1821
Journal of Cell Science 110, 1821-1829 (1997)
Printed in Great Britain © The Company of Biologists Limited 1997
JCS3600
Glyceraldehyde 3-phosphate dehydrogenase is bound to the fibrous sheath
of mammalian spermatozoa
Doris Westhoff and Günter Kamp*
Institut für Zoophysiologie, Westfälische Wilhelms-Universität Münster, Hindenburgplatz 55 D-48143 Münster, Germany
*Author for correspondence (e-mail: [email protected])
SUMMARY
Evidence is provided that the glycolytic enzyme glyceraldehyde 3-phosphate dehydrogenase is covalently linked
to the fibrous sheath. The fibrous sheath is a typical
structure of mammalian spermatozoa surrounding the
axoneme in the principal piece of the flagellum. More than
90% of boar sperm glyceraldehyde 3-phosphate dehydrogenase activity is sedimented after cell disintegration by
centrifugation. Detergents, different salt concentrations or
short term incubation with chymotrypsin do not solubilize
the enzyme, whereas digestion with trypsin or elastase does.
Short term incubation with trypsin (15 minutes) even
resulted in an activation of glyceraldehyde 3-phosphate
dehydrogenase. Purification on phenyl-Sepharose yielded a
homogeneous glyceraldehyde 3-phosphate dehydrogenase
as judged from gel electrophoresis SDS-PAGE and native
gradient PAGE. The molecular masses are 41.5 and 238
kDa, respectively, suggesting native glyceraldehyde 3phosphate dehydrogenase to be a hexamer.
Rabbit polyclonal antibodies raised to purified glyceraldehyde 3-phosphate dehydrogenase show a high specificity for mammalian spermatozoal glyceraldehyde 3phosphate dehydrogenase, while other proteins of boar
spermatozoa or the muscle glyceraldehyde 3-phosphate
dehydrogenase are not labelled. Immunogold staining
performed in a post-embedding procedure reveals the
localization of glyceraldehyde 3-phosphate dehydrogenase
along the fibrous sheath in spermatozoa of boar, bull, rat,
stallion and man. Other structures such as the cell
membrane, dense fibres, the axoneme or the mitochondria
are free of label.
During the process of sperm maturation, most of the
cytoplasm of the sperm midpiece is removed as droplets
during the passage through the epididymis. The labelling
of this cytoplasm, in immature boar spermatozoa and in
the droplets, indicates that glyceraldehyde 3-phosphate
dehydrogenase is completely removed from the midpiece
during sperm maturation in the epididymis.
The inverse compartmentation of the glycolytic enzyme
and mitochondria in the mammalian sperm flagella
suggests that ATP-production in the principal piece mainly
occurs by glycolysis and in the midpiece by respiration.
INTRODUCTION
also been identified in human spermatozoa (Wallimann et al.,
1986; Huzar and Vigue, 1990), but other mammalian spermatozoa (boar and bull) did not contain these enzymes (Kamp et
al., 1996) and consequently lack a PCr/CK-shuttle.
In contrast to sea urchin spermatozoa which preferentially
oxidize fat, mammalian spermatozoa use carbohydrates (for
review see Mann and Lutwak-Mann, 1981). This allows ATP
production in the cytoplasm independent of mitochondria. It
has also been reported that boar spermatozoa produce lactate
even under aerobic conditions which indicates the essential
role of glycolytic ATP-production (see Mann and LutwakMann, 1981). We have examined whether aerobic lactate production (aerobic glycolysis) is a consequence of an insufficient
PCr/CK-shuttle in mammalian spermatozoa. A high glycolytic
capacity in the principal piece of the flagellum could make up
for the lack of transfer of energy rich phosphate from the mitochondria to the distal dynein-ATPases. This paper presents
results on the localization of a central glycolytic enzyme, the
glyceraldehyde 3-phosphate dehydrogenase (GAPDH, EC
Fertility requires sperm motility and consequently ATP production. Oxidative phosphorylation in mitochondria is the most
efficient way to produce ATP but in the case of spermatozoa,
the compartmentation of mitochondria in the midpiece of the
flagellum causes the problem that ATP availability for the
dynein-ATPases of the mitochondrion-free part of the
flagellum (principal piece) is limited. Evidence for an energy
transport system through the principal piece has been presented
for sea urchin and rooster sperm (Tombes and Shapiro, 1985;
Kaldis et al., 1996). In these sperm, the energy-rich phosphate
of mitochondrial ATP is transferred to creatine by a specific
mitochondrial creatine kinase (CK) and the product phosphorylcreatine (PCr) diffuses to the dynein-ATPases, where local
ATP pools are regenerated from PCr by a cytoplasmic CK.
Creatine (Cr) returns to the mitochondria while ADP remains
as local pools both at the site of ATP consumption and the site
of ATP production. Mitochondrial and cytoplasmic CK have
Key words: Creatine kinase, Glycolysis, Glyceraldehyde 3-phosphate
dehydrogenase, Triosephosphate isomerase, Fibrous sheath,
Metabolic channelling, Spermatozoa
1822 D. Westhoff and G. Kamp
1.2.1.12) in the flagellum of mammalian spermatozoa as
defined by immunogold labelling using specific antibodies
against sperm GAPDH.
MATERIALS AND METHODS
Samples and chemicals
Fresh ejaculates of fertile boar, bull, stallion, turkey, carp and trout
were provided by local breeders associations. Samples of rat and
human spermatozoa were given by the Institute of Reproductive
Medicine (University of Münster, D-48149-Münster, FRG). Sea
urchin spermatozoa were collected at Friday Harbor (University of
Washington, WA 98250, USA) and lugworm sperm (Arenicola
marina, Polychaeta) at the German North Sea coast near Carolinensiel.
