Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

The Development-associated Cleavage of Lens Connexin 45.6 by Caspase-3-like Protease Is Regulated by Casein Kinase II-mediated Phosphorylation

2001; Elsevier BV; Volume: 276; Issue: 37 Linguagem: Inglês

10.1074/jbc.m106073200

ISSN

1083-351X

Autores

Xinye Yin, Sumin Gu, Jean X. Jiang,

Tópico(s)

Intraocular Surgery and Lenses

Resumo

Gap junctions are important in maintaining lens transparency and metabolic homeostasis. In this paper, we report that the gap junction-forming protein, connexin (Cx) 45.6, was specifically truncated during lens development and that the majority of the truncated fragments were located in the differentiated lens fibers. When isolated lens membranes were treated by caspase-3, the truncated fragments of Cx45.6 were reproduced, and this truncation occurred at the COOH terminus of Cx45.6. Moreover, when primary lens cells were treated with apoptosis-inducing reagents, Cx45.6 was cleaved similarly as the in vitro treatment by caspase-3, and this cleavage was blocked by a caspase-3 inhibitor. These results suggest that caspase-3 is responsible for the development-associated cleavage of Cx45.6. The cleavage site of Cx45.6 was identified between amino acid residues Glu367 and Gly368. We have shown previously that Ser363 is an in vivophosphorylated site by casein kinase II, and this specific phosphorylation leads to a rapid turnover of Cx45.6. Interestingly, we found here that when Ser363 was phosphorylated by casein kinase II, the cleavage of Cx45.6 catalyzed by caspase-3 was inhibited. This study, for the first time, demonstrates that a connexin can be a direct target of an apoptotic protease and that cleavage by caspase-3-like protease leads to the development-associated truncation of a lens connexin. Finally, caspase-3-mediated cleavage can be regulated by casein kinase II-mediated phosphorylation, suggesting that Cx45.6 turnover and specific cleavage by caspase-3-like protease is alternatively modulated. Gap junctions are important in maintaining lens transparency and metabolic homeostasis. In this paper, we report that the gap junction-forming protein, connexin (Cx) 45.6, was specifically truncated during lens development and that the majority of the truncated fragments were located in the differentiated lens fibers. When isolated lens membranes were treated by caspase-3, the truncated fragments of Cx45.6 were reproduced, and this truncation occurred at the COOH terminus of Cx45.6. Moreover, when primary lens cells were treated with apoptosis-inducing reagents, Cx45.6 was cleaved similarly as the in vitro treatment by caspase-3, and this cleavage was blocked by a caspase-3 inhibitor. These results suggest that caspase-3 is responsible for the development-associated cleavage of Cx45.6. The cleavage site of Cx45.6 was identified between amino acid residues Glu367 and Gly368. We have shown previously that Ser363 is an in vivophosphorylated site by casein kinase II, and this specific phosphorylation leads to a rapid turnover of Cx45.6. Interestingly, we found here that when Ser363 was phosphorylated by casein kinase II, the cleavage of Cx45.6 catalyzed by caspase-3 was inhibited. This study, for the first time, demonstrates that a connexin can be a direct target of an apoptotic protease and that cleavage by caspase-3-like protease leads to the development-associated truncation of a lens connexin. Finally, caspase-3-mediated cleavage can be regulated by casein kinase II-mediated phosphorylation, suggesting that Cx45.6 turnover and specific cleavage by caspase-3-like protease is alternatively modulated. connexin casein kinase II chicken embryonic fibroblast Dulbecco's modified Eagle's medium glutathione S-transferase polyacrylamide gel electrophoresis polymerase chain reaction benzyloxycarbonyl fluoromethyl ketone 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid embryonic day postnatal day The vertebrate lens is one of the most important model systems used in the study of the function and regulation of gap junctions. The lens is an avascular organ composed of an anterior epithelial cell layer and highly differentiated fibers ranging from the outer cortex toward the central core region. The terminal differentiation and aging of lens fibers are marked by dramatic morphological changes. As new cells arise on the outside of the lens, older cells are pushed inward where nuclei and organelles are lost (1Rafferty N. Esson E. J. Ultrastruct. Res. 1974; 46: 239-253Crossref PubMed Scopus (39) Google Scholar). Because these cells are never lost from the lens, the most central cells are as old as the organism itself. The survival of lens cells relies on the intercellular communications between these cells and the cells at the lens surface through a large network of gap junctions that facilitate the exchange of ions and metabolites throughout the organ (2Mathias R.T. Rae J.L. Sperelakis N. Cole W.C. Cell Interactions and Gap Junctions. CRC Press, Inc., Boca Raton, FL1989: 29-50Google Scholar, 3Cooper K. Mathias R.T. Rae J.L. Peracchia C. Biophysics of Gap Junction Channels. CRC Press, Inc., Boca Raton, FL1991: 57-74Google Scholar). Gap junctions are