| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 35, 36931-36942, August 27, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the
Institute of Anatomy and ||Cell Biology, University Hospital Essen, 55 Hufelandstrasse, 45122 Essen, Germany, Department of Physiology, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Canada, ¶Institute of Genetics, University of Bonn, 53117 Bonn, Germany, and **Laboratoire d'Oncologie Virale et Moleculaire, UFR de Biochimie, Université Paris, 75005 Paris, France
Received for publication, April 13, 2004 , and in revised form, May 25, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
- and 
-tubulin (18), and c-Src (19)). So far, the functional consequence of these protein-protein interactions is mostly unknown. McLeod et al. (20) demonstrated that two members of the CCN gene family, CCN1 (Cyr61) and CCN3 (NOV: nephroblastoma overexpressed) were up-regulated in Cx43-transfected C6 glioma cells. The CCN family now consists of six members, which belong to a group of structurally related, secreted, extracellular matrix-associated proteins (21). These proteins are characterized by an N-terminal secretory signal followed by four domains with different functions (for a review, see Refs. 22-24). The C terminus is able to bind target molecules like fibulin-1c (25), Notch-1 (26), integrins (27, 28), and Rpb7, a subunit of RNA polymerase II (22). All members of the CCN family are involved in diverse fundamental biological processes like proliferation, differentiation, adhesion, migration, and angiogenesis and are expressed in a number of tissues such as kidney, heart, liver, and spleen. Although these secretory proteins show a high structural similarity, they seem to exhibit different cellular functions. NOV may be involved in cell cycle control, because NOV expression induces growth inhibition in embryonic chicken fibroblasts (29, 30). Moreover it has been shown that less aggressive glioma and astrocytoma express high levels of NOV, which suggests an antiproliferative effect of NOV (31).
In the present study, we have investigated the possibility of growth reduction of the malignant trophoblast cell line Jeg3 upon connexin induction. This communication-deficient cell line (32, 33) serves as a model for the early placental trophoblast (34). Since Cx40 and Cx43 are involved in the regulation of proliferation (35) and differentiation (36-38) in human trophoblast cells, we have transfected this tumor cell line with these tissue-specific gap junction channels. We could demonstrate that Cx43, but not Cx40 and a Cx43 construct lacking most parts of the C terminus (trCx43), is able to reduce cell growth in vitro and tumor growth in nude mice. Upon induction of Cx43, we revealed an up-regulation of the growth-regulatory gene NOV. This up-regulation was accompanied by a change in localization of NOV from the nucleus to the membrane, where NOV colocalizes with Cx43. Furthermore, we could confirm a direct interaction of NOV with the carboxyl terminus of Cx43, both in vitro and in transfected cells.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Animals
Athymic male nude mice (Han:NMRI nu/nu) were obtained from the animal facility of the University Hospital, Duisburg-Essen, Germany. The mice were maintained in a pathogen-free barrier unit with regulated light/dark cycles (12 h/12 h). Mice had access to food and water ad libitum. The treatment of the animals was in accordance with the German law for animal protection and with permission of the state.
Plasmid Construction and Transfection
For construction of the plasmids based on the tet-on system, two different expression vectors containing rat Cx40 (pPH3) and rat Cx43 (pPH4) were used (32, 39). The coding region for rat Cx40 was excised from the vector pPH3 by HindIII/XbaI and subcloned into the SacII site of the rtTA-responsive plasmid pUHD10-3 (40), creating the plasmid pUHD/Cx40. The cDNA coding region for rat Cx43 was excised from plasmid pPH4 by EcoRI and subcloned into the EcoRI site of pUHD10-3, creating the construct pUHD/Cx43. For construction of the plasmid pUHD/trCx43, an expression vector containing trCx43 without most of the C-terminal region (pBEHpac18/Cx43 truncated, mutation at amino acid 258 (lysine) into a stop codon; C terminus of Cx43: amino acids 227-382) was used. The mouse Cx43 fragment was excised by EcoRI/BamHI and subcloned into the EcoRI/BamHI site of pUHD10-3, creating the construct pUHD/trCx43. The sequence and orientation of the inserts were analyzed by sequencing (MWG Biotech, Germany). To establish Cx40-, Cx43-, and truncated Cx43-inducible cell clones, the Jeg3 cell line was first stably transfected with the plasmid pUHD172-1neo (40), containing both a gene encoding rtTA and a gene encoding G418 resistance. Jeg3 cells (6 x 104) were transfected with 1.5 µg of pUHD172-1neo DNA using LipofectAMINE (Invitrogen) according to the recommended protocol and selected in medium supplemented with 500 µg/ml G418. Stable clones were screened for rtTA activity by transient transfection with the reporter plasmid pUHC13-3 (40). Cell clones expressing high levels of the rtTA gene and exhibiting only basic activity of the tet-on system (data not shown) were cotransfected with the connexin expression vectors pUHD/Cx40, pUHD/Cx43, and pUHD10-3/trCx43, respectively, and a plasmid encoding puromycin resistance (pBEHpac18) (41). Therefore, 6 x 104 Jeg3 cells were cotransfected with 1.5 µg of expression vector DNA and 1.5 µg of pBEH-pac18. In addition, Jeg3 cells were cotransfected only with pUHD10-3 and pBEHpac18 as the vector control. Cells were selected in medium supplemented with 0.5 µg/ml puromycin (Sigma). Resistant clones were isolated after 3 weeks.