All chemicals were of highest available purity. Biochemicals and
enzymes were purchased from Boehringer (D-68298 Mannheim),
immunochemicals from Sigma (D-82041 Deisenhofen), phenylSepharose CL-4B from Pharmacia (D-79111 Freiburg), and
immobilon transfer membrane Type PVDF (polyvinylidene difluoride) from Millipore (D-65731 Eschborn). Others were obtained
from Merck (D-64293 Darmstadt) and Fluka (D-89231 Neu-Ulm).
Microconcentrators (Centricon-100) were purchased from Amicon
(D-58453 Witten).
Enzyme activities
Maximum enzyme activities of various glycolytic enzymes including
glyceraldehyde 3-phosphate dehydrogenase (GAPDH, EC 1.2.1.12),
triose phosphate isomerase (TIM, EC 5.3.1.1) and phosphoglycerate
kinase (PGK, EC 2.7.3.2) were measured using coupled enzymatic
assays with NADH2 as indicator (Bergmeyer, 1983). Determinations
were performed at 25°C.
Extraction procedure
Sperm were separated from seminal plasma by centrifugation (3,000
g; 15 minutes) and resuspended in Tris-HCl buffer (10 mMol/l; pH
7.3) containing 5 mMol/l EDTA, 1 mMol/l dithiothreitol, 2 mMol/l
NaF, 0.01% silicon anti-foam, 0.01% macroglobulin. After homogenization with an Ultra-Turrax (3× 20 seconds, highest speed), the
samples were centrifuged at 4°C for 15 minutes at 20,000 g. The
sediment was resuspended in the same buffer as before and incubated
for 15 minutes at 20°C with trypsin (3 mg/g sperm wet weight).
Trypsin digestion was stopped with 20 µl 0.1 Mol/l phenylmethylsulfonyl fluoride (PMSF) per g sperm wet weight and the suspension
was centrifugated at 4°C for 15 minutes at 20,000 g. Commercial
trypsin (Sigma) was purified by affinity chromatography (Eberhardt,
1992).
Purification
Trypsin treatment of disintegrated sperm proved to be a very efficient
purification step. Solubilized GAPDH was bound to the hydrophobic
exchanger phenyl-Sepharose CL-4B (4°C) equilibrated with sodium
phosphate buffer (50 mMol/l; pH 7.0) containing 30% (NH4)2SO4.
Before application the extract was adjusted to 30% (NH4)2SO4. The
column was extensively rinsed with the same buffer until protein
elution reached a minimum. Reduction of the (NH4)2SO4 concentration to 25% resulted in GAPDH elution. Collected fractions of
GAPDH activity were concentrated with an Amicon microconcentrators (100 kDa) to 0.5 ml final volume and washed (3× 1.5 ml sodium
phosphate buffer 50 mMol/l; pH 7.0). Protein was determined
according to the method of Bradford (1976), as modified by Read and
Northcote (1981).
Electrophoresis and immunoblotting
Relative molecular mass was determined by electrophoreses which
were performed in a Hoefer Mighty-small apparatus. Blue native
polyacrylamide gel electrophoresis (BN-PAGE) at pH 7.0 was carried
out as described by Schägger et al. (1994) using a 4% stacking gel
and a running gel with a 5 to 25% gradient. Samples diluted with
twofold volumes of a buffer consisting of 15% glycerol, 50 mMol/l
BisTris, pH 7.0, were applied to the gel. Electrophoresis through the
stacking gel was performed at 100 V and through the running gel at
550 V (time 4.5 hours).
SDS-PAGE was carried out according to the method of Laemmli
(1970). Samples were incubated in closed tubes for 1 hour at 60°C in
buffer (62.5 mMol/l Tris-HCl, 0.1 Mol/l dithiothreitol, 0.02% Bromophenol Blue, 4.0% SDS, 40% glycerol, 10.2% 2-mercaptoethanol)
and applied to gels consisting of 3% stacking gel and 7% running gel.
The run was performed at 10°C and 14 mA for 1 hour.
For immunoblotting, proteins were transferred from SDS-gel to a
PVDF-membrane in a Biometra Fast-Blot apparatus (150 mA, 30
minutes, semi-dry method). Blots were washed 4 times with
phosphate buffer solution (PBS: NaCl 139 mMol/l, KH2PO4 3.6
mMol/l, Na2HPO4 12 mMol/l; pH 7.3) to remove SDS. Non-specific
protein binding sites on the PVDF-membrane were blocked with
bovine serum albumin (BSA: 3% in PBS buffer). After washing 4
times with BSA-free PBS, the membrane was incubated (1 hour,
20°C) with anti-boar sperm GAPDH serum diluted 1:3,000 with PBS
containing 1% BSA. Subsequently the membrane was rinsed in PBS
and incubated (1 hour, 20°C) with horseradish peroxidase-conjugated
goat anti-rabbit IgG diluted 1:40,000 (Sigma preparation A0545) in
PBS buffer with 1% BSA. The PVDF-membrane was again rinsed
with PBS buffer and then stained by 3,3-diaminobenzidine. A second
PVDF-membrane was stained for protein with Coomassie blue immediately after blotting.
