Báo cáo khoa học: Spermosin, a trypsin-like protease from ascidian sperm cDNA cloning, protein structures and functional analysis doc

Spermosin, a trypsin-like protease from ascidian sperm
cDNA cloning, protein structures and functional analysis
Eri Kodama
1
, Tadashi Baba
2
, Nobuhisa Kohno
2
, Sayaka Satoh
2
, Hideyoshi Yokosawa
1
and Hitoshi Sawada
1
1
Department of Biochemistry, Graduate School of Pharmaceutical Sciences, Hokkaido University, Japan;
2
Institute of Applied Biochemistry, University of Tsukuba, Tsukuba Science City, Japan
We have previously reported t hat two trypsin-like enzymes,
acrosin a nd spermosin, play key roles in sperm penetration
through the vitelline c oat of the ascidian (Urochordata)
Halocynthia roretzi [Sawada et al. (1984), J. Biol. C hem. 25 9,
2900±2904; Sawada et al. (1984), Dev. Biol. 105, 246±249].
Here, we show the amino-acid sequence of the ascidian
preprospermosin, which is deduced from the nucleotide
sequence of the isolated cDNA clone. Th e isolated ascidian
preprospermosin cDNA consisted of 1740 nucleotides, and
an open reading frame encoding 388 amino acids, w hich
corresponds to a molecular mass o f 41 896 Da. By sequence
alignment, it was suggested that His178, A sp230 and Ser324
make up a catalytic triad and that ascidian spermosin be

classi®ed as a novel trypsin family member. The mRNA of
preprospermosin is speci®cally expressed in ascidian gonads
but not in other tissues. P uri®ed spermosin c onsists of
33- and 40-kDa bands as determined by SDS/PAGE under
nonreducing conditio ns. The 4 0-kDa s permosin consists of a
heavy chain (residues 130±388) and a long light chain
designated L1 (residues 23±129), whereas the 33-kDa
spermosin includes the same heavy chain and a shorter light
chain d esignated L 2 ( residues 97±129). T he L1 chain
contains a proline-rich region, designated L1(DL2) which is
lacking i n L 2. Investigation with the glutathione-S-trans-
ferase (GST)±spermosin-light-chain fusion proteins, includ-
ing GST±L1, GST±L2, and GST±L1(DL2), revealed that the
proline-rich region in the L1 chain binds to the vitelline coat
of ascidian eggs. Thus, we propose that sperm spermosin is a
novel tryp sin-like p rotease that binds to the vitelline coat and
also plays a part in penetration of sperm through the vitelline
coat during ascidian fe rtilization.
Keywords: acrosin; ascidian; fertilization; lysin; spermosin;
trypsin-like protease; vitelline coat.
Fertilization is a pivotal event in the creation of new
individuals. In order to accomplish species-speci®c sperm±
egg fusion, sperm binding to and penetration through the
extracellular coat of the eggs (the vitelline coat in marine
invertebrates and zona pellucida in mammals), must be
precisely controlled, as the vitelline coat-free eggs would
allow gamete fusion with sperm from different species.
Upon p rimary binding of the sperm to the vitelline coat,
the sperm undergoes an acrosome reaction, which is an
exocytosis of the a crosomal vesicle l ocated on the s perm

head [1]. A lytic agent called a sperm lysin is exposed on the
surface of the sperm head and is partially released into the
surrounding seawater. In mammals, a trypsin-like enzyme
called acrosin (EC 3 .4.21.10) has long been believed to be a
zona-lysin [2,3], as the puri®ed acrosin is capable of
dissolving the zona pellucida in in vitro experiments [2,3].
However, recent studies with acrosin-knockout mice
convincingly demonstrated that acrosin is not essential for
in vivo sperm penetration through the zona pellucida [4,5]. It
is currently thought that acrosin is involved in the dispersal
of acrosomal contents during acrosome reaction [6]. These
results led us to propose that a sperm protease other than
acrosin m ay play a k ey role in the penetration of sperm
through the zona pellucida.
Ascidians (Urochordata) occupy a phylogenetic position
between invertebrates and segmented vertebrates [7].
Whereas all the ascidians are hermaphrodites that release
sperm and eggs simultaneously during the spawning season,
self-fertilization is strictly prohibited in several species
including Halocynthia roretzi [8]. As the vitelline coat-free
eggs of H. roretzi are self-fertile [8], the interaction between
sperm and the vitelline coat of the egg seems to be the
process of self±nonself recognition in ascidian fertilization.
Therefore, the vitelline coat lysin system seems to be
activated after the sperm recognizes the vitelline coat o f the
egg as nonself.
To investigate the biological functions of sperm proteases,
one of the largest solitary ascidians, H. roretzi,wasusedin
this study; fertilization experiments are more a ccessible than
those in mammals, and large amounts of s perm and egg are

