BIOLOGY OF REPRODUCTION 75, 492–500 (2006)
Published online before print 21 June 2006.
Adenylate Kinases 1 and 2 Are Part of the Accessory Structures in the Mouse
Wenlei Cao, Lisa Haig-Ladewig, George L. Gerton,3and Stuart B. Moss2,3
Center for Research on Reproduction and Women’s Health, Department of Obstetrics and Gynecology, University of
Pennsylvania Medical Center, Philadelphia, Pennsylvania 19104
Proper sperm function depends on adequate ATP levels. In the
mammalian flagellum, ATP is generated in the midpiece by
oxidative respiration and in the principal piece by glycolysis. In
locations where ATP is rapidly utilized or produced, adenylate
kinases (AKs) maintain a constant adenylate energy charge by
interconverting stoichiometric amounts of ATP and AMP with
two ADP molecules. We previously identified adenylate kinase 1
and 2 (AK1 and AK2) by mass spectrometry as part of a mouse
SDS-insoluble flagellar preparation containing the accessory
structures (fibrous sheath, outer dense fibers, and mitochondrial
sheath). A germ cell-specific cDNA encoding AK1 was charac-
terized and found to contain a truncated 30UTR and a different
50UTR compared to the somatic Ak1 mRNA; however, it
encoded an identical protein. Ak1 mRNA was upregulated
during late spermiogenesis, a time when the flagellum is being
assembled. AK1 was first seen in condensing spermatids and was
associated with the outer microtubular doublets and outer dense
fibers of sperm. This localization would allow the interconver-
sion of ATP and ADP between the fibrous sheath where ATP is
produced by glycolysis and the axonemal dynein ATPases where
ATP is consumed. Ak2 mRNA was expressed at relatively low
levels throughout spermatogenesis, and the protein was local-
ized to the mitochondrial sheath in the sperm midpiece. AK1
and AK2 in the flagellar accessory structures provide a
mechanism to buffer the adenylate energy charge for sperm
gametogenesis, sperm, sperm motility and transport,
Mammalian sperm motility is driven by the flagellum and is
dependent on the availability of an adequate and continued
supply of ATP. The flagellum can be segregated into three
domains, all of which contain the axoneme, a 9þ2 microtu-
bular structure. The midpiece contains the mitochondrial sheath
and 9 outer dense fibers (ODFs). The principal piece represents
the major portion of the tail and is comprised of the fibrous
sheath and 7 ODFs. A small end piece has only the axoneme
surrounded by the plasma membrane of the sperm. ATP is used
by the dynein ATPases that function as the flagellar motors and
in protein kinase A-mediated signal transduction pathways to
regulate motility. There are a number of factors that make the
generation and delivery of ATP difficult in a flagellum that
reaches a length of .100 lm in some species . The motile
wave is propagated along the length of the flagellum,
necessitating the availability of ATP to the dynein motors
present throughout the tail. As mentioned above, mitochondria
are present in the midpiece, which means that ATP generated
by oxidative phosphorylation must be transported to other
regions and/or that ATP is produced locally. Invertebrates such
as sea urchins utilize a phosphocreatine shuttle system to
actively move ATP down the length of the tail ; however,
such a system either is not present or is of limited importance in
mammalian sperm .
Mammalian sperm utilize glycolysis to generate ATP in the
principal piece of the tail. A variety of glycolytic enzymes are
localized to the fibrous sheath [4–7]. For example, hexokinase
is present in both the longitudinal columns and transverse ribs
of this structure . The importance of glycolysis was
elegantly demonstrated by the gene targeting of Gapdhs, the
sperm-specific isoform of glyceraldehyde 3-phosphate dehy-
drogenase . Gapdhs-null male mice have immotile sperm
and consequently are infertile. Furthermore, glycolysis is
required for mammalian sperm to acquire a hyperactivated
form of motility .
We recently performed a proteomic analysis of an SDS-
insoluble tail preparation containing the flagellar accessory
structures—fibrous sheath, outer dense fibers, and the
mitochondrial sheath—but lacking the axoneme and plasma
membranes of the flagella . This study identified a number
of proteins involved in the generation and utilization of ATP.
In addition to glycolytic enzymes, e.g., aldolase, triosephos-
phate isomerase, phosphoglycerate kinase, and glyceraldehyde
3-phosphate dehydrogenase, we determined that both adenylate
kinase 1 (AK1) and 2 (AK2) are present in the flagellar
accessory structures. Adenylate kinases (ATP:AMP phospho-
transferase, EC 18.104.22.168) are present in prokaryotes and
eukaryotes and participate in maintaining energy homeostasis
in cells [11–14]. These enzymes buffer the adenylate energy
charge by catalyzing the reaction, 2ADP , ATPþAMP, thus
producing either ADP or stoichiometric amounts of ATP and
AMP, depending on the concentrations of the three nucleotides
. Among the members of the AK family, the major AK
isoform, AK1, is found in the cytosol of tissues with high-
energy demands, e.g., skeletal muscle, while the more
ubiquitously expressed AK2 is localized mainly to mitochon-
Although there is a considerable body of literature on the
role of AK in heart and skeletal muscle tissues [17, 18],
relatively little is known about this enzyme family in
mammalian germ cells and sperm. Schoff et al.  showed
that AK activity exists in bovine sperm flagella and can
1Supported by NICHD HD06427 (S.B.M. and G.L.G.). The germ cell-
specific AK1 sequence described in this work has been assigned the
GenBank database accession number DQ486026.
2Correspondence: Stuart B. Moss, Center for Research on Reproduction
and Women’s Health, 1312 BRB II, 421 Curie Blvd., University of
Pennsylvania Medical School, Philadelphia, PA 19104.
FAX: 215 573 7627; e-mail: email@example.com
Received: 2 May 2006.
First decision: 26 May 2006.
Accepted: 13 June 2006.
? 2006 by the Society for the Study of Reproduction, Inc.
ISSN: 0006-3363. http://www.biolreprod.org
generate sufficient ATP to produce normal motility in
digitonin-permeabilized cells treated with MgADP. We found
that AK1 and AK2 are localized in different compartments of
the flagellum, suggesting that these enzymes play important
roles in buffering the adenylate energy charge in sperm.
