An Isozyme of Hexokinase Specific for the Human Red Blood Cell (HK,)
By Koko Murakami, Francine Blei, William Tilton, Carol Seaman, and Sergio Piomelli
The hexokinase (HK) of the human red blood cell (RBC)was
separated into two distinct major isozymes by fast protein
liquid chromatography using a linear salt gradient on a
MonoQ column. The first isozyme (HK,) eluted as a sharp
peak at the same position as HK, of human liver. The
second isozyme eluted between HK, and HK,, of human
white blood cells, and it appeared t o be unique t o the RBC
(it was designated HK,). From a gel filtration column, HK,
eluted before HK,, suggesting that it was larger than HK, by
several kilodaltons. In a mitochondria-enriched fraction
from human reticulocytes, no HK, was found; thus, HK,
was not a mitochondrial enzyme. Despite these differences
in chromatographic behavior, size, and mitochondrial binding, both forms behaved kinetically as HK,. RBC from
normal blood contained HK, and HK, at an equal activity,
but in reticulocyte-rich RBC, HK, dominated. When RBC of
increasingage was separated by buoyant density ultracentrifugation. the total HK activity decayed in a biphasic
manner, with half-lives respectively of -1 5 and -51 days.
When isolated by MonoQ column from each age-separated
fraction, HK, was the major form in the youngest RBC, and
decreased rapidly with cell age, with a t,,, of -10 days,
representing a negligible activity in the oldest RBC. Instead, HK, was relatively stable through the entire life span
of the RBC. with a t,,, of -66 days. Thus, HK, appears t o be
an RBC-specific isozymethat is predominant in the reticulocyte and is then rapidly degraded. During maturation of the
RBC, the fast decay of HK, contributes to the early sharp
decline of HK activity and the slow decay of HK, to the later
gradual decline.
0 1990 by The American Society of Hematology.
T
one or more of the other isozymes (HK,,, HK,,,, HK,,) may
be found in each tissue to constitute a specific set of isozymes.
In the RBC, the heterogeneous forms of HK, were
described originally by electrophoretic ~eparation.8.~
With
column chromatography, a subtype was well-separated from
HK, in rabbit RBC." However, in humans the isozymes of
the RBC have been reported by one group as heterogeneous
forms of HK, (multiple peaks in the HK, area),''-'* while
others have reported two isozyme^.'^ The apparent inconsistency in isozyme types between species and the contradictory
results with the human isozymes have hampered further
studies of the molecular origin and regulation of the HK
isozymes of human RBC. Nonetheless, one isozyme in rabbit
RBC has been shown to be prevalent in the reticulocyte and
to preferentially decline during the aging process.I4
We used a fast protein liquid chromatography (FPLC)
system with a MonoQ column (Pharmacia, Piscataway, NJ)
to separate the HK isozymes from human tissues. This
method was extremely efficient to characterize the human
RBC HK isozymes because of its high resolution and rapid
separation. Our results indicate that HK, (the HK isozyme
present in the human RBC along with HK,) is an isozyme
specific to the RBC, which rapidly decays during aging of the
RBC.
HE RETICULOCYTE loses most of its intracellular
organelles during maturation into the red blood cell
(RBC), which is incapable of protein synthesis. The RBC
then remains dependent on the preformed enzymatic machinery throughout its life span of approximately 120 days.
Therefore, the RBC appears to be an excellent model for
studies of the molecular aging of enzymes.
Previous studies from our laboratory suggested that in
senescent RBC there is a summation of metabolic failures,
rather than just molecular aging of a single enzyme.' These,
together with the progressive appearance of a senescent
antigen,2 may play an important role in the removal of the
senescent RBC from the circulation. HK (EC 2.7.1.1.)
appears to contribute most to the metabolic R B C
senescence,'~~-~
because (1) the energy production in the RBC
is derived entirely through glycolysis; (2) glucose transport
does not limit the rate of glucose utilization; (3) HK appears
to be a rate-limiting step because it has the lowest activity
among glycolytic enzymes and marked regulatory properties
with various intermediates; and (4)the H K activity sharply
declines during the aging processes.