obtainable from thousands of these animals which are
cultivated in Onagawa Bay for human consumption.
We have previously reported that H. roretzi sperm con-
tain a nove l trypsin-like protease called ascidian spermosin
Correspondence to H. Sawada, Department of Biochemistry, Graduate
School of Pharmaceutical Sciences, Hokkaido University, Sapporo
060-0812, Japan. Fax: + 81 11 706 4900, Tel.: + 81 11 706 3720,
E-mail:
Abbreviations:Boc,t-butyloxycarbonyl; L1, long light chain (residues
23±129); L 2, short light chain (residues 97±129); L1(DL2), L2-deleted
L1 (residues 23±96); MCA, 4-methylcoumaryl-7-amide; GST,
glutathione S-transferase.
Enzyme: ascidian sperm spermosin (EC. 3.4.21.99).
Note: nucleotide sequence reporte d in this paper has been submitted to
the DDBJ/Ge nBank/E BI Data Bank under accession number
AB052776.
(Received 28 August 2001, revised 16 November 2001, accepted 21
November 2001)
Eur. J. Biochem. 269, 657±663 (2002) Ó FEBS 2002
in addition to acrosin, an ascidian homologue of
mammalian acrosin. Ascidian spermosin has unique prop-
erties, especially in substrate speci®city: it hydrolyses only
t-butyloxycarbonyl (Boc)-Val-Pro-Arg-4-methylcoumaryl-
7-amide (MCA) among many ¯uorogenic substrates [9].
In addition, involvement of spermosin in ascidian fertiliza-
tion was r evealed by examining the effects o f leupeptin
analogues and anti-spermosin antibody on fertilization of
H. roretzi [10±12]. The presence of the spermosin-like
protease in the sperm of the other animals, including
mammals, has n ot yet been investigated. Therefore, t he

unique properties of ascidian spermosin led us to assume
that ascidian spermosin belongs to a novel trypsin-like
protease of sperm origin. In order to clarify this issue, we
attempted to isolate a c DNA clone encoding ascidian
spermos in. We found that ascidian spermosin consists of
two chains: a light chain encoded in the N-terminal portion
and a heavy chain encoded at the C-terminal portion. We
also found that there are two forms of spermosin i n sperm,
which share the same h eavy chain but are distinct i n the
length of light chains: one contains a long L1 chain and the
other contains a shorter L 2 chain. Furthermore, the proline-
rich region in the L 1 chain is capable of binding to the
vitelline coat, implying a r ole in sperm binding to the
vitelline coat.
MATERIALS AND METHODS
Biologicals
The solitary ascidian (Urochordata) Halocynthia roretzi
type C was used in this study. Sperm and eggs were
collected from dissected gonads as described previou sly
[13,14]. Mature oocytes wer e homogenized with ®vefold
diluted (20%) arti®cial seawater containing 0.1 m
M
diiso-
propyl¯uorophosphate. The homogenate was ®ltered
through a nylon mesh (pore size, 150 lm), and the vitelline
coats o n t he blotting cloth w ere w ashed e xtensively with
20% arti®cial seawater. Purity of the isolated vitelline coats
was examined under a light microscope.
Puri®cation and assay procedure of spermosin
The enzymatic activity of spermosin was determined using

Boc-Val-Pro-Arg-MCA as a substrate as described previ-
ously [9]. Spermosin was highly puri®ed from H. roretzi
sperm b y DEAE-cellulose chromatography, S ephadex
G-100 gel ®ltration, and soybean trypsin inhibitor-immobi-
lized Sepharose chromatography according to procedure
described previously [9].
Determination of N-terminal amino-acid sequences
SDS/PAGE was c arried out on a slab gel containing 12.5%
polyacrylamide as described previously [15]. Puri®ed sperm-
osin was subjected to SDS/PAGE under reducing and
nonreducing conditions and was then electrophoretically
transferred to a PVDF me mbrane (Millipore). The blotted
membrane was stained with 0.1% Coomassie brilliant blue
R-250 containing 1% acetic acid and 40% methanol. After
washing with 50% methanol, the bands were cut from the
membrane. The N-terminal sequence of t he puri®ed
spermosin was determined using a protein sequencer model
Procise 492 (Applied Biosystems). The solubilized vitelline
coat component, which is able to bind to the glutathione
S-transferase (GST)-L1 and GST-L1(DL2) fusion proteins,
was subjected to SDS/PAGE and transferred to a PVDF
membrane. The 28-kDa band on a membrane was subjected
to amino-acid sequence analysis.
Cloning of spermosin cDNA
The primers (sense and antisense) used for PCR were
designed from the N-terminal amino-acid sequence of
H. roretzi spermosin (heavy chain): sense primer, 5¢-AT
(T/C/A)GT(T/C/A/G)GG(T/C/A/G)GG(T/C/A/G)GC
(T/C/A/G)GA(A/ G)GC-3¢; and antisense primer, 5 ¢-AA
(T/C/A/G)GG(T/C/A/G)GG(T/C)GT(T/C/A)TA(T/C)