MATERIALS AND METHODS
Isolation of Germ Cells and Sperm
All animal procedures were approved by the University of Pennsylvania
Institutional Animal Care and Use Committee. Male germ cells were prepared
from decapsulated testes of adult mice (CD1 retired breeders, Charles River
Laboratories) by sequential dissociation with collagenase and trypsin-DNase I
[20, 21]. The cells were separated at unit gravity in a 2%–4% BSA gradient in
an enriched Krebs-bicarbonate medium to purify pachytene spermatocytes,
round spermatids, and condensing spermatids. Both the pachytene spermato-
cyte and round spermatid populations were at least 85% pure, while the
condensing spermatid population was approximately 40%–50% pure (contam-
inated primarily with anuclear residual bodies and some round spermatids).
Epididymal sperm were collected by mincing the caudae epididymides and
allowing the sperm to swim out in PBS. The sperm were collected by
centrifugation at 800 3 g for 5 min at room temperature, and SDS-resistant
head and tail structures were separated . Briefly, sperm were homogenized
in 1% SDS, 75 mM NaCl, 24 mM EDTA, pH 6.0 (S-EDTA), layered on 1.6 M
sucrose gradient in S-EDTA, and centrifuged at 5000 3 g for 1 h at room
temperature. The SDS-resistant tail structures were collected from the interface.
Generation of AK2 Antibody and
peptide Ac-RSYHEEFNPPKEPMKDDIC-amide (amino acids 150–167, acces-
sion number NP_058591) (Quality Controlled Biochemicals, Hopkinton, MA).
Proteins from epididymal sperm and tail preparations were separated by
10% SDS-PAGE and transferred to Hybond ECL nitrocellulose membranes
(Amersham Biosciences, Buckingham, UK). The membranes were blocked
with TBST (25 mM Tris-HCl, pH 8.0; 125 mM NaCl; 0.1% Tween 20)
containing 5% nonfat dry milk (NFDM) and incubated with primary antibody
[rabbit anti-mouse AK1, a generous gift of Dr. Edwin Janssen (Radboud
University, The Netherlands) 1:10000 dilution in 5% NFDM in TBST; rabbit
anti-mouse AK2, 1:1000 dilution in 5% NFDM in TBST]. After washing with
TBST, the blots were incubated with secondary antibody [donkey anti-rabbit
IgG conjugated with horseradish peroxidase (Amersham, Buckinghamshire,
UK), 1:5000 dilution in 5% NFDM in TBST]. The bound enzyme was detected
with the ECL kit according to the manufacturer’s directions (Amersham) and
exposed to film. As a control, the anti-AK1 was neutralized by adding a 10-fold
molar excess of the peptide in 200 ll 5% NFDM in TBST and incubating the
mixture at 48C overnight. After centrifugation to remove any particulate
material, the supernatant was used for immunoblot analysis.
Epididymal sperm were collected, attached to slides, fixed with 4%
paraformaldehyde for 15 min, and permeabilized with ?208C methanol for 2
min. In some cases, sperm were treated with S-EDTA to separate heads and
tails and to solubilize the membrane and axoneme prior to attaching to slides.
The slides were washed with PBS, and the samples were incubated with 10%
goat serum in PBS (blocking solution) for 30 min at room temperature and then
with the primary antibody (either anti-AK1 antibody, 1:2500 dilution in 10%
goat serum or anti-AK2 antibody, 1:50 dilution in 10% goat serum) in blocking
solution for 1 h at room temperature. For a control, PBS-goat serum was
substituted for the primary antibody. After washing with PBS, the samples were
incubated with the secondary antibody [donkey anti-rabbit IgG linked either
with Alexa Fluo-488 (for AK1) or with 568 (for AK2); Molecular Probes,
Eugene, OR] (1:500 dilution in 10% goat serum) in blocking solution for 1 h at
room temperature. After washing with PBS, the samples were mounted with
coverslips using Fluoromount-G (Southern Biotechnology Associates, Inc,
Birmingham, AL), examined using an inverted microscope (Nikon Eclipse TE
2000-u, Nikon Corp), and photographed with a CFW-1310C color digital
camera (Scion Corp., Frederick, MD).
For immunoelectron microscopy, sperm and SDS-insoluble accessory
structures were isolated and fixed with 4% paraformaldehyde and 0.25%
glutaraldehyde; all reagents were purchased from Electron Microscopy
Sciences (Fort Washington, PA). The samples were embedded in Lowicryl
K4M that was then polymerized with ultraviolet light (365 nm) for 5 days.
Ultrathin sections were cut and mounted on nickel grids coated with Formvar.
To prevent nonspecific binding, grids were incubated with blocking buffer
(PBS with 1.0% BSA and 2% normal goat serum) for 30 min at room
temperature. The sections were incubated with primary antibody (anti-AK1
antibody 1:1000; anti-AK2 antibody 1:50 in blocking buffer) for 1 h at room
temperature, washed in buffer, and incubated with 15 nm gold particle-labeled
anti-rabbit IgG (1:10 in blocking buffer) (Ted Pella, Inc., Redding, CA). After a
1 h incubation at room temperature, the grids are rinsed with buffer followed by
deionized water for 3 min, air-dried, and then examined with a FEI Tecnai G2
electron microscope. Images were collected with a Gatan Camera (Gatan, Inc.,
Isolation of Germ Cell AK1 cDNAs
A probe corresponding to a portion of an Ak1b cDNA (nucleotides 143–
492, accession # NM_021515) was amplified by PCR. After labeling with
[a32P]dCTP (3000 Ci/mmol) by random priming (Invitrogen, Piscataway, NJ),
the probe was used to screen a total of 500000 plaques from a mixed germ cell
cDNA library in the HybriZAP 2.1 vector (Stratagene, La Jolla, CA). The
library was plated and the plaques transferred to Hybond Nþ membranes
(Amersham Biosciences) that were treated with DNA denaturing solution (1.5
M NaCl, 0.5 M NaOH) and then DNA neutralizing solution (1.5 M NaCl, 0.5
M Tris-HCl, pH 8.0). Membranes were pre-hybridized in Rapid-hyb buffer
(Amersham Biosciences) for 15 min at 658C and then incubated overnight at
428C in Rapid-hyb buffer containing the radiolabeled probe. The membranes
were washed with 23SSPE/0.1% SDS at room temperature and then with 0.1
3SSPE/0.1% SDS at 658C and developed by exposure to film. Positive clones
were purified by two additional rounds of screening and converted to plasmids
according to the manufacturer’s directions. All clones were sequenced in both
directions using an ABI 3100 DNA sequence analyzer and the BigDye
chemistry (Applied Biosystems, Inc., Foster City, CA). The sequences were
analyzed using MacVector (Accelrys, Madison, WI) and Sequencher (Gene
Codes, Ann Arbor, MI) software.