Cells were initially screened for the appropriate connexin expression by immunocytochemistry following 48-h induction with 1 µg/ml doxycycline HCl (Dox)1 (Sigma).
Jeg3 Tet transfectants were cultivated in medium containing 500 µg/ml G418 sulfate and 0.5 µg/ml puromycin.
For the transfection of parental Jeg3 cells with NOV, the expression vector pCMV-NOVH (S) containing a G418 resistance gene was used (42). Transfection was performed as described above, and cells were selected in medium containing 500 µg/ml G418 sulfate.
Northern Blot Analysis
Total RNA was isolated from confluent cell monolayers and from solid tumors frozen in liquid nitrogen using a Qiagen RNeasy Kit (Qiagen). RNA samples (5 µg) were separated on a denaturing agarose gel before being transferred to a nylon membrane (Hybond-N; Amersham Biosciences). Hybridization at 42 °C and labeling of the specific probes with [-32P]dCTP by using the random primed labeling system (Amersham Biosciences) were performed as described previously (43). The following probes were used: rat Cx40 cDNA fragment (44) and rat heart Cx43 cDNA fragment (45). As a control, the blots were rehybridized with a glyceraldehyde-3-phosphate dehydrogenase-specific probe (GAPDH) (46).
Immunofluorescence and Microscopy
Indirect immunocytochemistry on cells and solid tumor sections was performed as described previously (47). The following primary antibodies were used: anti-Cx43 rabbit polyclonal antibody (1:300) (48), mouse monoclonal Cx43 antibody (1:100; Zymed Laboratories Inc.), mouse monoclonal Cx43-NT1 antibody (1:100; Fred Hutchinson Cancer Research Center, Seattle, WA), anti-Cx40 rabbit polyclonal antibody (1:200) (48), and rabbit polyclonal antibody against NOV (1:50) (42). The following appropriate secondary antibodies were used: fluorescein isothiocyanate-conjugated swine anti-rabbit and goat anti-mouse IgG (1:40; DAKO, Germany) and Cy3-conjugated goat anti-rabbit IgG (1:300; Dianova, Germany).
For double immunolabeling of connexin and NOV, the same protocol was used as described above, but both primary antibodies were added successively.
After immunolabeling of NOV, the DNA-specific dye 4',6'-diamidino-2-phenylindole hydrochloride (0.1 µg/ml; 15 min at 37 °C) was used to counterstain the nucleus of Jeg3 cells.
Reverse Transcriptase (RT)-PCR
2 µg of total RNA samples were digested with DNase I (Invitrogen) according to the protocol to prevent contamination with DNA. For reverse transcription, first strand synthesis was carried out using oligo(dT)12-18 primer (MWG Biotech) and Moloney murine leukemia virus reverse transcriptase (Invitrogen). 4 µl of the RT reaction were used for the following PCR experiments. For a semiquantitative measurement, -actin was used as an internal control. The PCR was performed using primers specific for NOV and -actin as shown in Table I. PCR was carried out in a 50-µl volume using BioTherm Taq polymerase (Genecraft, Germany). The PCR was performed for 40 cycles of 1-min denaturation at 94 °C, 1-min annealing at 59 °C, and 1.5-min elongation at 72 °C. The PCR amplification was followed by a 10-min final extension at 72 °C. All experiments were carried out up to four times in a thermocycler (Biometra, Germany). The PCR amplification products were electrophoresed on a 2% agarose gel stained with ethidium bromide. Densitometric analysis was performed using a Gel Imager (Intas, Germany). The expression of NOV was normalized to the expression of -actin and quantified with Gelscan Professional version 4.0 software. The gene expression data were analyzed for statistical significance by Student's t test. A p value of 
0.05 was considered to be significant.