AG(T/C)A T-3¢. The former and latter p rimers encoded t he
amino-acid sequences IVGGAEA and YDIXGGK, respec-
tively. The primers at concentrations of 10 l
M
were mixed
in PCR to amplify the H. roretzi gonad kgt11 cDNA library
as described p reviously [16]. A DNA band migrating at 78
base pairs was isolated, cloned into a pCRII v ector, and
transformed into Escherichia c oli DH5a. The spermosin
cDNA clones were isolated from 3 ´ 10
5
clones of H. rore-
tzi gonad kgt11 cDNA library by phage plaque hybridiza-
tion using the above PCR-ampli®ed DNA fragment
encoding spermosin as a probe. The probe was labelled
with [a
32
P]dCTP (BcaBEST kit, Takara) by the random-
priming procedure. Brie¯y, plaque lifts were prehybridized
at 55 °Cin5´ NaCl/Cit (1 ´ NaCl/Cit, 15 m
M
sodium
citrate p H 7.0 and 0.15
M
NaCl), 0.02% Ficoll 400, 0.02%
polyvinylpyrrolidone, 0.02% BSA, 0.1% SDS, and
0.1 mgámL
)1
salmon testis DNA. Hybridization was carried
out at 55 °C overnight in prehybridization buffer containing

32
P-labelled probe. The membranes were w ashed i n
2 ´ NaCl/Cit at room temperature for 10 min, in
2 ´ NaCl/Cit containing 0.1% SDS at 6 0 °C for 10 min,
and in 2 ´ NaCl/Cit at room temperature for 10 min before
autoradiography at )80 °C. The nucleotide s equence of
spermosin cDNA clone was determined by a Big Dye
Terminator Cycle Sequencing R eady Reaction using an
ABI 377A DNA Sequencing A pparatus (Applied Biosys-
tems).
Northern blot analysis
Total RNA was extracted from the H. roretzi gonad
according to the standard method of acid/guanidinium
thiocyanate/chloroform. Poly(A)
+
RNA was isolated from
total RNA by using oligotex-dT30 (Roche Diagnostics
Co.). Two microgrmas of poly(A)
+
RNA w ere subjected to
electrophoresis on 1.2% agarose gel containing 6% form-
aldehyde, and RNA bands were transferred to a Hybond-
N+ nylon membrane (Amersham). Stringency used for
hybridization and washing, and the probe were the same as
those, used in ÔcloningÕ. After washing, the blots were
autoradiographed at )80 °C.
Extraction of the vitelline coat
The v itelline coats were suspended i n arti®cial seawater
containing 0.5% Triton X-100. After stirring for 30 min at
4 °C, the suspension was centrifuged at 1 0 000 g for 30 min

to obtain the supernatant as the solubilized vitelline coat.
658 E. Kodama et al. (Eur. J. Biochem. 269) Ó FEBS 2002
Almost all the protein c omponents o f the vitelline coat were
solubilized under these conditions.
Expression of GST±spermosin light chain fusion proteins
DNA fragments corresponding to the L1, L2, and L1(DL2 ),
which is an L1 lacking L2 region, were ampli®ed by PCR
from the cDNAs using the following combinations of
forward and reverse p rimers designed to produce 5¢ BamHI
and 3¢ XhoI restriction sites to facilitate directional cloning
into the expression v ector pGEX-6P-1 (Amersham Phar-
macia): two forward primers (a, 5¢-GGATCCTCT
GAATCTACAAATCC-3¢;b,5¢-GGATCCTCTGAAG
GCCCGGTTC-3¢) a nd two reverse primers (c, 5¢-CTC
GAGTCATTTTCCTTTCTTTAG-3¢;d,5¢-CTCGAGT
CAATTTTCAGATTCCG-3¢) w ere designed and th e com-
bination of primers for L1, L2, and L1(DL2) were (a + c)
(b + c), and (a + d), respectively. All clones were
sequenced to con®rm the reading frame and sequence.
The GST fusion protein expression vectors were used for
expression in E. coli BL21. The e xpressed fusion proteins
were puri®ed using glutathione±agarose beads (Amersham)
according to the manufacturer's protocol.
Binding of GST±spermosin light chain fusion proteins
to the vitelline coat
The puri®ed GST±spermosin light chain fusion proteins
(3 lg) were mixed with the solubilized vitelline c oat i n
arti®cial seawater, and incubated at 4 °C for 1 h to form the
complex consisting of the fusion p rotein and the vite lline
coat. The mixture was applied onto a glutathione±agarose

column and washed four times with arti®cial seawater. The
GST±spermosin light chain f usion p rotein±vitelline coat
complex was eluted from the glutathione±agarose column
with 50 m
M
Tris/HCl (pH 8.8) containing 20 m
M
glutathi-
one. T he eluted proteins were subjected to SDS/PAGE
followed by silve r staining (Kanto Chemical Co. Inc.,
Tokyo, Japan) or by blotting to a PVDF membrane. The
bands detected were cut from the membrane and the
N-terminal amino-acid sequence was determined as
described above.
RESULTS
CDNA cloning of ascidian spermosin
Spermosin was highly puri®ed from H. roretzi sperm as
described previously [9]. Puri®ed spermosin was subjected to
SDS/PAGE under reducing conditions, f ollowed by blot-
ting to a PVDF membrane. The s equence of 33 amino acid
residues from the N terminus of th e 28-kDa spermosin
band, which corresponds to that of the heavy chain as
described b elow, was determined using a protein sequencer
(see Figs 1 and 3). The N-terminal sequence o f the
spermosin (heavy chain) was used to design the degenerate
oligonucleotide primers for P CR of H. roretzi gonad
cDNA. The PCR product was used as a probe for screening
the gonad kgt11 cDNA library to isolate a spermosin clone
(Fig. 1). A single open reading frame o f the spermosin clone
encodes 388 amino acid s. The deduced amino-acid sequence

