panels of cells and for certain cell types it is better to
perform T-ECA and OCR assays separately. Once the ATP,
OCR, T-ECA and L-ECA values are determined for the resting
and uncoupled cells, corresponding spare capacities and L-ECA/
T-ECA and OCR/L-ECA ratios can be analysed. After normalization of these values for protein content, comparison of
different cell types can be conducted. Practical use of this
approach and validation with different biological models are
illustrated below.
Analysis of FH-deficient MEFs
Genetic studies have revealed that the human cancer syndrome
hereditary leiomyomatosis and renal cell cancer (HLRCC)19 is
caused by inactivating mutations of the gene encoding the
Krebs cycle enzyme fumarate hydratase (FH) which converts
Integr. Biol., 2011, 3, 1135–1142
1137
Fig. 2 The involvement of fumarate hydratase in respiration/OxPhos in MEFs. (A) Raw OCRs values for the resting and uncoupled cells. Total
protein content in different samples is indicated above the bars (relative values). (B) OCRs normalized for total protein content. (C) Relative
changes in OCR (%, derived from A) in all cells compared to the wild type MEFs (Fh1+/+). Significant differences are indicated with asterisk
(p o 0.01) and hash (p o 0.001) signs.
fumarate to malate.20 FH is expressed predominantly in
mitochondria, but is also found at lower levels in the cytosol.21
It has been shown in mouse embryo fibroblasts (MEFs) that
are deficient for Fh1 (Fh1 / ) that mitochondrial function at
rest is strongly affected. Additionally, MEFs lacking Fh1
demonstrate reduced respiration and elevated glycolysis.7
We further investigated CEB in the following lines: Fh1 / ,
Fh1 / + FHDMTS (complemented with extramitochondrial
FH lacking the mitochondrial targeting sequence (MTS)),
Fh1 / + FH (complemented with full length FH) and
Fh1+/+ (wild type) MEFs.
Fig. 2A and B shows the main steps in the analysis of resting
OCR and Rmax values. First, raw data representing changes in
the phosphorescence lifetime of the O2-sensing probe (ms min 1)
were normalised for the total protein content in each sample
and averaged (N = 4). Then OCR values were calculated as %
relative to WT control. Fig. 2C shows that resting OCRs were
reduced to 10% in Fh1 / and 20% in Fh1 / + FHDMTS
cells, which is in agreement with previously published data.7
In these cell lines the Rmax was also decreased dramatically
(Fig. 2 and Table 1), being B15 times lower in Fh1 / cells
than in Fh1+/+. Interestingly, in Fh1 / + FH cells the Rmax
was still only 55% of control, indicating that their mitochondrial
function was not fully restored.
The analysis of L-ECA revealed a large activation of
glycolysis in resting Fh1 / and Fh1 / + FHDMTS cell lines
(Fig. 3A). Furthermore, the L-ECA ratio between the resting
and uncoupled cells was significantly higher for these cells
(Fig. 3B). Thus, L-ECA was the same for resting and
uncoupled Fh1 / MEFs, whereas in resting Fh1+/+ and
Table 1 Deficiencies in Fh1 and PNC1 decrease respiratory and
glycolytic spare capacities
Fh1
Spare
capacity
/
Fh1
Fh1
+
Fh+/+ + FH FHDMTS Fh1
Respiration 50
(% of Rmax)
Glycolysis
70
(% of Gmax)
1138
PNC1
/
/
Control PCNA1
shRNA shRNA2
20
0
0
65
0
65
30
5
70
50
Integr. Biol., 2011, 3, 1135–1142
Fh1 / + FH cells it ran at 30–40% of the maximal pace.
Resting Fh1 / + FHDMTS cells maintained their L-ECA
at B70% of the Gmax. Total ATP levels in four MEF lines
varied, reflecting Fh-dependent differences in cell phenotypes
(ESIw, Fig. S2A). However, in all the cells, total ATP levels
remained unchanged during the 60–90 min FCCP/oligomycin
treatment and measurement period, indicating that glycolysis
was able to compensate for the shortage of mitochondrial
ATP. This agrees with the observation that in all cells the
L-ECA reached similar rates upon uncoupling (Fig. 3A). The
T-ECA rates were also seen to be similar in uncoupled cells.
