This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Handa, K.
Right arrow Articles by Kiyono, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Handa, K.
Right arrow Articles by Kiyono, T.

 Previous Article  |  Next Article 

Journal of Virology, February 2007, p. 1379-1389, Vol. 81, No. 3
0022-538X/07/$08.00+0     doi:10.1128/JVI.01712-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

E6AP-Dependent Degradation of DLG4/PSD95 by High-Risk Human Papillomavirus Type 18 E6 Protein{triangledown}

Keisuke Handa,{dagger} Takashi Yugawa, Mako Narisawa-Saito, Shin-ichi Ohno, Masatoshi Fujita, and Tohru Kiyono*

Virology Division, National Cancer Center Research Institute, 5-1-1 Tsukiji, Chuo-ku, Tokyo, 104-0045, Japan

Received 8 August 2006/ Accepted 13 November 2006


arrow
ABSTRACT
 
In most cervical cancers, DNAs of high-risk mucosotropic human papillomaviruses (HPVs), such as types 16 and 18, are maintained so as to express two viral proteins, E6 and E7, suggesting that they play important roles in carcinogenesis. The carboxy-terminal PDZ domain-binding motif of the E6 proteins is in fact essential for transformation of rodent cells and induction of hyperplasia in E6-transgenic mouse skin. To date, seven PDZ domain-containing proteins, including DLG1/hDLG, which is a human homologue of the Drosophila discs large tumor suppressor (Dlg), have been identified as targets of high-risk HPV E6 proteins. Here, we describe DLG4/PSD95, another human homologue of Dlg, as a novel E6 target. DLG4 was found to be expressed in normal human cells, including cervical keratinocytes, but only to a limited extent in both HPV-positive and HPV-negative cervical cancer cell lines. Expression of HPV18 E6 in HCK1T decreased DLG4 levels more strongly than did HPV16 E6, the carboxy-terminal motif of the proteins being critical for binding and degradation of DLG4 in vitro. DLG4 levels were restored by expression of either E6AP-specific short hairpin RNA or bovine papillomavirus type 1 E2 in HeLa but not CaSki or SiHa cells, reflecting downregulation of DLG4 mRNA as opposed to protein by an HPV-independent mechanism in HPV16-positive cancer lines. The tumorigenicity of CaSki cells was strongly inhibited by forced expression of DLG4, while growth in culture was not inhibited at all. These results suggest that DLG4 may function as a tumor suppressor in the development of HPV-associated cancers.


arrow
INTRODUCTION
 
Infection with specific human papillomaviruses (HPVs), such as HPV type 16 (HPV-16) and HPV-18, is a major risk factor for human cancer of the uterine cervix. From the facts that E6 and E7 genes are expressed almost exclusively in cervical cancer cells and that they can inactivate the tumor suppressors p53 and Rb, respectively, they are believed to play key roles in carcinogenesis in the cervix. Accumulating lines of evidence suggest that the PDZ domain-binding motif is particularly important for transformation and tumorigenesis in cultured cells (16, 42) and hyperplasia and carcinogenesis in E6-transgenic mice (30). To date, several PDZ domain-containing proteins have been identified as E6 targets, namely, mammalian homologues of Dlg (DLG1/hDLG) and Scribble (Scrib/Vartul), MUPP1, MAGI-1, -2, and -3, and GIPC (5, 9, 16, 20, 21, 27). Both Dlg and Scribble function as cortical tumor suppressors in Drosophila and are essential for cortical localization of Lgl, which is required for asymmetric cortical localization of basal determinants, such as Prospero and Numb, in mitotic neuroblasts (32, 33). In mammalian cells, their functions are less well characterized, although some reports suggest that DLG1 may possess tumor suppressor activity (13, 23). Dlg-related proteins are associated with the cortical actin cytoskeleton and appear to have both structural and functional roles. Several homologues of Dlg have been identified in humans. Among them, four are closely related to Dlg, namely, DLG1 (hDLG, SAP97), DLG2 (PSD93, Chapsyn-110), DLG3 (NEDLG, SAP102), and DLG4 (PSD95, SAP90). DLG1, -2, -3, and -4 have 56 to 60% amino acid identity to Dlg and have 70 to 75% amino acid identity to each other, while DLG5 (P-DLG), which is highly expressed in placenta and prostate, has 40%, and DLG6 (MPP4), which is predominantly expressed in retina, has 33% identity to Dlg and belongs to the p55 (MPP1) MAGUK family. DLG1, -2, -3, and -4 are highly expressed in neuronal tissues, especially in postsynaptic densities. However, at least some of them are rather widely expressed in other tissues, though at lower levels, and seem to have functions in nonneuronal tissues, including colorectal epithelium for DLG1 (24) and esophageal epithelium for DLG3 (22). In Drosophila, Dlg is expressed in both neuronal and epithelial cells. Clearly, DLG2, -3, and -4 as well as DLG1 could be E6 targets in cervical epithelial cells. Here, we focused on DLG4 first, as it is the closest homologue of DLG1 and Dlg, and we isolated it as an E6-interacting protein in a yeast two-hybrid system (8).

Like DLG1, DLG4 has three PDZ domains in its amino-terminal half, a central Src homology 3 (SH3) motif, and a carboxy-terminal guanylate kinase (GUK)-homologous domain. Many lines of evidence suggest that DLG4/PSD95 in neural cells is associated with the NMDA receptor and has a role in synapse stabilization and plasticity (4, 18, 26). However, its significance in epithelial cells is mostly unknown. To explore the functions of DLG4 and the possibility that it is targeted by E6, we examined expression of DLG4 as well as some other known E6 target proteins in normal cervical epithelial cells as well as cervical cancer cell lines. We report that DLG4 is expressed in normal keratinocytes of both uterine cervix and skin and can be targeted for degradation by E6-E6AP complexes. Though DLG4 protein is thus down-regulated in HPV18 E6-positive HeLa cells, Dlg4 mRNA is down-regulated by an HPV-independent mechanism in the HPV16-positive cervical cancer cell lines SiHa and CaSki. In addition, expression of DLG4 in CaSki cells suppresses tumorigenicity to a remarkable degree, suggesting a tumor suppressor function in the development of HPV-associated cancers.


arrow
MATERIALS AND METHODS
 
Cell culture. The cervical cancer cell lines HeLa, CaSki, SiHa, C33A, OMC4, and Yumoto were maintained in Dulbecco's modified Eagle medium (DMEM; Sigma) supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum (FBS). The HPV16-positive cervical cancer cell lines QG-H and SKGIIIb were provided by Hiroshi Shirasawa and maintained in RPMI 1640 supplemented with 10% (vol/vol) FBS. HBEC1 (human bronchial epithelial cells) and HDK1 (human dermal keratinocytes) were purchased from Cell Applications Inc. (San Diego, CA), and HFF2 cells (human foreskin fibroblasts) were from Cambrex (Walkersville, MD). Normal human cervical keratinocytes (HCK1) were obtained with written consent from a patient who underwent a hysterectomy for a gynecological disease other than cervical cancer. They were infected with LXSH-hTERT retroviruses (28) encoding the catalytic subunit of human telomerase reverse transcriptase (hTERT), and the infected cells were selected in the presence of 20 µg/ml of hygromycin B. The resultant HCK1T cells were maintained in keratinocyte serum-free medium (KSFM; Invitrogen) supplemented with 5 ng/ml epidermal growth factor (EGF) (Sigma) and 50 µg/ml of bovine pituitary extract (Hammond Tech, Windsor, CA). HCK1T cells had been cultivated for more than 150 population doublings (PD), and the majority of the cells maintained almost normal diploid chromosomes at 10, 40, and 120 PD. All cells were incubated in a humidified atmosphere containing 5% CO2 at 37°C.

Plasmids. Wild-type and mutant HPV16 E6 (16E6) segments were described previously (16). Wild-type and mutant HPV18 E6 (18E6) segments, as well as a BPV1 E2 segment, were prepared in a similar manner here. These segments, which were attached with a small leader sequence, ACC, flanking the initiation codon for efficient translation (16), were cloned into retroviral vector plasmids and/or pGEM-4 (Promega) for in vitro translation with the Gateway system (Invitrogen). Briefly, the segments were first recombined with pDONR201 or pDONR221 by BP reaction to generate entry vectors. Then they were recombined with the retroviral vectors pDEST-CLXSN (designated pCLXSN-DEST previously), pDEST-CLMSCVpuro (designated pCLMSCVpuro-DEST previously) (19), and pDEST-GEM4-SP6 by the LR reaction (Invitrogen) according to the manufacturer's instructions to generate pCLXSN-16E6SD, -16E6SD{Delta}151, -18E6SD, and -18E6SD{Delta}158, pCLMSCVpuroBPV1-E2, and pGEM4-18E6, -18E6{Delta}158, and -18E6-151L. pDEST-GEM4-SP6 was constructed by inserting a modified rfA segment containing the ccdB gene (Invitrogen) between the EcoRI and BamHI sites of pGEM-4 so that the recombined segments were transcribed by SP6 polymerase. The other 16E6 segments inserted in pGEM-4 (Promega) for in vitro translation were described previously (16). Construction of the destination vector pDEST-CL-SI-MSCVpuro (designated pSI-CMSCVpuroDEST previously) for retroviral expression of short hairpin RNAs (shRNAs) and the entry vector pENTR-H1R-stuffer has been described previously (36). The E6AP-specific shRNA constructs pCL-SI-MSCVpuro-H1R-E6APRi3 and -4 were generated by a previously described method (36). The E6AP shRNA-targeted sequences were 5'-AGCTGCAAAGCATCTAATA-3' and 5'-GAAATCTAGTGAATGATGA-3'.