contains a region f rom residue 130 to residue 162 t hat
corresponds to the N-terminal amino-acid sequence deter-
mined by sequencing of the puri®ed spermosin (heavy
chain) protein. The molecular mass of preprospermosin was
estimated to be 41 896 Da. T he N-terminal sequence (22
residues) of the preprospermosin corresponds to a signal
peptide for a nascent protein destined for initial transfer to
the endoplasmic reticulum. Thus, the pro-form of spermosin
Fig. 1. Nucleotide and d educed amino-acid se quences of H. roretzi
spermosin. The a mino-acid sequence as d etermined by a p rotein
sequencer is underlined. The conserved catalytic triads in the serine
protease are indicated by boxes. The presumed cleavage sites in
preprospermosin and prospermosin are indicated by an arrow and two
arrowheads, respectively. Note that the cleavage of prospermosin at
the second arrowhead yields the L1 light chain and heavy chains, while
the two cleavages at two arrowheads yields the L2 light chain and
heavy chain (see Fig. 3C).
Ó FEBS 2002 Cloning and characterization of ascidian spermosin (Eur. J. Biochem. 269) 659
may start from Ser23 (see F igs 1 and 3 ). The active site
residues in serine proteases, histidine, aspartic acid, a nd
serine, were located at residues 178, 230, and 324,
respectively, in preprospermosin. This indicates that sperm-
osin is classi®ed i nto a family S1 (trypsin family) of clan SA
in serine proteinases [17]. The amino-acid sequence of
spermosin (heavy chain) showed 32% homology to that
of mouse plasma kallikrein and 27% homology to those of
mouse and ascidian acrosin (see Fig. 5).
A dendrogram analysis showed that ascidian spermosin is
classi®ed as a n ovel member of the S1 trypsin family (data
not shown).

Expression of spermosin mRNA in ascidian gonads
Northern blotting was c arried ou t with the same probe used
for cDNA cloning. A single transcript of approximately
1.9 kb was detected in the gonad, but not in the hepato-
pancreas, intestine, or branchial basket, of H. roretzi
(Fig. 2).
The presence of two forms of spermosin in ascidian
sperm
SDS/PAGE of the puri®ed spermosin gave a s ingle band of
28 kDa under reducing conditions, whereas it showed two
bands of 33 and 4 0 kDa under nonreducing c onditions
(Fig. 3A). The N-terminal sequence determination of these
bands (Fig. 3B) revealed that the 28-kDa protein is the heavy
chain of spermosin, while the 33-kDa spermosin is made up
of the heavy chain (residues 130±388) and the light chain
designated as L2 (residues 97±129), and the 40-kDa sperm-
osin consists of the heavy chain (residues 130±388) and the
light chain designated as L1 (residues 2 3±129) (Fig. 3C).
From these results, it was concluded that there are two forms
of spermosin in ascidian sperm and that the amount of the
33-kDa form is higher than that of the 40-kDa form. Models
for protein s tructures o f preprospermosin a nd spermosin
type 1 and type 2 are depicted in Fig. 3C.
Binding of spermosin light chains to the vitelline coat
Comparison of th e sequences between L1 a nd L2 light
chains revealed that the L1 but not the L2 had a region
1.9 kb
Gn Hp In
Bb
Fig. 2. Tissue-speci®c e xpression of spermosin mRNA i n H. roretzi.

Gn, Gonad; Hp, hepatopancreas; In, intestine; Bb, branchial basket.
Northern blots of poly(A)
+
RNA (2 lg each) from H. roretzi tissues
were hybridized with radiolabelled cDNA probes. A 1.9-kb mRNA
signal was detected only in the gonad.
SEST
40 kDa: IVGG
33 kDa:
SEGP
IVGG
28 kDa: IVGG
BA
29
45
2-ME
+ -
(kDa)
28
33
40
(kDa)
C
Preprospermosin
Spermosin type 1
Spermosin type 2
Signal
peptide
Light chain
Heavy chain

S
S
SEST IVGG
S S
SEGP IVGG
1
23 97 130 388
Cleavage site
H D S
178 230 324
28 kDa
(33 kDa)
L1
28 kDaL2
(40 kDa)
Fig. 3. The presence of two forms of spermosin in H. roretzi sperm. (A)
SDS/PAGE of the puri®ed spermosin. SDS/PAGE gave a band of
28 kDa under red ucing conditions and two b ands of 33 and 40 kDa
under nonreducing co nd itions. 2-ME, 2-mercaptoeth anol. (B) The
N-terminal amino-acid seq uences of three bands. T he 28-kDa protein
had a single amino-acid sequence and was determined as a heavy chain
of spermosin, whereas the 40-kD a protein con siste d of a he avy chain
(residues 130±388) and a light chain designated as L1 (residues 23±129,
designated L1), and the 33-kDa protein consisted of the heavy chain
(residues 130 ±388) and a light chain designated as L2 (residues
97±129). There were no distinc t bands of L1 and L 2 chains on SDS/
PAGE (12.5% gel) under the reducing conditions (see also Fig. 2 in
[9]), probably because of their high electrophore tic mobility that is
indistinguishable from the migration front. Alternatively, the low
molecular mass p roteins (11-kDa L1 chain and 3-kDa L2 chain) m ight