intercellular channels between two adjacent cells, which allow passage of small molecules (≤1000 Da) such as small metabolites, ions, and second messengers. The structural components of gap junctions are the members of a protein family called connexins, which consist of four conserved transmembrane domains and two extracellular loops, whereas their cytoplasmic regions are unique. The COOH terminus, the most variable region among connexins, contains several kinase and protease consensus sequences. Connexins are highly dynamic proteins that undergo rapid turnover both in cell lines (4Musil L.S. Cunningham B.A. Edelman G.M. Goodenough D.A. J. Cell Biol. 1990; 111: 2077-2088Crossref PubMed Scopus (594) Google Scholar) and in animal organs (5Dermietzel R. Traub O. Hwang T.K. Beyer E. Bennett M.V.L. Spray D.C. Willecke K. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 10148-10152Crossref PubMed Scopus (457) Google Scholar, 6Beardslee M.A. Laing J.G. Beyer E.C. Saffitz J.E. Circ. Res. 1998; 83: 629-635Crossref PubMed Scopus (370) Google Scholar, 7Fallon R.F. Goodenough D.A. J. Cell Biol. 1981; 90: 521-526Crossref PubMed Scopus (224) Google Scholar). The degradation has been identified to undergo either through the lysosome or proteasome pathway (8Guan X.J. Ruch R.J. Carcinogenesis. 1996; 17: 1791-1798Crossref PubMed Scopus (77) Google Scholar, 9Laing J.G. Beyer E.C. J. Biol. Chem. 1995; 270: 26399-26403Abstract Full Text Full Text PDF PubMed Scopus (205) Google Scholar, 10Laing J.G. Tadros P.N. Westphale E.M. Beyer E.C. Exp. Cell Res. 1997; 236: 482-492Crossref PubMed Scopus (217) Google Scholar, 11Musil L.S. Le A.-C.N. VanSlyke J.K. Roberts L.M. J. Biol. Chem. 2000; 275: 25207-25215Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar, 12Berthoud V.M. Tadros P.N. Beyer E.C. Methods. 2000; 20: 180-187Crossref PubMed Scopus (36) Google Scholar). Phosphorylation has been demonstrated to play an important role in regulation of gap junction stability and turnover (13Hertlein B. Butterweck A. Haubrich S. Willecke K. Traub O. J. Membr. Biol. 1998; 162: 247-257Crossref PubMed Scopus (57) Google Scholar, 14Berthoud V.M. Bassnett S. Beyer E.C. Exp. Eye Res. 1999; 68: 475-484Crossref PubMed Scopus (34) Google Scholar, 15Hossain M.Z. Ao P. Boynton A.L. J. Cell. Physiol. 1998; 174: 66-77Crossref PubMed Scopus (73) Google Scholar, 16Elvira M. Diez J.A. Wang K.K.W. Villalobo A. J. Biol. Chem. 1993; 268: 14294-14300Abstract Full Text PDF PubMed Google Scholar). We have recently shown that an in vivo phosphorylated lens connexin undergoes a faster turnover than its unphosphorylated counterpart (17Yin X. Jedrzejewski P.T. Jiang J.X. J. Biol. Chem. 2000; 275: 6850-6856Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar). Differentiating lens fibers share a number of morphological and biochemical characteristics with cells undergoing apoptosis such as nucleus degeneration, loss of organelles, and activation of members of a cysteine protease family named caspase (18Wride M.A. Parker E. Sanders E.J. Dev. Biol. 1999; 213: 142-156Crossref PubMed Scopus (76) Google Scholar, 19Wride M.A. Apoptosis. 2000; 5: 203-209Crossref PubMed Scopus (79) Google Scholar, 20Dahm R. Ophthalmic Res. 1999; 31: 163-183Crossref PubMed Scopus (95) Google Scholar, 21Lee A. Morrow J.S. Fowler V.M. J. Biol. Chem. 2001; 276: 20735-20742Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar). However, unlike apoptotic cells, which are rapidly digested, the organelle-free lens fibers retain their basic cell integrity and metabolism throughout the lifetime of the organ. Poly(ADP-ribose) polymerase, an enzyme involved in DNA repair and maintenance of genomic integrity, is cleaved by caspases in the developing lens (18Wride M.A. Parker E. Sanders E.J. Dev. Biol. 1999; 213: 142-156Crossref PubMed Scopus (76) Google Scholar). Caspase-3 is predominantly activated during staurosporine-induced apoptosis in lens cells (22Andersson M. Sjöstrand J. Petersen A. Honarvar A.K.S. Karlsson J. Invest. Ophthalmol. Vis. Sci. 2000; 41: 2623-2632PubMed Google Scholar). The function of caspases has been reported to be responsible for nucleus degradation during the later stage of lens fiber differentiation (18Wride M.A. Parker E. Sanders E.J. Dev. Biol. 1999; 213: 142-156Crossref PubMed Scopus (76) Google Scholar). Although activation of caspases has been demonstrated during lens fiber differentiation, only a couple of caspase substrates have been identified, such as cytoskeletal proteins named α- and β-spectrins (21Lee A. Morrow J.S. Fowler V.M. J. Biol. Chem. 2001; 276: 20735-20742Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar). Connexin (Cx)1 45.6 is one of the connexins expressed in chick lens fibers (23Jiang J.X. White T.W. Goodenough D.A. Paul D.L. Mol. Biol. Cell. 1994; 5: 363-373Crossref PubMed Scopus (81) Google Scholar, 24Jiang J.X. White T.W. Goodenough D.A. Dev. Biol. 1995; 168: 649-661Crossref PubMed Scopus (55) Google Scholar). Cx45.6, which is different from the other two chick lens connexins, Cx43 and Cx56, seems to be involved in lens development and differentiation. Mutations in Cx50, the mammalian ortholog of Cx45.6, have been identified in human cataract families (25Shiels A. Mackay D. Ionides A. Berry V. Moore A. Bhattacharya S. Am. J. Hum. Genet. 