At rest the T-ECA was notably increased in Fh1 / and
Fh1 / + FHDMTS cells (Fig. 3C), although the difference
between them and wild type control was smaller than for the
L-ECA (Fig. 3D). Further analysis revealed that this difference
was due to the contribution of ‘non-glycolytic’ sources of ECA.
Comparison of the L-ECA and T-ECA values showed that
their ratio was significantly increased in Fh1 / and Fh1 / +
FHDMTS cells, suggesting a higher contribution of glycolysis
to the T-ECA (Fig. 3C). From Fig. 3E, the ‘non-glycolytic’
ECA component can be worked out. Finally, a large decrease in
the OCR/L-ECA ratio for Fh1 / and Fh1 / + FHDMTS
cells reflects a strong shift in CEB towards glycolytic ATP
production (Fig. 3F).
Effects of PNC1 protein on CEB of HeLa cells
The mitochondrial pyrimidine nucleotide carrier 1 (PNC1)
controls mitochondrial DNA replication. This protein is
important for mitochondrial function: in HeLa cells in which
PNC1 has been knocked out, the OCR was decreased and
glycolysis activated.6 We further investigated the role of PNC1
in maintenance of CEB by analysing HeLa cells stably
overexpressing shRNA PCN1 (shRNA2 clone with PNC1
transcription reduced by 60%, as shown previously6) and
control HeLa cells overexpressing non-specific shRNA.
We found that in the cells with reduced PNC1 levels, OCR
at rest was reduced by 45% (Fig. 4A and B). Uncoupling did
not increase OCR in shRNA2 cells, showing that their spare
respiratory capacity was low or none. Indeed, Rmax for
shRNA2 cells was B80% lower than for control HeLa cells
This journal is
c
The Royal Society of Chemistry 2011
(ESIw, Fig. S2B). However, when the cells were treated with
FCCP/oligomycin for 3 h, ATP levels decreased significantly in
control cells, but not in shRNA2 cells (Fig. 4H). In contrast,
following incubation for 4 h in galactose(+) medium (ATP
production via glycolysis is inhibited), control cells showed no
changes, while shRNA2 cells showed a 20% decrease in ATP
levels. FCCP/oligomycin treatment in galactose(+) medium
greatly reduced the ATP levels in both cells, with a more
profound change in the PNC1-deficient cells.
Discussion
Fig. 3 The effect of fumarate hydratase on the ECA and CEB. (A)
Normalized L-ECA values (left) show a significant difference between
Fh1 / , Fh1 / + FHDMTS and Fh1 / + FH, Fh1+/+ cells at rest.
(B) The difference becomes more apparent when presented as a ratio of
resting and uncoupled cells. (C and D) T-ECA measurements also
reveal differences between these groups of cells (smaller than for the
L-ECA). (E) L-ECA/T-ECA ratio shows the relative contribution of
glycolysis to T-ECA (%) for the resting cells. F. A strong decrease in
the OCR/L-ECA ratio demonstrates a large imbalance between
OxPhos and glycolysis in Fh1 / and Fh1 / + FHDMTS MEFs,
and a shift towards glycolytic ATP production in these cells. Asterisks
indicate significant difference from wildtype control (Fh1+/+).
(Fig. 4A and B). The reduction in OCR was coupled to a
significant increase in L-ECA (Fig. 4C), which was also the
main contributor to the elevation of T-ECA (Fig. 4D). Fig. 4E
shows that for resting shRNA2 cells L-ECA accounted for
50–55% of the Gmax, and for control HeLa cells, 30–35%, and
that the contribution of L-ECA to T-ECA was increased from
50–55% to nearly 70% (Fig. 4E). A 70% decrease in the OCR/
L-ECA ratio for the cells expressing shRNA PCN1 indicates
that their CEB is biased towards glycolytic ATP production
(Fig. 4G).
Neither reduction in PNC1 expression nor FCCP/oligomycin
treatment for 60–90 min affected the ATP levels in HeLa cells
This journal is
c
The Royal Society of Chemistry 2011
The four-parametrical platform for the assessment of CEB and
bioenergetic abnormalities presented in this study demonstrates
universal features and applicability to different cell systems and
pathologies. It allows detection of various metabolic abnormalities
in a simple set of experiments. Thus, the lack of Fh1 in the MEFs
and the substantial reduction in PNC1 expression in HeLa
cells lead to rather similar effects on CEB. In both cases, we
observe a substantial decrease in the respiratory activity at
rest. Reduction in mitochondrial ATP flux is compensated by
a strong shift towards glycolytic ATP production (Fig. 2–4).
