Fig 2: The human and mouse ZNF521/Zfp521 promoter regions contain SPI1 and HOXC13 predicted binding sites. a: Human chromosome 18q11.2. Scale bar = 2.5 Kb. The GC% in 5‐base windows is shown at the top. The human ZNF521 promoter region used in this study is shown as a black box labelled promoter. ZNF521 exon 1 is annotated. H3K27Ac marks, DNaseI hypersensitivity hot spots in K562 erythroleukemia cells, and DNaseI hypersensitivity hot spots in GM12878 lymphocyte cells are shown. Sequence conservation among 100 vertebrates is shown at the bottom. The locations of the predicted SPI1 (Sa and Sb) and HOXC13 (H) binding sites are labeled. Sites Sa and Sb overlap with K562 DNase I hypersensitivity hot spots. b: Mouse chromosome 18qA1. Scale bar = 2.5Kb. The GC% in 5‐base windows is shown at the top. The mouse Zfp521 region used in this study and in experimental constructs is shown as a black box labeled promoter. Zfp521 exon is annotated. DNaseI hypersensitivity hotspots in CD19+ B‐cells are annotated. Sequence conservation among placental mammals is shown at the bottom. The locations of the predicted SPI1 (Sa and Sb) and HOXC13 (H) binding sites are labelled. Sites Sa and Sb overlap with CD19+ B‐cell DNase I hypersensitivity hot spots. c: The mouse Zfp521 promoter sequence. Predicted SPI1 binding sites are shown in gray, and the predicted HOXC13 binding site is underlined. d: Zfp521 reporter constructs used in transfection assays. Gray box represents predicted HOXC13 binding site, black boxes represent SPI1 binding sites. e: Reporter assay measuring Zfp521 promoter activation following HEK293 cell transfection with SPI1 (light gray bars), HOXC13 (medium gray bars), or SPI1 and HOXC13 cotransfection (dark gray bars). Relative expression levels are shown as normalized to transfection with reporter alone (black bars). Cotransfection of SPI1 and HOXC13 shows a statistically significant activation of the 1Kb reporter construct when compared to vector control (t‐test; p < 0.05). Error bars represent standard deviation of 3 independent transfections, each with three technical replicates

Fig 3: Protein‐protein interaction between Spi1 and HOXC13 tagged constructs. a: Co‐IP demonstrating that SPI1 and HOXC13 proteins have a physical interaction. The truncated form of SPI1 (aa1‐168) does not interact with HOXC13, but a missense mutant (K219A) of SPI1 does. b: Input samples. Mutant forms of SPI1 are shown (aa1‐168 and K219A). Vector = cotransfection with FLAG‐HOXC13 and empty vector lacking SPI1 sequence as negative control

Fig 4: Analysis of Zfp521 transcriptional activation by SPI1 and HOXC13. a: SPI1 and HOXC13 protein isoforms encoded by expression vectors. The full-length isoforms are shown on top and truncation mutants shown below. Both truncation mutants remove the DNA binding domain from the protein. b: Reporter assay measuring Zfp521 promoter activation following HEK293 cell cotransfection of wild type SPI1 and HOXC13 (black bar), transfection of truncated SPI1 and wild type HOXC13 (charcoal gray bar), transfection of wild type SPI1 and truncated HOXC13 (medium gray bar), or truncated SPI1 and truncated HOXC13 (light gray bar). Results were normalized to transfection with an empty expression vector (white bar). Only cotransfection of both wild-type proteins demonstrated statistically significant gene activation when compared to empty expression vector control (t-test; p < 0.05). Error bars represent standard deviation of 3 independent transfections, each with three technical replicates. c: SPI1 and HOXC13 regulation of the Zfp521 promoter is dose-dependent. Decreasing the amounts of transfection plasmid of either SPI1 or HOXC13 protein causes a decrease in promoter activation in HEK293 cells. The graph shows relative expression levels normalized to transfection with reporter construct only. d: Evaluation of SPI1 and HOXC13 binding to the Zfp521 promoter. Transfection & IP: Protein-DNA cross-linking followed by immunoprecipitation of FLAG-SPI1 or FLAG-HOXC13 with an anti-FLAG antibody allows amplification of the Zfp521 promoter region, but immunoprecipitation using irrelevant antibody (IgG) does not precipitate the Zfp521 promoter. Extracts from HEK293 cell transfections of the FLAG empty vector control do not show amplification of the Zfp521 promoter following immunoprecipitation. The GFP vector control DNA is not detected in any samples following immunoprecipitation. Right panel: Input samples show the presence of Zfp521 and GFP vector control DNA samples

Fig 5: Knockdown of SPI1 and HOXC13 reduces ZNF521 expression, and expression of Spi1 and HOXC13 can rescue Zfp521 knockdown cell defects. a: SPI1 and HOXC13 were knocked down by shRNA transfection in THP-1 cells, either in combination (black bars) or individually (gray bars). A control vector with a scrambled noneffective shRNA sequence was used as a control (light gray bars). Gene expression of ZNF521, SPI1, and HOXC13 were assayed in each knockdown condition (labels on x-axis). Knockdown of both SPI1 and HOXC13 resulted in significantly reduced ZNF521 expression as compared to control shRNA transfection (t-test; p < 0.05). Individual knockdowns had a less profound reduction in ZNF521 expression levels. Expression levels of SPI1 and HOXC13 confirm that the shRNA transfections reduced gene expression of each gene as expected. b: Using a knockdown rescue assay, we compared the ratio of viable cells on day 7 post-transfection for cells that were initially transfected with Zfp521 knockdown shRNA plasmids or control plasmids to cells that had the original knockdown plus a transfection on day 3 of Spi1 and HOXC13 expression constructs (day 7+). Cell viability is recovered when Spi1 and HOXC13 are cotransfected into BCL1 cells after Zfp521 shRNA transfection, and shows a significant increase when compared to cells with initial mock transfection (t-test; p < 0.05). Cells with initial empty vector or a scrambled Zfp521 shRNA sequence transfection do not show a similar increase in cell viability after rescue transfection. c: The percentage of dead cells identified by trypan blue staining in BCL1 cell cultures following Zfp521 knockdown or control transfections on day 3 (left) is lower than the percentage of cells on day 7 (middle). When Spi1 and HOXC13 expression constructs are cotransfected on day 3, the percentage of dead cells on day 7+ (right) is reduced as compared to cells on day 7 that did not receive rescue plasmid transfections (t-test, p < 0.05). d: Caspase 3/7 activity was measured in BCL1 cells 7 days after transfection with Zfp521 shRNA or control plasmids (left). Introduction of Spi1 and HOXC13 expression vectors by cotransfection on day 3 resulted in a significant reduction in Caspase 3/7 activity on day 7+ (middle), with no significant difference in cells initially transfected with Zfp521 shRNA as compared to cells with any other initial transfection condition. Cells transfected with a control empty vector on day 3 did not show a similar reduction in Caspase 3/7 activity (day 7 + C; right). NS = nonsignificant. In all panels error bars represent standard deviation of 3 independent transfections, each with three technical replicates