Báo cáo khoa học: Recombinant bovine zona pellucida glycoproteins ZP3 and ZP4 coexpressed in Sf9 cells form a sperm-binding active hetero-complex ppt

A mixture of bovine recombinant ZP3 rZP3 and rZP4 coexpressed in Sf9 cells exhibited inhibitory activity for bovine sperm–ZP binding similar to that of a native bovine ZPG mixture, where

Recombinant bovine zona pellucida glycoproteins ZP3 and ZP4 coexpressed in Sf9 cells form a sperm-binding active hetero-complex

Saeko Kanai1, Naoto Yonezawa1, Yuichiro Ishii1, Masaru Tanokura2 and Minoru Nakano1

1 Graduate School of Science and Technology, Chiba University, Japan

2 Graduate School of Agriculture and Life Science, The University of Tokyo, Japan

Mammalian oocytes are surrounded by the zona

pellu-cida (ZP), a transparent envelope that mediates several

critical aspects of fertilization, including species-selective

sperm recognition, blocking of polyspermy, and

protec-tion of the oocyte and embryo until implantaprotec-tion [1–3]

The ZP consists of three or four kinds of glycoproteins

(ZPGs) Human and rat ZPs consist of four ZPGs (ZP1,

ZP2, ZP3, and ZP4) [4,5], whereas porcine and bovine

ZPs comprise three ZPGs (ZP2, ZP3, and ZP4) that cor-respond to ZPA, ZPC, and ZPB, respectively, in other nomenclature [6] Murine ZP also consists of three ZPGs (ZP1, ZP2, and ZP3) [7] Porcine, bovine and murine ZPs have ZP2 and ZP3 in common, whereas ZP1 and ZP4 are products of distinct genes [8] All ZPGs contain a domain that consists of  260 amino acids and contains eight conserved Cys residues [9]

Keywords

baculovirus-Sf9; fertilization; glycoprotein;

zona pellucida; ZP domain

Correspondence

M Nakano, Graduate School of Science,

Chiba University, 1-33 Yayoi-cho, Inage-ku,

Chiba 263-8522, Japan

Fax: +81 43 290 2874

Tel: +81 43 290 2794

E-mail: mnakano@faculty.chiba-u.jp

(Received 18 April 2007, revised 27 July

2007, accepted 24 August 2007)

doi:10.1111/j.1742-4658.2007.06065.x

The zona pellucida (ZP) is a transparent envelope that surrounds the mam-malian oocyte and mediates species-selective sperm–egg interactions Por-cine and bovine ZPs are composed of the glycoproteins ZP2, ZP3, and ZP4 We previously established an expression system for porcine ZP glyco-proteins (ZPGs) using baculovirus in insect Sf9 cells Here we established a similar method for expression of bovine ZPGs The recombinant ZPGs were secreted into the medium and purified by metal-chelating column chromatography A mixture of bovine recombinant ZP3 (rZP3) and rZP4 coexpressed in Sf9 cells exhibited inhibitory activity for bovine sperm–ZP binding similar to that of a native bovine ZPG mixture, whereas neither bovine rZP3 nor rZP4 inhibited binding An immunoprecipitation assay revealed that the coexpressed rZP3⁄ rZP4 formed a hetero-complex We examined the functional domain structure of bovine rZP4 by constructing ZP4 mutants lacking the N-terminal domain or lacking both the N-termi-nal and trefoil domains When either of these mutant proteins was coexpressed with bovine rZP3, the resulting mixtures exhibited inhibitory activity comparable to that of the bovine rZP3⁄ rZP4 complex Hetero-com-plexes of bovine rZP3 and porcine rZP4, or porcine rZP3 and bovine rZP4, also inhibited bovine sperm–ZP binding Our results demonstrate that the N-terminal and trefoil domains of bovine rZP4 are dispensable for formation of the sperm-binding active bovine rZP3⁄ rZP4 complex and, furthermore, that the molecular interactions between rZP3 and rZP4 are conserved in the bovine and porcine systems

Abbreviations

ACA, Amaranthus candatus agglutinin; BO, Brackett and Oliphant; FITC, fuorescein isothiocyanate; Fuc, fucose; GNA, Galanthus nivalis agglutinin; LC, liquid chromatography; LCA, Lens culinaris agglutinin; Man, mannose; MOI, multiplicity of infection; PA, pyridylamino; PHA, Phaseolus vulgaris agglutinin; PSA, Pisum sativum agglutinin; RCA120, Ricinus communis agglutinin; rZP2, recombinant ZP2; rZP3,

recombinant ZP3; rZP3FLAG, FLAG-tagged rZP3; rZP4, recombinant ZP4; rZP4FLAG, FLAG-tagged rZP4; rZPG, recombinant ZPG; ZP, zona pellucida; ZPG, zona pellucida glycoprotein.

In mice, ZP3 is thought to be involved in gamete

recognition [1–3] ZP assembly is controlled by short,

hydrophobic sequences in the C-terminal propeptides

of ZPG precursors, and requires the ZP domains of

ZP2 and ZP3 [10,11] The molar ratio of murine

tran-scripts is estimated at ZP1⁄ ZP2 ⁄ ZP3 ¼ 1 : 4 : 4 [12], a

ratio that is consistent with a suggested model in

which a ZP2⁄ ZP3 heterodimer forms filaments that are

crosslinked by a ZP1 dimer [13] However, the molar

ratio of ZPGs in the murine ZP does not seem to

correspond to the molar ratio of their transcripts

[7] In pigs, the estimated protein molar ratio of

ZP2⁄ ZP3 ⁄ ZP4 is 1 : 6 : 6 [14] Although neither ZP3

nor ZP4 exhibits porcine sperm-binding activity by

itself, a high molecular mass ZP3⁄ ZP4 hetero-complex

does exhibit this activity [15,16]