Preparation of RNA and RT-PCR
RNA was prepared from germ cells, testis, and various somatic tissues
using Tri Reagent (Sigma; St. Louis, MO). Reverse transcription using 1 lg
mRNA was performed using SuperScript II Reverse Transcriptase according to
the manufacturer’s instructions (Invitrogen Corp., Carlsbad, CA). Products
were amplified using Extaq DNA Polymerase (Takara Co., Toyko, Japan) and
the appropriate primers (Table 1). Amplicons were cloned into the pCR2.1-
TOPO (Invitrogen). Plasmid DNA was prepared and sequenced as described
For quantitative RT-PCR assays, primers were designed using Primer
Express 1.5 Taqman Primer Design software (Applied Biosystems) (Table 2).
Products were amplified with the SYBR Green PCR Master Mix and analyzed
with the ABI 7900 HT Sequence Detection system. The following PCR
protocol was used: 1) denaturation (508C for 2 min, 958C for 10 min), 2)
amplification and quantification (958C for 15 sec, 608C for 1 min) repeated for
40 cycles, 3) a dissociation curve program (958C for 15 sec, 608C for 15 sec,
958C for 15 sec), and, 4) cooling at 48C. Amplicons were analyzed by
generating a dissociation curve and determining the threshold cycle (Ct) value
for each transcript. The relative quantification of gene expression was analyzed
by the 2-DDCTmethod . The mRNA corresponding to ribosomal protein S16
(accession number: BC082286) was used as a control .
Two-dimensional Gel Electrophoresis and
Two-dimensional gel electrophoresis was performed as previously
described except that 7-cm Immobiline DryStrips (pH 3–10, nonlinear) were
used . Equipment and reagents were from Amersham Bioscience (Uppsala,
Sweden). Briefly, samples (60 ll containing 100 lg protein) were mixed with
65 ll rehydration buffer (8 M urea, 2% w/v CHAPS, 0.5% IPG buffer [pH 3–
10, nonlinear], 2 mg/mL DTT) and loaded in the IPGphor strip holder. The
DryStrips were placed in the holder and overlaid with ;2 ml DryStrip cover
fluid. Strips were hydrated at 30 V for 14 h and then focused for a total of 80
kVh at 208C on the IPGphor IEF system. After electrophoresis, each strip was
equilibrated with 5 ml equilibration buffer A (6 M urea, 100 mM Tris-HCl pH
8.8, 30% v/v glycerol, 1% w/v SDS, 1% DTT) by rocking for 15 min, and then
with 5 ml equilibration buffer B (8 M urea, 100 mM Tris-HCl pH 6.8, 30% v/v
glycerol, 1% w/v SDS, 2.5% iodoacetamide) for an additional 15 min. For the
second dimension, the strips were placed on top of a Novex Bis-Tris 4%–12%
gel (Invitrogen). After electrophoresis, proteins were stained with Colloidal
Coomassie Blue for subsequent protein identification.
Two-dimensional gels were scanned with a Typhoon 9400 scanner
(Amersham), and the spots analyzed with DeCyder software (Amersham)
and picked manually. After digestion with trypsin, a portion of the digest was
ADENYLATE KINASE IN FLAGELLAR ACCESSORY STRUCTURES
analyzed directly by MALDI-TOF for molecular weight determination and
MALDI-TOF/TOF for sequence information using a Voyager 4700 proteomics
analyzer mass spectrometer (Applied Biosystems). Another portion of the
digest was subjected to nanoLC/Qstar-XL analysis (Applied Biosystems). The
data were acquired and analyzed with Analyst QS. The protein identification
and database search were performed with Mascot dll script of Analyst QS; the
combined MS and MS/MS data were used for the Mascot database search. A
protein score of .70 with a confidence identification of .95% was considered
Our previous proteomic analysis of the SDS-insoluble
flagellar accessory structures identified a number of proteins
involved in the generation and utilization of ATP . In
particular, we identified both cytoplasmic AK1 and mitochon-
drial AK2 as part of these structures. Two highly homologous
isoforms of somatic AK1 have been described. AK1 (accession
number: AAH14802) and AK1b (accession number:
NP_067490) are 194 and 210 amino acids long, respectively,
with AK1b having an additional 16 amino acids at its amino
terminal end. AK2 (accession number: NP_058591) is a 239
amino acid protein with ;30% identity to AK1 and AK1b. All
three proteins contain an ATP-binding site and the AK
We confirmed that AK1 and AK2 were present in sperm and
the SDS-insoluble tail fraction by immunoblotting. A band of
the predicted size for AK1 (Mr;22000) was seen in proteins
from both samples when probed with anti-AK1 (Fig. 1A); this
band was absent when the immunoblot was probed with
normal rabbit serum (data not shown). An abundant AK1-
immunoreactive band of Mr;56000 also was present in
sperm, but not in the SDS-insoluble tail preparation, which
initially suggested to us that a larger, soluble isoform of AK1
might be present in these cells. Using anti-AK2, a band of the
expected size (Mr;28000) was seen in protein extracts from
sperm and an SDS-insoluble tail preparation (Fig. 1B).
Preabsorption of the antibody with the peptide used as the
antigen eliminated immunoreactivity (data not shown). Both
AK1 and AK2 were not enriched in the SDS-insoluble tail
preparation compared to levels of the proteins in intact sperm,
indicating that portions of these proteins were solubilized under
these extraction conditions. The most abundant axonemal
protein, tubulin, is not seen in the tail preparation, indicating
that our SDS-insoluble preparation contained accessory
structures but not components of the axoneme .