|
cDNA synthesis and synthesis of biotinylated cRNA were performed as described previously (49). Fragmentation of cRNA, hybridization to HG_U95Av2 microarrays (Affymetrix) with 6000 genes and 6000 expressed sequence tags, washing, and staining as well as scanning of the arrays in a GeneArray scanner (Agilent) were carried out as recommended in the Affymetrix Gene Expression Analysis technical manual. Three cell clones per connexin construct were analyzed on an individual gene array. Scaling across all probe sets of a given array to an average intensity of 1000 units was included to compensate for variations in the amount and quality of the cRNA samples and other experimental variables. Absolute and comparison analyses (single pairwise and cross-pairwise comparisons) were performed using the Affymetrix Microarray Suite 5.0 software. In the absolute analysis, the signal intensities and detection calls of the analyzed genes were determined. In the comparison analysis, the alteration of gene expression (signal log ratio and -fold change, respectively) and the change call were identified. The "cut-off" level in -fold change values was defined as 1.5 and 2.5 (cross-pairwise comparisons, experimental groups + versus + Dox) using the Affymetrix Data Mining Tool 3.0 Software (DMT 3.0). The -fold change level of 1.5 represents a generally accepted -fold change difference for oligonucleotide microarray analysis (50).
The on-line supplemental material is comprised of data lists of upand down-regulated genes in Jeg3 connexin transfectants analyzed by gene arrays (supplemental Tables I and II).
In Vitro and in Vivo Coimmunoprecipitation
Plasmid Construction—pUHD/Cx43 and pCMV-NOVH (S) were used as a template for PCR to amplify Cx431-382, Cx43
1-257, Cx43CT257-382, and NOV1-357. The 5'-primer contained an EcoRI site and ATG sequence, and the 3'-primer contained a TAG sequence and a SalI site. For in vitro coimmunoprecipitation, the PCR products were digested with EcoRI and SalI. Cx43 fragments were ligated into the EcoRI and XhoI sites within pcDNA3 (Invitrogen) to form the pcDNA Cx43 constructs. For in vivo coimmunoprecipitation, the Cx43-amplified fragments were inserted into pFlag-CMV2 vector (Sigma) at EcoRI and SalI. For both experimental approaches, the NOV fragment was ligated into the EcoRI and XhoI sites within pcDNA-His6 (Invitrogen) at EcoRI and XhoI to form the pHis-NOV construct.
All of the constructs have been sequenced to confirm the proper reading frame and cDNA sequences for Cx43 and NOV.
In Vitro Coimmunoprecipitation—pcDNA Cx43 constructs were synthesized with the TNT reticulocyte lysate (Promega) in the presence of [35S]methionine. pHis-NOV was synthesized without [35S]methionine. The translation products of Cx43 and NOV were incubated together with anti-His antibody (H15; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and protein A-Sepharose 4B (Zymed Laboratories Inc.) in 700 µl of EBC buffer (50 mM Tris (pH 8), 120 mM NaCl, 0.5% Nonidet P-40) followed by five washes with NETN buffer (20 mM Tris (pH 8), 100 mM NaCl, 0.5% Nonidet P-40, 1 mM EDTA). The proteins bound to the Sepharose beads were separated on 12% SDS-PAGE and detected by autoradiography. Immunoblotting was conducted with anti-His antibody to confirm the translation of His-tagged NOV.
In Vivo Coimmunoprecipitation—293T cells were transient transfected with the FLAG-Cx43 expression vectors and pHis-NOV. Whole cell lysate were collected in NETN buffer (see above). Immunoprecipitations were performed using M2 Sepharose beads (Amersham Biosciences) incubated with 800 µl of 1 mg/ml whole cell lysate at 4 °C for 2 h. After the beads were thoroughly washed with NETN buffer, associated proteins were eluted with FLAG peptide (Sigma) for 30 min on ice and separated on 10% SDS gel. Proteins were transferred to polyvinylidene difluoride membrane and blotted with anti-FLAG and anti-His.