When subjected to nonreducing SDS⁄ PAGE, bovine

ZPGs form a band at an average apparent molecular

mass of 74 kDa, which is broad owing to heterogeneity

in glycosylation [17] After

endo-b-galactosidase-cata-lyzed removal of N-acetyl-lactosamine repeats at the

nonreducing ends of carbohydrate chains, bovine ZP2,

ZP3 and ZP4 migrate as three distinct bands of

appar-ent molecular masses of 72, 45 and 58 kDa,

respec-tively, under nonreducing conditions [17] Under

reducing conditions, the apparent molecular masses of

the endo-b-galactosidase-digested components shift to

76, 63 and 21 kDa for ZP2, to 47 kDa for ZP3, and to

68 kDa for ZP4 [17] Processing of bovine ZP2 occurs

at a specific site upon fertilization, and yields

disulfide-bonded polypeptides of 63 and 21 kDa [17,18] A large

fraction of ZP2 obtained from unfertilized eggs is

already processed, probably as an artefact of the

prep-aration, but the 76 kDa band of ZP2 completely

dis-appears upon fertilization [17,18]

The amino acid sequences of porcine and bovine

ZP2, ZP3, and ZP4, which were previously determined

by cDNA cloning and sequencing [6,19–21], are 77%,

85% and 75% identical, respectively The mature

por-cine and bovine ZP4 polypeptides differ in that an

N-terminal region corresponding to residues 1–135 of

bovine ZP4 (with the translation initiation Met

num-bered 1) is lacking in the porcine protein [19,21,22]

(Fig 1A) The estimated protein molar ratio of bovine

ZP2⁄ ZP3 ⁄ ZP4 is 1 : 2 : 1 [21], which differs

signifi-cantly from the porcine molar ratio, suggesting that

the structures of the bovine and porcine ZPs are

differ-ent

In a previous study, we partially separated an

endo-b-galactosidase-digested bovine ZPG mixture into

three components by RP-HPLC [21] Of the three

components, ZP4 exhibited the strongest

sperm-bind-ing activity ZP2 and ZP3 exhibited much weaker

activity [21] The components were not completely resolved by HPLC, indicating cross-contamination; thus, whether each bovine ZPG has sperm-binding activity by itself is not yet clear A previous report that bovine sperm–egg binding is inhibited in the presence

of anti-porcine ZP3 or ZP4 suggests that both ZP3 and ZP4 are involved in sperm–ZP binding [23]

In mice, in vitro studies have proposed that sperm ligands consist of O-linked carbohydrate chains linked

to Ser332 and Ser334 of ZP3 [24,25] Nevertheless, a recent structural analysis using MS did not show evi-dence for glycosylation [26] The in vivo studies per-formed to date using transgenic mice lacking each glycosyltransferase gene do not support the involve-ment of carbohydrate chains of mouse ZP in sperm binding [27–29] In pigs, neutral tri-antennary and tetra-antennary complex-type chains have the strongest sperm-binding activity of the N-linked chains of ZP [30], and O-linked chains also have sperm-binding activity [31] The nonreducing terminal b-galactosyl

A

Fig 1 Recombinant bovine ZP proteins (A) Schematic representa-tion of the rZP2, rZP4, rZP4136)464, rZP4182)464and rZP3 polypep-tides These recombinant polypeptides were expressed with His-and S-tags at their N-termini Open square, region specific to ZP2, ZP4, or ZP3; dotted square, trefoil domain; filled square, ZP domain Arrows indicate the putative furin cleavage sites that con-stitute the C-termini of the expressed polypeptides The calculated molecular masses of the polypeptide moieties of the recombinants, excluding extra peptides derived from the transfer vector, are shown in kDa to the right of each polypeptide (B, C) SDS ⁄ PAGE and immunoblot analyses of rZP2 (lane 1), rZP4 (lane 2), rZP4136)464(lane 3), rZP4182)464(lane 4), and rZP3 (lane 5) The pro-teins were expressed in Sf9 cells, secreted into the culture med-ium, isolated using metal-chelation column chromatography, and detected by SDS ⁄ PAGE (B) or by immunoblot analysis using anti-bodies specific for each of the ZPGs (C) Arrowheads indicate the recombinant protein bands Molecular mass markers are indicated

in kDa on the left of each panel.

residues of the complex-type N-linked chains are

involved in sperm binding [32] In cows, the major

neutral N-linked chain of ZP consists of only one

structure, a high-mannose-type chain containing five

mannose residues [33] Thus, the structures of the

por-cine and bovine neutral chains are quite different

a-Mannosyl residues at nonreducing termini are

essen-tial for the sperm-binding activity of bovine ZP [34],

although the participation of O-linked chains in sperm

binding has not yet been investigated Recently, we

reported that porcine recombinant ZPGs (rZPGs)

expressed in insect Sf9 cells have pauci-mannose and

high-mannose-type chains and bind to bovine sperm

but not to porcine sperm [16] This result supports a

significant role for a-mannosyl residues in bovine

sperm recognition and also demonstrates the utility of

rZPG expression in Sf9 cells

In this study, we used the Sf9 expression system to

obtain each of the bovine rZPGs without the

possibil-ity of contamination by the other rZPGs and examined

the sperm-binding activity and complex formation of

these rZPGs We also created deletion mutants of

recombinant (r)ZP4 to examine whether its N-terminal

region and trefoil domain are necessary for sperm–ZP

binding activity

Results

Expression of bovine rZP2, rZP3, rZP4 and rZP4

mutants in Sf9 cells infected with recombinant

baculoviruses

Native ZPGs are synthesized as transmembrane

proteins, processed at a site N-terminal to their

trans-membrane regions, and then secreted as mature

poly-peptides without their transmembrane regions Here,

His- and S-tagged recombinant polypeptides

corre-sponding to bovine ZP2 (Ile36 to Arg637), ZP3 (Arg32

to Arg348) and ZP4 (Lys25 to Arg464) were expressed

in Sf9 cells (Fig 1A) The N-termini of these rZPGs

correspond to those previously reported for mature

native bovine ZPGs [17] We presume that the

N-ter-mini of the native ZP3 and ZP4 polypeptides are

blocked and that the N-termini reported previously

might have been a result of degradation [17] Thus, the

N-termini of rZP3 and rZP4 expressed here are likely

to closely correspond to the N-termini of their native

counterparts

The C-termini of the mature bovine ZP2, ZP3 and

ZP4 polypeptides have not yet been determined The

immature proteins have putative furin-processing sites

at Arg634 to Arg637, Arg345 to Arg348, and Arg461

to Arg464, respectively Recent studies have revealed

that porcine, murine and human ZPGs are processed

at consensus sites for furin or furin-like processing enzymes [35–38] In at least three murine ZPGs and in porcine ZP3 and ZP4, this processing is followed by removal of the basic amino acid residues in the consen-sus sites by a carboxypeptidase [26,35] We presume that bovine ZPGs are processed similarly