A full-length human DLG4 cDNA was isolated from HeLa cell RNA by reverse transcription-PCR (RT-PCR). After amplification of the open reading frame with specific primers containing attB1 and attB2 sequences, 5'-aaaaagcaggctccaccATGTCCCAGAGACCAAGAG-3' and 5'-agaaagctgggTCAGAGTCTCTCTCGGGCTGGA-3' (sequences specific to DLG4 cDNA are capitalized), and attB adaptor primers, 5'-GGGGACAAGTTTGTACAAAAAAGCAGGCT-3' and 5'-GG GGACCACTTTGTACAAGAAAGCTGGGT-3', they were recombined into pDONR221 by BP reaction (Invitrogen). One cloned cDNA (clone 34) whose coding capacity was identical to that of the reported sequence (NCBI accession no. U83192) was recombined with pDEST-GEM4-SP6 and pDEST-15 glutathione S-transferase (GST) fusion vector (Invitrogen) by LR reaction (Invitrogen). Deletion mutants from the full-length DLG4 cDNA were produced by PCR with various sets of primers so as to encode the desired products (see Fig. 4B). These segments as well as wild-type p53 cDNA were similarly recombined with pDEST-GEM4-SP6 by LR reaction. The full-length DLG4 cDNA and various E6 segments were cloned into a destination vector, pDEST-EF6/mycHisA, to generate pEF6/DLG4 and various pEF6/E6 plasmids. pDEST-EF6/mycHisA was constructed by inserting a modified rfA segment (Invitrogen) into the EcoRI site of pEF6/mycHisA (Invitrogen) so that the recombined segments could be transcribed under the control of the EF1{alpha} promoter. The full-length DLG4 cDNA was also cloned into p3XFLAG-CMV10 (Sigma) to generate p3XFLAG-CMV10-DLG4 so that the FLAG tag was fused to DLG4 in frame. The hemagglutinin (HA)-tagged E6AP expression plasmid, pCMV-HA-E6AP, was a gift from Shunsuke Nakagawa, and the dominant negative mutant plasmid, pCMV-HA-E6AP-C833A, was generated by in vitro mutagenesis. The His6-tagged ubiquitin expression plasmid pCMV-His6-Ub was a gift from Dirk Bohmann.


Figure 4
View larger version (68K):
[in this window]
[in a new window]

  		FIG. 4. Binding of DLG4 and high-risk mucosotropic HPV E6 proteins. (A) A 5-µg aliquot of various MBP fusion proteins bound by amylose-Sepharose beads was mixed with 100 µg of LXSN-DLG4-infected C33A cell lysates. The captured DLG4 was detected by immunoblotting. As a control, a 5% equivalent of the input lysate was loaded. Aliquots of MBP fusion proteins were loaded on a gel and visualized by Coomassie brilliant blue (CBB) staining (lower panel; ß-gal, ß-Gal {alpha}-peptide). (B) Schematic representation of DLG4 and the mutant proteins. The locations of the three repeats of the PDZ domain (PDZ1 to -3), a Src homology 3 (SH3) domain, and a guanylate kinase-like domain (GUK) are indicated. (C) Various segments of DLG4 as well as DLG1 were translated in vitro in the presence of [35S]methionine. Radiolabeled proteins were incubated with MBP-16E6, -18E6, or -ß-Gal bound by amylose-Sepharose beads, and the captured proteins were resolved by SDS-PAGE and visualized with a BAS2500 image analyzer (Fujifilm Co. Ltd., Tokyo, Japan). As a control, a 10% equivalent of the input product was loaded. (D) Wild-type and mutant E6 proteins were translated in vitro in the presence of [35S]methionine and [35S]cysteine and mixed with GST-DLG4 bound to glutathione-Sepharose beads. The captured proteins were analyzed as for panel C. "{Delta}151" and "{Delta}158" indicate deletion of the carboxy-terminal 151st amino acid; "V," "I," "P," and "ETQV" indicate mutants with the carboxy-terminal leucine replaced with valine, isoleucine, proline, and glutamic acid-threonine-glutamine-valine, respectively; 16E6-152P is a 16E6 mutant with an additional proline at the carboxy-terminus; "L" indicates a mutant with the carboxy-terminal valine replaced with leucine.

Preparation and infection of retroviruses. The production of recombinant retroviruses was described elsewhere (29). Briefly, the retroviral vector plasmids were cotransfected with the packaging construct, pCL-10A1, into 293FT cells (Invitrogen) using TransIT-293 (Mirus Co., Madison, WI) according to the manufacturer's instructions, and culture fluid was harvested 48 to 72 h posttransfection. Titers of the recombinant viruses were greater than 2 x 105 drug-resistant CFU/ml with HeLa cells, except that with LXSN-DLG4 viruses it was 1 x 104 to 4 x 104 CFU/ml. A half milliliter of viral fluid and 2 ml of growth medium were added to cultured cells seeded on 35-mm dishes (BD Falcon 4046) in the presence of Polybrene (4 µg/ml for HCK1T and 8 µg/ml for cervical cancer cell lines). Following the inoculation, infected cells were selected in the presence of 50 µg/ml G418 or 0.5 µg/ml puromycin for HCK1T and 800 µg/ml G418 and 0.5 µg/ml puromycin for cervical cancer cell lines unless otherwise stated.

Immunoblotting. Whole-cell proteins were extracted in lysis buffer (0.5% NP-40, 1 mM dithiothreitol, 50 mM NaCl, 25 mM Tris-HCl [pH 8.0], 0.02% NaN3) supplemented with 5% (vol/vol) protease inhibitor cocktail (Nacalai Tesque Inc., Kyoto, Japan). Gels were loaded with 20 µg of total cell lysate per lane as described previously. Monoclonal antibodies against DLG4 (K28/43; Upstate), DLG1 (clone 12; BD transduction lab), E6AP (clone 330; Sigma), BPV1 E2 (clone 5E11; Quattromed, catalog no. 235), HPV16 E7 (clone 8C9; Zymed), and p53 (Ab6; Oncogene Science) and polyclonal antibodies against ZO1 (Z-R1; Zymed), Scrib (C-20; Becton Dickinson), MAGI-1 (M-5691; Sigma), HPV18 E7 (sc-1590; Santa Cruz), and ß-actin (sc-1616; Santa Cruz) were used as probes with horseradish peroxidase-conjugated anti-mouse, anti-rabbit (Jackson Immunoresearch Laboratories), or anti-goat (sc-2033; Santa Cruz) immunoglobulins as the secondary antibodies. All the antibodies were used at a dilution of 1:1,000 for Western blotting except E7 antibodies, which were used at a dilution of 1:100. A LAS3000 charge-coupled device imaging system (Fujifilm Co. Ltd., Japan) was employed for detection and quantification of proteins visualized by Lumi-light plus Western blotting substrate (Roche).

Quantitative RT-PCR analysis. Total RNA was obtained from cultured cervical cancer cell lines using an RNeasy kit (QIAGEN) according to the manufacturer's protocol. Single-stranded cDNAs were synthesized from 500 ng of RNA using a reverse transcription system (Promega) with random hexamers. Fifty nanograms of cDNA template was subjected to PCR amplification with primer sets specific to Dlg4 or acidic ribosomal phosphoprotein P0 (36B4) using a SYBR Green PCR core reagents kit (Applied Biosystems; Foster City, CA). The expression level of the Dlg4 gene was then normalized to RNA content for each sample using 36B4 as a control. PCR analysis was performed under the following conditions: denaturation at 95°C for 10 min, followed by 40 cycles (95°C for 15 s, 58°C or 60°C for 10 s for Dlg4 and 36B4 genes, respectively, and 72°C for 10 s). Amplified products were detected with a iCycler iQ real-time PCR detection system (Bio-Rad). Sequences of primer pairs used were as follows: 5'-CTCGATGTCTCGGCCAATG-3' and 5'-CTCTGTGATCCGCTTGTTAATCTC-3' for Dlg4 and 5'-TCGACAATGGCAGCATCTAC-3' and 5'-GCCTTGACCTTTTCAGCAAG-3' for 36B4.