Dye Transfer Assay
This assay is based on transfer of calcein from preloaded donor cells to recipient acceptor cells (51). Twenty-four hours before the experiment, the acceptor cells were seeded into 6-well plates (5 x 105 cells/well) to obtain confluent cultures. For loading the donor cells with the membrane-permeable dye calcein AM (C-3100; Molecular Probes, Inc., Eugene, OR) the cells were trypsinized, resuspended in 2 ml of staining solution (5 µM calcein AM in phosphate-buffered saline with 0.2% glucose), and agitated for 30 min at 37 °C in an incubator. After three washes with phosphate-buffered saline, the cell number was measured in a cell counter (CASY 1; Schaerfe System, Germany). Loading of the acceptor cells with the permanent membrane dye DiI (C-7000; Molecular Probes) was done directly in cell culture dishes using a solution of 5 µM DiI in phosphate-buffered saline with 0.2% glucose for 15-30 min. In each experiment, 9 x 104 donor cells were added to a confluent monolayer culture of DiI-stained acceptor cells. After an incubation time of 4 h at 37 °C, these cocultures were analyzed by flow cytometry. Up to three independent experiments of either combination were performed. Cells were analyzed on a FACScan flow cytometer (BD Biosciences) using Cell Quest software. The fluorescence intensity of the cells was recorded with an argon laser set up for an excitation wavelength at 488 nm. Each sample was measured up to three times with a total of 10,000 cells. Samples were analyzed by gating on the acceptor cell population, characterized by an intermediate to high forward scatter and low to intermediate side scatter and bright staining for DiI. The percentage of calcein-positive (communicating) acceptor cells as compared with the respective negative control represented by the DiI-stained acceptor cell population in the absence of donor cells was determined. A mixture of calcein-stained, DiI-stained, and unstained cells were used to optimize signal detection and to compensate for overlap of calcein fluorescence with the DiI channel.
Results are reported as the mean ± S.D. Levels of significance were determined at the p value of 0.05 using Student's t test.
In Vitro Proliferation Assay
For analysis of cell proliferation, the Jeg3 Tet transfectants and the NOV-transfected Jeg3 cells were plated at a density of 5 x 103 cells/well in 12-well plates in triplicate and were counted at day 2, 5, 7, and 10 after plating using a CASY cell counter (Casy1; Schaerfe System). The results were analyzed for statistical significance by Student's t test. A p value of 0.05 was considered to be significant.
Tumorigenicity Assay
Athymic male nude mice (Han:NMRI nu/nu) were subcutaneously inoculated with 2.5 x 106 cells/flank in 200 µl of isotonic NaCl solution. 6-9 animals/cell clone were analyzed. For a direct comparison, each nude mouse received an injection with the cell transfectant in the right flank and the vector control in the left flank. Induction of connexin expression in the Cx Tet tumors was maintained by feeding animals with sucrose water containing Dox (200 µg/ml Dox, 3% sucrose; protected from light degradation by dark feeder bottles) and by injection of 1 mg of Dox/mouse in isotonic NaCl solution every third day. The water was exchanged every 3 days.
Tumor volume was measured every 3 days using calipers in two perpendicular dimensions and was calculated assuming a spherical shape: tumor volume = 4/3(
)abc (a, b, c = half-axis). The tumors were harvested after 3 weeks. Statistical significance of the obtained data was analyzed by Student's t test. A p value of 0.05 was considered to be significant. Immediately after removal, the tumor tissues were frozen in liquid nitrogen, and 7-µm cryostat sections of the tumors were prepared for immunohistochemistry.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
0.05) of cell coupling after induction of the connexins. The percentage of cell coupling in the untreated cell clones derived from the small leak of the tet promoter in the absence of Dox (52). This small leak of connexin expression in the uninduced cell clones could be verified using RT-PCR analysis, demonstrating at least some connexin mRNA after 36 cycles (data not shown). The analysis of calcein transfer also confirmed the functionality of the C-terminal truncated Cx43 channel and showed that the truncated Cx43 channel is able to exchange the dye after treatment with Dox (Table II). The functionality of this mutated Cx43 construct has already been demonstrated by calcein transfer in HeLa cell transfectants.2 The trCx43 HeLa transfectants showed similar levels of calcein transfer as HeLa cells transfected with wild type Cx43.
|
The increase in intercellular coupling could be correlated with the induction of connexin expression on transcript and protein level (compare Fig. 1 with Table II).
