ANALYTICAL BIOCHEMISTRY 161,508-5 13 (I 987)

Enzymatic Assays for 2-Deoxyglucose and 2-Deoxyglucose 6-Phosphate*,’

MAGGIE M.-Y. CHI, MARY ELLEN FLJSATERI,JOYCE G. CARTER, BEVERLY J. NORRIS,

DAVID B. MCDOUGAL, JR., AND OLIVER H. LOWRY
Departments of Pharmacology and of Neurology and Neurosurgery, Washington University
School of Medicine, St. Louis, Missouri 63110

Received November 3. I986

Methods for 2-deoxyglucose (2-M;) and 2-deoxyghtcose 6-phosphate (DG6P) are described
which are based on the fact that DG6P is oxidized by glucose-6-phosphate dehydrogenase
(G6PDH). but at a rate lOOO-fold slower than for ghrcose 6-phosphate, whereas hexokinase
phosphorylates 2DG and glucose at comparable rates. Therefore, by adding the two enzymes in
a suitable order, and in appropriate concentrations, 2DG. glucose, DG6P, and glucose 6-P can
all be separately measured. To avoid a side reaction from the use of a high level of G6PDH,
when measuring DG6P, glucose is first removed with glucose oxidase plus aldose reductase.
0 1987 Academic Press, Inc.
KEY WORDS:2-deoxyglucose; 2-deoxyglucose 6-phosphate; glucose-6-phosphate dehydroge-
nase kinetics; enzymatic assay.

2-Deoxyglucose (2DG)’ has become a very
useful tool in the study of brain metabolism.
Glucose and 2DG are similarly transported
into brain and phosphorylated by hexoki-
nase. However, 2-deoxyglucose 6-phosphate
(DG6P) cannot be either converted to an an-
alog of fructose 6-phosphate, or rapidly me-
tabolized otherwise. Therefore, after 2DG
administration, nearly all the DG6P that is
formed accumulates, and this accumulation
can provide a measure of glucose consump-
tion. Because glucose is almost the exclusive
fuel for brain, glucose consumption is a use-
ful measure of brain energy metabolism and
therefore of neuronal activity. In 1977 Soko-
loff et al. described how to exploit this fact,
with [14C]2DG, in a classic paper (1). This
* This paper is dedicated to the memory of Nathan 0.
Kaplan.
’ Supported by grants from the American Cancer So-
ciety (BC-4-29). and the U.S. PHS (NS-08862.
NS-18387, NS-20762).
* Abbreviations used: 2DG, 2-deoxyghrcose; DG6P.
2-deoxyglucose 6-phosphate; G6P, glucose 6-phosphate;
G6PDH, glucose-6-phosphate dehydrogenase; 6PGDH,
6-phosphogluconate dehydrogenase; BSA. bovine serum
albumin: PTZ, pentylenetetrazole.
now widely used radioautographic method
requires a 30- to 45-min time lapse after
2DG injection for dissipation of most of the
unphosphorylated 2DG. The time lag limits
the method to the study of events which
occur on a long time scale.
This report describes sensitive enzymatic
methods for measuring both 2DG and
2DG6P as well as glucose and glucose 6-
phosphate (G6P) in the same tissue sample.
With these methods, it has proven possible to
measure changes in brain glucose consump-
tion on a time scale of 1 to 5 min.
The methods are based on the fact that
both G6P and DG6P can be oxidized by G6P
dehydrogenase (G6PDH) but at rates that
differ by a factor of more than a thousand.
Therefore, the NADH or NADPH produced
with a low level of enzyme is a measure of
G6P, while that produced with a much
higher G6PDH level is DG6P. Glucose and
2DG are similarly distinguished after phos-
phorylation by hexokinase.
For higher levels of these compounds (0.5
nmol or more) the fluorescence of the
NADPH generated is measured directly. For
lower levels (down to a few femtomoles if