Protein interaction assays. Various E6 and DLG4 proteins were synthesized using the in vitro transcription and translation-coupled reticulocyte lysate (TNT) system (Promega) in the presence of [35S]cysteine and/or [35S]methionine. Preparation and purification of maltose-binding protein (MBP) and GST fusion proteins were as described previously (16). The proteins immobilized to amylose- and glutathione-Sepharose (New England Biolabs and Amersham) were preincubated in NTN300 buffer (1% NP-40, 50 mM Tris-HCl [pH 7.5], and 300 mM NaCl) containing 2% bovine serum albumin for at least 10 min, mixed with 4 to 20 µl of the programmed reticulocyte lysates containing in vitro-translated products or cell lysates, and incubated at 4°C for 2 h under constant rotation. The Sepharose beads were then washed with NTN300 buffer five times. The captured proteins were eluted by adding sodium dodecyl sulfate (SDS) sample buffer, separated by SDS-polyacrylamide gel electrophoresis (PAGE), and detected with a BAS2500 image analyzer (Fujifilm Co. Ltd.) or by immunoblotting.

Anchorage-independent cell growth assay. Cells were seeded at 1 x 104 per 35-mm plate in 70% (vol/vol) DMEM-30% (vol/vol) FBS with 0.4% agarose and a 0.7% agarose underlay. Fresh medium was overlaid once a week.

Tumorigenicity in nude mice. Tumorigenicity was tested by injecting 5 x 106 cells subcutaneously into BALB/cAJcl-nu female mice (Nihon Clea, Japan) at 6 weeks of age and measuring the sizes of the resulting tumors at least once a week for 2 months. Animal protocols were approved by the committee for Ethics of Animal Experimentation and were in accordance with the Guideline for Animal Experiments in the National Cancer Center.

Statistical analysis. The data were analyzed using Student's t test, with a P value of <0.05 considered significant.


arrow
RESULTS
 
Expression of DLG4 in cervical cancer cell lines and normal cervical keratinocytes. We first examined the expression levels of DLG4 as well as some other known E6 targets in cervical cancer cells and normal human cell strains, including hTERT-transduced cervical keratinocytes (HCK1T) by immunoblotting (Fig. 1A). HPV-positive cervical cell lines examined expressed very little DLG4, while HCK1T as well as normal dermal keratinocytes (HDK1) expressed easily detectable levels of DLG4. The expression level of DLG4 in HCK1T was about 30 and 50% of those in a neuroblastoma cell line, IMR32, and a glioblastoma cell line, U87, respectively (data not shown), suggesting a significant function of DLG4 in epithelial cells. We confirmed the result with another antibody (Calbiochem clone 6G6-1C9), though it was less sensitive (data not shown). Among HPV-negative cervical cancer cell lines, C33A and OMC4 (but not the Yumoto line) expressed low levels of DLG4. It is noteworthy that all the normal human cell strains, including nonepithelial fibroblasts (HFF2), expressed DLG4. Unexpectedly, some HPV-positive cervical cancer cell lines expressed rather higher levels of DLG1 and Scrib than HCK1T. Expression levels of ZO1, a MAGUK family protein similar to DLG1 and DLG4 but not an E6 target, were quite variable.


Figure 1
View larger version (63K):
[in this window]
[in a new window]

  		FIG. 1. Expression of DLG4 in cervical cancer cell lines and primary human cells and the effects of various E6 proteins in HCK1T cells. (A) Expression levels of DLG4, DLG1, ZO1, Scrib, and p53 analyzed by immunoblotting. All cell proteins were extracted from subconfluent cultures. The same amounts of each cell extract (20 µg protein) were loaded per lane (ß-actin levels were not used as loading controls, as they were different in each cell line) (data not shown). HPV18-positive (HeLa), HPV16-positive (CaSki, SiHa, QG-H, SKGIIIb), and HPV-negative (OMC4, Yumoto, C33A) cervical cancer cell lines, an immortalized cervical keratinocyte cell line (HCK1T), normal human dermal keratinocytes (HDK1), human bronchial epithelial cells (HBEC1), and normal human foreskin fibroblasts (HFF2) were used. (B) HCK1T cells were transduced with the indicated genes by retroviral gene transfer. Cell proteins were extracted from subconfluent cultures, and expression levels of the indicated proteins were analyzed by immunoblotting. "{Delta}151" indicates deletion of the carboxy-terminal 151st amino acid; "V" indicates replacement of the carboxy-terminal leucine with valine; "L" indicates replacement of the carboxy-terminal valine with leucine;16E6SAT is an E6 mutant with an N-terminal three-amino-acid substitution which is defective in p53 inactivation; LXSN is the vector control.

DLG4 expression is down-regulated by E6 proteins in HCK1T. We next examined the effects of 16E6, 18E6, and their mutants on the expression of DLG4 as well as p53 as a control in HCK1T cells (Fig. 1B). The E6 mutants included the 8S9A10T substitution mutant of 16E6 (16E6SAT), which is unable to target p53 for degradation but is still able to bind E6AP (15), and carboxy-terminal deletion mutants (16E6{Delta}151 and 18E6{Delta}158), which are unable to bind DLG1 (16). Expression levels of DLG4 were decreased by wild-type E6 proteins of both HPV16 and -18 but not by their carboxy-terminal deletion mutants. However, 18E6 was more potent than 16E6 in its activity. The difference seemed largely dependent on the difference in the last amino acid of the two proteins, as 16E6-151V, which had the same C terminus as 18E6, showed activity similar to that of 18E6, and 18E6-158L showed activity similar to that of 16E6. As expected, p53 levels were unchanged by the 16E6SAT mutants, but the DLG4 levels again clearly depended on the difference in their last amino acids. As the 16E6SAT mutants showed slightly weaker activities than their counterparts, degradation of p53 might be indirectly involved in the regulation of DLG4 expression. Unexpectedly, down-regulation of DLG1 and Scrib was not sufficiently clear to allow an unequivocal conclusion, probably because their basal levels were relatively low in subconfluent cultures (Fig. 2B).


Figure 2
View larger version (51K):
[in this window]
[in a new window]

  		FIG. 2. Cell density-dependent regulation of DLG4 and other proteins. (A) Cells were harvested under subconfluent conditions (SC) or 7 days postconfluence (PC). Expression levels of the indicated proteins were analyzed by immunoblotting. The same amounts of each cell extract (20 µg protein) were loaded in each lane, with ß-actin being used as a loading control. Only a 170-kDa form of MAGI-1 detected in C33A and HCK1T is shown. (B) HCK1T cells transduced with 16E6, 18E6, and backbone vector (LXSN) were harvested under subconfluent (SC) or confluent (Conf) conditions or 9 days postconfluence (PC), and expression levels of the indicated proteins were analyzed by immunoblotting.

Cell density-dependent regulation of DLG4. Based on the notion that PDZ domain-containing proteins play roles in signal transduction through cell-cell contact and/or in regulation of cell polarity, DLG4 levels at different cell densities were examined. DLG1 was remarkably up-regulated in postconfluent cultures of HCK1T, CaSki, and SiHa but not HeLa cells (Fig. 2A). A similar increase of DLG4 was seen in postconfluent cultures of C33A and less clearly in HeLa and SiHa, while a decrease was seen in HCK1T cells. Likewise, a significant decrease of Scrib was observed in postconfluent cultures of HeLa, SiHa, and C33A cells, while a significant increase was observed in HCK1T cells, suggesting different regulatory mechanisms in DLG1, DLG4, and Scrib. Next, we examined the effects of E6 on their expression at different cell densities. Maximum levels of DLG4 were observed at confluence in HCK1T cells, whereas in the DLG1, Scrib, and MAGI-1 cells they were observed in postconfluent cultures (Fig. 2B). E-cadherin and involucrin, which is an early differentiation marker protein of keratinocytes, were also up-regulated in confluent cultures of HCK1T but very little in the cervical cell lines (Fig. 2A), and levels remained low in HCK1T cells expressing 16E6 and 18E6 (Fig. 2B). Down-regulation of DLG4 by both E6 proteins was observed at different cell densities (Fig. 2B). Interestingly, upregulation of DLG1 and MAGI-1 was inhibited by both E6 proteins and that of Scrib was inhibited by 16E6 only under postconfluent conditions. Surprisingly, down-regulation of p53 by 18E6 was alleviated at confluence and almost cancelled postconfluence (Fig. 2B). Taking into account that 18E6 is less efficient in promoting degradation of p53 than 16E6 (Fig. 1B and 2B) (37), the result implies the possibility that PDZ domain-containing proteins are better substrates for 18E6-E6AP ubiquitin ligase complexes than p53 and that accumulated proteins after confluence sequester the enzyme complexes from p53, though the idea remains to be validated.