These effects are best seen in Fh1 / and Fh1 / + FHDMTS
cells, which become almost independent of OxPhos energetically.
As a result, treatment with FCCP/oligomycin leads to a minor
increase in L-ECA in these cells. Conversely, Fh1 / + FH and
Fh1+/+ cells robustly increase glycolysis upon inhibition of
OxPhos (Fig. 3A and B).
A decrease in PNC1 expression is known to reduce UTP
levels in the mitochondria,22 thus affecting mitoDNA replication,
ROS production, AMPK-PGC-1a signalling pathway and
mitochondrial biogenesis.6 Here we show that a 60% decrease
in PNC1 expression causes 80% reduction of OCR at rest
(Fig. 4), which is compensated by increased glycolysis. As a
result (and similar to the Fh1 model), treatment of PNC1deficient cells with FCCP/oligomycin causes a significantly
smaller increase in L-ECA, than in control cells.
A comparative analysis of L-ECA and T-ECA in resting
and uncoupled cells shows that the lack of Fh1 in MEFs, or
deficit of PNC1 in HeLa cells, causes a bias in the ECA
towards a glycolytic component. For both models (wild type
cells), the non-glycolytic ECA comprised approximately 50%
of the T-ECA, suggesting a high level of CO2 production by
the Krebs cycle and other pathways (see below) and high
activity of carbonic anhydrases. Interestingly, for the MEF
model, the non-glycolytic flux was not affected by the lack of
Fh1. This can be explained by complex regulation of the Krebs
cycle, redundancies of metabolic pathways and contribution
of other CO2-producing pathways to the ECA. Thus, FH
deficiency is associated with increased activity of pentose
phosphate cycle and expression of malic enzyme (ME1) of
the pyruvate/malate cycle.23 These pathways contribute to
NADPH production and generation of CO2.
Importantly, cellular ATP levels remain unchanged during
the first 60–90 min FCCP/oligomycin treatment for all cell
types in both the models compared (ESIw, Fig. S2). This
indicates that glycolysis is capable of supplying the cells with
energy and cells do not undergo energy starvation. On the
other hand, the Gmax in both models remains similar for all the
Integr. Biol., 2011, 3, 1135–1142
1139
Fig. 4 The role of PNC1 in the CEB of HeLa cells. (A and B) Resting OCR and Rmax are markedly reduced in the cells with low PNC1
expression. (C and D) Reduced OxPhos resulted in significant increase in L-ECA (C) and T-ECA. In resting shRNA2 cells they comprise 450% of
the maximal levels observed in uncoupled cells (E). (F) The ratio of L-ECA/T-ECA in shRNA2 cells is significantly decreased, thus indicating that
PNC1 deficiency leads to increased glycolytic contribution to the T-ECA. (G) The OCR/L-ECA ratio shows that a decrease in PNC1 expression
causes a noticeable shift towards glycolytic ATP production. (H) Cellular ATP levels substantially decrease in shRNA2 cells grown for 4 h in
galactose(+) medium and in control HeLa cells treated with FCCP/oligomycin for 3 h. This confirms that PNC1 deficient cells are less dependent
on OxPhos and rely mostly on glycolysis as an ATP source. Asterisks indicate significant difference.
cell lines compared (Fig. 3A, B and 4C), since in uncoupled
cells glycolysis becomes the main source of ATP. In contrast to
Gmax, Rmax and spare respiratory capacity are reduced
dramatically in MEFs lacking Fh1 and in HeLa cells deficient
in PNC1, therefore upon uncoupling no increase in OCR is
observed in these cells (Fig. 3 and 4, Table 1). Finally, for both
models a decrease in the OCR/L-ECA ratio (Fig. 3F and 4G)
may serve as a good indicator of metabolic changes in
mutant cells.
Experimental
Materials
The O2-sensitive probe MitoXpress,24 pH-sensitive probe
pH-Xtra13 (both used extracellularly) and sealing oil were from
Luxcel Biosciences (Ireland). Carbonyl cyanide 4-(trifluoromethoxy)
1140
Integr. Biol., 2011, 3, 1135–1142
phenylhydrazone (FCCP), oligomycin, collagen IV, growth
media and other reagents were from Sigma-Aldrich. BCATM
Protein Assay kit was from Pierce (Rockford, Ill), CellTiter-Glos
Assay for ATP measurement was from Promega (Madison, WI).