not be suciently ®xed within the g el under our experimental condi-
tions. (C) Protein structure s of ascidian preprospermosin and sperm-
osin type 1 and 2. The N-terminal amino-acid sequences of L1, L2, and
heavy chains are shown above the respective models. The putative
disul®de bond is indicated by analogy to the oth er trypsin family [17].
660 E. Kodama et al. (Eur. J. Biochem. 269) Ó FEBS 2002
containing high amounts of proline residues (residues
28±88, see Fig. 1). It has been suggested that the proline-
rich regions located in the C t erminals of human and
porcine proacrosins play a key role in interaction between
proacrosin and the zona pellucida [18]. To investigate the
binding ability of the proline-rich region in ascidian
spermosin to the vitelline coat, three GST fusion proteins
(Fig. 4A), i ncluding L1, L2, and L1(DL2), an L1 lacking
the L2 region, were expressed and puri®ed by glutathione±
agarose chromatography. The puri®ed GST fusion
proteins were incubated with the solubilized vitelline coat
and the complex formed was adsorbed to the glutathione±
agarose beads. After washing, the vitelline c oat protein
components, which can interact with the f usion proteins,
were eluted with 20 m
M
glutathione and analysed by SDS/
PAGE. By comparison of protein patterns in SDS/PAGE,
it was found that th e 28-kDa band was detected only with
the GST-L1 or GST-L1(DL2) fusion proteins, but not with
the GST-L2 fusion protein (Fig. 4B), indicating that the
28-kDa protein of the vitelline coat has an ability to bind
to the proline-rich region present in the L1(DL2) domain of
the spermosin light chain. The predicted N-terminal

amino-acid sequence (SAXARNQNFG) showed no
appreciable identity with an y proteins by
FASTA
and
BLAST
database search analyses.
DISCUSSION
The present study demonstrated the amino-acid sequence of
spermosin from sperm of the ascidian H. roretzi for the ®rst
time. Here we show that there are two molecular forms of
the molecule made up of a common h eavy chain and either
a short or a long light chain depending on the processing
sites. We previously reported that ascidian spermosin is a
novel sperm trypsin-like protease, distinct from acrosin, a
well-known sperm trypsin-like protease that is widely
distributed in mammalian s perm, in terms of substrate
speci®city and inhibitor susceptibility [9,12]. The present
study clearly showed that ascidian spermosin is a novel
protease and is distinct from ascidian acrosin on the basis of
amino-acid sequence [16]. Whereas the acrosin family has a
C-terminal extension as a pro-piece of proacrosin, sperm-
osin has n o such C-terminal region. In contrast, the
N-terminal region of the light chain of ascidian spermosin
contains a p roline-rich region, which is observed in the
C-terminal region of proacrosin of nonrodent mammals
[19]. We identi®ed the vitelline coat component, to which the
proline-rich region of the L1 chain of spermosin binds, for
the ®rst time.
The sequence alignments of trypsin family proteases
(Fig. 5), suggested that the active site residues, histidine,

aspartic acid, and serine, are located at residues 178, 230,
and 324, respectively, in preprospermosin (see Figs 1 and 5 ).
Although spermosin heavy chain showed the highest
homology to mouse plasma kallikrein (32% ide ntity)
among the trypsin family, spermosin is unlikely to be a
functional homologue of kallikrein: spermosin is not able to
hydrolyse Pro-Phe-Arg-MCA [9], a preferred substrate for
plasma kallikrein, but is able to ef®ciently hydrolyse Boc-
Val-Pro-Arg-MCA [9], a preferred substrate for thrombin.
Ascidian spermosin is initially synthesized as a 388-residue
preproprotein with a 22-residue signal peptide in the
terminus (Figs 1 and 3). The N-terminal region from
residue N-23 to residue 129 (i.e. the L1 light chain) in the
pro-form of ascidian spermosin, which precedes the IVGG
sequence, is thought to be cleaved off at the bond between
Lys129 and Ile130 by the action of a putative trypsin-like
protease. As the sequence around the scissile bond is
A
B
GST-L1
GST-L1 (∆L2)
GST-L2
23 129
9623
97 129
GST
GST
GST
GST alone
GST-L1(

