Molecular Cloning and Hormonal Control in the Ovary of Connexin 31.5 Mrna and Correlation with the Appearance of Oocyte Maturational Competence in Red Seabream

Gap junctions are specialized regions of the plasma membrane of animal cells characterized by intercellular cytoplasmic communication channels that allow direct cell-to-cell passage of ions, metabolites, secondary messengers and small molecules up to 1000 Da in mass, thereby providing a pathway for both electrical and chemical communication between networks of coupled cells (Stagg and Fletcher, 1990). Moreover, gap junctions participate in the regulation of diverse functions, including the contraction of cardiac and smooth muscle, the transmission of neuronal signals at electrotonic synapses, and metabolic cooperation during the development of avascular organs (Kumar and Gilula, 1986; Bruzzone et al., 1996; Grazul-Bilska et al., 1997). The homologous protein found in gap junctions is called connexon, and each connexon molecule is made up of six connexin (Cx) subunits (Beyer et al., 1990; Bruzzone et al., 1996). The predicted molecular mass is used in combination with the generic term to distinguish members of the connexin family. Thus, the 27 kDa protein from rat liver is termed rat Cx27 (Hertzberg et al., 1985; Beyer et al., 1987).

The significance of gap junctions between the oocyte and adjacent granulosa cells has been investigated in the ovarian follicles of several species of fish (Spray et al., 1981; Teranishi et al., 1983). After injecting the fluorescent dye Lucifer Yellow into an oocyte, the dye is normally transferred to the overlying granulosa cells (Cerda et al., 1993). In teleost fish, heterocellular gap junctions were first recognized by electron microscopy in cortical alveolar follicles just prior to the onset of vitellogenesis (Iwamatsu and Ohta, 1981; Kessel et al., 1985). The ovarian follicles of the Atlantic croaker (Micropogonias undulatus) contain many heterocellular and homocellular associations and gap junctional contacts during vitellogenesis. Few of these are present after the completion of vitellogenesis, but they can be induced again following the stimulation of maturational competence by human chorionic gonadotropin (hCG) (York et al., 1993). This indicates that gap junction communication plays an important role both during vitellogenesis and in the acquisition of maturational competence.

Yoshizaki et al. (1994) isolated and characterized two ovarian connexin cDNAs from the Atlantic croaker, Cx32.2 and Cx32.7. The increased production of connexin mRNA appeared to be selective because the level of Cx32.2 increased with gonadotropin (GtH) induction of maturational competence. Cx32.2 is considered essential in the acquisition of oocyte maturational competence (Yoshizaki et al., 1994). However, gonadotropin control of connexin expression in relation to maturational competence has not been examined in any other vertebrate species (Yoshizaki et al., 1994).

Tanaka et al. (1993) recently purified two distinct gonadotropins from the pituitary gland of the red seabream (Pagrus major), designated PmGtH-I and PmGtH-II. Both stimulated the in vitro production of oestradiol-17β in the vitellogenic follicles of the red seabream, although the biological activity of PmGtH-I was much lower than that of PmGtH-II (Tanaka et al., 1993). Moreover, it is known that priming with an injection of gonadotropin influences the induction of oocyte maturation in vitro (Patino and Thomas, 1990; Degani and Boker, 1992). In the oocytes of Atlantic croaker (Patino and Thomas, 1990) and red seabream (Kagawa et al., 1994), actinomycin D, an inhibitor of RNA and protein synthesis, has been found to inhibit gonadotropin-induced germinal vesicle breakdown. These reports suggest that the production of new proteins through RNA synthesis induced by gonadotropin is essential for the development of maturational competence in oocytes (Patino and Thomas, 1990; Kagawa et al., 1994). Except for Atlantic croaker connexin 32.2 mRNA (Yoshizaki et al., 1994), however, gene activity related to oocyte differentiation or maturational competence has not been directly investigated in red seabream or in any other vertebrate species.