Plasticware was from Sarstedt (Ireland) and Greiner Bio One
(Frickenhausen, Germany).
Cell lines, tissue culture and experimental layout
Mouse embryonic fibroblasts (MEFs) were generated as
described previously.6 These comprised Fh1-deficient (Fh1 / )
MEFs, Fh1 / MEFs that had been stably transfected with
human FH with and without the mitochondrial targeting
sequence (Fh1 / + FH and Fh1 / + FHDMTS, respectively)
and wild-type MEFs (Fh+/+).7 Human cervical carcinoma cells
(HeLa) obtained from the American Tissue Culture Collection
and transfected with non-specific shRNA were used as a wild
This journal is
c
The Royal Society of Chemistry 2011
type control. PNC1 deficient HeLa cells were generated by
stable transfection of a specific shRNA for PNC1, as described
in ref. 6. All cells were cultured in Dulbecco’s Modified Eagle’s
medium (DMEM) supplemented with 4.5 g L 1 glucose, 2 mM
1
L-glutamine, 10% fetal bovine serum (FBS), 100 U ml
1
penicillin and 100 mg ml streptomycin (P/S), at 5% CO2.
For the ECA and ATP experiments, the cells were seeded at
5 104–1.5 105 cells/well on standard 96-well plates
(Sarstedt) pre-coated with 0.01% collagen IV, and left to
attach for 3 h. For the OCR experiments, MEFs were seeded
at 7.5 104 cells/well, and HeLa cells, which have much lower
respiration rates, at 3 105 cells per well. Typically, cells were
seeded in the middle part of the plate to allow uniform
temperature equilibration, and all remaining wells were filled
with medium to maintain uniform temperature and humidity.
OCR assay
Kinetic measurements were conducted as described,24,25 using
100 ml of air-equilibrated DMEM, supplemented with 1 mM
pyruvate, 10 mM glucose, 20 mM HEPES (pH 7.4) and
100 nM of MitoXpress probe. Spare respiratory capacity
was assessed at 1 mM FCCP/10 mM oligomycin added to the
medium 5 min before the measurement. Optimal FCCP
concentrations were determined for both cell models in a set of
OCR experiments. At the end of these experiments, ATP levels
were also measured to check for non-specific cytotoxicity of
FCCP. An equivalent amount of carrier dimethyl sulfoxide
(DMSO) was added to ‘untreated’ cells. Appropriate controls
(e.g. wells without probe and without cells) were also incorporated.
Mitochondrial-independent OCR tested for all the cells treated
with antimycin A (Complex III inhibitor) was undetectable.
Experimental wells were quickly sealed with 150 ml of mineral
oil pre-warmed to 37 1C and the plate was monitored on a
TR-F reader Victor 2 (PerkinElmer Life Sciences) at 37 1C
using a Samarium filter set (340 nm excitation and 642 nm
emission). Each sample well was measured repetitively every
1–5 min over 60–90 min, by taking two intensity readings at
delay times of 30 and 70 ms and gate time of 100 ms. Measured
TR-F intensity signals for each sample well were converted into
phosphorescence lifetime (t) values as follows: t = (t1
t2)/
ln(F1/F2), where F1 and F2 are the TR-F intensity signals at
delay times t1 and t2. From the resulting t profiles, the initial
slopes were calculated (ms min 1) which reflect sample OCR.26
L-ECA and T-ECA Assays
The ECA was measured according to a published modified
method.13 Three hours after cell seeding, the medium was
changed to 200 ml of OCR DMEM (see above) and the plate
was incubated under CO2-free conditions, 95% humidity, at
37 1C for 3 h to release absorbed CO2. The cells were then
washed with unbuffered DMEM (without NaHCO3 or
HEPES), and 100 ml of this medium containing 1 mM pH-Xtra
probe and the stimulants (oligomycin, FCCP/oligomycin or
DMSO) were added to experimental wells. For the T-ECA
assay, 150 ml of mineral oil pre-warmed to 371C were dispensed
to each well, while for the L-ECA assay the wells were left
unsealed. The plate was then measured kinetically on the Victor
2 plate reader at 37 1C for 60–90 min in the TR-F mode using a
This journal is
c
The Royal Society of Chemistry 2011
Europium filter set (340 nm excitation and 615 8.5 nm
emission). Two TR-F intensity signals were measured at delay
times of 100 and 300 ms and a measurement window of 30 ms.
