Figure 1.
The CSB-PGBD3 fusion protein is abundantly expressed by alternative splicing and polyadenylation of the CSB transcript.
(A) The CSB-PGBD3 fusion protein is expressed by alternative splicing of CSB exons 1–5 to the PGBD3 transposase 3′ splice acceptor site, whereas solitary PGBD3 transposase is expressed from a cryptic promoter in CSB exon 5. (B) As a result, the primate CSB locus generates three proteins: full-length CSB, the CSB-PGBD3 fusion protein, and solitary PGBD3 transposase. pA, polyadenylation signal; 5′ ss, 5′ splice donor site; 3′ss, 3′ splice acceptor site.
Figure 2.
The CSB-PGBD3 fusion protein binds to MER85 elements in vivo.
(left) ChIP-PCR in wild type HT1080 cells using antibodies for the N-terminus of CSB pulls down 6 representative MER85 elements with good matches to the Repbase consensus. N-terminal antibodies pull down both CSB-PGBD3 fusion protein and full-length CSB, whereas C-terminal antibodies pull down full-length CSB only (LTG unpublished). No, no antibody control; Ig, anti-mouse IgG nonspecific antibody control; N, CSB N-terminal antibody; C, CSB C-terminal antibody (right) Paired-end ChIP-seq shows enrichment for the same five out of six MER85 elements in CSB-null UVSS1KO cells stably expressing the CSB-PGBD3 fusion protein. Table S1 gives the positions and sequences of all MER85 elements. The 5′ and 3′ ends of MER85s are defined as the same orientation as the transposase ORF in parental PGBD3 elements.
Figure 3.
The PGBD3 transposase binds to the 5′ end of MER85s in vitro.
(A) An electrophoretic mobility shift assay (EMSA) using MER85s and MER85 DdeI restriction fragments. MER85-360 and MER85-427 were excised from plasmid clones using BamHI and EcoRI, then digested with DdeI or left intact. The restriction fragments were end-labeled and mixed with purified PGBD3 transposase. Restriction fragments derived from the 5′ end of each MER85 are marked by an asterisk. (B) Partial sequences of MER85-360 and MER85-427 with the DdeI restriction site indicated. The 5′ MER85 sequences that shifted upon incubation with transposase are highlighted in green. TSD, target site duplication; 5′-TIR, 5′ terminal inverted repeat.
Figure 4.
Fragment overlaps in the vicinity of the PGBD3 transposon reveal strong binding near each of three palindromes.
Occupancy of CSB-PGBD3 near the PGBD3 transposon was assessed by counting the number of overlapping ChIPed fragments at each position. TIR, terminal inverted repeat; ORF, open reading frame. Ordinate indicates the number of overlapping fragments.
Figure 5.
Mutations in the palindromic region reduce PGBD3 transposase binding affinity for MER85s.
Synthetic 42 bp MER85 fragments were mixed with purified PGBD3 or no protein, and used for an electrophoretic mobility shift assay. The binding affinities of the transposase for synthetic 42 bp fragments were normalized to the Repbase consensus sequence (100%) and a scrambled sequence (0%). Only sequence mismatches are displayed; positions that match the Repbase consensus are indicated by periods. TSD, target site duplication; TIR, terminal inverted repeat.
Figure 6.
The CSB-PGBD3 fusion protein binds preferentially to the 5′ palindromic sequence of all bound MER85s in the human genome.
Paired-end sequence reads near bound MER85s were used to reconstruct the location of immunoprecipitated fragments relative to the 5′ target site duplication (TSD) of each element. Cumulative fragment overlaps were calculated by summing the number of fragments from each element that overlapped each position relative to the 5′ TSD. TIR, Inverted Terminal Repeat.
Figure 7.
Non-MER85 peaks are enriched for TRE, TEAD1, and CTCF binding site motifs.
(left) Analysis using Multiple Em for Motif Elicitation (MEME). Sequences within 50 bp of non-MER85 peak summits were submitted to MEME to identify overrepresented motifs. (right) Analysis using Tomtom motif comparison tool. Position specific frequency matrices for the motifs identified by MEME were submitted to TOMTOM to identify matching transcription factor binding sites. The most significant matches for each result are shown. *AP-1 motif was annotated jundm2_secondary, Jun dimerization protein 2 secondary motif (UniPROBE mouse database); TEAD1, TEA domain family member 1 (JASPAR core 2009 database); CTCF, CCCTC binding factor (JASPAR core 2009 database).