Identification of cDNAs in Germ Cells Encoding Adenylate
To identify cDNAs encoding AK1 in spermatogenic cells,
we screened a mixed germ cell cDNA library. Two different
cDNA clones, a 1961 nt long form (Ak1_v1) and a 819 nt short
form (Ak1_v2), corresponding to AK1 were isolated and
sequenced. Ak1_v1 was not a full-length cDNA, as it lacked the
nucleotide sequence encoding the amino-terminal end of the
protein and the 50UTR; however, it was identical to the Ak1/
Ak1b cDNAs throughout the remaining open reading frame
and 30UTR. In comparison, the Ak1_v2 cDNA was a full-
length transcript, as it contained an in-frame stop codon
upstream of the initiator methionine. This sequence has been
assigned the GenBank database accession number DQ486026.
While this cDNA contained a different 50UTR and a truncated
30UTR when compared to Ak1_v1, the protein encoded by
Ak1_v2 was identical to somatic AK1. Alignment of Ak1_v2,
Ak1_v1, and various Ak1 cDNAs and ESTs to the Ak1 genomic
sequence showed that Ak1_v2 shared most of its exons with
TABLE 1. Primers used for RT-PCR.
GenBank accession no.Primer Amplicon size (bp)
AK RIKENNM_001033874 318
* Also known as Ak4.
?Homologous to human AK6.
TABLE 2.Primers used for quantitative RT-PCR.
GeneForward primerReverse primerConcentration (nM)Amplicon size (bp)
* Reference .
CAO ET AL.
Ak1_v1 (Fig. 2). However,itsfirstexon(containingthemajority
of the 50UTR) was upstream of the 50UTR of Ak1_v1,
suggesting that Ak1_v2 was transcribed from an alternative
promoter. A cDNA (accession number: AK046613) was
characterized from adipose tissue of a 4-day-old neonate and
had the same 50UTR as Ak1_v2; however, an amplicon
corresponding to this cDNA was not present in adult adipose
tissue when assayed by RT-PCR (Fig. 3). Several ESTs from
testis and round spermatids were identical to regions of Ak1_v2.
Testis-Specific Expression of Ak1_v2
To determine whether the Ak1_v2 cDNA was testis-specific,
RT-PCR was performed using primers that would only amplify
RNA corresponding to this transcript. An amplicon of the
predicted size was detected when testicular RNA, but not RNA
from a variety of somatic tissues, was used (Fig. 3A). When
primers corresponding to common regions of Ak1 were utilized,
heart and skeletal muscle, in addition to testis (Fig. 3B). In
3C). An amplicon corresponding to b-actin was present in all
samples, confirming the integrity of the RNA (data not shown).
Consistent with the RT-PCR results, immunoblot analysis
showed a restricted pattern of AK1 expression in testis and
somatic tissues, i.e., a band of Mr;22000 was present in
heart, skeletal muscle, and testis (Fig. 4). In contrast, a band of
Mr;28000 corresponding to AK2 was present in nearly all
tissues analyzed, although at varying levels. Higher molecular
weight bands that were immunoreactive with anti-AK2 were
also detected in testis and liver. Preabsorption of the antibody
with the peptide used as the antigen eliminated immunoreac-
tivity of these bands (data not shown).
AK1 and AK2 mRNAs and Proteins are Differentially
Expressed During Spermatogenesis
To examine the expression of adenylate kinases 1 and 2
during spermatogenesis, we used quantitative real-time RT-
PCR and immunoblotting. Ak1 mRNA was not detected in
pachytene spermatocytes (Fig. 5A). It was first detected at low
levels in round spermatids, indicating that the mRNA is post-
meiotically expressed. Transcript levels became much more
abundant in condensing spermatids compared to round
spermatids (.20-fold). Ak1 mRNA in condensing spermatids
was present at ;30-fold higher levels compared to Ak2 mRNA
(compare Fig. 5, A and B). AK1 was only detected in
condensing spermatids, demonstrating that it is post-meiotical-
ly transcribed and translated (Fig. 5C).
In contrast, Ak2 mRNA was present at low but relatively
equivalent levels in pachytene spermatocytes, round sperma-
tids, and condensing spermatids (Fig. 5B). AK2 also was
present in these cell types (Fig. 5D). Its lower level in
condensing spermatids compared to the other cell types most
likely reflects the loss of flagella from condensing spermatids
as a consequence of the enzymatic treatment required to
prepare purified spermatogenic cell populations.
In addition to AK1 and AK2, other murine adenylate
kinases have been identified using the UniGene database. To
organization of Ak1_v1 and Ak1_v2 to other cDNAs and RIKEN clones
encoding AK1 are compared. The arrow shows the position of start codon
and the asterisk shows the position of stop codon.
Alignment of Ak1 sequences at the genomic level. The exon
RNA was prepared from testis (Te), brain (Br), liver (Li), oviduct (Ov),
adipose (Ad), heart (He), skeletal muscle (Sk), spleen (Sl), and lung (Lu),
and used for RT-PCR analysis (see Table 1 for primers used). (-) No
template. A) Expression of the Ak1_v2 mRNA. RT-PCR was performed
using a primer corresponding to a region in the unique 50UTR of the
Ak1_v2 cDNA and a primer corresponding to a portion of the common
open reading frame of Ak1. B) Expression of Ak1 mRNA. RT-PCR was
performed with primers corresponding to a portion of the common open
reading frame of Ak1. C) Expression of Ak2 mRNA. RT-PCR was performed
with primers corresponding to a portion of the common open reading
frame of Ak2. Amplicons of the predicted size are indicated (arrows).
Ak1 and Ak2 mRNAs are expressed in testis and somatic tissues.
Proteins from epididymal sperm (Sp) and SDS-insoluble tails (Ta) were
prepared and processed for immunoblot analysis. A) Immunoblot probed
with anti-AK1 antiserum (a) and normal rabbit serum (b). Arrow points to
an immunoreactive band of the predicted size for AK1. Arrowhead points
to an unknown immunoreactive band of Mr;56000 that is only seen in
sperm protein extracts. B) Immunoblot probed with anti-AK2 antiserum (a)
and the antiserum preabsorbed with peptide (b). Arrow points to an
immunoreactive band of the predicted size for AK2.
