Figure 1.
NF2 suppresses T-antigen protein expression and T-antigen mediated activation of JCVE and JCVL promoters.
U-87 MG glioblastoma cells were transiently transfected with luciferase reporter constructs encoding either the JCVE (a, c) or JCVL (b, d) promoters, and plasmids encoding JCV T-antigen or HA tagged NF2 (HA-NF2). Immunoblotting of protein lysates revealed a dose-dependent decline in the expression of T-antigen, in the presence of increasing amounts of HA-NF2 (a, b lanes 5–8). Extracts were also used to conduct luciferase assays to determine JCVE (c) or JCVL (d) activation in response to these various transfection conditions. Results are shown here as fold change in luciferase activity, relative to the activity of the promoter alone. NF2 does not appreciably affect the activities of either promoter, while T-antigen greatly enhances the activity of both promoters. In combination, NF2 does prevent T-antigen from enhancing the activities of JCVE and JCVL.
Figure 2.
NF2 leads to accumulation of T-antigen mRNA, but does not affect its alternative splicing.
Schematic showing all alternatively spliced forms of JCV T-antigen, including the pre-mRNA and both large and small T-antigen mRNAs, and location of primers used for PCR analysis are shown (A, black primers). PCR amplification of cDNA derived from T-antigen expressing BSB8 cells (B, lane 3), detected both large and small T-antigen mRNAs, but not pre-mRNA. Co-transfection of U-87 MG cells with expression plasmids for T-antigen and T7-SF2, an inhibitor of splicing, resulted in an accumulation of pre-mRNA and a reduction in large and small T-antigen mRNAs (B, lane 4). Co-transfection with T-antigen and increasing amounts of HA-NF2 plasmids did not alter the pattern of mRNAs detected (C, compare lanes 6 to 7–9). Q-PCR analysis was also conducted for T-antigen copy number (A, grey primers), expressed as copies of T-antigen mRNA per copy number of the control mRNA GAPDH*, with standard deviations of the mean shown here (D). This assay revealed a dose dependent accumulation of T-antigen mRNA in the presence of NF2. *Note that GAPDH primers used in Panel C only detect human and not mouse GAPDH, therefore a band representing GAPDH from the BSB8 mouse cell line was not detected (C, lane 3).
Figure 3.
NF2 utilizes a proteasomal-dependent mechanism to promote the degradation of ubiquitin-bound T-antigen.
To determine if NF2 utilizes the proteasome to enhance T-antigen protein turnover, U-87 MG cells, transfected with HA-NF2 and T-antigen, were treated with the general proteasome inhibitor, MG132 (A, right), or vehicle (A, left). As shown earlier (Figures 1A and 1B), enhanced expression of NF2, suppresses that of T-antigen (A, left top and bottom). However, this affect was abolished with the addition of MG132 (A, right top and bottom). Band intensity representing T-antigen expression was quantified, normalized to its loading control, and shown as fold change relative to lane 1 (A, bottom left and right). p21, a protein known to be subjected to proteasomal degradation, was immunoblotted for in extracts of untransfected U-87 MG cells or those transfected with HA-NF2 or T-antigen alone, exposed to increasing concentrations of MG132 (B). Expression of this protein increased with drug exposure (B, compare lanes 1, 4, and 7 to subsequent lanes), indicating the effectiveness of MG132 to inhibit degradation of proteins subjected to proteasomal degradation. The ability of T-antigen to bind to ubiquitin, a key post-translational protein attached to proteasome targeted proteins, was assessed by Co-IP. U-87 MG cells were transfected with combinations of Myc tagged p53 (Myc-p53), T-antigen, and HA tagged ubiquitin (HA-Ub), protein complexes were immunoprecipitated with an anti HA antibody or normal mouse serum. Subsequent immunoblotting demonstrated that p53 and T-antigen were not precipitated by the HA antibody alone (C, lanes 2 and 6), but did when co-transfected with HA-Ub (C, lanes 3–4 and 7). Collectively, these results provide the first evidence that T-antigen can bind ubiquitin and suggest that T-antigen is subjected to proteasome-mediated degradation.
Figure 4.
NF2 requires its FERM domain to bind and suppress T-antigen expression.
To elucidate the domain of NF2 responsible for the downregulation of T-antigen protein expression, U-87 MG cells were cotransfected with T-antigen and increasing amounts of either HA-FERM or HA-ΔFERM NF2 truncation mutants (A). Immunoblotting revealed a dose-dependent decline in T-antigen expression in the presence of HA-FERM (A, lanes 1–3), but not HA-ΔFERM (A, lanes 4–6), indicating that the FERM domain is necessary for this mechanism. Band intensity representing T-antigen expression was quantified, normalized to its loading control (B). Lysates from U-87 MG cells co-transfected with T-antigen, HA-FERM, and HA-ΔFERM were utilized for Co-IP with T-antigen antibodies (C). In this assay, we demonstrate the ability of T-antigen to bind to the FERM (C, lane 5), but not ΔFERM (C, lane 10) region of NF2, further implicating FERM to be the necessary region facilitating the interaction among NF2 and T-antigen.
Figure 5.
NF2, T-antigen, and p53 complex in human glioblastoma cells and are detected in high grade glioma samples.
Immunoprecipitation of HA-NF2, T-antigen, and Myc-p53 proteins was performed using extracts from transiently transfected U-87 MG cells with antibodies against T-antigen (A), p53 antibody (B), or normal mouse serum as negative controls (A and B, lane 8), followed by immunoblotting to detect the presence of T-antigen, HA-NF2, or Myc-p53. As previously described, T-antigen binds NF2 (A, lane 4) and p53 (A and B, lane 6), individually. These proteins form a ternary complex (A and B, lane 7), however NF2 and p53 do not bind in the absence of T-antigen (B, lane 5). Protein lysates were prepared from fresh tissue specimens of surgically resected high grade gliomas utilized for immunoblotting revealed the presence of NF2 and p53 in an anaplastic oligodendroglioma (A, lane 2), as well as three glioblastomas (C, lanes 3–5). T- antigen was also detected in two of these tumor samples (C, lanes 2 and 4). Further, tumor tissue obtained from surgically resected high grade gliomas tumor, was paraffin embedded and used for immunohistochemistry. In the glioblastoma sample shown in Panel C, lane 4 (Gli1105), we detected cytoplasmic NF2 expression (Figure 5D). In comparable sections of the same glioblastoma sample T-antigen (E) and p53 (F) were localized to the nucleus. Panels D-F, original magnification x400; insets, original magnification, x1000; hematoxylin counterstain.