GST-L1 GST-L1(∆L2)GST-L2
28 kDa
Fig. 4. Expression of recombinant H. roretzi
spermosin light chain-GST fusion proteins.
(A) GST fusion prote ins inc luding L1, L2,
and L1(DL2). (B) The fusion proteins were
expressed in clones generated by PCR from
the spermosin c DNA as described in Mate-
rials and methods. GST±spermosin light
chain fusion proteins (3 lg each), which had
been previously puri®ed by glutathione±
agarose chromatography, were mixed with
the solubilized vitelline coat in arti®cial sea-
water. After incub ation at 4 °C for 1 h, the
fusion protein was applied to a column of
glutathione±agarose beads. The proteins
eluted with 20 m
M
glutathione w ere sub-
jected to SDS/PAGE and visualized by silver
staining. The N-terminal amino-acid
sequence of the 28-kDa protein, which was
detected in b oth cases of G ST-L1 and GST-
L1(DL2) fusion p roteins, was determined as
describedinMaterialsandmethods.
Ó FEBS 2002 Cloning and characterization of ascidian spermosin (Eur. J. Biochem. 269) 661
Lys-Lys-Gly-Lys-Ile-Val-Gly-Gly(126±133), together with
the fact that spermosin hydrolyses only Boc-Val-Pro-Arg-
MCA among peptidyl-MCA substrates, autocatalytic acti-
vation of prospermosin seems unlikely. Acrosin is a

candidate protease that cleaves the Lys129±Ile130 bond of
prospermosin.
Here we showed the existence of two f orms of spermosin
in ascidian sperm: the type 1 form containing the L1 light
chain and the type 2 form containing the L2 light chain in
addition to the heavy chain ( Fig. 3C). The Asn96±Ser97
bond should be cleaved to produce the lower molecular
weight form, and therefore, a putative sperm endop eptidase
that cleaves the C-terminal side of the asparagine residue
may be responsible for this processing. It seems unlikely that
spermosin type 1 is an inactive precursor of spermosin type
2, as spermosin type 1 and 2 are capable of binding to a
soybean trypsin inhibitor-immobilized Sepharose column.
With respect to the disul®de bond between the light chain
and the heavy chain of ascidian spermosin, it is plausible
that the C ys116 residue in the light chain is disul®de-bonded
to the C ys251 residue in the heavy chain by analogy to other
trypsin family proteins: the nearest cysteine residue to the
active s ite aspartic acid residue is known to b e d isul®de-
bonded to the light chain in many serine proteases including
acrosin, thrombin, kallikrein, factor X and plasmin [17] (see
Figs 1, 3C and 5). Although light and h eavy chains of
mammalian acrosin are bonded by two disul®de bridges, a
single S±S bridge between light and heavy chains is rather
common in serine proteases including mouse testis-speci®c
proteases (TESP-I and -II) [20], ascidian acrosin [16], and
mammalian proteases such as thrombin, kallikrein, factor X
and plasmin [17].
Homology between ascidian spermosin and human
acrosin, and that between ascidian spermosin and ascidian

acrosin were 27% in both cases. In contrast with acrosin,
spermosin did not have a consensus sequence o f an N-linked
sugar attachment and paired basic residues in the
N-terminal region, t he latter of which is proposed to be
responsible for the binding of (pro)acrosin to the zona
pellucida [21]. In place of the paired basic residues,
spermosin contained a proline-rich region in the L1 light
chain. We demonstrated that the proline-rich region in the
L1 chain binds to the vitelline coat, as is the case for the
paired basic residues in the N termini and the proline-rich
regions in the C termini of mammalian p roacrosins. In
addition, we found that the 28-kDa vitelline coat protein is
capable of binding to the above proline-rich region.
We have reported previously that spermosin inhibitors,
Z-Val-Pro-Arg-H [ 10] and Dns-Val-Pro-Arg-H [12 ], and
anti-spermosin antibody [11] are capable of inhibiting
fertilization in a concentration-dependent manner indicat-
ing that spermosin plays an important extracellular role in
ascidian fertilization and that the proteolytic activity of
spermosin is required for ascidian fertilization. As a
proline-rich region (residues 28±88) of spermosin L1 light
chain is able to associate with the vitelline coat of the egg, it
is inferred that spermosin is involved not only in the sperm
penetration of the vitelline coat but also in th e sperm
binding to the vitelline coat. Whether s permosin or its
homologue is present in mammalian sperm is an intriguing
issue, as a sperm protease(s) other than acrosin i s
considered to play an essential role in the sperm penetra-
tion of the zona pellucida in mammals. In connection with
this, it should be noted that a 27-kDa p rotein in mouse