AK1 and AK2 are present in sperm and SDS-insoluble tails.
ADENYLATE KINASE IN FLAGELLAR ACCESSORY STRUCTURES
determine whether additional Ak mRNAs were present in
spermatogenic cells, RT-PCR was performed using mixed
germ cell RNA and specific primers to Ak3–5, 7, Taf9
(homologous to human AK6) , and a RIKEN clone
(accession number: NM_001033874) that encodes a protein
with an AK signature domain and an ATP binding site. All of
these transcripts were present in mixed germ cells (Fig. 6).
Based on the amount of input template cDNA required for
amplification, the levels of these adenylate kinase mRNAs
were in the general range of Ak2 and not nearly as abundant as
The Mr;56000 Sperm Protein Is Not an AK1 Isoform
An abundant AK1-immunoreactive band of Mr;56000
was detected in sperm extracts (Fig. 1); however, we never
identified a cDNA encoding a protein of this size. Quantita-
tive RT-PCR showed that the Ak1 mRNA was expressed post-
meiotically (Fig. 5A). On the other hand, the Mr;56000
band appeared earlier during spermatogenesis, i.e., in
pachytene spermatocytes, indicating it is not an AK1 isoform
(Fig. 5C). To identify this immunoreactive band, we separated
sperm protein extracts by 2D gel electrophoresis. Three spots
corresponding to the Mr;56000 band were cored from the
gel and identified after trypsin digestion and mass spectrom-
etry. Two spots were identified as b-tubulin, an abundant
sperm protein that has a molecular size and pI consistent with
the location of the cored spot in the 2D gel (Table 3).
However, our AK1 antibody did not immunoreact with
purified tubulin (data not shown). The third spot was
identified as ATP synthase b-chain (EC 22.214.171.124; accession
number: P56480). This protein shares homologous regions
with AK  and has been reported to be present in rat sperm
. We conclude that the antibody is most likely reacting
with ATP synthase b-chain; since this protein is not seen in
the SDS-flagellar protein preparation (Fig. 1), it is not part of
the accessory structures.
testis and somatic tissues. Protein was
prepared from brain (Br), heart (He), intes-
tine (In), kidney (Ki), liver (Li), skeletal
muscle (Sk), spleen (Sl), testis (Te), and
sperm (Sp), and analyzed by immunoblot-
ting with anti-AK1 and anti-AK2. A) Immu-
noblot probed with anti-AK1 polyclonal
antibody. Arrow points to an immunoreac-
tive band of the predicted size for AK1.
Arrowhead points to an unknown immu-
noreactive band of Mr;56000 that is only
seen in testis and sperm protein extracts. B)
Immunoblot probed with anti-AK2 poly-
clonal antibody. Arrow points to an immu-
noreactive band of the predicted size for
AK1 and AK2 are expressed in
encoded proteins are differentially ex-
pressed in germ cells. A, B) Quantitative RT-
PCR of Ak1 (A) and Ak2 (B) from pachytene
spermatocytes (PS), round spermatids (RS),
and condensing spermatids (CS). Results
were normalized to mRNA corresponding
to ribosomal protein S16. C, D) Immunoblot
analysis of AK1 (C) and AK2 (D) from mixed
germ cells (MGC), pachytene spermatocytes
(PS), round spermatids (RS), condensing
spermatids (CS), and epididymal sperm (Sp).
Arrows point to immunoreactive bands of
the predicted size for AK1 and AK2 in each
respective immunoblot. Arrowhead points
to the immunoreactive band of Mr;56000
that is present in all germ cells and sperm.
Ak1 and Ak2 mRNAs and their
CAO ET AL.
AK1 and AK2 are Differentially Localized in Sperm
AK1 was localized to the entire length of the sperm
flagellum in a punctate pattern (Fig. 7A). Because we were
concerned that the larger Mr;56000 band recognized by anti-
AK1 might confound any results seen in sperm, the localization
was repeated using SDS-insoluble tail structures, which do not
contain this protein (Fig. 1). A similar pattern was observed,
i.e., a punctate pattern throughout the length of the sperm tail
(Fig. 7C). Immunoelectron microscopy showed that AK1 was
associated with both the outer dense fibers and the outer
microtubular doublets of intact sperm (Fig. 7, E and F). AK1
was also retained in the ODFs of the SDS-insoluble tail
structures, which lack the axonemal components (Fig. 7, G and
H). The association of AK1 with the ODFs and axoneme
would allow the interconversion of ATP and ADP between the
fibrous sheath where ATP is produced by glycolysis and the
axoneme where ATP is consumed by the dynein ATPases.
As expected from previous studies in somatic cells that
showed that AK2 is found in mitochondria, AK2 localized to
the midpiece of the sperm flagellum, the site of the
mitochondrial sheath (Fig. 8A). A similar localization pattern
was found when SDS-insoluble tail structures were analyzed,
confirming that AK2 is part of the accessory structures (data
not shown). Immunoelectron microscopy showed that AK2
was present in the mitochondrial sheath in both sperm and
SDS-insoluble accessory structures (Fig. 8C–F). The protein
was found clustered in the inner aspect of the sheath; this
clustering indicates that AK2 functions in spatially restricted
microdomains. In addition, the mitochondrial localization of
AK2, close to the outer dense fibers, suggests that AK1 and
AK2 work in concert to maintain the adenylate energy charge
in the sperm flagellum.
A substantial and continuing supply of ATP is necessary for
the various events associated with mammalian sperm function.
The hydrolysis of ATP by the axonemal ATPases in the long
flagellum supplies the energy required to regulate sperm
movement in a coordinated manner. We, and others, have
hypothesized that ATP is generated by multiple mechanisms
and that this production is compartmentalized along the
flagellum so that ATP can function in a localized fashion [3,
5, 8]. Because oxidative phosphorylation generates more ATPs
than glycolysis per glucose equivalent, the ATP necessary for
mammalian sperm motility was thought to be generated by
mitochondrial respiration in the midpiece (the most proximal
region of the sperm tail, where all the mitochondria are
located). This ATP either would have to diffuse to other
regions of the tail or be actively transported down the flagellum
from the midpiece, possibly by a phosphocreatine shuttle
system similar to that seen in sea urchin sperm . However,
the length of the mammalian flagellum probably precludes
efficient diffusion , and the phosphocreatine system is
either poorly developed or absent in these sperm .