Rabbit antiserum
Antibodies against purified boar sperm GAPDH were produced as
described by Harlow and Lane (1988). A rabbit was inoculated with
1 mg of the purified GAPDH emulsified in complete Freund’s
adjuvant. Incomplete Freund’s adjuvant with 0.5 mg antigen was used
for booster injection after 61 and 103 days. Serum with GAPDH-antibodies was obtained from coagulated blood by two step centrifugation at 3,000 g and 20,000 g (15 minutes each).
Electron microscopy and immunogold labelling
Immunoelectron microscopy was performed with a post-embedding
procedure. Semen samples were prefixed by paraformaldehyde (4%
in PBS) for 30 minutes at 20°C. Samples rinsed in PBS were dehydrated with ethanol (50% and 2× 70% ethanol in water) and embedded
in LR White, for 48 hours at 60°C.
Thin sections (80-90 nm) were attached to nickel grids. To prevent
unspecific binding of anti-GAPDH, grids were first incubated in PBS
with 1% BSA at 20°C for 30 minutes then the antibodies were added
in dilutions of 4,000 to 6,000 times and incubated for 1 hour at 20°C.
After washing 5 times with 1% BSA in PBS, gold (10 nm)-conjugated
goat anti-rabbit IgG in dilutions of 1:40 and 1:50 (Sigma G7402) was
added for 1 hour at 20°C and removed by washing with 1% BSA in
PBS and finally with Millipore-Q water (5 times each). Thin sections
were contrasted with filtered uranyl acetate (2% in water), and washed
5 times with Millipore-Q deionized water. A Philips electron microscope TM 10 (80 kV) was used for producing the micrographs.
RESULTS
Solubilization and purification
After sperm disintegration in various salt buffer solutions
(Na/K-phosphate, Tris-HCl or imidazole; 100 mMol/l; pH 7.3)
more than 90% of total boar sperm GAPDH activity was found
in the sediment after centrifugation of the cell homogenate at
Structural binding of GAPDH in spermatozoa 1823
3,000 g. The same behavior was observed with spermatozoa of
other mammals (bull, stallion, rat and man) but not with those
from birds (turkey), fishes (carp and trout), echinoderms (sea
urchin) or annelids (lugworm).
Changing the pH (between 6.5 and 8.0) or the ionic strength
(between 0.01 and 0.6 Mol/l) did not affect the binding of
GAPDH to boar sperm fragments. Incubation with ATP (1.5
mMol/l) or NADH2 (1 mMol/l) which would solubilize
GAPDH from erythrocyte membranes (Shin and Carraway,
1973) had no effect. Detergents such as Triton X-100 (1%) or
Chaps (0.5%) were also ineffective in GAPDH solubilization.
However, incubation (15 minutes) with trypsin or elastase (3
mg/g cell wet weight each) did solubilize the enzyme and even
resulted in an activation of GAPDH (4.7±0.3 fold for trypsin,
n=4; 3.3±0.3-fold for elastase, s.e.m. of n=3). After incubation,
trypsin digestion was stopped by adding PMSF. The degree of
solubilization and activation of GAPDH were dependent on
temperature, the ratio of trypsin to cell protein concentration
and the incubation time. Proteolytic inactivation was observed
with a pure trypsin preparation only after more than 15 minutes
incubation at 20°C. Traces of chymotrypsin (1% of trypsin
activity), however, caused rapid GAPDH inactivation (synergistic digestion). Pure chymotrypsin (trypsin activity <0.01%)
in the same concentration as used for trypsin did not significantly affect GAPDH solubilization or activity within 15
minutes incubation time.
The proteolytic solubilization of GAPDH was quite specific
and therefore an efficient step in the purification procedure (cf.
Fig. 4, lane 5 extract before and lane 4 after trypsin treatment).
A minor part of the total activities of the triosephosphate
isomerase (TIM, <20%), phosphoglycerate kinase (PGK,
<30%) and a major part of pyruvate kinase (PK, >50%) were
also solubilized by trypsin.
For further purification, solubilized boar sperm GAPDH was
applied to a phenyl-Sepharose column. Most contaminating
protein (mainly trypsin) did not bind to phenyl-Sepharose if
the extract and the column were adjusted to 30% (NH4)2SO4
(Fig. 1) whereas GAPDH, TIM and PGK were completely
bound. Reducing the (NH4)2SO4 concentration to 25% resulted
in the elution of GAPDH together with TIM, whereas PGK
was retained to 20% (NH4)2SO4 (Fig. 1). The activity ratio of
GAPDH and TIM in each fraction was fairly constant
(GAPDH/TIM = 0.6±0.1; n=19) indicating identical elution
behavior. The enzyme preparation contained 105 U GAPDH
and 147 U TIM per mg protein. This could either mean that
both enzymes bind separately but with similar affinities to
phenyl-Sepharose or that both enzymes form a complex that
binds to the phenyl-Sepharose. The following experiments
suggest the latter be the case: experiments with different ion
exchangers (DEAE-Sepharose, Q-Sepharose) and color ligand
chromatography were unsuccessful in separating GAPDH and
TIM activities. Ultrafiltration on a 100 kDa cut off membrane
resulted in a concentration of both activities though one would
have expected that TIM would pass through, assuming the
same molecular mass as for TIM of skeletal muscle (e.g. 53
kDa rabbit muscle TIM; Norton et al., 1970). The molecular
mass of muscle GAPDH is about 145 kDa (e.g. pig skeletal
muscle; Harris and Perham, 1968).