Two N-terminal deletion mutants of bovine rZP4 were also expressed in this study The rZP4136)464 mutant lacks residues Lys25 to Pro135 and consists of the trefoil and ZP domains of rZP4 The rZP4182)464 mutant lacks residues Lys25 to Tyr181 and thus con-sists only of the ZP domain (Fig 1A)

The apparent molecular masses of the recombinant proteins, as determined by SDS⁄ PAGE, agreed with the molecular masses predicted from their encoded amino acid sequences, and immunoblots with specific antibodies to ZPG confirmed the presence of the pro-teins (Fig 1B,C) The absorbance at 280 nm of the eluted fractions was used to estimate the yield of the recombinant proteins; about 15 lg of each rZPG was obtained from 200 mL of culture medium

Sperm-binding activity of bovine rZPGs

We examined the inhibitory activity of the bovine rZPGs towards binding of bovine sperm to plastic wells coated with solubilized bovine ZP (Method 1; Fig 2) In the presence of 2 lgÆmL)1 of solubilized bovine ZP, sperm binding to solubilized ZP-coated wells was reduced to its plateau level, which was about 10% of the level observed in the absence of solubilized

ZP In contrast, none of the bovine rZPGs significantly inhibited binding

Sf9 cells were coinfected with the appropriate re-combinant viruses to form rZP3⁄ rZP4, rZP2 ⁄ rZP4, rZP2⁄ rZP3 and rZP2 ⁄ rZP3 ⁄ rZP4 mixtures Expression

of the mixtures was confirmed by SDS⁄ PAGE (Fig 3A) and immunoblot analysis (data not shown) Bovine sperm binding to solubilized bovine ZP-coated wells was not significantly inhibited by rZP2, rZP3, or rZP4 (Fig 3B; see also Fig 2), but it was inhibited by the rZP3⁄ rZP4 mixture The mixture reduced binding to a level similar to that observed with solubilized bovine ZP (Fig 3B) The rZP2⁄ rZP4 and rZP2 ⁄ rZP3 mixtures did not significantly inhibit binding (Fig 3B) When rZP3 and rZP4 were expressed separately in Sf9 cells and then mixed, the mixture did not inhibit binding (Fig 3B)

To assess the effect of rZP2 on the inhibitory activ-ity of the rZP3⁄ rZP4 mixture, we compared the inhibi-tory activity of the rZP3⁄ rZP4 mixture to that of the rZP2⁄ rZP3 ⁄ rZP4 mixture The total amount of rZP3 and rZP4 in the mixtures was the same and was equal

to 0.2 or 0.4 lg (Fig 3C) The inhibitory activity of

rZP2⁄ rZP3 ⁄ rZP4 was not significantly different from

that of rZP3⁄ rZP4

In a previous study, we examined the inhibitory activity of each bovine ZPG for the binding of sperm

to ZP-encased eggs using an in vitro competition assay (Method 2 [21]) Recently, we established a competi-tion assay using solubilized ZP-coated plastic wells (Method 1 [16]) In Method 1, washing to remove sperm loosely attached to ZP does not require mouth pipetting; therefore, Method 1 is technically much eas-ier and more reproducible than Method 2 The inhibi-tory activity of a larger number of ZPGs can be examined at one time in Method 1 than in Method 2 However, Method 2 is an accepted assay system that has been used to evaluate the inhibitory activity of materials for sperm–ZP binding in many species, including mouse, cow, and pig Thus, we determined whether Method 2 yields parallel results to Method 1

In Method 2, bovine sperm binding to bovine eggs was not inhibited by rZP3 or rZP4, whereas binding was reduced by the rZP3⁄ rZP4 mixture to a level simi-lar to that observed with solubilized native bovine ZP (Fig 4) Thus, the two competition assay systems gave similar results

We examined whether the incubation of bovine sperm with solubilized bovine ZP or rZP3⁄ rZP4 induced the acrosome reaction of the sperm by using fluorescein isothiocyanate (FITC)-conjugated Pisum sativum agglu-tinin (PSA) (FITC-PSA) This lectin binds to the acro-somal area of acrosome-intact, acrosome-damaged and

Fig 3 Inhibitory effects of various bovine rZPG mixtures on bovine sperm-solubilized ZP binding (A) rZP2 ⁄ rZP4 (lane 1), rZP2 ⁄ rZP3 (lane 2), rZP3 ⁄ rZP4 (lane 3), rZP3 ⁄ rZP4 136 )464(lane 4), rZP3⁄ rZP4 182 )464(lane 5) and rZP2⁄ rZP3 ⁄ rZP4 (lane 6) mixtures were expressed by simulta-neous infection of Sf9 cells with the two or three corresponding recombinant viruses The rZPGs were collected from the culture superna-tant using metal-chelation column chromatography and detected by SDS ⁄ PAGE with silver staining Arrowheads indicate the recombinant protein bands Molecular mass markers are indicated in kDa (B) Bovine sperm were incubated with 0.2 lg of solubilized native ZP, 0.4 lg

of each rZPG, 0.27 lg of each bi-component rZPG coexpressed mixture, or a mixture of 0.4 lg of rZP3 and 0.4 lg of rZP4 that were sepa-rately expressed, purified and mixed (rZP3 + rZP4) for 30 min, and the inhibitory effect of the proteins was determined by Method 1 as described in the legend to Fig 2 The number of sperm binding to the ZP in the absence of inhibitors is designated 100% Assays were per-formed at least three times, and the data shown represent means ± SD (C) Bovine sperm were incubated for 30 min with a coexpressed mixture of rZP3 and rZP4 or a coexpressed mixture of rZP2, rZP3, and rZP4 The total amount of rZP3 and rZP4 was 0.2 or 0.4 lg, and the inhibitory effect of the rZPG mixtures was determined by Method 1 as described in the legend to Fig 2.