1998; 62: 526-532Abstract Full Text Full Text PDF PubMed Scopus (317) Google Scholar, 26Berry V. Mackay D. Khaliq S. Francis P.J. Hameed A. Anwar K. Mehdi S.Q. Newbold R.J. Ionides A. Shiels A. Moore T. Bhattacharya S.S. Hum. Genet. 1999; 105: 168-170Crossref PubMed Scopus (114) Google Scholar). Mice deficient in Cx50 not only develop cataracts, but also exhibit microphthalmia with a decrease of lens size (27White T.W. Goodenough D.A. Paul D.L. J. Cell Biol. 1998; 143: 815-825Crossref PubMed Scopus (298) Google Scholar). Primary cultures of chick lens closely mimic in vivolens cell differentiation processes in which the monolayer of the lens epithelial cells differentiate into structures called "lentoids" displaying features of differentiated lens fibers (28Piatigorsky J. Rothschile S.S. Milstone L.M. Dev. Biol. 1973; 34: 334-345Crossref PubMed Scopus (65) Google Scholar, 29Menko A.S. Klukas K.A. Johnson R.G. Dev. Biol. 1984; 103: 129-141Crossref PubMed Scopus (95) Google Scholar). We have recently shown that overexpression of Cx45.6 in lens primary cultures stimulates lens cell differentiation and formation of fibers (30Gu S. Yin X. Jiang J.X. Mol. Biol. Cell. 1999; 10: 405Google Scholar). Additionally, Cx45.6 is subjected to post-translational modificationin vivo. Cx45.6 is a phosphoprotein (23Jiang J.X. White T.W. Goodenough D.A. Paul D.L. Mol. Biol. Cell. 1994; 5: 363-373Crossref PubMed Scopus (81) Google Scholar, 31Jiang J.X. Goodenough D.A. Eur. J. Biochem. 1998; 255: 37-44Crossref PubMed Scopus (23) Google Scholar). We have recently shown that Cx45.6 is phosphorylated in vivo by CKII at Ser363 located at the COOH-terminal region, and this phosphorylation facilitated the turnover of Cx45.6 (17Yin X. Jedrzejewski P.T. Jiang J.X. J. Biol. Chem. 2000; 275: 6850-6856Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar). A previous report shows that Cx50, the ovine ortholog of Cx45.6, is also truncated at the COOH terminus in the center of the lens by a protease called calpain (32Lin J.S. Fitzgerald S. Dong Y. Knight C. Donaldson P. Kistler J. Eur. J. Cell Biol. 1997; 73: 141-149PubMed Google Scholar). Although post-translational modifications, such as phosphorylation and specific truncation of lens connexins, have been investigated previously, the physiological significance and regulatory mechanism of these modifications in lens differentiation and development are less certain. In this report, we identified a specific proteolytic cleavage of Cx45.6 associated with lens development. Our experimental evidence suggests that the protease involved in this cleavage is caspase-3, an apoptotic protease. Moreover, the cleavage site by caspase-3 in Cx45.6 was identified to be close to the phosphorylated site by CKII. This specific phosphorylation protects Cx45.6 from cleavage by caspase-3. This is the first report showing that the apoptotic protease, caspase-3, is responsible for cleavage of a connexin, and this specific cleavage is regulated by a specific phosphorylation. Fertilized chicken eggs were obtained from SPAFAS (Roanoke, IL) and Tyson Hatchery (Gonzalez, TX). CKII and caspase-3 were obtained from Sigma. Caspase-3 inhibitor Z-DEVD-FMK was from Calbiochem. [γ-32P]ATP and [35S]methionine were from PerkinElmer Life Sciences. Tissue culture reagents were purchased from Life Technologies, Inc. SDS-PAGE standards were from Bio-Rad. QuikChangeTMSite-directed Mutagenesis kit was obtained from Stratagene (La Jolla, CA). Chemiluminescence kit, ECL, was from Amersham Pharmacia Biotech.X-Omat AR film was obtained from Eastman Kodak. Bacterial fusion proteins containing GST fused with various portions of Cx45.6 were prepared as described (23Jiang J.X. White T.W. Goodenough D.A. Paul D.L. Mol. Biol. Cell. 1994; 5: 363-373Crossref PubMed Scopus (81) Google Scholar, 33He D.S. Jiang J.X. Taffet S.M. Burt J.M. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 6495-6500Crossref PubMed Scopus (163) Google Scholar). The fusion proteins generated included the following: GST fused with partial COOH terminus of Cx45.6, GST-Cx45.6F1 (amino acids 307–389) and its corresponding site mutants, GST-Cx45.6F1(D364A) and GST-Cx45.6F1(E367A); GST fused with entire COOH terminus of Cx45.6, GST-Cx45.6F2 (amino acids 237–400); and GST fused with intracellular loop region of Cx45.6, GST-Cx45.6(L) (amino acids 98–148). Briefly, DNA fragments containing various portions of Cx45.6 or its mutants were produced by PCR with specific oligonucleotide primers. The codon GAT encoding Asp364 was changed to GCT encoding Ala and GAA encoding Glu367 was altered to GCA encoding Ala in one of the two primer sequences required for mutant synthesis. Each DNA fragment generated was inserted into the expression vector pGEX-2T. The recombinant fusion proteins were expressed inEscherichia coli, induced by isopropyl-thio-β-d-galactoside, and isolated with GST beads. To ensure the correct sequence all constructs generated were sequenced at the Institutional DNA Sequencing Facility. Fusion protein containing GST plus intracellular loop portion of Cx45.6 was used to raise polyclonal antisera in rabbits (Pocono Rabbit Farm and Laboratory Inc, Canadensis, PA). The antisera generated were immunoaffinity-purified by passage through two Sepharose CL-4B columns, GST-conjugated and GST-Cx45.6 fusion protein-conjugated, respectively, as described (23Jiang J.X. White T.W. Goodenough D.A. Paul D.L. Mol. Biol. Cell. 1994; 5: 363-373Crossref PubMed Scopus (81) Google Scholar, 33He D.S. Jiang J.X. Taffet S.M. Burt J.M. