Fig 2: Phosphorylation of Thr143 and Thr149 stabilizes the interface between CycT1 and CDK9.(A) Model of P-TEFb where CycT1 is phosphorylated at Thr143 and Thr149. The model was created by adding phosphates to Thr143 and Thr149 in the published crystal structure of P-TEFb (PDB ID 3MI9), followed by energy minimization and molecular dynamics (MD) simulations. Residues predicted to interact with Thr143 and Thr149 are Gln73 in CycT1 and Lys68 in CDK9, respectively. (B) Gln73 is targeted by phosphorylated Thr143 in CycT1. WT CycT1(280) or mutant CycT1(280)Q73A proteins and CDK9 were coexpressed in the presence of bortezomib (+/- signs on top) in 293T cells. Co-IPs with CDK9 are presented in panels 1 and 2. Panels 3 and 4 contain input levels of CycT1 and CDK9 proteins. (C) Lys68 in CDK9 is targeted by phosphorylated Thr149 in CycT1. WT CDK9 or mutant CDK9K68A proteins and CycT1(280) were coexpressed in the presence of bortezomib (+/- signs on top) in 293T cells. Co-IPs with CDK9 are presented in panels 1 and 2. Panels 3 and 4 contain input levels of CycT1 and CDK9 proteins. (D) Mutations of K68A in CDK9 and Q73A in CycT1 attenuate cooperatively the binding between CycT1 and CDK9, equivalently to the mutant CycT1TT143149AA protein. WT CycT1(280) or mutant CycT1(280)Q73A proteins and WT CDK9 or mutant CDK9K68A proteins were coexpressed in the presence of bortezomib (+/- signs on top) in 293T cells. Co-IPs with CDK9 are presented in panels 1 and 2. Panels 3 and 4 contain input levels of CycT1 and CDK9 proteins.

Fig 3: Free CycT1 proteins are unstable but stabilized by proteasomal inhibitors in cells. (A) Levels of endogenous CycT1 proteins in resting CD4 + T cells are increased by proteasomal inhibitors without mRNA induction and cell activation. Resting CD4 + T cells were untreated (lane 1) or treated with 100 nM bortezomib (lane 2) or PMA/PHA (lane 3) for 24 h before cell lysis. Levels of CycT1 (panel 1), CDK9 (panel 2), and the loading control actin (panel 3) proteins were detected with indicated antibodies, respectively, by WB. Relative mRNA levels of CycT1 in bortezomib or PMA/PHA treated cells are presented in the middle bar graph as -fold change above levels of untreated cells (white bars, set as ‘1’). Error bars represent standard error of average (n = 3). Percentage of CD69 + cells in CD4 + T cells with indicated treatment are presented in the bottom table by flow cytometry with PE-conjugated anti-CD69 antibodies. These and subsequent WBs are representative of at least 3 independent experiments. The quantification of all WBs is presented in the Results section as fold-changes. (B) Graphic representation of mutant CycT1 proteins that do not interact with CDK9. Top diagram is a schematic presentation of the WT and mutant CycT1 proteins used in this study. WT CycT1 protein contains 726 residues, in which two cyclin boxes are found the N-terminal region (30–248) and a PEST motif is located at its C-terminus (709–726). The mutant CycT1 protein (CycT14MUT) contains four substitutions, E137D, T143A, T149A and T155A. CycT1(708) and CycT14MUT(708) represent two truncated proteins lacking the PEST motif. WT or indicated mutant CycT1 and CDK9 proteins were expressed in 293T cells treated with bortezomib and co-IPed with CDK9 by anti-Flag antibodies (bottom panels 1 and 2). Bottom panels 3 and 4 indicate input levels of CycT1 and CDK9 proteins. (C) Mutant CycT1 proteins that do not bind to CDK9 are degraded via the proteasome. WT CycT1 or two indicated mutant CycT1 proteins were expressed in 293T cells, which were untreated (lanes 1, 5, and 9) or treated with 2 µM MG132 (lanes 2, 6, and 10), bortezomib (lanes 4, 8, and 12), 10 µM MLN4924 (lanes 3, 7, and 11) before cell lysis. Levels of CycT1 (panel 1), CDK9 (panel 2) and the loading control actin (panel 3) proteins were detected with indicated antibodies, respectively, by WB. (D) C-terminally truncated mutant CycT1 proteins are unstable. WT CycT1 or five indicated mutant CycT1 proteins (CycT14MUT, CycT14MUT(580), CycT14MUT(480), CycT14MUT(380) and CycT14MUT(280)) were expressed in 293T cells in the presence of 2 µM bortezomib (lanes 2, 4, 6, 8, 10 and 12) or not (lanes 1, 3, 5, 7, 9 and 11) before cell lysis. Levels of CycT1 (panel 1) and the loading control actin (panel 2) proteins were detected with indicated antibodies, respectively, by WB. (E) Mutant CycT14MUT(280) proteins are degraded via proteasomal pathways. WT CycT1(280) (left panels) or mutant CycT14MUT(280) (right panels) proteins were expressed in 293T cells, which were untreated (lanes 1) or treated with 2 µM MG132 (lanes 2), bortezomib (lanes 4), 10 µM MLN4924 (lanes 3) before cell lysis. Levels of CycT1 (panels 1) and the loading control actin (panels 2) proteins were detected with indicated antibodies, respectively, by WB.