Fig 2 Inhibitory effects of rZP2, rZP3, rZP4 and solubilized bovine

ZP on bovine sperm-solubilized ZP binding Solubilized native bovine

ZP was adsorbed to each well of a 96-well plate (0.2 lg per well;

Method 1) Bovine sperm (4 · 10 5 ) were incubated with 0.2, 0.4 or

0.6 lg of solubilized ZP (·), rZP2 (r), rZP3 (m), or rZP4 (j) for

30 min, and then transferred to the coated wells After incubation

for 2 h, the wells were washed and 50 lL of glycerol ⁄ NaCl ⁄ P i was

added to each well The sperm that bound to the ZP were

recov-ered from the wells by vigorous pipetting, and the number of

sperm in 0.1 lL of the suspension was determined The number of

sperm binding to the ZP in the absence of inhibitors is designated

100% Assays were repeated at least three times, and the data

shown represent means ± SD.

partially acrosome-reacted bovine sperm but not to

acrosome-reacted bovine sperm [39] We performed this

experiment four times, and in each experiment, 100

sperm were observed for each incubation condition The percentages of sperm positively stained with FITC-PSA were 97.8 ± 0.9% for the sperm before incubation with the zona proteins, 93.8 ± 2.2% for the sperm after 3 h

of incubation in the absence of the zona proteins, 94.2 ± 3.6% for the sperm after 3 h of incubation with solubilized bovine ZP, and 92.8 ± 1.3% for the sperm after 3 h of incubation with rZP3⁄ rZP4 This indicates that the percentages of acrosome-reacted sperm, which were not stained with FITC-PSA, increased significantly but only slightly after 3 h of incubation in the absence and also in the presence of zona proteins, and therefore neither solubilized bovine ZP nor rZP3⁄ rZP4 induced the acrosome reaction of bovine sperm under the experi-mental conditions used in this study Neither solubilized bovine ZP nor rZP3⁄ rZP4 affected sperm motility

as compared to the sperm incubated without the zona proteins (data not shown)

The binding of sperm to rZPGs and to solubi-lized ZP was compared by indirect immunofluores-cence detection of rZPG-bound sperm Solubilized, native bovine ZP and the rZP3⁄ rZP4 mixture bound

to the acrosomal region of bovine sperm, as shown

by fluorescent staining, but rZP2, rZP3 and rZP4 did not bind to sperm (Fig 5) These results suggest that the inhibition of sperm–ZP binding by the rZP3⁄ rZP4 mixture is due to specific binding of rZP3⁄ rZP4 to the acrosomal area of sperm, but not due to

Fig 4 Inhibitory effects of various bovine rZPGs on bovine sperm–

egg binding Bovine sperm were incubated with 0.7 lg of solubilized

native ZP, rZP3, rZP4 or rZP3 ⁄ rZP4 mixture for 30 min and then

incubated with bovine eggs The inhibitory effects of the proteins

were determined by Method 2 The number of sperm binding to

eggs in the absence of inhibitors is designated 100% Assays were

performed six times, and the data shown represent means ± SD.

Fig 5 Indirect immunofluorescence staining of sperm-bound bovine rZPGs Suspensions of bovine sperm (50 lL at 2 · 10 6 mL)1) were incubated with 0.2 lg of rZP2, rZP3, rZP4, rZP3 ⁄ rZP4, rZP3 ⁄ rZP4136)464, rZP3⁄ rZP4182)464or solubilized native ZP for 30 min The proteins that bound to sperm were detected using a mixture of anti-porcine ZP2, ZP3, and ZP4 as the primary antibodies, and Alexa Fluor 546-conju-gated goat anti-(rabbit IgG) as the secondary antibody The sperm were observed using fluorescence microscopy As a control, the sperm were incubated without solubilized native ZP or rZPGs and then treated with the antibodies Insets, magnified fluorescence images of the sperm head Phase, phase-contrast image; fluorescence, fluorescence image.