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 6495-6500Crossref PubMed Scopus (163) Google Scholar). The fusion proteins and crude membrane samples were loaded in each lane of a 10% SDS-PAGE. For Western blotting, samples on SDS-PAGE were transferred to nitrocellulose membranes according to the method of White et al. (34White T.W. Bruzzone R. Goodenough D.A. Paul D.L. Mol. Biol. Cell. 1992; 3: 711-720Crossref PubMed Scopus (274) Google Scholar). Membranes were probed with the affinity-purified preimmune or Cx45.6 antibody (1:500 dilution). The primary antibodies were detected either with alkaline phosphatase-conjugated secondary anti-rabbit antiserum or peroxidase-conjugated secondary anti-rabbit antiserum and followed by chemiluminescence reagent kit (ECL) (Amersham Pharmacia Biotech) according to the manufacturer's instruction. The membranes were exposed to X-Omat AR films (Eastman Kodak) and detected by fluorography. Retroviral constructs and high titer viruses were prepared based on our protocol described previously (35Jiang J.X. Bruzzone R. Giaume C. Connexin Methods and Protocols. Humana Press Inc., Totowa, NJ2000: 159-174Google Scholar). Briefly, cDNA fragment containing the wild type Cx45.6 was made by PCR and was constructed into the retroviral vector RCAS(A) as described (17Yin X. Jedrzejewski P.T. Jiang J.X. J. Biol. Chem. 2000; 275: 6850-6856Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar, 36Jiang J.X. Goodenough D.A. Invest. Ophthalmol. Vis. Sci. 1998; 39: 537-543PubMed Google Scholar). With the wild type RCAS(A)-Cx45.6 DNA construct as a template, retroviral constructs of Cx45.6 mutants containing point mutations were generated with the QuikChangeTM Site-directed Mutagenesis kit according to the manufacturer's instructions. The codon alteration for site mutants, RCAS(A)-Cx45.6(D364A) and RCAS(A)-Cx45.6(E367A), were described above. RCAS(A)-Cx45.6(E367D) was made by changing codon GAA encoding Glu367 to GAT encoding Asp. All constructs generated were sequenced at the Institutional DNA Sequencing Facility. The high titer retroviruses (1–5 × 108 colony forming units/ml) containing the wild type and the mutant RCAS(A) Cx45.6 constructs were prepared as we have described previously (35Jiang J.X. Bruzzone R. Giaume C. Connexin Methods and Protocols. Humana Press Inc., Totowa, NJ2000: 159-174Google Scholar, 36Jiang J.X. Goodenough D.A. Invest. Ophthalmol. Vis. Sci. 1998; 39: 537-543PubMed Google Scholar). The crude cell membranes were isolated and prepared as described (37Jiang J.X. Goodenough D.A. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 1287-1291Crossref PubMed Scopus (230) Google Scholar). Briefly, lenses from embryonic day 9 or CEF cells infected with retroviruses were lysed in lysis buffer (5 mm Tris, pH 8.0, and 5 mm EDTA/EGTA) plus 2 mm phenylmethylsulfonyl fluoride, 10 mmN-ethylmaleimide, and 100 μm leupeptin. Crude membranes were pelleted at 35,000 rpm for 30 min (Beckman SW60 rotor). The crude membranes from CEF cells and lenses or the purified GST fusion proteins were subjected to in vitro cleavage by caspase-3. Crude membranes or 2.5 μg of various GST fusion proteins were incubated at 37 °C in 25 μl of caspase-3 reaction buffer (50 mm Tris-HCl, pH 8.0, 0.5 mm EDTA, 0.5 mm sucrose, and 5% glycerol) with and without 0.3 μg of caspase-3 at various times. Reactions were terminated by the addition of electrophoresis sample buffer (50 mm Tris, pH 6.8, 1% SDS, 2% β-mercaptoethanol, and 35% glycerol) and boiled for 5 min. Primary lens cultures were prepared according to the procedure described previously (17Yin X. Jedrzejewski P.T. Jiang J.X. J. Biol. Chem. 2000; 275: 6850-6856Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar, 38Musil L.S. Beyer E.C. Goodenough D.A. J. Membr. Biol. 1990; 116: 163-175Crossref PubMed Scopus (298) Google Scholar). Briefly, chick lenses dissected from embryonic day 11 embryos were ruptured, and the cells that dissociated from the lenses were plated at 1 × 106 cells per 35-mm tissue culture plates in medium 199 plus 10% fetal bovine serum. Eight days after cell plating, primary cultures were treated either with 5 μm staurosporine dissolved in culture medium or with 2 mm EGTA for 4.5 h at 37 °C. When caspase-3 inhibitor Z-DEVD-FMK was used, cultures were preincubated with this inhibitor for 1 h before exposure to staurosporine or EGTA. CEF cells were plated at 1 × 105 cells in 35-mm tissue culture plates with DMEM plus 10% fetal calf serum and 5% CO2. The 2nd day after cell plating CEF cells were infected with high titer retroviruses (5 μl/dish) carrying either the wild type or the mutant Cx45.6 cDNAs. When CEF cells reached confluence, they were collected either for membrane preparation or for treatment with apoptosis-inducing reagents. CEF cells were treated in