Genome-scale functional profiling of the mammalian AP-1 signaling pathway

Abstract

Large-scale functional genomics approaches are fundamental to the characterization of mammalian transcriptomes annotated by genome sequencing projects. Although current high-throughput strategies systematically survey either transcriptional or biochemical networks, analogous genome-scale investigations that analyze gene function in mammalian cells have yet to be fully realized. Through transient overexpression analysis, we describe the parallel interrogation of approximately 20,000 sequence annotated genes in cancer-related signaling pathways. For experimental validation of these genome data, we apply an integrative strategy to characterize previously unreported effectors of activator protein-1 (AP-1) mediated growth and mitogenic response pathways. These studies identify the ADP-ribosylation factor GTPase-activating protein Centaurin alpha1 and a Tudor domain-containing hypothetical protein as putative AP-1 regulatory oncogenes. These results provide insight into the composition of the AP-1 signaling machinery and validate this approach as a tractable platform for genome-wide functional analysis.

Fig. 3.

siRNA-mediated functional mapping of AP-1 growth effectors. (A) HEK293 cells were transfected with indicated siRNAs (columns) and subsequently transfected in a pairwise manner with indicated cDNAs (rows), along with an AP-1(PMA)-dependent luciferase reporter construct and a constitutive β-galactosidase reporter vector. After 48 h of incubation, fractional luciferase/β-galactosidase values were determined for each well and further normalized to the activities of three nonspecific siRNA control wells for each transfected cDNA. The resulting data were analyzed by using hierarchical clustering, visualized in treeview (Stanford University), and depict the average of at least three experiments. Effects of siRNA molecules on PMA-stimulated AP-1 activities are also shown. (B) shRNAs targeting predicted growth, and proliferation modulators transitional epithelia response 1 and BLK, and a nonspecific control (CD29), were transfected into HEK293 cells. Cultures were treated with PMA (100 ng/ml) for 24 h as indicated, and cell lysates were interrogated for c-Fos induction through Western blot analysis. Levels of α-tubulin protein were also detected to serve as loading controls.

Fig. 1.

Genomewide functional annotation of cancer-related signalsomes. (A) Approximately 20,000 cDNAs were cotransfected in parallel into 384-well tissue culture plates with appropriate luciferase reporter plasmids. After a 48-h incubation period, cells underwent homogenous lysis, and relative luciferase activation levels were determined. (B) Frequency distributions of reported activities from assays by using AP1(PMA)-, p53-, and Epo-dependent reporters in HEK293, HCT116, and HepG₂ cells, respectively, are shown on normal (black curve) and enlarged (olive curve) scales. Results depict the average of screens assayed in duplicate and are plotted against fold change and standard deviations from the respective experimental means. (C) Approximately 15–30 activators identified in each screen were assayed for specificity by evaluating activities across AP-1, p53, and Epo-responsive reporters. Results were subsequently normalized and assessed by a hierarchical clustering algorithm. Relative activation by modulators (rows) in each reporter assay (columns) is depicted by a red (higher than median activity) to yellow (lower than median activity) continuum color scheme. Clusters of pathway-specific and likely cross-modulatory activities are indicated. (D) Statistics of identified pathway effectors based on characteristics reported or inferred from the literature are shown. Numbers inside circles represent the percentage of total validated pathway activators.

Fig. 2.

Functional validation of AP-1 modulators. (A) The activities of 25 activators of AP-1(PMA) signaling were profiled in triplicate against AP-1-related reporter constructs sensitive to growth and proliferation stimuli [AP-1(PMA), SRE, and NFATc response elements (plus 2 μm of ionomycin)] and stress/cytokine induction (cAMP response element-binding protein response elements, AP-1, IFN response elements), as described. To account for the differential dynamic range of activity inherent to each response element, reporter (columns) and gene (rows) activities were normalized and clustered by using a hierarchical algorithm and are depicted by a red (high relative activity) to yellow (low relative activity) continuum color scheme. Predicted response pathways are based on clustered gene activities on a subset of reporter constructs. Gene annotations reflect official gene symbols designated by locuslink (www.ncbi.nlm.nih.gov/LocusLink), with the exception of previously unannotated genes, which have been assigned interim gene symbols indicated in Table 5. Encoded proteins with known or inferred activities related to predicted pathways, respectively, are indicated by an asterisk (*). (B) Induction of endogenous c-Fos and c-Jun protein levels by transfection of “growth” and “general” responsive subsets of indicated AP-1 modulators, as determined in A, into HEK293 cells was assessed through Western blot analysis. Cell lysates from parental vector (pCMVSport6) transfected HEK293 cells were used as a negative control, and c-FOS- and c-JUN-transfected and PMA-treated cell extracts were used as positive controls. Levels of ERK1 and ERK2 were detected to account for loading. (C) Coexpression based on microarray analysis of canonical (red text) and identified (black text) AP-1 pathway members in primarily lymphoid cell types [WB (whole blood), CD33+ (myeloid, bone marrow), CD14+ (monocyte, peripheral blood), BDCA4+ (dendritic, peripheral blood), CD58+ (natural killer, peripheral blood), CD4⁺ (T cell, peripheral blood), CD8+ (T cell, peripheral blood), CD19+ (B cell, peripheral blood), CD105+ (endothelial, bone marrow), CD34+ (hematopoietic progenitor, bone marrow), MOLT4 (lymphoblastic leukemia), B_lymph (lymphoblastic leukemia), Raji (Burkitt's lymphoma), HL60 (promyelocytic leukemia), Daudi (Burkitt's lymphoma), K562 (chronic myelogenous leukemia), thymus, tonsil (adult, pooled), and LymNd (lymph node, adult)] are shown by a red (high relative expression) to green (low relative expression) continuum color scheme. Relative mRNA levels shown were derived from comparisons of >50 random primary tissue and cell types (http://expression.gnf.org).

Fig. 4.

Identification of candidate oncogenes associated with AP-1 proliferation. (A) Primary CEF expressing predicted modulators of AP-1-mediated proliferation were plated at a density of ≈5 × 10⁵ in six-well plates, and cell counts were initially taken after 24 h (day 0) and at the indicated time points. Ectopically expressed genes (solid line) are indicated on each graph and compared with RCAS only control (dashed line). Results are shown as the function of initial seeding density (day 0) and depict the average of three experiments. (B) CEF cells were infected with BLK, c-JUN, or control RCAS retroviruses, and after 24 h, cells were overlaid with agar medium and maintained in culture for 10–14 days. (C) CEF cells were infected with retroviruses encoding c-JUN, TYRO3, Centaurin-α 1, TIAP, and RCAS control, as described above. Cells were maintained in soft agar for 4 weeks.