Fig 2: The preferred genomic-binding site for SETMAR is the ancestral TIR sequence.A, the program rGADEM (21) was used to perform a de novo motif analysis of SETMAR-binding sites identified through ChIP-Seq analysis; the ancestral TIR sequence was identified as the most common motif bound by SETMAR. The SETMAR-binding sites identified by subtracting the R371A from the WT ChIP-Seq peaks were analyzed using rGADEM to determine a consensus DNA motif for SETMAR binding. Nucleotides 11 to 29 of the consensus motif represent the core 19 bp binding element. The only nucleotide involved in major groove hydrogen bonding interactions for which there is degeneracy is the C/T at position 27. B, the program PAVIS (59) was used to analyze the ChIP-Seq data; the majority of SETMAR-binding sites are located in intergenic regions (56.6%) or introns (33.7%). Only 9.8% of sites are located within 10,000 base pairs or less upstream of a transcription start site. C, the distribution of distances to the nearest TSS from the 5' end of the TIR sites with two or fewer mutations. Distances were capped to 20,000 bp shown in 100 bp bins. D, distribution of the distances to the nearest TSSs from the 5' end of TIR sequences. The distribution of distances to the nearest TSS from the 5' end of the best TIRs (those identified as matches using a PWM) is plotted in 100 kb bins and exhibits a central peak with ~22% of TIRs falling within 100 kb of the closest TSS. ChIP-Seq, chromatin immunoprecipitation sequencing; PWM, position-weighted matrix; TIR, terminal inverted repeat; TSS, transcription start site.

Fig 3: Loss of SETMAR expression alters alternative splicing (AS). The pie chart and schematic diagram depict the distribution of 255 AS events common between both SETMAR KO clones.

Fig 4: A position-weighted matrix was generated from all SETMAR-bound TIR sequences with two or fewer mismatches compared with the perfect 19-base pair sequence identified by ChIP-Seq in human HEK293T cells. Using the PWM, 5253 TIR sequences were identified in the human genome, 763 of which were perfect, 2361 had one mismatch, and 1352 had two mismatches. The heat map shows the distribution of SETMAR motif ChIP peak hits among the human Genome Evolutionary Rate Profiling (GERP) constrained elements across the 27 selected primates. Each row in the heat map represents a block of human genome sequence containing all those GERP constrained elements that are not farther than 10 Kb from the adjacent ones (http://ftp.ensembl.org/pub/release-98/bed/ensembl-compara/90_mammals.gerp_constrained_element/). These blocks were sorted on their positions on the chromosomes ordered 1 to 22, X and Y, each represented by a different color on the left side bar. The data represent the number of SETMAR motif ChIP peak hits within each such block in the human genome and the corresponding homologous regions in the other selected primate genomes as indicated on the color key. This clustering recapitulates current phylogenetic classification of primates. ChIP-Seq, chromatin immunoprecipitation sequencing; HEK293T, human embryonic kidney 293T cell line; PWM, position-weighted matrix; TIR, terminal inverted repeat.

Fig 5: Structural basis for SETMAR–TIR interactions. A schematic of the SETMAR protein is shown with the SET domain and two domains derived from the mariner transposase gene MAR, DBD referring to the DBD derived from the Hsmar1 transposase at the top. A, SETMAR recognizes TIR-DNA through interactions mediated by HTH1, HTH2, and AT hook motifs. The SETMAR DBD dimerizes through interactions of HTH1. Residues that interact with DNA are shown as space filling models. A semitransparent surface model (pink for the C chain and cyan for the B chain) is shown along with a stick model for the TIR-DNA. B, the 25-mer DNA duplex derived from the Hsmar1 TIR used for cocrystallization with SETMAR DBD is shown with its core 19 bp binding element shaded in gray. A schematic diagram of SETMAR DBD is shown above with the relative positions of HTH1 and HTH2 motifs. Red letters are sequence-specific interactions identified from the crystal structure. Nucleotides are numbered according to the TIR complex monomer structure. C, close-up view of base-specific contacts made by R371 in HTH1 motif. Two hydrogen bonds are formed between the side chain and the base bG5 (position 5 guanine of chain b, see Fig. 1B for number scheme). The DNA is shown as a stick model with a semitransparent surface model in light gray in (C–E). D and E, close-up view of base-specific contacts in HTH2 motif of TIR complex. Key residues (R417, H427, S428, and R432) make seven hydrogen bonds with DNA nucleobases. Details of AT hook interactions involving R392 and R395 are shown in Fig. S3. DBD, DNA-binding domain; HTH, helix–turn–helix; TIR, terminal inverted repeat.