induction of the acrosome reaction of sperm by

rZP3⁄ rZP4

Effect of N-terminal deletions of rZP4 on the

sperm-binding activity of rZP3⁄ rZP4

Neither rZP4136)464nor rZP4182)464 significantly

inhib-ited bovine sperm-solubilized ZP binding (data not

shown) Mixtures of rZP3 with each of these

N-termi-nal deletion mutants were prepared by coinfection of

Sf9 cells with the corresponding baculoviruses, and

protein expression was confirmed by SDS⁄ PAGE

(Fig 3A) The rZP3⁄ rZP4136)464 mixture exhibited

inhibitory activity similar to that of solubilized native

ZP and rZP3⁄ rZP4 (Fig 3B), indicating that residues

25–135 of rZP4 are not necessary for the

sperm-bind-ing activity of rZP3⁄ rZP4 The rZP3 ⁄ rZP4182)464

mixture was slightly less inhibitory than the

rZP3⁄ rZP4136)464mixture Although statistically

signif-icant, this difference was very small, indicating that the

trefoil domain of rZP4 is not essential for the

sperm-binding activity of rZP3⁄ rZP4

The rZP3⁄ rZP4136)464 and rZP3⁄ rZP4182)464

mix-tures exhibited significant binding to the acrosomal

region, as shown by fluorescent staining (Fig 5), in a

manner similar to the rZP3⁄ rZP4 mixture, suggesting

that the inhibition of sperm-solubilized ZP binding by

the mixtures is due to specific binding of the mixtures

to the acrosomal area of sperm

Complex formation of FLAG-tagged rZP3

(rZP3FLAG) with rZP4

To examine whether rZP3 associates with rZP4, we

prepared rZP3 whose N-terminal His-tag was changed

to FLAG-tag (rZP3FLAG) and investigated whether

rZP4 (without FLAG-tag) was coimmunoprecipitated

with rZP3FLAG using anti-FLAG M2 gels rZP3FLAG

expressed alone in Sf9 cells was precipitated with

anti-FLAG gels and detected by antibody to anti-FLAG

(Fig 6A, lane 6 in the right panel) but not by antibody

to His (Fig 6A, lane 6 in the left panel) The bands

indicated by closed circles in Fig 6 were detected in

the culture supernatants both in the absence and in the

presence of baculovirus infection, and therefore were

unrelated to rZPGs rZP4 expressed alone was not

pre-cipitated by the anti-FLAG gels, as rZP4 was not

detected by antibody to His in the pellet (Fig 6A,

lane 2 in the left panel), although the rZP4 was

precip-itated using S-protein agarose from the supernatant of

the immunoprecipitation from the anti-FLAG gels

(Fig 6A, lane 3 in the left panel) When the

coex-pressed rZP3⁄ rZP4 mixture was subjected to the

immunoprecipitation, neither rZP3 nor rZP4 was pre-cipitated by the anti-FLAG gels (Fig 6A, lane 4 in the left panel), but they were precipitated using S-protein agarose from the supernatant of the immunoprecipita-tion with anti-FLAG gels (Fig 6A, lane 5 in the left panel) Antibody to FLAG detected rZP3FLAG (Fig 6A, lanes 6 and 7 in the right panel) but not rZP3 or rZP4 (Fig 6A, lanes 3 and 5 in the right panel) When the rZP3FLAG⁄ rZP4 mixture coexpressed

in Sf9 cells was subjected to immunoprecipitation, rZP4 and rZP3FLAG were coprecipitated and detected

by immunoblots with antibody to His (Fig 6A, lane 7

in the left panel) and antibody to FLAG (Fig 6A, lane 7 in the right panel), respectively These results indicate that there was no nonspecific binding of rZP4

or rZP3⁄ rZP4 mixture to the anti-FLAG gels and that rZP4 was pulled down by the anti-FLAG gels through the FLAG-tag of rZP3FLAG Thus, we found that the immunoprecipitation assay using FLAG-tag is useful for examining complex formation between rZPGs When rZP3FLAG and rZP4 were expressed separately

in Sf9 cells and the culture supernatants were mixed, incubated overnight, and subjected to immunoprecipi-tation using anti-FLAG gels, rZP3FLAG was pulled down, as revealed by the detection with antibody to FLAG (Fig 6B, lane 4 in the right panel), but rZP4 was not coprecipitated with rZP3FLAG (Fig 6B, lane 4

in the left panel) This result indicates that the sepa-rately expressed rZP3 and rZP4 did not form a com-plex

When the rZP3FLAG⁄ rZP4182)464 mixture coex-pressed in Sf9 cells was subjected to immunoprecipi-tation, rZP4182)464 and rZP3FLAG were coprecipitated

by anti-FLAG gels and detected by antibody to His (Fig 6C, lane 4 in the left panel) and antibody to FLAG (Fig 6C, lane 4 in the right panel), respec-tively When the coexpressed rZP3⁄ rZP4182)464 mix-ture was subjected to immunoprecipitation, neither rZP3 nor rZP4182)464 was detected in the pellet (Fig 6C, lane 1 in the left panel) but both were pulled down by S-protein agarose from the superna-tant of the immunoprecipitation (Fig 6C, lane 2 in the left panel), indicating that rZP3FLAG and rZP4182)464 formed a complex and that the complex was pulled down through the FLAG-tag of rZP3FLAG

These results of the immunoprecipitation assay indi-cate that complex formation between rZP3 and rZP4

is correlated with the inhibitory activity of the rZP3⁄ rZP4 mixture for sperm–ZP binding In addition, these results indicate that the N-terminal and trefoil domains of rZP4 are dispensable for complex forma-tion of rZP4 with rZP3

Glycosylation of rZPGs

The carbohydrate moieties of the rZPGs were analyzed

by digestion with glycopeptidase F The mobility of

rZP3 on SDS⁄ PAGE increased as digestion progressed

(Fig 7A), and three bands with higher mobilities

appeared, indicating that rZP3 has three N-linked

chains Although the mobilities of rZP2 and rZP4 also

increased after digestion with glycopeptidase F,

indi-cating that rZP2 and rZP4 have N-linked chains

(Fig 7A), the resulting bands were not sufficiently

resolved to deduce the number of N-linked chains in

these proteins Native bovine ZP2 has three N-linked

chains [40], but the numbers of N-linked chains in

native bovine ZP3 and ZP4 have not been reported

Therefore, whether the N-linked glycosylation

charac-teristics of the recombinant proteins are similar to

those of their native counterparts cannot be

deter-mined at present

We examined the carbohydrate structures of

rZP4136)464by liquid chromatography (LC)⁄ MS

analy-sis of its pyridylaminated chains This protein was

cho-sen for MS analysis because its yield was the highest

among the bovine rZPGs described here Only one

major peak was observed by LC, and was assigned

as Man3-GlcNAc-(Fuc-)GlcNAc-pyridylamino (PA)

(Man, mannose; Fuc, fucose) from m⁄ z ¼ 1135.5

([M + H]+) [41–43] Two minor peaks were also observed by LC, and were assigned as Man2 -GlcNAc-(Fuc-)GlcNAc-PA and Man3-GlcNAc-GlcNAc-PA from m⁄ z ¼ 973.3 ([M + H]+) and 989.4 ([M + H]+), respectively [41–43] The calculated m⁄ z ([M + H]+) values of these structures were 1135.4, 973.4, and 989.4, respectively

We also compared the carbohydrate structures of the recombinant and native ZPGs using five different lectins The two ZP4 deletion mutants and all three rZPGs were recognized by Galanthus nivalis agglutinin (GNA) and Lens culinaris agglutinin (LCA) (Fig 7B), but not by Ricinus communis agglutinin (RCA120), Phaseolus vulgaris agglutinin (PHA-L4), or Amaranthus