Our findings suggest that DLG4 as well as DLG1 has a specific role in normal epithelial cells, even though the expression levels are lower than those in neuronal tissues. Though calcium was reported to enhance differentiation of keratinocytes, the levels of DLG1 and DLG4 were unaffected by calcium concentration in the medium between 0.035 and 1 mM (data not shown), supporting a more critical role of cell-cell contact in keratinocyte differentiation (17).

Restoration of DLG4 by repression of E6 and E7 in HeLa cells expressing BPV1 E2. Previous studies showed that BPV1 E2 proteins can inhibit the E6/E7 promoters in the LCRs of HPV16 and -18 in HPV-positive cervical cancer cell lines so as to repress their expression and induce growth arrest (2, 3, 35). E6 can target DLG1, Scrib, MUPP1, and MAGIs for degradation (7, 9, 20, 27, 40). Therefore, we tested whether repression of high-risk E6 and E7 expression by BPV1 E2 can restore the expression of DLG4 and other E6 target proteins. In HeLa cells, expression levels of DLG4 and DLG1 proteins as well as p53 were restored by BPV1 E2 expression while the ZO1 level was rather reduced (Fig. 3, lanes 3 and 4). HeLa cells exogenously expressing 16E6 also demonstrated restored expression of DLG4 to an extent similar to that observed in parental HeLa cells with E2 expression, whereas the expression level of p53 was not restored (Fig. 3, lanes 1 and 2). DLG4 expression levels were very low in the 16E6-positive cervical cancer cell lines CaSki and SiHa and were not elevated by E2 expression, whereas p53 was clearly up-regulated (Fig. 3, lanes 5 to 8). DLG1 levels were relatively high in these cells and were increased very little by E2 expression. No change in expression levels of these proteins was observed in the control (the HPV-negative cervical cancer cell line C33A). These data together with those depicted in Fig. 1 suggest that DLG4 and DLG1 are efficient targets for degradation by 18E6 but not by 16E6 in vivo and that DLG4 down-regulation in HPV16-positive cervical cancer cell lines (CaSki and SiHa) depends not on either 16E6 or 16E7 function but on other mechanisms.


Figure 3
View larger version (86K):
[in this window]
[in a new window]

  		FIG. 3. E6-dependent and -independent down-regulation of DLG4. HPV18-positive (HeLa), HPV16-positive (CaSki, SiHa), and HPV-negative (C33A) cervical cancer cell lines, as well as HeLa cells infected with LXSN-16E6SD, were harvested 7 days after infection with MSCVpuro-BPV1E2 (E2) or the control vector MSCVpuro and selected in medium supplemented with 1 µg/ml of puromycin (2 µg/ml for SiHa cells) from day 2. Mock-infected cells died before harvesting (data not shown). Expression levels were analyzed by immunoblotting. The same amounts of cell extracts (20 µg protein) were loaded in each lane.

Different Dlg4 mRNA levels in cervical cancer cell lines. To address the mechanisms whereby DLG4 is down-regulated in the HPV16-positive cervical cancer cell lines SiHa and CaSki, mRNA levels were analyzed by quantitative RT-PCR. As expected, in CaSki and SiHa cells they were relatively low. In contrast, Dlg4 mRNA levels in HeLa cells were higher than those in HCK1T, C33A, and OMC4 cells (Table 1), all of which showed higher levels of DLG4 protein than HeLa cells (Fig. 1A). This discrepancy between mRNA levels and protein levels of DLG4 supports the notion that down-regulation of DLG4 in HeLa cells is posttranscriptional and caused by enhanced degradation of the protein by 18E6. The mRNA results also suggested that the down-regulation of Dlg4 in SiHa and CaSki is due mainly to reduced mRNA levels by an HPV-independent mechanism.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Dlg4 mRNA levels in cervical cancer cell lines determined by quantitative real-time RT-PCR

Specific interactions between E6 proteins of high-risk mucosotropic HPVs and DLG4. To further clarify the mechanisms whereby E6 down-regulates expression of DLG4, we examined whether E6 proteins can physically interact with DLG4 in vitro. Various MBP-fused E6 proteins immobilized on amylase-Sepharose beads were mixed with lysates from DLG4-overexpressing C33A cells. DLG4 was efficiently bound by E6 proteins of HPV16 and -18 but not by those of HPV11, -5, or -1 or by HPV16 E7 (Fig. 4A). To delimit the segment of DLG4 necessary for binding to 16E6 or 18E6, deletion mutants of DLG4 were prepared by in vitro transcription and translation in reticulocyte lysates. DLG1 was used as a positive control for binding to E6. As expected, DLG4 as well as DLG1 was efficiently bound by E6 proteins of both HPV16 and -18 but not by ß-galactosidase (ß-Gal). Deletion mutants containing PDZ domain 2 (PDZ2) alone bound as efficiently as the full-length DLG4 to both E6 proteins, while the mutant containing PDZ1 alone demonstrated faint binding (Fig. 4C). These results indicate that 16E6 and 18E6 bind to DLG4 mainly through PDZ2 and possibly also through PDZ1. A series of E6 mutants were prepared to examine the critical role of the carboxy-terminal region in binding to DLG4 in vitro (Fig. 4D). As expected, mutations which destroyed the class I PDZ-domain binding motif, XS/TXL/I/V, abrogated binding to DLG4. Addition of the motif of HPV18, ETQV, to the carboxy-terminal end of the one-amino-acid-deleted mutant (16E6{Delta}151) restored the binding affinity. Replacement of the last amino acid, leucine, of HPV16 E6 with valine (16E6-151V) greatly increased the binding affinity for DLG4, and a change to isoleucine (16E6-151I) did so to a lesser extent. Conversely, replacement of the last amino acid, valine, of HPV18 E6 with leucine (18E6-158L) decreased the affinity. These results indicated that the PDZ domain-binding motif in the carboxy-terminal motif of E6 proteins is critical for interaction with DLG4 and that differences in the last amino acid, whether leucine or valine, are critical for the affinity, accounting for the variation between 16E6 and 18E6 activities shown in Fig. 1B. The importance of the last amino acid residue on 16E6 and 18E6 has also been shown for other PDZ targets (7, 39, 41).

E6-induced degradation of DLG4 in vivo and in vitro. We next examined E6-induced degradation of DLG4 on transient transfection. Expression of 16E6 and 18E6 reduced the steady-state levels of DLG4 in 293FT cells, and 18E6 was more potent than 16E6. Again the activity proved dependent on the last amino acids of both E6 proteins (Fig. 5A). The half life of DLG4 was less than 15 min in 18E6-expressing cells, whereas it was more than 120 min without the E6 protein (Fig. 5B). These results clearly indicate that E6 can strongly enhance degradation of DLG4 in vivo. We could also confirm E6-dependent degradation of DLG4 in vitro (Fig. 5C). However, surprisingly, 16E6 was more potent than 18E6 in degradation of DLG4 in a reticulocyte lysate system, and the difference in the last amino acid was rather trivial in this system. The observed difference is not due to expression levels of E6 proteins, as 18E6 (Fig. 5C, lane 2), which has 10 cysteines, was rather more abundant than 16E6 (lane 5), which has 14 cysteines. The stability of these E6 proteins was similar, as we observed similar amounts of E6 proteins to remain after 3 h of incubation (data not shown).


Figure 5
View larger version (44K):
[in this window]
[in a new window]

  		FIG. 5. E6-dependent degradation of DLG4 in vivo and in vitro. (A) 293FT cells in 35-mm dishes were transfected with pEF6/DLG4 (100 ng), pEGFP-C1 (200 ng; Clontech), and wild-type or mutant E6 expression plasmids (100 ng). Total cell proteins were harvested at 46 h posttransfection, and expression levels of the indicated proteins were analyzed by immunoblotting. (B) 293FT cells in 35-mm dishes were transfected with pEF6/DLG4 (100 ng) and pEGFP-C1 (200 ng), with or without pEF6/18E6 (100 ng). At 24 h posttransfection, cycloheximide (CHX) was added at a concentration of 50 µg/ml. Total cell proteins were harvested at the indicated times, and expression levels of proteins were analyzed by immunoblotting. (C) DLG4, p53, and various E6 proteins were prepared in the TNT system with SP6 polymerase. E6 proteins were translated in the presence of [35S]cysteine. The programmed reticulocyte lysates of DLG4 (1 µl) and p53 (1 µl) were mixed with that of individual E6 (10 µl) and fresh rabbit reticulocyte lysate (5 µl) in the presence of 2.5 mM ATP, incubated for 180 min at 30°C, and then resolved on SDS-PAGE followed by immunoblotting. E6 proteins were visualized with a BAS2500 image analyzer (Fujifilm Co. Ltd.).