0003-2697187 $3.00 508

Copyright Q 1987 by Academic Press, Inc.
Al! rights of reproduction in any form reserved.
 ENZYMATIC ASSAYS FOR 2-DG AND DG6P 509

necessary), the NADPH is amplified by en- Direct Assays (0.5 to IO nmol

zymatic cycling (2,3). in Each Aliquot)
Two problems were encountered, both re-
sulting from the use of such a high level of
G6PDH. Commercial preparations of The analysis of a solution such as a tissue
G6PDH almost invariably contained traces extract which contains all four components,
of 6-phosphogluconate dehydrogenase. This 2DG, 2DG6P, glucose, and G6P, is made in
caused errors in DG and DG6P readings two parts.
from partial oxidation of the 6-phosphoglu- Part A. An aliquot of the solution is added
conate that had been formed from G6P or to a fluorometer tube containing 1 ml of a
glucose in previous steps. Fortunately, it was reagent composed of 50 mM Tris-HCl, pH
possible to destroy 6-phosphogluconate de- 8.1 (acid:base, l:l), 300 CIM ATP, 1 mM
hydrogenase in G6PDH from baker’s yeast MgC&, and 100 j.tM NADP+.
by brief incubation at pH 4.6. The second Step 1 (G6P). After taking an initial read-
problem was the slow oxidation of glucose by ing, 0.2 cLg(0.03 U) of G6PDH from baker’s
all three commercial types of G6PDH. This yeast is added in a volume < 10 ~1, and a
may be a side reaction of the G6PDH en- second reading is made 20 min later.
zymes themselves. This caused an error only Step 2 (glucose). Without delay, 10 pg of
when measuring DG6P, since DG was yeast hexokinase is added in a similar vol-
always measured together with glucose or ume and a third reading is made 20 min
after glucose had already been converted to later.
6-phosphogluconate. The cure was to de- Step 3 (NADPH destruction). Again with-
stroy the glucose with glucose oxidase before out delay, 25 ~1 of 2 M HCl is added and well
the DG6P reaction. Because the oxidase only mixed. This is followed after at least 10 min
attacks a-glucose, aldose mutarotase had to by 25 ~1 of 2 M NaOH.
be included also. Step 4 (2DG + DG6P). After an initial
reading is made, 75 pugof G6PDH are added
in a volume of 10 J or less, and a final read-
ing is made 50 min later.
MATERIALS AND METHODS
Part B. A similar aliquot of the solution is
added to 1 ml of reagent composed of the
Enzymes, as well as 2-deoxyglucose and same buffer used in Part A (no ATP or Mg),
2-deoxyglucose 6-phosphate, were from plus 100 pM NADP+, 50 pg/ml of glucose
Sigma Chemical Co. or Boehringer-Mann- oxidase from Aspergilfus niger (about 5 IU),
heim. The G6PDH was treated to destroy and 1 pg/ml of aldose mutarotase.
contaminating 6-phosphogluconate dehy- Step 1 (G6P). The same as step 1 of Part A.
drogenase (6PGDH) by incubating for 60 Step 2 (DG6P). Without delay, after the
min at 38°C in 50 mM pH 4.6 sodium ace- second reading for Step 1,50 pg of G6PDH is
tate buffer (acid:base, 1:l). This was then added and a third reading is made 50 min
neutralized with a 0.03 vol of 1 M imidazole later. (The lower level of G6PDH than in
base. 6PGDH destruction was partially Step 4 of Part A is sufficient, because of the
blocked by levels of (NH&SO4 as low as 0.1 absence of ATP which is somewhat inhibi-
M. Therefore G6PDH suspensions in strong tory.1
(NH&SO4 were centrifuged and dissolved in Calculations are based on standards and
the original volume of the sodium acetate blanks carried through the procedures from
buffer for treatment. At least 95% of the the beginning. 2DG is calculated by the dif-
6PGDH should be removed without loss of ference between the sum of 2DG + DG6P
G6PDH. (Part A, Step 4) and DG6P (Part B, Step 2).
510 CHI
Indirect Assay (25 to 500 pmol