Glycolysis is a major ATP generator for mammalian sperm
motility. When oxidative phosphorylation in mouse sperm
mitochondria is suppressed, sperm remain motile as long as
glucose is present in the medium . However, motility is
greatly diminished when sperm are incubated in a medium
without glucose. Glycolytic enzymes have been localized to the
fibrous sheath in the principal piece, a more distal region of the
tail to the mitochondria, indicating that glycolysis could supply
localized ATP to the ATPases along the entire length of the
principal piece. Male mice carrying a targeted deletion of the
gene encoding the sperm-specific glyceraldehyde 3-phosphate
dehydrogenase are infertile; their sperm are immotile and have
very low cellular ATP levels . Other studies demonstrated
that glycolysis is necessary for hyperactivated motility .
We previously showed that glucose is necessary for the
tyrosine phosphorylation of a subset of sperm proteins
associated with capacitation, while uncouplers of oxidative
phosphorylation do not affect this process .
In addition to oxidative phosphorylation and glycolysis,
another way to generate cellular ATP is via adenylate kinase,
where two molecules of ADP can be interconverted to one
molecule of ATP and one molecule of AMP. Because AKs
have an equilibrium constant approaching 1, they maintain the
adenylate energy charge at a relatively constant value . In
locations where ATP is being used rapidly, e.g., the sperm
flagellum, the reaction is in the direction of ATP production,
allowing the cell to scavenge energy from ADP during peaks of
net energy consumption.
Our proteomic analysis identified both AK1 and AK2 in the
accessory structures surrounding the axoneme, leading us to
hypothesize that these enzymes are in positions to buffer the
liver (Li), brain (Br), and mixed germ cells (MGC) were prepared and used
for RT-PCR with unique primers corresponding to Ak3–7 and the RIKEN
clone (Table 1). (-) No template. Amplicons of the predicted size are
Other Ak mRNAs are expressed in mixed germ cells. RNA from
TABLE 3. Identification of the Mr;56000 protein.
accession no. Predicted pIPredicted MW No. of peptidesCoverage (%)Mascot Score
450 ATP synthase b chain
ADENYLATE KINASE IN FLAGELLAR ACCESSORY STRUCTURES
adenylate energy charge in the flagellum . AK1 and 2 are
differentially localized in the sperm flagellum, with AK2 being
in the mitochondrial sheath and AK1 located at the outer dense
fiber-outer microtubular doublet interface. This localization is
reminiscent of SPAG4, which has been postulated to act as part
of the link between the ODF and axoneme . SPAG4 binds
ODF1 (but not ODF2), but whether AK1 binds either SPAG4
or ODF1 is not known.
There are at least seven murine genes (Ak1–7) and a RIKEN
clone (accession number: NM_001033874) predicted to encode
microtubular doublets of the sperm flagellum. A) Indirect immunofluo-
rescence of sperm probed with anti-AK1. B) The corresponding Nomarski
image. C) Indirect immunofluorescence of a SDS-insoluble tail prepara-
tion probed with anti-AK1. D) The corresponding Nomarski image. E)
Immunoelectron microscopy of sperm probed with anti-AK1 showing a
region of the midpiece. F) Immunoelectron microscopy of sperm probed
with anti-AK1 showing a region of the principal piece. G) Immunoelec-
tron microscopy of a SDS-insoluble tail preparation probed with anti-AK1
showing a region of the midpiece. H) Immunoelectron microscopy of a
SDS-insoluble tail preparation probed with anti-AK1 showing a region of
the principal piece. I) Immunoelectron microscopy of sperm probed with
secondary antibody alone. J) Immunoelectron microscopy of SDS-
insoluble tail preparation probed with secondary antibody alone. ODF,
outer dense fibers; FS, fibrous sheath; MS, mitochondrial sheath; Ax,
axoneme. Bar ¼ 10 lm (A–D), 0.2 lm (E–I), 0.5 lm (J).
AK1 is associated with the outer dense fibers and outer
flagellum. A) Indirect immunofluorescence of sperm probed with anti-
AK2. B) The corresponding Nomarski image. Arrows demarcate the
boundaries of the midpiece where the mitochondrial sheath is located. C,
D) Immunoelectron microscopy of sperm probed with anti-AK2. E, F)
Immunoelectron microscopy of SDS-insoluble tails probed with anti-AK2.
G) Immunoelectron microscopy of sperm probed with secondary antibody
alone. H) Immunoelectron microscopy of SDS-insoluble tails probed with
secondary antibody alone. ODF, outer dense fibers; MS, mitochondrial
sheath; Ax, axoneme. Bar¼10 lm (A, B), 0.2 lm (C–E, G, H), 0.5 lm (F).
AK2 is present in the mitochondrial sheath of the sperm
CAO ET AL.
adenylate kinases. (The approved gene symbol for Ak4 is
Ak3L1.) These isozymes have different expression and/or
localization patterns. While we focused on AK1 and AK2 in
this study because of their locations in the flagellar accessory
structures, we also found mRNAs corresponding to all the
known AK isoforms in mixed germ cells. This differs from
recent analyses using the less sensitive Northern blot technique
that did not find all of the Ak genes expressed in the testis .
Future experiments need to be directed towards determining
whether all the different Ak transcripts are translated and if (and
where) the proteins are present in sperm.
The multiplicity of AKs expressed in murine germ cells is
similar to that of the flagellated protozoan parasite, Trypano-
soma brucei, which expresses 7 adenylate kinase genes .