E6-induced degradation of DLG4 depends on E6AP in vivo. E6AP can recognize p53 as a substrate for ubiquitination only when making complexes with E6 proteins. Likewise, E6AP is reported to be required for degradation of another E6 target, Scrib (27), although another ubiquitin ligase was suggested to be necessary for degradation of DLG1 (10, 34). In order to examine whether E6AP is involved in the degradation of DLG4, E6AP-specific shRNAs were used to reduce the E6AP expression in the transient-transfection assay. Although the apparent reduction of E6AP levels was not very obvious, probably owing to imperfect transfection efficiency and the stability of E6AP (15), both E6AP-specific shRNAs inhibited the E6-dependent degradation of DLG4, and the additional expression of E6AP but not the dominant negative mutant of E6AP reversed the effect (Fig. 6A). E6AP-specific shRNA was stably expressed in three HPV-positive cell lines, HeLa, CaSki, and SiHa, and an HPV-negative cell line, C33A, and the steady-state levels of the E6 targets were measured by immunoblotting (Fig. 6B). As expected, steady-state levels of p53 were increased in the HPV-positive cell lines but not in C33A harboring mutant p53, indicating a specific effect of E6AP shRNA. Levels of DLG4, DLG1, and Scrib in HeLa were increased by the E6AP shRNA expression. In CaSki and SiHa cells, the levels of Scrib and DLG1 were significantly increased, while those of DLG4 were not. In C33A cells, no stabilization of DLG4, DLG1, or Scrib by E6AP shRNA was observed. These results thus suggest that E6AP is involved in degradation of DLG4, DLG1, and Scrib in HeLa cells and in that of DLG1 and Scrib in CaSki and SiHa cells.


Figure 6
View larger version (74K):
[in this window]
[in a new window]

  		FIG. 6. E6- and E6AP-dependent degradation and ubiquitination of DLG4. (A) 293FT cells in 35-mm dishes were transfected with pEF6/DLG4 (100 ng), pEGFP-C1 (200 ng; Clontech) and the expression plasmid pEF6/18E6 (300 ng), shRNA expression plasmid (1.2 µg), pCMV-HAE6AP (50 ng or 200 ng), or pCMV-HAE6APC833A (DN; 200 ng), as indicated at the top, and total cell proteins were harvested at 72 h posttransfection and analyzed by immunoblotting. Relative amounts of endogenous E6AP are indicated at the bottom. Note that the anti-E6AP monoclonal antibody cannot detect exogenously expressed HA-E6AP and that E6AP shRNA3 cannot target it, because it lacks the first 108 amino acid residues. (B) E6AP-specific shRNA4 was expressed in cervical cancer cell lines by retroviral gene transfer. Cells were harvested from 50%-confluent cultures, and expression levels of the indicated proteins were analyzed by immunoblotting. (C) 293FT cells in 90-mm dishes were transfected with pEGFP-C1 (500 ng; Clontech), pCMV-His6-Ub (1 µg), p3XFLAG-CMV10-DLG4 (500 ng), pEF6/18E6, pEF6/18E6{Delta}158 (4 µg), pCMV-HAE6AP (4 µg), pCMV-HAE6APC833A (4 µg), or pCL-SI-MSCVpuroH1R-E6APRi4 (4 µg), as indicated at the top. MG132 was added at a concentration of 50 µM for 4 h before harvest, and cells were collected at 48 h posttransfection. Expression levels of the indicated proteins were analyzed by immunoblotting. Aliquots of cells were used for in vivo ubiquitination assay as described previously (8). Briefly, after boiling in 4% SDS in phosphate-buffered saline, lysed cells were diluted in 6 volumes of 1% Triton X-100 in phosphate-buffered saline and immunoprecipitated with anti-FLAG M2 affinity resin (Sigma), and the bound proteins were analyzed by immunoblotting with anti-His tag antibody (sc-803; Santa Cruz). (D) DLG4, p53, and various E6 proteins were prepared in the TNT system with SP6 polymerase in the presence of [35S]methionine. The programmed reticulocyte lysate of DLG4 (2 µl) or p53 (2 µl) was mixed with that of E6 (20 µl) with 10 µl of fresh rabbit reticulocyte lysate in the presence of 4 mM ATP{gamma}S [adenosine 5'-(3-thiotriphosphate)], incubated for 120 min at 30°C, and then resolved on SDS-PAGE followed by visualization with a BAS2500 image analyzer (Fujifilm Co. Ltd.). (E) DLG4, p53, and various E6 proteins were prepared in the TNT system with SP6 polymerase. The programmed wheat germ extract (WGE) of DLG4 (1 µl) was mixed with that of individual E6 (5 µl), HA-E6AP, or HA-E6APC833A (5 µl), as indicated at the top, and fresh wheat germ extract lysate (6 µl) in the presence of 2.5 mM ATP was incubated for 3 h at 30°C and then resolved on SDS-PAGE followed by immunoblotting with anti-DLG4 antibodies.

E6-dependent ubiquitination of DLG4 in vivo and in vitro. To further investigate the potential role of E6 and E6AP in degradation of DLG4, we examined E6- and E6AP-dependent ubiquitination of DLG4 in vivo. Polyubiquitinated forms of FLAG-tagged DLG4 were increased by 18E6 coexpression, concomitantly with the reduced steady-state level of DLG4, and they were further increased by E6AP coexpression but decreased by coexpression of E6AP shRNA4 or a dominant negative mutant of E6AP (Fig. 6C). We next examined E6-dependent ubiquitination of DLG4 in vitro using proteins translated in rabbit reticulocyte lysates and wild-type and mutant E6 proteins of HPV16 and HPV18. As expected, formation of high-molecular-weight forms of p53, which could be considered to be ubiquitination, was enhanced by addition of 18E6{Delta}158 as well as the wild-type E6 proteins but not by 16E6SAT and 16E6SAT{Delta}151. In a parallel experiment, formation of high-molecular-weight forms of DLG4 was enhanced by 16E6SAT as well as the wild-type E6 proteins but not by the carboxy-terminal deletion mutants (Fig. 6D). We did not observe significant difference in the activities of 16E6 and 18E6 in the reticulocyte lysate system. In the wheat germ extract system, where the functional ubiquitin ligase E6AP is absent (11), both E6 and E6AP were required for the formation of high-molecular-weight forms of DLG4 (Fig. 6E). 18E6 was clearly more potent than 16E6 in putative ubiquitination, concordant with the in vivo degradation assay results. Our findings as a whole indicate that E6AP is involved in E6-mediated ubiquitination and degradation of DLG4 both in vivo and in vitro.

Tumor-suppressive effects of DLG4 in CaSki cells. To address the biological significance of the observed down-regulation of DLG4 in cervical cancer cell lines, we decided to examine the biological effect(s) of exogenous DLG4 expression in CaSki cells. As carboxy-terminal PDZ domain-binding motif-dependent transformation of 3Y1 cells is evident only after cells become confluent (16), we hypothesized that close cell-cell contact is required for any negative growth signal(s) involved with PDZ-domain containing proteins. We chose CaSki cells as the targets, as they express more E-cadherin than SiHa cells (Fig. 2A) and retain morphological features of epithelial cells. We prepared pooled populations of DLG4-expressing CaSki cells (CaSki/DLG4) as well as the control cells (CaSki/LXSN) by retroviral gene transfer. As expected, exogenously expressed DLG4 was not rigorously degraded by HPV16 E6 expressed in the cells and was easily detected by immunoblotting, though the expression levels were decreased at confluence (Fig. 7A). We first analyzed exponential growth of these pooled cells on plastic dishes, but no significant differences were observed (Fig. 7B). Then, we compared anchorage-independent growth of these pooled cells in soft agarose medium and found that CaSki/DLG4 formed smaller and fewer colonies than CaSki/LXSN (Fig. 7C). Finally, tumorigenicity in nude mice was examined. Surprisingly, DLG4 markedly suppressed tumor growth of CaSki cells in a pilot experiment. The result was confirmed in repeated experiments using the same pooled cells (Fig. 7D; Table 2), and four more sets of independently established pooled cells (Table 2); the tumor-suppressive effect was not detected with the second set of pooled cells, but here the control cells (CaSki/LXSN) formed no or small tumors, in contrast to the other experiments. These results together with the findings for anchorage-independent growth in soft agarose medium indicate that overexpressed DLG4 exerts tumor-suppressive effects on CaSki cells.