the presence of 1 mm staurosporine for 8 h at 37 °C. The isolated embryonic lens membranes were incubated in 20 μl of CKII reaction buffer (10 mmMgCl2, 50 mm MOPS, pH 7.0, 150 mmNaCl, and 20 μCi of [γ-32P]ATP) plus 2 μl of CKII (119 units/ml) at 30 °C for 2 h. The reaction was stopped by the inactivation of CKII by incubating at 56 °C for 30 min. Fifty microliters of caspase-3 reaction buffer were added to the in vitro phosphorylated lens membrane sample, and the reaction mixture was then divided into two tubes. In each tube, 3 μl of caspase-3 or caspase-3 storage buffer (50 mm HEPES, pH 7.4, 100 mm NaCl, 10 mm dithiothreitol, 1 mm EDTA, 0.1% CHAPS, and 10% (w/v) glycerol) was added and incubated at 37 °C for 6 h. The reaction was terminated by boiling the samples for 5 min in the presence of 0.6% SDS prior to the immunoprecipitation with affinity purified Cx45.6 antibody as previously described (31Jiang J.X. Goodenough D.A. Eur. J. Biochem. 1998; 255: 37-44Crossref PubMed Scopus (23) Google Scholar, 37Jiang J.X. Goodenough D.A. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 1287-1291Crossref PubMed Scopus (230) Google Scholar). To control for the immunoprecipitation assay, embryonic day 9 lenses were rinsed three times with culture medium deficient in methionine prior to the incubation with 1 ml of35S-labeled medium (methionine-free DMEM with 5% dialyzed fetal calf serum, 20 μm methionine, and 0.5 mCi of [35S]methionine) for 3 h. The radioactive medium was removed, and lenses were incubated in non-radioactive medium (DMEM supplemented with 0.5 mm methionine and 10% fetal calf serum) for various chase periods. 35S-Labeled lens membranes were isolated and treated at the identical condition as the 32P-labeled sample except the omission of ATP in CKII reaction mix, followed by the treatment by caspase-3, and immunoprecipitation as described above. Chick lenses at embryonic (E) days 8, 12, 15, and 18 and postnatal (P) days 1, 15, 30, and 60 were immunoblotted with affinity-purified Cx45.6 antibody that recognized the intracellular loop region of Cx45.6. The full-length Cx45.6 was phosphorylated in vivo and migrated as three close bands at the positions of 53, 56, and 58 kDa (Fig.1) as demonstrated previously (23Jiang J.X. White T.W. Goodenough D.A. Paul D.L. Mol. Biol. Cell. 1994; 5: 363-373Crossref PubMed Scopus (81) Google Scholar). At E8, a full-length Cx45.6 was solely present in the lens (Fig. 1,lane 1). Starting at E12, an additional band appeared migrating at around 46 kDa (empty arrowhead) (Fig. 1,lane 2) and its concentration level increased during the development of the lens as shown at E15 (Fig. 1, lane 3), E18 (Fig. 1, lane 4), and postnatal stages (Fig. 1,lanes 5–8). In addition, a 48-kDa fragment (solid arrowhead) closely migrating with the 46-kDa fragment appeared at E15, which is a lesser amount than 46-kDa fragment at embryonic stages. The level of this fragment, however, increased dramatically at the later developmental stages and reached a similar level as the 46-kDa fragment approaching P15 (Fig. 1, lane 6). No protein bands could be detected using pre-immune antibody (data not shown). To exclude the possibility that these cleavages may have occurred during sample preparation, we incubated lysates from embryonic and postnatal lenses at room temperature for 16 h and did not detect any accumulation of these fragments (data not shown). To determine if one of the closely migrated fragments was the phosphorylated form of another fragment, crude membranes from lens at P30 were treated with alkaline phosphatase. The levels of these two fragments were not altered (data not show), suggesting that neither of the fragments is the phosphorylated form of the other. The exterior portions of the 2-month postnatal chick lens containing epithelial cells and the fibers of the outer cortex were separated from the central core portion of the lens composed of mature lens fibers. The whole lens lysate contained both the full-length Cx45.6 and the cleaved fragments (Fig.2, lane 3, arrowheads). However, the cleaved fragments predominantly existed in the mature fibers of the central core (Fig. 2, lane 1, arrowheads) and to a much lesser extent in the outer cortex which, in contrast, was mainly composed of full-length Cx45.6 (Fig. 2, lane 2). Because cells are retained in the lens and new cells continually arise on the outside of the lens, most cells in the central core regions are differentiated mature lens fibers (1Rafferty N. Esson E. J. Ultrastruct. Res. 1974; 46: 239-253Crossref PubMed Scopus (39) Google Scholar). Consistent with the observation in Fig. 1 that truncated Cx45.6 fragments were accumulated associated with lens development, the results indicate that the specific cleavage of Cx45.6 in differentiated lens fibers may be related to the process of lens cell differentiation and formation of mature fibers. To identify the potential protease(s) involved in the cleavage of Cx45.6 during lens development, we first experimented with Ca2+-activated, cysteine protease calpain. This protease cleaves at the COOH terminus of Cx50, the ovine counterpart of Cx45.6 in the central region of the lens (32Lin J.S. Fitzgerald S. Dong Y. Knight