Of note, three of the AK proteins are present in the flagellum
either in the axoneme or the extra-axonemal paraflagellar rod, a
structure that is required for motility in this organism. Multiple
AKs are also found in the flagellum of Chlamydomonas
reinhardtii [34, 35]. Similar to mammalian sperm, Chlamydo-
monas lacks creatine kinase, eliminating the possibility of ATP
transport by a phosphocreatine shuttle system. In addition, the
finding that motility is reactivated in the presence of ADP
alone suggests the presence of AKs . A Chlamydomonas
adenylate kinase with three AK domains is anchored by two
outer dynein arm proteins, ODF5p and Oda10p . Another
Chlamydomonas protein, CPC1, is found in the central pair of
the axoneme and contains an unusual AK domain . The
presence of two proteins with AK activity explains why
mutation of either protein alone reduces but does not totally
eliminate motility in Chlamydomonas. Homologues of Cpc1
are found in a number of mammalian organisms having motile
flagella, suggesting that additional proteins with AK function
are present in the axoneme . However, such proteins would
not be identified in our proteomic analyses, as they would be
solubilized during the isolation of the accessory structures.
The disruption of the adenylate kinase gene, adk1, in
Saccharomyces pombe is lethal . When this gene is
disrupted in Saccharomyces cerevisiae, there is a compensatory
metabolism mechanism to keep the cells alive; however, they
have a petite phenotype and can not grow under non-
fermentative conditions [38, 39]. The ablation of the murine
Ak1 gene resulted in mice which are viable, although the
phenotype indicated that AK1 is critical for the maintenance of
skeletal muscle energetic economy [17, 40]. In particular, the
potential of myofibers to sustain nucleotide ratios is decreased,
and a higher ATP turnover rate is seen. Metabolic stress
amplified this reduction in energy homeostasis. Importantly,
glycolysis and the guanylate and creatine kinase phospho-
transfer pathways are upregulated, presumably compensating
for the loss of AK1. Similar results are seen in hearts of AK1-
null animals, i.e., an upregulation of alternative high-energy
transfer pathways prevents gross abnormalities in the absence
of metabolic stress .
The function of AK1 and AK2 in sperm remains unclear.
AK activity was detected in bovine sperm flagellum and can
generate sufficient ATP to produce motility in digitonin-
permeabilized cells treated with MgADP . When this
enzyme activity is inhibited by P1,P5-di(adenosine 50)-
pentaphosphate, an AK-specific inhibitor, motility is disrupted.
Surprisingly, a fertility defect was not reported in Ak1-null
male mice, although a detailed assessment of sperm function
was not performed . It is possible that the Ak1-null sperm
compensate by other mechanisms, e.g., upregulation of other
AK proteins and/or glycolysis in the fibrous sheath of the tail.
The Ak2 gene has not been eliminated by targeting, although its
widespread expression pattern suggests that the Ak2-null
phenotype would be embryonic lethal. Future experimentation
should offer insights into the role of the adenylate kinases in
We are grateful to Drs. Be ´ Wieringa and Edwin Janssen for generously
providing the AK1 antibody. Mass spectrometry and peptide micro-
sequencing were provided by the Proteomics Core Facility of the
Geonomics Institute and the Abramson Cancer Center at the University
of Pennsylvania. We thank Dr. Chao-Xing Yuan and Christine Busch of this
facility for their expertise and guidance during the course of this work.
1. Cummins JM, Woodall PF. On mammalian sperm dimensions. J Reprod
Fertil 1985; 75:153–175.
2. Tombes RM, Shapiro BM. Metabolite channeling: a phosphorylcreatine
shuttle to mediate high energy phosphate transport between sperm
mitochondrion and tail. Cell 1985; 41:325–334.
3. Ford WC. Glycolysis and sperm motility: does a spoonful of sugar help the
flagellum go round? Hum Reprod Update 2006; 12:269–274.
4. Westhoff D, Kamp G. Glyceraldehyde 3-phosphate dehydrogenase is
bound to the fibrous sheath of mammalian spermatozoa. J Cell Sci 1997;
5. Travis AJ, Foster JA, Rosenbaum NA, Visconti PE, Gerton GL, Kopf GS,
Moss SB. Targeting of a germ cell-specific type 1 hexokinase lacking a
porin-binding domain to the mitochondria as well as to the head and
fibrous sheath of murine spermatozoa. Mol Biol Cell 1998; 9:263–276.
6. Mori C, Nakamura N, Welch JE, Gotoh H, Goulding EH, Fujioka M,
Eddy EM. Mouse spermatogenic cell-specific type 1 hexokinase (mHk1-s)
transcripts are expressed by alternative splicing from the mHk1 gene and
the HK1-S protein is localized mainly in the sperm tail. Mol Reprod Dev
7. Bunch DO, Welch JE, Magyar PL, Eddy EM, O’Brien DA. Glyceralde-
hyde 3-phosphate dehydrogenase-S protein distribution during mouse
spermatogenesis. Biol Reprod 1998; 58:834–841.
8. Miki K, Qu W, Goulding EH, Willis WD, Bunch DO, Strader LF,
Perreault SD, Eddy EM, O’Brien DA. Glyceraldehyde 3-phosphate
dehydrogenase-S, a sperm-specific glycolytic enzyme, is required for
sperm motility and male fertility. Proc Natl Acad Sci U S A 2004; 101:
9. Bone W, Jones NG, Kamp G, Yeung CH, Cooper TG. Effect of ornidazole
on fertility of male rats: inhibition of a glycolysis-related motility pattern
and zona binding required for fertilization in vitro. J Reprod Fertil 2000;
10. Cao W, Gerton GL, Moss SB. Proteomic profiling of accessory structures
from the mouse sperm flagellum. Mol Cell Proteomics 2006; 5:801–810.
11. Kishi F, Tanizawa Y, Nakazawa A. Isolation and characterization of two
types of cDNA for mitochondrial adenylate kinase and their expression in
Escherichia coli. J Biol Chem 1987; 262:11785–11789.
12. Nakazawa A, Yamada M, Tanaka H, Shahjahan M, Tanabe T. Gene
structures of three vertebrate adenylate kinase isozymes. Prog Clin Biol
Res 1990; 344:495–514.