Figure 7
View larger version (26K):
[in this window]
[in a new window]

  		FIG. 7. Characterization of CaSki cells expressing DLG4. (A) DLG4-expressing CaSki cells (CaSki/DLG4) and control cells (CaSki/LXSN) were harvested either under subconfluent conditions (SC) or 7 days postconfluence (PC), and expression levels of the indicated proteins were analyzed by immunoblotting, as for Fig. 2A. (B) Growth curves for CaSki/DLG4 and CaSki/LXSN. Ten thousand cells were seeded on 35-mm dishes (BD Falcon 4046) and fed with fresh medium every 3 or 4 days. Cells were trypsinized and counted in duplicate on the indicated days after seeding. (C) Anchorage-independent growth of CaSki/DLG4 and CaSki/LXSN. Cells were seeded in soft agarose medium, and colonies over 50 µm in diameter were counted after 4 weeks. Five photographs of randomly selected areas in each dish were taken at a magnification of 40x, and the numbers of colonies were measured with the COLONY program (Fujifilm Co. Ltd.). The assay was performed in triplicate and repeated three times. Means plus standard deviations for three experiments are shown. *, P < 0.05 versus vector-transduced cells. (D) Tumorigenicity of CaSki/DLG4 and CaSki/LXSN cells in nude mice. Five million cells were subcutaneously injected above the hind legs of nude mice, and tumor size was measured at least once a week. In each case, pooled cells were injected at two sites per mouse. The tumor volume (mm3) was approximated by multiplication of the major axis, the minor axis and the height of each lesion. Each point is the mean of all the data for CaSki/DLG4 and CaSki/LXSN (n = 15), and bars represent standard deviations. Note the statistical significance at day 21 (P < 0.01) and subsequently.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Tumorigenicity of CaSki cells expressing DLG4


arrow
DISCUSSION
 
Identification of DLG4 as a novel E6 target. In this study, we identified the Dlg4 gene product, DLG4/PSD95, as a novel E6 target and showed that it was down-regulated in five out of five HPV-positive cervical cancer cell lines, whereas the known E6 targets DLG1 and Scrib were not (Fig. 1A). The result led us to consider that DLG4 might be an important E6 target in the development of HPV-positive cervical cancers. We found that E6 proteins of both HPV16 and -18 were able to bind DLG4 through the carboxy-terminal PDZ domain-binding motif in vitro (Fig. 4A and D) and down-regulate expression of DLG4 (Fig. 1B) through enhanced ubiquitination (Fig. 6C, D, and E) and degradation (Fig. 5A, B, and C) of DLG4 both in vivo and in vitro. Actually, accumulation of DLG4 as well as p53 in HeLa cells was observed with expression of both BPV1 E2 (Fig. 3) and E6AP-specific shRNA (Fig. 6A and B). Our results clearly indicate that the E6 proteins of HPV16 and -18 have the potential to induce proteolytic degradation of DLG4 as well as DLG1. Recently, Grm and Banks reported that degradation of DLG1 and MAGIs by human papillomavirus E6 is E6AP independent (10). However, our results clearly indicate that degradation of DLG4 and DLG1 by E6 in HeLa cells is, at least in part, E6AP dependent, consistent with a recent report by others (25), though we cannot rule out the possibility that a ubiquitin ligase(s) other than E6AP is also involved. Low expression levels of DLG4 in the HPV16-positive cell lines CaSki and SiHa as well as in one of three HPV-negative cervical cancer cell lines tested, Yumoto, were due mainly to an HPV-independent reduction in Dlg4 mRNA levels (Table 1).

Differential effects of HPV16 and HPV18 E6 proteins on DLG1, DLG4, and Scrib. We and others have described potential differences between 16E6 and 18E6 in affinity and effects on DLG1 (16, 23). In this report, we show the second PDZ domain of DLG4 to be critical for binding to both E6 proteins, similar to DLG1 (16, 23), but, unlike with DLG1, the third domain was not found to be involved in the binding to 18E6 (16, 23). We also show 18E6 to be more potent, at least in vivo, for targeting DLG4 as well as DLG1 than 16E6, which in contrast has a stronger influence on Scrib. This difference appeared to be primarily dependent on the last amino acids of these two proteins, leucine for 16E6 and valine for 18E6, because swapping of these last amino acids alone can change the targeting specificity (Fig. 1B and 4D). The same conclusion about DLG1 and Scrib was drawn by others recently (41). E6 proteins of most high-risk mucosotropic types end with valine like 18E6, and only that of HPV33, which has close homology to that of HPV16, ends with leucine. It is likely that these two groups of E6 proteins target overlapping PDZ domain-containing proteins with different specificities so as to lead to similar overall consequences. This notion again underlines the importance of functions dependent on the carboxy termini of E6 proteins. 16E6 appeared to have a weaker capacity for binding and inducing degradation of DLG4 both in vivo and in vitro (Fig. 1B, 4C and D, 5A, and 6A). However, in a reticulocyte lysate system, 16E6 showed a higher activity for inducing degradation of DLG4 (Fig. 5C) together with similar activity for inducing its ubiquitination (Fig. 6D), and surprisingly, the difference in the last amino acid (leucine or valine) had little effect on degradation (Fig. 5C). Although we have no data to explain the apparent discrepancy, one possibility is that the rabbit orthologue of E6AP or a putative ubiquitin ligase suggested for DLG1 degradation (10) favors 16E6 in the degradation of DLG4.

Tumor-suppressive effects of DLG4 in CaSki cells. If DLG4 functions as a tumor suppressor in the development of cervical cancer, restoration of its expression might be expected to change the carcinogenic phenotype(s) of cervical cancer cell lines. Based on this notion, we introduced DLG4 into CaSki cells (Fig. 7A). This resulted in reduced colony forming ability in soft agarose medium (Fig. 7C). More strikingly, CaSki/DLG4 cells were much less tumorigenic than control CaSki/LXSN cells (Fig. 7D and Table 2). However, CaSki/DLG4 cells did not show any difference in morphology and exponential growth in plastic dishes (Fig. 7B). This indicates that DLG4 would have no toxic effect on normal cell growth, as inferred from its expression in normal epithelial cells (Fig. 1), and that the effect on CaSki cells is specific to anchorage-independent growth and tumorigenicity in nude mice. One may argue that the reduced tumorigenicity by overexpression of DLG4 would depend on sequestering 16E6 from p53. However, the fact that the levels of p53 in the CaSki/DLG4 cells were as low as those in CaSki/LXSN cells (Fig. 7A) suggests this to be unlikely. Another possibility is that endogenous DLG4 has transforming/oncogenic activity like DLG1 (6) and that overexpressed DLG4 functions as a dominant-negative (DN) inhibitor of this activity, though this interpretation does not explain the reduced RNA levels of DLG4 in CaSki and SiHa cells. As DLG4 is expressed in not only cervical keratinocytes but also other epithelial cells, such as dermal keratinocytes and bronchial epithelial cells, we speculate that it might also function as a tumor suppressor in other tissues. Indeed, with immunoblotting of lung cancer cell lines, we found that three out of nine lung cancer cell lines examined expressed undetectable levels of DLG4 (K. Handa, unpublished data).

Though enhanced long-term potentiation and impaired learning are reported for mice lacking DLG4 (26), no association with cancer has been reported. However, as DLG1 to -4 are close homologues of Dlg, they might compensate for each other in epithelial tissues. Alternatively, it is possible that DLG4 has more functions in humans than in mice, as human DLG4 is also expressed in a wide range of nonneural tissues (38), possibly because of a unique genome organization. In humans, but not in mice, the Dlg4 and VLCAD (very-long-chain acyl coenzyme A dehydrogenase) genes are arranged in a head-to-head orientation and share a 245-bp overlapping region that contains part of Dlg4 exon 1 and the entire exon 1 of VLCAD, including 62 bp of the protein-coding sequence. Despite the overlap of their 5' ends, Dlg4 and VLCAD exhibit peak mRNA expression in different tissues, suggesting that they are independently regulated at the transcriptional level (43). In CaSki, SiHa, and Yumoto cells (Table 1), Dlg4 mRNA was downregulated. However, as the promoter region is not a CpG island, promoter methylation is not a likely cause of the downregulation. Indeed, Dlg4 mRNA levels were down-regulated in these cell lines but easily detected by RT-PCR.