Cx43, but Not Cx40 or trCx43, Reduces in Vitro Proliferation of Jeg3 Cells—To determine whether the expression of different connexins alters the cell growth properties of Jeg3 Cx transfectants, we analyzed their in vitro cell proliferation (Fig. 2). This investigation displayed different proliferation properties between the Cx transfectants. As shown in Fig. 2A, the Cx43 cell clones revealed a significantly reduced cell growth (p 0.05) upon induction compared with uninduced cells from day 7 after plating onwards. The proliferation rate varied between the clones. Whereas clone #36 showed a clear reduction in proliferation already on day 5 of cultivation, clone #20 exhibited only moderate reduction in proliferation after 1 week. The Dox-induced cell clones grew also at a significantly slower rate than the vector control cells at day 10 (Fig. 2D, clone #36). In contrast, Jeg3 cells transfected with Cx40 demonstrated no differences in proliferation after induction with Dox versus untreated cells (Fig. 2B). The Cx40 cell clones exhibited the same proliferation characteristics as the vector control represented in Fig. 2E. Interestingly, similar results were obtained with the truncated Cx43 cells (trCx43), which showed no reduction of cell growth after Dox treatment (Fig. 2C). Thus, in contrast to Cx40 and trCx43 gap junctions, Cx43 channels significantly reduce cell proliferation in Jeg3 cells.
|
|
67% when compared with the vector control tumors. In contrast, the Cx40 (Fig. 3G) as well as the trCx43 cell clones (Fig. 3H) revealed no reduction of tumor growth. Mostly, the Cx40 and trCx43 tumor cells developed even larger tumors; especially the trCx43/#16 (Fig. 3H) seemed to support tumor growth. Further analyses of additional trCx43 clones missing the C terminus of Cx43 are needed to verify this phenomenon.
To summarize, the reduced tumorigenic potential of Jeg3 cells transfected with Cx43, but not with Cx40 and trCx43, could be confirmed not only in analysis of cell monolayers but also in three-dimensional growth experiments after subcutaneous injection into nude mice.
NOV Shows an Up-regulation in Cx43-transfected Jeg3 Cells—To identify genes responsible for the different cell physiological behavior of the connexin transfectants, gene array analysis was performed. Therefore, absolute and comparison analyses of the cDNA expression pattern of Jeg3 Tet Cx transfectants (3 cell clones/Cx construct) before and after Dox treatment were generated. Beside single pairwise comparisons of each cell clone, induced versus uninduced, we established cross-pairwise comparisons (see Supplementary Data). In the comparative gene array analysis of the Dox-induced versus uninduced Cx transfectants, we identified an average of 54 genes that were differently regulated at least 1.5-fold. A total of 37 genes were up-regulated, and 17 genes were down-regulated in the induced versus uninduced connexin cell clones.
We focused on the expression pattern of those genes using the obtained gene array data, which are already described in previous published results (20, 31). Among those genes, we could identify a 1.5-fold up-regulation of the NOV transcript in the Dox-induced Cx43 transfectants versus uninduced cells. In contrast, the Cx40 cell clones revealed a slight down-regulation with a median of 1.4-fold, whereas the trCx43 transfectants displayed no clear regulation of NOV. Moreover, a mean 2.8-fold increased expression in signal intensity of NOV in the Cx43 induced transfectants versus uninduced cell clones was found (data not shown).
Using semiquantitative RT-PCR analysis before and 48 h after Dox treatment of three cell clones of each Cx transfectant, we were able to confirm the results of the gene arrays (Fig. 4). NOV expression was significantly increased in induced Cx43 transfectants compared with the uninduced clones and the Cx40- and trCx43-transfected cells. The Cx40 transfectants showed only in part a weak but no significant down-regulation of NOV expression (see Cx40/#13), whereas in the trCx43 transfectants no regulation could be found. In addition, the expression of NOV in the induced Cx43 transfectants was also significantly increased compared with parental Jeg3 cells (see Fig. 9A). Thus, the increased expression of NOV in Cx43-transfected Jeg3 cells obtained by gene array analysis could be verified by RT-PCR.