The SDS-PAGE of the sperm GAPDH preparation showed
only one band corresponding to a molecular mass of 41.5 kDa
(Fig. 2) which is slightly higher than that of boar muscle
GAPDH (39 kDa, Harris and Waters, 1976). Despite the high
TIM activity, no protein band corresponding to TIM was
visible, which is probably due to the higher specific activity of
TIM as compared to GAPDH (e.g. 7,800 U/mg for rabbit
muscle TIM; Norton et al., 1970; cf Discussion).
The molecular mass of the native GAPDH could not be
determined by normal gradient PAGE, as the migration was
less than 10 mm in 24 hours at 160 V. When coated with
Coomassie Blue (BN-PAGE), electrophoretic analysis was
possible and revealed one single protein band (Fig. 3). The
molecular mass of this was estimated to be 238 kDa, which is
5.7 times higher than that of the subunit, indicating a hexameric
structure for the native GAPDH.
3.0
5
30 % (NH4)2SO4
25 % (NH4)2SO4
20 % (NH4)2SO4
2.5
3
1.5
2
1.0
1
0.5
0
0.0
0
20
40
60
80
100
fractions
120
140
160
180
200
enzyme activity [ U ]
2.0
extinction (280 nm)
Fig. 1. Elution profile of
boar sperm GAPDH, TIM
and PGK activities from
phenyl-Sepharose.
Solubilized by trypsin,
GAPDH (s), TIM (u) and
PGK (n) bind to phenylSepharose at 30%
(NH4)2SO4 while most
other proteins (d) do not.
At 25% (NH4)2SO4 both
GAPDH and TIM
activities are eluted. PGK
elutes at 20%. The
activities are expressed in
units of 1 µmol
substrate/minute at 25°C
per fraction (5.2 ml).
4
1824 D. Westhoff and G. Kamp
Fig. 2. SDS-PAGE of purified
boar sperm GAPDH. Lane 1
shows standard proteins: glycogen
phosphorylase (94 kDa), albumin
(67 kDa), ovalbumin (43 kDa),
and carbonicanhydrase (30 kDa),
lanes 2 and 3 the purified GAPDH
preparation with 1 and 2 µg
protein, respectively.
Specificity of rabbit serum antibodies
Fig. 4 shows western blots of boar sperm extracts and various
glycolytic enzymes with the rabbit serum. Sperm GAPDH was
labelled (lanes 1 and 11, corresponding to protein staining lanes
6 and 12) while boar muscle TIM (lane 13), PGK (lane 14) and
GAPDH (lanes 15 and 16) were not (lanes 7-10). The extract of
sperm homogenate was not treated with trypsin and consequently did not contain GAPDH; it revealed no reaction with the
antibodies (lane 2) but many protein bands (lane 5). After
trypsin-treatment the extract contained GAPDH. Before application of western blots trypsin was removed from the extract by
four washing steps with 5-fold volumes of extraction buffer on
a microconcentrator (cut off membrane 30 kDa). Lane 4 shows
a few protein bands, one of which corresponds to the GAPDH
being stained (lane 3). Hence, the immunological response
indicates high specificity of the antibodies in the serum against
sperm GAPDH, which is a prerequisite for the identification and
localization of the enzyme in the cell by immunogold labelling.
Immunogold labelling
EM-sections of boar sperm flagella were embedded and
incubated with anti-GAPDH-TIM. The fibrous sheath was
specifically labelled by the antiserum (Fig. 5A and D), but no
gold particles were found at the dense fibres or the axoneme.
The midpiece of the flagellum was completely free of any label
(Fig. 5B and E). After extraction of GAPDH by trypsin, sperm
fragments appeared fuzzy with gaps in the fibrous sheath and
immunogold labelling was not possible (Fig. 5C and F). This
confirms the specificity of the antiserum against the GAPDH.
Like boar, spermatozoa from bull, rat, stallion, and man also
Fig. 4. Western blots of boar sperm extracts and purified
glycolytic enzymes. Proteins were transferred from the
SDS-gel to the PVDF-membrane and either stained with
boar sperm anti-GAPDH (a and c) or for protein (b and
d). Lanes 1 and 11 show the immune response of purified
boar sperm GAPDH (1 µg protein, corresponding protein
staining lanes 6 and 12). Trypsin untreated extract shows
various protein bands (lane 5) but no labelling with antiGAPDH (lane 2). Trypsin-treated extract (lane 4) reveals
a prominent protein band and an immune response at the
same position. This protein corresponds to purified
GAPDH (lane 1). GAPDH of boar muscle (2 µg protein
each, lanes 15 and 16) as well as TIM (2 µg protein, lane
13) and PGK (3.6 µg protein, lane 14) do not react with
boar sperm anti-GAPDH (lane 7-10).
Fig. 3. BN-PAGE of purified
boar sperm and muscle
GAPDH. Lane 1 shows
standard proteins:
thyroglobulin (669 kDa),
ferritin (440 kDa), catalase
(232 kDa), lactate
dehydrogenase (140 kDa) and
albumin (67 kDa), lane 2 the
boar muscle GAPDH (0.5 µg
protein, commercial product
Boehringer Mannheim), lane
3 the boar sperm GAPDH
preparation containing PGK (1.5 µg protein), lane 4 the boar sperm
GAPDH preparation (1 µg protein) without PGK. The arrow
indicates a second minor protein band in lane 3 which may be
assigned to a GAPDH-TIM-PGK complex.
revealed a very specific labelling of the fibrous sheath with
anti-GAPDH (Fig. 6). No response was detected in electron
micrographs of turkey or carp spermatozoa which do not form
a fibrous sheath (not shown).