Possible roles of DLG4 in epithelial cells. We hypothesized that DLG4, as well as DLG1, Scrib, and MAGI-1, might accumulate when HCK1T cultures reached confluence and that this might be altered in cervical cancer cell lines and E6-expressing HCK1T cells. Indeed, DLG4 did accumulate in HCK1T cells at confluence, but unlike DLG1 and others, it was no longer evident 7 days thereafter (Fig. 2B). DLG4 might thus have a specific function different from that of other PDZ domain-containing target proteins. PDZ domains act as protein-protein binding domains to make complexes with other PDZ domain-containing proteins in certain combinations (1, 12, 14, 31). It is likely that DLG4 contributes to complex formation. Indeed, in Drosophila, both Dlg and Scribble are required for cortical translocation of another cortical tumor suppressor, Lgl. We first examined the effect of DLG4 expression in HeLa as well as CaSki cells, but only mild repression of tumorigenicity was observed (Handa, unpublished), although the steady-state level of the ectopic DLG4 in HeLa cells was about one-third of that in CaSki or C33A cells, probably because of 18E6-mediated degradation. Unlike CaSki cells, HeLa cells do not express E-cadherin and do not form a pavement-like sheet at confluence. It is possible that PDZ domain-containing proteins like DLG4 can form functional complexes only when cell-cell contact is sufficiently tight and that a certain threshold level of other PDZ domain-containing proteins is also required for DLG4 to function. Recently, DLG1, MAGI-1, and MUPP1 were shown to suppress transformation of baby rat kidney cells by combinations of E7 or E1A and EJ-ras (41). In this regard, it will be interesting to examine whether other E6 targets can also act as tumor suppressors in the development of cervical cancer. Clearly, further investigations are required to clarify the underlying mechanism(s) whereby these PDZ domain-containing E6 targets could function as tumor suppressors.


arrow
ACKNOWLEDGMENTS
 
This work was supported in part by a Grant-in-Aid for Cancer Research from the Ministry of Health Labor and Welfare and a Grant-in-Aid for Scientific Research on the Priority Area "Cancer" from the Ministry of Education, Culture, Sports, Science, and Technology of Japan to T.K.

We thank Hiroshi Shirasawa (Chiba University) for providing cervical cancer cell lines, Shunsuke Nakagawa (University of Tokyo) for the E6AP plasmid, Dirk Bohmann (University of Rochester) for the His6-Ub plasmid, and Takako Ishiyama, Kei Haga, and Yuki Kakuta for providing expert technical assistance.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: 5-1-1 Tsukiji, Chuo-ku, Tokyo, 104-0045, Japan. Phone: 81-3-3542-2511. Fax: 81-3-3543-2181. E-mail: tkiyono{at}gan2.res.ncc.go.jp. Back

{triangledown} Published ahead of print on 22 November 2006. Back

{dagger} Present address: Department of Operative Dentistry and Endodontology, School of Dentistry, Health Sciences University of Hokkaido, 1757 Ishikari-Tobetu, Hokkaido, 061-0293, Japan. Back