|
|
|
To analyze the correlation between Cx43 and NOV protein disappearance from the cell membrane, we performed time courses between 0 and 48 h after Dox removal. The immunocytochemistry revealed that up to 14 h after Dox removal, a colocalization of NOV and Cx43 at various points along the lateral cell membranes between adjacent cells could still be observed (Fig. 6A). After 24 h, however, Cx43 and NOV have disappeared from the membranes, and only nuclear and cytoplasmic staining of NOV was found. These observations correlated to the results from RT-PCR analysis of Cx43 (Fig. 6B) and NOV mRNA expression (Fig. 6C). The NOV transcripts decreased simultaneously with the reduction of Cx43 expression during this time course. Whereas the Cx43 expression decreased rapidly to 0 after 48 h, the NOV expression declined to a basic level that corresponded to the nuclear and cytoplasmic staining and the loss of membrane staining (Fig. 6, compare A and C).
|
Direct interaction of NOV with the C terminus of Cx43 has been confirmed using an in vitro coimmunoprecipitation assay. A Cx43 construct with wild type Cx43 (Cx431-382) and two mutant constructs, C-terminal truncated Cx43 (Cx431-257, trCx43) and a construct with most of the C terminus (Cx43CT257-382), were generated (Fig. 7A), and a His-tagged NOV construct was used. In the in vitro translation procedure, the Cx43 constructs were labeled with [35S]methionine. The pHis-NOV was synthesized without [35S]methionine. As shown in Fig. 7B (c), lane 1, wild type Cx43 (Cx431-382) as well as the Cx43 mutant with only the C-terminal tail (Cx43CT257-382, lane 3) coimmunoprecipitated with full-length NOV, whereas C-terminal truncated Cx43 (Cx431-257, lane 2) failed to coprecipitate with NOV. These findings confirmed that the direct interaction of NOV and Cx43 requires the carboxyl terminus of Cx43 and does not need any further cofactor molecule.
|
1-257) did not show a binding with NOV (Fig. 8).
|
0.05) in cell proliferation compared with parental Jeg3 cells after 9 days of culture (Fig. 9D). Whereas the parental Jeg3 cells demonstrated a strong proliferative activity from day 4 of culture onward, the NOV-transfected Jeg3 clones first started to proliferate at day 7 but never reached the proliferative capacity of the parental Jeg3 cells. In conclusion, these findings showed that NOV itself exhibits an antiproliferative activity on Jeg3 cells. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Our results further revealed that only Cx43, but not Cx40 or trCx43, is able to reduce cell growth in vitro and tumor growth in vivo, which points out that the carboxyl terminus of Cx43 is essential to achieve growth reduction capacities in these tumor cells. In addition, this finding verifies isoform-specific functions of the connexins that could be correlated with the observation in the human placenta (35-38). Experiments with a similar C-terminal truncated Cx43 mutant (Cx43-256M) as used in this study also exhibited no change in cell proliferation in neuroblastoma cells (14). The results of the above mentioned proliferation behavior of the different transfectants strengthen the finding that the C terminus of Cx43 mediates the different signal cascade leading to the growth reduction of Cx43 transfectants. Among all differently regulated genes obtained from gene array data, NOV showed an up-regulation in Cx43 transfectants by gene array analysis and RT-PCR. NOV, a member of the CCN family of growth regulator proteins, is expressed in many tissues and has been shown to regulate several cellular functions like proliferation, adhesion, and differentiation (22-24, 56). Interestingly, NOV was already found to be up-regulated in C6 glioma cells after transfection with Cx43 (20). These Cx43 transfectants also revealed a reduced cell proliferation in vitro (57). As in glioma cells (20, 31), in choriocarcinoma cells NOV and Cx43 were colocalized at the cell membrane. Uninduced or parental Jeg3 cells expressed NOV, but localization was exclusively found in the nucleus and in perinuclear regions and to a lesser extent in the cytoplasm. Upon induction of the different connexins, only in the Cx43 transfectants a switch in NOV expression pattern from the nucleus to the membrane was observed. The finding that association with NOV was not present in Cx40 and in C-terminal truncated Cx43 transfectants enforces the assumption that the C terminus of Cx43 is involved in binding NOV. The shift in NOV expression pattern upon Cx43 induction could be validated by the time course experiments in induced Cx43 transfectants after Dox removal. By 24 h after Dox removal, NOV had disappeared from the membrane and was localized in the nucleus and in the cytoplasm. In parallel, there is a simultaneous decrease in NOV and Cx43 mRNA expression during this time course. Thus, Cx43 and NOV up-regulation and membrane localization are clearly dependent on the presence of Cx43. The function of NOV seems to be dependent on its structure and subcellular localization. It is hypothesized that an N terminus-truncated form of NOV, which is accumulated in the nucleus, leads to a growth stimulation, whereas the full-length protein, which is secreted or remains at the cell membrane, inhibits proliferation (22-25). The N-truncated NOV isoform was detected in the nuclei of HeLa and osteosarcoma cells (58). Since we found that the expression of NOV in parental Jeg3 cells, Cx40, trCx43, and uninduced Cx43 transfectants is predominantly restricted to the nucleus, we must consider whether different NOV isoforms may play also a role in our system.