Four isozymes of H K are found in mammalian tissues.'
These are called HK,, HK,,, HK,,,, and HK,, according to
their order of elution from a DEAE-anionic exchange column. HK, is generally present in most tissues; in addition,
From the Division of Pediatric Hematology/Oncology. Columbia
University College of Physicians and Surgeons, New York, NY.
Submitted July 27, 1989; accepted October 3. 1989.
Supported in part by National Institutes of Health Grant
AM26793.
Address reprint requests to Sergio Piomelli. MD, Pediatric
Hematology. Columbia University College of Physicians and Surgeons. 630 W 168th St, New York. N Y 10032.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
"advertisement" in accordance with 18 U.S.C.section 1734 solely to
indicate this fact.
0 1990 by The American Society of Hematology.
0006-4971/90/7503-0023$3.00/0
770
MATERIALS AND METHODS
Preparation of human '@we RBC" suspensions. Blood samples
from healthy adult volunteers were defibrinated in a syringe with
glass beads. The blood was diluted with five volumes of phosphatebuffered saline containing 10 mmol/L glucose (BSG), and filtered
through two layers of Whatman #2 paper.' After filtration, the cells
were washed twice with BSG and packed by centrifugation. The
resulting suspension was essentially free of platelets, contained less
than 1% of the original leucocytes, but retained an unchanged
percentage of retic~locytes.'~
Reticulocyte-rich samples, obtained
from the blood of patients with sickle cell anemia discarded at time
of therapeutic exchange transfusion, were similarly filtered.
Preparation of HK from RBCs. RBC suspensions were lysed
with a hypotonic buffer (5 mmol/L Tris-CI, pH 7.4, 0.5 mmol/L
dithiothreitol, 0.5 mmol/L glucose, and 1 mmol/L ethylenediamine
tetra acetic acid [EDTA] ). After centrifugation at 30,000 x gfor 30
minutes, the supernatant was loaded on DE52 (Whatman, Clifton,
Blood, Vol 75, No 3 (February 1). 1990: pp 770-775
77 1
HEXOKINASE ISOZYME SPECIFIC FOR HUMAN RBC
NJ) equilibrated with column buffer (20 mmol/L Tris-C1, pH 7.4,
0.5 mmol/L dithiothreitol, 0.5 mmol/L glucose, and 0.1 mmol/L
EDTA). The hemoglobin was first washed out with column buffer,
then the protein fraction containing HK was eluted with 0.5 mol/L
KCI in the same buffer, and precipitated by 65% saturated ammonium sulfate. After dialysis against column buffer, the protein
fraction was chromatographed on a BioGel 0.5m column (1.5 x 100
cm) (Biorad, Richmond, CA) equilibrated with the same buffer. The
peak fractions containing HK activity were combined, passed
through a millipore filter (200 pm) (Millipore, Bedford, MA), and
stored at -7OOC.
Preparation of a mitochondria-enriched fraction from human
reticulocytes. The reticulocyte-rich (1 6% to 25%) RBC suspension, obtained from sickle cell patients as described above, was lysed
with the hypotonic buffer. The isotonicity was then rapidly restored
by addition of the appropriate amount of a concentrated NaCl
solution, and the lysate was centrifuged at 1,600 x g for 10 minutes.
From the resulting supernatant, the particulate fraction was obtained by centrifugation at 30,000 x g for 20 minutes. This sediment
was then washed twice with a mannitol buffer containing MgCI,
(0.25 mol/L mannitol, 20 mmol/L Tris-HC1, pH 7.4, 0.5 mmol/L
dithiothreitol, 0.5 mmol/L glucose, 1 mmol/L MgCI,). The final
pellet, enriched with mitochondria, was colorless, and thus presumably free of hemoglobin and cytoplasmic protein. Measurements of
hexokinase activity showed that approximately 5% of the total lysate
activity was contained in this fraction. It was then divided into two
aliquots, one suspended in the mannitol buffer and the other in the
mannitol buffer containing 0.2 mmol/L glucose-6-phosphate (G6P).