13. Lee Y, Kim JW, Lee SM, Kim HJ, Lee KS, Park C, Choe IS. Cloning and
expression of human adenylate kinase 2 isozymes: differential expression
of adenylate kinase 1 and 2 in human muscle tissues. J Biochem (Tokyo)
14. Villa H, Perez-Pertejo Y, Garcia-Estrada C, Reguera RM, Requena JM,
Tekwani BL, Balana-Fouce R, Ordonez D. Molecular and functional
characterization of adenylate kinase 2 gene from Leishmania donovani.
Eur J Biochem 2003; 270:4339–4347.
15. Atkinson DE. The energy charge of the adenylate pool as a regulatory
parameter. Interaction with feedback modifiers. Biochemistry 1968; 7:
16. Noma T. Dynamics of nucleotide metabolism as a supporter of life
phenomena. J Med Invest 2005; 52:127–136.
17. Janssen E, Dzeja PP, Oerlemans F, Simonetti AW, Heerschap A, de Haan
A, Rush PS, Terjung RR, Wieringa B, Terzic A. Adenylate kinase 1 gene
deletion disrupts muscle energetic economy despite metabolic rearrange-
ment. Embo J 2000; 19:6371–6381.
18. Pucar D, Janssen E, Dzeja PP, Juranic N, Macura S, Wieringa B, Terzic A.
Compromised energetics in the adenylate kinase AK1 gene knockout heart
under metabolic stress. J Biol Chem 2000; 275:41424–41429.
19. Schoff PK, Cheetham J, Lardy HA. Adenylate kinase activity in ejaculated
bovine sperm flagella. J Biol Chem 1989; 264:6086–6091.
ADENYLATE KINASE IN FLAGELLAR ACCESSORY STRUCTURES
20. Romrell LJ, Bellve AR, Fawcett DW. Separation of mouse spermatogenic
cells by sedimentation velocity. A morphological characterization. Dev
Biol 1976; 49:119–131.
21. Bellve AR, Cavicchia JC, Millette CF, O’Brien DA, Bhatnagar YM, Dym
M. Spermatogenic cells of the prepuberal mouse. Isolation and
morphological characterization. J Cell Biol 1977; 74:68–85.
22. O’Brien DA, Bellve AR. Protein constituents of the mouse spermatozoon.
I. An electrophoretic characterization. Dev Biol 1980; 75:386–404.
23. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using
real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods
24. Jeong YJ, Choi HW, Shin HS, Cui XS, Kim NH, Gerton GL, Jun JH.
Optimization of real time RT-PCR methods for the analysis of gene
expression in mouse eggs and preimplantation embryos. Mol Reprod Dev
25. Ren H, Wang L, Bennett M, Liang Y, Zheng X, Lu F, Li L, Nan J, Luo M,
Eriksson S, Zhang C, Su XD. The crystal structure of human adenylate
kinase 6: an adenylate kinase localized to the cell nucleus. Proc Natl Acad
Sci U S A 2005; 102:303–308.
26. Garboczi DN, Fox AH, Gerring SL, Pedersen PL. Beta subunit of rat liver
mitochondrial ATP synthase: cDNA cloning, amino acid sequence,
expression in Escherichia coli, and structural relationship to adenylate
kinase. Biochemistry 1988; 27:553–560.
27. Kawa G, Yamamoto A, Yoshimori T, Muguruma K, Matsuda T,
Moriyama Y. Immunohistochemical localization of V-ATPases in rat
spermatids. Int J Androl 2000; 23:278–283.
28. Nevo AC, Rikmenspoel R. Diffusion of ATP in sperm flagella. J Theor
Biol 1970; 26:11–18.
29. Mukai C, Okuno M. Glycolysis plays a major role for adenosine
triphosphate supplementation in mouse sperm flagellar movement. Biol
Reprod 2004; 71:540–547.
30. Williams AC, Ford WC. The role of glucose in supporting motility and
capacitation in human spermatozoa. J Androl 2001; 22:680–695.
31. Travis AJ, Jorgez CJ, Merdiushev T, Jones BH, Dess DM, Diaz-Cueto L,
Storey BT, Kopf GS, Moss SB. Functional relationships between
capacitation-dependent cell signaling and compartmentalized metabolic
pathways in murine spermatozoa. J Biol Chem 2001; 276:7630–7636.
32. Shao X, Tarnasky HA, Lee JP, Oko R, van der Hoorn FA. Spag4, a novel
sperm protein, binds outer dense-fiber protein Odf1 and localizes to
microtubules of manchette and axoneme. Dev Biol 1999; 211:109–123.
33. Ginger ML, Ngazoa ES, Pereira CA, Pullen TJ, Kabiri M, Becker K, Gull
K, Steverding D. Intracellular positioning of isoforms explains an
unusually large adenylate kinase gene family in the parasite Trypanosoma
brucei. J Biol Chem 2005; 280:11781–11789.
34. Zhang H, Mitchell DR. Cpc1, a Chlamydomonas central pair protein with
an adenylate kinase domain. J Cell Sci 2004; 117:4179–4188.
35. Wirschell M, Pazour G, Yoda A, Hirono M, Kamiya R, Witman GB.
Oda5p, a novel axonemal protein required for assembly of the outer
dynein arm and an associated adenylate kinase. Mol Biol Cell 2004; 15:
36. Hyams JS, Borisy GG. Isolated flagellar apparatus of chlamydomonas:
characterization of forward swimming and alteration of waveform and
reversal of motion by calcium ions in vitro. J Cell Sci 1978; 33:235–253.
37. Konrad M. Molecular analysis of the essential gene for adenylate kinase
from the fission yeast Schizosaccharomyces pombe. J Biol Chem 1993;
38. Konrad M. Analysis and in vivo disruption of the gene coding for
adenylate kinase (ADK1) in the yeast Saccharomyces cerevisiae. J Biol
Chem 1988; 263:19468–19474.
39. Bandlow W, Strobel G, Zoglowek C, Oechsner U, Magdolen V. Yeast
adenylate kinase is active simultaneously in mitochondria and cytoplasm
and is required for non-fermentative growth. Eur J Biochem 1988; 178:
40. Janssen E, de Groof A, Wijers M, Fransen J, Dzeja PP, Terzic A, Wieringa
B. Adenylate kinase 1 deficiency induces molecular and structural
adaptations to support muscle energy metabolism. J Biol Chem 2003;
CAO ET AL.