This study reports that the histone acetyltransferases p300 and CBP directly interacted with and acted as transcriptional coactivators for transcription factor REL. Also, p300 and REL were found in a complex in nuclear extracts from both REL-transformed chicken spleen cells and the DLBCL cell line RC-K8, which has constitutively nuclear REL activity. Given that REL’s oncogenic activity depends on its ability to activate transcription, these results suggest that the REL-p300 interaction plays a role in REL-dependent transformation. Furthermore, we show that the RC-K8 cell line exclusively expressed a C-terminally truncated version of p300, due to a genomic loss of 3′ sequences from one allele of EP300. These results represent the first identification of a p300 truncation in a B-lymphoma cell line.
The RC-K8 DLBCL cell line is noteworthy because of the presence of mutations in genes encoding four components of the NF-κB/REL pathway: REL, IκBα, A20, and p300. We know of no other example of a tumor cell line that has four mutations which impact a single signaling pathway. First, RC-K8 cells express a C-terminally truncated REL protein, REL-NRG, in which sequences of the RHD are fused to non-REL gene sequences [28
]; because the TAD sequences are missing in REL-NRG, the protein can no longer activate transcription [5
]. Second, RC-K8 cells contain inactivating mutations in the REL inhibitor IκBα and no IκBα protein can be detected in these cells; ectopic expression of functional IκBα inhibits the proliferation of RC-K8 cells [5
]. Third, there are inactivating mutations in the gene encoding A20, an upstream negative regulator of NF-κB, and expression of wild-type A20 in RC-K8 cells induces apoptosis [30
]. Finally, as we show herein, RC-K8 cells express a truncated version of p300, a coactivator for REL-dependent transactivation. Given that chronic, low level REL transactivation is optimal for avian B-cell transformation [7
], it is likely that RC-K8 cells have fine-tuned their level of REL-dependent target gene transactivation by balancing mutations that positively influence REL activity (mutations in IκBα and A20) with ones that reduce REL’s transactivation ability (REL-NRG and p300ΔC). In this model, inactivation of IκBα provides the major impetus for constitutively nuclear REL activity. However, because high level expression of REL is toxic in certain settings [32
], mutations that created the DNA-binding competent but transcriptionally inactive REL-NRG protein and the p300ΔC mutant protein may have been selected to dampen REL-driven transactivation of target genes. In addition, the interaction of p300ΔC with REL may block the ability of CBP to compensate for the loss of normal p300 co-activating function in RC-K8 cells. No RelA-containing nuclear complexes are detected in RC-K8 cells, suggesting that the constitutive NF-κB activity is exclusively due to REL complexes [5
]. p300ΔC was also expressed in RC-K8 cells at a much lower level than wild-type p300 in both A293 cells and five B-lymphoma cell lines ().
A20 acts as a negative upstream regulator of NF-κB activity by deubiquinating TRAF6 and modifying the ubiquitination of RIP1 [33
]. Whether A20 mutations contribute to the oncogenic state of RC-K8 cells by promoting unrestrained NF-κB activity is not clear. That is, expression of wild-type A20 was shown to induce apoptosis in RC-K8 cells [30
]; however, it was conspicuously not shown that expression of A20 reduces nuclear NF-κB complexes in these cells. Indeed, given that there is no IκBα protein in RC-K8 cells, the reconstitution of normal A20 upstream activity would not be expected to reduce the amount of nuclear REL protein. Thus, at least in RC-K8 cells, A20 mutations may either contribute to oncogenesis by affecting NF-κB/REL at a downstream step or by affecting other pathways. Indeed, A20, through TRAF6 deubiquitination, also negatively regulates MAPK activity [35
Based on Northern blotting and quantitative RT-PCR of mRNA and genomic DNA, sequences from exons 18–31 appear to be missing from one of the EP300
loci in RC-K8 cells. These results indicate that p300ΔC contains only those amino acid residues encoded by EP300
exons 1–17 (aa 1-1047); therefore, p300ΔC would be missing the HAT domain (aa 1224-1669) [36
] and the entire C-terminal transactivation domain (). As such, p300ΔC would be expected to be unable to enhance transcriptional activation by REL because other studies have shown that the HAT domain is required for p300′s transcriptional enhancement ability [10
]. Also, the N- and C-terminal regions of p300 form a bipartite transctivation domain that recruits TBP, TFII
B, and RNA polymerase [10
], and this function would also be abolished in p300ΔC. Based on the size of the p300ΔC protein on SDS-polyacrylamide gels (~160 kDa), we suggest that p300ΔC is fused to approximately 45 kDa of non-p300 sequences, because exons 1–17 contain 1047 codons, which would yield approximately 115-kDa of p300 polypeptide.
Although our RT-PCR experiments indicate that some wild-type p300 mRNA is expressed in RC-K8 cells (), we have not been able to detect any wild-type p300 protein in these cells. Of note, both wild-type and mutant IκBα mRNAs are also expressed in RC-K8 cells, but no wild-type IκBα protein could be detected [5
]. These results raise the intriguing possibility that the expression of certain mutant mRNAs (e.g., for p300 and IκBα) in RC-K8 cells can suppress translation from the corresponding wild-type mRNAs.
To our knowledge, this report provides the first evidence of a p300 truncation in a B-cell lymphoma. Nevertheless, inactivating mutations in p300, many of which result in C-terminal truncations, have been identified in other types of human tumors and tumor cell lines. Previous reports of p300 truncating mutations have been mostly in cancers of epithelial origin, including colorectal, gastric, breast, pancreatic, cervical, and ovarian [9
]. Most cell lines containing p300 truncations have deleted or silenced second EP300
alleles, and the colorectal cancer cell line HCT116 contains different mutations in its two EP300
]. Expression of wild-type p300 slows the growth of two cancer cell lines with biallelic inactivating mutations in EP300
]. Moreover, p300 knockout mice develop histiocytic sarcomas [41
]. Based on these types of data, p300 has been suggested to be a tumor suppressor. Alterations of p300 have been less frequently found in hematopoietic malignancies. Translocations resulting in fusions of p300 to monocytic leukemia zinc-finger protein (MOZ) in acute myeloid leukemia patients and to mixed lineage leukemia (MLL) in myelodysplastic syndrome patients have been reported [42
Future work is needed to identify the precise genetic alteration that results in expression of the p300ΔC protein in RC-K8 cells, to determine whether p300ΔC contains non-p300 sequences, and to determine whether the p300ΔC protein contributes to the growth, survival, or malignant phenotype of RC-K8 cells. Along these lines, it will be interesting to determine whether overexpression of wild-type p300 or knockdown of p300ΔC can affect the growth or survival of RC-K8 cells or the interaction of REL with CBP.