Preparation of platelet and white blood cell (WBC) suspensions.
Blood collected in EDTA (1 8 mL) was centrifuged at 600 x g for 3
minutes to separate the platelet-rich plasma (PRP). From the PRP
the platelets were precipitated by centrifugation at 1,800 x g for 15
minutes: these preparations contained essentially platelets (platelet
suspension). The sedimented mixture of RBC and WBC was then
loaded on a dextran solution (1.25% dextran and 0.8% NaEDTA in
0.9% NaCI) and allowed to separate by gravity for 30 minutes at
37OC. The supernatant dextran layer contained 72% and 33%,
respectively, of the original WBC and platelets. These were concentrated by centrifugation at 1,800 x g for 15 minutes (WBC plus
platelet suspension). Contaminating RBC was removed from both
preparations by lysis with the hypotonic buffer for 1 minute, followed
by restoration of isotonicity with NaCI. The colorless pellet of cells
was then suspended in 2 mL of hypotonic buffer, and lysed by three
cycles of freezing (-70°C) and thawing. The supernatant, after
centrifugation at 30,000 x g for 30 minutes, was directly applied on
a MonoQ column.
Preparation of liver extracts. Human liver obtained at autopsy
was homogenized in three volumes of column buffer with an Omni
mixer (Dupont, Wilmington, DE), then centrifuged at 30,000 x g
for 30 minutes. From the resulting supernatant, the HK fraction was
obtained by the same procedure described above for RBC.
MonoQ column chromatography. A MonoQ anionic exchange
column (0.5 x 5 cm) was used in an FPLC system (Pharmacia).
Using a GP250 gradient programmer, a linear gradient of KCI was
established, using two buffers (no-salt buffer: 50 mmol/L Tris-C1,
pH 7.4,O.S mmol/L dithiothreitol, 0.5 mmol/L glucose, and 0.05%
Na Azide; high-salt buffer: 1.0 mol/L KCl added to no-salt buffer).
After the injection of the enzyme-containing solution, the proteins
were first eluted with 95% no-salt buffer and 5% high-salt buffer for
10 minutes, then with a linear gradient of 5% to 35% high-salt buffer
(50 to 350 mmol/L KCl) over 10 to 40 minutes. The flow rate was
1.0 mL/min; 0.5 mL was collected in each fraction. The column was
run at room temperature, and the eluate was collected on ice. The
concentration of KCI in column fractions was verified using a
CDC8O conductivity meter (Radiometer American, Westlake, OH).
Repeated analyses of several preparations were reproducible without
need for regenerating the MonoQ. The recovery of HK activity was
70% to 80%.
Enzyme assays. HK activity was measured at 37OC by coupling
with glucose-6-phosphate dehydrogenase, and pyruvate kinase activity was measured at 25OC by coupling with lactate dehydrogenase as
described previously.’
Protein and hemoglobin determination. Protein concentrations
were measured by the method of Lowry et all6 using bovine serum
albumin as reference. Hemoglobin was quantitated after conversion
to cyanmethemoglobin.”
Buoyant density ultracentrifugation of RBC. Discontinuous
density gradients (d 1.0973, 1.1010, 1.1048, 1.1085, 1.1122) of
arabinogalactane were prepared in a cellulose nitrate tube (1.6 x 10.2
cm) as previously reported.’* Two milliliters of RBC suspensions
diluted in BSG to a hematocrit of approximately 70% (equivalent to
-4 mL blood) was placed on top of the gradient, which was then
ultracentrifuged in a Spinco model (Beckman, Somerset, NJ), using
the SW27 swinging bucket rotor at 60,000 x g, for 45 minutes. Six
layers of cells were at the interfaces. They were isolated by slicing
the tube, and washed with BSG. The hemoglobin content of each
fraction was measured and used to establish the cumulative distribution function of each layer; from this the average cell age for each
layer was estimated as previously p~b1ished.l~
RESULTS
HK isozyme projile of human RBC and other human
tissues. It has been difficult until now to reproducibly
separate the HK isozymes of RBC due to their similar size
and charge. This study used the MonoQ anionic exchange
column in an FPLC system (Pharmacia), a technique that is
particularly successful in analyzing various isozymes because
of its fine and rapid resolution.