The emission lifetime of the pH-Xtra probe and its increase over
time (ms min 1) were calculated as described for the OCR.
Duplexed OCR and T-ECA assay
When feasible, the OCR and T-ECA assays were multiplexed
in one well. In this case, the cells were prepared as for the ECA
assay in 100 ml of ECA DMEM containing pH-Xtra and
MitoXpress probes (at the same working concentrations),
stimulants and oil seal. Measurements on the Victor 2 reader
were carried out using a two-label protocol, individual label
settings were the same as above. Spectral and decay characteristics
of these probes allow their selective detection without any
cross-talk.13
Total ATP Assay
Total cellular ATP was quantified using the CellTiter-Glos
assay, following the manufacturers’ protocol. Briefly, at
certain time intervals after the addition of FCCP/oligomycin,
the cells were lysed with the CellTiterGlos reagent. After
intensive shaking for 2 min, the samples were transferred into
wells of white 96-well plates (Greiner Bio One) and read on
a Victor 2 (PerkinElmer) plate reader under standard luminescence settings.
Total protein assay
After L-ECA measurement, medium in the sample wells was
replaced with 20 ml of the cell lysis buffer containing 50 mM
HEPES, pH 7.0, 150 mM NaCl, 1 mM EDTA, 1% NP-40 and
protease inhibitors, pH 7.5. Cells were lysed for 15 min at RT,
lysate was collected, pooled for the same sample type and then
analysed for protein concentration using the BCATM Protein
Assay kit as per manufacturers’ protocol. Total protein
content was used then to normalise ATP, T-ECA, L-ECA
and OCR values for the samples analysed. For simplicity,
the protein content in the control cell type was regarded as
1 arbitrary unit (a.u.).
Statistics
All experimental results were evaluated for statistical difference
using a two-tailed Student t-test. The level of confidence of
0.01 was accepted as statistically significant. All processed data
are presented as mean values standard deviation (error bars
on the plots). Each experiment was repeated several times to
ensure consistency of results.
Conclusions
The four-parametrical platform proposed in this study provides
simple high throughput analysis of metabolic abnormalities
affecting the main pathways of ATP production in the cell. It
enables us to probe the involvement of different genes in CEB
and thus represents a useful tool for researchers working in
bioenergetics and cell biology.
Integr. Biol., 2011, 3, 1135–1142
1141
Acknowledgements
Financial support of this work by the Science Foundation of
Ireland, grant (07/IN.1/B1804) and by the European Commission
(FP7 grant NMP4-SL-2008-214706) is gratefully acknowledged.
We thank Dr James Hynes for the useful discussions of results
and Dr Ruslan Dmitriev for technical assistance.
References
1 D. E. Amacher, Curr. Med. Chem., 2005, 12, 1829–1839.
2 M. T. Lin and M. F. Beal, Nature, 2006, 443, 787–795.
3 W. E. Muller, A. Eckert, C. Kurz, G. P. Eckert and K. Leuner,
Mol. Neurobiol., 2010, 41, 159–171.
4 L. D. Marroquin, J. Hynes, J. A. Dykens, J. D. Jamieson and
Y. Will, Toxicol. Sci., 2007, 97, 539–547.
5 A. V. Zhdanov, R. I. Dmitriev and D. B. Papkovsky, Cell. Mol.
Life Sci., 2011, 68(5), 903–917.
6 C. Favre, A. Zhdanov, M. Leahy, D. Papkovsky and
R. O’Connor, Oncogene, 2010, 29, 3964–3976.
7 L. O’Flaherty, J. Adam, L. C. Heather, A. V. Zhdanov, Y. L. Chung,
M. X. Miranda, J. Croft, S. Olpin, K. Clarke, C. W. Pugh,
J. Griffiths, D. Papkovsky, H. Ashrafian, P. J. Ratcliffe and
P. J. Pollard, Hum. Mol. Genet., 2010, 19, 3844–3851.
8 K. Bartlett and S. Eaton, Eur. J. Biochem., 2004, 271, 462–469.
9 D. A. Tennant, R. V. Duran and E. Gottlieb, Nat. Rev. Cancer,
2010, 10, 267–277.
10 D. A. Ferrick, A. Neilson and C. Beeson, Drug Discovery Today,
2008, 13, 268–274.