epididymis extract i s speci®cally recognized by anti-ascidian
spermosin antibody on the basis of Western blot analysis
(E. Kodama, H. Yokosawa & H. S awada, unpublished
data). Further studies are needed to search for a spermosin
homologue in mammalian sperm and to elucidate its role
in mammalian fertilization.
HrSpermosin 130 IVGGAEAVPNSWPYAAAFGTYDISGGKLEVSQMCGSTIITPRHALTAAHCFMMDPDIDQT
MmKallikrein 391 IVGGTNASLGEWPWQVSLQV KLVSQTHLCGGSIIGRQWVLTAAHCFDGIPY PD
MmAcrosin 40 IVSGQSAQLGAWPWMVSLQI FTSHNSRRYHACGGSLLNSHWVLTAAHCFDNKKK VY
HrAcrosin 36 IVGGEMAKLGEFPWQAAFLY KHVQVCGGTIIDTTWILSAAHCFDPHMYNLQS
**.* * . :*: .:: **.::: *:*****
HrSpermosin 190 YYIFMGLHDETT YKGVRP-NKIVGVRYHPKTNVFTDDPWLVYDFAILTLRKKVI
MmKallikrein 444 VWRIYGGILSLS EITKETPSSRIKELIIHQEYKVSEGN YDIALIKLQTPLN
MmAcrosin 96 DWRLVFGAQEIEYGRNKPVKEPQQERYVQKIVIHEKYNVVTEG NDIALLKITPPVT
HrAcrosin 88 IKKEDALIRVADLDKTDDTDEGEMTFEVKDIIIHEQYNRQTFD NDIMLIEILGSIT
: : * : : . *: :: : :
HrSpermosin 243 ANFAWNYACLP-QPKKIPPEGTICWSVGWGVTQNTGGDNV LKQVAIDLVSEKRCK-E
MmKallikrein 495 YTEFQKPICLP-SKADTNTIYTNCWVTGWGYTKEQGETQN ILQKATIPLVPNEEC-QK
MmAcrosin 152 CGNFIGPCCLPHFKAGPPQIPHTCYVTGWGYIKEKAPRPSP-VLMEARVDLIDLDLCNST
HrAcrosin 144 YGPTVQPACIP-GANDAVADGTKCLISGWGDTQDHVHNRWPDKLQKAQVEVFARAQC
*:* * *** :: * :. : :. *
HrSpermosin 298 EYRSTITSKSTICGGTTPG-QDTCQGDSGGPLFCKEDGK WYLQGIVSYGPSVCG-
MmKallikrein 551 KYRDYVINKQMICAGYKEGGTDACKGDSGGPLVCKHSGR WQLVGITSWG-EGCGR
MmAcrosin 211 QWYNGRVTSTNVCAGYPEGKIDTCQGDSGGPLMCRDNVDSP FVVVGITSWG-VGCAR
HrAcrosin 200 LATYPESTENMICAGLRTGGIDSCQGDSGGPLACPFTENTAQPTFFLQGIVSWG-RGCAL
:*.* * *:*:******* * : : **.*:* *.
HrSpermosin 351 SGPMAAYAAVAYNLEWLCCYMP NLPSCEDIECDESGEN
MmKallikrein 605 KDQPGVYTKVSEYMDWILEKTQ SSDVRALETSSA
MmAcrosin 267 AKRPGVYTATWDYLDWIASKIG PNALHLIQPATPHPPTTRHPMVSFHPPSLRPPW
HrAcrosin 259 DGFPGVYTEVRKYSSWIANYTQHLLQDRNADVATFTITGDPCSSNGSIISGSEGDFSSPG

*: . .*:
MmAcrosin 322 YFQHLPSRPLYLRPLRPLLHRPSSTQTSSSLMPLLSPPTPAQPASFTIATQHMRHRTTLS
HrAcrosin 319 FYSGSYTDNLDCKWIIQIPDIGSRIQLSFTEFGVEYHTFCWYDDVKVYSGAVGNIASADA
MmAcrosin 382 FARRLQRLIEALKMRTYPMKHPSQYSGPRNYHYRFSTFEPLSNKPSEPFLHS
HrAcrosin 379 ADLLGSHCGMNIPSDLLSDGSSMTVIFHSDYMTHTLGFRAVFHAVSADVSQSGCGGIREL
HrAcrosin 439 LTDHGEFSSKHYPNYYDADSNCEWLITAPTGKTIELNFLSFRLAGSDCADNVAIYDGLNS
HrAcrosin 499 SQLPKNN
Fig. 5. Sequence alignment of ascidian
spermosin heavy chain with those of trypsin-
family proteases including ascidian acrosin by
the
CLUSTALW
program. Identical residues in
the sequenc es amon g se rine protease are
indicated by asterisks. Three co nserved
active site residues in the S1 subfamily [17] of
trypsin-like pr otease are indicated by closed
triangles. The loc ations of paired basic resi-
dues in the N-terminal regions of mouse and
ascidian acrosin are indicated by c los ed
circles. Asterisks indicate the positions o f
fully conserved amino-acid residue s. Double
and single dots represent the ``strong'' (score
of Gonnet Pam250, > 0.5) and ``weak''
(£ 0.5) consensus positions, respectively. For
details of CLUSTAL W and Gonnet
Pam250, access the URLs o f http://
hypernig.nig.ac.jp/homology/clustalw.shtml
and ewsletter/
archives/2/clustalw17.html, respectively.