Immature spermatozoa were also found in ejaculates as
shown in the case of boar sperm in Fig. 7. The spermatozoa
are characterized by a cytoplasmic droplet at the midpiece
which is normally shed before the spermatozoa leave the epididymis. The cytoplasm of the midpiece (Fig. 7A), which will
be removed or already was removed by the droplets (Fig. 7BC), contained GAPDH. Hence, the lack of GAPDH in the
midpiece of mature spermatozoa is a consequence of sperm
maturation in the epididymis.
DISCUSSION
We have shown that GAPDH in mammalian spermatozoa is
bound to the fibrous sheath. This structure surrounds the dense
fibers and the axoneme in the principal piece of the flagellum
like the mitochondrial sheath in the midpiece. The function of
this structure remains unclear; probably it gives a certain elasticity to the flagellum. However, the localization of a glycolytic
enzyme along this structure led us to propose a metabolic
function for this structure to supply distal dynein ATPases with
energy (see below). GAPDH is the first catalytically active
enzyme proved to be bound to the fibrous sheath. This result may
also explain the very low GAPDH activities found in buffalo
Structural binding of GAPDH in spermatozoa 1825
Fig. 5. Immunogold electron
microscopy of boar sperm
flagellum. Longitudinal (A)
and cross (D) sections of the
principal piece show the
labelled fibrous sheath (fs)
while the dense fibres (df)
and the axoneme (ax) are
unlabelled. No label is
visible in the sections of the
middle piece (B and E)
where the fibrous sheath is
replaced by the
mitochondrial sheath (mi).
Trypsin-treated spermatozoa
are highly damaged (C and
F) and the fibrous sheath is
perforated (arrows) and
unlabelled after incubation
with anti GAPDH. Bars: 38
nm (A,C,D); 72 nm (B); 53
nm (E,F).
spermatozoa by Gandhi and Anand (1982). These authors
measured the soluble part of GAPDH activity and probably
removed the major insoluble part by centrifugation. The problem
of the solubilization of glycolytic enzymes in mammalian spermatozoa has also been described by Harrison (1971) but the
localization of the enzymes has not been demonstrated before.
Microcompartmentation of glycolytic enzymes has been
described for other cell types, providing evidence for a structural organization of glycolytic enzymes in various tissues or
cells (for review see Masters et al., 1987; Srere, 1987). In these
cases, however, the enzymes are associated with the cytoskeleton, the contractile fibres (muscle) or the plasma membrane
(erythrocytes) through weak ionic or hydrophobic interactions
which are readily broken during extraction in hypotonic or
hydrophobic buffer solutions. Because enzyme association was
reversible and some enzymes depending on enzyme phosphorylation, the enzyme-structure association has been interpreted
to be a regulatory mechanism (Minaschek et al., 1992). In
contrast, the GAPDH in boar spermatozoa appears irreversibly
linked to cell structures under physiological conditions and the
proteolytic solubilization by trypsin and elastase suggests a
covalent linkage which does not constitute a regulatory
mechanism of GAPDH activity.
teolysis appears to be fairly specific because the solubilized
GAPDH proved to be a rather homogenous protein preparation in both SDS- and native-PAGE. This suggests that trypsin
hydrolyzes the peptide chain by which GAPDH is linked to
cellular structures. The substrate specificity of trypsin
requires one of the two positively charged amino acids,
arginine or lysine (Brandon and Tooze, 1991) to be located
at an exposed position along the proposed linker (cf. Fig. 8).
Amino acids with small uncharged side chains at that position
may explain the cleavage by elastase. The fact that chymotrypsin did not solubilize GAPDH indicates that the
proposed linker does not contain aromatic amino acids at an
exposed position. A specific peptide linker for GAPDH at the
fibrous sheath may be the best explanation for the GAPDH
compartmentation along that fibrous sheath and for the proteolytic solubilization. Alternatively, it cannot be completely
excluded that GAPDH is only trapped in pockets of the
fibrous sheath which could be hydrolyzed by proteases and
thus release GAPDH. Nevertheless, the binding of sperm
GAPDH must be different from that of certain glycolytic
enzymes to actin or tubulin in somatic cells or the band 3
protein of erythrocyte membranes (for review see Knull and
Walsh, 1992).
GAPDH binding site
Solubilization of GAPDH by limited trypsin treatment
activated rather than inactivated the enzyme. The tryptic pro-
Comparison of GAPDH from muscle and sperm
Some properties of sperm GAPDH indicate a specific enzyme
modification: (i) the molecular mass of subunits (41.5 kDa) is
1826 D. Westhoff and G. Kamp
Fig. 6. Immunogold
electron microscopy of rat,
human, stallion and bull
sperm flagella. The rat
fibrous sheath is selectively
labelled as shown in the
cross sections through the
proximal (A) and the distal
part (C) of the principal
piece as well as in the
superficial section through
the fibrous sheath (B). The
longitudinal section of the
midpiece of a rat
spermatozoon (A) is
unlabelled. A cross section
of a human sperm flagellum
(D) shows the proximal part
of the principal piece with
labelled fibrous sheath and
fairly unlabelled dense
fibres. Stallion sperm cross
sections (E) through the
midpiece (left) and the
principal piece (right)
reveal a very specific
labelling of the fibrous
sheath which is also true for
bull spermatozoa (F).