The present study was designed to measure changes in the mRNA in the ovarian follicles of red seabream induced by treatment with hCG during the acquisition of oocyte maturational competence. The objectives of the study were to clarify whether 17α,20β-dihydroxy-4-pregnen-3-one (DHP; Sigma, St Louis, MO, USA), human insulin-like growth factor-I (IGF-I; human recombinant, Bachem Inc., Torrance, CA, USA), PmGtH-I or PmGtH-II can induce the acquisition of oocyte maturational competence at the molecular level during the final stage of maturation in the red seabream and to generalize the assumption that the synthesis of mRNA for connexin is essential to this process.

Mature red seabream (Pagrus major), weighing 2.0–3.0 kg, were purchased from fishermen in Gokasho Bay, Mie Prefecture, Japan, and kept in a flow-through outdoor tank (3000 l) under natural conditions of temperature and photoperiod. Fish were anaesthetized with 2-phenoxyethanol (Sigma) and killed by decapitation at 08:00 h (for incompetent ovaries) and 16:00 h (for competent ovaries). Incompetent ovaries were removed and place in ice-cold Ringer’s solution (pH 7.4, supplemented with 10 g l⁻¹ NaCl, 0.22 g l⁻¹ KCl, 0.25 g l⁻¹ CaCl₂, 0.74 g l⁻¹ MgSO₄·7H₂O, 1.19 g l⁻¹ Hepes, 0.9 g l⁻¹ glucose, 0.1 g l⁻¹ streptomycin and 100 000 i.u. l⁻¹ penicillin G calcium). The ovaries were cut into small pieces in ice-cold Ringer’s solution, and ovarian follicles were dispersed by pipetting the fragments.

Clusters of ovarian follicles were removed by passing the suspension through a stainless-steel wire mesh (pore diameter approximately 600 µm). Ovarian follicles at the tertiary yolk stage (approximately 500 µm in diameter; Matsuyama et al., 1988) were collected on a wire mesh (pore diameter approximately 420 µm). Approximately 60 ovarian follicles were transferred into individual wells of 24-well culture plates, each of which contained 1 ml of Leibovitz’s L-15 medium (Gibco-BRL, Gaithersburg, MD, USA) supplemented with 2.4 g l⁻¹ Hepes, 0.1 g l⁻¹ streptomycin and 100 000 i.u. l⁻¹ penicillin G calcium (pH 7.6). The oocytes were pre-incubated in 1 ml of Leibovitz’s L-15 medium only for 0 h (control 0) or 12 h (control 12). Moreover, the oocytes were pre-incubated in 1 ml of Leibovitz’s L-15 medium in the presence of DHP (10 ng ml⁻¹), IGF-I (10 µmol ml⁻¹), PmGtH-I (100 ng ml⁻¹ or 300 ng ml⁻¹) or PmGtH-II (100 ng ml⁻¹ or 300 ng ml⁻¹) for 12 h, washed three times with fresh Leibovitz’s L-15 medium, transferred to Leibovitz’s L-15 medium containing DHP (10 ng ml⁻¹) and/or actinomycin D (1 µg ml⁻¹; Sigma) and incubated for a further 12 h at 20 °C.

The levels of oocyte maturation were determined by counting the number of oocytes that had undergone germinal vesicle breakdown. The occurrence of germinal vesicle breakdown was used as the criterion of oocyte maturation. After incubation, all the ovarian follicles were stored at −80 °C until used.

Data are expressed as means ± S.E.M. Analysis of variance (ANOVA) followed by Duncan’s multiple-range tests were applied for statistical analysis.

