Identification of ELK1 interacting peptide segments in the androgen receptor

Abstract

Prostate cancer (PCa) growth requires tethering of the androgen receptor (AR) to chromatin by the ETS domain transcription factor ELK1 to coactivate critical cell proliferation genes. Disruption of the ELK1-AR complex is a validated potential means of therapeutic intervention in PCa. AR associates with ELK1 by coopting its two ERK docking sites, through the amino-terminal domain (A/B domain) of AR. Using a mammalian two-hybrid assay, we have now functionally mapped amino acids within the peptide segments 358-457 and 514-557 in the A/B domain as required for association with ELK1. The mapping data were validated by GST (glutathione S-transferase)-pulldown and BRET (bioluminescence resonance energy transfer) assays. Comparison of the relative contributions of the interacting motifs/segments in ELK1 and AR to coactivation of ELK1 by AR suggested a parallel mode of binding of AR and ELK1 polypeptides. Growth of PCa cells was partially inhibited by deletion of the upstream segment in AR and nearly fully inhibited by deletion of the downstream segment. Our studies have identified two peptide segments in AR that mediate the functional association of AR with its two docking sites in ELK1. Identification of the ELK1 recognition sites in AR should enable further structural studies of the ELK1-AR interaction and rational design of small molecule drugs to disrupt this interaction.

Keywords: D-box; ELK1; ERK; FxFP motif; androgen receptor; prostate cancer.

Figure 1:. Mapping regions of the A/B domain of AR required for interaction with ELK1 by deletion analysis.

A. Schematic of the structural organization of AR including functional domains and sub-domains. B. HeLa cells harboring the GAL4-TATA-LUC promoter-reporter and expressing a Gal4-ELK1 fusion protein, were transfected with expression plasmids for AR(A/B)-VP16 fusion constructs containing no deletion or consecutive, overlapping deletions within the A/B domain. Cells were harvested 48h post-transfection for preparation of cell lysates to measure luciferase activities that were normalized to protein expression from western blots and plotted. *P < 0.03, **P < 0.02. C. From a parallel set of wells transfected as described for Panel B, cell lysates were probed by western blotting with antibodies against AR or GAPDH (loading control). To ensure detection of all the deletion mutants by western blot, two separate AR antibodies with different epitopes were used. Antibody #1 corresponds to Abcam catalog number ab133273, and antibody #2 corresponds to Abcam catalog number ab9474. D. HeLa cells harboring GAL4-TATA-LUC (but not expressing Gal4-ELK1) were transfected with double fusion constructs of the A/B domain of AR (Gal4-AR[A/B]-VP16) containing no deletion or the same consecutive deletions within the A/B domain used in Panel B. Cells were harvested 48h after transfection and cell lysates prepared to measure luciferase activities that were normalized to protein expression from western blots and plotted. *P < 0.02, **P = 0.006. E. From a parallel set of wells transfected as described for Panel D, cell lysates were probed by western blotting with antibodies against AR or GAPDH (loading control). To ensure detection of all the deletion mutants by western blot, two separate AR antibodies with different epitopes were used. Antibody #1 corresponds to Abcam catalog number ab133273, and antibody #2 corresponds to Abcam catalog number ab9474.

Figure 1:. Mapping regions of the A/B domain of AR required for interaction with ELK1 by deletion analysis.

A. Schematic of the structural organization of AR including functional domains and sub-domains. B. HeLa cells harboring the GAL4-TATA-LUC promoter-reporter and expressing a Gal4-ELK1 fusion protein, were transfected with expression plasmids for AR(A/B)-VP16 fusion constructs containing no deletion or consecutive, overlapping deletions within the A/B domain. Cells were harvested 48h post-transfection for preparation of cell lysates to measure luciferase activities that were normalized to protein expression from western blots and plotted. *P < 0.03, **P < 0.02. C. From a parallel set of wells transfected as described for Panel B, cell lysates were probed by western blotting with antibodies against AR or GAPDH (loading control). To ensure detection of all the deletion mutants by western blot, two separate AR antibodies with different epitopes were used. Antibody #1 corresponds to Abcam catalog number ab133273, and antibody #2 corresponds to Abcam catalog number ab9474. D. HeLa cells harboring GAL4-TATA-LUC (but not expressing Gal4-ELK1) were transfected with double fusion constructs of the A/B domain of AR (Gal4-AR[A/B]-VP16) containing no deletion or the same consecutive deletions within the A/B domain used in Panel B. Cells were harvested 48h after transfection and cell lysates prepared to measure luciferase activities that were normalized to protein expression from western blots and plotted. *P < 0.02, **P = 0.006. E. From a parallel set of wells transfected as described for Panel D, cell lysates were probed by western blotting with antibodies against AR or GAPDH (loading control). To ensure detection of all the deletion mutants by western blot, two separate AR antibodies with different epitopes were used. Antibody #1 corresponds to Abcam catalog number ab133273, and antibody #2 corresponds to Abcam catalog number ab9474.