Abbreviations: (df) dense
fibers, (fs) fibrous sheath,
(mi) mitochondria. Bars:
120 nm (A); 50 nm (B); 74
nm (C); 53 nm (D); 46 nm
(E); 35 nm (F).
different from that of boar muscle GAPDH (38 kDa, Fig. 2; 35
kDa, Bode et al., 1975) and also of the native enzyme (boar
sperm GAPDH 238 kDa, Fig. 3; boar muscle GAPDH 120-144
kDa, Allison and Kaplan, 1964; Harris and Perham, 1968). The
ratio of molecular mass suggests that boar sperm GAPDH
exists as a hexamer while boar muscle GAPDH is a tetramer.
To our knowledge, no hexameric GAPDH has been found in
other tissues. (ii) The electrophoretic mobility of sperm
GAPDH differs from that of the muscle enzyme. In contrast to
skeletal muscle GAPDH, sperm GAPDH does not migrate in
the native gradient PAGE. In BN-PAGE, Coomassie dye
surrounds the protein and thus provides for electrophoretic
mobility of sperm GAPDH (Fig. 3). The hydrophobic nature
of sperm GAPDH explains also the affinity to phenylSepharose. (iii) The immunological response of the antibodies,
which were produced in rabbits, is specific for the sperm
GAPDH. The enzyme of boar muscle was not labelled.
The specific activities of boar sperm and muscle GAPDH
are approximately the same (105 U/mg protein at 25°C for
sperm GAPDH; 92-170 U/mg protein at 25°C for rabbit muscle
GAPDH; Ferdinand, 1964; Bode et al., 1975).
Multienzyme complex of sperm GAPDH with TIM
and PGK
Due to the hydrophobic nature of sperm GAPDH this enzyme
binds to phenyl-Sepharose. While PGK was finally separated
from GAPDH by a stepwise elution from the phenyl-
Sepharose, the elution profile of TIM and GAPDH activities
were identical. This result can be explained either by very
similar hydrophobic properties of both enzymes or by the fact
that the enzymes are aggregated and eluted as a complex. The
latter is supported by gel filtration and ultrafiltration which did
not separate the two enzyme activities. Native GAPDH in
sperm probably has a considerably higher molecular mass than
TIM as judged from the situation in other cell types. TIM has
a relative molecular mass of 53 kDa in muscle for instance
(Norton et al., 1970).
On the other hand, a heteromeric enzyme complex of
GAPDH and TIM should reveal two protein bands in the
SDS-PAGE, but these were not detected (Fig. 2). This discrepancy can be explained by taking into account that TIM
of somatic and probably also of sperm cells have very high
specific activities (7,800 U/mg protein, Norton et al., 1970).
Given such a high specific activity, the TIM activity measured
in the purified GAPDH preparation would account for only
2% of total protein and thus may escape detection in the SDSgel.
Like TIM, native sperm PGK is likely to have a considerably lower relative molecular mass than GAPDH (PGK of
rabbit muscle 47 kDa; Krietsch and Bücher, 1970) but it was
not separated from GAPDH by ultrafiltration. A separation of
GAPDH and PGK, however, was achieved on phenylSepharose when the enzymes were eluted by a stepwise
gradient. A linear gradient resulted in a preparation consisting
Structural binding of GAPDH in spermatozoa 1827
of the mammalian flagellum which adjoins the middle piece
containing the mitochondrial sheath (Fawcett, 1970; Fouquet
and Kann, 1994). Both the mitochondrial sheath in the middle
piece and the fibrous sheath in the principal piece surround the
dense fibers which run parallel to the axoneme (Figs 5 and 6)
and may be stiff elastic structures protecting sperm against
damage (Baltz et al., 1990). The segmented appearance formed
by two longitudinal columns connected by semi-circular ribs
confers upon the fibrous sheath the flexibility which is essential
for tail movement. The physiological function of the fibrous
sheath, however, is still unclear though various reports have
revealed its protein composition by SDS-electrophoresis (Kim
et al., 1995). Our results show that GAPDH is the first enzyme
identified at the fibrous sheath and lead us to propose that this
structure is glycolytically active.
Fig. 7. Immunogold electron microscopy of immature boar
spermatozoa. (A) The midpiece of an immature boar spermatozoon
with many gold particles in the peripheral area of the cytoplasm.
Correspondingly the droplets (B and C) which remove the cytoplasm
from the midpiece were also labelled. Abbreviations: (ax) axoneme,
(df) dense fibers, (fs) fibrous sheath, (mi) mitochondria. Bars, 180 nm.
of GAPDH, TIM and PGK activities and it revealed an additional minor protein band at 385 kDa in the native BN-PAGE
(Fig. 3, lane 3). It seems likely that this protein band represents
a complex of PGK with GAPDH and TIM. In contrast to the
proposed covalent binding of GAPDH to the fibrous sheath,
the aggregation among the enzymes may be due to the
hydrophobic nature of all three enzymes. Possibly, the three
enzymes form a multienzyme complex channelling triosephosphates to phosphoglyceric acid and producing ATP along the
fibrous sheath, the possible physiological significance of which
is discussed below.
The fibrous sheath
The fibrous sheath is exclusively found in the principal piece
Immature spermatozoa
A characteristical phenomenon of mammalian sperm maturation in the epididymis is the migration of a cytoplasmic droplet
from the proximal to the distal midpiece. During ejaculation
most droplets are shed and remain normal constituents of the
semen. The droplets remove all fragments of the Golgi
apparatus and the endoplasmic reticulum from the spermatozoa (Garbers et al., 1970) and also contain glycolytic enzymes
like hexokinase, glucose phosphate isomerase and lactate
dehydrogenase (Harrison and White, 1971). Our result on
GAPDH supplements the collection of glycolytic enzymes
which are removed from the midpiece by the droplets. In
addition, it also demonstrates that this protein is completely
disposed from the midpiece. This result was surprising and
suggests that the immature but not the mature spermatozoa
need a glycolytic capacity in the midpiece. One explanation
may be that immature spermatozoa show biosynthetic activity
which preferentially takes place around the nucleus and in the
midpiece where the endoplasmic reticulum and the Golgi
apparatus is concentrated. Biosynthesis requires both energy
and carbohydrates which may be produced by glycolysis or by
gluconeogenesis. Mature spermatozoa, however, lost their
biosynthetic activity and consequently the energy demand
shifted towards cell motility. In contrast to the biosynthetic
machinery, the dynein ATPases are distributed along the
axoneme not only of the midpiece but also of the principal
piece.