C. Donaldson P. Kistler J. Eur. J. Cell Biol. 1997; 73: 141-149PubMed Google Scholar). Western blot of the membranes from embryonic (E) day 9 chick lens revealed a full-length Cx45.6 (Fig. 3 A, lane 2). Treatment by calpain led to a cleavage from a full-length Cx45.6 into a major 33-kDa fragment (Fig. 3 A, lane 3, open arrowhead), different from the 46-kDa fragment observed in the 2-month P chick lens (Fig. 3 A, lane 1, solid arrowhead). This result suggests that unlike ovine Cx50 calpain is unlikely involved in the in vivo truncation of Cx45.6 in chick lens. Caspase-3-like protease involved in apoptosis has been reported to be activated when lens epithelial cells differentiate into fibers in vivo (39Ishizaki Y. Jacobson M.D. Raff M.C. J. Cell Biol. 1998; 140: 153-158Crossref PubMed Scopus (260) Google Scholar), and caspase-3 expression is localized in the outer fibers (18Wride M.A. Parker E. Sanders E.J. Dev. Biol. 1999; 213: 142-156Crossref PubMed Scopus (76) Google Scholar). Caspase-3 is activated between embryonic days 8 and 12 in chick lens (18Wride M.A. Parker E. Sanders E.J. Dev. Biol. 1999; 213: 142-156Crossref PubMed Scopus (76) Google Scholar), which is at the similar time frame as when the cleavage of Cx45.6 was observed. To determine whether caspase-3 is the protease responsible for the cleavage, embryonic lens membranes were incubated with caspase-3. As shown in Fig. 3 B, after 2-h incubation, the full-length Cx45.6 was partially cleaved into fragments around 46 kDa (Fig. 3 B, lane 1, arrowhead), which migrated to the same position as the truncated Cx45.6 of the lens membrane from the newborn chick (Fig. 3 B, lane 5, arrowhead). After 16-h incubation with caspase-3, the majority of the full-length Cx45.6 proteins was cleaved (Fig. 3 B, lane 3, arrowhead). These results indicate that Cx45.6 may be a direct substrate for caspase-3. Alternatively, Cx45.6 can be cleaved by other protease(s) present in the lens membrane that could be activated by caspase-3. To examine these possibilities further, purified fusion proteins containing GST fused with portions of Cx45.6 were used for in vitro caspase-3 enzymatic digestion. The fusion parts of Cx45.6 include the following: the intracellular loop of Cx45.6 (GST-45.6L) (amino acids 98–148) (Fig. 3 C, lanes 3 and 4), partial COOH terminus (GST-45.6 F1) (amino acids 307–389) (Fig.3 C, lanes 5 and 6), and entire COOH terminus (GST-45.6F2) (amino acids 237–400) (Fig. 3 C, lanes 7 and8). These fusion proteins, together with the GST protein itself (Fig. 3 C, lanes 1 and 2), were treated with caspase-3. Both GST-45.6 F1 (Fig. 3 C, lane 6) and GST-45.6F2 (Fig. 3 C, lane 8) were cleaved by caspase-3 (arrowheads), whereas GST protein (Fig. 3 C, lane 2) and GST-45.6L (Fig. 3 C, lane 4) were not cleaved. These results suggest that Cx45.6 is a substrate of caspase-3 and that the cleavage site is located within the COOH terminus between amino acids 307 and 389. To examine whether Cx45.6 can be cleaved by caspase-3 in lens cells, primary lens cultures were incubated with apoptosis-inducing reagents staurosporine (Fig.4, lane 2) (22Andersson M. Sjöstrand J. Petersen A. Honarvar A.K.S. Karlsson J. Invest. Ophthalmol. Vis. Sci. 2000; 41: 2623-2632PubMed Google Scholar) or EGTA (Fig.4, lane 4) (40McGinnis K.M. Wang K.K.W. Gnegy M.E. J. Neurochem. 1999; 72: 1853-1863Crossref PubMed Scopus (78) Google Scholar). These two compounds are known chemical inducers for caspase-mediated apoptosis (22Andersson M. Sjöstrand J. Petersen A. Honarvar A.K.S. Karlsson J. Invest. Ophthalmol. Vis. Sci. 2000; 41: 2623-2632PubMed Google Scholar, 40McGinnis K.M. Wang K.K.W. Gnegy M.E. J. Neurochem. 1999; 72: 1853-1863Crossref PubMed Scopus (78) Google Scholar). Treatments by both reagents regenerated fragments around 46 kDa (Fig. 4, lanes 2 and 4, arrowhead). The cleavage was significantly inhibited by Z-DEVD-FMK, a caspase-3 inhibitor (Fig. 4, lane 5). The results confirm that caspase-3 is likely to be involved in the cleavage of Cx45.6 in vivo. The amino acid sequence Asp364-Glu-Val-Glu367-Gly was identified in the COOH-terminal region of Cx45.6 (see Fig.5 A). This sequence is similar to a known caspase-3 consensus sequence (41Talanian R.V. Quinlan C. Trautz S. Hackett M.C. Mankovich J.A. Banach D. Ghayur T. Brady K.D. Wong W.W. J. Biol. Chem. 1997; 272: 9677-9682Abstract Full Text Full Text PDF PubMed Scopus (775) Google Scholar) with the exception of substituting Asp with Glu at position 367. If cleavage by caspase-3 occurs between Glu367 and Gly368, the predicted molecular weight of one of the cleaved fragments would match that of the 46-kDa fragment. The two aspartate residues in the consensus sequence of caspase-3 are essential, and cleavage cannot occur if either residue is altered to alanine (42Barkett M. Xue D. Horvitz H.R. Glimore T.D. J. Biol. Chem. 1997; 272: 29419-29422Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar). Since glutamate like aspartate is an acidic residue, this segment of the sequence in Cx45.6 is a potential target for caspase-3. As illustrated in Fig.5 A, mutants were constructed with alterations in the conserved amino acid residues corresponding to the consensus sequence of