Fig 4: Siah1 and Siah2 preferentially interact with free CycT1. (A) Siah1 preferentially interacts with free CycT1 protein. DN Siah1C75S proteins were co-expressed with WT CycT1 or mutant CycT14MUT proteins in the presence of bortezomib in 293T cells, followed by co-IPs with CycT1 (anti-HA antibodies) (panels 1 and 2), or with Siah1 (anti-Flag antibodies) (panels 5 and 6). Panels 3 and 4 indicate input levels of CycT1 and Siah1 proteins. (B) Siah1 preferentially interacts with free CycT1(280) protein. DN Siah1C75S proteins were co-expressed with WT CycT1(280) or mutant CycT14MUT(280) proteins in the presence of bortezomib in 293T cells, followed by co-IPs with CycT1 (anti-HA antibodies) (panels 1 and 2), or with Siah1 (anti-Flag antibodies) (panels 5 and 6). Panels 3 and 4 indicate input levels of CycT1 and Siah1 proteins. (C) Siah2 preferentially interacts with free CycT1 protein. WT Siah2 proteins were co-expressed with WT CycT1 or mutant CycT14MUT proteins in the presence of bortezomib in 293T cells, followed by co-IPs with CycT1 (anti-HA antibodies) (panels 1 and 2), or with Siah2 (anti-Flag antibodies) (panels 5 and 6). Panels 3 and 4 indicate input levels of CycT1 and Siah2 proteins. (D) Siah2 preferentially interacts with free CycT1(280) protein. WT Siah2 proteins were co-expressed with WT CycT1(280) or mutant CycT14MUT(280) proteins in the presence of bortezomib in 293T cells, followed by co-IPs with CycT1 (anti-HA antibodies) (panels 1 and 2), or with Siah2 (anti-Flag antibodies) (panels 5 and 6). Panels 3 and 4 indicate input levels of CycT1 and Siah2 proteins.

Fig 5: Knockdown of Siah2 increases levels of free CycT1 in 293T cells and inhibition of Siah1/2 stabilizes CycT1 in primary T cells. (A) Knockdown of Siah2 stabilizes free CycT1 in 293T cells. 293T cells were transfected with scrambled shRNA (lane 1) or mixed shRNA targeting Siah2 (lane 2). 48 h after transfection, WT CycT1 or mutant CycT14MUT proteins were expressed by transient transfection for another 48 h before cell lysis. Levels of Siah2 (panel 1), CycT1 (panels 3 and 5) and the loading control actin (panels 2, 4 and 6) proteins were detected with indicated antibodies, respectively, by WB. (B) Ectopic expression of PHYL(130) stabilizes CycT1 in primary CD4 + T cells. Top diagram presents the experimental scheme. Primary T cells collected at days 3 and 7 (bottom right panels, lanes 1 and 2) were lysed and subjected to WB. Levels of CycT1 (panel 1), Siah1 (panel 2), Siah2 (panel 3) and the loading control actin (panel 4) proteins were detected with indicated antibodies, respectively. Primary T cells without electroporation (right panels, lane 1) and those electroporated with the empty plasmid vector (right panels, lane 2) or plasmids containing PHYL(130) fragments (right panels, lane 3) were collected at day 7, and cell lysates were subjected to WB. Levels of CycT1 (panel 1), PHYL (panel 2) and the loading control actin (panel 3) proteins were detected with indicated antibodies, respectively.