Two highly conserved regions of human Cx32 (Kumar and Gilula, 1986), rat Cx43 (Beyer et al., 1987) and Xenopus laevis Cx30 (Gimlich et al., 1988), Cx38 (Ebihara et al., 1989; Gimlich et al., 1990) and Cx43 (Gimlich et al., 1990) were used to design mixed primers for the polymerase chain reaction (PCR). One of the regions is located at the predicted junction of the first extracellular and second transmembrane domains [5′-TTCCCT(CT)AT(ACT)TC(ACT)CA(CT)(AG)T(CGT)CG-3′] and the other at the junction of the second extracellular and fourth transmembrane domains [5′-GT(CT)TT(CT)TC(ACGT)(AG)TGGG(ACGT)C(GT)(AGT) GA-3′] (Yoshizaki et al., 1994). Total RNA was extracted from incompetent and competent ovaries using a total RNA extraction kit (Pharmacia Biotech, Uppsala, Sweden). Total RNA (5 µg) was used for cDNA synthesis using a first-strand cDNA synthesis kit (Pharmacia Biotech). PCR was carried out using 25 ng of cDNA as template, 10 µmol l⁻¹ forward and reverse primers, 200 µmol l⁻¹ of each dNTP and Taq DNA polymerase (5 units ml⁻¹; Takara, Japan) in 50 µl of buffer (50 mmol l⁻¹ KCl, 10 mmol l⁻¹ Tris-HCl, pH 9.0, 0.1 % Triton X-100, 1.5 mmol l⁻¹ MgCl₂). Thirty cycles were conducted, each cycle consisting of 1 min at 94 °C, 45 s at 50 °C and 1 min at 72 °C, except that the first denaturation was carried out for 2 min and the last elongation reaction for 5 min. A sample (5 µl) of each PCR product was electrophoresed in 2 % agarose gels, with the DNA molecular mass marker X (Boehringer Mannheim, Germany) as a reference to estimate the molecular masses of the amplified fragments. The remaining PCR products were electrophoresed in 2 % agarose gels, and fragments were isolated, ligated, cloned into pGEM-T Easy Vector (Promega, Madison, WI, USA) and sequenced using a thermo sequenase fluorescent-labelled primer cycle sequencing kit with 7-deaza-dGTP (Amersham, Buckinghamshire, UK).

Two specific primers (SP) were designed, SP1 (G5′-CAAGGAGCTGGAACACATCA-3′) and SP2 (G5′-AGCTCTGTGGTGGACTTACA-3′). The level of the PCR product (358 bp) of ovarian mRNA increased significantly during the acquisition of oocyte maturational competence. First-strand cDNA synthesis was initiated at the poly(A)⁺ RNA using the oligo(dT) anchor primer 5′-AACTGGA-AGAATTCGCGGCCGCAGGAAT₁₈-3′. Specific cDNA was amplified by PCR using specific primers (SP1 or SP2) that anneal to a region of known sequence and an adaptor primer (5′-AACTGGAAGAATTCGCGGCCGCAG-3′). PCR was carried out with 0.2 µg of cDNA as template, 25 µmol l⁻¹ forward and reverse primers, 0.25 mmol l⁻¹ of each dNTP and Ex Taq polymerase (Takara, Japan) in 50 µl of buffer. Additional controls that amplify dC-tailed cDNA using each primer individually (either the adaptor primer or SP2) were used to identify nonspecific products. PCR was carried out for 30 cycles as follows: 94 °C for 30 s for denaturing, 52 °C for 45 s for primer annealing and 72 °C for 45 s for extension, followed by one cycle of 3 min at 70 °C for extension. The final PCR product was amplified and cloned into pGEM-T Easy Vector and sequenced using the ALF DNA sequencer (Pharmacia Biotech).

The cDNA fragment generated by the 3′RACE procedures was used to generate full-length cDNA using a 5′RACE system (version 2.0 kit; Gibco-BRL). Total RNA (1 µg) was reverse-transcribed according to the kit protocol using a gene-specific primer (GSP) (GSP1; 5′-GCAGGGCAGTCTCAGAAA-GCAG-3′) located within the coding sequence. Two further GSPs were designed, GSP2 (5′-GACACTGCCAGC-ATGAACACAGTG-3′) and GSP3 (5′-CTTCTGGTTCTT-GATGTGTTCCAGCTCCTT-3′). The full length of the oligo(dG) anchor primer was included with the kit. The final PCR product was amplified and cloned into pGEM-T Easy Vector and sequenced. The DNA sequence was analyzed using the GENETYX-WIN (Software Development Co., Japan) software package.