11 Y. Yang, V. A. Valera, H. M. Padilla-Nash, C. Sourbier,
C. D. Vocke, M. A. Vira, M. S. Abu-Asab, G. Bratslavsky,
M. Tsokos, M. J. Merino, P. A. Pinto, R. Srinivasan, T. Ried,
L. Neckers and W. M. Linehan, Cancer Genet. Cytogenet., 2010,
196, 45–55.
12 L. C. Chao, K. Wroblewski, Z. Zhang, L. Pei, L. Vergnes,
O. R. Ilkayeva, S. Y. Ding, K. Reue, M. J. Watt,
C. B. Newgard, P. F. Pilch, A. L. Hevener and P. Tontonoz,
Diabetes, 2009, 58, 2788–2796.
1142
Integr. Biol., 2011, 3, 1135–1142
13 J. Hynes, T. C. O’Riordan, A. V. Zhdanov, G. Uray, Y. Will and
D. B. Papkovsky, Anal. Biochem., 2009, 390, 21–28.
14 J. Hynes, E. Natoli, Jr. and Y. Will, Curr. Protoc. Toxicol, 2009,
40, 2.16.1–2.16.22.
15 A. S. Divakaruni and M. D. Brand, Physiology, 2011, 26, 192–205.
16 M. Jastroch, A. S. Divakaruni, S. Mookerjee, J. R. Treberg and
M. D. Brand, Essays Biochem., 2010, 47, 53–67.
17 M. M. Wamelink, E. A. Struys and C. Jakobs, J. Inherited Metab.
Dis., 2008, 31, 703–717.
18 W. F. Boron, Am. J. Physiol.: Adv. Physiol. Educ., 2004, 28,
160–179.
19 V. Launonen, O. Vierimaa, M. Kiuru, J. Isola, S. Roth,
E. Pukkala, P. Sistonen, R. Herva and L. A. Aaltonen, Proc. Natl.
Acad. Sci. U. S. A., 2001, 98, 3387–3392.
20 I. P. Tomlinson, N. A. Alam, A. J. Rowan, E. Barclay,
E. E. Jaeger, D. Kelsell, I. Leigh, P. Gorman, H. Lamlum,
S. Rahman, R. R. Roylance, S. Olpin, S. Bevan, K. Barker,
N. Hearle, R. S. Houlston, M. Kiuru, R. Lehtonen, A. Karhu,
S. Vilkki, P. Laiho, C. Eklund, O. Vierimaa, K. Aittomaki,
M. Hietala, P. Sistonen, A. Paetau, R. Salovaara, R. Herva,
V. Launonen and L. A. Aaltonen, Nat. Genet., 2002, 30, 406–410.
21 C. Knox, E. Sass, W. Neupert and O. Pines, J. Biol. Chem., 1998,
273, 25587–25593.
22 S. Floyd, C. Favre, F. M. Lasorsa, M. Leahy, G. Trigiante,
P. Stroebel, A. Marx, G. Loughran, K. O’Callaghan,
C. M. Marobbio, D. J. Slotboom, E. R. Kunji, F. Palmieri and
R. O’Connor, Mol. Biol. Cell, 2007, 18, 3545–3555.
23 H. Ashrafian, L. O’Flaherty, J. Adam, V. Steeples, Y. L. Chung,
P. East, S. Vanharanta, H. Lehtonen, E. Nye, E. Hatipoglu,
M. Miranda, K. Howarth, D. Shukla, H. Troy, J. Griffiths,
B. Spencer-Dene, M. Yusuf, E. Volpi, P. H. Maxwell, G. Stamp,
R. Poulsom, C. W. Pugh, B. Costa, C. Bardella, M. F. Di Renzo,
M. I. Kotlikoff, V. Launonen, L. Aaltonen, M. El-Bahrawy,
I. Tomlinson and P. J. Pollard, Cancer Res., 2010, 70, 9153–9165.
24 J. Hynes, S. Floyd, A. E. Soini, R. O’Connor and
D. B. Papkovsky, J. Biomol. Screening, 2003, 8, 264–272.
25 J. Hynes, T. C. O’Riordan, J. Curtin, T. G. Cotter and
D. B. Papkovsky, J. Immunol. Methods, 2005, 306, 193–201.
26 T. C. O’Riordan, A. V. Zhdanov, G. V. Ponomarev and
D. B. Papkovsky, Anal. Chem., 2007, 79, 9414–9419.
This journal is
c
The Royal Society of Chemistry 2011