HrSpermosin, H. roretzi spermosin (DDBJ/
GenBank/EBI accession number,
AB052776); MmKallikrein, Mus musculus
plasma kallikrein (M58588); MmAcrosin,
Mus musculus acrosin (D00574); HrAcrosin,
H. roretzi acrosin (AB052635).
662 E. Kodama et al. (Eur. J. Biochem. 269) Ó FEBS 2002
ACKNOWLEDGEMENTS
This work was supported in p art by Grant-in-aids for Scienti®c
Research from the Ministry of Education, Science, Sports, and Culture
of Japan and the A kiyama Foundation. We are grateful t o C. C.
Lambert (California State University Fullerton) for his c ritical reading
of this manuscript and valuable advice.
REFERENCES
1. Hoshi, M., Takizawa, S. & Hirohashi, N. (1994) Glycosidases,
proteases and ascidian fertilization. Semin. De v. Bio l. 5, 201±208.
2. Mu
È
ller-Esterl, W. & Fritz, H. (1981) Sperm acrosin. Methods
Enzymol. 80, 621±632.
3. Urch, U.A. (1986) The action of acrosin on the zona pellucida.
Adv. Exp. Med. Biol. 207, 113±132.
4. Baba, T., Azuma, S., Kashiwabara, S. & Toyoda, Y. (1994) Sperm
from mice carrying a targeted mutation of the acrosin gene can
penetrate the oocyte zona pellucida and eect fertilization. J. Biol.
Chem. 269, 31845±31849.
5. Adham, L.M., Nayernia, K. & Engel, W. (1997) Spermatozoa
lacking acrosin protein show delayed fertilization. Mol. Reprod.
Dev. 46, 370±376.
6. Yamagata, K., Mura yama, M., Okabe, M., Toshimori, K.,

Nakanishi, T., Kashiwabara, S . & Baba, T. ( 1998) Ac rosin
accelerates the dispersal o f sperm acrosomal proteins during
acrosome reaction. J. Biol. Chem. 273, 10470±10474.
7. Swalla, B.J. (2001) Phylogeny of the Urochordates: Implications
for chordate evolutio n. In The Biology of Ascidians (Sawad a, H .,
Yokosawa, H. & Lambert, C.C., eds), pp. 219±224. Springer-
Verlag, Tokyo.
8. Fuke, T.M. (1983) Self and non-self recognition between gametes
of the a scidian, Halocynthia roretzi. Roux's Arch. Dev. Biol. 192 ,
347±352.
9. Sawada, H., Yokosawa, H. & Ishii, S. (1984) Puri®cation and
characterization of two types of trypsin-like enzymes from sperm
of the ascidian (Prochordata) Halocynthia roretzi. E viden ce for the
presence of spermosin, a novel acrosin-lik e enzy me. J. Biol. Chem.
259, 2900±2904.
10. Sawada, H., Yokosawa, H., Someno, T., Sain o, T. & Ishii, S.
(1984) Ev idence for the participation o f two sp erm proteases,
spermosin and acrosin, in fertilization of the ascidian, Halocynthia
roretzi: inhibitory eects of leupept in analogs on enzyme activitie s
and fertilization. Dev. Biol. 105, 246±249.
11. Sawada, H., Iwasaki, K., Kihara-Negishi, F., Ariga, H. &
Yokosawa, H. (1996) Localization, expression, and the role in
fertilization of s permosin, an ascidian sperm trypsin-like protease.
Biochem. Biophys. Res. Commun. 222, 499±504.
12. Sawada, H. & Someno, T. (1996) Substrate speci®city of ascidian
sperm trypsin-like proteases, spermosin and acrosin. Mol. Reprod.
Dev. 45, 240±243.
13. Hoshi, M., Numakunai, T. & Sawada, H. (1981) Evidence for
participation of sperm proteinases in fertilization of the solitary
ascidian, Halocynthia ro retzi: e ects of protease inhib itors. Dev.

Biol. 86, 117±121.
14. Sawada, H., Yokosawa, H., Hoshi, M. & Ishii, S. (1982) Evidence
for acrosin-like enzyme in sperm extract and its involvement in
fertilization of the ascidian, Halocynthia roretzi. Gamete Res. 5,
291±301.
15. Laemmli, U .K. (1970) Cleavage of stru ctural protein during
the assembly of the head o f bacteriophage T4. Nature 227,
680±685.
16. Kodama, E., Baba, T., Yokosawa, H. & Sawada, H. (2001) cDNA
cloning and functional analysis of ascidian sperm proacro sin.
J. Biol. Chem. 276, 24594±24600.
17. Barrett, A.J., Rawlings, N .D. & Woessner, J.F. (1998) Handbook
of Proteolytic Enzymes.AcademicPress,SanDiego,CA.
18. Urch, U.A. & Patel, H. (1991) The interaction of boar sperm
proacrosin with its natural substrate, the zo na pellucida, and with
polysulfated polysaccharides. De velopm ent 11 1 , 1165±1172.
19. Baba, T., Kashiwabara, S., Watanabe, K., Ito h, H., Michikawa,
Y., Kimura, K., Takada, M., Fukamizu, A. & Arai, Y. (1989)
Activation and maturation m echanisms of boar acrosin zymogen
based on t he deduced primary struc ture. J. Biol. Chem. 264,
11920±11927.
20. Kohno, N., Yamagata, K., Yam ada, S., Kashiwabara, S., Sakai,
Y. & Baba, T. (1998) Two novel testicular se rine proteases, TESP1
and TESP2, are present in the mouse sperm acrosome. Biochem.
Biophys. Res. Commun. 245, 658±665.
21. Richardson, R.T. & O'Rand, M.G. (1996) Site-directed muta-
genesis of rabbit proacrosin. Identi®cation of residues involved in
zona pellucida binding. J. Biol. Chem. 271, 24069±24074.
Ó FEBS 2002 Cloning and characterization of ascidian spermosin (Eur. J. Biochem. 269) 663