To clarify whether isolated mRNA from ovarian follicles incubated with each hormone was differentially expressed specifically during the maturation of oocytes, northern blot analysis was performed. Poly(A)⁺ RNA was purified using an mRNA purification kit (Oligotex-dT30 Super, Japan). Messenger RNA (5 µg) was electrophoresed according to the formaldehyde gel method (Sambrook et al., 1989) and transferred to nylon membranes (Hybond N⁺) (Amersham) according to the manufacturer’s instructions. The samples were prehybridized at 65 °C in rapid hybridization buffer (Amersham) for 10 min, and then hybridized with the probe at 65 °C for 3 h. The PCR product label used for the oligolabelling kit (Amersham) was [³²P]dCTP-labelled cDNA probe. Post-hybridization washing was carried out in 2× standard saline citrate (SSC) for 15 min at 65 °C, in 1× SSC/0.5 % sodium dodecyl sulphate (SDS) for 15 min at 65 °C, and finally in 0.1× SSC/0.5 % SDS for 15 min at 65 °C.

The hybridization signals were analyzed using a FUJIX BAS 1000 bio-imaging analyzer (Fuji Film, Japan). The expression level of each sample was normalized to the red seabream β-actin signal and expressed as percentage of the control level.

One major fragment (358 base pairs) was amplified from the competent oocytes. Homology analyses using the GenBank and EMBL general database searches indicated that the nucleotide sequence of the 358 base pair (bp) cDNA has a high homology with bovine Cx32 (Simonic et al., 1997) (69.1 % identity) and human Cx32 (Kumar and Gilula, 1986) (87.3 % identity). As shown by PCR, levels of amplified cDNA fragments of 358 bp increased simultaneously with the acquisition of oocyte maturational competence (data not shown). The connexin cDNAs generated by the 3′RACE and 5′RACE procedures may be combined to generate a full-length cDNA sequence. The 2400 bp cDNA had an open reading frame of 825 bp that began with the first ATG codon at position 359 and ended with a TGA stop codon at position 1184 (Fig. 1). A putative polyadenylation signal AATAAA (Proudfoot and Brownlee, 1976) occurred at position 2366 (Fig. 1). This open reading frame encodes a polypeptide that contains 275 amino acid residues (Fig. 1) with a predicted molecular mass of 31393 Da. We designated this protein red seabream P. major (Pm) Cx31.5 on the basis of the nomenclature system of Beyer et al. (1987).

The GenBank and EMBL databases showed that, among all previously reported connexin genes, PmCx31.5 has the highest homology to human Cx32 (66.3 % similarity and 88.0 % identity) (Simonic et al., 1997) and rat Cx32 (65.3 % similarity and 88.3 % identity) (Kumar and Gilula, 1986) (Fig. 2). Hydropathicity analysis of the predicted amino acid sequences using the method of Kyte and Doolitte (1982) indicated that PmCx31.5 has four major hydrophobic regions and four major hydrophilic regions. Hydrophobic regions predict transmembrane domains, while hydrophilic regions indicate extracellular or intracellular domains (Fig. 3). Thus, the identify and similarity between the positively established domains of rat Cx32 (Haefliger et al., 1992) and Atlantic croaker Cx32.2 (Yoshizaki et al., 1994) and the corresponding putative domains of PmCx31.5 can be determined individually (Table 1).

With the exception of the intracellular loop and C-terminal domain, all other domains of PmCx31.5 showed relatively high homology with that of Atlantic croaker Cx32.2 (44.2 % identity and 67.6 % similarity for Cx31.5). Connexin consensus sequences containing three cysteine residues were identified in each of the extracellular domains (loops): C(D,N)TXQPGCX₂VCYD in the first loop (Fig. 1; filled circles) and CX_3∼4PCX₃(L,I,M,V)(D,E,N)C(F,Y)(L,I,V,M)(S,A)(K,R)P (Yoshizaki et al., 1994) in the second loop (Fig. 1; filled triangles). Further, a consensus sequence found in the C-terminal cytoplasmic domain of most connexins, QNX_6or9S (Hoh et al., 1991), was identified in Cx31.5 at amino acid residues 242–253 (Fig. 1; asterisks). The amphipathic helix consensus of the third transmembrane domain, TX₃SX₃(K,R)X₃E (Hoh et al., 1991), was present at positions 134–145 (Fig. 1; open circles). As described above, PmCx31.5 is similar to the connexin family in its general features, and all the consensus sequences are present.