in Each Aliquot)
It is assumed that each aliquot consists of
50 ~1 of a 0.02 M HCl tissue extract.)
Part A. Step 1 (G6P). The aliquot is added
to a 7 X 75-mm glass tube. To each aliquot is
added 50 ~1 of 80 mM Tris-HCl (acid:base,
20:60) containing 600 FM ATP, 2 mM
MgC12, 200 pM NADP+, and 0.4 pg/ml of
G6PDH and 0.04% bovine serum albumin
(BSA). After 20 min a lo-p1 aliquot is trans-
ferred into another tube of the same size,
which already contains 50 ~1 of 0.025 M
NaOH.
Step 2 (G6Pplus glucose). Within 10 or 15
min, 1 ~1 of a 1 mg/ml solution of hexoki-
nase is added to the remaining 90 ~1 from
Step 1, and after 20 min another IO+1 ali-
quot is transferred into a tube containing
NaOH as in Step 1.
Step 3. The NADPH is now destroyed as
in the direct assay by first adding 2 ~1 of 2 M
HCl to the remaining 80 ~1, followed after at
least 10 min by 2 ~1 of 2 M NaOH.
Step 4 (2DG + DG6P). To each tube is
added 5 ~1 of 1.5 mg/ml G6PDH, and after
50 min a lo-p1 aliquot is transferred into a
tube with NaOH as in Step 1.
Part B. Step I (G6Pplus DG6P). The 50-~1
aliquot is added to 50 ~1 of 80 mM Tris buffer
(acid:base, 20:60) which contains 50 Mg/ml
glucose oxidase, 1 p/ml mutarotase, 200 PM
NADP+, and 0.02% BSA. After 50 min, 5 ~1
is added of 1 mg/ml G6PDH. After incuba-
tion for 40 min, a lo-p1 aliquot is added to a
tube with NaOH, as in Part A.
Amplification Procedure
The four tubes with NaOH plus aliquots
from Steps 1, 2, and 4 of Part A, and Step 1
of Part B are heated for 5 min in a water bath
at about 95°C. A lo-ccl aliquot is added to
100 ~1 of NADP cycling reagent with enzyme
levels to give about 2000-fold amplification
in 1 h at 25°C. The cycling is terminated by
heating for 3 min in a water bath at about
95°C. See Refs. (3) and (4) for details of the
ET AL.
cycling procedure, and the final indicator
step to measure the 6-phosphogluconate
produced in the cycle.
Calculations are based on standards and
blanks carried through all steps. Note that
although the 1O-p1 aliquots from the different
steps represent slightly different fractions of
the original sample (because of addition of
G6PDH, acid, and alkali), these differences
affect the standards and samples alike.
To keep the procedures which require am-
plification as simple as possible, three of the
components are calculated indirectly: glu-
cose by subtracting G6P from G6P + glu-
cose, DG6P by subtracting G6P from G6P
+ DG6P, and DG by subtracting DG6P
from DG6P + DG. This is likely to be quite
satisfactory for glucose and DG6P, because
G6P is usually very low in brain. The calcu-
lation of DG by difference may be more sub-
ject to error, since in some cases DG6P may
equal or exceed DG.
Each of these three components can be
measured separately if preferred: Glucose
and DG6P can be measured directly by first
destroying the NADPH from G6P by merely
reversing Steps 2 and 3 of parts A and B,
respectively.
DG direct measurement is more compli-
cated:
Step 1. The reagent is the same as for Part
A, but G6PDH is increased to give a concen-
tration of 75 pg/ml, and incubation is for
50 min.
Step 2. HCl is added as in Step 3 of Part A,
but the samples are heated for 5 min at 95°C
before neutralizing with NaOH. This is to
destroy the high level of G6PDH, as well as
the NADPH from G6P and DG6P.
Step 3. Hexokinase and G6PDH are added
to give concentrations of 10 and 0.2 &ml
respectively, with incubation for 20 min.
Step 4. The NADPH from glucose is de-
stroyed with HCl, but without heat. This is
then neutralized with NaOH as before.
Step 5. A high level of GGPDH is added, as
in Step 1, to yield NADPH from DG alone.
 ENZYMATIC ASSAYS FOR 2-DG AND DG6P 511