arrow
REFERENCES
 
    1
  1. Christopherson, K. S., B. J. Hillier, W. A. Lim, and D. S. Bredt. 1999. PSD-95 assembles a ternary complex with the N-methyl-D-aspartic acid receptor and a bivalent neuronal NO synthase PDZ domain. J. Biol. Chem. 274:27467-27473.[Abstract/Free Full Text]
    2. 2
  2. Dostatni, N., P. F. Lambert, R. Sousa, J. Ham, P. M. Howley, and M. Yaniv. 1991. The functional BPV-1 E2 trans-activating protein can act as a repressor by preventing formation of the initiation complex. Genes Dev. 5:1657-1671.[Abstract/Free Full Text]
    3. 3
  3. Dowhanick, J. J., A. A. McBride, and P. M. Howley. 1995. Suppression of cellular proliferation by the papillomavirus E2 protein. J. Virol. 69:7791-7799.[Abstract]
    4. 4
  4. El-Husseini, A. E., E. Schnell, D. M. Chetkovich, R. A. Nicoll, and D. S. Bredt. 2000. PSD-95 involvement in maturation of excitatory synapses. Science 290:1364-1368.[Abstract/Free Full Text]
    5. 5
  5. Favre-Bonvin, A., C. Reynaud, C. Kretz-Remy, and P. Jalinot. 2005. Human papillomavirus type 18 E6 protein binds the cellular PDZ protein TIP-2/GIPC, which is involved in transforming growth factor beta signaling and triggers its degradation by the proteasome. J. Virol. 79:4229-4237.[Abstract/Free Full Text]
    6. 6
  6. Frese, K. K., I. J. Latorre, S. H. Chung, G. Caruana, A. Bernstein, S. N. Jones, L. A. Donehower, M. J. Justice, C. C. Garner, and R. T. Javier. 2006. Oncogenic function for the Dlg1 mammalian homolog of the Drosophila discs-large tumor suppressor. EMBO J. 25:1406-1417.[CrossRef][Medline]
    7. 7
  7. Gardiol, D., C. Kuhne, B. Glaunsinger, S. S. Lee, R. Javier, and L. Banks. 1999. Oncogenic human papillomavirus E6 proteins target the discs large tumour suppressor for proteasome-mediated degradation. Oncogene 18:5487-5496.[CrossRef][Medline]
    8. 8
  8. Gewin, L., H. Myers, T. Kiyono, and D. A. Galloway. 2004. Identification of a novel telomerase repressor that interacts with the human papillomavirus type-16 E6/E6-AP complex. Genes Dev. 18:2269-2282.[Abstract/Free Full Text]
    9. 9
  9. Glaunsinger, B. A., S. S. Lee, M. Thomas, L. Banks, and R. Javier. 2000. Interactions of the PDZ-protein MAGI-1 with adenovirus E4-ORF1 and high-risk papillomavirus E6 oncoproteins. Oncogene 19:5270-5280.[CrossRef][Medline]
    10. 10
  10. Grm, H. S., and L. Banks. 2004. Degradation of hDlg and MAGIs by human papillomavirus E6 is E6-AP-independent. J. Gen. Virol. 85:2815-2819.[Abstract/Free Full Text]
    11. 11
  11. Huibregtse, J. M., M. Scheffner, and P. M. Howley. 1993. Cloning and expression of the cDNA for E6-AP, a protein that mediates the interaction of the human papillomavirus E6 oncoprotein with p53. Mol. Cell. Biol. 13:775-784.[Abstract/Free Full Text]
    12. 12
  12. Im, Y. J., J. H. Lee, S. H. Park, S. J. Park, S. H. Rho, G. B. Kang, E. Kim, and S. H. Eom. 2003. Crystal structure of the Shank PDZ-ligand complex reveals a class I PDZ interaction and a novel PDZ-PDZ dimerization. J. Biol. Chem. 278:48099-48104.[Abstract/Free Full Text]
    13. 13
  13. Ishidate, T., A. Matsumine, K. Toyoshima, and T. Akiyama. 2000. The APC-hDLG complex negatively regulates cell cycle progression from the G0/G1 to S phase. Oncogene 19:365-372.[CrossRef][Medline]
    14. 14
  14. Jelen, F., A. Oleksy, K. Smietana, J. Otlewski, Y. J. Im, J. H. Lee, S. H. Park, S. J. Park, S. H. Rho, G. B. Kang, E. Kim, S. H. Eom, C. Nourry, S. G. Grant, J. P. Borg, K. S. Christopherson, B. J. Hillier, W. A. Lim, and D. S. Bredt. 2003. PDZ domains—common players in the cell signaling. Acta Biochim. Pol. 50:985-1017.[Medline]
    15. 15
  15. Kao, W. H., S. L. Beaudenon, A. L. Talis, J. M. Huibregtse, and P. M. Howley. 2000. Human papillomavirus type 16 E6 induces self-ubiquitination of the E6AP ubiquitin-protein ligase. J. Virol. 74:6408-6417.[Abstract/Free Full Text]
    16. 16
  16. Kiyono, T., A. Hiraiwa, M. Fujita, Y. Hayashi, T. Akiyama, and M. Ishibashi. 1997. Binding of high-risk human papillomavirus E6 oncoproteins to the human homologue of the Drosophila discs large tumor suppressor protein. Proc. Natl. Acad. Sci. USA 94:11612-11616.[Abstract/Free Full Text]
    17. 17
  17. Kolly, C., M. M. Suter, and E. J. Muller. 2005. Proliferation, cell cycle exit, and onset of terminal differentiation in cultured keratinocytes: pre-programmed pathways in control of C-Myc and Notch1 prevail over extracellular calcium signals. J. Investig. Dermatol. 124:1014-1025.[CrossRef][Medline]
    18. 18
  18. Kornau, H. C., L. T. Schenker, M. B. Kennedy, and P. H. Seeburg. 1995. Domain interaction between NMDA receptor subunits and the postsynaptic density protein PSD-95. Science 269:1737-1740.[Abstract/Free Full Text]
    19. 19
  19. Kyo, S., M. Nakamura, T. Kiyono, Y. Maida, T. Kanaya, M. Tanaka, N. Yatabe, and M. Inoue. 2003. Successful immortalization of endometrial glandular cells with normal structural and functional characteristics. Am. J. Pathol. 163:2259-2269.[Abstract/Free Full Text]
    20. 20
  20. Lee, S. S., B. Glaunsinger, F. Mantovani, L. Banks, and R. T. Javier. 2000. Multi-PDZ domain protein MUPP1 is a cellular target for both adenovirus E4-ORF1 and high-risk papillomavirus type 18 E6 oncoproteins. J. Virol. 74:9680-9693.[Abstract/Free Full Text]
    21. 21
  21. Lee, S. S., R. S. Weiss, and R. T. Javier. 1997. Binding of human virus oncoproteins to hDLG/SAP97, a mammalian homolog of the Drosophila discs large tumor suppressor protein. Proc. Natl. Acad. Sci. USA 94:6670-6675.[Abstract/Free Full Text]
    22. 22
  22. Makino, K., H. Kuwahara, N. Masuko, Y. Nishiyama, T. Morisaki, J. Sasaki, M. Nakao, A. Kuwano, M. Nakata, Y. Ushio, and H. Saya. 1997. Cloning and characterization of NE-dlg: a novel human homolog of the Drosophila discs large (dlg) tumor suppressor protein interacts with the APC protein. Oncogene 14:2425-2433.[CrossRef][Medline]
    23. 23
  23. Massimi, P., N. Gammoh, M. Thomas, and L. Banks. 2004. HPV E6 specifically targets different cellular pools of its PDZ domain-containing tumour suppressor substrates for proteasome-mediated degradation. Oncogene 23:8033-8039.[CrossRef][Medline]
    24. 24
  24. Matsumine, A., A. Ogai, T. Senda, N. Okumura, K. Satoh, G. H. Baeg, T. Kawahara, S. Kobayashi, M. Okada, K. Toyoshima, and T. Akiyama. 1996. Binding of APC to the human homolog of the Drosophila discs large tumor suppressor protein. Science 272:1020-1023.[Abstract]
    25. 25
  25. Matsumoto, Y., S. Nakagawa, T. Yano, S. Takizawa, K. Nagasaka, K. Nakagawa, T. Minaguchi, O. Wada, H. Ooishi, K. Matsumoto, T. Yasugi, T. Kanda, J. M. Huibregtse, and Y. Taketani. 2006. Involvement of a cellular ubiquitin-protein ligase E6AP in the ubiquitin-mediated degradation of extensive substrates of high-risk human papillomavirus E6. J. Med. Virol. 78:501-507.[CrossRef][Medline]
    26. 26
  26. Migaud, M., P. Charlesworth, M. Dempster, L. C. Webster, A. M. Watabe, M. Makhinson, Y. He, M. F. Ramsay, R. G. Morris, J. H. Morrison, T. J. O'Dell, and S. G. Grant. 1998. Enhanced long-term potentiation and impaired learning in mice with mutant postsynaptic density-95 protein. Nature 396:433-439.[CrossRef][Medline]
    27. 27
  27. Nakagawa, S., and J. M. Huibregtse. 2000. Human scribble (Vartul) is targeted for ubiquitin-mediated degradation by the high-risk papillomavirus E6 proteins and the E6AP ubiquitin-protein ligase. Mol. Cell. Biol. 20:8244-8253.[Abstract/Free Full Text]
    28. 28
  28. Nakamura, H., H. Fukami, Y. Hayashi, T. Kiyono, S. Nakatsugawa, M. Hamaguchi, and K. Ishizaki. 2002. Establishment of immortal normal and ataxia telangiectasia fibroblast cell lines by introduction of the hTERT gene. J. Radiat. Res. (Tokyo) 43:167-174.[Medline]
    29. 29
  29. Naviaux, R. K., E. Costanzi, M. Haas, and I. M. Verma. 1996. The pCL vector system: rapid production of helper-free, high-titer, recombinant retroviruses. J. Virol. 70:5701-5705.[Abstract/Free Full Text]
    30. 30
  30. Nguyen, M. L., M. M. Nguyen, D. Lee, A. E. Griep, and P. F. Lambert. 2003. The PDZ ligand domain of the human papillomavirus type 16 E6 protein is required for E6's induction of epithelial hyperplasia in vivo. J. Virol. 77:6957-6964.[Abstract/Free Full Text]
    31. 31
  31. Nourry, C., S. G. Grant, and J. P. Borg. 2003. PDZ domain proteins: plug and play! Sci. STKE 2003:RE7.[Medline]
    32. 32
  32. Ohshiro, T., T. Yagami, C. Zhang, and F. Matsuzaki. 2000. Role of cortical tumour-suppressor proteins in asymmetric division of Drosophila neuroblast. Nature 408:593-596.[CrossRef][Medline]
    33. 33
  33. Peng, C. Y., L. Manning, R. Albertson, and C. Q. Doe. 2000. The tumour-suppressor genes lgl and dlg regulate basal protein targeting in Drosophila neuroblasts. Nature 408:596-600.[CrossRef][Medline]
    34. 34
  34. Pim, D., M. Thomas, R. Javier, D. Gardiol, and L. Banks. 2000. HPV E6 targeted degradation of the discs large protein: evidence for the involvement of a novel ubiquitin ligase. Oncogene 19:719-725.[CrossRef][Medline]
    35. 35
  35. Romanczuk, H., F. Thierry, and P. M. Howley. 1990. Mutational analysis of cis elements involved in E2 modulation of human papillomavirus type 16 P97 and type 18 P105 promoters. J. Virol. 64:2849-2859.[Abstract/Free Full Text]
    36. 36
  36. Sawada, M., T. Kiyono, S. Nakashima, J. Shinoda, T. Naganawa, S. Hara, T. Iwama, and N. Sakai. 2004. Molecular mechanisms of TNF-alpha-induced ceramide formation in human glioma cells:P53-mediated oxidant stress-dependent and -independent pathways. Cell Death Differ. 11:997-1008.[CrossRef][Medline]
    37. 37
  37. Scheffner, M., J. M. Huibregtse, and P. M. Howley. 1994. Identification of a human ubiquitin-conjugating enzyme that mediates the E6-AP-dependent ubiquitination of p53. Proc. Natl. Acad. Sci. USA 91:8797-8801.[Abstract/Free Full Text]
    38. 38
  38. Stathakis, D. G., K. B. Hoover, Z. You, and P. J. Bryant. 1997. Human postsynaptic density-95 (PSD95): location of the gene (DLG4) and possible function in nonneural as well as in neural tissues. Genomics 44:71-82.[CrossRef][Medline]
    39. 39
  39. Thomas, M., B. Glaunsinger, D. Pim, R. Javier, and L. Banks. 2001. HPV E6 and MAGUK protein interactions: determination of the molecular basis for specific protein recognition and degradation. Oncogene 20:5431-5439.[CrossRef][Medline]
    40. 40
  40. Thomas, M., R. Laura, K. Hepner, E. Guccione, C. Sawyers, L. Lasky, and L. Banks. 2002. Oncogenic human papillomavirus E6 proteins target the MAGI-2 and MAGI-3 proteins for degradation. Oncogene 21:5088-5096.[CrossRef][Medline]
    41. 41
  41. Thomas, M., P. Massimi, C. Navarro, J. P. Borg, and L. Banks. 2005. The hScrib/Dlg apico-basal control complex is differentially targeted by HPV-16 and HPV-18 E6 proteins. Oncogene 24:6222-6230.[CrossRef][Medline]
    42. 42
  42. Watson, R. A., M. Thomas, L. Banks, and S. Roberts. 2003. Activity of the human papillomavirus E6 PDZ-binding motif correlates with an enhanced morphological transformation of immortalized human keratinocytes. J. Cell Sci. 116:4925-4934.[Abstract/Free Full Text]
    43. 43
  43. Zhou, C., and B. Blumberg. 2003. Overlapping gene structure of human VLCAD and DLG4. Gene 305:161-166.[CrossRef][Medline]

Journal of Virology, February 2007, p. 1379-1389, Vol. 81, No. 3
0022-538X/07/$08.00+0     doi:10.1128/JVI.01712-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Lee, M., Vasioukhin, V. (2008). Cell polarity and cancer - cell and tissue polarity as a non-canonical tumor suppressor. J. Cell Sci. 121: 1141-1150 [Abstract] [Full Text]  
  • Garcia-Mata, R., Dubash, A. D., Sharek, L., Carr, H. S., Frost, J. A., Burridge, K. (2007). The Nuclear RhoA Exchange Factor Net1 Interacts with Proteins of the Dlg Family, Affects Their Localization, and Influences Their Tumor Suppressor Activity. Mol. Cell. Biol. 27: 8683-8697 [Abstract] [Full Text]  
  • Cooper, B., Brimer, N., Vande Pol, S. B. (2007). Human Papillomavirus E6 Regulates the Cytoskeleton Dynamics of Keratinocytes through Targeted Degradation of p53. J. Virol. 81: 12675-12679 [Abstract] [Full Text]  
  • Yugawa, T., Handa, K., Narisawa-Saito, M., Ohno, S.-i., Fujita, M., Kiyono, T. (2007). Regulation of Notch1 Gene Expression by p53 in Epithelial Cells. Mol. Cell. Biol. 27: 3732-3742 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Handa, K.
Right arrow Articles by Kiyono, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Handa, K.
Right arrow Articles by Kiyono, T.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»