Northern blot analysis of the PmCx31.5 transcript with the β-actin probe as a positive control was carried out using 5 µg of poly(A)⁺ RNA extracted from ovarian follicles cultured for 12 h in normal medium, DHP (10 ng ml⁻¹), IGF-I (10 µmol ml⁻¹), PmGtH-I (100 ng ml⁻¹ or 300 ng ml⁻¹) or PmGtH-II (100 ng ml⁻¹ or 300 ng ml⁻¹) (Fig. 4). PmCx31.5 mRNA was strongly expressed in the presence of PmGtH-II (300 ng ml⁻¹), but not in the presence of DHP, IGF-I or PmGtH-I. Fig. 5 shows the effects of control 12, DHP (10 ng ml⁻¹), IGF-I (10 µmol ml⁻¹), PmGtH-I (100 ng ml⁻¹ or 300 ng ml⁻¹) and PmGtH-II (100 ng ml⁻¹ or 300 ng ml⁻¹) on the percentage in vitro germinal vesicle breakdown of oocytes and on the PmCx31.5 transcription level. PmGtH-II induced a high level of oocyte germinal vesicle breakdown. Only in the presence of GtH-II (300 ng ml⁻¹) was the transcription level of PmCx31.5 significantly different (P<0.01) from the control value. Northern blot analysis of poly(A)⁺ RNA (5 µg) extracted from maturationally competent and incompetent ovaries is shown in Fig. 6. PmCx31.5 mRNA was strongly expressed after, but not before, the acquisition of oocyte maturational competence.

We have shown that GtH-II induced oocyte maturational competence in the ovarian follicles of the red seabream. Kagawa et al. (1998) reported that GtH-II, but not GtH-I, is involved in the final maturation of red seabream oocytes. Moreover, in the oocytes of the Atlantic croaker, actinomycin D, an inhibitor of RNA synthesis, blocked the induction of oocyte maturational competence by gonadotropin (Patino and Thomas, 1990). Kagawa et al. (1994) reported that hCG induces maturational competence in red seabream oocytes via de novo synthesis of RNA and protein. Newly synthesized proteins are therefore probably involved in hCG (or GtH-II)-induced oocyte maturation. However, if hCG (or GtH-II) directly stimulates an increase in oocyte maturational competence in ovarian follicles, this increase should not require de novo protein synthesis. In the present study, connexin cDNA was expressed during the acquisition of oocyte maturational competence in ovarian follicles after treatment with GtH-II.

This study demonstrated that the amount of connexin mRNA in the gap junctions of ovarian follicles increases as a result of treatment with GtH-II. We are not sure what role the gap junction plays in the function of ovarian follicles. Electron microscopy has demonstrated that gap junctions form between microvilli between the oocyte and granulosa cells of Xenopus laevis (Gimlich et al., 1988; Ebihara et al., 1989), medaka (Oryzias latipes) (Iwamatsu and Ohta, 1981) and zebrafish (Brachydanio rerio) (Kessel et al., 1985). Furthermore, gonadotropin treatment increases the number of gap junctions in Xenopus laevis (Yoshizaki and Patino, 1995) and Atlantic croaker (Yoshizaki et al., 1994) ovarian follicles. Reports suggest that the intercellular movement of Lucifer Yellow from oocytes to granulosa cells is promoted by gonadotropin in Xenopus laevis (Browne et al., 1979) and Bufo arenarum (Villecco et al., 1996).