The HK was adsorbed onto the MonoQ in the no-salt
buffer, then eluted by a linear gradient from 50 to 350
mmol/L KCl in the same buffer. Figure 1 A shows that the
HK of human RBC was well-separated into two peaks, one
eluting at 120 mmol/L KCl (HK,) and another at 160
mmol/L KCl (HK,). The elution was completed within 30
minutes, at a flow rate of 1 mL/min. T h e resolution and
rapidity of the MonoQ column exceeded by far those of
conventional columns, such as DEAE-cellulose, phosphocellulose, etc. Although much of the data reported in this
report result from the analysis of partially purified enzymes
on DEAE-cellulose, it must be noted that a similar pattern of
separation was obtained with crude lysate; thus there was no
preferential loss of any isozyme during preparation.
To identify the specificity of the RBC enzyme, the HK of
human platelet suspensions, of mixed WBC and platelet
suspensions, and of liver extracts was similarly analyzed by
MonoQ column chromatography. T h e platelet HK consisted
only of HK, that eluted at the same salt concentration (120
mmol/L KCl) as HK, from RBC (Fig IB). T h e WBC plus
platelets preparation contained both HK, and HK,,, eluting
a t 120 and 185 mmol/L KCI, respectively (Fig 1B). Therefore, the second form (HK,) of RBC HK, which eluted at
160 mmol/L KCl, a lower salt concentration than for HK,, of
WBC, cannot represent either W B C or platelet contamination. T h e H K from liver was separated by the MonoQ
772
MURAKAMI ET AL
-
c
2
.2
1
.1
2
.2
1
.1
I
I
E
\
I
3
Y
E
v
y 2
I
1
FRACTION NUMBER
Fig 1. HK isozyme profiles from a MonoQ column (FPLCI in
various tissues.
column into two major forms: HK, that eluted at 120
mmol/L KCI, and HK,,, that eluted at 205 mmol/L KCI,
later than HK,, from WBC (Fig lC), and was inhibited by
glucose at a high concentration. A subtype of HK, was
constantly observed in the liver extract, which eluted as a
preceding shoulder to HK,. Thus, the second isozyme of RBC
(HK,) eluting at 160 mmol/L KCI between HK, and HK,,
appeared different from any HK isozyme in other blood cell
types or other tissues tested, and it most likely represents a
unique isozyme of the RBC.
The kinetic parameters of HK, and HK, of RBC were
compared. The Km (Michaelis constant) for glucose was
0.064 mmol/L (95% confidence limits, 0.04 to 0.1 mmol/L
for HK,, and 0.05 to 0.09 mmol/L for HK,) for both
isozymes; the Km for adenosine triphosphate was also
comparable: 1.38 mmol/L (95% confidence limits, 0.95 to
2.3 mmol/L) for HK, and 1.27 mmol/L (95% confidence
limits, 0.9 to 3.33 mmol/L) for HK,. Thus, the RBC-specific
isozyme behaved kinetically as HK,.
HK from rabbit RBC similarly prepared was also analyzed on the MonoQ column. Figures 1A and 1D compare
the isozyme profiles of human and rabbit RBC. In both cases
there were two distinct peaks; however, while both first peaks
eluted at 120 mmol/L KCl, the second peak eluted at 160
mmol/L KCI in the case of human RBC, while in the case of
rabbit RBC it eluted later at 180 mmol/L KCI. Thus,
although there were two isozymes in both species, the second
one was much more clearly separated from the first in the
case of the rabbit RBC than in the case of human RBC.
The RBC-specijic isozyme of human HK (HK,) is larger
than HK,. When the hemolysate of human RBC was first
directly applied on DEAE-cellulose to wash out hemoglobin
and later chromatographed on a BioGel column, the HK
eluted in a sharp peak in fractions 38 to 41 (Fig 2, inset). The
fractions were not pooled but were instead individually
analyzed on MonoQ columns. In the first of the BioGel
fractions (no. 38), HK, represented most of the activity with
a small percentage of the activity due to HK, (Fig 2). On the
other hand, HK, dominated in the last BioGel fraction (no.