A

B

C

Fig 6 Complex formation between rZP3 FLAG and rZP4 (A) Immu-noprecipitation of the coexpressed mixture of rZP3FLAG⁄ rZP4 Cul-ture supernatants without rZPGs (lane 1 in each panel), containing rZP4 expressed alone (lanes 2 and 3 in each panel), containing coexpressed rZP3 ⁄ rZP4 mixture (lanes 4 and 5 in each panel), con-taining rZP3 FLAG expressed alone (lane 6 in each panel), or contain-ing coexpressed rZP3 FLAG ⁄ rZP4 mixture (lane 7 in each panel), as indicated above each panel, were subjected to anti-FLAG immuno-precipitation The rZPGs pulled down by the anti-FLAG gels (F) were detected by immunoblotting with antibody to His (left panel) and with antibody to FLAG (right panel) The rZP3 and rZP4 remain-ing in the supernatant after the immunoprecipitation were sub-jected to pull-down by S-protein agarose (S) to examine the expression of the rZPGs (B) Immunoprecipitation of rZP3 FLAG ⁄ rZP4 mixture individually expressed and then combined Culture superna-tants containing rZP4 expressed alone (lanes 1 and 2 in each panel), rZP3 FLAG expressed alone (lane 3 in each panel), or a mix-ture of rZP3FLAGand rZP4 individually expressed, mixed, and incu-bated overnight (lane 4 in each panel), as indicated above each panel, were subjected to anti-FLAG immunoprecipitation The rZPGs pulled down by anti-FLAG gels (F) were detected by immu-noblotting with antibody to His (left panel) and with anti-FLAG M2 (right panel) rZP4 remaining in the supernatant after the immuno-precipitation was subjected to pull-down by S-protein agarose (S)

to examine the expression of rZP4 (C) Immunoprecipitation of rZP3 FLAG ⁄ rZP4182)464 mixture coexpressed in Sf9 cells Culture supernatants containing coexpressed rZP3 ⁄ rZP4182)464(lanes 1 and

2 in each panel), rZP3FLAGexpressed alone (lane 3 in each panel),

or coexpressed rZP3 FLAG ⁄ rZP4182)464(lane 4 in each panel), as indi-cated above each panel, were subjected to anti-FLAG immunopre-cipitation The rZPGs pulled down by anti-FLAG gels (F) were detected by immunoblotting with antibody to His (left panel) and with anti-FLAG M2 (right panel) rZP3 and rZP4182)464remaining in the supernatant after the immunoprecipitation were subjected to pull-down by S-protein agarose (S) to examine the expression of the rZPGs (lane 2 in each panel) The rZP3 and rZP3 FLAG bands are indicated by arrowheads in (A), (B), and (C) The rZP4 band is indi-cated by an arrow in (A) and (B) The ZP4 182 )464band is indicated

by an asterisk in (C) Bands detected by the antibodies but unre-lated to rZPGs are indicated by closed circles in (A), (B), and (C) Molecular mass markers are indicated in kDa on the left of each panel in (A), (B), and (C) IB, immunoblot.

candatus agglutinin (ACA) (data not shown) In

con-trast, all tested lectins recognized native bovine ZP2,

ZP3, and ZP4 This latter result is consistent with the

known structures of the bovine ZP [33]; a native

bovine ZPG mixture has a high-mannose-type chain

and acidic di-antennary, tri-antennary, and

tetra-anten-nary complex-type chains The lectin staining results

for rZP4136)464 are consistent with the above MS

assignments N-linked chains of similar structure to

those of rZP4136)464; i.e pauci-mannose-type chains

with or without fucose, may be abundant in rZPGs,

and these chains were recognized by GNA and LCA

Since rZPGs were not recognized by RCA or PHA-L4,

complex-type chains may not be abundant in rZPGs

The lectin-staining results for rZPGs and the MS

results for rZP4136)464 are consistent with the major

structures of N-linked chains found in recombinant

glycoproteins expressed in Sf9 cells, i.e

pauci-man-nose-type chains with or without fucose residues linked

to the innermost GlcNAc residue [41–43]

Sperm-binding activity of interspecific mixtures

of porcine and bovine rZP3 and rZP4

Recently, we reported that a porcine rZP3⁄ rZP4

mix-ture coexpressed in Sf9 cells binds bovine, but not

porcine, sperm, owing to the presence of

pauci-mannose-type and high-pauci-mannose-type chains on

por-cine rZP3⁄ rZP4 [16] In this study, we obtained

inter-specific rZP3⁄ rZP4 mixtures by coinfection of Sf9 cells

with baculoviruses encoding either bovine ZP3 and

porcine ZP4, or porcine ZP3 and bovine ZP4 We examined these mixtures for inhibitory activity towards bovine sperm-solubilized ZP binding after confirming expression by immunoblotting (Fig 8A) Both of the interspecific rZP3⁄ rZP4 mixtures inhibited binding to

an extent similar to that observed for the bovine rZP3⁄ rZP4 mixture (Fig 8B) None of the interspecific rZP3⁄ rZP4 mixtures coexpressed in Sf9 cells was immunoprecipitated by anti-FLAG gels (Fig 8C,D, lane 1 in the left panels), whereas both interspecific rZP3⁄ rZP4 mixtures were precipitated by S-protein agarose from the supernatants of the immunoprecipita-tion assays (Fig 8C,D, lane 2 in the left panels) When bovine rZP4 whose N-terminal His-tag was changed to FLAG-tag (rZP4FLAG) and porcine rZP3 were coex-pressed and subjected to the immunoprecipitation using anti-FLAG gels, porcine rZP3 and bovine rZP4FLAG were coprecipitated and detected by body to His (Fig 8C, lane 3 in the left panel) and anti-body to FLAG (Fig 8C, lane 3 in the right panel), respectively When bovine rZP3FLAG and porcine rZP4 were coexpressed and subjected to immunoprecipita-tion, bovine rZP3FLAG and porcine rZP4 were copre-cipitated and detected by antibody to FLAG (Fig 8D, lane 3 in the right panel) and antibody to His (Fig 8D, lane 3 in the left panel), respectively These results indicate that porcine rZP3⁄ bovine rZP4FLAG and bovine rZP3FLAG⁄ porcine rZP4 complexes were formed and immunoprecipitated through FLAG-tag