Experiments with NOV-transfected Jeg3 cells revealed that the transfected full-length NOV is located exclusively at the lateral cell membrane, similar to the observation in Cx43 transfectants after induction of Cx43, and induces growth reduction in Jeg3 cells independent from Cx43 expression. These findings correlated with the data of the Cx43 transfectants, which displayed also a reduced cell proliferation and showed that NOV itself exhibited an antiproliferative effect on Jeg3 cells. Reduced proliferative activity of NOV has already been described in chicken embryonic fibroblasts (29). Gupta et al. (31) could demonstrate that transfection of glioma cells with full-length NOV alone had antiproliferative effects on cell proliferation and tumor growth in nude mice. Furthermore, our findings point to the fact that NOV can act independent of the presence of Cx43; however, we suggest that the Cx43 protein seems to be able to increase full-length NOV expression and to change its location from the nucleus to the membrane by binding NOV. By in vitro and in vivo coimmunoprecipitation, we were able to confirm the direct interaction of NOV with the C terminus of Cx43 in the region between 257 and 382 amino acids. This binding is apparently independent from the requirement of any cofactor molecule. Direct interactions between the C-terminal region of Cx43 and the tight junction-associated protein ZO-1 (16), - and -tubulin (18), and c-Src (19) have already been demonstrated. It is well known that the last five amino acids of the C terminus of Cx43 (DDLEI) interact with the second PDZ domain of ZO-1 (16). The binding of -tubulin occurred in the 35-amino acid juxtamembrane region of the Cx43 tail, which contains a presumptive tubulin binding motif (18). Other interactions of Cx43 with -catenin (59) and caveolin-1 (60) could be observed by colocalization and immunoprecipitation. Since NOV possesses binding motives such as (T/S)XV (61) for binding PDZ domain proteins, it could therefore interact via such binding sites with potential PDZ domains in the C terminus of Cx43 (60). In the opposite way, NOV could bind to the C-terminal region of Cx43 via own presumptive PDZ or Src homology 2/Src homology 3 domains.
Our results has identified not only a new interaction partner of Cx43, the growth regulator NOV, but this binding between NOV and Cx43 is associated with a growth reduction in Jeg3 cells. The disappearance of NOV from the nucleus may cause NOV to no longer be available as a transcriptional regulator. This mechanism could be responsible for the change in cell physiology of the Cx43-expressing Jeg3 cells, especially the decrease in proliferation rate and the reduction in tumor growth. Regarding changes in cell cycle, it could be demonstrated in several publications that Cx43 inhibits cell growth by delaying exit from the G1/G0 phase of the cell cycle (62-64). Up to now, we do not have any hints of a clear change in expression of cell cycle genes at transcription levels using gene array analysis (see Supplementary Data).
There is accumulating evidence that growth regulation seems not only dependent on channel properties but on the C terminus-induced signal cascades by interacting with other proteins in an isoform- and cell type-specific manner. In conclusion, this paper shows for the first time that Cx43 seems to be responsible to regulate cell growth in Jeg3 cells via the up-regulation of the growth regulator NOV accompanied by a shift of localization of NOV from the nucleus to the membrane and a direct binding of NOV to the C terminus of Cx43.
|
FOOTNOTES |
|---|
The on-line version of this article (available at http://www.jbc.org) contains supplemental Tables I and II.
To whom correspondence should be addressed. Tel.: 49-201-723-4387; Fax: 49-201-723-5635; E-mail: e.winterhager{at}uni-essen.de.
1 The abbreviations used are: Dox, doxycycline HCl; RT, reverse transcriptase.
2 K. Maass, personal communication.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