41). Thus, since HK, clearly migrated faster than HK, on
BioGel, HK, appeared to be larger than HK, by several
kilodaltons.
HK,, but not HK,, is eluted by G6P from the particulate
fraction of human reticulocytes. The particulate-bound
HK in a mitochondria-enriched fraction from human reticulocytes, which contained approximately 5% of the total
hexokinase activity in the original lysate, was solubilized
with 0.2 mmol/L G6P for 10 minutes at 3OOC. The supernatant from a l-mL suspension (pel1et:buffer ratio = 1:l)
*Fr
mFr
oFr
oFr
38
39
40
41
E
BIOGEL
36 30 40 4
FRACTION NUMBER
Fig 2. HK isozyme profiles of BioGel column fractions. The
high-salt eluate from a DEAE-cellulose column (0.4 U HK, 25 mg
protein) was analyzed on BioGel A 0.5m (inset). BioGelfractions 38
to 41 (10 to 20 mu) were individuallyanalyzed on MonoQ columns.
773
HEXOKINASE ISOZYME SPECIFIC FOR HUMAN RBC
contained 10.2 mU of HK in the G6P-containing preparation
and 2.4 m U in the untreated one. Thus, the bound form of
RBC HK eluted specifically with G6P, as shown for bovine
brain HK by Rose and Warms.” The G6P-solubilized HK
was analyzed on a MonoQ column; no HK, was found; thus,
also in the human RBC, HK, is not a mitochondria-bound
isozyme. The G6P-solubilized HK eluted in a closely overlapping peak with HK,. We could not clearly identify an
additional mitochondrial-bound form of HK that was separable from HK, by ionic exchange high performance liquid
chromatography (HPLC) as Stocchi et alZ0 recently described in rabbit RBC.
HK, ispredominant in the reticulocyte. The HK isozyme
patterns of normal and reticulocyte-rich RBC suspensions
were compared (Fig 3A). RBC from normal blood contained
approximately equal activities of HK, and HK,. On the other
hand, HK, was predominant in reticulocyte-rich RBC suspensions, representing 75% to 80% of the total HK activity.
RBC of increasing mean cell age can be separated by
buoyant density ultracentrifugation on a discontinuous gradient of arabinogalactane.” Six layers of cells were isolated,
and their respective mean age was estimated.” From each
layer the cells were lysed and centrifuged, and the resulting
supernatants were divided into two aliquots. One was directly
applied on the MonoQ column. The hemoglobin content in
the void volumes of each eluate was measured and the HK
activities were expressed as a specific activity against the
hemoglobin content. In the other aliquot the total activity of
HK was measured.
Figure 3B shows the HK activity profiles from a MonoQ
column of each layer in a typical experiment. In the lightest
cells ( L l , mean age = 22 days), HK, predominated,
constituting 82% of total activity. HK, then decreased
dramatically in the successive layers, in inverse proportion to
the age of the cells. In the heaviest cells (L6, mean age = 79
days), HK, was negligible. Thus it was clear that HK, was
*Normal RBC
0 Reticulocyte-rich RBC
A
relatively stable and that HK, rapidly decayed during aging
of the cells. The specific activities of each HK isozyme were
then plotted against the mean age of the cells in each layer.
As shown in the inset, HK, activity decreased exponentially
from 50 to 10 mU/pmole hemoglobin from L1 to L3, with an
apparent tIl2 of -10 days (95% confidence limits, 8 to 12
days). For this estimate, only the specific activities of HK, in
L1, L2, and L3 were used since the contribution of HK, into
the HK, peak became overwhelming in L4, L5, and L6. The
apparent
of HK, activity could be precisely estimated in
all layers using only the left limb of the HK, peak, which did
not include any contribution by HK,. The decay rate of HK,
was much slower, with an estimated tl/2 of -66 days (95%
confidence limits, 60 to 84 days).