caspase-3. Fusion proteins GST-Cx45.6F1, and its mutants GST-Cx45.6F1(D364A) and GST-Cx45.6F1(E367A), were subjected to caspase-3 cleavage in vitro (Fig. 5 B). GST-Cx45.6F1 was cleaved by caspase-3 into two closely migrating bands (Fig. 5 B, lane 2). Both GST-Cx45.6F1(D364A) (Fig. 5 B, lane 4) and GST-Cx45.6F1(E367A) (Fig. 5 B, lane 6) exhibited resistance to cleavage by caspase-3. These data suggest that either Asp364 or Glu367 seems to be crucial in the cleavage by caspase-3. The COOH terminus of Cx45.6 in GST fusion protein may not have similar conformation as compared with the one in the membrane-spanning full-length molecule. Therefore, full-length Cx45.6 and its mutants carrying alterations of conserved residues corresponding to caspase-3 (see Fig. 5 A) were expressed in CEF cells through retroviral infection, and the isolated cell membranes were subjected to caspase-3 treatment (Fig. 5 C). Similar to the results shown for fusion proteins, mutations in D364A and E367A prevented cleavage of Cx45.6 by caspase-3 (Fig. 5 C, lanes 4 and 6). Furthermore, the mutation of E367D, producing a cleavage sequence in Cx45.6 identical to the consensus sequence for caspase-3, increased cleavage (Fig. 5 C, lane 8). The ratio of the fragments around 46 kDa to the full-length form of Cx45.6(E367D) (Fig. 5 C, lane 8) is greater than that to the full-length wild type Cx45.6 (Fig.5 C, lane 2). Similar results were observed when CEF cells expressing wild type and mutant Cx45.6 were treated by the apoptosis-inducing reagent staurosporine (Fig. 5 D). Neither Cx45.6(D364A) nor Cx45.6(E367A) was cleaved (Fig. 5 D, lanes 4 and6), whereas the wild type Cx45.6 was (Fig. 5 D, lane 2). In comparison to the wild type Cx45.6, the cleavage of Cx45.6(E367D) was further enhanced (Fig. 5 D, lane 8). Together, the results obtained from fusion proteins and full-length Cx45.6 expressed in cells suggest that the cleavage site of Cx45.6 by caspase-3 is between Glu367 and Gly368. We have reported that Ser363, an amino acid residue close to the caspase-3 cleavage site, is phosphorylated in vivo by CKII (17Yin X. Jedrzejewski P.T. Jiang J.X. J. Biol. Chem. 2000; 275: 6850-6856Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar). Phosphorylation of certain substrates has been shown to be important in regulating their cleavage by caspase-3 (42Barkett M. Xue D. Horvitz H.R. Glimore T.D. J. Biol. Chem. 1997; 272: 29419-29422Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar, 43Walter J. Schindzielorz A. Grünberg J. Haass C. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 1391-1396Crossref PubMed Scopus (109) Google Scholar). To determine whether phosphorylation at Ser363 affects the ability of Cx45.6 to serve as a substrate for caspase-3, Cx45.6 in the membranes of embryonic lenses was treated in the presence of CKII and followed by proteolytic digestion by caspase-3 (Fig.6, lane 2). The results indicate that Cx45.6 phosphorylated by CKII was protected from caspase-3 digestion. To ensure that Cx45.6 in both full-length and truncated forms could be immunoprecipitated, embryonic lenses were metabolically labeled with [35S]methionine and chased in non-radioactive medium immediately (Fig. 6, lanes 3 and4) and after 6 h of incubation (Fig. 6, lanes 5 and 6). The isolated membranes labeled with [35S]methionine were treated in the presence (Fig. 6,lanes 4 and 6) and absence (Fig. 6, lanes 3 and 5) of caspase-3 and followed by immunoprecipitation. The Cx45.6 antibody we used here was capable of immunoprecipitating the full-length as well as the truncated (arrowhead) Cx45.6 cleaved by caspase-3. Together, the data suggest that the specific cleavage of Cx45.6 by caspase-3 is likely to be regulated by CKII-mediated phosphorylation. In this report, we identified caspase-3 as the protease responsible for the cleavage of Cx45.6 during chick lens development. Additionally, the cleavage on Cx45.6 was localized to the site close to the amino acid residue phosphorylated by CKII. Our previous studies (17Yin X. Jedrzejewski P.T. Jiang J.X. J. Biol. Chem. 2000; 275: 6850-6856Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar) have shown that Cx45.6 is phosphorylated in vivo by CKII at Ser363, and this phosphorylation may be a signal for Cx45.6 degradation mediated by proteasome. Here, we found that phosphorylation of Ser363 by CKII protected Cx45.6 from cleavage by caspase-3, suggesting that the specific cleavage of Cx45.6 is regulated by CKII-mediated phosphorylation. Based on our experimental evidence, a model is proposed for the involvement of CKII-mediated phosphorylation in the regulation of Cx45.6 turnover and specific cleavage by caspase-3-like protease. In the outer cortex of the lens, when a full-length Cx45.6 is phosphorylated by CKII at Ser363, it becomes unstable and possibly undergoes degradation mediated through the proteasome pathway (17Yin X. Jedrzejewski P.T. Jiang J.X. J. Biol. Chem. 2000; 275: 6850-6856Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar). The CKII phosphorylation site Ser363 is located within a PEST domain (17Yin X. Jedrzejewski P.T. Jiang J.X. J. Biol. Chem. 2000; 275: 6850-6856Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar, 44Liu Z.P. Galindo R.L. Wasserman S.A. Genes Dev. 1997; 15: 3413-3422Crossref Scopus (60) Google Scholar), and this domain has been reported to be associated with a rapid protein turnover via proteasome-mediated degradation (45Rechsteiner M. Rogers S.W. Trends Biochem. Sci. 1996; 21: 267-271Abstract Full Text PDF PubMed Scopus (1406) Google Scholar). Additionally, phosphorylation is also known to signal protein degradation by proteasome (46Lange C.A. Shen T. Horwitz K.B. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 1032-1037Crossref PubMed Scopus (388) Google Scholar). When Ser363 is not phosphorylated by CKII, Cx45.6 renders itself into a substrate for caspase-3, which cleaves at Glu367 at the COOH terminus of Cx45.6 and generates truncated fragments of Cx45.6. These fragments were localized to the central core region of the lens. We observed two fragments derived from Cx45.6 in the size of 46 and 48 kDa. These fragments were accumulated during lens development. The 46-kDa fragment was predominant during the embryonic stage, whereas the level of the 48-kDa fragment increased and reached a similar level as compared with the 46-kDa form at the postnatal stage. Formation of two fragments was also observed when GST-Cx45.6F1 was treated with caspase-3. However, mutation of a potential cleavage site in GST-Cx45.6F1 blocked the production of either fragment, suggesting that this site initiates generation of these two fragments. Because of the closeness of their migration, these two fragments sometimes were less distinctive on the SDS-PAGE. We also observed a potential third band larger than 48 kDa at P60 samples, and this band could be generated from additional cleavage or certain type of posttranslational modifications. Our data suggest that caspase-3-like protease is responsible for thein vivo cleavage of Cx45.6. In contrast to our observation in chicken, a previous study has shown that a large portion of the COOH terminus of Cx50, the ovine ortholog of Cx45.6, is removed by calpain (32Lin J.S. Fitzgerald S. Dong Y. Knight C. Donaldson P. Kistler J. Eur. J. Cell Biol. 1997; 73: 141-149PubMed Google Scholar). This discrepancy could account for the difference in expression level or accessibility of these two proteases expressed in these two species, because the responsibility of one protease could be replaced by the other during evolution. It is also likely that the cleaved fragment of Cx50 observed in ovine is caused by in vitroexperimental manipulation. We found if lysates of lens primary cells were handled improperly, Cx45.6 would be cleaved into a 33-kDa fragment, an identical size fragment as the fragment cleaved by calpain (data not shown). We have also shown that the truncated Cx45.6 was accumulated during lens development and was localized to the differentiated lens fibers. The activation of caspases-3 also initiates in the differentiated lens fibers and is involved in later stage cell differentiation, such as nucleus degeneration (18Wride M.A. Parker E. Sanders E.J. Dev. Biol. 1999; 213: 142-156Crossref PubMed Scopus (76) Google Scholar). Thus, the caspase-3-truncated Cx45.6 is likely to form a unique type of gap junctions, different from the ones formed by full-length Cx45.6. These intercellular channels may facilitate the passage of factors important for fiber differentiation and/or maintain the physiological functions of mature fibers. Functional gap junction channels are, indeed, formed by truncated Cx50 of ovine lens (32Lin J.S. Fitzgerald S. Dong Y. Knight C. Donaldson P. Kistler J. Eur. J. Cell Biol. 1997; 73: 141-149PubMed Google Scholar). These channels are characterized to be less sensitive to low pH, compared with those formed by full-length Cx50. Studies by Le and Musil (47Le A.N. Musil L.S. Dev. Biol. 1998; 204: 80-96Crossref PubMed Scopus (53) Google Scholar) have shown that lens primary cultures differentiate normally even if the channels are blocked with the gap junction inhibitor 18β-glycyrrhetinic acid. Since 18β-glycyrrhetinic acid has been tested for gap junctions formed solely by full-length connexins (48Guan X.J. Wilson S. Schlender K.K. Ruch R.J. Mol. Carcinog. 1996; 16: 157-164Crossref PubMed Scopus (163) Google Scholar), the channels consisting of truncated connexins may not be similarly inhibited by this chemical. Lens Cx45.6 is the first connexin identified as a substrate for caspase. In addition to lens fibers other cell types, such as skin keratinocytes, also undergo degeneration of nuclei and organelles during cell differentiation. Caspase activation has been identified in the differentiation of epidermal keratinocytes and is required for the loss of the nucleus (49Weil M. Raff M.C. Braga M.M. Curr. Biol. 1999; 9: 361-364Abstract Full Text Full Text PDF PubMed Scopus (176) Google Scholar). It is of great interest to determine whether connexins in those cells are also subjected to regulation by caspases. The physiological significance of the regulation of connexins by caspases requires further investigation. We thank D. Adan-Rice and L. Wang for technical assistance and Z. Dong for valuable discussions. We thank members of the Jiang laboratory for critical reading of the manuscript.

Referência(s)