The role of the gap junction in the maturation of oocytes can be inferred from the observation that only GtH-II induces increases in PmCx31.5 mRNA levels, while GtH-I and DHP do not. Moreover, Yoshizaki et al. (1994) reported that levels of mRNA transcripts of Atlantic croaker Cx32.2 were increased when hCG treatment was used to stimulate the acquisition of oocyte maturational competence. This phenomenon was independent of the production of maturation-inducing steroid, which has the same effect as hCG on the acquisition of oocyte maturation competence in the Atlantic croaker (Chang et al., 1999). In the present study, PmCx31.5 showed a similar expression pattern to Atlantic croaker Cx32.2, which is also expressed after hCG (GtH-II) treatment.

GtH-II was found to increase the number of gap junctions between the oocyte and granulosa cell via increased levels of PmCx31.5, leading to the acquisition of oocyte maturational competence. However, although IGF-I effectively induces germinal vesicle breakdown, it does not affect the level of PmCx31.5 transcripts (Figs 4, 5). It would appear, therefore, that the induction of oocyte maturational competence by treatment with IGF-I results from an action different from that of gonadotropin. This agrees with a previous report (Kagawa et al., 1994) that IGF-I transcription is not involved in the acquisition of oocyte maturational competence. The mechanism for the induction of oocyte maturational competence by IGF-I remains unexplained.

It is noteworthy that PmCx31.5 possesses protein kinase C (PKC) phosphorylation sites, (S,T)X(R,K) (Kishimoto et al., 1985; Woodget et al., 1986), and casein protein kinase II phosphorylation sites, (S,T)X₂(D,E) (Pinna, 1990) (Fig. 1). These kinases regulate phosphorylation in many systems (Hoh et al., 1991). Takeda et al. (1987) reported that rat Cx32 was phosphorylated by PKC in vitro. Atlantic croaker Cx32.2 also has PKC phosphorylation sites, and the increase in ovarian Cx32.2 mRNA production coincides with the induction of oocyte maturational competence (Yoshizaki et al., 1994). Moreover, phorbol 12-myristate 13-acetate (PMA), a PKC activator, inhibits the induction of oocyte maturational competence by gonadotropin (Stagg and Fletcher, 1990; Chang et al., 1999). It is conceivable, therefore, that the cyclic AMP/protein kinase A (PKA) transduction pathway is involved in the activation of the Cx31.5 gene by gonadotropin in the red seabream, particularly since cyclic AMP response elements have been identified in the 5′ upstream region (Miller et al., 1988; Yoshizaki et al., 1994). In the nibe (Nibea mitsukuri), oocyte maturation is induced through the activation of PKA (G. Yoshizaki, personal communication). Furthermore, PKC acutely inhibits and PKA acutely enhances cell–cell junctional coupling (Stagg and Fletcher, 1990). In our study, a northern blot analysis was carried out using 5 µg of poly(A)⁺ RNA extracted from incompetent and competent ovaries (Fig. 6). PmCx31.5 mRNA was strongly expressed after, but not before, the acquisition of oocyte maturational competence. These results suggest that the production of Cx31.5 in the red seabream ovary is related to the maturation of ovarian follicles to produce competent oocytes.

In summary, this study found that the induction of oocyte maturational competence by GtH-II requires de novo synthesis of ovarian RNA and protein and coincides with an increase in the production of ovarian PmCx31.5 mRNA and the establishment of granulosa cell–oocyte gap junctions in fully grown ovarian follicles (York et al., 1993). Thus, it is possible that PmCx31.5 is the protein synthesized during the gonadotropin-induced acquisition of maturational competence in red seabream ovaries. We have not determined the tissue distribution of PmCx31.5 gene expression, nor have the mechanisms of PmCx31.5 activation with hCG treatment been addressed in the present study. Moreover, the physiological roles of PmCx31.5 in the acquisition of oocyte maturational competence were not investigated. Further studies are necessary to identify the second-messenger systems that regulate the PmCx31.5 gene under the influence of gonadotropin in the ovary.

The authors would like to thank to Dr John F. Leatherland, Department of Biomedical Sciences, Ontario Veterinary College, University of Guelph, Canada, for his helpful comments on this manuscript. We appreciate the useful advice of Dr Akihiro Takemura, Sesoko Station, Tropical Biosphere Research Center, University of the Ryukyus, Japan.