The above results indicating a biphasic rate of decay were
in agreement with the results of Rogers et
but in conflict
with our previous report of a linear decline of the total HK
activity of the RBC.’ Therefore we re-examined several
separations recently performed in our laboratory to obtain a
topmost layer of only 2% to 3% of the RBC population. (This
was in contrast to our earlier study in which the topmost
layer usually contained more than 10% of the RBC population to provide sufficient yield to measure many enzymes
simultaneously.) When the HK activity of each layer was
plotted against its cumulative distribution function, it was
apparent that the rate of decay was clearly biphasic with two
half-lives of -15 (95% confidence limits, 8 to 18 days) and
-51 days (95% confidence limits, 46 to 57 days), respectively. In the same layers, the activity of pyruvate kinase
decayed very linearly, with a single slope and a t,/2 of -30
days (95% confidence limits, 28 to 32 days) (Fig 4).
DISCUSSION
Conflicting reports are present in the literature with
regard to the HK isozymes of human RBC. Stocchi et alZ2
reported the separation of three closely overlapping peaks
100-
LI
L2
E
’N
3
120,
60
I0
AGE (DAYS),/’
/
/
h
I
I
.L
’/
.2 -
,
/
0
Y
,’
I
.I
20
30
40 10
20
FRACTION NUMBER
30
4c
Fig
Human RBC HK
isozyme profiles from a MonoQ
column (FPLC). (A) Normal and
reticulocyte-rich RBC. (B) RBC
fractions of increasing ages: L1
(22 days). U (32 days). L3 (43
days), L 4 (65 days), L5 (66days),
L6 (79 days). The inset shows
the decay rates of HK, and HK,
activities.
774
MURAKAMI ET AL
other HK isozyme present in other tissues. Therefore, the
HK, isozyme appears to be unique to the RBC.
In normal RBC, the activities of HK, and HK, were
approximately equal. In reticulocyte-rich RBC suspensions,
HK, was predominant. When RBC were separated on
buoyant density gradients into age-dependent layers, HK,
was shown to be predominant in the youngest cell population,
and to be negligible in the oldest RBC. Thus, HK,, the
RBC-specific isozyme, appears to be present a t a high level in
the reticulocyte and to degrade rapidly. From the relative
specific activity in the layers, the t1/2of both isozymes could
be estimated. While HK, decreased very rapidly, with a tIl2
of -10 days, HK, was much more stable with a tl/2of -66
days. Therefore it appears that the rapid in vivo decay of HK,
observed by us and others, reflects the rapid decay of HK,.
Our findings for human RBC HK closely parallel the
findings by Magnani et all4 for rabbit RBC. In that species as
well the second RBC isozyme decreased during cell aging.
Therefore, it appears likely that the presence of an isozyme of
HK in RBC precursors, which rapidly decays on entry into
the circulation, is a general phenomenon. It has been recently
suggested by B e ~ t l e rthat
~ ~ buoyant density separation of
I 1 1 1 1 1 1 1
I
I
I
t
RBC only results in a redistribution of reticulocytes and
1
IO
50
90
99
therefore estimates of half-life with these techniques could be
PERCENTILE SCALE
artifactual if the change in enzyme activity takes place only
at the reticulocyte-mature RBC transition. However, this
Fig 4. The decay rate of HK and pyruvate kinase during aging
view is not shared by
of the RBC.
It has been shown by others that HK decays in a biphasic
manner during aging of RBC." On reanalysis of several
using a very narrow DEAE cellulose column. Gahr,I3 using
separations performed recently in our laboratory, in which
2%to 3% of the youngest RBC were isolated in the topmost
the same column with a finer gradient, could show only two
layer, a biphasic decay was similarly apparent. Our previous
peaks. In our laboratory, using the same DEAE columns as
study that had erroneously suggested a single slope of decline
Stocchi et al, we also obtained two peaks, as did Gahr. To
was based on separations that did not contain many fractions
clarify this issue, we used an FPLC system with a MonoQ
with predominantly young RBC.' The present results suggest
column that offered superior resolution. With this technique
that, during maturation and aging of the RBC, the initial
we could clearly separate the human RBC HK into two
rapid decay of HK reflects the rapid decay of HK,, and the
peaks. These two forms were also distinguished from each
later slow decrease reflects the slow decay of HK,.