Energy metabolism of mature spermatozoa
Energy metabolism of boar and other mammalian spermatozoa
is dependent on extracellular fructose or glucose, and lactate is
released from the spermatozoa even when sufficient oxygen is
available (for review see Mann and Lutwak-Mann, 1981). Consequently, glycolytic flux exceeds pyruvate oxidation in the
mitochondria. The reason for this aerobic lactate fermentation
may be the compartmentation of mitochondria in the midpiece
and an insufficient energy transport system from the mitochondria to the distal dynein ATPases. As mentioned before, a
PCr/CK system is not present in boar and bull spermatozoa and
probably is also of relatively low capacity in other mammalian
spermatozoa (Kamp et al., 1996; Kaldis et al, 1997). Glycolytic
ATP-production in the principal piece may compensate an
insufficient transport system producing ATP close to the distal
dynein ATPases.
On the other hand, mitochondrial ATP production appears also
1828 D. Westhoff and G. Kamp
Fig. 8. Scheme of the proposed
compartmentation of the two different
ATP-producing reactions in the
flagellum of mammalian spermatozoa.
Motility requires ATP for the function
of dynein-ATPases located along the
whole axoneme. In the midpiece
(upper part) mitochondria are present
and the FoF1-ATP-synthase
regenerates ATP from ADP and
inorganic phosphate (Pi). In the
principal piece (lower part) ATP
regeneration occurs by glycolysis; the
location of the central enzyme
complex, GAPDH together with TIM
and PGK, along the fibrous sheath has
been proposed in this paper.
essential for cell motility and cannot be replaced by glycolysis.
Correspondingly, boar spermatozoa are not tolerant against
anoxia (for review see Mann and Lutwak-Mann, 1981) despite
the relatively high glycolytic capacity (Kamp et al., 1996). They
immediately loose their forward motility under anoxic conditions
(Mann and Lutwak-Mann, 1981). Probably mitochondria are
neccessary for the ATP production in the midpiece.
Fig. 8 summarizes the proposed compartmentation of the
two different ATP producing reactions, one in the midpiece
(mitochondrial ATP synthase reaction) and the other in the
principal piece (glycolytic reactions catalyzed by GAPDH and
PGK). Work is in progress to prove whether a lactate dehydrogenase isoenzyme is connected to GAPDH. This would
allow a rapid reoxidation of NADH. A sperm specific LDH
isoenzyme is known to be located in the mitochondria (called
LDH-X or LDH-4, McIndoe and Mitchell, 1978; Gallina et al.,
1994). The latter probably catalyses lactate oxidation to
pyruvate which is the substrate for respiratory ATP-production.
Other LDH isoenzymes are also present but their localization
is unknown. The localisation of other glycolytic enzymes will
follow to strengthen the hypothesis of a structure-function relationship between glycolysis and the fibrous sheath in the
principal piece of mammalian spermatozoa.
We thank Marita Koch and Barbara Hasert for excellent assistance
in protein biochemistry and histochemistry, respectively and Dr J. R.
Harris and Prof. Dr G. Wegener (Institute of Zoology, University
Mainz) for critically reading the manuscript. Supported by grants of
the NRW-government and the Deutsche Forschungsgemeinschaft (Ka
583/4-1).
REFERENCES
Allison, W. S. and Kaplan, N. O. (1964). The comparative enzymology of
triosephosphate dehydrogenase. J. Biol. Chem. 239, 2140-2152.
Bergmeyer, H. U. (1983). Methods of Enzymatic Analysis. Vol. 2 Verlag
Chemie, Weinheim.
Baltz, J. M., Williams, P. O. and Cone, R. A. (1990). Dense fibers protect
mammalian sperm against damage. Biol. Reprod. 43, 485-491.
Bode, J., Blumenstein, M. and Raftery, M. A. (1975). Nonidentical alkylation
sites in rabbit muscle glyceraldehyde-3-phosphate dehydrogenase. Biochem.
14, 1146-1152.
Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye
binding. Anal. Biochem 72, 248-254.
Brandon, C. and Tooze, J. (1991). Introduction to protein structure. Garland
Pub. Inc., New York.
Eberhardt, J. (1992). Isolation and characterization of five serine proteases
with trypsin-, chymotrypsin- and elastase-like characteristics from the gut of
the lugworm Arenicola marina (l.) (Polychaeta). J. Comp. Physiol. 162B,
159-167.
Fawcett, D. W. (1970). A comparative view of sperm ultrastructure. Biol.
Reprod. 2, 90-127.
Ferdinand, W. (1964). The isolation and specific activity of rabbit muscle
glyceraldehydephosphate dehydrogenase. Biochem. J. 92, 578-585.
Fouquet, J. P. and Kann, M. L. (1994). The cytoskeleton of mammalian
spermatozoa. Biol. Cell 81, 89-93.