Figure 8.
CSB-PGBD3 peak summits coincide with the TRE, TEAD1, and CTCF motifs.
Average fragment overlaps in the vicinity of TRE, TEAD1, and CTCF motifs were plotted for the CSB-PGBD3 ChIP-seq data. The overlaps peak sharply and symmetrically around the motifs, consistent with tethering of the CSB-PGBD3 fusion protein to the corresponding transcription factors through protein-protein interactions.
Figure 9.
c-Jun co-immunoprecipitates with the CSB-PGBD3 and CSB-eGFP proteins, but not with eGFP-PGBD3.
Nuclear lysates from UVSS1KO cells stably expressing FLAG-HA-tagged CSB, CSB-PGBD3, CSB-eGFP, and eGFP-PGBD3 were immunoprecipitated using anti-FLAG, anti-c-Jun, and a nonspecific antibody. (A) Western blots probed with anti-c-Jun antibodies. c-Jun is immunoprecipitates with anti-FLAG antibodies in cells expressing FLAG-HA-tagged CSB-PGBD3 or CSB-eGFP, but not full-length CSB or eGFP-PGBD3. (B) Western blots probed with anti-FLAG antibodies. FLAG-HA-tagged CSB-PGBD3 and CSB-eGFP immunoprecipitate with anti-c-Jun antibodies. * Denotes lane with uncharacteristically high background. The same nonspecific antibody was used for all negative control samples, which leads us to believe this band is an artefact due to contamination rather than a true IP of CSB-eGFP. IP, antibodies used for immunoprecipitation; Ig, anti-mouse IgG nonspecific antibody control; FL, mouse monoclonal anti-FLAG antibody; c-Jun, anti-c-Jun antibody.
Figure 10.
CSB-PGBD3 and CSB-eGFP co-immunoprecipitate with RNA polymerase II (RNAPII).
(A) HT1080 whole cell lysates were immunoprecipitated using anti-RNAPII CTD antibodies, N-terminal CSB antibodies, or nonspecific antibodies. CSB and CSB-PGBD3 were detected by western blotting with antibodies against the N-terminus of CSB. (B) UVSS1KO cells expressing FLAG-HA tags only, FLAG-HA-tagged CSB, or FLAG-HA-tagged CSB-PGBD3 were immunoprecipitated using antibodies for FLAG tags or a nonspecific antibody control. RNAPII was detected by western blotting with antibodies against the CTD of RNAPII. (C) UVSS1KO cells expressing FLAG-HA-tagged CSB-PGBD3, CSB-eGFP, or eGFP-PGBD3 were immunoprecipitated with antibodies against the CTD of RNAPII or a nonspecific antibody control. CSB-PGBD3, CSB-eGFP, and eGFP-PGBD3 were detected by western blotting with anti-FLAG antibodies. Ig, anti-mouse IgG nonspecific control; Pol, anti-RNAPII CTD; N, anti-CSB N-terminus; FL, anti-FLAG.
Figure 11.
CSB-LacI and eGFP-PGBD3 induce partial up-regulation of genes regulated by CSB-PGBD3.
Average signal log ratios from quantitative PCR (QPCR) of genes regulated by CSB-PGBD3, CSB-eGFP, or eGFP-PGBD3 expression in UVSS1KO cells compared to cells expressing FLAG-HA-tags alone. Orange cells: increased expression; Blue cells: decreased expression; No color: expression change was less than 2-fold (1 signal log ratio); darker color indicates a larger change in expression.
Figure 12.
ChIP–seq data suggest multiple roles for the CSB-PGBD3 fusion protein in gene regulation.
Top, Transcription factor binding sites before the MER85 replicative burst. Middle, We anticipated that the CSB-PGBD3 fusion protein would bind to MER85 elements throughout the genome and regulate nearby genes through interactions mediated by the N-terminal CSB domain. Bottom, Our ChIP-seq data revealed that CSB-PGBD3 binds over TRE, CTCF, and TEAD motifs, and regulates genes near TRE motifs in CSB-null cells. Full-length CSB may facilitate or suppress these interactions through chromatin remodeling or competition for factors that also bind the N-terminal CSB domain of CSB-PGBD3. The CSB-PGBD3 fusion protein does bind to MER85 elements as anticipated, but these sites may function as a reservoir for CSB-PGBD3 protein or mediate chromatin looping, perhaps by interaction with CTCF.