Other Variations in the Assay
Aside from greater sensitivity with fluoro-
metric rather than spectrophotometric mea-
surement, the direct assay is simpler in fluo-
rometer tubes, since the long duration of
some of the steps would make it awkward to
carry out a large experiment in the spectro-
photometer without a prohibitive number of
cells. However, standardizations are best
made spectrophotometrically, and since each
component is measured separately there is
no need for glucose oxidase and mutarotase.
The methods have been adapted and used
satisfactorily to measure much smaller
quantities than those measured by the proce-
dures given here. Adaptation to measure
levels in the 0. I pmol range simply requires
reduction in volumes to the microliter and
submicroliter level. For these small volumes
we have used the oil-well technique (2). Full
description of these modifications would be
out of place here, but the assay principles and
style are the same as presented.
EXPERIMENTAL
Kinetics of G6PDH from Baker’s Yeast
For analytical applications the levels
G6P and DG6P will be well below their
spective K,‘s. Therefore, the significant
netic value is the first-order rate constant
the chosen NADP+ concentration. Table
gives the first-order rate constants for en-
zyme levels of 1 &ml, with the NADP+
concentration extrapolated to infinity (&&J.
This constant is 1500 times greater for G6P
than for DG6P. The difference is due to a
combination of a much greater V,,, and a
much smaller K,,, (Table I).
To achieve 99% conversion kt = 4.6 (i.e.,
-In O.Ol), where k = kSP X pg/ml G6PDH.
Therefore for 99% conversion in 20 min with
G6P and in 50 min with DG6P would re-
quire a minimum of 4.6/(20 X 2.7) = 0.09
pg/ml G6PDH and 4.6/(50 X 0.00 18) = 5 1
pg/ml G6PDH, respectively. With 0.2 pg/ml
G6PDH for 20 min, as recommended for
G6P, there is a calculated 0.7% oxidation of
DG6P. This is analytically insignificant with
levels of the components likely to be en-
countered in biological material. It is better
to err slightly on this side rather than on the
other, for the following reason. Glucose will
usually be much higher than 2DG. If even a
few percent of the G6P from glucose fails to
be oxidized at the low G6PDH step, it will
cause a large percentage error in the 2DG
evaluation.
of
re- The Kinetics of Hexokinase with 2DG
ki-
at The kinetics of baker’s yeast hexokinase
I were measured with both glucose and ~DG