Gallina, F. G., Gerez de Burgos, N. M., Burgos, C., Coronel, C. E. and
Blanco, A. (1994). The lactate/pyruvate shuttle in spermatozoa: Operation in
vitro. Archs. Biochem. Biophys. 308, 515-519.
Gandhi, K. K. and Anand, S. R. (1982). Regulation of glycolysis/fructolysis
in buffalo spermatozoa. J. Reprod. Fert. 64, 145-150.
Garbers, D. L., Wakabayashi, T. and Reed, P. W. (1970). Enzyme profile of
the cytoplasmic droplet from bovine epididymal spermatozoa. Biol. Reprod.
3, 327-337.
Harlow, E. and Lane, D. (1988). Antibodies, a Laboratory Manual. Cold
Spring Harbor Laboratory Press, New York.
Harris, J. I. and Perham, R. N. (1968). Glyceraldehyde 3-phosphate
dehydrogenase from pig muscle. Nature 219, 1025-1028.
Harris, J. I. and Waters, M. (1976). Glyceraldehyde 3-phosphate
dehydrogenase In The Enzymes XIII: Oxidation-Reduction, part C (ed. P. D.
Boyer, H. Lardy and K. Myrbäck), pp. 1-49. Academic Press, New York.
Harrison, R. A. P. (1971). Glycolytic enzymes in mammalian spermatozoa.
Biochem. J. 124, 741-750.
Harrison, R. A. P. and White, I. G. (1971). Glycolytic enzymes in the
spermatozoa and cytoplasmic droplets of bull, boar and ram, and their
leakage after shock. J. Reprod. Fert. 30, 105-115.
Huzar, G. and Vigue, L. (1990). Spermatogenesis-related change in the
synthesis of the creatine kinase B-type and M-type isoforms in human
spermatozoa. Mol. Reprod. Dev. 25, 258-262.
Kaldis, P., Stolz, M., Wyss, M., Zanolla, E., Rothen-Rutishauser, B.,
Vorherr, T. and Wallimann, T. (1996). Identification of two distinctly
localized mitochondrial creatine kinase isoenzymes in spermatozoa. J. Cell
Sci. 109, 2079-2088.
Kaldis, P., Kamp, G., Piendl, T. and Wallimann, T. (1997). Functions of
creatine kinase isoenzymes in spermatozoa. Dev. Biochem. (in press).
Kamp, G., Büsselmann, G. and Lauterwein, J. (1996). Spermatozoa: models
for studying regulatory aspects of energy metabolism. Experientia 52, 487494.
Kim, Y. H., McFarlane, J. R., Almahbobi, G., Stanton, P. G., TempleSmith, P. D. and de Kretser, D. M. (1995). Isolation and partial
characterization of rat sperm tail fibrous sheath proteins and comparison with
Structural binding of GAPDH in spermatozoa 1829
rabbit and human spermatozoa using a polyclonal antiserum. J. Reprod.
Fertil. 104, 107-114.
Knull, H. R. and Walsh, J. L. (1992). Association of glycolytic enzymes with
the cytoskeleton. Cur. Top. Cell. Regul. 33, 15-30.
Krietsch, W. K. G. and Bücher, T. (1970). 3-Phosphoglycerate kinase from
rabbit skeletal muscle and yeast. Eur. J. Biochem. 17, 568-580.
Laemmli, E. K. (1970). Clearage of structural proteins during the assembly of
the head of bacteriophage T4. Nature 227, 680-685.
Mann, T. and Lutwak-Mann, C. (1981). Male Reproductive Function and
Semen. Springer Verlag, Berlin.
Masters, C. J., Reid, S. and Don, M. (1987). Glycolysis - new concepts in an
old pathway. Mol. Cell. Biol. 76, 3-14.
McIndoe, W. M. and Mitchell, G. G. (1978). Lactate dehydrogenase
isoenzymes in the spermatozoa of the domestic fowl, Gallus domesticus and
turkey, Meleagris gallopavo. Comp. Biochem. Physiol. 61B, 433-437.
Minaschek, G., Gröschel-Stewart, U., Blum, S. and Bereiter-Hahn, S.
(1992). Microcompartmentation of glycolytic enzymes in cultured cells. Eur.
J. Cell Biol. 58, 418-428.
Norton, I. L., Pfuderer, P., Stringer, C. D. and Hartmann, F. C. (1970).
Isolation and characterization of rabbit muscle triosephosphate isomerase.
Biochemistry 9, 4952-4958.
Read, A. L. and Northcote, W. J. (1981). Minimization of variation in the
response to different proteins of the Coomassie blue dye-binding assay for
protein. Analyt. Biochem. 116, 53-64.
Schägger, H., Cramer, W. and von Jagow, G. (1994). Analysis of molecular
masses and oligomeric states of protein complexes by blue native
electrophoresis and isolation of membrane protein compexes by twodimensional native electrophoresis. Anal. Biochem. 217, 220-230.
Srere, P. A. (1987). Complexes of sequential metabolic enzymes. Annu. Rev.
Biochem. 56, 89-127.
Tombes, R. M. and Shapiro, B. M. (1985). Metabolic channeling: a
phosphorylcreatine shuttle to mediate high energy phosphate transport
between sperm mitochondrion and tail. Cell 41, 325-334.
Wallimann, T., Moser, H., Zurbriggen, B., Wegmann, G. and Eppenberger,
H. M. (1986). Creatine kinase isozymes in spermatozoa. J. Muscle Res. Cell.
Motil. 7, 25-34.
(Received 7 May 1997 - Accepted 4 June 1997)