In the interspecific rZP3⁄ rZP4 mixtures, complex for-mation was parallel to sperm-binding activity

Fig 7 N-glycans of bovine rZPGs (A) The rZP2, rZP3 and rZP4 proteins were digested with glycopeptidase F for 0 min or 24 h (for rZP2 and rZP4), or for 0, 1 or 5 min or 24 h (for rZP3), and the mobility shifts of the rZPGs on SDS⁄ PAGE (8% separating gel) were examined After 1 min of digestion, the rZP3 sample yielded three bands (indicated by bars) of higher mobility than undigested rZP3 (0 min), indicating that rZP3 contains three N-linked chains After 24 h of digestion, rZP2 and rZP4 also migrated faster than undigested rZP2 and rZP4 (0 min), indicating that rZP2 and rZP4 contain N-linked chain(s) as well The bands were not sufficiently resolved, however, to allow determination of the number of N-linked chains Molecular mass markers are indicated in kDa on the left of each panel (B) GNA and LCA recognized the endo-b-galactosidase-digested native bovine ZPGs (lane 1 in each panel), as expected from the reported structures of the major N-linked chains [33] rZP2 (lane 2), rZP4 (lane 3), rZP4136)464(lane 4), rZP4182)464(lane 5) and rZP3 (lane 6) were also recognized by GNA and LCA Molecular mass markers are indicated in kDa on the left of each panel.

We previously reported that native bovine ZP3 and

ZP4 partially purified by RP-HPLC each has

sperm-binding activity, although the activity of ZP3 is much

weaker [21] Native ZP2 also has weak sperm-binding

activity, but whether this activity is significant is

unknown We also reported that a mixture of native

ZP3 and native ZP4 proteins has sperm-binding

activ-ity that is slightly stronger than that of ZP4 alone,

sug-gesting that ZP3 promotes binding of ZP4 to sperm

[21]

In this study, we found that none of the bovine

rZPGs bound to sperm when assayed alone, as

revealed by two kinds of in vitro competitive inhibition

assays and indirect immunofluorescence staining Of

the three possible dual combinations of the three

rZPGs, only the rZP3⁄ rZP4 mixture bound to sperm rZP3 and rZP4 coexpressed in Sf9 cells formed a het-ero-complex When rZP3 and rZP4 were expressed separately in Sf9 cells and then mixed, the mixture did not inhibit sperm–ZP binding, and an interaction between rZP3 and rZP4 was not detected As complex formation between rZP3 and rZP4 was parallel to the sperm-binding activity of the rZP3⁄ rZP4 mixture, sperm binding to the bovine ZP in vitro is mediated by

a hetero-complex of rZP3 and rZP4 This conclusion obtained using the rZPGs further suggests that the pre-viously reported sperm-binding activity of partially purified native ZP4 [21] was due to contamination with ZP3 The weak sperm-binding activities that we reported for native ZP2 and ZP3 [21] may be also ascribed to contamination with both ZP3 and ZP4

or with ZP4, respectively In pigs, native ZP4

D

Fig 8 Inhibitory effect of heterospecific porcine ⁄ bovine rZP3 ⁄ rZP4 mixtures on bovine sperm-solubilized ZP binding (A) Mixtures of porcine rZP3 and bovine rZP4 (rpZP3 ⁄ rbZP4, lane 1 in each panel) or of bovine rZP3 and porcine rZP4 (rbZP3 ⁄ rpZP4, lane 2 in each panel) were expressed by simultaneous infection of Sf9 cells with the two corresponding recombinant viruses The rZPGs were collected from the cul-ture supernatant using metal-chelation column chromatography and detected by SDS ⁄ PAGE (left panel) and immunoblotting (right panel) using a mixture of antibodies specific for each ZPG Arrowheads indicate the rZPG bands Molecular mass markers are indicated in kDa on the left of each panel (B) Bovine sperm were incubated with 0.4 lg of the rbZP3 ⁄ rbZP4, rpZP3 ⁄ rbZP4 or rbZP3 ⁄ rpZP4 mixtures for 30 min The assay (Method 1) was performed as described in the legend to Fig 2 The number of sperm binding to the solubilized ZP in the absence

of inhibitors was designated 100% (without inhibitors) Assays were performed at least three times, and the data shown represent means ± SD (C) Immunoprecipitation of rpZP3 ⁄ rbZP4 FLAG mixture coexpressed in Sf9 cells Culture supernatants containing coexpressed rpZP3 ⁄ rbZP4 (lanes 1 and 2 in each panel) or coexpressed rpZP3 ⁄ rbZP4 FLAG

(lane 3 in each panel), as indicated above each panel, were sub-jected to anti-FLAG immunoprecipitation rZPGs pulled down by anti-FLAG gels (F) were detected by immunoblotting with antibody to His (left panel) and with anti-FLAG M2 (right panel) The rpZP3 and rbZP4 remaining in the supernatant after the immunoprecipitation were sub-jected to pull-down by S-protein agarose (S) to examine the expression of the rZPGs (lane 2 in each panel) (D) Immunoprecipitation of rbZP3 FLAG ⁄ rpZP4 mixture coexpressed in Sf9 cells Culture supernatants containing coexpressed rbZP3 ⁄ rpZP4 (lanes 1 and 2 in each panel)

or coexpressed rbZP3 FLAG ⁄ rpZP4 (lane 3 in each panel), as indicated above each panel, were subjected to anti-FLAG immunoprecipitation The rZPGs pulled down by anti-FLAG gels (F) were detected by immunoblotting with antibody to His (left panel) and with anti-FLAG M2 (right panel) The rbZP3 and rpZP4 remaining in the supernatant after the immunoprecipitation were subjected to pull-down by S-protein aga-rose (S) to examine the expression of the rZPGs (lane 2 in each panel) The rpZP3, rbZP3 FLAG and rbZP3 bands are indicated by arrowheads

in (C) and (D) The rbZP4, rbZP4 FLAG and rpZP4 bands are indicated by arrows in (C) and (D) The bands detected by the antibodies but unre-lated to rZPGs are indicated by closed circles in (C) and (D) Molecular mass markers are indicated in kDa on the left of each panel in (C) and (D) IB, immunoblot.