TABLE 1
KINETICS OF G6PDH FROM BAKER’S YEASTY

Constant substrate Kttl VILX kSp

mm
(PM) Variable substrate (fiM) (firno mg-’ min-‘) (min-‘)
G6P, 500 NADP+ 4.4 70
G6P, 5 NADP+ 10.8 2.1
NADP+, 500 G6P 19 57
DG6P, 5 NADP + 1.4 0.0018
NADP+, 500 DG6P 540 0.63
n Measurements were made at about 23°C in 50 mM Tris-HCl, pH 8.1, with 0.02% BSA. The G6PDH was about
30% pure. The Km’s are for the variable substrate. V’,,,= is the calculated value for the variable substrate extrapolated
to infinity. The constant kg= is the first-order rate constant for G6P or DG6P with 1 &ml of G6PDH and the
NADP+ concentration extrapolated to infinity.
512 CHI ET AL.

TABLE 2 either glucose or DG in 2 or 3 min with the

KINETICSOFHEXOKINASE' recommended hexokinase level.

Vmm
(firno mg-’ G’ax Reproducibility and Illustrative Data
Substrate min-‘) (min-‘)

Glucose 111 77 0.69 In the 1 to 10 nmol, direct-reading range,

2DG 595 210 0.35 reproducibility for DG and DG6P is limited
only by care in pipetting, instrumental stabil-
“Measured in 50 mM Tris-HCI buffer, pH 8.1, con- ity, etc, and the coefficient of variation (CV)
taining 300 pM ATP at 23’C. The velocities with glucose
should not exceed 2%. CV’s in the 10 pmol
were followed directly in the fluorometer by the
NADPH formed with G6PDH. The 2DG velocities were range have averaged 3 to 4%, and in the 0.5
measured indirectly by stopping the reaction with HCl pmol range have averaged about 6%. In an
and measuring the DG6P in a separate step. Because the experiment with 100 ng samples (dry weight)
ATP concentration was only about twice the K,,,, the from 4 areas of the hippocampus, 2DG levels
true Vmar‘s would be about 50% greater than the V,lmax’~ for quadruplicate samples had an average
shown. Vi=, is V,,,,,,/K,,,,and represents the first-order
rate constant with a hexokinase concentration of 1 cv of 4.7%.
a/ml. An illustration of the applicability of the
procedures for all four substances is given in
Table 3. Mice were given a convulsive dose
under analytical conditions. The I’,,,,, for of pentylenetetrazole (PTZ) 1 min after an
Ix; was actually higher than for glucose by a injection of 2DG, and were frozen 0.5 to 3
factor of 2.7 (Table 2). However, the K,,, was min later in liquid N2. Samples of the cere-
also higher for DG by an even greater margin bral cortex weighing about 50 mg were dis-
(5 .Cfold). Therefore, the analytically signifi- sected out at -20°C and extracted at this
cant first-order rate constant is about half as temperature with 1 ml of 0.1 M HCl in abso-
large for 2DG as for glucose. There is clearly lute methanol. After dilution with aqueous
no problem in complete phosphorylation of 0.05 M HCl, the samples were centrifuged,

TABLE 3

Min after 2DG DG6P Glucose G6P

DG PTZ (rmollkd

1 0 448 + 36 180 + 20 145Ok 170 115? 16

1.5 0 445 +40 251 f 16 134Ok 90 87+ 7
1.5 0.5 470+21 227e 16 1540+ 80 look 9
2 0 454+ 16 345* I1 123Ok 40 80+ 6
2 1 379+ 18 389k 36 115Ok 60 76? 2
3 0 459k 9 430+31 1610f 140 84+ 11
3 2 214+31 750+67 79Ok 40 65+ 5
4 0 437+ 7 477k37 172Ok 100 79k 4
4 3 248+ 10 845 f 62 96Ok 70 70+ 8

’ Mice were injected with 1 mmol/kg of 2DG in a tail vein at zero time and with 100 mg/kg of pentylenetetrazole
intraperitoneally 1 min later. They were killed by immersion in liquid N2 at the times shown. Each entry is the
average & SE for four mice at 1 to 3 min and three mice at 4 min atier 2DG injection.
 ENZYMATIC ASSAYS FOR 2-DG AND DC6P 513

and the supematants were heated for 5 min One warning may be in order. In many
at 95°C and analyzed by the direct fluoro- tissues, including the brain, brief ischemia
metric procedure. After a lag period of about will cause a rapid increase in glycolysis and
1 min following PTZ injection, there was a therefore conversion of 2DG to DG6P at a
dramatic rise in DG6P and fall in 2DG and rate far in excess of that corresponding to
glucose, indicating that glucose consumption normal glucose consumption. This possibil-
had exceeded the rate of transport of 2DG ity should therefore be kept in mind in apply-
and glucose into the brain. ing the proposed enzymatic methods in en-
The methods, with appropriate adjust- ergy metabolism studies.
ment of the procedures, have been success-
fully applied to the analysis of brain samples REFERENCES
ranging from 50 mg wet wt to 0.1 hg dry wt 1. Sokoloff, L., Reivich, M., Kennedy, C., Des Rosiers,
(equal to 0.5 pg wet wt). M. H., Patlak, C. S., Pettigrew, K. D., Sakurada.
O., and Shinohara, M. (1977) J. Neurochem. 28,
897-916.
DISCUSSION 2. Lowry, 0. H., and Passonneau, J. V. (1972) A Flex-
ible System of Enzymatic Analysis, Academic
The presentation of the proposed methods Press, New York.
has been aimed at measurements in brain in 3. Lowry, 0. H. (1980) Mol. Cell. Biochem. 32,
general, and mouse brain in particular. 135-146.
There is, of course, no reason why they are 4. Chi, M. M.-Y., Lowry, C. V., and Lowry, 0. H.
limited to brain. An early 2DG study by (1978) Anal. Biochem. 128, 186-190.
5. Hintz, C. S., Chi, M. M.-Y., and Lowry, 0. H.
Kipnis and Cori was made with rat dia- (1978) Anal. Biochem. 89, 119-129.
phragm (6), and the Sokoloff procedure has 6. Kipnis, D. M., and Cori. C. F. ( 1959) J. Biol. Chem.
been used with many other types of tissue. 234, 171-177.