other by molecular size on a gel filtration column and by
Both isozymes of human RBC HK'0913*22
appear to behave
affinity for the mitochondrial fraction. Thus, it can be
kinetically as HK,, as do also the two isozymes of rabbit
concluded that human RBC contains two forms of HK
RBC. Magnani et aI2' indicated that all HK isozymes of
isozymes.
human RBC were cross-reactive with a polyclonal antibody
The HK of rabbit RBC had been separated into two
isozymes by DEAE-cellulose column ~hromatography."*'~ against HK,. In studies of rabbit RBC, they showed that only
HK, bound to the mitochondria and speculated that this
When the HK isozyme profile from a MonoQ column of
binding stabilized this isozyme, while the unbound one
human and rabbit RBC were compared, the two isozymes of
rapidly d e c a ~ e d . They
' ~ also suggested that the subtypes of
human RBC eluted at 120 mmol/L and 160 mmol/L KC1,
RBC HK resulted from posttranslational modifications of
whereas those of rabbit RBC eluted at 120 mmol/L and 180
HK,. Our study has shown that (1) HK, is an isozyme
mmol/L KCI, respectively. Thus, the two rabbit RBC
specific to the RBC, and is not found in any other tissues and
isozymes were even further apart from each other than those
cells; (2) HK,, but not HK,, binds to the mitochondrial
of human RBC. Previous studies that used DEAE-cellulose
fraction; (3) HK, is clearly larger by several kilodaltons than
chromatography could not resolve the isozymes as efficiently
HK,; and (4) the two forms independently decay at different
as the MonoQ column. While the two forms of rabbit RBC
rates.
HK could be easily separated on DEAE-cellulose, this
Thus, our data suggest that HK, and HK, are separate
technique could not adequately resolve the much closer two
gene products that are independently regulated, and that the
peaks of human RBC HK, and hence yielded conflicting
HK, gene is expressed only in the RBC. Although HK, is
results.
smaller than HK,, it does not appear to be a degradative
Our study demonstrates that HK,, the second form of
intermediate of HK,. In fact, as HK, decreases, HK, does
human RBC HK, is eluted in a position different from any
775
HEXOKINASE ISOZYME SPECIFIC FOR HUMAN RBC
not increase and the two forms decay at the independent
rates. Altay et a12* showed in their electrophoretic studies
that the fast-moving band (presumably our HK,) was
increased in hemolytic anemia and deficient in a patient with
HK deficiency. These results are consistent with our conclusion that HK, and HK, are products of two separate genes.
Studies of the molecular mechanisms for the regulation of
the HK, gene and of the molecular structure of both HK,
and HK, may clarify the process of aging of the RBC.
ACKNOWLEDGMENT
We thank Drs E. Beutler and L. Corash for their thoughtful
review of the manuscript and helpful suggestions.