Figure 2:. Further deletion analysis of the region spanning amino acids 457 to 514 within AR[A/B].

A. HeLa cells harboring the GAL4-TATA-LUC promoter-reporter and stably expressing a Gal4-ELK1 fusion protein, were transfected with expression plasmids for AR(A/B)-VP16 fusion constructs containing no deletion or deletions within the region aa457 to aa514 which lies between the two putative ELK1 interaction segments mapped in Fig. 1. Cells were harvested 48h post-transfection and cell lysates prepared for measurement of luciferase activities that were normalized to protein expression from western blots and plotted. *P < 0.003. B. From a parallel set of wells transfected as described for Panel A, cell lysates were probed by western blotting with antibodies against AR or GAPDH (loading control).

Figure 3:. Validation of putative ELK1 interacting segments using full-length AR and testing the orientation of AR and ELK1 in their binary complex.

A. HeLa cells harboring the GAL4-TATA-LUC promoter-reporter and constitutively expressing a Gal4-ELK1 fusion protein, were transfected with expression plasmids for either wild type or mutant AR containing deletions of the putative ELK1 interacting segments, identified in Figure 1. At the time of transfection, cells were treated with testosterone (10 nM) to allow nuclear translocation of AR or vehicle control. 48h post- transfection, cells were harvested for preparation of lysates that were used for measurement of luciferase activity. *P = 0.001, **P = 0.00002. B. From a parallel set of wells transfected as described for Panel A, cell lysates were probed by western blotting with antibodies against AR or GAPDH (loading control). C. HeLa cells co-transfected with a plasmid containing GAL4-TATA-LUC and expression plasmids for wild type AR. Expression plasmids for either wild type or mutant Gal4-ELK1 fusion proteins containing either a deletion of the D-box segment (Δ308-321) or a deleterious mutation in the FxFP motif in which FQFP was mutated to FQLA were also co-transfected. At the time of transfection, cells were treated with testosterone (10 nM) to allow nuclear translocation of AR or vehicle control. Forty-eight hours after transfection, cell lysates were prepared for measurement of luciferase activity. *P = 0.0005, **P = 0.0003. D. From a parallel set of wells transfected as described for Panel A, cell lysates were probed by western blotting with antibodies against AR or GAPDH (loading control). E. Schematic illustrating parallel mode of binding of AR and ELK1.

Figure 4:. Validation of putative ELK1 interacting segments using orthogonal and in situ assays.

A. GST pulldown assay for binding of an AR peptide fragment containing the ELK1-interacting segments with wild type ELK1. Purified His-tagged ELK1 and a purified GST-tagged AR fragment consisting of amino acids 334-566 were incubated with GST beads for two hours. Following washing, the proteins were probed by western blotting with antibodies against ELK1, and GST. Coomassie staining was used to monitor protein input. B. BRET assay to confirm binding between wild type ELK1 and a mutant AR protein with regions not critical for interaction with ELK1 deleted. HEK293T cells were transfected with different ratios of TurboELK1 fusion and Rluc-AR-V7 truncated variant protein with amino acids 50-357 additionally deleted. Two days post-transfection, cells were replated in 96 well plates. Two hours later, coelenterazine was added and BRET emissions were measured. The 635 nm/535 nm ratios are normalized to the ratios obtained from cells transfected with RLuc construct alone, with representative western blot showing AR and ELK1 protein expression.

Figure 5:. Effect of deleting ELK1 interaction segments on the ability of AR to support colony growth in PCa cells.

22Rv1 cells stably expressing ectopic AR mutants with deletion of either one of the two ELK1 interaction segments or vector control transduced cells were used. Each line was infected with lentivirus carrying either a control (nontargeting) shRNA plasmid or an AR-targeted shRNA plasmid. In Panel A, 72 hours after infection, the cell lysates were probed by western blotting with an antibody to detect both the endogenous AR and the shorter ectopic mutant ARs, using GAPDH as the loading control. The blot western blot shows virtually complete knockdown of the endogenous AR by the AR shRNA with residual ectopic mutant ARs whose levels are comparable to or greater than the original level of endogenous AR seen in the control cells. As the shRNA did not target the endogenous AR-V7 variant in the 22Rv1 cells, both AR-V7 and GAPDH serve as loading controls for each recombinant cell line. In Panel B, the cells treated in parallel as described in Panel A, and 72 hours after infection they were plated for colony formation in conditioned media. After 7-8 days, the colonies were stained and counted. * P < 0.05.

Figure 6:. Amino acid sequences of the two ELK1-interacting peptide segments in AR.

The red font denotes hydrophobic and charged amino acid residues that are known to contribute to a more ordered structure. Stretches of amino acid residues contributing disordered (flexible) structural elements (amino acids G, A, S, T, N, Q and P) are underlined.