uncontaminated with ZP3 exhibits no sperm-binding

activity, and only the ZP3⁄ ZP4 hetero-complex has

sperm-binding activity [15] Recently, we reported a

parallel result for porcine rZPGs; neither rZP3 nor

rZP4 has physiologically significant sperm-binding

activity, but rZP3⁄ rZP4 coexpressed in Sf9 cells does

have activity [16] Thus, in both the porcine and

bovine systems, sperm binding to the ZP is mediated

by a ZP3⁄ ZP4 hetero-complex Furthermore, all three

ZPGs are shared in the porcine and bovine systems

The molecular mechanisms by which sperm interact

with the ZP appear to be similar for pigs and cows

Neither solubilized bovine ZP nor rZP3⁄ rZP4

signif-icantly induced the acrosome reaction of bovine sperm

in this study However, this does not mean that

solubi-lized bovine ZP does not have acrosome

reaction-inducing activity Previous reports have shown that

30–35% of bovine sperm complete the acrosome

reac-tion after incubareac-tion with 50 ngÆlL)1 of solubilized

bovine ZP as compared to about 10% after incubation

with unrelated glycoproteins [44,45] The induction of

the acrosome reaction is only 3–4% in those reports at

9 ngÆlL)1 of solubilized bovine ZP, however, which is

the concentration examined in the present study As

the concentrations of the zona proteins examined

in the competitive inhibition assays in the present

study were lower than 9 ngÆlL)1, it could be concluded

that the acrosome reaction of bovine sperm was not

significantly induced under the experimental conditions

used in this study Because in mice a recent report

sug-gested that an intact porous structure of ZP is

neces-sary for mechanical induction of the acrosome reaction

of mouse sperm [46], it remains to be clarified whether

solubilization of bovine ZP reduces its acrosome

reac-tion-inducing activity for sperm According to previous

reports, 4 h of incubation is necessary for complete

capacitation of bovine sperm [44,45] Then, it is also

possible that the bovine sperm used in this study were

not completely capacitated after 30 min of incubation,

and therefore the acrosome reaction was not induced

significantly by incubation with the zona proteins

Native bovine, porcine and murine ZP2 are

pro-cessed at a specific site by an unidentified enzyme upon

fertilization [17,47,48] This processing plays a role in

blocking polyspermy by the ZP [49] Specific

proteo-lysis of bovine ZP2, together with formation of

intra-molecular and interintra-molecular disulfide linkages, is

involved in ZP hardening [18], but the role of ZP2 in

sperm binding is not yet clear Because, here, a bovine

rZP2⁄ rZP3 ⁄ rZP4 mixture coexpressed in Sf9 cells

inhibited bovine sperm–ZP binding at a level similar to

that of rZP3⁄ rZP4, we conclude that rZP2 does not

affect the sperm-binding activity of rZP3⁄ rZP4

Neither rZP2⁄ rZP4 nor rZP2 ⁄ rZP3 coexpressed in Sf9 cells exhibited sperm-binding activity Thus, we found

no evidence for involvement of ZP2 in sperm–ZP bind-ing In mice, a ZP consisting of mouse ZP1, human ZP2 and mouse ZP3 was made using transgenic mice [49] Human ZP2 in the chimeric ZP remained unc-leaved after fertilization, and mouse sperm continued

to bind to the ZP On the basis of these observations,

a model was proposed in which mouse sperm recognize the supramolecular structure of the ZP but not the car-bohydrate structure of the ZP [3,49] Additionally, sperm cannot recognize the supramolecular structure modulated by ZP2 processing Considering this model,

it remains to be clarified whether processed ZP2 inhib-its the sperm-binding activity of the ZP3⁄ ZP4 complex

in cows

The mature bovine ZP4 polypeptide consists of a unique N-terminal region, a trefoil domain, and a ZP domain Although porcine and bovine ZP4 are homol-ogous, the mature porcine ZP4 polypeptide lacks the N-terminal region found in the bovine protein [21,22] The trefoil domain was first discovered in proteolysis-resistant trefoil factor peptides that play roles in muco-sal defense and healing [50] As trefoil factor peptides are expressed in association with mucins, they are likely to interact with mucins through carbohydrate or polypeptide moieties [50] The roles of the N-terminal region and trefoil and ZP domains of bovine ZP4 have not yet been clarified; however, in mouse, the ZP domain is essential for the assembly of ZP2 and ZP3 [10] In this study, both coexpressed rZP3⁄ rZP4136)464 and coexpressed rZP3⁄ rZP4182)464 mixtures showed sperm-binding activity similar to that of the rZP3⁄ rZP4 mixture, as revealed by a competitive inhibition assay (Method 1) and indirect immunofluorescence staining Moreover, rZP3 and rZP4182)464 formed het-ero-complexes These data indicate that the N-terminal region and trefoil domain of rZP4 are not necessary for the sperm-binding activity and hetero-complex for-mation of rZP3⁄ rZP4

a-Mannosyl residues at the nonreducing termini of high-mannose-type chains of the bovine ZP are essen-tial for sperm binding, as previously shown by the fact that a-mannosidase treatment greatly reduces the inhibitory activity of native ZP against sperm–egg binding [34] Porcine rZPGs expressed in Sf9 cells have pauci-mannose-type and high-mannose-type chains with or without fucose at the innermost GlcNAc, and

do not have detectable amounts of complex-type chains [16] Porcine rZP3⁄ rZP4, which binds to bovine sperm but not to porcine sperm, loses most of its inhibitory activity towards bovine sperm–ZP binding upon a-mannosidase treatment [16] Here, MS and