REFERENCES
1. Seaman C, Wyss S, Piomelli S: The decline in energetic
metabolism with aging of the erythrocyte and its relationship to cell
death. Am J Hematol8:31, 1980
2. Bosman GJCGM, Kay MMB: Erythrocyte aging: A comparison of model systems for simulating cellular aging in vitro. Blood
Cells 14:19, 1988
3. Chapman RG, Hennessey MA, Watersdorph AM, Huennekens FM, Gabrio BW: Erythrocyte metabolism: Level of glycolytic
enzymes and regulation of glycolysis. J Clin Invest 41:1249, 1962
4. Brok F, Ramot B, Zwang E, Danon D: Enzyme activities in
human blood cells of different age groups. Isr Med J 2:291, 1966
5. Piomelli S, Wyss SR Metabolic death of the red blood cell.
Blood 38:833a, 1971 (abstr)
6. Rapoport TA, Heinrich R, Jacobasch G, Rapoport S: A linear
steady-state treatment of enzymatic chains: A mathematical model
of glycolysis of human erythrocyte. Eur J Biochem 42:107, 1974
7. Grossbard L, Schimke R T Multiple hexokinases of r it tissues.
Purification and comparisons of soluble forms. J Biol Chem 244:
3546,1966
8. Holmes EW Jr, Malone JI, Winegrad AI, Oski FA: Hexokinase isoenzymes in human erythrocytes: Association of type I1 with
fetal hemoglobin. Science 156:646, 1967
9. Kaplan JC, Beutler E: Hexokinase isoenzymes in human
erythrocytes. Science 159:215, 1968
10. Stocchi V, Stulzini A, Magnani M: Chromatographic fractionation of multiple forms of red blood cell hexokinases. J Chromatogr
237:330, 1982
1 1 . Magnani M, Serafini G, Stocchi V: Hexokinase type I
multiplicity in human erythrocytes. Biochem J 254:617, 1988
12. Rijksen G, Vansen G, Kraaijenhagen RJ, van der Vlist MJM,
Vlug AMC, Staal GEJ: Separation and characterization of hexokinase I subtypes from human erythrocytes. Biochim Biophys Acta
659:292, 1981
13. Gahr M: Different biochemical properties of foetal and adult
red cell hexokinase isoenzymes. Hoppe-Seyler’s Z Physiol Chem
361:829, 1980
14. Magnani M, Stocchi V, Dacha M, Fornaini G: Rabbit red
blood cell hexokinase. Mol Cell Biochem 61:83, 1984
15. Piomelli S, Corash LM, Davenport DD, Miraglia J, Amorosi
EL: I n vivo lability of glucose-6-phosphate dehydrogenase in GdAand GdMediterraneandeficiency. J Clin Invest 47:940,1968
16. Lowry OH, Rosenbrough NJ, Farr AL, Randall RJ: Protein
measurement with the folin phenol reagent. J Biol Chem 193:265,
1951
17. van Kampen EJ, Zijlstra WG: Standardization of hemoglobinometry 11. The hemoglobin-cyanide method. Clin Chim Acta
6538,1961
18. Corash LM, Piomelli S, Chen MC, Seaman C, Gross E
Separation of erythrocytes according to age on a simplified density
gradient. J Lab Clin Med 84:147, 1974
19. Rose IA, Warms JVB: Mitochondrial hexokinase. Release,
rebinding, and location. J Biol Chem 242:1635, 1967
20. Stocchi V, Magnani M, Piccoli G, Fornaini G: Hexokinase
microheterogeneity in rabbit red blood cells and its behaviour during
reticulocytes maturation. Mol Cell Biochem 79:133, 1988
21. Rogers PA, Fisher RA, Harris H: An examination of the
age-related pattern of decay of the hexokinases of human red cells.
Clin Chim Acta 69291, 1975
22. Stocchi V, Magnani M, Canestrari F, Dacha M, Fornaini G:
Multiple forms of human red blood cell hexokinase. Purification,
characterization and age dependence. J Biol Chem 257:2357, 1982
23. Beutler E: The relationship of red cell enzymes to red cell
life-span. Blood Cells 14:69, 1988
24. Danon D, Marikovsky Y: The aging of the red blood cell. A
multifactor approach. Blood Cells 14:7, 1988
25. Lutz HU: Red cell density and red cell age. Commentary to
Beutler E: The relationship of red cell enzymes to red cell life-span.
Blood Cells 1476,1988
26. Piomelli S: Commentary to Beutler E: The relationship of red
cell enzymes to red cell life-span. Blood Cells 14:81, 1988
27. Magnani M, Chiarantini L, Serafini G, Stocchi V, Dacha M,
Fornaini G: Human erythrocyte hexokinases are immunologically
related. Arch Biochem Biophys 45540, 1986
28. Altay C, Alper CA, Nathan DG: Normal and variant
isoenzymes of human blood cell hexokinase and the isoenzyme
patterns in hemolytic anemia. Blood 36:219, 1970